|Home | About | Journals | Submit | Contact Us | Français|
Circadian rhythm abnormalities are strongly associated with bipolar disorder, however the role of circadian genes in mood regulation is unclear. Previously, we reported that mice with a mutation in the Clock gene (ClockΔ19) display a behavioral profile that is strikingly similar to bipolar patients in the manic state.
Here, we utilized RNA interference (RNAi) and viral-mediated gene transfer to knock-down Clock expression specifically in the ventral tegmental area (VTA) of mice. We then performed a variety of behavioral, molecular and physiological measures.
We found that knock-down of Clock specifically in the VTA results in hyperactivity and a reduction in anxiety-related behavior which is similar to the phenotype of the ClockΔ19 mice. However, VTA specific knock-down also results in a substantial increase in depression-like behavior, creating an overall mixed-manic state. Surprisingly, VTA knock-down of Clock also altered circadian period and amplitude, suggesting a role for Clock in the VTA in the regulation of circadian rhythms. Furthermore, VTA dopaminergic neurons expressing the Clock shRNA have increased activity compared to controls, and this knock-down alters the expression of multiple ion channels and dopamine-related genes in the VTA which could be responsible for the physiological and behavioral changes in these mice.
Taken together, these results suggest an important role for CLOCK in the VTA in the regulation of dopaminergic activity, manic and depressive-like behavior, and circadian rhythms.
Bipolar disorder is a severe psychiatric illness that is characterized by extreme changes in mood. Several studies have found irregularities in daily rhythms in activity, body temperature, blood pressure, the secretion of different metabolites in urine, and in circulating hormone levels in bipolar patients (1–6). Furthermore, bipolar patients appear to have an unstable and non-adaptive clock, and in fact Interpersonal and Social Rhythm Therapy (IPSRT) which maintains patients on a regular sleep and social schedule is very effective in reducing the number of manic and depressive episodes (7). Interestingly, treatment with the mood stabilizer lithium leads to a lengthening of the circadian period that is observed across several species and this effect on rhythms may underlie its therapeutic efficacy (8, 9). In addition, the use of other chronobiological tools such as bright light therapy have proven to be effective for the modulation of both circadian rhythms and mood in bipolar disorder and other affective disorders (4). Thus, it has been hypothesized that abnormalities in the circadian clock contribute to the development and progression of mood disorders, including bipolar disorder.
Circadian rhythms are controlled by a conserved group of core clock genes (10). The proteins, Circadian Locomotor Output Cycles Kaput (CLOCK) and Brain and Muscle ARNT-Like Protein 1 (BMAL1) bind to cis-regulatory elements in several genes including the Period genes (Per1, 2 and 3) and the Cryptochrome genes (Cry 1 and 2) (11, 12). The PER and CRY proteins suppress the activity of CLOCK and BMAL1, creating a negative feedback loop which cycles over the course of approximately twenty-four hours (13). The master circadian clock is located in the suprachiasmatic nucleus (SCN), however these genes are widely expressed throughout the brain and in other organs where they can form additional pacemakers which may cycle independently from the SCN.
Previously we found that mice with a mutation in the Clock gene (ClockΔ19) which creates a dominant-negative protein (11, 14), display a complete behavioral profile that is very similar to human mania (15). Furthermore, chronic treatment with lithium restored the majority of their abnormal behaviors to wild type levels. Interestingly, when we recorded from dopamine neurons in the VTA we found that the ClockΔ19 mice have and increase in dopaminergic activity (15, 16). Since the ClockΔ19 animals express the mutant protein throughout the brain and through all stages of development, we wanted to determine the importance of CLOCK in the VTA alone in adult animals. Thus we developed a new strategy to knock-down the expression of CLOCK in the VTA of adult mice using a CLOCK-specific shRNA expressed in adeno-associated virus (AAV).
8–10 week old male C57BL/6J mice (bred at UT Southwestern) were used for all the studies. Mice were group housed on a 12h light/dark cycle with lights on at 7am and lights off at 7pm with ad lib access to food and water for most studies. For the circadian rhythm studies, mice were singly housed with access to a running wheel and ad lib food and water first in a 12/12 L/D cycle and then they were put into constant darkness to measure free running rhythms. All other behavioral tests were performed at ZT3-6. All mouse experiments were in compliance with protocols approved by our institutional animal care committee.
Viral production and shRNA design were carried out using a helper-free triple transfection method in HEK 293 cells (ATCC) as described in Hommel et al., 2003 (17) and was purified according to Zolotukhin et al., (18). Sequences and other details are in Supplement 1.
