Long Non-coding RNAs: key players in brain and central nervous system development

Jie Lv<sup>1</sup>; Hui Liu<sup>1</sup>; Hongbo Liu<sup>1</sup>; Qiong Wu<sup>1</sup>; Yan Zhang<sup>2</sup>

Long Non-coding RNAs: key players in brain and central nervous system development

Jie Lv¹

, Hui Liu¹

, Hongbo Liu¹

, Qiong Wu¹

, Yan Zhang²

1. School of Life Science and Technology, Harbin Institute of Technology, Harbin, 150001, China
2. College of Bioinformatics Science and Technology, Harbin Medical University, Harbin 150081, China

Author

Correspondence author
Computational Molecular Biology, 2014, Vol. 4, No. 5 doi: 10.5376/cmb.2014.04.0005
Received: 09 Mar., 2014 Accepted: 26 Apr., 2014 Published: 04 Jul., 2014

This is an open access article published under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Abstract

Regulatory long non-coding RNAs have been emerged as a major contribution of cognitive evolution in mammalian central nervous system and brain tissues. Though proteins have relatively conserved during evolution, the lncRNAs have evolved rapidly to cope with essential and widespread cellular regulation, partly by directing generic protein function. Long non-coding RNAs, highly yet specifically expressed in mammalian brain, provide tissue- and neuronal activity-specific epigenetic and transcriptional regulation. lncRNAs have been documented to be essential for brain development and be involved in brain related diseases. We suggest that lncRNAs are important to modulate diverse central nervous system processes and are the major factor that is important to the brain development, which may be employed to develop novel diagnostic and therapeutic strategies to treat brain related diseases. Moreover, animal models with altered lncRNA expressions and high-throughput approaches would help to understand the mechanisms of lncRNAs in brain development and the etiology of lncRNA-driven human neurological diseases.

Keywords

Long Non-coding RNAs; Central nervous system; Neurogenesis; Brain development; RNA-Seq

Background

The central nervous system (CNS) has been under high evolution and brain is an advanced animal organs. CNS includes distinct categories of neuronal and glial cell types. The amazing cognitive and behavioral functions in brain may involve in neural networks comprised by billions of neurons (Graff and Mansuy, 2008). It is still unknown of the molecular mechanisms about the cooperation among these neurons, though advances in epigenetic areas have been increasing (MacDonald and Roskams, 2009). Based on current view of points and accumulating evidences, epigenetic factors are considered to affect mammalian development and cell differentiation. Furthermore, aberrant epigenetic modification changes by DNA methylation and histone modifications have key roles in human diseases (Kaut et al., 2014; Coppieters et al., 2013; Besingi and Johansson, 2014; Zykovich et al., 2013; Bryant et al., 2014; Sanchez-Mut et al., 2013; Robertson, 2005; MacDonald and Roskams, 2009; Liu et al., 2014; Lv et al., 2010; Lv et al., 2012; Liu et al., 2011; Zhang et al., 2010). For example, the enzymes and complexes such as Polycomb proteins and Trithorax-group proteins, are basal for developmental processes (Kouzarides, 2007; Ringrose and Paro, 2007). However, the mechanisms of loci specificity have only started to be discovered. Recent evidences suggested that the chromatin associated proteins are guided by non-coding RNAs (ncRNAs) (Khalil et al., 2009; Dinger et al., 2008; Mattick, 2009).

The spatio-temporal expression patterns of ncRNAs seem important for CNS function. ncRNAs are implicated in a variety of biological processes including structural (for example, ribosomal RNAs), regulatory (for example, long and micro non-coding RNAs) and catalytic processes. In mammalian brain, ncRNAs are implicated in brain patterning, neuro- genesis, synaptic and neuron connectivity (Mehler and Mattick, 2007) and CNS disease (Taft et al., 2010).

