An insight into the sprawling microverse of microRNAs in depression pathophysiology and treatment response

Abstract

Stress-related neuropathologies are pivotal in developing major depressive disorder (MDD) and are often governed by gene-regulatory changes. Being a stress-responsive gene-regulatory factor, microRNAs (miRNAs) have tremendous biomolecular potential to define an altered generegulatory landscape in the MDD brain. MiRNAs’ regulatory roles in the MDD brain are closely aligned with changes in plasticity, neurogenesis, and stress-axis functions. MiRNAs act at the epigenetic interface between stress-induced environmental stimuli and cellular pathologies by triggering large-scale gene expression changes in a highly coordinated fashion. The parallel changes in peripheral circulation may provide an excellent opportunity for miRNA to devise more effective treatment strategies and help explore their potential as biomarkers in treatment response. This review discusses the role of miRNAs as epigenetic modifiers in the etiopathogenesis of MDD. Concurrently, key research is highlighted to show the progress in using miRNAs as predictive biomarkers for treatment response.

Introduction

Major depressive disorder (MDD) is a global concern and one of the leading causes of mental illnesses worldwide. MDD has been described as a key risk factor for suicidal ideation and attempts, with a 4% or more lifetime prevalence rate [1]. Recent data show that the rate of suicide is 20 times higher in MDD patients than in the general population [2]. It is estimated that ~60% of the people who die by suicide are MDD patients [2]. Moreover, half of the MDD patients are resistant to the currently available antidepressants, and about 20% show a refractory response to any intervention [3]. Thus, it is critical to understand the role of neurobiological correlates in MDD pathogenesis for developing new therapeutic approaches and identifying potential strategies to monitor the treatment responses.

The development and progression of MDD are mainly contributed by stress-associated brain changes. Recent research has found that stress, in the form of life adversities received at the early stage, may trigger depressive symptoms in adulthood [4]. Severe stressful life events, whether received at an early or later stage, can be associated with an increased vulnerability to developing MDD. Data suggest that both first and recurrent episodes of depression are preceded by severe stressful life events [5]. Identifying neurobiological factors that are instrumental in the development and progression of stress-induced MDD pathogenesis has become increasingly challenging. Based on our present understanding, dysfunctionality in cellular signaling pathways, ranging from plasticity-related changes to immuno-inflammatory modulations significantly contributes to molecular pathologies in the MDD brain. Previous data from our lab and others have shown neuroinflammatory changes in MDD patients with an increased risk of suicide [6-8]. Abnormalities in pro-inflammatory cytokine levels are often linked to neuropathologies of MDD and suicide brain. For example, irregularities in the levels of interleukin (IL)-1β, IL-6, tumor necrosis factor (TNF)-α, transforming growth factor (TGF)-β1, vascular endothelial growth factor (VEGF), kynurenic acid (KYN), and lower IL-2, IL-4, and interferon (IFN)-γ are predominantly seen in specific brain regions of postmortem brain from MDD and suicide subjects [6]. Earlier studies have also suggested a concurrent change in the stress response system and disruption in inflammatory pathways in the brains of depressed suicide subjects. In these studies, changes were noted not only in inflammatory cytokines but also in B and T-cell functions [9, 10]. Many times the vulnerability to these immunomodulators were found to be associated with complex gene regulatory changes [8, 11]. Now with the advent of nextgeneration sequencing platforms and higher-end gene expression analysis tools, the role of complex gene expression networks in gene regulatory changes has become prominent in the pathogenesis and progression of MDD and increased vulnerability to suicide [7].

Due to the polygenic nature of this disease, the idea behind the complex array of gene regulatory networks and their associated changes to influence multiple signaling pathways has recently been validated in the MDD brain [7]. However, it is yet to be ascertained how the regulatory changes in gene expression induce pathologies in the MDD brain, notably the integration of multiple interactions among various genes. The idea that gene expression regulation happens more often at the upstream level, which later converges at the downstream level to achieve common regulatory pathways, is gaining momentum. This type of regulation is a complex molecular process involving the integration of various inputs from underlying molecular circuitries leading to improper stress response and, consequently, maladaptive changes in the brain [12].

Method

The primary criteria used to search the miRNA-associated reports were based on studies highlighting the role of miRNA in MDD. With the help of pubmed database the search was performed to include original articles from 2012-2022 using various key words to increase the reterival of articles falling under the criteria mentioned previously. The following keywords were used for retrieval: “microRNAs” AND “Major depressive disorder” OR “MDD” OR “Major Depression” AND “Neuroplasticity” OR “Synaptic plasticity” AND “Stress” AND “Treatment” AND “Biomarker.”

Integration of miRNAs to MDD Etiopathogenesis

Recently, the role of miRNAs has gained considerable importance in MDD pathogenesis. MicroRNA is one of the smallest members of the noncoding RNA family inherently programmed to mask coding gene expression at the post-transcriptional level. The gene expression inhibition is the end result of a perfect interaction between miRNAs and target mRNAs through characteristic sequence complementarity. Decade-long research on miRNAs stems from their role as the epigenetic master regulators of many biological functions in the brain. This has set a clear path to articulate their role in various psychiatric disorders, including MDD [13]. According to the miRbase registry, currently, little over 2500 mature miRNA sequences have been cataloged for humans. For initial years, miRNAs were considered modulators of a single protein-coding gene. Recent evidence, however, suggests that a mature miRNA can be a regulatory hub at the center of a complex gene regulatory network. This network could be a ramification of direct and indirect association between hub miRNAs and their target genes [13]. Despite the paucity of structural variation and limited potential to go through the exon splicing procedure and restricted size (~22 nucleotides), this small form factor (here we name microverse) exhibits enormous diversity in targeting a wide array of RNA molecules ranging from protein-coding (mRNA) to long noncoding RNAs (lncRNA) [14, 15]. In the growing vocabulary of epigenetics, miRNAs have quickly taken a spotlight for their persuasive gene regulatory role in disease pathogenesis. Lately, the pathogenic role of miRNAs has been associated with various psychiatric conditions, including MDD [13]. Besides, changes have been found for miRNA expression in peripheral blood as a proxy to the MDD brain [16]. This notable feature of miRNAs has made them a suitable choice as potential biomarkers in MDD diagnosis and treatment response.

Recent understanding of the pathogenic role of miRNAs as epigenetic modifiers has provided a clear understanding of their ability to shape the gene regulatory landscape in the brain's stress axis [17]. Given their ability to target genes both directly and indirectly, miRNAs, as individuals or in unison with other miRNAs, can destabilize a gene regulatory network or redefine a network from a preexisting one [18]. It is known that miRNA-mediated epigenetic perturbations are at the cross-section of stress-induced biological changes in the brain [17]. Due to dynamic and quick responsiveness to environmental stimuli, miRNAs can cause a magnitude of gene regulatory changes at individual or network levels under stressful conditions [13]. In the MDD brain, vulnerability to alter gene ontological functions associated with plasticity and HPA axis activity is due to pervasive changes in miRNA expression [13, 18].

