What can a clock mutation in mice tell us about bipolar disorder?
Bipolar disorder, also known as manic-depressive illness, is characterized by episodes of mania and episodes of depression usually interspersed with periods of relatively normal mood (1). During the manic phase, affected individuals exhibit elevated mood, irritability, increased activity, reduced sleep, hypersexuality, and increased goal-directed activities. Bipolar disorder in its various forms affects >3% of the population and is associated with a high risk for suicide, substance abuse, and vocational disability (2). Although several animal models for major depressive disorder have been developed, there are no plausible models for bipolar disorder (3). In this issue of PNAS, Roybal et al. (4) describe the results of a systematic analysis of the behavior of a mouse with a deletion of exon 19 in the Clock gene, which shows remarkable parallels to the symptoms observed in individuals in an episode of mania (1). The Clock mutant mice exhibit hyperactivity, decreased sleep, reduced anxiety, and increased response to cocaine, sucrose, and medial forebrain bundle stimulation. Furthermore, many of these behaviors can be reversed by transfection of the ventral tegmental area (VTA) dopaminergic neurons with WT Clock gene or by treatment with therapeutic doses of lithium (Li+), a commonly prescribed mood stabilizer.
Considerable evidence accumulated over the last 30 years supports the notion that bipolar disorder involves a fundamental disruption in circadian rhythms (5). The episodes of mania and depression in bipolar disorder generally develop a regular periodicity, often linked to the seasons of the year (6, 7). Within an episode, disrupted circadian rhythms including sleep–wake cycle, hormonal secretions, and diurnal variation in mood are evident (8–10). Current treatments to prevent the recurrence of episodes of mania/depression emphasize maintaining a stable diurnal pattern of activity along with pharmacologic treatment (11).
The circadian clock has been shown by genetic analysis in Drosophila and mammals to consist of a time-delayed transcription–translation feedback loop (12). In mammals, a heteromeric dimer of the transcriptional activators, CLOCK and BMAL1, induces the expression of several genes by interacting with the enhancer elements of their promoters known as the E-box. These genes include Per1 (Period), Per2, Cry1 (Cryptochrome), and Cry2, the protein products of which translocate to the nucleus to inhibit the activity of the CLOCK–BMAL1 complex, thereby repressing their own expression. Recent studies have identified a polymorphism in the 3′ flanking region of Clock that is associated with more frequent episodes of mood disturbances and reduced need for sleep in bipolar subjects (13, 14). Nievergelt et al. (15) have reported a suggestive association of two other circadian genes, Per3 and ARNTL (BmaL1), with bipolar disorder. Mansour et al. (16) replicated the association of BmaL1 with bipolar disorder and also found an association with Timeless. Thus, clock genes are implicated as potential risk genes in this disorder of complex (non-Mendelian) genetics.
The effective treatment of bipolar disorder must address two issues: the management of the presenting mood disturbance and the prevention of the recurrence of subsequent episodes (1). Atypical antipsychotic drugs, whose primary mechanism of action is blockade of dopamine D2 receptors, are particularly effective in treating manic psychosis (17), although mood stabilizers are generally started at the same time. Thus, the finding of McClung et al. (18) of hyperactivity of the VTA dopaminergic neurons in the Clock mutant is consistent with the clinical response to antipsychotic drugs. With regard to depressive episodes, biogenic amine uptake inhibitors are typically prescribed along with a mood stabilizer. Li+, valproic acid (VPA), and lamotrigine are the most commonly used mood stabilizers and are administered chronically after the resolution of the acute mood disturbance to diminish the risk of occurrence of subsequent episodes (19).
Because of their low potency and absence of identifiable receptor interactions, the mechanism of therapeutic action of the mood stabilizers has proved elusive. However, recent studies suggest that antidepressants and mood stabilizers may be acting through interconnected intracellular signaling pathways that promote neurogenesis and synaptic plasticity (20). The antidepressants, by inhibiting the presynaptic uptake–inactivation of the biogenic amines, norepinephrine, and/or serotonin, activate adenylyl cyclase via their G protein-coupled receptors (GPCRs), ultimately causing activated cAMP response binding protein (CREB) to translocate to the nucleus and increase the expression of BDNF, TrkB, and Bc1–2.
