Lethality associated with complete knockout models of BAF complexes has posed major challenges in assessing the biological significance of loss of BAF subunits during development. Recently it was found that conditional knockout mouse models of BAF complex subunits, BAF155 and BAF170, resulted in dissociation of other BAF complex subunits, and hence disassembly of the entire complex. The free subunits were then degraded by the ubiquitin-proteasome system resulting in loss of BAF complex functionality (Narayanan et al., 2015). By applying the above mouse model system, the activity of chromatin remodeling BAF complex has been shown to be vital for the overall formation of nervous system structures, because loss of their functions caused severe neural tissue agenesis (Narayanan et al., 2015; Bachmann et al., 2016; Nguyen et al., 2016).

Following loss of BAF complex due to FoxG1-Cre-mediated double conditional deletion of the scaffolding subunits, BAF155 and BAF170, in the telencephalon, the entire forebrain and related structures including the olfactory bulb failed to develop (Bachmann et al., 2016). Similarly, though Emx1-Cre-mediated ablation of BAF complex also had severe effects on specification of cortical structure, the cortex formed thereof was flimsy and markedly malformed (Narayanan et al., 2015).

Induction of FoxG1-Cre activity via tamoxifen injection 3 days after the formation of primary NSCs caused loss of Sox2⁺ cells in olfactory placode and epithelium of BAF155/BAF170_dcKO mutants that were immunohistologically examined at developmental stages E9.5, E10.5, and E11.5. Consistently, there was an absolute lack of HuCD⁺, Tuj⁺, and Ctip2⁺ neurons in the olfactory placode/epithelium, hence confirming that olfactory epithelium (OE) was virtually not specified. Given that Pax6-dependent transcriptional activity was greatly reduced in dcKO of BAF155 and BAF 170 in olfactory NSCs, it has been mechanistically postulated that the BAF complex is essential for conventional activation of Pax6-dependent transcriptional activity in neural precursor cells (Ninkovic et al., 2013; Tuoc et al., 2013a; Bachmann et al., 2016).

The mammalian nervous system consists of billions of neurons and glia that are predominantly derived from multipotent NSCs. NSCs initially exit as neuroepithelial cells (NECs), which later in early neural development (~E9.5–E10.5) transform into the so called multi-/unipotent radial glia cells (RGCs) located in the germinal zone of the brain and spinal cord. During development of the brain and spinal cord, RGCs proliferate via symmetric mitotic cell division to amplify their stemness pool from which subsequent neuronal and glial cell type diversity originates (Guillemot, 2007; Kriegstein and Alvarez-Buylla, 2009; Taverna et al., 2014; Tuoc et al., 2014).

Epigenetic alterations, including chromatin remodeling at specific genomic locations, have been correlated with differential expression of transcription factors, leading to temporal regulation of specific genes that define cell autonomous paradigms needed to establish NSC fate (Hirabayashi and Gotoh, 2010; Juliandi et al., 2010; Coskun et al., 2012; Ronan et al., 2013; Narayanan and Tuoc, 2014; Watson and Tsai, 2016; Yao et al., 2016). These alterations are necessary to maintain NSCs in the cell cycle so as to generate the appropriate number of progeny that will later differentiate to afford the right number of neurons and glia needed for normal neural tissue development. The BAF complex has emerged as a key determinant in the selective activation of developmental signals that ensure NSCs proliferate within a developmental time window before switching to differentiation mode.

Several experiments have led to the concept of the existence of specific npBAF complex (Lessard et al., 2007; Bachmann et al., 2016). The npBAF complex is exclusively functional in neural stem/progenitor cells, and is composed of either a BRG1 or BRM core ATPase, a BAF155::BAF155 homodimer or BAF155::BAF170 heterodimer, and BAF250 paralogs (BAF250A or BAF250B). This is in contrast to esBAF complex that is made up of only BRG1 ATPase with functionally non-redundant BAF155 and BAF170 subunits. In both cases however, subunits such as BAF45a/d, BAF53a, and BAF55a remain unchanged, and during neural induction are capable of specifying self-renewal and proliferation (Lessard et al., 2007; Staahl et al., 2013; Bachmann et al., 2016). Thus, esBAF complex is plausibly reconstituted in NSCs to form npBAF complex in order to confer multipotency, while maintaining proliferative ability. However, as neural development advances, say neural progenitors differentiating into neurons, npBAF complex switches subunits to form neuron-specific nBAF complex (Lessard et al., 2007; Bachmann et al., 2016), which is discussed further in the next section.