At the molecular level, miRNA regulates the mRNA transcripts by targeting a characteristic sequence element at the 3’ untranslated region (UTR) called miRNA Response Elements (MRE). The complementary binding between the MRE and miRNA seed region may lead to several regulatory reactions. Overall, this can directly reduce mRNA's transcriptional stability and hence, their expression [19]. Molecular sponging is another mechanism miRNAs apply via lncRNAs to achieve their tight gene regulatory activity. lncRNAs act as a molecular decoy to prevent the binding of miRNAs with their targets [15]. Overall, the data on miRNA changes in MDD brains from clinical and preclinical settings is compelling. A few recent studies that have determined MDD-associated miRNA expression changes at the miRNome level have shown a high degree of epigenetic interference at the gene expression level [20, 21].

Programmed Biogenesis of miRNAs and their Complex Regulatory Role in MDD Pathogenesis

Mammalian miRNA biogenesis is a synchronized molecular process in generating mature miRNAs, which are only 22 nts in length (Figure 1). The process starts in the nucleus with the transcription of primary miRNAs (pri-miRNA) by RNA Polymerase II or III. The genomic location of the long pri-miRNA transcriptional unit could be intra- or intergenic. After transcription, while still in nucleus, pri-miRNAs become the target of endonucleolytic cleavage by RNase III endonuclease Drosha. Drosha is a double-stranded (ds) RNA-specific ribonuclease that can create staggered cuts on the RNA helix. The products of the enzymatic cleavage are small hairpin stem-loop structure (60-100 nts long) called precursor miRNAs (pre-miRNAs) [22]). It is interesting to note that the enzymatic process cannot be done by Drosha alone unless supported by an ancillary protein called DiGeorge syndrome critical region 8 (DGCR8). Together, they form a microprocessor complex to help remove the flanking segments of the stem-loop structure and further crop them into shorter pre-miRNAs. Additionally, the cleavage facilitates the generation of 3’ protruding termini, which ultimately help in the transportation of pre-miRNAs to cytosol with the help of exportin-5 protein. The protruded 3’termini of premiRNAs further help their recognition by another RNase III enzyme, Dicer, in the cytoplasm. With the help of co-factors TRBP (TAR RNA-binding protein) and or PACT (PKR-activating protein), Dicer binds (PAZ domain assisted) to the 2 nucleotides of 3’overhang of doublestranded pre-miRNAs to form 22 nts miRNA/miRNA* duplexes. Finally, the effector molecule, also known as RNA-induced silencing complex (RISC), is formed once one of the strands from miRNA/miRNA* duplexes is loaded onto the Argonaute (Ago) protein [23].

Figure 1: Programmed biogenesis of miRNAs.

A stepwise description of mature miRNA biogenesis has been presented with a schematic diagram. The diagram shows the involvement of various ancillary protein factors and processing enzymes to support the transcription of primary miRNA transcripts and their subsequent processing through precursor miRNA to mature form. MicroRNA biogenesis following the canonical pathway begins in the nucleus. Successful transcription of primary miRNA (pri-miR) depends on the RNA polymerase II machinery in the nuclear environment. Following a canonical path, the pri-miRNA is processed by the nuclear RNase III enzyme Drosha to produce precursor miRNA (pre-miRNA). Pre-miRNA is transported to cytosolic environment with the help of Ran-GTP and Exportin5 transporter complex. In cytosol the pre-miRNA is processed by the RNase III enzyme Dicer to generate mature miRNAs. Mature miRNAs are incorporated into the RNA-induced silencing (RISC) complex, which regulates gene expression by pairing primarily to the 3′ untranslated region of protein-coding mRNAs to repress target mRNA expression. The arrest in target expression can be caused by any one of the three mechanisms: i) translational blockage by inhibiting the access of elongation factor/ribosome complex; ii) transcript degradation by decapping and exonuclease activity, and iii) transcript degradation by deadenylation. The figure was prepared using software from Biorender.com.

As members of the RNase III family, Drosha and Dicer play a significant role in miRNA processing. Interestingly, animals share some commonality with plants and fungi in miRNA processing mediated by RNase III. It is worthwhile to mention that the ancestral role of RNase III in bacteria was restricted to processing ribosomal RNA (rRNAs). However, in higher metazoans, including mammals, this role has evolved to include a more specialized function to process mi- and si-RNAs (small interfering RNA) [24].

As mentioned earlier, the microprocessor complex formed by Drosha and DGCR8 plays a critical role in cleaving the pri-miRNAs. The three-dimensional structural analysis reveals that the site-specific cleavage of pri-miRNA following the recognition by microprocessor complex is activated by the higher order hairpin loop structure of pri-miRNA transcript. The basal and apical junctions of the pri-miRNA hairpin loop determine the binding and cleavage specificity of the microprocessor complex. Empirical evidence shows that conserved sequence motifs at both the apical loop and basal single-stranded overhang are essential for the enzymatic activity of the microprocessor complex. The protein crystal structure of the microprocessor complex reveals that the cleavage sites are located 11 base pair (bp) from basal and 22 bp from the apical junction of pri-miRNA helical stem-loop. Upon recognition by the two RNase III binding domains, Drosha cleaves off the phosphodiester bonds on two helices of double-stranded pri-miRNA. Post-nuclear processing of pre-miRNAs is facilitated by the Dicer in the cytoplasm. In doing so, Dicer recognizes pre-miRNAs with a certain degree of affinity towards 5′ phosphate and 3′ overhang at the base of the miRNA stem-loop [25]. The successful interaction between Dicer and the double-stranded pre-miRNA molecule defines their loading to RISC. The mature miRNA incorporated in the RISC identifies a short sequence based on 3' untranslated region (3'UTR) of target genes. The interaction between the two complementary sequences between the miRNA seed region and targeted 3’UTR may result in imperfect binding and can trigger various biomolecular reactions. Most of the time, the imperfect binding based on seed region complementarity (2-7 nucleotides) leverages a single miRNA to target several mRNA transcripts and helps to achieve a divergent gene regulatory network central to any molecular pathways. Other mechanisms to cause gene expression inhibition by miRNA involve an interaction of RISC with two other components (60S ribosome and eIF6) from eukaryotic translational initiation machinery. The impact on 60S ribosome and eIF6 can significantly impede the polysome assembly, leaving the nascent transcript vulnerable to various nucleolytic activities. Due to polysome disassembly, target transcript destabilization can also be persuaded via deadenylation of poly A tail [25, 26].

Besides masking the target transcript expression at the post-transcriptional level, miRNAs can also achieve a complex level of gene expression regulation via targeting various transcription factors (TFs). This non-canonical mode of regulation can result in a delicate balance of gene transcription due to altered promoter occupancy of targeted TFs. Thus, it is reasonable to say that any miRNA expression changes in the brain may result in homeostatic imbalances due to the destabilization of the gene regulatory network, which is central to MDD pathogenesis [26, 27].

