INTRODUCTION

Epigenetic regulation, including DNA methylation, is an essential process for normal organismal activity [1]. DNA methylation directly regulates gene expression, the function of regulatory elements and chromatin stability, also playing an important role in cell differentiation and tissue specialization during embryogenesis, including in the central nervous system (CNS, Fig. 1). DNA methylation in eukaryotes is implemented by a covalent transfer of a methyl group to the C-5 position of cytosine (C) in those DNA sites where it is followed by guanine (G), the so-called CpG sites, which occur with high frequency in genomic regions (CpG islands, CGIs) localized in the gene promoter region [1]. DNA methylation disorders, both its hyper- and hypomethylation, are associated with various CNS diseases, including autism, schizophrenia, bipolar disorder, and depression [1] (Table 1). However, the molecular mechanisms of DNA methylation and its role in the pathogenesis of these disorders remain poorly understood. Experimental animal models provide an important tool to explore pathological processes at different systemic levels, specifically, using highly translational model organisms, primarily rodents and primates.

Along with laboratory mammals, the zebrafish (Danio rerio), a small freshwater teleost fish, has gained popularity in recent years as an effective translational model system [2]. The wide application of zebrafish in neurobiological studies owes a handful of their biological features, including a close similarity of their CNS physiology to that of humans [2], rapid development (outside the maternal organism), high fecundity, transparency of their embryos, mammalian-like responses to pharmacological agents and ease of their administration to fish by immersion [3, 4]. Zebrafish are highly amenable to genetic manipulations, allowing creation of mutations that cause defects in various organ systems, recapitulating human pathologies [5]. To date, the zebrafish genome has been fully mapped and sequenced [6], revealing quite a high (~70%) genetic homology with humans [7]. More than 80% of human disease-related genes have at least one ortholog in zebrafish, 69% of which fall on linearly homologous genes [8], while among the orthologous genes, 47% have an unambiguous linkage to a zebrafish ortholog [5]. The presence of orthologs associated with human diseases provides the possibility to investigate key expression pathways of these genes and their downstream targets [9]. The constellation of these and other assets enables zebrafish to be used as an effective model for studying the role of DNA methylation in CNS functioning.

Fish genetic studies are somewhat complicated by an additional cycle of whole genome duplication compared to the rest of fishes [5]. As a result, zebrafish has 26206 coding genes, which is more than in any other vertebrate sequenced previously. The zebrafish genome contains more species-specific genes than in humans and mouse, which is also a consequence of duplication [10], while the duplicated genes encode potentially more diverse and zebrafish-specific proteins [11]. On the other hand, carrying out genetic modifications in these fish is by far easier and simpler than in rodents [12].

Modern genetics has got enough tools for editing the zebrafish genome, of which zinc finger nucleases (ZFNs), transcription activator effector-like nucleases (TALENs), and clustered regularly interspaced short palindromic repeats (CRISPR)/CRISPR-associated protein 9 (Cas9), which allow inducing targeted mutagenesis, are of particular interest [13]. Currently, the application of such technologies has enabled engineering zebrafish genetic models with mutations characteristic of various CNS diseases. They include the models of Parkinson’s disease created by PINK1 gene knockout, peripheral neuropathy and dominant optic atrophy created by MFN2 and KIF5A gene knockouts, Leigh syndrome with a mutation in the COA6 gene, Dravet syndrome with a mutation in the SCN1A gene, Kosteff syndrome with a mutation in the OPA3 gene, and congenital myasthenic syndrome or severe combined DL-2 hydroxyglutaric aciduria with a mutation in the SLC25A1 gene [14]. Thus, zebrafish-based genetic models have been created and successfully used to study the molecular pathogenetic mechanisms of various human brain diseases, develop novel therapeutic approaches, and run preclinical trials.

Fig. 1.
figure 1

Common control mechanisms of gene transcription via DNA methylation. The upper two diagrams show the pathways of gene silencing by blocking the interaction of transcription factors with the promoter or enhancer regions, respectively. Promoter methylation recruits an activating protein (AP), which prevents transcription factors from binding to the promoter sequence. The bottom diagram exemplifies the interaction of methyl-CpG-binding protein (MBP) with a region enriched with methylated CpG sites. After MBP binding to DNA, it recruits histone deacetylase (HDA) to trigger histone deacetylation, leading to chromatin condensation at the given site and promoter closing for binding transcription factors, hence to blocking gene transcription.

