Current and future perspectives of non-coding RNAs in brain function and neuropsychiatric disease

Abstract

Non-coding RNAs (ncRNAs) represent the majority of the transcriptome and play important roles in regulating neuronal functions. ncRNAs are exceptionally diverse in both structure and function and include enhancer RNAs (eRNAs), long non-coding RNAs (lncRNAs) and microRNAs (miRNAs), all which demonstrate specific temporal and regional expression in the brain. Here, we review recent studies demonstrating that ncRNAs modulate chromatin structure, act as chaperone molecules, and contribute to synaptic remodeling and behavior. In addition, we discuss ncRNA function within the context of neuropsychiatric diseases, particularly focusing on addiction and schizophrenia, and the recent methodological developments that allow for better understanding of ncRNA function in the brain. Overall, ncRNAs represent an underrecognized molecular contributor to complex neuronal processes underlying neuropsychiatric disorders.

Introduction to the diversity of brain ncRNAs

The last decade has seen an unprecedented expansion of our knowledge of how complex genomic regulatory systems interact to shape the developing and adult brain (1–3). Recently, areas of the genome that do not code for proteins have been shown to confer risk for disease states across the clinical spectrum (4). The non-coding genome regulates gene transcription through multiple mechanisms, including the production of non-coding RNAs (ncRNAs). These ncRNAs participate in diverse cellular functions including intracellular signaling, chromatin modification, and alteration of transcriptional activity. Unraveling the molecular functions of ncRNAs is an important step towards understanding the contribution of the non-coding genome to human disease (5–7).

Neuropsychiatric disorders demonstrate particular complexity marked by a high degree of risk-associated common variation located in noncoding regions (8–10). The past decade has seen an increased appreciation for the varied molecular roles of ncRNAs on brain structure and function (11–14). ncRNAs are known to form molecular scaffolds on chromatin, organizing three-dimensional chromatin structure to optimize gene expression (15, 16). In addition to their actions in the nucleus, ncRNAs also modify dendritic macrostructure and synaptic signaling (17–19). Additionally, preclinical and clinical studies suggest that ncRNAs confer risk for neuropsychiatric diseases including addiction and schizophrenia (4, 9, 20, 21).

Here, we aim to contextualize recent breakthroughs in our understanding of the contribution of ncRNAs to basic neurobiological processes and their aberrant functioning in the context of addiction and schizophrenia, focusing on enhancer RNAs (eRNAs), long non-coding RNAs (lncRNAs), and microRNAs (miRNAs) (Figure 1). Our literature search strategy is described in Supplementary Information (Supplementary Tables 1–2). Involvement of ncRNAs in the pathogenesis of additional neuropsychiatric diseases, as well as other RNA species that are less well-studied, are outside the scope of this review. We instead direct the reader to recent reviews of lncRNAs in neurodegenerative disease (22), eRNAs and enhancers in brain disease (23), and brain circular RNAs (24). We close with a proposed translational pipeline to guide discovery and validation of ncRNA transcripts that may represent therapeutic targets in neuropsychiatric disease.

Figure 1. Diversity of ncRNA species in the brain.

While many different types of ncRNA exist in brain and other tissue types, this model focuses on enhancer RNAs (eRNAs), long noncoding RNAs (lncRNAs) and microRNAs (miRNAs). The basic biogenesis of these RNA species and their impacts on target gene expression are shown in the model. eRNAs and lncRNAs are typically minimally processed post-transcriptionally, while miRNAs undergo tightly regulated processing involving multiple steps and nuclear-to-cytosolic export.