Another exciting aspect of miRNA-mediated expression regulation underlies their ability to regulate their own transcriptional fate via a feedback loop [28]. To achieve expression homeostasis, miRNAs control the cellular availability of specific TFs directly or indirectly. In Figure 2, we have presented one example to highlight the key steps governing the feedback mechanism for miRNA expression stabilization and destabilization process. In the nucleus, promoter-specific miRNA transcription can happen independently from a genomic locus with the recruitment of RNA polymerase II (RNA PolII), specific TF, and other transcriptional activators (Figure 2A). In Figure 2B, miRNA seed sequence-specific target scanning has been shown, followed by the recruitment of RISC. Ago2 is one of the major components in RISC and helps drive the gene silencing process following miRNA-dependent target binding near the 3’UTR. In this instance, the target could be a specific TFs responsible for miRNA transcriptional regulation. The cellular fate of the complex enzymatic interaction leads to the destabilization of target transcripts. Figure 2C is a schematic drawing of the 4W5N (from the protein database, PDB), a crystalized structure of Ago2 locking miRNA-target transcript complex near the PAZ (Piwi/Argonaute/Zwille) domain. The tight interaction in the proximity of the PAZ domain determines the catabolic fate of the target transcripts or the specific TFs following their deadenylation. This helps achieve a balance for further regulation of miRNA transcription in the nucleus [29].

Figure 2: Conceptualizing of miRNA-target feedback loop.

The diagram collectively describes the framework of miRNAs being regulated by TFs which may in turn, be recruited as direct miRNA targets. This dynamic process provides tight regulatory control and helps to achieve a bipartite relationship between miRNA and its’s cognate target(s). A) The diagram shows one such example highlighting key steps that govern the feedback mechanism for miRNA expression stabilization and destabilization process. In the nucleus, promoter-specific miRNA transcription can happen independently from a genomic locus with the recruitment of RNA polymerase II (RNA PolII), specific TF, and other transcriptional activators. B) miRNA seed sequence-specific target scanning has been shown, followed by RNA Induced Silencing Complex (RISC) recruitment. The Ago2 is one of the major components in RISC and helps drive the gene silencing process following miRNA-dependent target binding near the 3’UTR (untranslated region). In this instance, the target could be a specific set of TFs responsible for miRNA transcriptional regulation. The cellular fate of the complex enzymatic interaction leads to the destabilization of target transcripts. C) The schematic is about a 4W5N (from the protein database, PDB), which is a crystalized structure of Ago2 locking miRNA-target transcript complex near the PAZ (Piwi/Argonaute/Zwille) domain. The tight interaction in the proximity of PAZ domain determines the catabolic fate of the target transcripts or the specific TFs following their deadenylation and helps achieve a delicate balance for further regulating the miRNA transcription in the nucleus. The figure was prepared using software from Biorender.com.

Outlining the Role of miRNA in the Brain

Like any coding gene, a programmed spatio-temporal regulation is key to achieving diverse miRNA expression changes across different tissue types in the human body. The brain does that job exceptionally well, and any significant perturbation in that regulatory process may mark the onset of neuropsychiatric disorders, including MDD [30]. Even compartmentalized distribution of miRNAs can be accountable for MDD pathogenesis [18]. MiR-128, miR-134, and miR-132 are examples of this type of compartmentalization where their redistribution has been shown between synaptic spaces and neuronal soma [31]. The localized distribution of miRNAs in synaptic and/or dendritic areas is generally associated with activity-dependent neuronal response in the brain. These activity-dependent changes happen to perform specific functions and are regulated mainly by intra- and extra-synaptic environmental cues [32]. Additional complexities are also noted regarding genetic loci encoding for miRNAs. For example, coding units for brainenriched miRNAs such as miR-124, miR-7, miR-9, miR-128, miR-129, miR-133a, miR-138, miR-153, and miR-218 have been found on different chromosomes or the same chromosome but on different arms [26]. Transcribing the same miRNA from two separate anatomical loci may seem redundant but could be part of a complex gene regulatory process controlled by specific neuronal inputs. This kind of tight regulation is attributed to neuromodulators that regulate the expression of miRNAs in order to achieve a higher-order gene regulatory network for complex behavioral manifestations [16]. Our lab recently reported a similar regulatory dependency for miR-124-3p. The expression of miR-124-3p was tested in the prefrontal cortex of rats modeling stress-induced depression phenotype. Under corticosterone treatment, the transcription of miR-124 located on chromosome 3 was increased, whereas the other two loci, positioned on chromosomes 2 and 15, did not show transcriptional changes. In addition, the promoter region on chromosome 3, proximal to the miR-124 transcriptional unit, showed a discernable change in promoter methylation in parallel to behavioral changes. Together, these data suggest precise epigenetic regulation of miRNA expression orchestrated by changes in underlying neural circuitry [27, 33].

Desynchronized miRNA expression is often the signature of the neuropsychiatric brain. Understanding the coordinated mRNA expression regulation is one way to look at this desynchronization. This is partly associated with the regulatory interdependency of miRNAs localized in clusters. In one of our miRNA studies in locus coeruleus (LC) of the MDD brain, we observed that miRNAs loci that were in close proximity from the same chromosome (e.g., miR-20b-5p, miR-106a-5p, miR-890 on chromosome X, miR- 330-3p, miR-99b-3p on chromosome 19, and miR-409-5p, miR-1197 on chromosome 14) had similar expression changes with same directionality. Similar findings were noted for miRNAs with transcriptional units mapped in the opposite orientation (miR-330-3p and miR-99b-3p on chromosome 19). This intriguing phenomenon is not surprising given the anatomical peculiarity of LC, where a cohesive gene regulatory network is required to maintain a higher level of molecular coordination. Our miRNA data in the LC of MDD subjects supports the idea of transcriptional integration through a highly coordinated regulation process at the molecular level [26]. Thus, miRNA needs to be studied collectively to understand their relative involvement in regulating complex brain functions under pathological changes.

Plasticity Associated miRNAs in MDD

The epigenetic role of miRNAs has been well-appreciated in regulating the processes of neurogenesis, synaptic development, axon guidance, and neuronal plasticity. All these processes are critical in maintaining normal brain functions, and any disruption in these processes may lead to psychopathological abnormalities [7, 21]. In the MDD brain, plasticity-related dysregulation leads to maladaptive responses to stressful stimuli, leading to improper integration of neuronal inputs. One of the hallmarks of the MDD brain is reduced synaptic plasticity due to an array of activity-dependent biomolecular changes, causing disarray in local protein translation. At the same time, it is now becoming clear that miRNAs can be localized in the synaptic compartment with a function to regulate protein translation locally [34]. Therefore, plasticity-related dysfunctioning of the MDD brain can be epigenetically associated with an improper miRNA response [7, 18, 20, 33]. In many instances, miRNA synthesizing and processing machinery participate in miRNA-mediated altered synaptic plasticity [18]. A study in Dicer knock-out mice has demonstrated a deficiency in reconciling synaptic plasticity in the forebrain neurons [35]. Changes in dendritic spines and branches have also been noticed at the morphological level of the Dicer knock-out mice [36]. At the molecular level, the changes could be rooted in the enzymatic deficiency of Dicer endonuclease to process the pre-miRNA molecules. This finding is further supported in DGCR8 haploinsufficient mice [37]. As mentioned in this review, the DGCR8 participates in the microprocessor complex with Drosha to help liberate pre-miRNA from primary hairpin miRNA transcripts. In the mouse model, the DGCR8 genetic deformity is caused by a microdeletion at chromosome 22q11.2. This microdeletion profoundly impacts dendritic development, especially the dendritic spine and tree growth. [37, 38]. Another study showed that an increased miRNA expression in the CA1 region of mouse hippocampi caused decreasing synaptic plasticity, abnormal behavioral manifestation, and changes in long-term potentiation (LTP) and long-term depression (LTD) with significant effects on synaptic excitability [39]. These studies were the first to highlight the role of improper miRNA biogenesis in causing plasticity-related changes in mouse brains and linking it with behavioral deficiencies.