Full size image
Table 1. Selected human brain disorders associated with aberrant genome methylation
Full size table

DNA METHYLATION AS A MECHANISM OF EPIGENETIC GENOME REGULATION

The role of epigenetic mechanisms in multicellular organisms is reduced to modulating gene expression profiles in cells and tissues [15]. The basic epigenetic processes include DNA methylation and demethylation, histone acetylation and deacetylation, as well as phosphorylation and dephosphorylation of transcription factors [16]. In contrast to acetylation (which facilitates the process of transcription), DNA methylation can both induce and inhibit gene transcription depending on the number of methyl groups attached (Figs. 1, 2). For example, the addition of one group attenuates electrostatic attraction between DNA and histones and thus enhances transcription, whereas the addition of 2–3 methyl groups entails DNA site blockade and gene repression [17].

The specific enzymes that catalyze, recognize, and remove DNA methylation are subdivided into several groups (Table 2). Writer enzymes catalyze the addition of methyl groups to cytosine residues, eraser enzymes modify or remove methyl groups, while reader enzymes recognize methyl groups and bind to them, thus influencing gene expression [18]. DNA methylation in eukaryotes is catalyzed by a family of DNA methyltransferases (Dnmts), which are writer enzymes. Three of them (Dnmt1, Dnmt3a, and Dnmt3b) are active during embryogenesis [17]. Dnmt3a and Dnmt3b, known as de novo Dnmt, can establish a novel methylation pattern of unmodified DNA. In vivo experiments have shown the crucial role of Dnmt3b at early developmental stages and Dnmt3a for normal cell differentiation [17] (Figs. 1, 2).

Dnmt1 is one of the most well-studied enzymes of this family. It copies the DNA methylation pattern from the parental DNA strand onto the newly synthesized daughter strand, thus participating in DNA replication, and controls cell differentiation and division during embryogenesis [19]. Dnmt3L is expressed early in development and does not perform a catalytic function on its own, but binds to Dnmt3a and Dnmt3b, stimulating their methyltransferase activity and altering the global RNA expression of genes involved in neuronal differentiation [20, 21]. Dnmt3L synthesis is suppressed during the latter process and is not observed in the postnatal human brain [22]. Dnmt expression diminishes by the time of terminal differentiation of most cells in embryogenesis, however, in the adult mammalian brain neurons continue expressing Dnmts, suggesting the importance of DNA methylation for brain performance throughout ontogenesis [23].

The class of reader enzymes includes MBDs (Table 1), ubiquitin-like proteins UHRF (see below), and proteins containing a Cys2His2 (C2H2)-like zinc finger motif. All of them share a high affinity for 5-methylcytosine and inhibit the binding of transcription factors [24]. MBD family proteins contain a conserved methyl-CpG-binding domain that provides a high affinity for single methylated CpG sites. Proteins in this family are expressed in the mammalian brain to a greater extent than in any other tissue and are required for normal neuronal development and functioning. This family includes MeCP2, MBD1, MBD2, MBD3, and MBD4. Besides their leading role in transcriptional repression, they are also involved in maintaining DNA methylation [24]. Ubiquitin-like proteins are represented by UHRF1 and UHRF2, multidomain proteins that bind to half-methylated DNA during S-phase and recruit DNA methyltransferase DNMT1 to regulate chromatin structure and gene expression [25]. The DNA methylation structure in the genome undergoes dynamic changes throughout embryonic development due to both its de novo methylation and demethylation.