Enhancer RNAs in brain function

Enhancers are broadly conceptualized as short regions of DNA that increase the likelihood of gene transcription at one or more cis loci, typically by facilitating chromatin remodeling and promoting transcription factor (TF) binding (11). Enhancers are often identified in genome-scale assays by their enrichment in active, accessible chromatin marks including histone 3 lysine 27 acetylation (H3K27ac) and H3K4me1 (25). Enhancer elements drive transcription of target genes in multiple ways, including three-dimensional chromatin remodeling, enhancer-promotor looping which increases the concentration of transcriptional regulatory machinery near target transcription start sites (TSS), and phase separation of proteins to create transcriptionally active intracellular condensates (26). Our understanding of enhancer biology was further strengthened by the discovery of active transcription at enhancer sites (13). eRNAs were first discovered in neuronal cells, with many of the observed eRNAs acting on immediate early genes (IEGs) such as Fos and activity-regulated cytoskeleton-associated protein (Arc) (13, 27, 28). eRNAs range in size but are typically 50–2000kb in length, are bidirectionally transcribed from enhancer sites, and are generally not polyadenylated or capped (although a small subset is polyadenylated and exhibits greater stability) (26). Interestingly, the production of eRNAs from a specific enhancer site may be a more sensitive indicator of enhancer activity than enhancer-specific chromatin modifications (29).

The exact function of eRNAs continues to be debated, as they were initially conceptualized as passive readouts of open chromatin and promiscuous RNA polymerase activity (13, 30). However, recent studies have demonstrated active molecular roles for specific eRNAs. eRNAs promote looping between enhancers and their conjugate promoter regions, facilitating mRNA transcription (27, 31). eRNA transcripts bind to chromatin-modifying proteins such as CREB binding protein (CBP) independently of RNA sequence to promote CBP-mediated deposition of H3K27ac and increased transcriptional activity (32). eRNAs can also bind to the E-subunit of the negative elongation factor (NELF) protein, which normally serves to prevent poised Pol II from transcribing mRNA (21, 27). N⁶-methyladenosine (m⁶A) modification leads to eRNA degradation, and loss of m⁶A on eRNAs and other chromatin-associated RNAs leads to accumulation of open chromatin (33).

In addition to tissue-specific expression (34), eRNAs are expressed in a brain region-specific and cell type-specific manner (35). eRNAs acting on IEGs are induced more rapidly than their conjugate mRNA after depolarization (13, 27, 36). The Fos locus contains five separate eRNAs that are induced in a stimulus-specific manner to neuronal activation, brain-derived neurotrophic factor (BDNF) treatment, or activation of cAMP (28). Modulation of the conserved Fos eRNA-2 locus by enzymatically dead Cas9 (dCas9) fused to the epigenetic activator p300 leads to recruitment of bromodomain-containing protein 4 (BRD4) and Fos mRNA transcription (37). Interestingly, BRD4 and the mediator complex TF MED1 occupy eRNA-producing super enhancers and promote phase separation of transcriptional condensates in the nuclei in concert with local RNA-RNA interactions (38, 39). Our emerging understanding of neuronal eRNAs suggest that these molecules are critical in optimizing three-dimensional chromatin structure and gene expression patterns in response to a diverse array of stimuli (Figure 2).

Figure 2. Enhancer RNAs (eRNAs) control distinct molecular processes in brain cells.

eRNAs are bidirectionally transcribed from genomic enhancer sites and go on to regulate numerous processes including enhancer-promoter looping and transcriptional activation at the promoter sites of their target genes.

Role of enhancer RNAs in addiction and schizophrenia

Recent research has linked specific eRNAs and enhancer loci to neuropsychiatric diseases and phenotypes in both clinical and preclinical studies (40). Adolescence is a critical maturational period, and exposure to alcohol or other drugs of abuse during this developmental stage leads to an increased risk for addiction and affective disorders mediated in part by epigenetic mechanisms (41). Adolescent alcohol exposure leads to decreased Arc eRNA and mRNA expression in the amygdala, which increases risk for anxiety-like behavior in adult rats. Specifically, decreased CBP and KDM6B at the synaptic activity response element (SARE) enhancer site upstream of Arc produces condensed chromatin architecture due to decreased H3K27ac, resulting in decreased Arc eRNA and mRNA expression in the adult rat amygdala after adolescent alcohol exposure (21). Inhibition of Arc eRNA (−) in the CeA of alcohol-naïve control rats mimics adolescent alcohol exposure-induced anxiety and decreased Arc expression via increased NELF binding at the promoter (21). Genomic deletion of the proopiomelanocortin (Pomc) enhancer leads to lower alcohol consumption and widespread alterations in the hypothalamic transcriptome, affecting genes involved in opioid and oxytocin/vasopressin signaling (42). Additionally, a recent study linked a conserved enhancer of the galanin (GAL) gene with alcohol use and anxiety in males in the large human UK Biobank cohort. CRISPR-mediated knockout of this enhancer in mice leads to decreased GAL expression in the amygdala and hypothalamus and decreased alcohol intake (43).