More recently, the role of miRNAs in synaptic plasticity has been greatly appreciated due to the identification of a complete set of miRNA synthesizing machinery in the synaptic compartments of neurons. Drosha and DGCR8 enzymes can be traced back in the synaptic microenvironment along with pri-miRNA fractions. The localization of miRNA synthesizing machinery in the synaptic compartment is a paradigm-shifting phenomenon from their previously known role in synthesizing miRNAs canonically (nuclear/cytosolic microenvironment). The non-canonical miRNA synthesis is a programmed molecular act of directional targeting to achieve a specialized function, such as activity-dependent modulation of gene repertoire meant to regulate synaptic functions. The regulated proteomic output at the synapse could be an epigenetic rescaling of local protein expression by the synaptic pool of miRNAs [40]. In fact, it has been suggested that many psychopathologies are the direct result of abnormal gene expression changes in the synaptic environment [41]. Parallel changes in miRNA expression, as determined by some recent studies, have drawn significant attention to correlating their possible role in activity-dependent gene expression modulation in the synaptic microenvironment [40]. This synaptic modulation of local miRNA repertoire may also help to answer the plasticity-related compromised landscape of neuronal architecture as seen in neuropsychiatric conditions such as MDD. This phenomenon is further supported by recent sequencing data from our lab, which showed the differential distribution of somatodendritic miRNAs has a pivotal role to play in the MDD brain by targeting various neuronal functions related to synaptic plasticity [18]. More specifically, we recently identified the expression of 351 miRNAs in the synaptosome purified from the human prefrontal cortex (PFC) [18]. Of them, 24 were uniquely expressed at the synapse. In addition, miR-215-5p, miR-192-5p, miR-202-5p, miR-19b-3p, miR-423-5p, miR-219a-2-3p, miR-511-5p, and miR-483-5p significantly changed in the synaptic fraction obtained from dorsolateral (dl)PFC of MDD subjects. Cell culture-based confirmation and bioinformatic analysis highlighted their role in synaptic plasticity, nervous system development, and neurogenesis. A shift in expression ratios (synaptic vs. total fraction) of miR-19b-3p, miR-376c-3p, miR-455-3p, and miR-337-3p was also observed in the MDD group. Moreover, the precursor (pre-miR-19b-1, pre-miR-199a-1, and pre-miR-199a-2) miRNAs were found to be inversely correlated with the mature miRNA (miR-19b-3p, miR-199a-3p) expression. Our study provides new insight into synaptic miRNA repertoire and their local regulatory landscape in reshaping synaptic functionality in MDD pathogenesis [18].

Additional studies have also highlighted the role of brain-enriched miRNAs in reconstituting a diverse array of behavioral responses by targeting many genes critical to plasticity. For example, the role of miR-132, miR-134, miR-138, and miR-124 have been demonstrated in regulating various plasticity genes, including ARC, CaMKIIα, LimK1, FMRP, CREB, and BDNF in preclinical and clinical models [42]. The inhibitory effect of miR-134 on LimK1 was related to altered dendritic spine morphology. It was also thought to cause impeded excitatory synaptic transmission at the postsynaptic region [43].

Neurotrophin-regulated role of miRNAs in synaptic plasticity has also been suggested in the hippocampal CA1 brain area of the mouse brain. An induced expression of miR-134 in synaptosome was found to be connected with reduced Sirt1 gene expression in the mouse hippocampus [44]. Interestingly, the Sirt1 gene-regulated miR-134 expression changes caused decreased BDNF and CREB protein expression in hippocampal neurons [44]. Dendritic spine morphology could be affected by several other genes, such as DHX36 and LimK1. Changes in DHX36 expression were negatively related to spine morphogenesis in hippocampal neurons. Interestingly, the expression changes of these genes were under the direct control of miR-134 localized in the dendritic environment [45]. MiR-134-mediated change in DHX36 protein was also associated with altered neurocognitive performance [45].

Neurotrophin-mediated synaptic plasticity is directly related to miR-132. MiR-132 modulates synaptic functions in an activity-dependent manner where CREB acts as an upstream regulator of miR-132 expression with a functional role in hippocampal spinogenesis [46]. Data also suggest that reduced expression of miR-132 in hippocampal neurons occurs due to low CREB activity leading to an inadequate response in memory acquisition and fear conditioning [47]. Diminished expression of miR-132 also causes altered synaptic plasticity and spinogenesis of hippocampal neurons. This expression attenuation of miR-132 and subsequent dysfunctionality associated with synaptic plasticity is directly related to the induced expression of target protein p250GAP, causing inhibition in neuronal outgrowth and sprouting [48]. Analysis at the molecular level also identified BDNF as a responsible gene for regulating miR-132 expression. These results further demonstrate the involvement of the BDNF-ERK-CREB signaling pathway in influencing the synthesis of postsynaptic proteins and neuronal outgrowth via miR-132 [49]. Additional evidence supporting the role of miR-132 in plasticity modulation was examined in an overexpression mouse model. In the miR-132 overexpression model, hippocampal MeCP2 was identified as a direct target known for increasing the dendritic spine density [48-50].

Homeostatic synaptic plasticity (HSP) has an enduring effect on synaptic strength by changing the synaptic AMPARs response and enhancing synaptic transmission. Recently, miR-124 has been shown to modulate HSP by tightly regulating GluA2 expression (AMPA receptor family member). Activity-dependent changes in miR-124 expression can be reduced with selective repression of neuronal activity at the synapse. Conversely, increased miR-124 activity can counterbalance the homeostatic response by inducing calcium-permeable AMPA (CPAMPA) [51]. Another report has suggested that serotonin moderates miR-124 expression in a CREB-dependent manner. This CREB-mediated miR-124 repression enhanced 5HT-dependent long-term facilitation of sensory-motor neurons [52].