Differentiated cells produce a stable and unique DNA methylation pattern that regulates tissue-specific gene transcription. At the same time, neuronal activity can modulate this pattern in response to physiological stimuli and environmental exposures and can be inherited [26]. For example, the differences in DNA methylation patterns (described in Mongolians traditionally leading a nomadic lifestyle) include hypomethylation of the promoter regions of the PM20D1 gene, which encodes a mitochondrial fission protein and is involved in cellular anti-ROS defense and thermoregulation [27].Mounting evidence also supports the important role of DNA methylation in brain functions. For example, in patients with Parkinson’s disease, the development of depressive symptoms is accompanied by changes in DNA methylation in nigral neurons [28]. However, there is a correlation between the improvement of some human cognitive skills during meditation and DNA methylation in SLC1A2 (encodes excitatory amino acid transporter 2, EAAT2, the main glutamate transporter), FBXO38 (ubiquitination and antitumor activity regulator), HMP19 (pancreatic cancer suppressor), PPP1R9A (encodes NRB1, a key protein involved in shaping synapses), and in some other genes encoding functionally different proteins in the brain [29].

Fig. 2.
figure 2

A brief summary of the role of DNA methylation and demethylation in the regulation of neuronal functions. Native cytosine is methylated by three main methyltransferases, Dnmt3a, Dnmt3b, Dnmt3L, with Dnmt1 being responsible for maintaining continuous methylation. Passive demethylation is driven by various external factors. Аctive demethylation is catalyzed by TET-methylcytosine-dioxygenases (TETs) and thymine-DNA-glycosylase (TDG, see also Table 2).

Full size image
Table 2. Enzymes for genome methylation and demethylation in rodents and zebrafish (also see Figs. 1, 2)
Full size table

ZEBRAFISH GENOME METHYLATION IN MODELING BRAIN DISORDERS

Zebrafish is a logical model for studying DNA methylation since the basic epigenetic mechanisms are evolutionarily conserved across all vertebrates [30] (Fig. 3). To date, there are data on the effect of DNA methylation on the development of zebrafish nervous system, myelination of oligodendrocytes, and differentiation of nerve cells during neural tube formation (e.g., morphine-induced decrease in DNA methylation and H3K27 hyperacetylation) [31]. DNA methylation also depends on neuroprotection, neuroplasticity (e.g., morphine-induced decrease in neurotrophin BDNF production), proliferation of neural precursors (e.g., due to dysfunction of lysine (K)-specific methyltransferase 2A, KMT2A, which maintains the expression of various genes), as well as tumorigenesis in the zebrafish CNS (e.g., due to compromised expression of the MYCN and ALK genes encoding the proto-oncogenic protein N-Myc and receptor tyrosine kinase, respectively) [31].

Zebrafish have also been used to study aberrant epigenetic mechanisms in offspring exposed to ethanol in embryogenesis [32], and to model anxiety-like behavior in fish exposed to polycyclic aromatic hydrocarbons (PAHs) [33]. Fetal valproate syndrome, a developmental disorder of the fetal nervous system, caused by exposing a fetus to the histone deacetylase inhibitor sodium valproate, whose symptoms resemble autism and whose pathophysiological mechanisms are poorly understood, has also been successfully modeled in zebrafish. Specifically, the serotonergic deficit, characteristic of this syndrome, has been investigated and found to be associated with Ascl1 proneural gene silencing [34].

Zebrafish are actively used to study the role of DNA methylation in the pathogenesis of complex polygenic psychiatric disorders (for examples see Fig. 4). Since schizophrenia in humans is associated with a single-nucleotide polymorphism in the gamma-aminobutyric acid type A receptor subunit beta-2 (GABRB2) gene, the manifestation of schizophrenic symptoms (social isolation and cognitive dysfunction) in zebrafish is modeled by the administration of I-methionine (MET). Meanwhile, the level of DNA methylation in general and GABRB2 promoter in particular is significantly increased, recapitulating the pathological pattern observed post mortem in the brain tissue of patients with schizophrenia [35]. Zebrafish models of neurodegenerative diseases, such as pharmacological models of parkinsonism induced by 1-methyl-4-phenylpyridinium (MPP+), have also been engineered [36], demonstrating the disruption of epigenetic mechanisms and a therapeutic efficacy of trans-2-phenylcyclopropylamine, an inhibitor of histone demethylation enzymes [37].

