|Home | About | Journals | Submit | Contact Us | Français|
Restraint stress exposures evoke progressively larger increases in 2-arachidonoylglycerol (2-AG) in limbic brain regions as the number of repetitions increases. The Porsolt swim test usually involves two swim exposures, separated by 24 hours and we asked whether the 2-AG response differed between the first and second exposures.
Four groups of male C57/Bl6N mice were studied: control; exposed to a single 6 minute swim and killed immediately; exposed to a single 6 minute swim and killed 24 hours later; and exposed to 2 swims, separated by 24 hours and killed after the second swim. Outcomes were swim behavior, serum corticosterone, and 2-AG and 2-oleoylglycerol (2-OG) contents in amygdala, hippocampus and prefrontal cortex.
Mean 2-AG contents were not significantly different among the four treatment groups in any brain region and did not correlate with immobility in either forced swim exposure. However, 2-AG contents in all three brain regions only of the mice exposed to two swims were significantly, positively correlated with serum corticosterone concentrations measured at the same time. 2-OG is present in brain and exhibits a striking regional heterogeneity in control mice. 2-OG concentrations in prefrontal cortex were significantly reduced in the mice killed on the second day compared to the mice killed on the first day. As the target of 2-OG in brain is not known, the significance of these observations await further studies.
Although prior exposure to swim stress does not alter brain 2-AG contents upon re-exposure, 2-AG concentrations in brain become significantly correlated with the HPA axis response to stress when prior exposure to the stress has occurred. These data suggest that even a single exposure to a short period of intense stress sensitizes the 2-AG response to re-exposure to that situation and are consistent with a role for endocannabinoid signaling in modulating stress responses.
Endocannabinoid signaling in the CNS is an important mechanism for short- and long-term changes in synaptic strength. The endocannabinoid system (ECS) consists of presynaptic cannabinoid receptors (CB1R) and two lipid ligands, the endocannabinoids (Freund et al., 2003). The endocannabinoids are the arachidonates 2-arachidonoylglycerol (2-AG) and N-arachidonylethanolamine (AEA, also called anandamide). Both are synthesized from phospholipid precursors in response to several triggers, including increased neuronal activation (Hillard, 2000) and metabotropic receptor activation (Maejima et al., 2001, Varma et al., 2001). There is no evidence for vesicular storage of the endocannabinoids, thus their concentrations in brain tissue provide an estimate of the relative rates of their synthesis and degradation (Buczynski and Parsons, 2010).
Recent studies support the hypothesis that the endocannabinoid signaling system in brain is an important modulator of stress responses, particularly when the stress is repeated (Patel and Hillard, 2008). While a single episode of 30 minute restraint does not reliably alter 2-AG contents in brain, repeated exposure of mice to the same restraint stress increases 2-AG contents in limbic brain regions after 5–10 repetitions (Patel et al., 2005, Rademacher et al., 2008). Patel and colleagues have recently demonstrated that repeated restraint also sensitizes 2-AG mediated signaling at inhibitory synapses in the amygdala (Patel et al., 2009), suggesting that elevated tissue concentrations of 2-AG reflect increased endocannabinoid signaling. Further work from this group demonstrated that repeated restraint down-regulates monoacylglycerol lipase activity, resulting in decreased 2-AG clearance (Sumislawski et al., 2011).
The Porsolt forced swim test consists of two short periods of forced swim separated by 24 hours (Porsolt et al., 1978). Our goal in this study was to use this well-characterized behavioral test as a short, intense stress exposure to determine whether 2-AG signaling was different immediately after the second stress exposure compared to the first. This approach has allowed us to more fully characterize the interactions between stress and 2-AG concentrations in limbic regions of the brain.
There is very little in the literature about 2-oleoylglycerol (2-OG), a structural analog of 2-AG. 2-OG is liberated from triacylglycerols during digestion in the intestinal lumen (Mu and Hoy, 2004). It is present in the circulation and is significantly elevated in the serum from hibernating woodchucks (Marmota monax) compared to active animals (Vaughn et al., 2010). 2-OG has recently been reported to act as an agonist of a class 1 orphan G protein coupled receptor, GPR119 (Hansen et al., 2011). The GPR119 transcript is found in brain (Overton et al., 2006); however, nothing is known about its function. Because it is chemically similar to 2-AG, we have also measured its concentration in brain as an exploratory project.