The locomotor response to novelty, circadian locomotor rhythms, elevated plus maze, open field, dark light, forced swim test and learned helplessness test were all performed as described previously (15,16). Detailed methods can be found in Supplement 1.
Frozen brains from mice infected with AAV scrambled or Clock shRNA were sliced into 7µm sections and placed on slides (Arcturus). Slides were stored at −80°C until further processing on the LCM (Arcturus). LCM slides were dehydrated in 100% ethanol for 20 sec followed by xylene for 20 sec. Slides were then air dried and mounted in the LCM. The transfected regions could be visualized by fluorescent microscopy since all viruses used co-expressed GFP. 3000 transfected cells were laser-captured from the VTA of each mouse. Following sample collection, RNA was purified using the PicoPure RNA extraction kit (MDS Analytical technologies) as described by the manufacturer and processed with the RiboAmp kit (MDS Analytical technologies). One round of RNA amplification was performed before the amplified RNA was reverse-transcribed using Superscript III (Invitrogen) and quantified by qPCR (Applied Biosystems).
Real time PCR was performed as described previously (16). Details including primer lists can be found in Supplement 1. The amount of gene expression was quantified using the ΔΔCt method as previously described (20).
All data are expressed as mean ± standard error of the mean. Significance for two group comparisons in behavioral assays and qPCR analysis was determined by an independent Student’s T-test. One-way ANOVA was used to determine significance in the period calculation for circadian experiments. Differences in electrophysiology were analyzed by two-way ANOVA. In all experiments P<0.05 is considered significant.
To knock-down the expression of CLOCK in the ventral tegemental area (VTA) of mice, we designed a shRNA sequence that was targeted to a specific region of mouse Clock mRNA (Clock shRNA) and another shRNA containing scrambled sequences (Scr shRNA) with no similarity to any known genes. These shRNAs were cloned into an adeno-associated virus (AAV) vector (17) that was designed to express both enhanced green fluorescent protein (EGFP) and a U6 promoter-driven shRNA. Use of this vector enabled us to identify only the neurons that express viral-mediated shRNA. To specifically deliver the AAV containing Scr shRNA or Clock shRNA into the VTA of mice, we performed stereotaxic injections. Immunostaining of AAV transfected mouse brain sections with anti-tyrosine hydroxylase (TH) showed a clear co-localization of EGFP and TH (Figure 1A–F), demonstrating that we can distinctively target Clock shRNA into the VTA. Two weeks following AAV injection, brains were removed and the cells expressing EGFP were collected using laser capture microscopy (LCM). Quantitative analysis of Clock mRNA levels by real time PCR revealed that Clock shRNA significantly (P< 0.001) reduced the levels of Clock in the VTA of mice when compared to mice injected with the Scr shRNA (Figure 1G). These results demonstrate that the Clock shRNA is able to significantly knock-down the expression of Clock in the VTA of mice. To determine if the knock-down of Clock in the VTA leads to a compensatory increase in expression of Npas2 as is seen in Clock knock-out mice which are missing Clock expression throughout development (22), levels of Npas2 expression were measured in the VTA. In agreement with previous results (23), Npas2 expression was not detectable in either the Scr shRNA injected animals or in the Clock shRNA injected animals in the VTA, while it was readily detectible in striatum and cortical regions, demonstrating that our primer set was functional (data not shown). Thus, there is no measurable compensatory increase in NPAS2 following several weeks of adult knock-down of Clock in VTA.
After successfully knocking-down Clock expression in the VTA, we wanted to determine how this manipulation would influence behavior. We first determined the effects of Clock knock-down in the VTA on locomotor activity. Clock shRNA injected mice exhibit a significant (P < 0.001) increase in their total locomotor activity in response to a novel environment when compared with Scr shRNA injected mice over a two hour period (Figure 2A and 2A inset). The effect is somewhat delayed, indicating a lack of habituation rather than an immediate hyperactive response to novelty. These results indicate that CLOCK protein expression in the VTA plays a role in regulating locomotor activity over time in response to a novel context.
