Epigenetic mechanisms, trauma, and psychopathology: targeting chromatin remodeling complexes

SWI/SNF family

A complex taxonomy of SWI/SNF chromatin remodeling complexes is based on the conserved ATPase subunit that represents the SWI2/SNF2 protein family. In humans, the most common ATPase subunits are hBRM and BRG1 (Vignali et al., 2000; Becker and Hörz, 2002). A complicated machinery of different protein subunits is formed around the ATPase subunit (Table 2). One of those proteins are the BRG1 or HBRM-associated factors (BAF) complexes. The mutations of the polymorphic BAF complexes are often associated with altered cognitive functions. Notably, BAF complexes are expressed by postmitotic neurons [known as neuronal BRG1 or HBRM-associated factors (nBAF)]. Mutations in the neuron-specific subunit known as BAF53b (actin-related protein) result in long-term memory deficits and altered hippocampal synaptic plasticity (Vogel-Ciernia et al., 2013). Transgenic mice with a deletion of the hydrophobic domain of BAF53b perform significantly worse than healthy control mice on the Object Location Memory Test (OLM) and Object Recognition Memory Test (ORM) as well as in a long-term memory version of both tests. In the short-term memory versions of OLM and ORM, mutants did not differ significantly from wild-type (control) individuals. Similar results were obtained in fear conditioning tasks in the lateral amygdala neurons (Yoo et al., 2017). Genetic manipulation in the BAF53b subunit results in an altered gene expression and leads to an impairment of hippocampal synaptic plasticity (Vogel-Ciernia et al., 2013) and a blunted dendritic outgrowth in the hippocampus (Wu and Liu, 2007). The overexpression of BAF53b does not enhance the already existing memory but plays a crucial role in consolidating and forming a memory trail. BAF53b operates in a late phase of memory creation (after up to 24 h after training) and plays an essential role in creating long-term memory as found in the paradigm of fear conditioning (Yoo et al., 2017). Important recent reports by Marom et al. (2017) confirm the key role of the BAF complex in developing intact cognitive functioning. A pathogenic variant of the actin-like 6A (ACTL6A) gene (which encodes another protein component of the BAF complex) was described in three subjects with developmental disabilities, affecting mainly language and memory (Marom et al., 2017). Although the size of a group does not allow us to draw unequivocal conclusions, this report is part of a series of studies confirming the relevance of ATP-dependent complexes and their genetic impact on cognitive functions. Altered hippocampal cytoarchitecture, especially among dendrites, is also observed during the manipulation in other components of BAF, especially the SWI/SNF-related, matrix-associated, actin-dependent regulator chromatin A2 (SMARCA2) gene. Some polymorphisms of the SMARCA2 gene are linked to the animal model of schizophrenia (Loe-Mie et al., 2010) and human schizophrenia (Walsh et al., 2008), whereas mutations in SMARCA2 are linked to the Coffin-Siris syndrome (Santen et al., 2012; Tsurusaki et al., 2012), Nicolaides-Baraitser syndrome (Van Houdt et al., 2012), and general intellectual disability (Hoyer et al., 2012). In addition, SMARCA2 (and its protein product) seems to be one of the ‘network centers’ among the SWI/SNF (BAF) complexes. The SMARCA2 gene is modified by the down-regulation of REST/RFSF, which results in alterations of other interactors and consequences in the development of abnormal dendritic spines. Mouse models revealed impaired social interaction and prepulse inhibition in Smarca2 knockout mice. Importantly, deficits in both domains – social interaction and prepulse inhibition – are observed in patients with schizophrenia. Moreover, it is possible to manipulate the expression of the SMARCA2 gene by the application of psychoactive drugs that down-regulate this complex and antipsychotic drugs that in turn up-regulate it. Therefore, SMARCA2 is considered to be involved in the pathophysiology of schizophrenia (Koga et al., 2009; Loe-Mie et al., 2010). Several mutations in the SMARCA2 gene seem to be related to the Coffin-Siris syndrome, which is characterized by intellectual disability, growth deficiency, microcephaly, and a number of dysmorphic features (Tsurusaki et al., 2012). Loe-Mie et al. (2010) revealed how these schizophrenia-related genes, and their protein products, are evolutionary novel. Risk alleles in the SMARCA2 gene are derived from a point mutation in the conservative region. This mutation leads to altered protein trafficking in the cytoplasm-nucleus pathway. This mechanism is hypothesized to influence both robust primate evolution and schizophrenia. Finally, there is a positive selection among primates and humans toward the SMARCA2 risk alleles, which reinforce the appearance of this evolutionally novel molecular phenomena (Loe-Mie et al., 2010).

