Current status and future prospects for epigenetic psychopharmacology

Abstract

Mounting evidence suggest that epigenetic regulation of brain functions is important in the etiology of psychiatric disorders. These epigenetic regulatory mechanisms, such as DNA methylation and histone acetylation, are influenced by many pharmaceutical compounds including psychiatric drugs. It is therefore of interest to investigate how psychiatric drugs are of influence and what the potential is of new epigenetic drugs for psychiatric disorders. With this targeted review we summarize the current state of knowledge in order to provide insight in this developing field. Several traditional psychiatric drugs have been found to alter the epigenome and in a variety of animal studies, experimental compounds with epigenetic targets have been investigated as potential psychiatric drugs. After discussion of the most relevant epigenetic mechanisms we present the evidence for epigenetic effects for the most relevant classes of drugs.

Keywords: :

• BDNF
• DNA Methylation
• DNMT inhibitors
• epigenetics
• HDAC inhibitors
• Histone Modifications
• neuropharmacology
• psychiatry
• psychotropics
• treatment

Introduction

In the mid 1950s, C.H. Waddington revolutionized the field of developmental biology by merging the theories of genetics and development. Previously developmental genetics and its mechanisms were studied separately but Waddington focused on the interplay between genetic factors and embryological growth and differentiation. He coined the term “epigenetics” to describe this novel area.¹ Today, “epigenetics” is used to describe “the study of stable alterations in gene expression potential that arise during development and cell proliferation.” Epigenetic processes underlie embryonic cellular development and differentiation, but they are also important in mature mammals, either at random or under influence of the environment (for review see²). One area in which this is of special importance is the field of psychiatry as emerging evidence suggest that epigenetic mechanisms influence normal patterns of neurodevelopment and cerebral function (for review see ref. 3), as well as the aberrant neurodevelopment in psychiatric disorders.⁴

The epigenetic machinery is extensive and complex and to date not fully understood. Central to our understanding is that epigenetic mechanisms influences transcription of DNA my modifying access to the DNA sequences. Figure 1 depicts the epigenetic mechanisms that are important for this review. Within the nucleus of a cell, the DNA sequence lies wrapped around histone proteins. The complex of DNA, histones and non-histone proteins, such as scaffold and polycomb proteins, forms a highly condensed structure called chromatin. The basic unit of chromatin is the nucleosome. Each nucleosome is connected to the next by a short segment of linker DNA to which a histone H1 protein binds. The string of nucleosomes is folded into a strongly condensed fiber bringing about a strong compaction. Several details on the folding of the nucleosome string are still elusive, and in particular the higher order structure of chromatin is not fully understood. Within the chromatin structure, the DNA sequence is methylated at certain nucleotides. The methylation pattern of DNA across the genome together with the chromatin structure is referred to as the epigenome. As in other cells, epigenetic mechanisms are essential to the development of the nervous system. The epigenetic machinery drives both embryonic and postnatal neural development. It is involved in neurogenesis,^5,6 neuronal differentiation, cell fate specification⁷ and development of dendrites.⁸ Interestingly, the course of development of the epigenetic profile is influenced by environmental factors in utero. In different species environmental factors such as temperature or the presence of predators, have been shown to affect the phenotype of the offspring. In humans and mice, the physiology of the baby is affected by the nutritional state of the mother⁹ and maternal stress in rats also alters the phenotype of their offspring.¹⁰ Such epigenetic developmental plasticity may involve preparing the offspring for the type of environment in which they are likely to live¹¹ but aberrant epigenetic regulation may also lead to disease. Emerging evidence now suggests that vulnerability to psychiatric disorders such as depression and schizophrenia is modulated by epigenetic processes (for review see refs. 12 and 13).

Figure 1. AUTHOR: Please provide figure legend.

It was previously thought that the epigenetic marks obtained in utero remain the same over time. However, it is now clear that these mechanisms are dynamically regulated. Epigenetic remodeling takes place throughout adult life, under the influence of environmental factors such as nutrition, drugs, and chemical, physical and psychosocial factors (for review see ref. 14). The reversible nature of epigenetic marks and their susceptibility to environmental influences makes them potential targets for therapeutic interventions. Indeed, in the past decade many drugs including known psychotropic drugs were found to exert an influence on the epigenome. Therefore, the aim of this review is to summarize the current state of knowledge regarding drug-induced alterations of the epigenome in order to provide insight in this developing field.