In a recent study, it was found that olfactory neural stem cells (oNSCs) express a specific BAF complex (onscBAF) that is essential for their proliferation and maintenance during the development of the OE, a peripheral aspect of the olfactory system (Bachmann et al., 2016). Even though it was apparent (as per Ascl1 and Ngn1 progenitor-specific, and HuC/D, Tuj, and Lhx2 immature neuron-specific immunostainings) that neurogenesis at early (~E10.5) but not later (~E13.5) stages in FoxG1 BAF155 mutants was not significantly affected, there was striking reduction in Pax6⁺, Nestin⁺, and Sox2⁺ oNSCs in the OE basal layer. By applying various cell cycle markers, it was ascertained that progenitor proliferation and cell cycle exit index was greatly reduced in BAF155 mutants. Interestingly, proliferation of apically located Pax6⁺, Sox2⁺, Otx2⁺, REEP6⁺, and K18⁺ glial-like sustentacular cells were unaffected by loss of BAF155 in the olfactory placode or epithelium (Bachmann et al., 2016). Explanatorily, loss of BAF155 subunit during prenatal development of the mouse OE caused precocious regression of NSCs proliferative capacity, leading to reduced generation of neural precursors and hence neurons. Consequently, neurogenesis was impaired at later developmental stages and resulted in dysgenesis of the OE. The function of BAF155 in NSCs to a large extent seem to be evolutionally conserved as its homolog, psa-4, together with psa-1 (BRG1) in worms are known to be important for asymmetric division of NSCs (Bultman et al., 2000; Sawa et al., 2000; Kim et al., 2001).

Direct observation of neurosphere formation in culture indicated impaired proliferative and self-renewal capacity in NSCs lacking BRG1. To that end, In vivo Cre-directed deletion of BRG1 under Nestin promoter in mouse NSCs at E10.5 led to reduced brain size, consequent to diminished proliferation and depletion of neural precursor population. Perinatal lethality was thus inevitable (Matsumoto et al., 2006; Lessard et al., 2007). Compelling evidence obtained from knockdown investigations indicates that other BAF subunits, such as BAF45a/d, BAF53a, and BAF55a are also indispensable for NSC proliferation and self-regeneration (Lessard et al., 2007; Staahl et al., 2013).

Various signaling pathways known to drive cell proliferation are engaged during BAF complex-mediated transcription regulation of neural development (Zhan et al., 2011; Vasileiou et al., 2015). For instance, BRG1-containing npBAF interacts with components of the Wnt signaling pathway to promote neural progenitor proliferation in the telencephalon (Vasileiou et al., 2015). Also during neural patterning, npBAF complex activates aspects of the notch signaling cascade leading to proliferation of CNS precursor cells, while suppressing the sonic hedgehog signaling pathway (Zhan et al., 2011).

During embryonic and early post-natal neural development, proliferative and multipotent NSCs acquire unipotent fate and ultimately differentiate to generate neurons in the process of neurogenesis. In the mammalian embryonic brain, two major neurogenic (germinal) zones, ventricular and subventricular zones (VZ/SVZ), are the main sites of neurogenesis. The VZ abuts the lateral ventricles whereas the SVZ is atop the VZ (Guillemot, 2007; Kriegstein and Alvarez-Buylla, 2009; Taverna et al., 2014; Tuoc et al., 2014). Some neurogenic centers persist in the adult brain to ensure what is known as adult neurogenesis (Zhao et al., 2008; Ming and Song, 2011).

Neuronal generation, identity establishment, and migration to final functional areal position are under multiple regulatory factors including chromatin remodelers. In this section, we will describe the crucial involvement of the chromatin remodeling BAF complex in orchestrating the transition of neural stem/progenitor cells to fully differentiated neurons, and their final localization during early and late neural development.

Accompanying the generation of neurons from progenitors during neurogenesis, immature neurons usually undergo different forms of migration (e.g., radial, tangential, and multipolar) from their birthplaces to their final functional location in the developing brain (Evsyukova et al., 2013). In the developing neocortex, newly born neurons predominantly migrate radially under glial guidance out of the VZ or SVZ to form the laminated cortical plate.