Increased G protein α palmitoylation has been reported to trigger dendritic shrinkage and reduced synaptic transmission through the Rho-dependent signaling axis [53]. Studies have pointed out the involvement of miR-138 as a master regulator in these synaptic changes. It has been suggested that miR-138 modulates the Rho-pathway by masking the expression of Acyl Protein Thioesterase1 (APT1), which regulates the palmitoylation-mediated change in Gα protein activity of hippocampal CA1 and DG neurons [31, 54]. In addition, miR-138 also targets decapping mRNA 1B (DCP1B), coding for a local synaptic protein known to regulate plasticity functions in the hippocampus and prefrontal cortex [55]. Interestingly, miR-138 was found to be connected with the Sirtuin1 gene via a regulatory loop. The expression of the Sirtuin1 gene directly contributes to the axonal regeneration of hippocampal dorsal root ganglion neurons. Increased expression of Sirtuin1 acts as an inhibitory factor to reverse the miR-138 expression level [56]. In addition, studies show the epigenetic role of miR-9, miR-125a/b, and miR-188 by regulating a host of genes (REST, FXR1P, CAMKK2-AMPK, PSD-95, Bcl-W, Syn-2, Nrp-2, 2-Ag, and Bace1) which are associated with synaptic plasticity [49]. Lastly, studies highlighting the cooperative role of RNA binding protein FMRP with specific miRNAs (miR-132-3p, miR-125b-5p, miR-138-5p, and miR-124-3p) have demonstrated the augmented inhibition on plasticity-related genes Arc and CaMKIIα in mouse brain [42].

Stress-Associated miRNAs in the MDD Brain

In the introduction section, we have highlighted the role of stress in the psychopathology of MDD. Susceptibility to stress is a major risk factor for developing MDD [27]. Environmental adversities remain a key to stress-associated maladaptive changes in the brain. These maladaptive responses are closely linked with epigenetic modifications of genes in specific brain areas. miRNAs play a critical role in these epigenetic modifications [27]. In the corticosterone-induced rodent model of depression, we noted transcriptome-wide changes in miRNA expression in the PFC, a brain area considered important for glucocorticoid feedback inhibition and HPA axis equilibrium [27]. Chronic corticosterone administration in rats caused significant alterations in 26 miRNAs. Of them, 19 miRNAs were upregulated (let-7i, miR-19b, miR-29c, miR-101a, miR-124, miR-137, miR-153, miR-181a, miR-181c, miR-203, miR-218, miR-324-5p, miR-365, miR-409-5p, miR-582-5p, miR-155, miR-29a, miR-30e, miR-721, miR-699) and 7 were downregulated (miR-146a, miR-200c, miR-351, miR-155, miR-678, miR-764-5p, miR-135a). The global miRNA changes were thought to be responsive to the dysfunctionality of the HPA axis as these miRNAs targeted genes that were ontologically associated with inflammation, synaptic plasticity, cell differentiation, cell survival, and cell adhesion [27]. It is interesting to mention that a subgroup of rats from the stress model displayed hopelessness and were susceptible to depression-like behavior. It was more of a surprise to see those rats showing a blunted miRNAs response in PFC compared to resilient rats. This suggests that vulnerability to stress could be related to the aberrant miRNA response in the brain [27]. Another preclinical model of chronic unpredictable stress showed a significant expression variability in miRNA let-7a, which was related to the altered expression of the serotonin Htr4 gene in the hippocampus [57]. Studies have also reported miR-124- and miR-18a-mediated down-regulation of glucocorticoid receptor (GR) gene transcription and their role in susceptibility to stress [58, 59]. Chronic stress-induced changes in miRNA expression were found to be brain region specific. A study with a chronic stress/recovery paradigm showed a significant decrease in miR-709 expression in the cerebellum, which was not evident in either hippocampus or frontal cortex. Differential responsiveness was further detected for the same miRNA in the cerebellum during the recovery phase [60]. In another study, rats undergoing chronic and acute immobilization stress showed alterations in dendritic spine morphology and concomitant changes in the expression of miR-134, miR-17-5p, and miR-124 in the CA1 area of the hippocampus and central nucleus of the amygdala [43].

MiRNA changes in response to stress susceptibility and resiliency are key to understanding the epigenetic perturbation in brain areas associated with mood regulation. Studies in our lab using rat stress models have shown changes in miRNA expression in PFC and their potential to develop stress-induced depression [27]. In another study, transcranial injection of miR-124 in murine hippocampal neurons, manifested stress resiliency when exposed to chronic unpredictable stress regimes [61]. Contrasting changes were noted when anti-miR-124 oligo was infused into the brain. The result demonstrated increased vulnerability to stress as the endogenous miR-124 expression was alienated with antagomir oligo infusion. The role of miR-124 was thought to be associated with spatio-temporal changes that are key to alter the functions of downstream genes or any other epigenetic constructs pertinent to stress coping mechanisms [61]. These findings were highly useful in conceptualizing miRNAs' modulatory role in perturbing the downstream molecular pathways to the variable stress responses.

With the increasing use of next-generation sequencing and big data analysis, substantial differences have been noted in the transcriptomic profile between males and females [62, 63]. The sex-biased transcriptomic differences are even more significant in MDD subjects, given that MDD disproportionately affects females compared to males [64]. We also believe that the same is true for miRNA expression profile as some recent studies on miRNA expression profiling supported this idea [65-68]. These studies indicate that there are biological determinants that significantly help to define the sexually dimorphic nature of neuropsychiatric disabilities, including MDD [69]. However, the underlying mechanism for this dimorphism is not very clearly known at this point. In a recent report, an interesting finding was made to highlight the role of the cholinergic system in acquiring some of the differences through sex-specific miRNA regulation [70]. The study specifically highlighted the female-specific changes in miR-10 and mir-199 expression; both associated with cholinergic activities [70].

In a recent study, our lab highlighted sex-specific changes in miRNAs in early-life and acute stress. Most susceptibilities were seen in the limbic area of the brain, with males expressing more significant changes in the hypothalamus following postnatal stress than females [65]. We also found that with early chronic variable stress, miRNAs that were enriched in PFC had less expression variability in females than in males undergoing maternal separation (MS). The behavioral changes were also less pronounced in females, with signs of anhedonia noted under early life stress. Interestingly, in female animals, miR-493 showed increased expression under MS; however, in the socially reared stressed animals, miR-493 was decreased [65, 66]. These novel findings outline the sex-specific changes in miRNA response to early life stress. Additionally, the findings also serve as the bridge between the studies that rely exclusively on either male or female animals. In a separate study, we implicated the interaction of miRNA with the serotonin pathway in the early life stress (ELS) model [71]. It was found that the cognate interaction between specific miRNAs and genes from serotonergic pathways has a role in increasing susceptibility to depression-related reward deficits. We showed a reduction in miR-320-5p following MS, with sex differences in its response to enrichment. Its predicted target, serotonin receptor 1A (Htr1a), showed opposite expression changes while validating this target gene in the cellular model. We also found significant increases in serotonin transporter Slc6a4 expression in the PFC. Altogether, this study showed that enrichment might be a viable tool for preventing anhedonia following ELS via miRNA and serotonergic gene targets [71]. In keeping with our previous studies, where we had highlighted the involvement of the hypothalamus in early-life and acute stress [66], we observed that sex plays a critical role in hypothalamic miRNA response [65, 66]. Changes in miRs-29, – 124, – 132, – 144, – 504 expressions were significantly notable in MS male rats. However, environmental enrichment was able to reverse the downregulation of miR-29b-1-5p and -301b-3p in MS animals. Moreover, enrichment significantly altered the target genes (MAPK6 and MMP19) of the corresponding miRNAs [66]. These studies show that miRNAs play a role in causing susceptibility to develop depression associated with early life trauma and this miRNA-mediated susceptibility can be reversed using environment enrichment.