Method and scope

Two major types of epigenetic mechanisms that regulate gene expression in the nervous system have been studied to date; (i). Posttranslational histone modification and (ii) DNA methylation. Other types of epigenetic mechanisms such as non-coding RNA’s that can regulate gene expression, and chromatin remodeling mediated by polycomb proteins (for review see ref. 15) are likely to also be major mediators of neurodevelopmental plasticity. However, the central mechanisms of these latter processes are largely unknown and will not be discussed in this review. We therefore conducted a systematic literature search from 1996 to July 2011 using Medline from with search terms: (antidepressant OR antipsychotic) AND (epigenetic OR DNA methylation OR histone OR chromatin) ; (HDAC inhibitor OR DNMT inhibitor) AND (depressive disorder OR anxiety OR psychosis OR schizophrenia OR bipolar disorder) ; lithium AND (depressive disorder OR anxiety OR psychosis OR schizophrenia OR bipolar disorder) AND (epigenetic OR methylation OR histone OR chromatin). Cross references were included where relevant.

Posttranslational histone modification

Posttranslational histone modifications take place at the N-terminal histone tail of the nucleosome. Figure 1 depicts the most common modifications. These include small covalent including acetylation, methylation and phosphorylation. Less frequent modifications are ubiquitination, sumoylation, ADP ribosylation and deimination. Their effects on transcriptional activity are not yet well understood and are beyond the scope of this review.

Histone acetylation

Histone acetylation occurs most frequently on the lysine residues at H3 and H4 of the NH₂-terminal, though it can also take place at other histones and in the globular domains. This dynamic process is controlled by specific enzymes that either add or remove the acetyl group. Histone acetyl transferases (HATs) catalyze the addition of acetyl groups. Over a dozen HATs have been identified, of which some can also acetylate non-histone proteins such as transcription factors. Several transcription factors even contain intrinsic HAT activity to activate gene expression. Histone deacetylases (HDACs) catalyze the removal of acetyl groups from histone proteins, as well as non-histone proteins among which p53, Sp1 and CREB (for review see ref. 16). There are four classes of HDACs, of which class I and II are the most relevant to this review, as the first is expressed throughout brain and body, and the second primarily in the brain, heart and muscle. Histone hyperacetylation is associated with decondensation of chromatin and an increase in gene expression, whereas hypoacetylation correlates with repression of chromatin and a decrease in gene expression. The balance between the opposing activity is important for regulating transcription and disbalance can lead to neurodegenerative diseases (for review see ref. 17).

Histone methylation

Histone methylation occurs on lysine residues of the histone tail. This modification can exist in a mono-, di-, or trimethylated state, each with a different effect on transcriptional activity through distinct coregulators. As with acetylation, methylation is regulated by specific enzymes that can exert effects on histone as well as on non-histone proteins. Histone methyltransferases (HMTs) add methyl groups to lysine residues and histone demethylases (HMDs) remove these methyl groups. There are distinct HMTs and HDMs for various lysine residues, each with specific abilities to catalyze mono-, di-, or trimethylated states (for reviews see^18,19). Interestingly, methylation of different lysine residues can achieve opposite effects on gene expression; it can cause both repression and activation depending on which lysine residue of the histone tail is methylated. In psychiatric epigenetics, methylation of histone H3 has gained attention. Trimethylation of histone H3 at lysine 4 (H3K4) is important for transcriptional activation, while the same methylation state is repressive at H3K9 and H3K27, and dimethylation of H3K9 and H3K27 are repressive as well.²⁰

Overall the methylation process gives rise to many unique possibilities to influence transcriptional activity, by combining a variety of enzymes with different lysine residues in different states.

Histone phosphorylation

Although histone phosphorylation is somewhat less well understood, several nuclear protein kinases and protein phosphatases are known that add or remove phosphate groups from the histone tail. The protein kinase MSK 1 and the protein phosphatase inhibitor DARRP‑32 have been shown to regulate phosphorylation in the brain. Phosphorylation of histones is associated with the promotion of transcriptional activity. It is the best characterized at serine 10 on histone H3, where it recruits a HAT to halt repressive methylation on lysine 9. This HAT in turn acetylates the lysine residue in a process called phosphoacetylation, which further increases transcriptional activity (for review see ref. 21).