Long non-coding RNAs (lncRNAs) are ncRNAs that are longer than 200 nt and are abundant in brain cell types (Mercer et al., 2008). The classical lncRNAs are transcribed through the same transcriptional machinery as other mRNAs, that is, RNA polymerase II (PolII) occupancy in lncRNA promoter and active histone modifications that are associated with lncRNA promoter and gene body (Ilott and Ponting, 2013). The number of all lncRNAs in mouse is estimated as at least 40,000, which is more than the number of protein-coding genes (Managadze et al., 2013). Most lncRNAs are poorly annotated, and their functions including the roles in CNS functions have not been widely studied. The functions of lncRNAs appear to associate with the genomic localization. For example, lncRNAs can be in close with development associated key genes. Neighboring protein-coding genes can exhibit concordant or discordant expression patterns with lncRNAs (Dinger et al., 2008; Ponjavic et al., 2009), implying the potentially regulatory roles of lncRNAs. Given most of lncRNAs are specifically expressed in brain, the tissue specificity and brain region specificity of lncRNAs seems to be exceptionally vital for regulating CNS functions (Mercer et al., 2008).

Some lncRNAs can regulate the epigenetic modifications of protein-coding genes by cis- or trans-acting fashions that need recruiting chromatin remodeling factors to particular genomic loci (Khalil et al., 2009; Redrup et al., 2009). One classical example of this kind is the HOXC loci where a lncRNA HOTAIR is transcribed and HOTAIR recruits Polycomb protein complex PRC2 to HOXD loci and represses HOXD in trans (Rinn et al., 2007).

1 lncRNAs in the central nervous system

The proximity of lncRNAs to genes related to regulatory development proteins implies that lncRNAs can play important roles in mammalian organ development. Actually, many transcriptomic studies have revealed the dynamic lncRNA expression profiles and their functions among developing, fetal and adult tissues, in additional to embryonic stem (ES) cells (Dinger et al., 2008; Sheik Mohamed et al., 2010), neural cell subtypes (Mercer et al., 2010; Aprea et al., 2013; Lin et al., 2011), and brain (Mercer et al., 2008; Ponjavic et al., 2009; Lv et al., 2013a; Lv et al., 2013b).

1.1 lncRNA expression in brain and neural differentiation

To quickly explore the brain developmental stage specificity and brain specificity, the Allen Brain Atlas (http://www.brain-map.org/) is an option. The Allen Brain Atlas covers in situ hybridization (ISH) data and is a constantly updating website, from which we are able to examine the expression of hundreds of lncRNAs in various tissues in adult and developing mouse brains (Ng et al., 2012a). ~ 64% of 1328 lncRNAs investigated by Allen Brain Atlas are detectable in adult mouse brain and are expressed selectively for specific brain regions especially in hippocampus and cerebellum (Mercer et al., 2008). The brain region specificity is expected as the expression is low in whole brain transcriptome profiling. Therefore, it is necessary to perform transcriptome studies on specific brain regions to improve the lncRNA detection power. In addition, in situ hybridization maps in Allen Brain Atlas revealed that the most lncRNA are expressed in CNS (Mercer et al., 2008). The lncRNAs expressed in CNS are complex, including imprinted transcripts, cis-antisense, intronic and bidirectional transcripts (Carninci et al., 2005). Furthermore, many lncRNAs expressed in CNS exhibited cross-species conservation, which is meaningful as conservation may indicate functionality. Ponjavic et al. have found over 200 lncRNAs that are detectable in developing and adult brain (Ponjavic et al., 2009), which are mainly located near transcriptional regulators with similar expression patterns and a large more conserved lncRNAs may await to be discovered in near future.

Particular lncRNAs which are differentially expressed during CNS differentiation are potential regulators in mediating neural functions. Sox2, an important transcription factor in ES cells, is necessary for neural development. One study has demonstrated that Sox2OT, a lncRNA containing Sox2 in its introns, is expressed in adult neurogenesis (Mercer et al., 2008). Another report indicated that Sox2OT might be responsible for modulating Sox2 expression (Amaral et al., 2009). Taken together, current evidences may suggest that lncRNAs can mediate the expression of other factors to orchestrate neural cell identity.

RNA sequencing (RNA-seq) followed by computational analysis has been widely used to identify tissue restricted expressed lncRNAs. Kaushik et al. had used this approach to identify lncRNA transcripts from five different tissues of adult zebrafish (Kaushik et al., 2013). They identified 442 predicted lncRNA transcripts and 77 differentially expressed lncRNAs. Within the differentially expressed lncRNAs, 61% are brain restricted expressed.