Epigenetic mechanisms regulated by ATP-dependent chromatin remodeling structures within the SWI/SNF family are being affected by other ongoing epigenetic processes or even occur in coordination with other chromatin remodeling complexes (Zaghlool et al., 2016). Gozes (2016) suggested that there is a possible interaction between activity-dependent neuroprotective protein (ADNP) and the SWI/SNF complexes. An increased activity of ADNP (at the mRNA and protein level) has been reported in peripheral blood lymphocytes of patients with schizophrenia as well as in individuals with mild cognitive impairment (Gozes, 2016). The ADNP level has been associated with cognitive functions in patients with Alzheimer’s disease (Magen and Gozes, 2014) and correlates with the level of τ protein in an animal model of Alzheimer’s disease (Schirer et al., 2014). The altered interaction between ADNP and SWI/SNF is proposed as one of the potential indicators of ASD (Vandeweyer et al., 2014). Helsmoortel et al. (2014) showed that mutations in the ADNP gene occur in at least 0.17% of patients with ASD. This estimation may seem irrelevant, but in fact it is among the most prevalent ASD genes, taking into account the fact that no single gene is found to be mutated in more than 1% of patients with autism (Vandeweyer et al., 2014). These findings provide evidence of complex protein-protein and multigene interactions that could be a consequence of epigenetic mechanisms. A biochemical analysis performed by Mandel et al. (2008) revealed that proteins form the SWI/SNF chromatin remodeling complex family (BRG1, BAF250a, and BAF170) immunoprecipitated with ADNP. Furthermore, ADNP binds to SMARCA2, SMARCA4, and SMARCC2 by its C-terminal end and possibly positions the whole protein complex in the appropriate DNA location by its zinc finger and homeobox domain. This crucial relationship explains the impact of epigenetic processes, involving nBAF complexes and concomitant enzymes on altered cognitive abilities, intellectual disabilities, and ASD. A specific protein architecture among nBAF complexes regulates the functional characteristics and tissue specificity during neuronal development. Location shift of 3 of 15 subunits of BAF components initiate dendritic outgrowth and axonal development. Thus, mutations of the nBAF protein components (SAMRCB1, SMARCA4, SMARCA2, SMARCE1, ARID1A, and ARID1B) are reported in syndromic intellectual disability disorders (Santen et al., 2012; Vandeweyer et al., 2014). To date, the reports of exome sequencing analyses of patients with ASD reveal even more important mutations in genes encoding the nBAF proteins, including BAF155, BAF170, BAF180, BAF250b, and BAF100a (O’Roak et al., 2012; Basak et al., 2015).

ISWI family

A complex taxonomy of the ISWI chromatin remodeling complexes is based on the conserved ATPase subunit that belongs to the ISWI protein family. In humans, the most common ATPase subunit is SMARCA5 (hSNF2h; Vignali et al., 2000; Becker and Hörz, 2002). A complex machinery of protein is formed around the ATPase subunit. The most widely studied protein complex is ATP-utilizing chromatin assembly and remodeling factors (ACF) and its subunit bromodomain adjacent to zinc finger domain 1A (BAZ1A; Table 3). Animal models imply an overexpression of the ACF (and BAZ1A) gene in the nucleus accumbens (NAc), which results in greater reductions of social interaction and lower sucrose preference. As much as 65% of experimental mice with up-regulated ACF exhibit depression-related behavioral abnormalities, including social avoidance. The up-regulation of BAZ1A correlates with a greater susceptibility to social defeat stress in animal models (Sun et al., 2015). This leads to a disruption in nucleosome location and gene silencing (Sananbenesi and Fischer, 2009). The chronic social defeat stress (CSDS) experiment demonstrates the importance and impact of epigenetic processes on the development of pathological behavior (Golden et al., 2011; Sun et al., 2015). It is undeniable that some individuals are characterized by genetic vulnerability to develop depression-like behavior. However, this genetic background is not sufficient; thus, an interaction with disadvantageous environment is obviously required for the development of depression (e.g. mice placed in a cage with dominant and an aggressive individuals with high position in the hierarchy). In this experimental paradigm, epigenetic changes are observed, which includes upstream regulation of the ISWI family subunit (BAZ1A), likely due to altered ventral tegmental area neuronal projections to NAc as well as the activation of the BDNF protein (Sun et al., 2015). In the NAc, BAZ1A alters the activity of chromatin remodeling complexes. ATP-dependent ‘protein machinery’ is more likely to position nucleosomes within the transcription start sites (TSS regions), blocking translation and silencing genes (Sananbenesi and Fischer, 2009). Therefore, depressive-like behavior can be observed not due to the activation of some defective or mutated genes but due to epigenetic alterations of ATP-dependent chromatin remodeling complexes (Figure 1).