DNA methylation

In cytosine or DNA methylation, a cytosine nucleotide is methylated by transfer of a methyl group from S-adenosyl methionine (SAM), resulting in 5-methyl-cytosine. It is catalyzed by DNA methyltransferases (DNMT) and occurs in approximately 3% of cytosines in human DNA. Recently a new, less frequent, form of DNA methylation has been discovered; 5-methylcytosine²² that has the highest concentration in neuronal cells²³ suggesting a role in neuronal function. ”Traditional” DNA methylation primarily takes place at the dinucleotide sequence CpG. The CpG sequence occurs in low frequency throughout the genome and in high frequency in so-called CpG islands (CGIs). CGIs occur in promoter, intragenic, 3′utr and intergenic regions.²⁴ DNA methylation is associated with transcriptional repression, is enhanced by methyl-binding proteins and also silences genes via blocking binding of certain transcription factors ^25;26. The cooperation between methyl-binding proteins and chromatin-remodeling enzymes illustrates that different epigenetic mechanisms act in synchrony to influence genetic transcription.

Psychopharmacological compounds with putative epigenetic effects

The recent interest for the role of epigenetics in the brain has led researchers to explore the possibility that drugs can modify epigenetic processes involved in psychiatric disorders. So far, several psychiatric drugs have been found to influence epigenetic mechanisms, suggesting that this may partially explain their mechanism of action. Moreover, the therapeutic potential of experimental compounds that specifically target epigenetic mechanisms is extensively investigated. We aim to give a broad overview of the current knowledge of drugs that modify epigenetic changes in psychiatric disorders.

Antidepressants

Imipramine

Imipramine is a tricyclic antidepressant that has been used since the 1950s. Its primary mechanism of action is to inhibit the reuptake of serotonin and norepinephrine, thus elevating the levels of these neurotransmitters in the brain. Recently, epigenetic effects of imipramine have been uncovered, providing a deeper understanding into the therapeutic effect of this drug. Since dysfunction of brain-derived neurotrophic factor BDNF is often implicated in the etiology of depression, Tsankova et al.²⁷ investigated the effectiveness of imipramine on epigenetic regulation of the Bdnf gene in the hippocampi of mice. Chronic social defeat led to repression of Bdnf and an increase in histone dimethylation of H3K27. Chronic treatment with imipramine could not reverse this, even though the depression-like behaviors disappeared. However chronic imipramine did lead to hyperacetylation of H3 at the Bdnf promoters, mediated by downregulation of Hdac5. Moreover, the efficacy of imipramine was blocked by overexpression of Hdac5, suggesting that downregulation of Hdac5 is essential to the efficacy of imipramine. The authors suggest that since histone H3 hypermethylation was not affected by imipramine, this remains a possible target for antidepressant therapy. Imipramine did however largely reverse repression of dimethylation of H3K9 and H3K27 and increased phospho-CREB binding in the nucleus accumbens induced by social defeat stress.²⁸ The global pattern of H3 dimethylation in the nucleus accumbens of mice which did not develop depression-like behaviors strongly resembled the dimethylation profile of mice that received chronic imipramine treatment after social defeat. This indicates that resilient animals somehow naturally overcome the effects of stress on the accumbens. The genes that show dimethylation in resilient mice, but not in mice treated with imipramine, may therefore provide novel targets for antidepressant medication.²⁸

Amitriptyline

Amitriptyline is a tricyclic antidepressant. In rat astrocytes, amitriptyline induced partially reversible global DNA hypomethylation without affecting histone acetylation and reduced enzymatic activity of DNA (cytosine-5)-methyltransferase 1 (DNMT1) without altering DNMT1 protein levels.²⁹ In vitro amitriptyline did increase H3 acetylation by inhibiting HDAC activity.³⁰