Various factors regulate neuronal migration in the developing cortex (Evsyukova et al., 2013). Although much is yet to be known about the role of chromatin remodeling factors in neuronal migration, Wiegreffe et al. (2015) have elegantly provided striking evidence showing that the chromatin remodeling BAF complex subunit, BAF100a (Ctip1/Bcl11a), is crucial in regulating radial migration of PNs in the developing cortex. Importantly they indicated that Ctip1 expression by migrating neurons (Leid et al., 2004) in the intermediate zone (IZ) is required for multipolar to bipolar morphology transition, which is a critical stage in the process of radial migration (Wiegreffe et al., 2015). After electroporating Cre-GFP into BAF100a-flox/flox brains at E14.5, more GFP-positive multipolar neurons compared to bipolar neurons were found in the IZ of E17.5 neocortex. The multipolar cell accumulation in the IZ may signify failed or disrupted morphological switch to continue their radial migration to cortical layers. Most of the aberrantly migrating neurons due to Bcl11a deletion displayed disorientation with respect to the pial surface, which is an opposing phenotype to normal radially migrating neurons (Wiegreffe et al., 2015). At post-natal stages, BAF100a-deficient cortex displayed discernable dysplasia, especially in the upper layers with resultant reduction in inter-hemispheric connection. Mechanistically, it was revealed that BAF100a directly regulated its downstream effector Sema3c to control PN polarity and radial migration that ensured proper formation of upper cortical layers during neocorticogenesis (Wiegreffe et al., 2015).

After gaining subtype identity, immature neurons undergo specific differentiation processes to become fully matured and functional. During neural maturation, neurites are specified to become either axons or dendrites with which neurons make synaptic connections with adjoining neurons to afford circuit integration. Improper axonogenesis (axonal growth and pathfinding) and dendritogenesis (dendritic outgrowth, arborization, and refinement) results in anomalous neuronal morphology and synaptic connections that underscore several neurodevelopmental disorders. Many factors, including chromatin regulators play vital roles during neural morphogenesis (axonogenesis and dendritogenesis) (Whitford et al., 2002; Wu et al., 2007; Bachmann et al., 2016).

The BAF complex is important in neuronal morphogenesis during neural development (Whitford et al., 2002; Wu et al., 2007; John et al., 2012; Weinberg et al., 2013; Choi et al., 2015; Bachmann et al., 2016). BAF complex controls specification of neurites to differentiate into axons, usually one per neuron, and subsequent axonal fasciculation and pathfinding (Weinberg et al., 2013; Bachmann et al., 2016). For example, sensory neurons found in the dorsal root ganglion (DRG) extend axons that synapse with second order neurons in the dorsal spinal cord to allow relay of somatosensory information. It has been revealed that the BAF complex subunit, BAF100a, is required for terminal maturation and morphogenesis of dorsal spinal neurons during development of the spinal cord (John et al., 2012). Loss of BAF100a expression in spinal neurons was shown to cause central axon dysgenesis, leading to denervation of the dorsal horn by primary somatosensory neurons in the DRG.