In a recent report, miR-218 has been highlighted for its role in early life stress [72]. An abnormal miR-218 expression in adolescent mice was considered an early prognosticator of lifetime stress vulnerability in male mice [72]. In this study, in addition to adulthood (PND 75), the samples from mice were collected at early adolescence (postnatal day 21) and mid-adolescence (PND 35) age. It was reported that there was a trend of increasing miR-218 expression and, at the same time, decreased expression of target gene Dcc in the mPFC from early adolescence to adulthood. This illustrates the role of miR-218 as an epigenetic regulator of early life stress throughout development as in adulthood [73]. Similar findings were made in other studies, including a report showing a change in miR-16 expression [74]. The miR-16 change was closely associated with maternal deprivation and was related to the repression of Bdnf gene expression in the hippocampus [74]. A recent study reported that maternal deprivation and CUS were repressive for miR-504 expression and simultaneously negatively impacted D1 and D2 dopamine receptor expression in the nucleus accumbens (NAc) of rats. This epigenetic interplay between dopamine transmission and miR-504 in maternally deprived pups caused increased vulnerability to stress during adulthood [75]. Another animal study showed the role of miR-34c in the central amygdala with increased anxiety-like behavior in mice [76]. Under the chronic unpredictable stress model, the anhedonic behavior of rats was associated with a parallel change in let-7a expression in the hippocampus, possibly mediated through the repression of Htr4 gene expression [57]. In chronically stressed rats, epigenetic regulation of glutamatergic receptor Gria4 by induced miR-124-3 in the frontal cortex was partially responsible for the induction of depression-like behavior [33]. Studies have also reported miR-124 and miR-18a mediated downregulation of GR gene, shown to be significant in stress susceptibility [33].

Besides, a transgenerational epigenetic inheritance has recently been highlighted to understand the miRNA’s contribution to MDD, which can be passed on from one generation to the next [21, 77, 78]. In a recent report, paternal stress exposure has been found to alter sperm miRNA content, and an altered miRNA pool can be inherited to reprogram the HPA axis in offspring [78]. The study's outcome suggested that miRNA-mediated epigenetic inheritance could potentially impact the development of MDD.

To provide an integrated view of the miRNAs and their functions mentioned above in the context of their roles in plasticity and stress responsiveness, we ran a target prediction analysis. A list of miRNAs and networks for plasticity and stress are provided in Figures 3A and 3B, respectively. First, we built two target prediction-driven miR-Tar networks, which were then filtered and enriched with a high confidence level for biological functions related to synaptic/neural plasticity and maladaptive stress behavior. As can be seen, a large number of genes (e.g., Akt1, Akt3, MAPK, SIRT1, CREB, WNT1, WNT3, NFAT, TGFB, PI3K, GSK3B, NFKB, GRIA1, FAS, BCL2) appeared in these networks that have earlier been reported to regulate various brain functions including plasticity. These genes have also been shown to be abnormally expressed or functionally impaired in stress-related disorders. Further analysis, based on functional clustering of predicted target genes, provided two separate gene ontology lists (Figures 3C and 3D); some of them shared common gene functions, whereas some ontological clustering was unique to either plasticity (dopaminergic, serotonergic, and sphingolipid pathway) or stress susceptibility (phosphatidylinositol, inositol biosynthesis, and synaptic vesicle cycle). This integrated view of miRNA-gene function provides a deeper understanding of how miRNAs can be involved in stress-related disorders.

Figure 3: miRNA-target gene base network and associated gene ontology for plasticity and stress-responsive miRNAs.

A) Plasticity-associated miRNA-target-based interaction network. Predicted targets based on synaptic plasticity-associated miRNAs (listed in Table 1) were used to map a miRNA-target interaction network with square nodes showing the miRNA names and round ones as connected targets. Target prediction was based on miRTarBase version 8.0. B) Stress-associated miRNA-target-based interaction network. Predicted targets based on cognate interaction with stress-responsive miRNAs (listed in Table 1) were used to map a miRNA-target interaction network with square nodes showing the miRNA names and round ones as connected targets. Target prediction was based on miRTarBase version 8.0. C) Gene ontology analysis for biological process conducted with predicted target genes associated with plasticity-related miRNAs. Grided bubble plot showing significant enrichment of terms in various categories associated with neuronal functions. The lower to higher p values are shown on a scale of gradient color, and the circle size means the number of gene counts in each GO term. D) Gene ontology analysis for biological process conducted with predicted target genes associated with stress responsive miRNAs. Grided bubble plot showing significant enrichment of terms in various categories associated with neuronal functions. The lower to higher p values are shown on a scale of gradient color, and the circle size means the number of gene counts in each gene ontology term.

Use of miRNAs in MDD Treatment Response

There is a growing interest in understanding if miRNA expression changes can be viewed as markers for antidepressant treatment response. Based on clinical and preclinical studies, it is becoming clearer that several antidepressants exert their effects by targeting miRNAs [79]. As miRNAs have been reported to play essential roles in depression, identifying miRNAs associated with antidepressant treatment responses will be pivotal in understanding the predictors and molecular mechanisms of antidepressant treatment. Moreover, identifying potential miRNA panels could serve as treatment response biomarkers for devising more effective therapeutic strategies [80]. Recent data from preclinical studies have shown altered miRNA response in the hippocampus of rats treated with the NMDA receptor antagonist ketamine. MiR-451 is one such miRNA that shows expression reversal in the early stress model of rats when treated with ketamine [81]. In the mouse model, changes in miR-16 and miR-135a expression were responsive to fluoxetine and escitalopram treatment, respectively, both being serotonin reuptake inhibitors (SSRIs) [82, 83]. The same miR-16 showed increased expression levels in depressed patients after four weeks of SSRI treatment compared to baseline [82]. Similar findings were noted for miR-135 expression under cognitive behavioral therapy; however, no change was noted in miR-135 expression when MDD patients were treated with escitalopram for 12 weeks. Other studies, including a preclinical report, have suggested changes in miR-212 expression level in rats' dentate gyrus in response to electroconvulsive (ECT) stimulation [84]. In a separate report, miR-24-3p was suggested as a tentative ECT response biomarker in treatment resistance depression (TRD) [85]. In a separate clinical study, patients with psychotic depression have been shown to respond with ECT, which normalized the elevated blood levels of miR-126-3p and miR-106a-5p [86]. Based on 11 independent clinical studies, 88 miRNAs were found to be differentially regulated in response to antidepressant treatments. Interestingly, a majority of them were upregulated (61) compared to 26 downregulated miRNAs [87]. Very recently, in a randomized control trial, a total of 228 miRNAs were found to be correlated with symptomatic remission after two weeks of SSRI treatment in MDD subjects. Out of 228 miRNAs, miR-483-5p exhibited the most robust treatment response outcome [88]. Furthermore, a relation of miR-1202 to citalopram treatment was studied in whole blood drawn from MDD patients. Interestingly, treatment responders showed lower baseline miR-1202 plasma levels compared to controls and non-responder patients. However, in responders, the miR-1202 expression peaked after eight weeks of antidepressant treatment [89].