Fluoxetine

Fluoxetine is a widely used antidepressant, belonging to the class of selective serotonin reuptake inhibitors (SSRI). In the hippocampus, fluoxetine reversed decreased histone H3K9 trimethylation but not H3K4 trimethylation induced by chronic restraint stress.³¹ However postnatal fluoxetine exposure did not changes the acetylation or trimethylation of histone H3 at the BDNF promoters.³² Chronic fluoxetine treatment of healthy rats decreases acetylation of H3 in three serotonin projection areas; the caudate putamen, the frontal cortex and the dentate gyrus of the hippocampus. In addition, expression of the methyl-binding proteins MeCP2 and MBD1 was increased, accompanied by increased Hdac2 expression, further inhibiting transcriptional activity in these brain regions.³³ In mice that have been subjected to social defeat stress, reduced H3 acetylation in the hippocampus was reversed by fluoxetine in the same way as histone deacetylases (HDAC) inhibitors, suggesting a similar mechanism of action.³⁴

Escitalopram

The SSRI escitalopram was able to reverse hypermethylation at the P11 (S100A10) locus in a genetic rodent model of depression. Expression of P11 is associated with depression suggesting a putative epigenetic mechanism of escitalopram efficacy.³⁵

MAO inhibitors

Monoamine oxidase (MAO) inhibitors form a group of highly effective antidepressants. They act by inhibiting the MAO enzymes, thereby preventing the breakdown of the monoaminergic neurotransmitters serotonin, dopamine, epinephrine and norepinephrine. The MAO inhibitors phenelzine and tranylcypramine both inhibit demethylation of histone H3K4, resulting in a global increase in H3K4 methylation by breaking down lysine-specific demethylase 1 (LSD1), a histone demethylase that is structurally similar to MAO A and B.^36,37 LSD1 specifically demethylates mono- and dimethylated H3K4 and H3K9, thus inhibiting transcription.³⁷ Interestingly, tranylcypromine was ten times more effective in inhibiting LSD1 than MAO A or B, suggesting that also other selective LSD1 inhibitors might be effective in treating depression.³⁸ Other experimental MAO inhibitors also increase H3K4 methylation,³⁶ but the effects of such drugs on H3K9 methylation have not been investigated to date.

Antipsychotics

Haloperidol

Haloperidol is a widely used typical antipsychotic that preferentially antagonizes dopamine D2 receptor activity. In addition, some evidence indicates that haloperidol induces histone modifications to achieve its therapeutic effects. In the striatum of mice infusion of haloperidol led to phosphorylation of histone H3 at serine 10, acetylation of H3K14 as well as phospho-acetylation. These processes were mediated by the NMDA receptor, since administration of an NMDA receptor antagonist blocked phosphoacetylation.³⁹ Similar results on phosphorylation were seen following administration of the D2 receptor antagonist raclopride.⁴⁰ Two studies also reported an increase in mouse striatal H3 phosphorylation in response to haloperidol, although acetylation levels remained unchanged.^40,41 One study reported that haloperidol decreased global DNA methylation levels in the brain of female rats.⁴² However, haloperidol did not influence methylation of the reelin and GAD67 gene promoters in the frontal cortex or striatum.⁴³

Clozapine

The atypical antipsychotic clozapine binds dopamine D1, D2, and 5HT2 receptors. In the rat cerebral cortex it increased trimethylation of H3K4 by 3-fold and accompanied by increased expression of the H3K4-specific histone methyltransferase gene Mll1. Mll1 occupancy at the GAD67 promoter had doubled.⁴⁴ In the prefrontal cortex of humans who had been treated with clozapine, H3K4 trimethylation at the GAD67 gene had increased by 2-fold.⁴⁴ GAD is an enzyme that catalyzes the decarboxylation of glutamate to GABA and CO_2. As the epigenetic modification at the GAD64 locus is associated with transcriptional activation, these findings suggest that the therapeutic action of clozapine relies at least in part on increased GABAergic activity mediated by histone methylation.⁴⁵ Further support for this hypothesis comes from the finding that methionine-induced reelin promoter hypermethylation was strongly decreased after administration of clozapine, together with an increase in promoter-associated H3K9 and H3K14 acetylation.⁴³ Clozapine reduced cytosine methylation at the Gadd45-β locus similar to the effects of brain-permeant mGlu2/3 receptor agonist LY379268 on hypermethylation.⁴⁶

