Researchers have long struggled to understand the molecular basis of psychiatric disorders (Figure 1). In the last half-century, these diseases were predominantly thought to be due to altered levels of certain monoamines. Current research, however, supports the hypothesis that these disorders are driven by disruptions in prominent biological pathways, particularly those contributing to synaptic plasticity. While research has made substantial progress in understanding the biological underpinnings of these disorders, here we will only provide a brief overview of some key findings and general themes (for more complete overviews of BPD, schizophrenia, and MDD, respectively, see (Belmaker & Agam, 2008; Goodwin & Jamison, 2007; Lang, Puls, Muller, Strutz-Seebohm, & Gallinat, 2007)).

Genetic, cytoarchitectural, and molecular findings all suggest that a multitude of factors influence the onset and progression of these illnesses. No particular disorder or behaviroal phenotype has been shown to correlate conclusively with any individual gene variant across multiple studies. Rather, it is thought that a multitude of genetic variants contribute to a predisposed state in which certain individuals are sensitive to certain environmental conditions. In recent years, the field of epigenetics has shown that certain environmental influences can produce lasting changes in DNA that alter gene expression and can ultimately influence long-term behavior (Weaver et al., 2004). In addition, familial, twin, and whole genome-wide association studies have firmly established that psychiatric illnesses show some degree of genetic heritability, and many genetic risk factors are thought to confer vulnerability to the various disorders (Belmaker & Agam, 2008; Sklar et al., 2008). Structural and functional MRI studies in individuals with psychiatric disorders have identified ventricular alterations and volumetric changes in regulatory regions such as the limbic system and prefrontal cortex, which are thought to be vital to proper mood regulation (Drevets, 2000; Drevets, Ongur, & Price, 1998; Drevets et al., 1997; Drevets, Savitz, & Trimble, 2008; Neumeister et al., 2005). Finally, molecular findings support several emerging themes associated with the biological underpinnings of psychiatric illnesses, including fundamental impairments in neuroplasticity (Pittenger & Duman, 2008), increased neuronal vulnerability to aberrant stimuli (Kato, 2008), and dysregulation in neurotrophic factor signaling (Duman, 2004; Shaltiel, Chen, & Manji, 2007).

Notably, although researchers have used genetic, in vivo neuroimaging, and molecular approaches to address questions surrounding the biological underpinnings of psychiatric illnesses, these findings have yet to yield a coherent overall image of the molecular basis of associated behaviors. One emerging area of particular interest is the study of microRNAs (miRNAs). miRNAs are small (about 21–23 nucleotides in length), non-peptide-coding RNAs. These molecules are partially complementary to messenger RNAs (mRNAs); thus they bind to targeted mRNAs and either suppress translation of these mRNAs to proteins, or promote mRNA degradation. Accumulating evidence indicates that miRNAs may act as key regulators in a large number of different cellular processes, including early development, proliferation, apoptosis, metabolism, and cell differentiation. A variety of miRNAs have been found in the central nervous system (CNS), and are believed to play critical roles in brain development and structural plasticity. Evidence from a handful of recent studies suggests that miRNAs may prove to be key to our understanding of the pathophysiology and, ultimately, the therapeutics of psychiatric disorders, due to their crucial role in development and broad genetic regulatory influence. MiRNAs could also serve as a unifying link between many diverse groups of findings; for instance, they influence development of prominent brain structures, and are highly responsive to medications used to treat individuals with psychiatric disorders (Zhou et al., 2008). It is possible that they could be differentially expressed in patients with psychiatric illnesses due to genetic predisposition. It should be cautioned, however, that this field is still relatively new and additional research is warranted to elucidate their possible role in psychiatric illnesses and potential targeting for treatment.