1.2 High-throughput approaches to study the lncRNAs in CNS development.

A study systematically found more than 1600 conserved lincRNAs in four mouse cell types based on chromatin signatures (Guttman et al., 2009). The cell types they investigated include neural precursor cells (NPCs). Their analysis found that those lncRNAs that are associated with “brain cluster” are related to some brain related biological processes, such as hippocampal development and oligodendrocyte (OL) myelination.

The results together with others (Lv et al., 2013a; Lv et al., 2013b; Ng et al., 2012b; Qureshi and Mehler, 2012) have highlighted the importance of lncRNAs in regulation of cellular fate in neural cells and brain. Increasing evidences suggested that lncRNAs can control epigenetic targeting via their ability to bind RNA, DNA and protein (Guttman and Rinn, 2012; Mercer and Mattick, 2013; Tsai et al., 2010). lncRNAs contain functional three-dimensional structures that can form scaffolds or molecular ‘sponges’ and in turn allow activity-dependent regulation (Tripathi et al., 2010; Mercer and Mattick, 2013; Tsai et al., 2010; Barry et al., 2013). Malat1, as an example, has been shown to relate with synapse formation by acting as splicing factor ‘sponge’, suggested that lncRNAs have alternative splicing functions in neural cells (Anko and Neugebauer, 2010). As an earlier mechanistic study, a lncRNA related to alternative splicing in neuronal cells was reported for Gomafu (Barry et al., 2013). The expression of Malat1 was generally stable during induction of stimulating neurons, implying that Malat1 plays a different role in human neuronal functions, or perhaps has regulatory functions in distinct subtypes of neural cells. In addition, lncRNAs are also associated with mRNA transcription, translation and decay (Tripathi et al., 2013; Mercer and Mattick, 2013). Altogether, the enormous regulatory potentials of investigated lncRNAs and even more candidates would call for more detailed studies about the distinct group of non-coding RNAs.

The differential lncRNA expression patterns should be interpreted by experimental or computational functional analysis. As a first step, Mercer et al. (Mercer et al., 2010) systematically analyzed lncRNAs that had significant changes in expression and found that several of these lncRNAs were part of or close to protein-coding gene loci with a known function in brain and CNS development. In addition, a software Scripture was used to reconstruct the transcriptome of mouse ES cells, neuronal precursor (NP) cells and lung fibroblast cells. The full-length transcript structures for most annotated genes and a large number of lncRNAs were construct (Guttman et al., 2010). Another study found that there were ~170 lncRNAs that are differentially expressed during lineage commitment of neuron and oligodendrocyte (OL), neuronal-glial transitions, and developmental stages of OL (Mercer et al., 2010). Recently, a study used RNA-seq to identify lncRNAs that may be important in neurogenic commitment process (Aprea et al., 2013). Some selected lncRNAs have been validated. Recently, Ramos et al. utilized high- throughput approaches including RNA-seq and ChIP-seq to identify lncRNAs related to distinct neural cell types and lncRNAs having important roles in embryonic and adult neurogenesis (Ramos et al., 2013).

In addition, more and more lncRNAs were associated with conserved enhancer elements that regulate the brain development. p300 and H3K4me1 marks have been employed in one work to identify enhancers in mouse that are mediated by neuronal activity (Kim et al., 2010). These predicted enhancers are rich in putative lncRNAs, expanding in either direction from the CBP binding positions and within 2000 bp from enhancer. The enhancer lncRNAs were also found in the intergenic region that are between the Dlx-5 and Dlx-6 loci within the Dlx loci. The region covers with a piece of conserved intergenic enhancer (Zerucha et al., 2000). Dlx-6 is a homeobox element and itself a transcription factor and is vital in embryonic brain developmnet (Wang et al., 2010).