Sulpiride and Amisulpiride

Two bezamine derivates with antipsychotic properties, sulpiride and amisulpiride, act by antagonizing the dopamine D2 receptor. A third benzamine derivate, MS-275 acts as HDAC inhibitor and is highly effective in increasing acetylation of H3 associated with Reelin and Gad67 gene promoters in the frontal cortices of mice. These drugs have the same effect in the hippocampus and the striatum, but much higher doses are needed, indicating there is different sensitivity to these drugs across different regions of the brain.⁴⁷ Indeed H3K9 and H3K14 acetylation was increased in the promoter of the Reelin gene after administration of sulpiride in the frontal cortex and striatum of mice that displayed methionine-induced hypermethylation of the reelin promoter, In addition sulpiride decrease methylation at this promotor.⁴³

Risperidone

Risperidone is another atypical antipsychotic that antagonizes D2 receptors, and binds to 5HT2 receptors. It induces global phospho-acetylation of H3 in the striatum, mediated by the NMDA receptor, suggesting that dopamine, serotonin and glutamate act in synchrony to influence chromatin regulation.⁴⁸

Mood stabilizers

Valproate

Valproic acid (VPA) is a short fatty-chain acid commonly used in the treatment of epilepsy and bipolar disorder. It increases GABAergic activity by inhibiting GABA transaminase, an enzyme that is involved in the synthesis and degradation of GABA. Moreover, VPA is the most extensively investigated compound in psychiatric epigenetics and a potent HDAC inhibitor of class I and II. It increases levels of acetylated histone H3 and H4 thus promoting gene expression. Much effort has been dedicated to elucidating the effects of VPA on neuronal processes in the brain. The therapeutic actions of VPA may be mediated through increase of tyrosine hydroxylase (TH) of which VPA increases mRNA and protein levels.⁴⁹ In term TH is involved in the biosynthesis of the catecholamines dopamine, norepinephrine and epinephrine, neurotransmitters long implicated in the mechanisms underlying psychiatric disorders. When VPA is administered to rat cortical neurons, mRNA levels and protein levels of exon 1–9-containing Bdnf are elevated, and in astrocytes, glial-derived neurotrophic factor is increased^50-52 accompanied by increased H3 and H4 acetylation.^50-52 In a study on rat glioma cells, VPA failed to induce any change in levels of Bdnf mRNA per se, but it enhanced the stimulatory effect of serotonin on Bdnf gene expression.⁵³ Changes in histone H3 and H4 acetylation were found after four weeks of VPA treatment in bipolar and schizophrenia patients⁵⁴ and increased glutamic acid decarboxylases (GAD) (involved in GABA synthesis) in response to HDAC inhibitors was found neuronal progenitor cells and lymphoblast of healthy humans.^55,56 Increased histone acetylation was associated with VPA-induced hippocampal neurogenesis in hippocampal extracts.⁵⁷ Further, in the frontal cortex of schizophrenic and bipolar disorder patients with psychosis, DNMT1 mRNA and protein expression was preferentially increased in layer I, II, and IV interneurons when patients received antipsychotic monotherapy, but not when they were treated with a combination of VPA and antipsychotics⁵⁸ and only in schizophrenia not bipolar disorder.⁵⁹ The potential role of epigenetic modification of GABA regulation is recently reviewed by Guidotti.⁴⁵

Another gene that is targeted by VPA is RELN which has been studied as candidate gene for bipolar disorder and schizophrenia. In neural progenitor cells VPA increased acetylation of H3 and H4 and induced expression of reelin.^60,61 This is most likely mediated by demethylation of the reelin promoter region, as treatment with VPA has been shown to stimulate promoter demethylation at the reelin and Gad67 genes.^61,62 Additionally, Dnmt1 and 3A and B protein levels were downregulated in response to VPA further increasing gene expression.^55,62 When VPA was administered together with methionine, it blocked the hypermethylation that is normally induced by methionine alone.^63,64 In addition, VPA attenuated behavioral effects that were induced by methionine. It was suggested that this was mediated by VPA induced hyperacetylation of histone H3 at the reelin and GAD67 promoters, which facilitated a state of euchromatin and enhanced DNA demethylase activity. In mice that exhibit methionine-induced hypermethylation, VPA induced demethylation of the reelin promoter in the frontal cortex, and increased acetylation of H3K9 and H3K14 associated with the promoter. When VPA was administered in conjunction with clozapine or sulpiride, both demethylation and acetylation were further enhanced.⁴³