The group of AID/APOBEC polynucleotide cytidine deaminases is a family of proteins able to introduce mutations into DNA and RNA [38]. AID/APOBEC provide amine deamination to a carbonyl group and 5-methylcytosine conversion to thymine, which is one of the mechanisms of active demethylation. Modeling of the mechanism of action of the AID/APOBEC complex in zebrafish demonstrated that AID/APOBEC overexpression promotes DNA demethylation in fish, whereas its knockdown or knockout inhibit DNA demethylation of various genes required for cellular reprogramming and development [39, 40].

Fig. 3.
figure 3

Effects of environmental factors on gene expression altered by DNA methylation in various cells of the zebrafish brain. TET—Tet-methylcytosine dioxygenases, DNA demethylation enzymes; DNMTs—DNA-methyltransferases, DNA methylation enzymes (see also Figs. 1, 2).

Full size image
Fig. 4.
figure 4

Effects of promoter methylation of selected marker genes on the development of CNS phenotypes in zebrafish. Hypermethylation of the glucocorticoid receptor gene nr3c1 promoter alters its expression and, hence, normal functioning of the hypothalamic–pituitary–adrenal axis, evoking anxiety-like behavior in fish [58]. The lack of promoter methylation of the peroxisome proliferator-activated receptor gamma (pparg) gene, a key transcription factor controlling the expression of many downstream genes involved in lipid transport, biogenesis and lipolysis, upregulates pparg expression, leading to rapid accumulation of ATP and abnormally increased locomotor activity in zebrafish larvae [59]. The glutamate ionotropic receptor NMDA type subunit 1 (grin1) gene plays an important role in the pathogenesis of memory impairment. Nicotine-induced hypermethylation of its promoter in fish (and also in rats) suppresses grin1 expression, leading to memory deterioration in zebrafish [60].

Full size image

PROSPECTIVE RESEARCH TRENDS

Importantly, brain cells, unlike other tissues, are able to change DNA methylation pattern during development, which is key for learning and memory processes. Therefore, disorders of DNA methylation mechanisms lead to impaired cognitive activity of the brain. CNS diseases associated with DNA methylation disorders (Table 1) include Rett syndrome (caused by mutated methyl-binding protein MeCP2 [41]) and hereditary sensory and autonomic neuropathy type 1, which develops into dementia and hearing loss in adulthood (due to autosomal dominant mutation in the Dnmt1 N-terminal regulatory domain sequence [42]). DNA methylation disorders are also associated with Martin-Bell or Fragile X-chromosome syndrome (resulting from abnormal methylation of a trinucleotide repeat in the FMR1 gene on the X chromosome), which is a common form of mental retardation [43], as well as Prader-Willi and Angelman syndromes caused by erroneous DNA methylation of an imprinted allele and manifested as significant mental disorders [44]. Since incorrect expression or dysfunction of even a single gene entails significant perturbations in brain activity, the importance of understanding the mechanisms underlying the impact of DNA methylation on CNS gene expression cannot be overestimated.

As mentioned above, environmental factors impact epigenetic mechanisms, including DNA methylation. For example, exposure to narcotic substances [45, 46], ethanol [47], and some drugs [48] can modulate DNA methylation. Stress is also such an environmental factor. For example, increased methylation of the glucocorticoid receptor promoter during neonatal stress in mice, leading to its under-expression, resembles similar changes in abused children [49], as well as in patients with schizophrenia, depression, and bipolar disorder [50]. In this regard, studying pharmacogenic and stress-induced modulation of DNA methylation in the zebrafish brain appears to be an extremely exciting and promising research trend.

In general, zebrafish are valuable translational objects for modeling human genetic disorders, while being practically as good as rodents in their potential to reproduce mental disorders [51, 52]. Although epigenetic processes in zebrafish are still poorly understood, the use of these model objects for epigenomic studies, specifically, in terms of organism’s viability extension has a considerable potential. One of the pressing priorities is exploring epigenetic mechanisms in the development of the normal zebrafish brain, namely the processes of shaping the epigenome landscape and functional diversification of neurons, formation of neuronal networks and brain functions in general and in different periods of the organism’s life. Among the future directions is the development of technologies sensitive to all forms of cytosine, primarily to 5-methylcytosine and 5-hydroxymethylcytosine [53], and the analysis of changes in cellular processes caused by their accumulation.