1.3 Regulation of lncRNA expression in the nervous system

How lncRNAs are regulated in CNS and what factors can influence lncRNA expression are not well understood. The main ideas are that lncRNAs are under similar regulatory mechanisms with that of protein-coding genes (Dinger et al., 2008; Guttman et al., 2009; Cawley et al., 2004; Mercer et al., 2010; Zhang et al., 2009). For instance, Pax2, a transcription factor, functions in formation of the mouse brain; while Ncrms is a lncRNA that is exactly mediated via Pax2 (Bouchard et al., 2005). Interestingly, Ncrms is the host gene for miR-135a (Rodriguez et al., 2004), a miRNA, which has reversed expression pattern in medulloblastoma, compared with normal brain (Ferretti et al., 2009). The evidences suggest that genetic and epigenetic factors can both mediate tumorigenesis. In another example, Sox2, which is a pluripotency related transcription factor, plays an important role in the preservation of the Neural Stem Cells (NSCs) in embryonic and adult brain (Pevny and Placzek, 2005). In Sox2 gene loci, a lncRNA exists, which is named by Sox2 overlapping transcript (Sox2OT). Genomic studies showed that it shares same transcriptional direction with the Sox2 gene. Sox2 and Sox2OT transcribe stably in mouse embryonic stem cells and are down regulated during stem cell differentiation. Amaral et al. detected that in the neurogenic region of the adult mouse brain Sox2OT is expressed and is under dynamic regulation during CNS development, suggesting that it can regulate the self-renewal and neurogenesis of stem cells (Amaral et al., 2009).

Nkx2.2as,whichis a lncRNA antisense to the Nkx2.2 gene, is transcribed in the embryonic brain and is necessary to oligodendrocyte development (Price et al., 1992). Aberrant transcription of Nkx2.2as in Neural Stem Cell (NSC) can induce the oligodendrocyte differentiation by Nkx2.2 upregulation, indicating that Nkx2.2as regulates NSC differentiation by increasing the expression of Nkx2.2 (Tochitani and Hayashizaki, 2008).

In addition, recent evidences imply that the perturbed epigenetic processes can alter the lncRNA expression patterns (Mattick, 2009). When treated with trichostatin A (TSA), OL development process is changed. OL maturation is inhibited by TSA which is a histone deacetylase inhibitor by suppressing OL-specific gene expression (Mercer et al., 2010). We summarized the examples of loss of gene function studies in brain and CNS in Table 1, which can be achieved by locally administered RNA interference (RNAi) reagents. Taken together, it is indicated that lncRNAs are regulated by similar transcriptional and epigenetic factors with protein-coding genes.

Table 1 lncRNAs involved in brain and CNS development and the resulting phenotypes in model animal systems

lncRNA	Process	Phenotype	Reference(s)
Dlx1os	Homeodomain transcription factor regulation in developing brain	Morphologically normal together with mild skull and neurological defects by gene inactivation	(Kraus et al., 2013)
Dlx6os1	Homeodomain transcription factor regulation in developing brain	Morphologically normal together with altered GABAergic interneuron development by gene inactivation	(Feng et al., 2006)
Malat1	Tumorigenesis	Normal animal development by gene inactivation	(Zhang et al., 2012)
Miat	Retina development	Defects in specification of retina cell types by knockdown and overexpression in neonatal retina	(Rapicavoli et al., 2010)
Six3os1	Retina development	Defects in specification of retina cell types by knockdown and overexpression in neonatal retina	(Rapicavoli et al., 2011)
Tug1	Retina development	Defects in differentiation of photoreceptor progenitor cells after knockdown in neonatal retina	(Young et al., 2005)
RNCR2	Retina development	Knockdown leads to the increase of amacrine cells and Müller glial cells in post-natal retina	(Rapicavoli et al., 2010)
Vax2os	Retina development	Defects in differentiation of photoreceptor progenitor cells after overexpression in neonatal retina	(Meola et al., 2012)

Note: Long non-coding RNAs: new players in cell differentiation and development

Though lncRNAs are expressed across various tissues, the functions in brain development can be explored if using a traditional knockout approach. For instance, mice with knockouts of lncRNAs Hotair (Li et al., 2013) and Xist (Marahrens et al., 1997) resulted in severe phenotypes, but mice with a knockout of the ubiquitously and highly expressed lncRNA Malat1 displayed no obvious phenotype (Eissmann et al., 2012). Regulation of synaptogenesis (Bernard et al., 2010), alternative splicing (Tripathi et al., 2010), control of cell cycles (Tripathi et al., 2013) and diseases (Gutschner et al., 2013) have been reported for Malat1, but it is still unknown what the precise role is for this abundant and broadly expressed lncRNA. The results indicated that further functional analyses are needed, which is helpful to uncover the functional roles within neural cells.