Figure 1:

Euchromatin (A) changes to heterochromatin (B) in NAc.

ATP-dependent chromatin remodeling complexes position nucleosomes close to the TSS region (B) and therefore unable transcription. In an animal model of CSDS, Sun et al. (2015) showed that such a phenomenon occurs more often among individuals susceptible to CSDS, and it leads to development depression-like phenotype.

The BAZ1B protein, with a similar structure to BAZ1A, works contradictory to BAZ1A and heightens responses to rewarding stimuli as well as promotes adaptive responses to aversive stimuli. BAZ1B appears to be closely related to chronic salient stimuli: cocaine administration and CSDS. Importantly, behavioral effects caused by cocaine and CSDS last longer than remodeled chromatin complex. BAZ1B levels return to control values within hours after cocaine administration or stress exposure, whereas behavioral repercussions of these treatments can be still observed for a long time afterward. Thus, it is hypothesized that epigenetic changes in chromatin induce further alterations in various genes, which exert longer behavioral effects (Sun et al., 2015). BAZ1A contribute to stress susceptibility maintenance of depressive-like phenotypes, whereas BAZ1B promotes adaptive responses to aversive stimuli, being hypothesized to serve as a stress resilience indicator. Other studies have revealed an interaction between the ACF (ISWI family) and BAF (SWI/SFN) components, particularly between BAZ1B and SMACRA genes, which can have clinical implications for future studies (Zaghlool et al., 2016). BAZ1B plays a pivotal role in neuron differentiation from pluripotent stem cells. Moreover, the haploinsufficiency of the BAZ1B gene plays a crucial role in the development of Williams syndrome (Lalli et al., 2016).

CHD family

In the CHD family, SNF2-like ATPase domain is located in the center of a protein network and the tandem chromodomains are located in the N-terminal region (Marfella and Imbalzano, 2007). There are several human diseases associated with the CHD protein family (e.g. dermatomyositis, Hodgkin’s lymphoma, neuroblastoma, and the CHARGE syndrome) and some reports suggest that it has an anticancer properties (Rother and van Attikum, 2017), but due to the scope of this article we focus on the CHD8 gene, which is strongly associated with ASD. This gene encodes a protein with the ATP-dependent chromatin remodeling activity. CHD8 binds to posttranslationally trimethylated histone H3 at active promoters and modulates the interaction between DNA and histone (Thompson et al., 2008; Cotney et al., 2015). Krumm et al. (2014) revealed that CHD8 represents genes with de novo loss-of-function mutations in large-scale resequencing studies of ASD. Combined with reports of the CHD8 role in the neurodevelopment of midfetal cortex, these findings suggest an important role of the CHD protein family in the development of cognitive impairments. Other research has confirmed that CHD8 activity targets risk genes for ASD. Indeed, a loss of the CHD8 gene significantly deregulates the function of ASD-related genes (Cotney et al., 2015). An important role of CHD8 in genomic homeostasis during neurodevelopment, along with reports on BAF protein and the ADNP enzymatic role (discussed in the ‘SWI/SNF family’ section), implies that understanding and explaining the pathophysiology of ASD might be achieved using epigenetic approaches.