To further determine whether suppression of Clock expression in the VTA leads to any alterations in locomotor activity or rhythms in activity, wheel running activity was monitored for an initial seven days under a 12 hour light/dark cycle and then under constant darkness for another 15 days. Amplitude and period were calculated from the dark/dark phase of the wheel running data. Figure 2B shows that the Clock shRNA injected mice display a behavioral rhythmicity throughout the 15 day dark/dark treatment. However, the Clock shRNA injected mice show an increase in their activity levels during the resting phase and less robust activity during the dark phase, indicating a lower rhythm amplitude (P<0.01) (Figure 2C). Interestingly, although Clock shRNA injected mice exhibit an increase in their locomotor activity in a novel environment (Figure 2A), when we examine the total activity in the home cage over 24 hours, the Clock shRNA injected mice display a significant (P<0.05) decrease in total activity (data not shown). Furthermore, Clock shRNA injected mice show a significant decrease (p<0.01) in the circadian period that was on average 15 min shorter (23.4 ± 0.05 hr) than the Scr shRNA injected mice which display a normal circadian period of 23.6 ± 0.01 hr (Figure 2C, inset). Together, these data suggest that expression of CLOCK protein in the VTA might play a critical role in the generation of normal circadian rhythms in locomotor activity.
To examine the contribution of Clock expression in the VTA to anxiety-related behavior, we subjected Clock and Scr shRNA injected mice to three separate measures: the elevated plus maze, light/dark box, and the open-field. Previous studies have validated these tests as measures of anxiety since behavior is significantly altered in wild type mice with the administration of anxiolytic and anxiogenic drugs (24). Furthermore, results from our lab and others show a clear separation of anxiety-related behavior from hyperactivity in these tests (15, 21). Although the Clock shRNA injected mice have an increase in activity over a two hour period in a novel environment, they are similar to control mice in response to this environment for the first 25 minutes (Figure 2A). Since anxiety-related measures are performed over 5–10 minute periods, the failure to habituate to novelty by the Clock shRNA injected mice should not influence these paradigms. In the elevated plus-maze we find that the Clock shRNA injected mice spend significantly more time (47.0 +/− 7.07 sec vs 23.8 +/− 7.16, P<0.05) on the open arm than the Scr shRNA injected mice. In addition, we find that the number of entries made by the Clock shRNA injected mice in the open arm are significantly higher as compared with Scr shRNA injected mice (13.1+/−1.4 vs 6.5 +/− 0.7, p<0.001; Figure 3A–B). Furthermore, in the light/dark test we find that the mice injected with the Clock shRNA not only spend significantly more time (189.6 +/−12.8, P<0.05) in the lighted area of the box (153.8 +/−11.8), but they also made significantly increased number of crosses to the lighted area (48.2+/−4.0 vs 30.1+/−3.4, p<0.001; Figure 3C–D). In the open field Clock shRNA injected mice spend significantly more time (11.4+/−1.1, P<0.05) in the center of the field than the scr shRNA injected mice (8.2+/−0.9 sec; Figure 3E). The results obtained with these three separate measures are consistent with each other and suggest that knock-down of Clock in the VTA leads to an overall reduction in anxiety-related behavior.
To determine whether knock-down of Clock mRNA expression in the VTA of mice alters depression-like behavior, we carried out forced swim (FST) and learned helplessness tests (LH). These measures have been validated extensively with antidepressant drugs (25, 26). Immobility in the FST and failure to escape in the LH test are indicative of depression-like behavior. Surprisingly, when the latency to immobility is measured in the FST, we find that Clock shRNA injected mice display significantly (P<0.01) shorter latency to immobility compared to Scr shRNA injected mice (Figure 4A). Furthermore, the total immobility period (81 +/− 8.2 sec) is significantly (P<0.05) longer in the Clock shRNA injected mice (Figure 4B). To further substantiate the results obtained with the FST, we carried out the LH test. We find that the Clock shRNA injected mice exhibit greater latency to escape in the LH test (Figure 4C) when compared to the Scr shRNA injected mice (8.0 +/− 1.06 vs 5.0 +/− 0.5 sec, P<0.05). Moreover, the percentage of failures to escape is significantly higher in Clock shRNA injected mice than in Scr shRNA controls (Figure 4C). These results are the opposite of those seen in the ClockΔ19 mice which display a decrease in depression-related behavior in both of these measures (Roybal et al., 2007). Thus, these tests show that CLOCK function in the VTA is necessary for proper mood-related behavior, however different manipulations of this protein lead to opposite effects.
Previously we found that the ClockΔ19 mice have an increase in dopaminergic cell firing which accompanies their overall manic-like phenotype (16). To determine how knock-down of Clock mRNA expression in the VTA of adult animals influences dopaminergic cell firing, we recorded the firing rate of individual dopamine neurons in the VTA from mice that were injected with the Scr or Clock shRNA. In both sets of animals we also recorded from neighboring dopaminergic cells that were not infected with a virus as a control. We find that Clock shRNA infected cells have a significant increase (P<0.01) in the firing rate of dopaminergic cells compared to both non infected cells and the AAV Scr shRNA infected cells (Figure 5). These results, together with the behavioral abnormalities seen in the Clock shRNA injected mice, strongly suggest that CLOCK is involved in modulating dopaminergic activity, and that this modulation is important in regulating mood, activity and anxiety-related behavior.