2 lncRNAs in diseases of the CNS and brain

Disruptions to genome-wide lncRNA-mediated functions could have negative consequences, which is particularly important in the mammalian brain and nervous system where most tissue-specific lncRNAs are expressed. Indeed, it is emerging that lncRNAs are involved in the pathology of neurological diseases related to imprinting, for instance, Prader–Willi syndrome (PWS) and Angelman syndrome (AS) (Koerner et al., 2009). Additionally, lncRNAs that are differentially expressed between ESCs and differentiated neurons are related to schizophrenia (SZ), bipolar disorder (BD) and even autism spectrum disorders (ASD) (Lin et al., 2011). We have summarized several lncRNAs involved in diseases of the CNS and brain in Table 2.

Table 2 lncRNAs involved in diseases of the CNS

lncRNA	Genomics	Evidence	Disease	Reference(s)
Ube3a-as	Antisense to Ube3a	responsible for repressing paternal Ube3a ex- pression; silencing of paternal Ube3a can occur in the absence of Ube3a-as	PWS-AS	(Vitali et al., 2010)
FMR4	share a bidirectional promoter with the FMR1 gene	is silenced in FXS; FMR4 does not simply regulate FMR1	FXS	(Khalil et al., 2008)
ASFMR1	antisense to the 5′UTR region of FMR1	is silenced in FXS	FXS	(Ladd et al., 2007)
Sox2OT	encompasses the entire Sox2 gene	implicated in modulating Sox2 expression	CNS developmental abnormalities	(Amaral et al., 2009)
A region in 2q11.2	2q11.2 chromosomal region that includes DGCR5, a REST regulated lncRNA	VCFS is caused by deletions of the region	velocardiofacial syndrome (VCFS)	(Johnson et al., 2009)
NRON	mediates the cyto- plasmic to nuclear shuttling of the NFAT	NRON is potentially associated with DS through NFAT	Down's syndrome (DS)	(Arron et al., 2006)
BACE1-AS	Antisense to BACE1	modulates BACE1 gene expression; BACE1-AS levels are increased in tissues from AD patients	Alzheimer's disease (AD)	(Faghihi et al., 2008)
BC200	Chromosome 11, p11.2, an ~600,000 bp region	Increased levels of BC200 were found in brain that are preferentially affected in AD	Alzheimer's disease (AD)	(Mus et al., 2007)
ATXN8OS	Antisense to ATXN8	implicated in the molecular pathophysiology of SCA8	spinocerebellar ataxia type 8 (SCA8)	(Daughters et al., 2009; Koob et al., 1999; Moseley et al., 2006)
An unnamed lncRNA	associated with the cyclin D1 gene promoter	Recruit FUS/TLS to repress cyclin D1	amyotrophic lateral sclerosis (ALS)	(Wang et al., 2008)
An unnamed lncRNA	lncRNA transcripts derived from the mouse T early α (TEA) promoter	Responsible in part for MS	Multiple sclerosis (MS)	(Huseby et al., 2012; Friese and Fugger, 2009)
M21981	nested within individual introns of the IL2RA gene	is upregulated with T-cell activation and is identified by genome- wide association studies (GWAS) to be susceptible to MS	Multiple sclerosis (MS)	(International Multiple Sclerosis Genetics et al., 2007)
Tmevpg1	Computational Molecular Biology • Volume 4 View Options . PDF(1819KB) . FPDF(win) . HTML . Online fPDF Associated material . Readers' comments Other articles by authors . Jie Lv . Hui Liu . Hongbo Liu . Qiong Wu . Yan Zhang Related articles . Long Non-coding RNAs . Central nervous system . Neurogenesis . Brain development . RNA-Seq Tools . Email to a friend . Post a comment Terms and Conditions \| Returns and Refund Policy \| Privacy Statement \| Information for Advertisers \| Contact BioSci Publisher Copyright © 2019 Sophia Publishing Group Inc.