All of these studies were carried out using mice housed on a 12:12-hour light:dark cycle (lights on at 0600) in American Association for Accreditation of Laboratory Animal Care (AAALAC) approved facilities. All experiments were carried out in accordance with the NIH Guide for the Use and Care of Laboratory Animals and approved by the Institutional Animal Care and Use Committee of the Medical College of Wisconsin. Male C57Bl/6N mice were ordered from Harlan Laboratories (Madison, WI, USA) and allowed to acclimate to on-site housing for at least 10 days. Experiments were performed between 0900–1200. All mice were given ad libitum access to standard mouse chow and water.
A modification of the Porsolt swim test (Porsolt et al., 1978) was used as the stress in this study. Mice were exposed to either one or two swim sessions; in all studies, the swim sessions lasted for 6 minutes. When mice were exposed to two swim sessions they occurred approximately 24 hours apart. The swim sessions occurred during the light phase under normal lighting conditions and behaviors during the swims were recorded using a digital camera. One-liter glass beakers (12 cm diameter and 15 cm height) were filled with 700 – 800 ml of tap water. This resulted in a depth of 9 – 10 cm, which did not allow the mice to touch the bottom of the beaker with their hind-limbs when their heads were above water. The water temperature was 24 ± 2°C. Swim behavior was analyzed by scorers blinded to the experimental group. Time immobile was defined as no movement other than that required to maintain the animal's balance or keep its head above water. Immobility was scored during the last 5 minutes of the swim test. The change in immobility was calculated in animals that received both swims as the immobility during swim one subtracted from immobility during swim two. Thus an animal with greater immobility during swim two has a positive value for the change in immobility.
The mice were divided into four groups; group 1 was not exposed to swim; groups 2 and 3 were exposed to a single swim; and group 4 was exposed to 2 swims separated by 24 hours. Brains and blood were harvested from the 4 groups of mice as follows: group 1 – immediately upon disturbing the home cages; group 2 – immediately after the first FST; group 3 – 24 hours after the first FST when the second FST would have been administered; and group 4 – immediately after the second FST. Mice were killed by decapitation immediately after cervical dislocation. Brains were removed and frozen in liquid nitrogen within 4 minutes of death. Brains were stored at −80°C until they were dissected. Dissections of hippocampus, amygdala and ventral medial prefrontal cortex were carried out on dry ice as shown (Fig. 1). The dissected brain regions were stored at −80°C until used for analysis.
Brain regions were weighed and extracted individually. Tissue was placed into borosilicate glass culture tubes with 2 ml of acetonitrile containing deuterated standard (46.5 pmol [2H8] 2-AG). Samples were homogenized with a glass rod and sonicated on ice for 30 minutes. After overnight incubation at −20°C to precipitate proteins, samples were centrifuged at 1000 g for 5 minutes. The resulting supernatant was transferred to a new glass tube and evaporated to dryness under nitrogen gas. Methanol (120 μl) was added along the sides of the tube to resuspend any particulates, and the sample was dried under nitrogen again. Once completely dry, lipid extracts were suspended in 20 μl acetonitrile, placed into an amber glass vial and frozen at −80°C until analysis.
Lipids were separated and quantified using isotope-dilution, liquid chromatography-mass spectrometry (LC/MS). Samples (5 μl) were separated on a reverse-phase C18 column (Kinetex, 100 × 2.1 mm, 2.6 μm diameter; Phenomenex) using mobile phase A (deionized water, with 0.005% acetic acid) and mobile phase B (acetonitrile with 0.005% acetic acid). Samples were eluted at 300 μl/minute flow rate by linear gradient (Solvent B increasing linearly from 62–67% over 7 minutes, then 67–100% over 5 minutes, then held at 100% for 7 minutes). Detection was made in positive ion mode. Selective ion monitoring was used to detect [2H8] 2-AG and 1(3)-AG (m/z 387; retention times = 9.4 and 10.2 minutes, respectively), 2-AG and 1(3)-AG (m/z 379; retention times = 9.7 and 10.5 minutes, respectively) and 2-OG and 1(3)-OG (m/z 357; retention time = 13.7 and 14.5 minutes, respectively). Monoacylglycerols are often observed as a doublet as they isomerize to 1(3)-AG or 1(3)-OG during extraction (Stella et al., 1997). Therefore, the areas of both peaks were combined to yield total 2-AG and 2-OG.