In an extensive study, Perisic and coworkers examined epigenetic effects of a variety of psychoactive compounds including VPA in hippocampal and cortical astrocytes. VPA was the only drug to exert effects on histone acetylation, inducing dose-dependent global hyperacetylation of histones H3 and H4, and decreased levels of inhibitory H3K9 dimethylation. Furthermore, the increased acetylation by VPA caused a 4-fold increase in glutamate transporter GLT-1 mRNA and reduced DNA methylation of the GLT‑1 gene, which supports the hypotheses of glutamate hypofunction in bipolar disorder. In addition, VPA induced global demethylation of CpG sites in astrocytes of the cortex and hippocampus. The effects on acetylation were transient, since they peaked during the first hour of treatment and then declined. Fourty-eight hours after VPA treatment acetylation levels had returned to baseline, suggesting that the effects of VPA are fully reversible.²⁹ In glioma cells VPA increased expression of HDAC 1, 2 and 3, and of MeCP2, suggesting VPA affects both histone acetylation and DNA methylation although the upregulation of HDACs and MeCP2 may constitute a compensatory mechanism to VPA-induced hyperacetylaton.⁶⁵ In these cells VPA increased melatonin MT1 receptor expression providing a possible mechanism of action of VPA in bipolar disorder. A final clue to an epigenetic mechanism of action is the finding that in rats, VPA heightened levels of histone H4 acetylation in the striatum, associated with the fosB promoter gene. Striatal fosB protein levels were also elevated, and since fosB is a ubiquitous transcription factor this indicates that VPA can enhance gene expression in the striatum.⁶⁶

HDAC inhibitors

A well-balanced regulation of HDACs and HATs is essential to gene transcription. Compounds that inhibit HDACs have been extensively studied in models of neurodegenerative disorders. HDAC inhibitors differ in their specificity for the separate classes of HDACs, although most target class I and II. HDAC inhibitors can be divided into categories based on their chemical structure and the most important are the hydroxamates, the short fatty-chain acids and the benzamides. Many more HDAC inhibitors exist that have been primarily investigated in the treatment of cancer, such as SAHA, trapoxin and LBH589 (for review see refs. 16 and 67). Unfortunately, HDAC properties of psychiatric drugs are largely unknown with the exception of valproate, imipramine and amitriptyline, but other HDAC inhibitors do show some potential.

Sodium butyrate

Sodium butyrate (SB) is a short fatty-chain acid that is used in many experimental paradigms to inhibit histone deacetylase. It selectively inhibits HDAC class I and II, and it is applied in a range of studies to elucidate the involvement of HDACs in development and disease (for review see ref. 68). Similar to VPA, SB upregulates BDNF and GDNF mRNA levels in astrocytes, with marked increases in GDNF promoter activity and promoter-associated histone H3 acetylation.⁵¹ As BDNF is involved in memory formation, SB might have beneficial effects on this process, as indeed was shown.⁶⁹ Treatment with SB also enhanced long-term memory for object recognition in mice, and improved memory formation in a fear conditioning task. Following fear conditioning, mice had elevated trimethylation levels of H3K4 and dimethylation of H3K9. Treatment with SB further elevated trimethylation while it diminished dimethylation, enhancing gene expression. This shows that while SB primarily is an HDAC inhibitor, it also affects histone methylation.⁷⁰ In a contextual fear model, fear extinction was highly accelerated by SB therapy, with fear decreasing eight times as fast as in the control condition, providing further support for learning and memory enhancement of SB.⁷¹ More inconsistent results have been reported on the effectiveness of SB on depression-like behavior in animals. Despite increased histone H3 and H4 acetylation in the hippocampus and the frontal cortex following SB administration SB only improved depression-like behavior in three of seven of depression-like behavior.^72,73 Further, SB induced a dose-dependent global increase in acetylation of histone H3 and H4.²⁹ In the central nucleus SB induces acetylation and phosphoacetylation of histone H3K14, accompanied by increased c-Fos expression.⁷⁴ Similar to VPA, SB increased TH mRNA and protein levels⁷⁵ suggesting a role in catecholaminergic pathways and bearing promise for a potential therapeutic role in psychiatric disorders.