Post-mitotic neuron specific BAF53b subunit (Olave et al., 2002) has been shown to play a key role in activity-dependent dendritic morphogenesis. Homozygous knockout of BAF53b subunit in mice largely led to early post-natal lethality, with just about 10% surviving up to adult stages, albeit they displayed behavioral disturbances such as hyperactivity (Lessard et al., 2007; Wu et al., 2007). Cultured cortical neurons lacking BAF53b (BAF53b^−/−) exhibited striking activity-dependent dendritogenic incompetence, which phenotype could only be rescued by exogenous BAF53b GOF but not its homolog, BAF53a (Wu et al., 2007). To mutually exclude the functional role of BAF53b from its homolog, BAF53a, in orchestrating dendritogenesis, two chimeric protein candidates, BAF53a with subdomain 2 of BAF53b and its reverse were generated. Intriguingly, the chimeric BAF53a protein containing subdomain 2 of BAF53b, but not its reverse candidate was able to affect activity-dependent dendrite outgrowth, hence suggesting the divergent subdomain 2 of BAF53b as indispensable for activity-dependent dendrite formation (Wu et al., 2007). That notwithstanding, phenotypes similar to BAF53b^−/− were observed when other subunits of the nBAF complex (i.e., BRG1, BAF45b, and BAF57) were knocked down using RNA interference (Lessard et al., 2007; Wu et al., 2007), suggesting multiple BAF subunit involvement in neurite development. The outcome of investigations in invertebrate models recapitulates BAF53b as a critical player in neuronal dendritogenesis. Loss of function of BAP55 (homolog of BAF53a/b) in Drosophila melanogaster interfered with normal dendritic targeting of olfactory projection neurons, which phenotype was rescuable with expression of mammalian BAF53a/b using GH146-Gal4 driver (Tea and Luo, 2011). Additionally, loss of BAP55 and other Drosophila homologs of BRM, BAF60c and BAF47, perturbed formation of sensory neural dendrites in the peripheral nervous system (Parrish et al., 2006). Plausible mechanisms underlying the role of BAF complex in orchestrating dendritic elaboration include the loss of BAF53b, which causes reduction in the entire BAF complex occupancy at promoters of dendritogenic genes such as GAP43 and Ephexin1, hence culminating in alterations in their expression (Wu et al., 2007). BAF complex is also known to mechanistically engage the calcium-mediated transcription activation function of calcium-responsive transactivator (CREST), which links activity responsive Ca²⁺ signaling and chromatin reorganization by BAF complex known to lead to dendritic outgrowth, arborization, and refinement (Aizawa et al., 2004; Wu et al., 2007).

Even though early olfactory receptor neuron (ORN) population identified by Lhx2⁺, Tuj⁺, and HuCD⁺ immunostaining did not change significantly in BAF155 knockout mice, axonogenesis was markedly defective at early embryonic stage (E13.5) (Bachmann et al., 2016). Likewise, the high expression of BAF170 in ORNs is essential for their maturation, because diminished number of OMP⁺ mature ORN was found in BAF170-deficit OEs (Bachmann et al., 2016). Traditionally, immature ORNs located in the developing OE project axons that synapse with the dendrites of mitral cells in the olfactory bulb during differentiation. The afferent pathway so formed then permits second order axonal projection from mitral cells to primary olfactory centers in the cortex. To confirm interrupted axonal pathfinding from ORNs to telencephalic regions due to the absence of BAF155 subunit, Dil crystal injection of OE was used to reveal axonal dysgenesis in BAF155 deficit ORNs as opposed to obvious visualization of Dil-labeled olfactory neuronal tract in controls (Bachmann et al., 2016). The observation thus consolidates the idea that BAF155 subunit is important for neuronal axonogenesis in the olfactory system. The significance of BAF complex in directing axonal pathfinding is also observed in Caenorhabditis elegans as deletion of the nematode homolog of BAF60 (ham-3) disturbed axonal projection during differentiation of some serotonin-producing neurons (Weinberg et al., 2013).

Thus, the BAF complex plays crucial roles in neuronal morphogenesis in different parts of the CNS (e.g., dorsal root ganglion, forebrain, and olfactory system) in both invertebrates and vertebrates.

Gliogenesis is a neurodevelopmental process that leads to the generation of non-neuronal glial cell types including astrocytes and oligodendrocytes from multipotent NSCs. The production of astrocytes (astrogenesis) and oligodendrocytes (oligodendrogenesis) have been extensively studied in the developing brain and spinal cord (Rowitch and Kriegstein, 2010; Gallo and Deneen, 2014). In many structures of the CNS such as the cortex, differentiation of NSCs initially primed to generate neurons at early stages become progressively restricted in the course of cortical development. Typically, NSCs in the cortex undergo intrinsic changes that make them acquire glial fate at later stages of neural development, with astrogenesis peaking at P0–P2 and oligodendrogenesis at late post-natal stages around P14 in mice (Freeman, 2010; Rowitch and Kriegstein, 2010; Gallo and Deneen, 2014). Epigenetic cell-type-specific regulation of gene expression is implied in changes in NSCs that determine neuro/gliogenic fate. These include transcription factor expression selectivity engendered by epigenetic alterations such as chromatin remodeling that control NSC fate (Hirabayashi and Gotoh, 2010; Juliandi et al., 2010; Coskun et al., 2012; Narayanan and Tuoc, 2014; Yao et al., 2016).