Enhancers and eRNAs additionally confer significant risk for neuropsychiatric disorders in clinical populations. A study exploring cell type-specific enhancer-promoter interactions in the brain found that sporadic risk variants for schizophrenia are overrepresented in neuron-specific enhancer regions (3). eRNA-encoding enhancer sites active in dopaminergic neurons are also enriched in genetic variants associated with addiction and schizophrenia (9). Similarly, a transcriptomic analysis identified 118 differentially regulated eRNAs in the prefrontal cortex of schizophrenia subjects compared to healthy controls, as well as over 900 enhancer elements whose expression is altered by genetic risk variants (44). Convergent evidence from molecular biology, animal models, and human genetic and phenotypic data suggest that enhancers and eRNAs optimize the brain transcriptome and confer risk for psychiatric disease.

Long non-coding RNAs in brain function and disease

LncRNAs are defined as RNA transcripts that are longer than 200bp that do not encode for proteins (45). In the human genome, lncRNAs represent approximately 96,308 genes that encode 172,216 unique lncRNA transcripts (46). lncRNAs have attracted attention for their roles in regulating gene transcription and translation in both homeostatic and disease states, particularly in the brain where lncRNAs are enriched and highly conserved (47–50).

Two main mechanistic modalities exist for lncRNAs, either acting in cis, defined as influencing the expression of nearby genes, or in trans where lncRNAs control gene expression, transcription, or other cellular functions away from a local particular transcriptional locus (45). LncRNAs broadly act as scaffolding molecules (particularly on chromatin), recruitment molecules, sponges (binding to miRNAs or other RNA classes), splicing regulators, or by binding to proteins to modulate their function (45, 51–53) (Figure 3). For example, the lncRNA HOTAIR provides a structural scaffold to link the histone demethylase lysine-specific demethylase 1 (LSD1) and the polycomb repressor complex 2 (PRC2), thus regulating chromatin dynamics and transcription (53). LncRNAs are transcribed from various genomic contexts, including from discrete genetic loci or as antisense transcripts (i.e., coded from the opposite DNA strand) of protein-coding genes such as BDNF (20, 45, 54) (Figure 3). While some antisense lncRNAs limit the expression of their conjugate sense mRNA (54), other lncRNAs such as Evf2 participate in TF recruitment during neural development (55). Unlike eRNAs, lncRNAs do not have a distinct histone posttranslational modification signature. Instead, similar to mRNA expression, canonical activating marks such as H3K27ac, H3K4me3, and H3K36me3 are correlated with increased lncRNA expression, whereas H3K27me3 is associated with decreased lncRNA expression (56). Conversely, DNA methylation analyses reveal a characteristic increase in methylation density downstream of the lncRNA TSS compared to mRNA TSS (56).

Figure 3. Long noncoding RNAs (lncRNAs) play diverse roles in brain function.

LncRNAs can be transcribed from discrete loci or as antisense transcripts of coding genes. LncRNAs then regulate multiple cellular functions including acting as sponges for other RNAs, regulating and recruiting epigenetic effectors, and controlling chromatin structure and gene expression.

