A Comprehensive Review on the Role of Non-Coding RNAs in the Pathophysiology of Bipolar Disorder

Aim: Bipolar disorder is a multifactorial disorder being linked with dysregulation of several genes. Among the recently acknowledged factors in the pathophysiology of bipolar disorder are non-coding RNAs (ncRNAs). Methods: We searched PubMed and Google Scholar databases to find studies that assessed the expression profile of miRNAs, lncRNAs and circRNAs in bipolar disorder. Results: Dysregulated ncRNAs in bipolar patients have been enriched in several neuron-related pathways such as GABAergic and glutamatergic synapses, morphine addiction pathway and redox modulation. Conclusion: Altered expression of these transcripts in bipolar disorder provides clues for identification of the pathogenesis of this disorder and design of targeted therapies for the treatment of patients.


Introduction
Bipolar disorder (BD) is a multifactorial disorder characterized by the occurrence of severe mood impairment episodes, neuropsychological complications, immunological alterations, and perturbation in personal/social functions [1]. As one of the main sources of disability all over the world [2], BD is associated with premature death from the coexistence of other medical conditions as well as suicide attempts [3,4]. Several genetic and environmental parameters have been recognized to modulate the risk of BD, yet most of them being liked with a number of other mental disorders as well. The causal link between a few of these risk factors and BD has been established [5]. Among the recently acknowledged factors in the pathophysiology of BD are non-coding RNAs (ncRNAs) [6]. These transcripts participate in the epigenetic marking of several genes through modulating chromatin configuration and RNA editing. Their binding with complementary sequences in the genome might alter methylation or RNA sites. Moreover, long ncRNAs (lncRNAs) are implicated in the complicated regulatory systems that control the expression of target genes [7]. Transcript profiling in autopsy samples of medial frontal gyrus from bipolar patients and non-psychiatric controls have shown differential expression of ten lncRNA transcripts and a global higher number of alternative spliced variants in these patients [6]. Other types of ncRNAs such as microRNAs (miRNAs) and circular RNAs (circRNAs) have also been dysregulated in brain tissues or peripheral blood of patients with BD [6,8]. We performed a comprehensive search in PubMed and Google Scholar databases to find studies that assessed expression profile of miRNAs, lncRNAs and circRNAs in BD. This study is study is a narrative review and studies have been selected and discussed based on preference/choices from the authors. We included studies that reported dysregulation of ncRNAs between BD patients and normal controls. We also included data regarding the expression profile of targets of ncRNAs whenever this data was provided in the original articles.

CircRNAs and BD
CircRNAs have a circular secondary structure. This structure is formed by the rearsplicing of a single-stranded linear RNA and the creation of a covalent link. These procedures lead to the formation of an encircled non-polyadenylated RNA structure. Compared with linear RNAs, circRNAs have higher stability resulting from a lack of accessible ends for exoribonucleases. They are about 0.1-10% of linear transcripts in eukaryotes [9]. CircRNAs participate in brain development and integrity of neurons [10,11]. A number of studies have demonstrated the release of brain-specific circRNAs into the peripheral blood in the course of neurological disease, potentiating these transcripts as biomarkers for showing disease progression or response to therapies [9]. Few studies have assessed the expression of circRNAs in brain samples of patients with BD obtained through autopsy. Luykx et al. have shown up-regulation of two circRNAs in brain tissues of bipolar patients. These transcripts were originated from the NEBL and EPHA3 loci, respectively [6]. The latter locus is involved in the development of CNS. Eph receptors of the protein-tyrosine kinase family participate in the production of neurotransmitters, construction of dendritic spines, and synaptic and postsynaptic events [12]. Moreover, they contribute to memory-associated functions [13] and modulation of anxiety [14], two functions that are disturbed in bipolar patients. Authors have suggested circRNA molecules as possible markers for the diagnostic assessment of patients with BD [6]. Zimmerman et al. have recently shown expression of circHomer1a, a neuron-associated circRNA in the frontal cortex. Notably, expression of this circRNA was decreased in both the prefrontal cortex (PFC) and induced pluripotent stem cell-originated neurons of patients with BD. CircHomer1a has been shown to regulate the expression of several splicing variants of genes participating in synaptic plasticity and psychiatric disorders. Thus, circHomer1a modulates synaptic gene activity and intellectual flexibility [15]. Tables 1 and 2 summarize the results of studies that assessed the expression of circRNAs in BD.