A separate area of research on the role of DNA methylation in zebrafish CNS relates to clarifying the influence of individual epigenetic factors on the organism’s formation. The perinatal period is known to be critical to normal brain development [54], so studies of DNA methylation and other epigenetic changes during this period, as well as their links to the subsequent pathogenesis of brain diseases are of particular importance. Although the canonical functions of DNA methylation are highly conserved, there is a wide range of DNA methylation levels and patterns in different animal species. Comparative analysis of DNA methylation patterns in vertebrates (specifically, in rodents and zebrafish) appears therefore promising, as the insight into these inter-taxa differences may help reveal new evolutionary aspects of CNS gene regulation.

A crucial aspect of studying the regulation of gene expression is the genome size. Zebrafish and rodents have significant differences in the amount of genetic information, which can influence gene diversity, location, and functional characteristics. A larger genome size may provide additional opportunities for evolution and diversity, whereas a smaller genome size may imply a more efficient organization of genetic information. The orchestration of genomic regions also plays a decisive role in cell functioning.

Zebrafish and rodents have differences in the structure of chromosomes, the arrangement of genes and other functional elements. They also exhibit differences in DNA methylation patterns, which may be reflected on the phenotype and specific biological processes. For example, a comparative DNA methylation profiling of brain cells in mice and zebrafish showed that on average 24% of CpGs were methylated in the mouse brain, while in zebrafish this index reached 70% [55]. These data confirm that DNA methylation is species specific. Furthermore, until recently, many DNA methylation studies focused on such genomic regions as CGIs and transcription start sites (TSSs), and only few recent studies assessed DNA methylation in other elements, such as intergenic regions and enhancers (Fig. 4). Nevertheless, a positive correlation between DNA demethylation at enhancers and the expression levels of nearby genes has been shown in zebrafish [56].

Due to high occurrence of intergenic regions in zebrafish genome, they also have a higher overall DNA methylation level than mice [55], which corresponds to the presence of a greater number of transposons in the intergenic regions of the zebrafish genome compared to mice. For example, given the genome size and the proportion of DNA repeats, transposons can account for up to 52% of the zebrafish genome and 45% of the mouse genome [57]. The understanding of the differences in gene regulation between zebrafish and rodents, including at the level of DNA methylation, may shed light on the molecular mechanisms underlying their biological differences, including the CNS, while further studies in this area may open up new avenues for the treatment and prevention of diseases associated with compromised CNS gene expression.

The analysis of sex differences in the zebrafish CNS [61] is another promising area of research in terms of the growing recognition of their role both in the functioning of the normal human brain and in its pathology [62], as well as the data on sex differences in DNA methylation in the human [63] and zebrafish [64] brain. Another important trend in zebrafish DNA methylation studies is concerned with analyzing the effects of various neurotropic drugs and toxins on gene expression of DNA methyltransferases and demethylases in the fish brain and the synthesis of the enzymes themselves, as well as evaluating their enzymatic activity. For example, arecoline, a natural psychoactive alkaloid with a partial agonism to nicotinic and muscarinic acetylcholine receptors, when administered chronically, exerts an anxiolytic effect on fish behavior against the background of increased gene expression in the brain. This suggests the involvement of epigenetic regulation in the long-term effects of arecoline via DNA methylation [65]. Thus, it is possible to identify at least several different DNA methylation-related mechanisms, due to which one or another neurotropic drug can exert its effect on the CNS (Table 3). Given the high throughput of zebrafish as a screening model, testing these drugs on zebrafish may lead to the identification of new classes of pharmacological agents that inhibit or enhance DNA methylation in the CNS.