Chronic psychosocial stress produces many changes at the cellular level mediated by glucocorticoids, including compromising cellular defenses and making them more susceptible to insults (ie. free radicals, seizures, etc), reducing cellular energy stores, and decreasing neurogenesis. Notably, however, miRNAs are hypothesized to play a specialized role in cellular responses to stress. During the cellular stress response, miRNAs have the capacity to change from translation suppressors to activators by forming new interactions between miRNA/Argonaute complexes and RNA-binding proteins that alter their subcellular localization (Leung & Sharp, 2007). Because of this modified activity, miRNAs could provide a pivotal role in mediating cellular adaptation to stress. In particular, miRNAs have been implicated in the way cells respond to oxidative stress (Leung, Calabrese, & Sharp, 2006; Marsit, Eddy, & Kelsey, 2006), nutrient deprivation (Bhattacharyya, Habermacher, Martine, Closs, & Filipowicz, 2006), and DNA damage (Crosby, Kulshreshtha, Ivan, & Glazer, 2009; Pulkkinen, Malm, Turunen, Koistinaho, & Yla-Herttuala, 2008).

Recent data investigating the manner in which miRNAs are regulated by chronic restraint stress in Wistar-Kyoto rats found that 13 miRNAs were upregulated and three miRNAs were downregulated in the hippocampus following chronic restraint stress for 21 days (Zhou R, et al 2009 in preparation). Interestingly, one of these miRNAs—miR-221—shares a common downstream target: protein phosphatase 3, regulatory subunit B, alpha isoforms (PPP3R1 or calcineurin B). Calcineurin B is a regulatory subunit of the calcium/calmodulin-regulated protein phosphatase calcineurin. Calcineurin, in turn, comprises a 19 kDa calcium-binding protein, calcineurin B, and a 61 kDa calmodulin-binding catalytic subunit, calcineurin A. PPP3R1 was further confirmed as a target of miR-221 in hippocampal cultures treated with mimics or inhibitors of miR-221 to either suppress or enhance translation. Calcineurin B was also shown to be downregulated by this chronic stress paradigm (Zhou R, et al 2009 in preparation). Interestingly, there is much support in the literature for calcineurin’s role in neuroplasticity (Malleret et al., 2001; Mulkey, Endo, Shenolikar, & Malenka, 1994; Wang & Kelly, 1997; Winder, Mansuy, Osman, Moallem, & Kandel, 1998).

Another recent study found that miR-18a downregulated glucocorticoid receptor (GR) protein in the paraventricular nucleus (PVN) in Fischer 344 (F344) but not Sprague Dawley (SD) rats undergoing repeated restraint stress (Uchida et al., 2008). Cell culture studies also demonstrated that miR-18a blocked translation of GR mRNA, and that miR-18a was upregulated in the PVN in F344 rats, but not in SD rats. Interestingly, the authors claimed that F344 rats did not show hypothalamic-pituitary-adrenal (HPA) axis habituation following 14 days of this stressor. Furthermore, the F344 rats had decreased hippocampal cell proliferation and increased anxiety-like behaviors compared with SD rats. Strain differences in GR protein levels were specific for the PVN and were not detected in the hippocampus or prefrontal cortex. The authors suggest that F344 rats may provide a good model for investigating the effects of chronic stress in situations where miR-18a confers some vulnerability to stress.

Another area where the interplay between stress and miRNAs has been explored is that of maternal care in early life, which has been shown to prime the brain to be either more or less sensitive to stress and, ultimately, to influence adult ability to adapt to stress. Two potential mechanisms that may underlie this phenomenon are (1) miRNA regulation and (2) epigenetics (Figure 1). Both of these processes, in turn, have been reported to modulate GR function. Epigenetic studies have noted that maternal behavior where mothers engage in a high degree of licking/grooming and arched-back nursing of their pups produces adult offspring with a more tolerant HPA axis stress response. These effects can be reversed by cross-fostering or by application of a histone deacetylase (HDAC) inhibitor (Weaver et al., 2004), and are due to altered GR function by acetylating histones H3-K9 and the methylation of the NGFI-A consensus sequence on the exon 1(7) promoter of the GR gene (Fish et al., 2004; Weaver et al., 2004). In humans, a very recent study found epigenetic differences in a neuron-specific glucocorticoid receptor (NR3C1) promoter between human postmortem hippocampus samples from suicide victims with a history of childhood abuse and those from either suicide victims with no history of childhood abuse or controls (McGowan et al., 2009).