Although the interest in astrocyte function has increased dramatically in recent years, our understanding of astrocyte development, especially under the control of epigenetic factors, has lagged behind as compared to other brain cell types. In addition, the molecular mechanisms underlying astrocyte specification and expansion are largely unknown. Phenotypic analyses of early loss of BRG1 in BRG1cKO mutants with Nestin-Cre indicated that the loss of this core BAF subunit led to a severe defect in the maintenance of NSC pool and consequently also in the production of neurons and glia cells (Matsumoto et al., 2006; Lessard et al., 2007). However, it is unknown whether chromatin remodeling via the BAF-complex regulatory function might control embryonic astrogenesis. In post-natal brain, BAF complex is apparently required in permitting neurogenic tendencies of NSCs, while suppressing their gliogenic commitments. Hence, when BRG1 (Ninkovic et al., 2013) or BAF170 (Tuoc et al., 2016) were abolished in adult NSCs, there was premature onset of gliogenesis.

Oligodendrocyte identity specification is critical to establishing myelination schemes during CNS development. The chromatin remodeling BAF complex also finely regulates generation and maturation of oligodendrocytes. The BRG1 ATPase core of the BAF complex is reported to be vital for the generation of oligodendrocyte lineage. Multi-stage chromatin immunoprecipitation (ChIP) sequencing studies reveal that co-occupancy of BRG1 and the oligodendrocyte-lineage specifying basic-helix-loop-helix transcription factor, Olig2, at enhancer elements synergistically regulate recruitment of transcription factors to gene loci central to oligodendrogenesis (Yu et al., 2013; Bischof et al., 2015). Matsumoto et al. (2006) went further to elucidate how BRG1 interacts with Olig2 to specify oligodendrocytic fate in neural progenitors at early cortical developmental stages. They reported that at about E14 in the mouse cortex, at which stage oligodendrocyte precursors are yet to colonize the cortex, BRG1 was found to interact with Olig2 promoter, leading to Olig2 transcriptional repression. As such, nestin-dependent deletion of BRG1 in early neural progenitor cells resulted in premature expression of Olig2 in the cortex, although such precocious Olig2 expressing cells could not differentiate into oligodendrocyte precursor cells (OPCs). That notwithstanding, BRG1 was found to interact with the said proximal Olig2 promoter in the ganglionic eminence where neural progenitors normally transform into Olig2⁺ OPCs, implying that the BRG1 component of the BAF complex is a repressor of Olig2 expression and an indirect suppressor of OPC specification in the early developing cortex. However, it is needed for the differentiation and maturation of OPCs (Matsumoto et al., 2006; Yu et al., 2013; Bischof et al., 2015).

Learning and memory (formation and storage) are cognitive functions that are underscored by persistent alterations in memory-related neural circuits due to transient environmental stimuli that result in behavioral changes (Day and Sweatt, 2010; Fischer, 2014; Watson and Tsai, 2016). The hippocampal formation is involved in cognitive functions such as memory formation, fear conditioning, and learning (Goncalves et al., 2016). The evidence thereof is mostly centered on hippocampal and DG neurogenesis (adult neurogenesis), dendritogenesis of hippocampal neurons, and various hippocampal-dependent activity experimentations in which synaptogenesis and synaptic alterations such as long-term potentiation (LTP) have been used to study memory formation and maintenance (Goncalves et al., 2016). Overall generation and integration of new neurons into complex neural circuits at pre-/post-natal developmental stages are also of great importance in memory formation and behavior (Deng et al., 2010; Urban and Guillemot, 2014). Given that the hippocampus is a major center of cognitive function in the brain, neurogenic activities that lead to its formation is of critical value to learning and memory. Hippocampal structures, such as the hippocampus proper and the DG are malformed in the presence of defective neurogenesis, which are strongly correlated with behavioral imbalances.

Various regulatory factors play importance roles in learning and memory. For instance, long-term memory processing requires fine interplay of intracellular signaling with transcriptional, translational, and epigenetic factors to regulate gene expression in cognitive function development (Deng et al., 2010; Urban and Guillemot, 2014; Goncalves et al., 2016). Epigenetic mechanisms such as methylation, acetylation, phosphorylation, ribosylation, and chromatin remodeling state (heterochromatin or euchromatin) regulate memory formation and maintenance (Day and Sweatt, 2010; Fischer, 2014; Watson and Tsai, 2016). Accumulating evidence from recent studies indicates the essentiality of the chromatin remodeling BAF complexes in memory formation and learning. Dysfunction of subunits in both npBAF and nBAF complexes in the CNS has been posited to underline various psychiatric perturbations (Koga et al., 2009; Neale et al., 2012; Yoo et al., 2017).