MiR-124 has shown drug-responsive changes in peripheral blood mononuclear cells (PBMC) of MDD patients treated with escitalopram. Higher miR-124 was found at baseline before escitalopram treatment, which sharply decreased after eight weeks of therapy, especially in responders [90]. This suggests the role of this miRNA as a biomarker for treatment efficacy. Using next-generation sequencing in MDD patients treated with duloxetine, another study found downregulation of several miRNAs which targeted WNT and mitogen-activated protein kinase (MAPK) signaling. However, the report also showed that six miRNAs were significantly changed in the placebo response group, and they were also present after 8 weeks of duloxetine treatment [91], suggesting that the alterations in the expression of this miRNA may be associated with metabolic processes that are independent of clinical outcome. After replication in patients treated with escitalopram and in the postmortem brains of depressed subjects, this study showed the downregulation of miR-146b-5p; miR-24-3p, and miR-425-3p were the most effective miRNAs for antidepressant treatment response [91].

Several technological developments have been made to understand the brain's response to the treatment directly in the peripheral circulation. In this regard, extracellular vesicle or exosome has become increasingly popular due to their ability to cargo a specific pool of miRNAs from the brain [92, 93]. In a recent study, the expression of miRNAs, including let-7e, miR-21-5p, miR-223, miR-145, miR-146a, and miR-155 was measured in the serum exosome obtained from patients with MDD before and after treatment with antidepressants. In the study, remitted patients receiving antidepressant treatment showed increased expression of let-7e, miR-21-5p, miR-145, miR-146a, and miR-155 compared with the non-remission group. On the contrary, the expression levels of the same miRNAs were significantly lower before the treatment started. This is an interesting report highlighting the role of brain-derived serum exosome miRNAs in predicting antidepressant response in the peripheral circulation [94].

Identification of unique miRNAs can also serve as a potential source of screening that may help early detection of the severity of depression and the treatment response. Recently, plasma exosome-based miRNA expression profiles in treatment-resistant depressed patients (TRD) showed significant differences in the expression of two miRNAs: miR-335-5p was significantly upregulated, while has-miR-1292-3p was significantly downregulated [93]. This suggests that measurements of these miRNAs will improve the diagnosis accuracy of TRD patients. Some of the antidepressant-responsive miRNAs are represented as a chord diagram in Figure 4. As can be seen, some miRNAs are commonly associated with various antidepressants, whereas some are unique to each antidepressant. Further genome-wide examination of antidepressant-responsive miRNAs will help find a panel that can be used in predicting treatment response.

Figure 4: Antidepressant-responsive miRNAs.

A hypothetical circular connectivity plot to highlight the key miRNAs which have been found to be consistently responsive to various antidepressant treatments. In the plot, miRNAs responsive to the respective antidepressant treatments are connected via cords (represented with different colors). Maximum number of miRNAs were found to be responsive to escitalopram treatment.

Outstanding Questions and Future Directions

MiRNA-related research in neuropsychiatry has significantly advanced our present understanding of their roles in disease pathogenesis. This is especially true for depression and other stress-related disorders. As pointed out in the previous sections, alterations in miRNAs expression or functions can give rise to specific phenotypes associated with stress resiliency to susceptibility. However, their potential value as biomarkers to predict the disease outcome is yet to be thoroughly tested in clinical settings. Like any other molecular biomarkers, the use of miRNAs in the diagnostic and prognostic assessment of neuropsychiatric disorders should be highly sensitive and specific. To ascertain the diagnostic value, a more heuristic approach is needed where an association of miRNAs needs to be established with various endophenotypes, given the heterogeneity associated with depression and other psychiatric disorders. For this, a panel of miRNAs showing consistent response must be studied with high-throughput screening [16]. On the other hand, the efficacious use of miRNAs in RNA therapeutics is currently limited by a couple of factors. One of them is their lower penetration ability to cross cell membranes, which can be a potential drawback in drug delivery methods [95]. The second factor is their ability to trigger innate immune responses upon their systemic delivery in vivo [95]. However, the possible use of exosomes (EVs) as vesicular cargo with an ability to cross the blood-brain barrier (BBB) is quite promising in targeted drug delivery [96]. We have discussed this and other potential delivery methods in detail later in this section.

Another critical point to consider is to evaluate miRNA-miRNA crosstalk. Much of the studies conducted so far have focused on examining a direct relationship of miRNAs with their target genes. However, it is important to note that miRNAs tend to work in a coordinated fashion, and a crosstalk examination can provide an integrative approach to understanding how miRNAs work cohesively to regulate a specific function or participate in particular disease pathogenesis. This crosstalk is not limited to a specific cell or tissue but across various brain regions. It is generally agreed upon that brain regions function through interconnected gene networks. Research has found clustered miRNA expression changes across different regions of the MDD brain. Given that individual brain regions cannot function in isolation, coordinated miRNA response across various brain regions under complex neuropsychiatric conditions like MDD is expected. Thus, in the future, consideration should be given to an integrative approach to finding interconnected miRNA networks across brain regions.

Certain limitations also exist in measuring miRNAs in blood plasma or serum. This approach may not provide reliable biomarkers, given that miRNAs are secreted in the blood from various sources. Exosomal (EV) miRNAs from body fluids seem to be a promising approach, which assures more reliable profiling of miRNA panels [97]. Depending on the size, EVs can cross the blood-brain barrier. These EVs contain various genetic materials, including miRNAs. However, EV subpopulations should be carefully screened to overcome the issues associated with the specificity and purity of samples. Studies have suggested that exosomal miRNAs offer a significant advantage compared to free-streaming miRNAs in peripheral body fluids. Exosomes contain a specific set of cell surface markers, which may represent their tissue of origin [79]. The neuron-derived exosomal fraction found in peripheral circulation can be selectively immuno-enriched with the help of an antibody (85). Similarly, glial and astrocytic exosomes can be immunoprecipitated using specific antibodies. However, due to a very low abundance of cell-type specific exosomes in peripheral body fluid, it is strongly recommended to apply highly sensitive miRNA detection assay methods such as low-input RNA next-generation sequencing or microarray (Gene-Chip) platforms [98].

The full potential of miRNA as a therapeutic target is another area of active research and could be an attractive option for developing neurotherapeutics against psychiatric disorders. In practice, miRNA-based therapies are an advancement of siRNA agents, where miRNAs can influence a regulatory sequence or multiple pathways concurrently. In this context, therapeutically, it is possible to supplement non-functional miRNAs as synthetic oligonucleotides through artificial antagonists, either as oligonucleotides or small molecules. However, often the specificity of oligo therapeutics depends on their successful delivery mechanism to the tissue or cell type. Currently, besides liposomal delivery system and cationic polymers conjugates, various nanoscale drug carriers, such as liposomes, dendrimers, viral capsids, polymeric nanoparticles, silicon nanoparticles, and magnetic/metallic nanoparticles, have been designed to enable effective drug delivery through the blood-brain barrier. However, only a few of them are tested for successful miRNA-based oligo delivery in the central nervous system. Advanced formulations are needed to achieve a higher degree of target-specific delivery while shielding RNA oligonucleotides from various degrading enzymes in systemic circulation.