In a different study, Perkins and colleagues analyzed 264 human miRNAs from postmortem prefrontal cortex of individuals with schizophrenia or schizoaffective disorder using a custom miRNA microarray. They found that 16 miRNAs were significantly regulated in the prefrontal cortex of schizophrenic patients (15 downregulated, one upregulated) compared to healthy controls (Perkins et al., 2007). Another analysis of miRNA expression in postmortem cortical grey matter from the superior temporal gyrus noted significant upregulation of let-7g and miR-181b expression in schizophrenia (Beveridge et al., 2008). Others have examined 22q11.2 microdeletions, which have previously been shown to place individuals at high risk for schizophrenia. Using mouse genetics to mimic the human 22q11.2 microdeletion, they found biogenic changes in brain miRNAs influenced by this microdeletion (Stark et al., 2008).

In conjunction with these clinical observations, many researchers have attempted to elucidate the molecular alterations associated with lithium’s delayed onset of action. Over the past decade, studies have identified a plethora of proteins targeted by lithium treatment, with changes reported in canonical proteins such as B-cell lymphoma 2 (Bcl-2), extracellular signal regulated kinase (ERK), brain derived neurotrophic factor (BDNF), and glycogen synthase kinase 3 (GSK-3β). Broadly, lithium treatment has also been associated with enhanced neuroplasticity, improved neuroprotection, increased rates of neurogenesis, and several other changes in biological function (Chen & Manji, 2006; Coyle & Duman, 2003). Overall, lithium alters a multitude of intracellular cascades, has a large number of direct and indirect targets, and induces system-wide changes (Chen & Manji, 2006; Coyle & Duman, 2003; Shaltiel et al., 2007). Thus, predicting its targets has proven challenging and, at times, almost random (though some major cascades have shown a coherent response).

In this context, it is noteworthy that miRNAs have recently emerged as key regulators of complex temporal and spatial patterns of gene/protein expression changes and, thereby, synaptic and neural plasticity. Researchers have therefore recently begun to investigate whether miRNA expression is altered by treatment with mood stabilizers. Zhou and colleagues identified eight miRNAs whose expression levels decreased in the hippocampus of rats following chronic treatment with lithium or valproate (two structurally dissimilar mood stabilizers); one miRNA increased in response to both drugs (see Table 2) (Zhou et al., 2008). This study went on to validate that decreases in miRNA levels can elicit a concurrent increase in a predicted protein target in cell culture following exogenous application of lithium or VPA. While this work highlights that miRNAs are responsive to mood stabilizer treatment, subsequent studies are needed to validate whether targeting any of these miRNAs individually can elicit mood stabilizer-like effects in rodent models of mania or depression.

More broadly, regulated alteration of miRNAs could prove to have a high therapeutic potential, or at least help to explain the therapeutic effects of current mood stabilizers. Any given miRNA can have hundreds of possible messenger RNA targets, so that altering one miRNA may lead to a diverse set of alterations in downstream protein expression. Given that lithium treatment alters some miRNA levels, it is possible that these molecules serve as upstream regulators of already identified protein targets. The accuracy of bioinformatic prediction software continues to improve as models incorporate new findings surrounding the targeting mechanisms of miRNAs (Chaudhuri & Chatterjee, 2007; Rajewsky, 2006; Sethupathy, Megraw, & Hatzigeorgiou, 2006). This improvement will help researchers to efficiently determine whether the multitude of reported effects from lithium could be due to the alteration of only a handful of miRNAs.