Memory of psychological trauma and ATP-dependent chromatin remodeling complexes

The definition of epigenetic processes emphasizes its crucial role in the mediation between environmental pressure and genotype (Waddington, 1968; Ptashne, 2007). Stressful and traumatic events, as part of environmental burden, have large impact on the development of the phenotype, somatic and mental health state, psychosocial functioning, and general environmental fitness (Weiner, 1992). Clinical studies reveal a number of posttranslational alterations in protein architecture that result in signaling cascade forming short-term memory. Then, epigenetic processes contribute to the conversion of unstable short-term memory to robust long-term memory trail. Thus, new gene expression is hypothesized to be crucial in creating long-lasting memory trail of a stressful event (Barrett and Wood, 2008; Kwapis and Wood, 2014). Animal studies of memory consolidation often use trauma paradigm or fear conditioning due to the necessity of strong stimulus appliance. There are several lines of evidence revealing how various traumatic events affect epigenetic processes, mainly due to histone acetylation, histone phosphorylation, and methylation, but there are also reports about DNA methylation and nucleosome remodeling complexes (Vogel-Ciernia et al., 2013; Kwapis and Wood, 2014; Sun et al., 2015). Epigenetic mechanisms of ATP-dependent chromatin remodeling complexes take part in memory consolidation through the nBAF complex and its BAF53b subunit. Although BAF53b seems to be pivotal for the consolidation of, particularly, visually contextual fear memory, it does not participate in the consolidation of auditory fear (Kwapis and Wood, 2014). The nBAF complexes play a crucial role in the consolidation of the hippocampus-related memory trails, whereas auditory amygdala-dependent trails are autonomous from the BAF53b complex. In animal studies, the CSDS paradigm is used to develop depressive-like features, manifesting in anxiety, anhedonia, and social avoidance in rodents (Golden et al., 2011). Considering the CSDS protocol, results obtained with this paradigm can easily be categorized as consequences of an exposure to trauma. In this protocol, a smaller mouse is placed in a cage with a bigger and more aggressive male. Interindividual interference that lasts for 5–10 min is pivotal for the protocol because, after the defeat of the smaller mouse, both rodents are separated by Plexiglass, allowing visual and olfactory, but not physical, contact for the following 24 h. The procedure is repeated for 10 consecutive days, each day with a different aggressor (Golden et al., 2011; Sun et al., 2015). The CSDS procedure attunes with modern trauma definitions that imply the occurrence of various types of traumatic events (Sideli et al., 2012; Misiak et al., 2017). Another important factor is interindividual interference (presenting in abuse, neglect, wars, human-made disaster, technological disasters, or assaults), considering there is a lower prevalence of posttraumatic stress disorder as an aftermath of natural disasters (Galea et al., 2005; Neria et al., 2008). Thus, taking into account all the variables discussed above, we here provide a crucial information about molecular alterations emerging after trauma. In animal models, applying the CSDS paradigm results in alterations among all four families of the ATP-dependent protein complexes (SWI/SNF, ISWI, CHD, and INO80) in the NAc. Particularly robust up-regulation of the BAZ1A protein was observed among stress-susceptible individuals, and the interaction between BAZ1A and SMARCA5 genes in executing transcriptional regulations was noted. Interestingly, the up-regulation of BAZ1A mRNA is observed in the NAc in postmortem brain examinations in individuals with depression (Sun et al., 2015). This allows us to hypothesize similar epigenetic mechanisms among CSDS-susceptible people and animal (rodent) models. Moreover, a significant up-regulation in the prefrontal cortex of BAZ1A genes was observed in genome-wide expression profiling of schizophrenia patients performed by Mistry et al. (2013). This is in agreement with several studies showing an increased risk of psychosis in individuals with a history of childhood trauma (Varese et al., 2012). In addition, novel theoretical models consider childhood trauma (and concomitant epigenetic alterations) to be essential components of the etiology of psychosis (Misiak et al., 2017). Early-life adversities cause significant alterations in NAc activity (Tidey and Miczek, 1996; Gambarana et al., 2001; Salamone and Correa, 2012), which is a structure of great dependency from the ATP-dependent chromatin remodeling complexes (Sun et al., 2015).