Following decapitation, blood was collected from the trunk within 4 minutes of the termination of FST exposure or disturbing the control mice in their home cages. Blood samples were centrifuged at 10,000 g for 30 seconds. Serum was removed and frozen at −20°C until assayed. Corticosterone concentrations were determined by radioimmunoassay (MP Biomedicals, Solon, OH). All measurements were carried out in a single assay; the intra-assay coefficient of variation for the assay (provided by the manufacturer) is 4.4–10.3%.
Statistical analyses were performed using GraphPad Prism software version 4.0a for Macintosh (San Diego, CA, USA). Data were analyzed using ANOVA, unpaired and paired t tests as appropriate. Pearson correlations were performed as needed. Data are presented as the mean ± SEM. Comparisons in which p values were less than 0.05 were considered statistically significant.
The endocannabinoid, 2-AG, and its congener, 2-OG, were quantified in lipid extracts of the prefrontal cortex, hippocampus and amygdala (Figure 2). The effects of FST exposure on each lipid were analyzed using two-way ANOVA with relationship to the presence or absence of a swim session (i.e. with or without; “exposure”) and the day of the swim protocol (i.e. one or two; “day”) as the two main factors. There were no significant differences in 2-AG content between any of the groups in the three brain regions analyzed according to two-way ANOVA (Figure 2A) [prefrontal cortex (day F1,21 = 3.1, p>0.05, exposure F1,21 = 0.2, p>0.05 and interaction F1,21 = 0.003, p>0.05), hippocampus (day F1,20 = 2.2, p>0.05, exposure F1,20 = 0.01, p>0.05 and interaction F1,20 = 0.2, p>0.05) and amygdala (day F1,20 = 2.3, p>0.05, exposure F1,20 = 1.5, p>0.05 and interaction F1,20 = 1.4, p>0.05)].
2-OG content was determined in the same lipid extracts (Figure 2B). Interestingly, the content of 2-OG showed a distinct heterogeneity among the brain regions examined. One way ANOVA of the data from the control mice indicates a significant difference among the groups (F2,14 = 5.4, p<0.05) and Tukey's Multiple Comparison test indicates that the concentration of 2-OG in the prefrontal cortex is significantly greater than in the amygdala (q=4.6, p<0.05). Two-way ANOVA of the data indicates a significant effect of day on 2-OG content in the prefrontal cortex (day F1,21 = 7.8, p<0.05). There was neither a significant effect of swim exposure (exposure F1,21 = 0.006, p>0.05) nor an interaction between day and swim exposure (F1,21 = 0.006, p>0.05) on the 2-OG measurements. A similar trend for 2-OG to decrease 24 hours after the initial swim exposure was observed in the amygdala (day F1,20 = 3.8, p=0.07, exposure F1,20 = 1.0, p>0.05 and interaction F1,20 = 0.13, p>0.05) and hippocampus (day F1,20 = 2.4, p=0.14, exposure F1,20 = 0.02, p>0.05 and interaction F1,20 = 0.24, p>0.05).
We examined the correlations between 2-AG and 2-OG concentrations in each brain region and each treatment group (Table 1). 2-AG and 2-OG concentrations were not correlated in any brain region examined in control mice; there was a significant correlation in the prefrontal cortex from animals killed immediately after swim one, that correlation is dependent upon one animal that had very high amounts of both lipids (Fig 3A). Interestingly, a significant, positive relationship between 2-OG and 2-AG concentrations is present in the amygdala of mice killed 24 hours after the first swim exposure (Fig 3B).