Specific lncRNAs have been implicated in the regulation of cellular and molecular plasticity in the brain. Early investigations into lncRNA function showed that Evf2 is critical in regulating GABAergic interneuron formation and plasticity in the hippocampus by coordinating a TF expression switch (55). Malat1 regulates a conserved set of synaptic plasticity-associated genes and thereby alters dendritic spine density (57). Neat1, a lncRNA involved in nuclear organization via phase separation of paraspeckles (58, 59), is involved in the response to oxidative stress in dopaminergic neurons (60), and Meg3 modulates the degree of AMPA receptor expression of the neuronal cell surface (61). The lncRNA GM12371 is involved in transcriptional regulation of synapse-associated genes in the hippocampus and is activated by the cAMP-PKA signaling cascade (19). The BDNF antisense lncRNA BDNF-AS is involved in the prevention of neurotoxicity in response to ketamine (62) as well as neurite outgrowth and neuronal differentiation during normal development (54). More recent work has connected specific lncRNAs including BC1 and AtLAS to synaptic function in vivo and altered social behaviors in animal models (17, 18), and Neat1 knockout mice also show decreased social interaction and altered stress-induced neuronal plasticity (63).

These critical roles played by lncRNAs in neuronal and behavioral processes suggest that these ncRNAs may be altered in neuropsychiatric disease states such as addiction and schizophrenia. One of the first studies to systematically examine transcriptomic changes in the postmortem brain of alcohol use disorder (AUD) subjects found numerous differentially expressed intergenic lncRNAs compared to control subjects (64). Studies examining specific lncRNAs show that MALAT1 is increased in multiple brain regions in AUD subjects (65). Our group recently demonstrated that BDNF-AS acts as repressor of BDNF by recruiting the PRC2 to deposit repressive H3K27me3 and thereby downregulates BDNF transcription in individuals with early onset AUD (individuals who began drinking before the age of 21) (20). Increased BDNF-AS appears to be regulated by decreased m⁶A methylation in the postmortem amygdala of subjects with early onset AUD (20). In humans and in animal models of cocaine dependence, genome-wide studies show numerous lncRNA transcripts that are dysregulated (66, 67). Chronic cocaine administration in mice causes a downregulation of the lncRNA Gas5 in the NAc, while overexpression of Gas5 in the NAc decreases cocaine induced place preference and self-administration in mice (68). RNA-seq of the NAc in both cocaine-dependent mice and in mice overexpressing Gas5 revealed substantial changes in synaptic plasticity-associated genes (68).

In addition to their emerging roles in addiction, lncRNAs are dysregulated in schizophrenia and related disorders. Copy number variations (CNVs) seen in patients with schizophrenia and autism show enrichment for brain-expressed lncRNAs (69). NEAT1 is decreased in multiple cortical regions in schizophrenia subjects and impacts oligodendrocyte precursor differentiation (70). Analysis of postmortem brain tissue has led to identification of lncRNA DGCR5 as a potential regulator of aberrant transcription of schizophrenia-associated genes including IGSF9B and NRGN (71). Additionally, the activity-regulated lncRNA Gomafu regulates the alternative splicing of genes involved in schizophrenia pathogenesis, specifically DISC1 and ERBB4 (72). The development of deep sequencing technologies in combination with precise genome editors such as CRISPR-Cas9, in addition to further evaluation of lncRNA expression in clinical populations, will allow for more granular knowledge about the role of lncRNAs in neuropsychiatric disorders.

microRNAs regulate brain functions and confer disease risk

miRNAs are perhaps the most well-studied ncRNA species in neuroscience. While previous authoritative reviews have covered miRNA biology relating to brain function and pathophysiology (73–75), here we aim to consolidate recent developments focusing specifically on interactions between miRNAs and other gene regulatory mechanisms as they relate to brain function and the pathophysiology of addiction and schizophrenia.