LncRNAs and BD
Dysregulation of numerous lncRNAs has been described in peripheral blood and brain tissues of patients with BD. Hu et al. have profiled transcriptome in post-mortem brain tissues of patients with schizophrenia and BDs as well as healthy subjects. They reported differential expression of several long intergenic RNAs in various brain regions of bipolar patients. They showed that these lncRNAs have brain region-specific signatures and are mostly enriched in some pathways including immune system development and oligodendrocyte differentiation. Altered expression of these lncRNAs in patients was explained by modification of DNA methylation alteration [16]. Ji et al. have reported up-regulation of the XIST gene, the principal regulator of X chromosome inactivation (XCI) in the lymphoblastoid cells and brain tissues of female subjects with either BD or major depressive disorder. This up-regulation was accompanied by over-expression of the XCI escapee gene KDM5C. Authors have suggested that up-regulation of XIST might cause or result from delicate changes in XCI [17]. Table 3 summarizes the studies which reported up-regulation of lncRNAs in bipolar patients.
Ghafelehbashi et al. have assessed the expression of IFNG-AS1 lncRNA, and IFNG and IL-1B mRNAs in peripheral blood of BD patients compared with healthy subjects. They reported down-regulation of IFNG-AS1 in patients and its correlation with IFNG expression. Moreover, expression of IL-1B was decreased in patients compared with controls. Thus, inflammatory lncRNAs might participate in the pathogenesis of BD [21]. Hu et al. have reported down-regulation of ENSG00000228794 in patients with BD. This lncRNA resides in a genomic region that is linked with BD. ENSG00000228794 is possibly implicated in calcium ion transport, thus it can modulate synaptic plasticity [16]. Table 4 summarizes the results of studies that reported down-regulation of lncRNAs in BD.

miRNAs and BD
The expression profile of miRNAs has been vastly assessed in different biological sources of patients with BD including whole blood, lymphoblastoid cell lines, brain tissues or extracellular vesicles. Squassina et al. have assessed miRNAs signature in lymphoblastoid cell lines from bipolar patients who deceased by suicide and those with low risk of suicide. They reported higher miR-4286 levels while lower miR-186-5p in lymphoblastoid cell lines obtained from suicide attempters compared with the low-risk group and healthy controls. Conversely, expression of miR-4286 was reduced in postmortem brains of bipolar patients who attempted suicide compared with controls, yet it could not yield the level of significance. Exposure of human neural progenitor cells with lithium down-regulates expression of miR-4286 [23]. Lee et al. have reported abnormal expression of a number of miRNAs in the serum samples of bipolar patients. Among dysregulated miRNAs has been miR-7-5p which was up-regulated in bipolar patients [8]. Notably, miR-7 has been previously shown to suppress the healing of damaged peripheral nerves by altering the migration and proliferation of neural stem cells [8]. Moreover, this miRNA has been over-expressed in the neocortex of superior temporal lobes of patients with Alzheimer's disease [24]. Choi et al. have extracted extracellular vesicles (EVs) from the anterior cingulate cortex. They reported over-expression of miR-149 in bipolar patients compared to controls. They also validated dysregulation of both miRNAs in EVs extracted from brains of an animal model of depressive-like manners [25]. Figure 1 shows the molecular mechanisms of participation of miR-34a in the pathogenesis of BD.