As already stated, BAF170 subunit in BRG1-containing BAF complex is expressed in RGL progenitors in the DG of the adult brain. For this reason, when hGFAP-Cre and Nestin-CreER-mediated conditional loss of BAF170 in RGL cells was performed, it caused abnormal proliferation and localization of RGL cells, with subsequent depletion of progenitor pools in the SGZ of the DG. Due to the lack of BAF170 in RGL progenitors, neurogenesis in DG was skewed to terminal astrogenesis. By using the Morris water maze test, BAF170 mutant mice were observed to display noticeable impairment in adaptive behavior (Tuoc et al., 2016).

The nBAF subunit, BAF100b (Ctip2/Bcl11b), is exclusively expressed in post-mitotic granule neurons in the SGZ of the DG but not in progenitors. Its expression is important for embryonic and adult hippocampal neurogenesis leading to proper functioning of the hippocampus in memory and learning. Newly born neurons lacking Ctip2 locate ectopically of the granule cell layer in the hippocampus. These Ctip2-deficient neurons characteristically display reduced expression of calbindin 1 (Calb1), which plays important roles in neuro-behavioral and memory formation (Simon et al., 2012; Soontornniyomkij et al., 2012). On the whole, loss of Ctip2 during development of the hippocampus leads to dramatic reduction in size and neuronal pool of the DG, and impairment of spatial memory formation (Simon et al., 2012, 2016; Lennon et al., 2017).

Another nBAF subunit, BAF53b (Vogel-Ciernia et al., 2013; Choi et al., 2015; Yoo et al., 2017) is also shown to play a strong role in cognitive processes in mice. Upon RNA sequencing of the dorsal hippocampus of BAF53b^+/− mutant mice, it was revealed that genes involved in memory formation (including cytoskeletal reorganizers and chromatin modifiers) were misregulated when compared with wild-type mice (Vogel-Ciernia et al., 2013). Consistently, alterations in the neuron-specific BAF complex subunit BAF53b expression in amygdala neurons modulated learning and memory formation in adult mice. Specifically, when Yoo et al. (2017) applied viral vector-mediated methods to knockdown or overexpress BAF53b functionality within the lateral amygdala in experimental auditory fear conditioning schemes, they found that inductive expression of BAF53b accompanied learning activities to augment memory consolidation and related neuroplasticity. Likewise, disrupted normal levels of BAF53b expression by genetic manipulations leading to BAF53b heterozygosity (BAF53b^+/−) in mouse CaMKIIa⁺ post-mitotic neurons, caused aberrant neuronal morphology acquisition such as defective spine formation, severe defects in hippocampal synaptic plasticity, and long-term memory impairment (Vogel-Ciernia et al., 2013).

Similar to phenotypes of BAF53b mutant mice, mutation in CREST (also known as SS18L1, a newly identified nBAF subunit) causes defects in learning and memory. Mechanistically, CREST acts together with the co-activator, CREB-binding protein (CBP), to regulate expression of immediate early genes including c-fos. Such early genes play significant roles in brain plasticity, activity dependent memory formation, and other complex behaviors.

All together, these findings suggest that BAF complex-driven chromatin remodeling in neuron generation, differentiation, synaptic development, and plasticity in related structures such as the hippocampus is essential for cognitive functions.

Drs. Grange S. Coffin and Evelyn Siris were the first to report CSS (OMIM #135900) in 1970 (Kosho et al., 2014a). It is a rarely occurring congenital abnormality with clinical characteristics such as intellectual disability (ID), progressive coarsening of the face, hypertrichosis, frequent infections, feeding difficulties, and hypoplasia of the 5th distal phalanges and fingernails (Levy and Baraitser, 1991; Schrier et al., 2012). Schrier et al. (2012) summarized 80 previously reported cases for both common and discriminating clinical hallmarks, and brought forward a differential diagnosis between CSS and other related syndromes, including Nicolaides-Baraitser syndrome (NCBRS) and Deafness, Onychodystrophy, Osteodystrophy, Mental Retardation, and Seizures (DOORS) syndrome. In recent years, comprehensive human exome sequencing and genome-wide association studies have revealed that among roughly 180 CSS cases reported, about 110 patients carry a mutation in genes encoding BAF complex subunits. These genes include BAF250A (ARID1A), BAF250B (ARID1B), BRM (SMARCA2), BRG1 (SMARCA4), BAF47 (SMARCB1), BAF57 (SMARCE1), and BAF200 (ARID2) (Santen et al., 2012; Tsurusaki et al., 2012; Bramswig et al., 2015, 2017; Miyake et al., 2016).