miRNAs are approximately 22–23 nucleotides in length, and typically function to regulate mRNA stability and translational activity (76, 77). miRNAs are transcribed from discrete genomic loci prior to stem-loop formation, and they are processed via the Drosha/Dicer enzymatic pathway and exported from the nucleus to the cytoplasm to form mature miRNAs (Figure 1) (78, 79). Mature miRNAs are incorporated into the RNA-induced silencing complex (RISC), where they function to target homologous sequences in 3’-untranslated regions (UTR) of target mRNAs leading to mRNA degradation or translation inhibition (76, 80). Post-transcriptional regulation of mRNA transcripts by miRNAs allows for tighter control of neuronal transcriptional output and adds further complexity to the epigenome (81).

miRNAs have been extensively studied in neuroscience, particularly within the field of alcohol and drug addiction (73, 75). Early mechanistic studies revealed a role for specific miRNAs such as miR-212 in the nucleus accumbens (NAc) and striatum during cocaine exposure, and these miRNAs appear to target critical pathways involved in drug addiction such as the CREB pathway and BDNF signaling (82, 83). A recent study used an in-silico approach to identify miR-495 in the NAc as a regulator of cocaine seeking behavior though targeting of synaptic plasticity-related genes including Bdnf and Arc (84). Interestingly, DNA methylation at the miR-124 promoter and subsequent miR-124 expression in microglia is regulated by cocaine exposure (85), and further work has shown that miR-124 regulates the epigenetic enzyme poly [ADP-ribose] polymerase 1 (Parp1) (86). miRNAs also play significant roles in opioid response and addiction (87), such as downregulating the μ-opioid receptor in the hippocampus following repeated morphine and fentanyl exposure (88). miR-218 is decreased in the NAc following repeated heroin administration, and lentiviral overexpression of miR-218 in this brain region inhibits heroin self-administration at least partially via targeting of the epigenetic effector MeCP2 (89).

miRNAs also serve to regulate the response to alcohol in the brain and development of alcohol addiction (90). Early reports in animal models showed that miR-9 regulates specific splice variants of BK channels to drive neuroadaptation to alcohol (91), and miRNAs alter neural progenitor cell proliferation in a model of fetal alcohol spectrum disorders (92). Similar to cocaine addiction, miRNAs involved in repeated alcohol exposure and dependence target CREB and BDNF signaling pathways. miR-206 is upregulated in the medial prefrontal cortex (mPFC) of alcohol dependent rats and targets BDNF for degradation (93). The reduction in Bdnf expression in mPFC during ethanol exposure also appears to be regulated by increases in miR-30a-5p levels, and this mechanism plays a role in the transition from moderate to excessive drinking in mice (94). BDNF-targeting miRNAs (miR-30a, miR-195, miR-191 and miR-206) are increased in peripheral blood in rats with rapid escalation of alcohol drinking (95). miR-494 is decreased in the rat amygdala and participates in the anxiolytic effects of acute alcohol exposure by leading to an upregulation of CBP and other CREB-related genes and subsequent transcriptional activation via increased histone acetylation (96). We showed that miR-137 negatively regulates LSD1 to drive altered histone methylation, increased anxiety-like behaviors, and increased alcohol consumption in adulthood following adolescent alcohol exposure (97). Using a combination of human postmortem brain tissue and rat hippocampal slice culture, Coleman and colleagues revealed a role for microglial-derived let-7 in alcohol-induced neuroinflammation via targeting of toll-like receptor 7 (TLR7) and other inflammatory processes (98). miRNAs also regulate addiction-relevant molecular phenotypes at the synapse, as inhibiting miR-411 in the prefrontal cortex of female mice with a history of alcohol consumption reduced further alcohol drinking and preference (99).

miRNAs have been shown in preclinical studies to contribute to the pathogenesis of schizophrenia. For example, carriers of the 22q11.2 microdeletion are at increased risk for schizophrenia (100, 101), and mouse models of this microdeletion demonstrate decreased expression of miR-185 and the RNA-binding protein Dgcr8 leading to altered dendritic spine development (102, 103). miR-382–3p and miR-674–3p are decreased after 22q11.2 microdeletion, and this leads to increased ventricular size by a dopamine D1 receptor (Drd1)-dependent mechanism (104). miR-219 is downregulated in a pharmacological model of NMDA receptor hypofunction relevant to schizophrenia and targets downstream calcium signaling via calcium/calmodulin-dependent protein kinase II (CaMKII) (105). Taken together, these emerging results suggest that miRNAs regulate addiction- and schizophrenia-relevant behaviors by targeting epigenetic processes, synaptic signaling, and cell-cell interactions in the brain (Supplementary Tables 1&2).