miRNAs and BD
The expression profile of miRNAs has been vastly assessed in different biological sources of patients with BD including whole blood, lymphoblastoid cell lines, brain tissues or extracellular vesicles. Squassina et al. have assessed miRNAs signature in lymphoblastoid cell lines from bipolar patients who deceased by suicide and those with low risk of suicide. They reported higher miR-4286 levels while lower miR-186-5p in lymphoblastoid cell lines obtained from suicide attempters compared with the low-risk group and healthy controls. Conversely, expression of miR-4286 was reduced in postmortem brains of bipolar patients who attempted suicide compared with controls, yet it could not yield the level of significance. Exposure of human neural progenitor cells with lithium down-regulates expression of miR-4286 [23]. Lee et al. have reported abnormal expression of a number of miRNAs in the serum samples of bipolar patients. Among dysregulated miRNAs has been miR-7-5p which was up-regulated in bipolar patients [8]. Notably, miR-7 has been previously shown to suppress the healing of damaged peripheral nerves by altering the migration and proliferation of neural stem cells [8]. Moreover, this miRNA has been over-expressed in the neocortex of superior temporal lobes of patients with Alzheimer's disease [24]. Choi et al. have extracted extracellular vesicles (EVs) from the anterior cingulate cortex. They reported over-expression of miR-149 in bipolar patients compared to controls. They also validated dysregulation of both miRNAs in EVs extracted from brains of an animal model of depressive-like manners [25]. Figure 1 shows the molecular mechanisms of participation of miR-34a in the pathogenesis of BD. Expression of this protein is correlated with CYLD levels [26,27]. CYLD is a deubiquitinase that targets PSD-95. The latter protein participates in the maturation and function of synapses [28]. Table 5. summarizes the studies which reported up-regulation of miRNAs in BD. Expression of miR-34a is increased in BD. This miRNA binds with 3' UTR of SHAK3 to decrease its expression. Expression of this protein is correlated with CYLD levels [26,27]. CYLD is a deubiquitinase that targets PSD-95. The latter protein participates in the maturation and function of synapses [28]. Table 5 summarizes the studies which reported up-regulation of miRNAs in BD.
High throughput analysis miR-193a-3p was upregulated in both BD and SZ. [30] miR-330-3p ---p < 0.05 High throughput analysis miR-330-3p has been over-expressed in the blood of subjects with BD and monopolar depression.
High throughput analysis miR-28a-3p is in the same family as miR-708, a miRNA that is associated with risk of BD.