Trichostatin A

Originally an antifungal antibiotic, two decades ago Trichostatin A (TSA) was discovered to possess HDAC-inhibiting properties. It belongs to the hydroxymates, targeting both class I and class II HDACs. In the brain, tissue specific effects have been reported for TSA on BDNF expression. In the hippocampus TSA induces transcription of promoter exon 1 but not of exon 4 of the BDNF gene, associated with hyperacetylation at H3K9 and H3K14 and increased BDNF protein levels. At the same time an increase in HDAC mRNA and protein levels was also observed, suggesting a compensatory mechanism in response to HDAC inhibition.⁷⁶ In astrocytes TSA upregulates both BDNF and GDNF mRNA levels, with marked increases in GDNF promoter activity and promoter-associated histone H3 acetylation.⁵¹ However, in glioma cells TSA does not induce any change in the basal levels of BDNF mRNA although TSA did strengthen the positive influence of serotonin on BDNF transcription in these cells,⁵³ suggesting that TSA may enhance BDNF expression in glia through serotonin-mediated pathways.

TSA also induced transcription of depression associated genes: the melatonin MT1 receptor gene in glioma cells⁶⁵ and the Glucocorticoid Receptor (GR) gene.⁷⁷ In a study on the influence of maternal care on GR expression, Weaver and colleagues suggested that poor mothering lowers GR mRNA levels through methylation of the NGFI-A binding site. This interferes with NGFI-A binding to the GR exon 1 promoter, inhibiting expression.⁷⁷ In adulthood, TSA reversed the epigenetic effects of poor maternal care. It increased hippocampal GR expression by increasing H3K9 acetylation, DNA demethylation and binding of NGFI-A to the GR exon 1 promoter, to levels comparable to rats that received normal fostering and reversed the behavioral effects of poor fostering.⁷⁸ They found that over 900 genes were regulated by maternal fostering behavior. Of these, less than 2% were affected by TSA treatment but included genes that have been implicated in learning and memory, X-linked mental retardation and neurodegenerative disorders. TSA administration also accelerated fear extinction by 8-fold in a contextual fear paradigm.⁷¹

Alteration of transcription of glutamate related genes may provide a functional link to psychosis. In astrocytes, TSA decreased DNA methylation of the GLT-1 gene leading to a 4-fold increase in GLT-1 mRNA.²⁹ Similar to VPA, TSA increases mRNA levels of the reelin and GAD67 genes in neural precursor and neural blastoma cells. By increasing acetylation of H3 and H4 and decreasing expression and enzymatic activity of DNMT1 and 3A and B, the reelin promoter is demethylated and more available for transcription.^55,60,61 The influence of TSA was also investigated in human lymphocytes. It induced an increase of nearly 4-fold in GAD67 mRNA levels, an almost 2-fold increase in acetylation of H3K9 and K14, and a 5-fold rise in attachment of acetylated H3K9 and K14 to the GAD67 promoter.⁵⁶

MS-275

MS-275 is a benzamide-based HDAC inhibitor that selectively targets class I HDACs, but it is not as popular in scientific research as the previous compounds. In mouse models for depression, infusion of MS-275 into the nucleus accumbens delivered strong antidepressant-like effects. In these mice, chronic defeat stress led to a unique global pattern of gene transcription in the nucleus accumbens which could be reversed by MS-275 treatment. The effects of MS-275 on gene expression were compared with the effects of fluoxetine, in search for novel antidepressant targets. A number of genes were only regulated by MS-275, including genes involved in gap junction formation and in adrenergic receptor function.⁷⁹ In addition, MS-275 increased mRNA levels of reelin and GAD67 in neural precursor and neural blastoma cells. By decreasing expression and enzymatic activity of DNMTs, the reelin promoter is demethylated and more available for transcription.⁵⁵ Simonini and colleagues⁴⁷ reported that MS-275 increased acetylation of histone H3 interacting with reelin and Gad67 gene promoters preferentially in the frontal cortices of mice. However, much higher dose of MS-275 were needed to have an effect in the hippocampus and the striatum, indicating different sensitivity to MS-275 across different regions of the brain.