The potential of miRNAs as circulating biomarkers of brain disease has been met with considerable optimism but has yet to substantially alter clinical practice (75). A small study found that subjects suffering from heroin use disorder showed increased let-7b-5p, miR-206 and miR-486–5p in the periphery, while those with methamphetamine use disorder showed increased miR-9–3p (106). miR-124 and miR-181 are increased in peripheral blood of females with cocaine use disorder (107). A group of 17 downregulated miRNAs stemming from a single imprinted locus were identified from peripheral blood mononuclear cells of schizophrenia patients (108). An additional study showed that miR-130b and miR-193a-3p are increased in plasma from subjects with schizophrenia (109). miR-9 is decreased in neural progenitor cells derived from schizophrenia patients (110). The number of miRNAs identified in these studies varies, and miRNA-mRNA interactions in converging biological pathways have been used as a method to identify rational treatment targets. For example, one of the top GWAS hits in schizophrenia patients is near the MIR137 gene (10), and miR-137 is known to affect genes in schizophrenia-relevant pathways including chromatin modifiers and synaptic proteins (111–113). Similarly, miRNA and mRNA expression data from the dorsolateral prefrontal cortex of schizophrenia subjects identified glial cell differentiation and synaptic genes as convergent downstream pathways affected by multiple differentially expressed miRNAs (114), and multiple studies have shown that DICER1, a mediator of miRNA biogenesis, is increased in this same brain region (115, 116).

A proposed ncRNA discovery pipeline for neuroscience

A vital future direction for neuroscience is the identification of transcripts, both coding and noncoding, that alter critical neuronal processes and confer risk for neuropsychiatric disease. As ncRNAs are only beginning to be widely studied, we propose here a conceptual and experimental framework for exploring relevant ncRNA biology (Figure 4).

Figure 4. A ncRNA discovery pipeline for neuroscience.

A proposed systematic approach that can be used in identifying and validating ncRNAs involved in neuropsychiatric diseases.

The first step is identification of ncRNA transcripts that may alter brain function or be involved in genetic risk for disease. Genome wide association studies (GWAS) have been utilized in psychiatry for decades but have only recently reached the required power to detect small effect sizes observed in many common variants typical of genetic risk structure (4, 10, 117). GWAS have implicated ncRNAs in the pathogenesis of brain diseases, such as the association of MIR137 with schizophrenia (10). Many of these loci identified in GWAS are in poorly understood ncRNA-containing regions or other genomic sites thought to contribute to transcriptional regulation (4), and future studies should continue to combine GWAS with epigenomic techniques to stratify risk variants and determine their functional significance (118–121). A recent study analyzed the UK Biobank cohort and found that allelic variation at an enhancer linked to the GAL gene is linked to alcohol intake (43). The authors then used CRISPR to knockout this enhancer in mice and found decreased alcohol consumption and anxiety-like behavior in males, mirroring results in a human discovery cohort (43). RNA-seq has been used in animal models such as cocaine place preference (66) to discover novel ncRNA transcripts associated with addiction pathogenesis. Transcriptome-level alterations in miRNA and lncRNA have additionally been observed in preclinical models of alcohol exposure (122–126), and in human postmortem brain tissue from AUD subjects (64, 127–129). Novel cell-sorting and genomic methods have significantly expanded the ncRNA discovery toolbox. For example, using laser capture microdissection of specific cell types followed by sequencing identified genetic variants associated with schizophrenia and addiction preferentially located in transcribed noncoding regions (9). Moving forward, neuroscientists should continue to utilize emerging technologies in both patients with neuropsychiatric disease and animal models (130, 131) to stratify ncRNAs for further investigation.