Since CLOCK is a transcription factor, we wanted to know what gene expression changes would occur following Clock knock-down in the VTA. We performed microarray analysis on RNA isolated from VTA tissue of Clock shRNA infected mice and scrambled controls. We find a number of changes that occur with Clock knock-down in the VTA (see Table S1 and Figure S1 in Supplement 1). 90% of the selected genes that we chose for RT/PCR confirmation (total of 12 genes selected) showed similar changes in expression, giving us great confidence in our microarray data (see Table S2 in Supplement 1 for genes). Interestingly, we noticed that the expression levels of a number of ion channels are altered following Clock shRNA infection (Table 1). Most of the channels were upregulated and there are a number of cholinergic channels and glutamatergic channels that show this regulation. An upregulation of these channels could lead to increased dopaminergic activity in the VTA. Fewer channels were downregulated and they included some potassium channels. We also found changes in a number of genes that are known to regulate dopaminergic activity (Table 2).
These studies indicate that expression of CLOCK in the adult ventral tegmental area (VTA) is involved in the modulation of mood, anxiety and locomotor behavior in mice. The adeno-associated virus (AAV) Clock short hairpin RNA (shRNA) infected animals show a prolonged hyperactive response to novelty and decrease in anxiety-related behavior that is similar to the ClockΔ19 animals. Furthermore, our previous studies found that restoration of a functional CLOCK protein specifically in the VTA of ClockΔ19 mice is sufficient to rescue their hyperactivity and anxiety-related behavior (15). Therefore, it seems clear that the response to novelty and anxiety-related behaviors are largely VTA driven, and that less activity of CLOCK in this region, regardless of the method used or duration of knock-down, results in an increased response to novelty and lowered levels of anxiety-related behavior.
Surprisingly, we find that shRNA-mediated knock-down of CLOCK in the VTA leads to increased levels of depression-like behavior, while the ClockΔ19 animals have a decrease in depression-like behavior (15). Interestingly, when we attempt to rescue the lowered depression-like behavior in ClockΔ19 mice by expressing a functional CLOCK protein in the VTA, we find that it is not sufficient to reverse the behavior (unpublished observations) while it does reverse the anxiety-related phenotypes (15). Depression and anxiety are often co-morbid, however studies are beginning to differentiate specific molecular mechanisms that control each state. For example, a recent study by Wallace et al., (2009) found that rats kept in social isolation developed both increased anxiety and increased depression-like behavior, and both were reversed by chronic antidepressant treatment (21). However, the anxiety-related behaviors and their reversal by antidepressants, were dependent upon expression of the transcription factor, CREB, in the nucleus accumbens shell, while the depression-like behavior was not controlled by a CREB-dependent mechanism (21). Furthermore, anxiety-related behavioral measures in mice may be a reflection of “risk taking” behavior, rather than feelings of anxiety. Patients with bipolar disorder, for example, can feel anxious about certain social or other situations while at the same time they engage in extremely risky behavior without concern about the consequences. Thus, there is a complex interaction between anxiety-related behavior and depression that warrants future study.
These differences in depression-like behavior observed in the ClockΔ19 mice and the Clock shRNA injected mice could come from the fact that the ClockΔ19 animals have the mutation throughout the brain while the Clock shRNA is injected locally in the VTA. It is possible that CLOCK function in another brain region is involved in regulating depression-like behavior. Some possibilities might include the hippocampus and prefrontal cortex since both have been implicated in mood regulation in previous studies (27, 28). There is also the possibility that CLOCK expression in the suprachiasmatic nucleus (SCN) is involved in this modulation of behavior.
There are other important differences between this local knock-down and the ClockΔ19 animals that might contribute to the opposing results in measures of depression-like behavior. First, the ClockΔ19 mice have a mutation in the protein which leads to dominant-negative activity while the shRNA affects Clock mRNA expression only. The dominant-negative protein not only affects CLOCK activity, but also likely alters the activity of any protein that interacts with BMAL1. Indeed there are behavioral differences in mice with the ClockΔ19 mutation versus mice which are null for the Clock gene. Clock knock-out animals have essentially normal circadian locomotor rhythms with only a slight alteration in their response to light while the ClockΔ19 animals have extremely long periods or are totally arrhythmic (11, 29). It will be interesting to determine in future studies how overexpression of the CLOCKΔ19 protein specifically in the VTA of adult animals affects depression-related behavior.