The 6q25 microdeletion syndrome is an uncommon congenital and genetic disease characterized by ID with developmental delay, microcephaly, hearing impairment, distinct dysmorphic features, visual impairment, and corpus callosum agenesis (Santen et al., 2014; Ronzoni et al., 2016). Mutations of BAF250b were reported to be linked with this syndrome (Pirola et al., 1998). The case presenting with agenesis of corpus callosum (ACC) had a deletion of 5.5–7.6 Mb, including BAF250b and 10 other genes. To date, more than 24 cases with clear deletion information have been reported. Most of the patients carry deletions in BAF250b and several other genes, with deletion sizes of about 58 kb–14.5 Mb. Phenotypes are very variable, and dependent on the size and site of the deletion (Sim et al., 2015). It has been proposed that the haploinsufficiency of the affected genes results in defects in normal brain development and is responsible for the phenotype. The essential region for the phenotype was mapped to a 6q25.3 region harboring the protein-coding genes ZDHHC14 and BAF250b (Michelson et al., 2012). Other studies (Halgren et al., 2012; Santen et al., 2014) further suggested that BAF250b was the critical gene responsible for the core phenotypes of the syndrome. Recently, Ronzoni et al., reported an evidence that loss of BAF250b mainly underpins 6q25 microdeletion syndrome (Ronzoni et al., 2016). In that study, a patient carrying a small 6q25 deletion containing only BAF250b exhibited the typical features of the syndrome including developmental delay, speech impairment, and agenesis of the corpus callosum (Hoyer et al., 2012; Mignot et al., 2016; Ronzoni et al., 2016).

NCBRS (OMIM #601358) is a disorder that shares remarkable clinical features with CSS. The syndrome is characterized by ID with impaired speech, sparse scalp hair, prominence of the interphalangeal joints and distal phalanges due to decreased subcutaneous fat, characteristic coarse facial features, microcephaly, and seizures (Sousa et al., 2009; Van Houdt et al., 2012). The hallmark differences between NCBRS and CSS come from features of the hands and feet. The representative NCBRS patients show prominent interphalangeal joints and distal phalanges, whereas those with CSS present hypoplasia/aplasia of the fifth fingernails with or without involvement of the terminal phalanges (Wieczorek et al., 2013).

Recently, exome sequencing studies have suggested that NCBRS is associated with BRM (SMARCA2) mutation. Among the 70 mutations reported in NCBRS patients, one is an in-frame 6-bp deletion, two are in-frame multi-exon deletions affecting the ATPase domain, and 62 are missense mutations. Of the 62 missense mutations, 60 are mapped to the ATPase domain, one is proximal but outside the ATPase domain, and the other is in the bromo domain (Van Houdt et al., 2012; Sousa et al., 2014; Ejaz et al., 2016). Notably, deletions encompassing human BRM locus have not been found in NCBRS (Christ et al., 1999). Because null BRM KO mice do not show major developmental defects (Koga et al., 2009; Magnani and Cabot, 2009) and none of the identified mutations cause truncated proteins, it has been proposed that the mutations in BRM cause GOF or dominant-negative effects. Hence, these mutations possibly interfere with the ability of ATP hydrolysis by native BRM protein (Van Houdt et al., 2012; Ejaz et al., 2016). In addition to BRM, missense mutations in BAF47 (p.Arg366Cys) was also reported to present with NCBRS (Wieczorek et al., 2013).