The second step is the functional validation of ncRNAs that are discovered and potentially involved in the pathogenesis of brain diseases. For example, ncRNA expression levels can be validated in vitro using technologies such as 5’ RNA-seq (132) (e.g., GRO-seq and CAGE-seq), and the examination of existing datasets can determine the regional specificity of ncRNA expression (35). In vitro model systems can prove especially useful as this stage, as evidenced by the discovery of the stimulus-specific induction of specific eRNA signatures near the Fos gene (28). The combination of in vivo functional characterization and in vitro inhibition of the Bdnf-as transcript in both neuronal culture and in mouse brain was critical in demonstrating the role of this lncRNA in the inhibition of neuronal proliferation and survival (54). Additionally, CRISPR-Cas9 editing and activation/inhibition (CRISPRa/CRISPRi) can modulate expression and function of ncRNA constructs. This has proven useful specifically in linking enhancers and eRNAs to their target promoters and mRNAs (133). An interesting fusion of the discovery and functional validation steps is the utilization of genome-scale screening technologies such as multiplexed CRISPRa of ncRNA regions (134) and subsequent measurement of downstream transcriptomic profiles. Recently, Carullo and colleagues took the novel approach of fusing specific eRNA sequences to dCas9, thereby allowing the researchers to investigate direct eRNA-induced gene regulation (133). Aside from probing the possible association of ncRNA expression with downstream molecular pathways involved in neuropsychiatric disorders, functional validation of ncRNAs additionally increases our overall knowledge of ncRNA biogenesis and functionality.

The third step in our proposed pipeline is the use of preclinical models to evaluate the effects of altered ncRNA expression. While targeting these transcripts in awake, behaving animals poses challenges, the technological advances mentioned above such as CRISPRa/CRISPRi have been successfully used to study ncRNA function in the intact rodent brain (135, 136). For instance, a transcriptomic approach linked the lncRNA Neat1 to epigenetic transcriptional repression, and overexpression of Neat1 in mouse hippocampus leads to impaired memory formation (135). MK-801, an NMDA receptor antagonist used to mimic symptoms of schizophrenia, leads to decreased miR-219 expression and downstream alteration of calcium signaling that is reversed by antipsychotic administration (104). Interrogation of these transcripts in animal models and downstream analysis of molecular pathways altered by ncRNA expression will be critical in prioritizing ncRNAs to investigate in clinical populations.

Lastly, we propose that specific ncRNA transcripts with high likelihood of involvement in neuropsychiatric disease based on previous studies should be investigated as possible diagnostic and/or prognostic markers in clinical populations. Although measurement of ncRNAs in the periphery and cerebrospinal fluid may not contribute to diagnosis or treatment in the immediate future, determining expression levels based on preclinical prioritization can provide insight into brain pathophysiology. For example, multiplexed measurement of miRNAs may be useful in diagnosing and risk-stratifying individuals with schizophrenia (137, 138). Given the alteration of the miRNA biogenesis machinery in schizophrenia (116), measuring members of this pathway in peripheral tissues could serve as a noninvasive biomarker (115). Taken together, a pipeline designed to translate prioritized candidate ncRNAs has the potential to advance our understanding of the brain and generate new treatment targets for psychiatric disease.

Conclusions

Here, we have outlined recent advances in our knowledge of ncRNAs related to brain function in normal and pathological states. eRNAs, lncRNAs, and miRNAs contribute to the maintenance of synaptic plasticity and other neuronal processes and are modulated in schizophrenia and addiction. The continued systematic investigation of specific ncRNAs using emerging techniques is likely to lead to further insights into ncRNA biology and will continue to unravel the complexity of the brain and the pathophysiology of neuropsychiatric disorders.