In conclusion, the current review summarizes the key aspects of miRNAs in shaping the epigenetic landscape of the MDD brain. More importantly, the review provides a narrative on miRNAs and their influence in orchestrating a vast array of neuromolecular pathways susceptible to changes under disease pathophysiology such as MDD. The review also offers future directives and roadmaps for miRNA-dependent biomarker discoveries and their promising role as therapeutic targets under neuropsychiatric conditions.

Table 1:

Plasticity and stress-associated miRNAs in MDD pathogenesis

miRNA biogenesis-associated enzymes
Name	Functions	References
Drosha, Dgcr8, Dicer, Trbp, Ago, and eIF2c	Primary (pri) and precursor (pre) miRNA processing, target identification, and loading	Ha et al. 2014; O’Brien et al. 2018; Davis and Hata 2009
Plasticity-associated miRNAs
miRNAs	Affected Functions	Reference
miR-132-3p, miR-125b-5p, miR-138-5p, and miR-124-3p	Synaptic plasticity, spine morphogenesis, memory acquisition, fear conditioning, dendritic spine density	Wayman et al. 2008; Edbauer et al. 2010; Impey et al. 2010; Yi et al. 2014
miR-134, pre-miR-134	Dendritic spine formation, excitatory synaptic transmission; dendritic transportation; synaptic plasticity, LTP and memory formation; neuronal sprouting; dendritic spine density	Schratt et al. 2006; Gao et al. 2010; Bicker et al. 2013
miR-138	Palmitoylation mediated change in G protein α activity; synaptic plasticity; axonal regeneration	Siegel et al. 2009; Liu et al. 2013; Schroder et al. 2014
miR-9, miR-125a/b and miR-188	Synaptic plasticity, axonal elongation, synaptic connectivity; synaptic strength, and dendritic spine stabilization	Ye et al. 2016
miR-124	Synaptic plasticity, dendritic arborization, homeostatic synaptic plasticity, facilitation of sensory-motor neurons, 5HT-dependent long-term facilitation of sensory-motor neurons	Rajasethupathy et al. 2009; Hou et al. 2015; Roy et al. 2017
let-7i, miR-19b, miR-29c, miR-101a, miR-124, miR-137, miR-153, miR-181a, miR-181c, miR-203, miR-218, miR-324-5p, miR-365, miR-409-5p, miR-582-5p, miR-155, miR-29a, miR-30e, miR-721, miR-699, miR-146a, miR-200c, miR-351, miR-155, miR-678, miR-764-5p, miR-135a*	Compromised HPA axis functionality, inflammation, synaptic plasticity, cell differentiation, cell survival, cell adhesion, and epigenetic modifications	Dwivedi et al. 2015
miR-96, miR-141, miR-182, miR-183, miR-183*, miR-198, miR-200a, miR-200a*, miR-200b, miR-200b*, miR-200c, and miR-429	Stress-induced depression phenotype in learned helpless rats	Smalheiser et al. 2011
miR-298, miR-130b, miR-135a, miR-323, miR-503, miR-15b, miR-532, and miR-125a, miR-7a, miR-212, miR-124, miR-139, and miR-182	Depression and response to antidepressants in the hippocampus	Cao et al. 2013
Stress and MDD-associated miRNAs
miR-709, miR-186	Stress-induced motor function impairments	Babenko et al. 2012
miR-134, miR-17-5p and miR-124	Dendritic spine morphology of neurons in the central nucleus of the amygdala	Schratt et al. 2006; Yu et al. 2008; Beveridge et al. 2009
miR-34c	Anxiety-like behavioral changes in the amygdala of stress-induced mice	Haramati et al. 2011
miR-16	Maternal deprivation associated changes in mice hippocampus inducing anxiety-like phenotype	Bai et al. 2012
let-7a	Anxiety-like behavioral changes in the amygdala of stress-induced mice	Bai et al. 2014
miR-124 and miR-18a	Glucocorticoid receptor regulation under stress susceptibility	Vreugdenhil et al. 2009; Herman et al. 2012
miR-504	D1 and D2 dopamine receptor modulation in nucleus accumbens of adult rats due to maternally deprivation	Y. Zhang et al. 2013
miR-135	Regulates anxiety-like behavior by modulating the 5HT system	Issler et al. 2014
miR-29c, miR-30a, miR-30c, miR-32, miR-193-5p, miR-204,miR-375, miR-5323p, and miR-698	Transgenerational transmission of paternal lifetime experiences via sperm miRNAs	Rodgers et al. 2015
miR-139-5p, miR-195, miR-320c and miR-34c-5p	Disrupting polyamine biosynthesis pathway in depressed suicide brains	Lopez et al. 2014a
miR-1202	Metabotropic glutamate receptor modulation in brains of depressed suicide subjects	Lopez et al. 2014b
miR-218	Depression susceptibility by increasing resiliency against stress-induced depression	Torres-Berrio et al. 2017
miR-144-5p, miR-320a, miR-451a, miR-17-5p, miR-223-3p, miR-335, let-7a-5p, let-7d-5p, let-7f-5p, miR-24-3p, miR-425-3p, miR-330-3p, miR-345-5p, let-7b and let-7c	Widespread association of miRNAs with etiopathology of MDD in peripheral tissues of depressed patients	Dwivedi. 2016
miR-16	Modulates 5HT system in MDD and counter influences antidepressant fluoxetine response	Baudry et al. 2010
miR-139-5p, miR-195, miR-320c and miR-34c-5p	Disrupt polyamine biosynthesis pathway in the brain of depressed suicide subjects	Lopez et al. 2014a
miR-215-5p, miR-192-5p, miR-202-5p, miR-19b-3p, miR-423-5p, miR-219a-2-3p; miR-511-5p, miR-483-5p	Involved in glutamatergic, ERK/MAPK, neuregulin, estrogen receptor, PI3K, telomeres signaling as well as axon guidance	Yoshino et al. 2020
miR-301b-3p, - 132- 3p, - 132-5p, -449a- 5p, -30e-5p, - 338- 3p, - 144-3p, let-7 g- 3p, miR-200a-3p, miR-322-5p, miR-200b-3p, miR-20b-5p, miR-34b-5p, miR-34c- 3p, miR-34c-5p, miR-145-5p and miR-203a-3p	Involved in synaptic functions by modulating MAPK6, MMP19, and serotonergic signaling	Lauren et al, 2021a, 2021b, and 2021c

Highlights.

• miRNAs play a critical role in regulating gene expression at post-transcriptional level.

• Recent studies showing the role of miRNAs in stress-related disorders, including depression, have transformed the field of neurobiology of affective disorders.

• miRNAs not only participate in the phenotypic characteristics of depression but also provide novel targets for therapeutic interventions.

• Blood-based miRNA profiling can serve as biomarkers in the diagnosis of depression and therapeutic response.