Discussion
Several studies have reported aberrant expression of ncRNAs in bipolar patients. Moreover, the expression of ncRNAs is influenced by drugs used for these patients. For instance, a combination of drugs including lithium, valproate, lamotrigine and quetiapine has been shown to alter the expression of several genes including miRNAs in cultured human neurons. Among the differentially expressed genes have been miR-128 and miR-378 whose targets are enriched in neuron projection development and axonogenesis [53]. Thus, ncRNAs not only are involved in the pathogenesis of BD but also they might participate in the determination of response to prescribed drugs. NcRNA profiling has revealed specific alterations in certain lncRNAs and miRNAs in the manic state indicating a possible role of these transcripts in the determination of disease status [18]. Notably, lncRNAs have been the largest group of differentially expressed ncRNAs [18]. Such state-specific transcript signature potentiates ncRNAs as preferable biomarkers for early diagnosis of BD.
Next generation sequencing technique has facilitated the identification of putative biomarkers for discrimination of bipolar patients from healthy subjects. A representative of this kind of experiment is the study conducted by Lee et al. which identified overexpression of miR-7-5p, miR-23b-3p, miR-142-3p, miR-221-5p, and miR-370-3p in bipolar patients compared with healthy individuals. The diagnostic accuracy of this panel of miRNAs was estimated to be 0.907 [8].
Dysregulated ncRNAs in bipolar patients have been enriched in several neuron-related pathways such as GABAergic and glutamatergic synapses, morphine addiction pathway, redox modulation as well as TGF-β, Wnt, Akt/PI3K, Hippo and FoxO pathways. Significance of a number of these pathways such as GABAergic and glutamatergic synapses signaling and TGF-β, Hippo and FoxO pathways have been recognized in the pathogenesis of BD [54,55]. The relevance of other pathways with this disorder should be appraised in future studies. Another functional annotation analysis of the differentially expressed coding and non-coding genes between patients with BD and healthy controls has shown remarkable enrichments of cellular pathways associated with angiogenesis and vascular system evolution [6]. The largest GWAS conducted in BD has reported that the most significant loci have been related to ion channels, neurotransmitter transporters and synaptic components. Yet, this study has not reported any indication for involvement of angiogenesis or vascular related loci in BD [56]. Pathway analysis revealed nine significantly enriched gene sets, including regulation of insulin secretion, circadian rhythm, and endocannabinoid signaling [56]. Notably, insulin resistance signaling pathway and circadian rhythm have been among the related pathways with dysregulated miRNAs in BD [8,23]. Finally, top genes existing in these pathways have been shown to encode Ca2+ and K+ channel subunits, MAPK and GABA-A receptor subunits [56], the latter being recognized as one of the most important pathways enriched among dysregulated ncRNAs in BD [8].
However, different studies have indicated abnormal activity of various signaling pathways in BD including immune response pathways [57], neuroplasticity, circadian rhythms and GTPase binding [58] and G protein-receptor dysregulation [59]. Such a heterogeneous range of biological pathways involved in BD might be related to distinct brain areas assessed in these investigations. Imminent investigations integrating particularly large sample sizes of patients with BD and comparison of transcriptome of coding and ncRNAs in different parts of the brain are required to find the most relevant pathways.
Dysregulation of ncRNAs has been reported in other brain disorders as well. For instance, assessment of lncRNA signature using high-throughput sequencing has led to the recognition of aberrantly expressed lncRNAs in acute ischemic stroke. ENSG00000226482 has been among up-regulated lncRNAs. This lncRNA has a potential role in activation of the adipocytokine signaling [60]. Moreover, another experiment in the animal model of blast traumatic brain injury has shown elevation of plasma amounts of a brain-enriched miRNA, namely miR-127. This study has concluded that levels of sphingolipids, miR-128, and the let-7 family can show the presence of could blast traumatic brain injury. Moreover, a number of other miRNAs have been shown to serve as markers for a global level of damage after blast injury [61]. Moreover, miRNAs have been shown to serve as diagnostic markers for cognitive impairment. Certain panels of miRNAs have high sensitivity and specificity values in this regard [62].
Notably, several dysregulated ncRNAs in BD, are also dysregulated in other neuropsychiatric conditions such as schizophrenia or Alzheimer's disease. Although this observation supports their potential roles in synaptic plasticity or neurodevelopment, it complicates the design of disease-specific diagnostic panels for BD.
Taken together, circRNAs, lncRNAs and miRNAs are regarded as potential contributors in the pathology of BD and putative biomarkers for diagnosis of this disorder. Their participation in the response of patients to the prescribed medications and their potential as therapeutic targets have been less studied. Thus, these research areas should be explored in future studies.
NcRNAs are superior to transcripts of standard genes in the field of biomarker study as they represent the final step of function of the gene. As transcripts of standard genes should be translated to proteins to exert their function, the transcript level might not reflect the final level of the functional molecule. Moreover, miRNAs represent important regulators of gene expression as they can target several transcripts.
Finally, studies reporting dysregulation of ncRNAs in BD have some limitations. For instance, they often suffer from various confounding factors. This is especially true for postmortem brain studies. Moreover, most of the studies reviewed in this manuscript may not have sufficient statistical power due to their small sample sizes. Analysis and interpretation of differences between data of postmortem brains and blood samples, differences between expression data on ncRNA and protein coding genes, and matching with GWAS-identified loci are other research fields that should be explored in future studies.

Conclusions
NcRNAs are potential markers for neurological disorders such as BD. Several ncRNAs have been found to be dysregulated in blood samples of bipolar patients. Molecular studies for identification of the mechanism of dysregulation of these transcripts in bipolar patients would facilitate the development of new therapeutic strategies.  Institutional Review Board Statement: All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards. Informed consent forms were obtained from all study participants. Informed consent forms were obtained from all study participants. The study protocol was approved by the ethical committee of Shahid Beheshti University of Medical Sciences (IR.SBMU.MSP.REC.1399.042, 12/09/2020). All methods were performed in accordance with the relevant guidelines and regulations.
Informed Consent Statement: Informed consent was obtained from all subjects involved in the study.

Data Availability Statement:
The analyzed data sets generated during the study are available from the corresponding author on reasonable request.

Conflicts of Interest:
The authors declare no conflict of interest.