It is traditionally believed that an increase in the methylation of CGIs at gene promoters leads to the suppression of gene expression, while a decrease in methylation, by contrast, evokes its enhancement. For example, an increase in CpG methylation at the promoter of the brain-derived neurotrophic factor (BDNF) gene correlates with the diminished BDNF synthesis in mouse neurons [66]. At the same time, for other genomic regions, there is no definite correlation with the level of DNA methylation and the direction of gene expression changes [67]. For example, various associations of DNA methylation with severe human depressive disorders occur in genomic regions beyond the promoter (e.g., overall hypomethylation of the synapsin gene (SYN2) in depression [68, 69]). Studies of autism spectrum disorders (ASD) using genome-wide sequencing with sodium bisulfide conversion of the placenta from mothers, whose children subsequently developed ASD, revealed a change in methylation of a previously uncharacterized non-coding region of the NHIP (Neuronal Hypoxia Inducible, Placenta Associated) RNA gene [70]. These data demonstrate that the link between DNA methylation and CNS developmental disorders is diverse and affects regions across the entire genome, but not only coding regions. In addition to blocking transcription factors during DNA methylation (Fig. 1), yet another mechanism of epigenetic regulation is based on the hydroxylation of 5-methylcytosine by TET demethylases (Fig. 2) [70, 71]. Currently, the differences between DNA methylation and hydroxymethylation are usually not accentuated, measuring both modifications indiscriminately. For this reason, in future DNA methylation studies [72] in the zebrafish brain, it is important to study these two processes, as well as their individual contributions to CNS modulation, separately. DNA methylation outside CpG regions occurs most frequently in mammalian neurons and glial cells (specifically, in mice and humans), representing a rare phenomenon in the frontal cortex of a human fetus, but significantly augmenting later in life against the background of synaptogenesis and increasing synaptic density [73]. DNA methylation at non-CGI regions may play an important role in the regulation of gene activity, and continue in the adult brain, acting similarly to CpG methylation when repressing transcription [74]. For example, binding of the classical MBP MeCP2 to unmethylated DNA sequences is crucial for BDNF expression, as shown in a mouse model of Rett syndrome [75]. In this regard, investigation of this issue in zebrafish models may be quite relevant.

The analysis of DNA methylation in the zebrafish CNS also greatly matters when studying the role of hormonal (e.g., steroid) regulation, as well as in the action of various endocrine disruptors. Specifically, methylation of the vitellogenin 1 gene flanking fragment in the zebrafish brain proves to be sensitive to chronic estrogen exposure [76]. The gonadoliberin 3 (GnRH3) gene promoter in the zebrafish telencephalon and olfactory bulb reveals a hypermethylation pattern upon dexamethasone exposure [77]. The effect of the toxic organophosphate TDCIPP on zebrafish larvae leads to increased anxiety-like behavior in adult animals against the background of increasing gene expression of three DNA-methyltransferases, as well as to hypermethylation at promoters of the BDNF and dopamine receptor D4b genes in the brain of female, but not male, zebrafish [78], thus demonstrating sex-dependent toxic effects on the CNS. In zebrafish, epigenetic regulation under exposure to a number of such substances may also be transgenerational in nature. For example, microcystin-leucine arginine, an intracellular toxic endocrine disruptor, enhances promoter methylation (and hence decreases expression) of the BDNF and glutamate carboxylase 1 genes in the brain of F1 generation upon drug administration to F0 generation individuals [79].

Table 3. Several putative mechanisms of action of neuroactive drugs on genome methylation-related CNS processes
Full size table

CONCLUSIONS

DNA methylation in the brain can alter gene expression of key signaling molecules, receptors, and growth factors, affecting structural changes in neurons and the shaping of new synapses, which is directly related to the plastic properties of the brain (Fig. 2). In addition, DNA methylation plays a pivotal role in brain development since its earliest stages. There has been created a number of genetic and pharmacological zebrafish models of various diseases caused by DNA methylation disorders (Fig. 3), including those induced by ethanol, narcotics, and stress. Zebrafish are also widely used to model psychiatric (schizophrenia, autism) and neurodegenerative (parkinsonism, dementia) diseases associated with DNA methylation disorders in humans. In general, modeling DNA methylation disorders in zebrafish is a promising tool to study and elucidate molecular mechanisms of epigenetic disorders in the pathogenesis of a number of diseases, also providing an opportunity to develop new approaches to their diagnosis and treatment. The accumulated knowledge on the role of DNA methylation in the zebrafish CNS provides a basis for further translational studies in this area, aimed at finding effective therapeutic strategies.