The other major difference between the two systems is that the ClockΔ19 mutation is expressed throughout development while the Clock shRNA is injected into adult animals. It is plausible that having this mutation throughout development leads to compensatory effects that alter the adult animal’s behavioral response. Indeed when CLOCK is absent throughout development, NPAS2 expression in the SCN becomes elevated and it can compensate for the loss of CLOCK to control circadian rhythms (22).
Interestingly, CLOCK knockdown in the VTA leads to an increase in the firing rate of dopaminergic neurons. These results are similar to the increased dopaminergic cell firing rate observed in ClockΔ19 mice (16). Increased dopaminergic cell firing has long been associated with hyperactive behavior and the reward value for pleasurable stimuli, however other studies find that stress also leads to increased dopamine release (30). Furthermore, a study by Krishnan et al., (2007) found that mice which developed depression-like behavior following chronic social defeat have an increase in dopamine cell firing while mice that are resistant have no change in dopaminergic activity (31). These studies, along with our current study, show that both manic-like and depressive-like behavior is associated with an increase in VTA dopamine cell firing. This could be an important mechanism that underlies bipolar disorder since patients show both mood states, however future studies are needed to determine how this increased firing is involved in mood regulation.
Since CLOCK is a transcription factor, any change in CLOCK expression or activity will influence the expression of many other genes. Using microarray analysis, our previous studies found that there are many differences in gene expression in the VTA of ClockΔ19 mice versus wild type controls including a number of genes that are known to regulate dopaminergic activity (16). We find a similar increase in levels of tyrosine hydroxylase (Th) in the Clock shRNA expressing cells as seen in the ClockΔ19 mice, as well as other interesting changes in genes involved in dopamine synthesis, metabolism and release. Some of these changes may be involved in a common mechanism leading to increased dopaminergic activity in both models, while some of the changes point to different mechanisms that may underlie the differences in depression-related behavior (32). Interestingly, we find differences in expression in a number of ion channels and receptors following VTA specific knock-down of Clock, specifically increases in cholinergic and glutamatergic receptors. The nicotinic acetylcholine (nACh) receptors expressed on cell bodies of dopamine neurons modulate the rate of action potential firing and subsequently levels of dopamine release at synapses (33). Stronger nACh receptor responses make VTA dopaminergic neurons more excitable (34, 35). Interestingly, there was an upregulation of certain GABAergic receptors in the VTA which could be compensatory. Future studies will determine the importance of these changes in receptor and ion channel expression in the increased dopaminergic activity and behavioral changes in these mice.
Another intriguing finding of this study is that CLOCK knock-down in the VTA alters circadian locomotor rhythms. VTA knock-down leads to a slightly shorter period and a decrease in amplitude. Consistent with our findings, the circadian period of homozygous Clock null mice are on average ~20 min shorter than the wild type mice (29). In contrast, the ClockΔ19 mice also have a decrease in rhythm amplitude but they often display a long period in constant darkness or completely arrhythmic behavior (14). It is uncertain as to how changes in the VTA modulate circadian period. It is also currently unclear as to how changes in circadian rhythms per se contribute to mood related phenotypes, but it is interesting that Clock knock-down in the VTA and ClockΔ19 mice have opposite changes in period length. Interestingly, chronic social defeat is also associated with a reduction in circadian amplitude of body temperature rhythms only in mice that show a depression-like phenotype, suggesting that these rhythm changes may be important in mood regulation (31).
In summary, our results find a significant role for CLOCK in the VTA in the regulation of depression-like and manic-like behavior. Both of these phenotypes are associated with increased dopaminergic activity in this region. While the hyperactivity and anxiety-related behaviors are mostly controlled through CLOCK expression in the VTA, it is likely that other brain regions contribute to the regulation of depression-like behavior. CLOCK in the VTA also regulates some aspects of circadian locomotor activity. These results are important in developing our understanding of the complexities of bipolar disorder, and how multiple mood states might develop with mutation in a single gene.
We would like to thank Drs. Mary Kay Lobo for help with the LCM and Scott Russo for the help with viral surgeries. We would also like to thank Drs. Caroline H. Ko and Ming-hu Han for valuable discussion. This work is supported by grants from NIDA the NIMH, NINDS, NARSAD, The McKnight Foundation, and the Blue Gator Foundation.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Disclosure: Dr. McClung is a consultant for Orphagen Pharmaceuticals, has received honorarium from GlaxoSmithKline and Pfizer, and has research funding from GlaxoSmithKline. All other authors report no biomedical financial interests or potential conflicts of interest.