Treatment with MS-275 stimulated promoter demethylation and histone H3 acetylation in reelin and GAD67 in the cortex and hippocampus of mice⁶² and reversed defeat induced behavior.⁷⁹ In neural progenitor cells MS-275 dose-dependently decreased methylation at the reelin and GAD67 promoters as well, and it upregulated expression of reelin and GAD67, correlating with the extent of inhibition of HDAC activity.⁵⁵ MS-275 also downregulated DNMT1, DNMT3A and DNMT3B protein levels. These DNMTs, together with MeCP2 and HDAC1 appear to constitute gene-specific repressor complexes at the reelin and GAD67 promoters, as activation of these genes by MS-275 is associated with dissociation of these proteins from the promoter regions. In addition, MS-275 decreased expression of MeCP2 and HDAC1. The synchronic involvement of DNMTs, MeCP2 and HDAC1 again show that DNA methylation and histone acetylation are intertwined mechanisms, controlling gene expression in synchrony.⁵⁵

DNMT inhibitors

DNMTs inhibitors are predominantly used to investigate the functional characteristics of DNMTs butrecent studies suggest that DNMT activity is important in social behavior of mice.⁸⁰ The compounds 5-aza-dC (AZA), zebularine and doxorubicin inhibit DNMT1 and DNMT3 and decrease DNA methylation of the reelin promoter in neural progenitor cells. This dramatically increased reelin and GAD67 mRNA levels, showing that the expression of the reelin and GAD67 genes was mediated by DNMTs. In addition, these DNMT inhibitors reduced DNMT enzymatic activity, decreased DNMT1 protein levels and increased H3 acetylation in the promoter area.^60,61,81 Treating hippocampal neurons with AZA leads to increased BDNF exon 4 mRNA, indicating that this splice variant is methylated by DNMT1 or DNMT3.⁸² These effects were replicated using systemic administration of AZA leading to dose-dependent antidepressant-like effect, followed by decreased DNA methylation and increased BDNF levels, in the hippocampus.⁸³

Limitations and recommendations

Current epigenetic research in psychiatry is primarily focused on a subset of mechanisms. In most studies, histone acetylation, methylation, phosphorylation and DNA methylation are investigated, as these mechanisms are best characterized and measurable with techniques such as chromatin immunoprecipitation (ChIP). Moreover these epigenetic mechanisms are generally studied in isolation despite mounting evidence that neither histone modifications nor DNA methylation act independently. It would therefore be desirable to measure these different levels of epigenetic modification simultaneously, to get better insight into relevant epigenetic interactions. The current state of evidence is further obscured by the fact that the effects several epigenetic mechanisms such as ubiquitination, sumoylation, RNA- and polycomb-based mechanisms are not fully understood. Despite that they have not been well characterized in the brain they are likely to be important in epigenetic regulation as well. Future understanding these mechanisms would improve epigenetic psychiatric research.

Conclusion

Table 1 gives an overview of the epigenetic mechanisms influenced by know and potential future psychopharmacological drug. The most prominent (but not all) epigenetic influences of psychiatric drugs are depicted in Figure 1. Compelling evidence from animal and human research indicate that some antidepressants most notably imipramine as well as some antipsychotics particularly clozapine and the mood stabilizer valproic acid alter epigenetic regulation of a variety of genes. In addition new compounds such as sodium butyrate and trichostatin A show some promise for future development of epigenetic psychopharmacology. Considering that the first converging evidence of behavioral and pharmacological efficacy of epigenetic compounds is emerging, the expectation is that further combined behavioral and pharmacological studies can pioneer this promising field. Although our understanding of the complex epigenetic regulatory mechanism is in the early stages, the preliminary evidence as presented here suggest we could be standing at the dawn of the development of epigenetic therapy for psychiatric disorders.

Table 1. Known epigenetic mechanisms of current and potential psychotropic drugs

Drug	Histone acetylation	Histone monomethylation	Histone dimethylation	Histone trimethylation	Histone phosphorylation	Histone phosphoacetylation	DNA methylation
Imipramine	27		28
Amitriptyline	30						29
Fluoxetine	33,34			31
Escitalopram							35
MAO inhibitors		36,37	36,37
Haloperidol	39				39–41	39	42
Clozapine	43			44			43,46
(Ami)sulpiride	43,47						43
Risperidone						48
Valproic acid	29,43,49,50,51,56,57,60,61,66		29				29,43,61–64
Sodium butyrate	29,51,72–74		70	70		74
Trichostatin A	51,55,56,60,61,76,77						29,55,60,61
MS-275	47,55,62						55,62
DNMT inhibitors	60,61,80,81						83

Acknowledgments

We thank Ingrid Janssen for drawing the figure. M.B. is supported by a NARSAD young investigator grant. R.A.O. gratefully acknowledges support of NIMH grant RO1 MG078075.