The mood stabilizers appear to act, in part, via the complementary signaling pathway, the Wnt (wingless) signaling pathway (21) (Fig. 1). Wnt binds to a cell surface GPCR, which activates an intermediary kinase, Dishevelled (Dvl). Activated Dvl inhibits glycogen kinase 3β (GSK3β). Active GSK3β phosphorylates β-catenin, thereby targeting it for degradation. Li+ directly inhibits GSK3B by competing with Mg2+ (22). Furthermore, chronic Li+ treatment activates Akt, which is also an inhibitor of GSK3β. Inhibition of GSK3β results in the accumulation of β-catenin, which stimulates axonogenesis and prevents apoptosis through activation of the Tcf/Lcf1 promoter family. Mai et al. (23), using a cell line that overexpresses GSK3β, found that lamotrigine and VPA at therapeutically relevant concentrations also protect against GSK3β-mediated apoptosis, although the precise mechanism remains unclear. VPA, a histone deacetylase inhibitor, also increases β-catenin expression through a direct activation of its transcription (24).
GSK3β sits at the intersection between the Wnt signaling and Clock signaling pathways. Li+ activates Akt, which inhibits GSK3β. Li+ directly inhibits GSK3β, whereas lamotrigine and VPA indirectly inhibit GSK3β. Inhibited GSK3β permits the accumulation of β-catenin, an activator of the Tcf/Lef1 promoter family, and reduces the phosphorylation of Per, reducing its translocation to the nucleus and slowing the circadian cycle. VPA also inhibits nuclear histone deacetylase (HDAC), increasing β-catenin expression.
GSK3β may serve as the bridge between the mood-normalizing effects of the mood stabilizers and their ability to attenuate subsequent mood cycling in bipolar disorder. The Drosophila ortholog of GSK3β, Shaggy, phosphorylates Timeless, thereby regulating circadian rhythms in the fruit fly (25). VPA and Li+ both lengthen the circadian cycle and increase arrhythmicity in Drosophila (26). In the mouse, GSK3β is expressed in the suprachiasmatic nucleus and liver where its phosphorylation exhibits robust circadian oscillations (27). Recombinant GSK3β phosphorylates Per2 in vitro. Treatment of cells with Li+ inhibits GSK3β's phosphorylation of Per2, thereby retarding its translocation into the nucleus and prolonging the circadian cycle, whereas GSK3β overexpression advances the phase of clock gene expression.
The high prevalence of coexisting substance abuse in bipolar disorder suggests that it may be an intrinsic component of the illness (28). In addition to the clock mutant mouse, it is noteworthy that two other circadian genes have been implicated in deviant responses to drugs of abuse. Mice with mutant Per2 exhibit increased spontaneous consumption of ethanol as compared with WT because of its enhanced rewarding effects (29). Ethanol consumption in this mutant can be attenuated by treatment with acamprosate, a glutamatergic drug with efficacy in preventing relapse in abstinent alcoholics. Furthermore, the Per2 mutant mice show significantly greater sensitization to cocaine as compared with WT mice (30). In contrast, Per1 null mutants show no place-preference for cocaine, a behavioral surrogate for cocaine's rewarding properties, and no sensitization after acute or repeated doses of cocaine.
An important limitation of the mutant clock model is that it recreates only the behavioral homologues of mania but not the mood oscillations characteristic of bipolar disorder. Persistent mania is clinically quite uncommon (1). The mutant Clock mouse characterized by Roybal et al. (4) was developed by using N-ethyl-N-nitrosourea-induced mutagenesis, resulting in a protein with exon 19 deleted that has dominant negative effects. Recently, a mouse with a null mutation of the Clock gene has been developed (31). Interestingly, the circadian rhythms of the Clock −/− mice are relatively well preserved, but light-induced phase advances were three times greater in clock −/− mice than in clock +/+ mice. It will be important to determine whether clock −/− mice exhibit a behavioral phenotype more consistent with the biphasic mood dysregulation of bipolar disorder.
The behavioral phenotypes emerging from mice with mutations of the various clock genes are “fractals” of the clinical features of bipolar disorder with disrupted circadian rhythms, manic-like symptoms, and increased substance- abuse vulnerability. The therapeutic effects of mood stabilizers were discovered serendipitously, and their mechanisms of action are complex and imperfectly understood. Thus, mice with mutations in their clock genes may ultimately prove to be particularly useful behavioral models for identifying novel pharmacologic approaches for treating bipolar disorder.
Footnotes
- *E-mail: joseph_coyle{at}hms.harvard.edu
-
Author contributions: J.T.C. wrote the paper.
-
Conflict of interest statement: J.T.C. serves on the Scientific Advisory Board of Abbott Pharmaceuticals and owns its stock. Abbott markets valproic acid, a drug discussed here; however, J.T.C. does not make any recommendations about its use.
-
See companion article on page 6406.
- © 2007 by The National Academy of Sciences of the USA