Kleefstra's syndrome (OMIM #610253) was known as a recognizable mental retardation syndrome caused by heterozygous mutation in the euchromatin histone methyl transferase 1 (EHMT1) gene (Kleefstra et al., 2006, 2009). However, not all the patients presenting core phenotypes of Kleefstra's syndrome carry an EHMT mutation. Indeed, de novo mutations in four epigenetic regulator-encoding genes, MLL3, MBD5, NR113, and BAF47 (SMARCB1) were also identified from four of nine EHMT1 mutation-negative patients (Kleefstra et al., 2012). These patients exhibit core features of Kleefstra syndrome; nevertheless they have very heterogeneous phenotypes, which are therefore referred to as Kleefstra syndrome spectrum (KSS). Common phenotypic features among these patients include ID, childhood hypotonia, behavioral anomalies, synophrys, and midface hypoplasia. Investigation with drosophila model revealed direct interactions between MLL3, MBD5, NR113, BAF47 (SMARCB1), and EHMT1 (Kleefstra et al., 2012). The findings suggested that mutations in any protein within the complex of MLL3, MBD5, NR113, BAF47, and EHMT1 could lead to KSS.

Autism is a neurodevelopmental disorder characterized by deficits in social skills, speech problems, and restricted and repetitive behaviors. It is well established that a combination of genetic and environmental factors cause autism (Chaste and Leboyer, 2012). However, the underlying genetic basis is unclear for at least 70% of cases. Recent exome sequencing analyses of autism patients have revealed mutations in the following BAF complex components: BAF155, BAF170, BAF180, and BAF250b (Neale et al., 2012; O'Roak et al., 2012). A de novo mutation of genes encoding for the above BAF subunits were identified in patients who showed severe social communication problems with low verbal intelligence quotient (IQ) and nonverbal IQ (Neale et al., 2012; O'Roak et al., 2012).

Recently, mutations in BCL11A that encodes BAF100a subunit were also linked to autism. Three patients with microdeletions of 2p15–p16.1 who presented with developmental delay also showed common features of ASD, including moderate to severe ID, facial dysmorphism, and hypotonia (Basak et al., 2015). Although the 2p15–p16.1 microdeletion encompasses ~15 protein-coding genes and six pseudogenes, BCL11A is the only common candidate gene in three reported patients with ASD (Basak et al., 2015).

Schizophrenia (SZ; OMIM #181500) is a chronic, severe, and disabling brain disorder that affects ~0.3–0.7% of the world's population. It is characterized by delusion, thought disorder, auditory hallucination, reduced emotional expression and social engagement, disorganized speech, and lacking motivation. The causes of SZ include genetic and/or environmental factors (Zai et al., 2017). In addition to genetic mutations, epigenetic alteration has also been shown to be involved in SZ (Tsankova et al., 2007). BRM was associated with SZ in a screening study of Japanese population (Koga et al., 2009). Mechanistically, the risk alleles of intronic polymorphisms (rs3763627 and rs3793490) were correlated with decreased expression of BRM in post-mortem prefrontal cortex (Koga et al., 2009). Another risk allele, a missense polymorphism (rs2296212, D1546E), located in a highly conserved functional region in mammalian species, reduces the nuclear localization of BRM (Koga et al., 2009). BRM knockout mice exhibit impaired social interaction and pre-pulse inhibition, which suggest that the impaired functions of BRM might play a role in SZ. In cultured mouse primary cortical neurons, knockdown of BRM resulted in impaired dendritic spine growth and abnormal morphology, which is considered an intermediate phenotype of SZ (Bellon, 2007; Lewis and Gonzalez-Burgos, 2007).

In addition to BRM, intronic SNPs in BCL11A were significantly associated with schizophrenic patients (Basak et al., 2015). Further studies will be necessary to elucidate the influence of the intronic SNPs on BCL11A expression and function.

HSCR is a congenital intestinal disorder characterized by defective migration of enteric neural cells. Recently, a de novo heterozygous frameshift mutation in exon 20 of BAF250b (c.5789delC p.Pro1930Leufs^*44) was identified in a patient with HSCR symptoms (Takenouchi et al., 2016). The patient also displayed some CSS features, including coarse facial features, impaired development, epilepsy, and hypoplasia of the corpus callosum and cerebellum. In addition, the patient exhibited congenital cataract disorder and defects in enteric nervous system, which are the cardinal pathologic characteristics of HSCR (Takenouchi et al., 2016). Interestingly, studies in drosophila model confirmed that dysregulation of the ARID1 subunit leads to impaired enteric neural cell migration (Eroglu et al., 2014).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. The reviewer IP and handling Editor declared their shared affiliation, and the handling Editor states that the process nevertheless met the standards of a fair and objective review.