Serum corticosterone concentrations were significantly higher after the swim exposure than before and this change was not affected by the number of swims according to two-way ANOVA (Fig. 4A; day F1,21 = 0.04, p>0.05, exposure F1,21 = 112.1, p< 0.0001 and interaction F1,21 = 0.01, p>0.05).
After the second swim exposure, there was a significant, positive correlation between serum corticosterone concentrations and 2-AG concentrations in the prefrontal cortex (Fig. 4B; p<0.05, Pearson r = 0.79), hippocampus (Fig. 4C; p<0.05, Pearson r = 0.76) and amygdala (Fig. 4D; p<0.05, Pearson r = 0.87). After the first swim exposure there was no correlation between serum corticosterone and 2-AG content in any brain region examined (Table 2). There was no correlation between serum corticosterone obtained before the first or second FST exposure and 2-AG content in any brain region examined. Serum corticosterone does not correlate with 2-OG concentrations in any of these three brain regions under any of the four experimental conditions (Table 2).
The mean times immobile during the first and second FST were compared using an unpaired t test. As expected, mice exhibited significantly more immobility during swim two than swim one when all the mice were compared (Fig. 5A; t25 = 2.7, p<0.05). In the mice that received both swim exposures (i.e. group 4), paired t test also confirmed significantly more immobility during swim two than swim one (Fig. 5B; t6 = 4.2, p<0.01). There was no correlation between immobility during swim one or two and serum corticosterone concentrations measured immediately after swim exposure (Table 3). The change in immobility was calculated in the mice that were exposed to both swim sessions (group 4). There was no correlation between the change in immobility and serum corticosterone concentrations measured immediately after the second swim exposure. There were no significant correlations between immobility during swim one, two or the change in immobility and monoacylglycerol content in any brain region examined (Table 3).
Although the mean brain regional contents of 2-AG were not different in mice exposed to swim stress compared to controls, 2-AG contents in three brain regions were significantly, positively correlated with circulating concentrations of corticosterone in mice only after the second swim exposure. These data support the hypothesis that limbic 2-AG concentrations and circulating glucocorticoids are mechanistically related under some circumstances.
We report for the first time that 2-OG is present in the prefrontal cortex, hippocampus and amygdala. 2-OG concentrations are not correlated with the concentrations of 2-AG in control mice, suggesting that the mechanisms regulating their synthesis and/or degradation are not overlapping. Interestingly, while 2-AG concentrations in control mice are not different among the three brain regions examined, 2-OG concentrations are quite variable. The concentrations of 2-OG are 100, 60 and 7 pmol/mg in prefrontal cortex, hippocampus and amygdala, respectively. The source of 2-OG in the brain is not known. While it is possible that 2-OG is synthesized locally, another possibility is that 2-OG enters the brain from the circulation. Remarkably, 2-OG concentrations in the prefrontal cortex are reduced by more than 50% in mice exposed to a swim 24 hours earlier, regardless of whether they were re-exposed to the swim stress. An interesting hypothesis that is suggested by these data is that swim stress alters circulating 2-OG concentrations, perhaps as a result of an effect on food intake or intestinal function, which are reflected as a change in prefrontal cortical concentrations. Recent data indicate that 2-OG is an agonist of the GPR119 receptor (Hansen et al., 2011). GPR119 is found at high concentrations in pancreas and the GI tract and its activation is associated with increased release of glucagon-like peptide 1, a physiological regulator of meal-induced glucagon secretion (Shah and Kowalski, 2010) and insulin secretion and satiety (Vrang and Larsen, 2010). GPR119 mRNA is expressed ubiquitously in brain (Overton et al., 2006), nothing is known about its function or protein distribution. Therefore, an outcome of this study is the identification of 2-OG as a lipid in brain that changes in response to stress and has a distinct distribution in the three brain regions examined.
Several other studies have examined 2-AG concentrations in brain tissue that is harvested immediately after exposure to stress, and the majority of these demonstrate that 2-AG contents in limbic brain regions do not change significantly after a single 30 minute restraint exposure (Carrier et al., 2005, Rademacher et al., 2008, Hill et al., 2009), which is consistent with the current results. Patel and colleagues also reported that amygdalar 2-AG is unchanged 20 minutes into a restraint episode (Patel et al., 2009). However, 2-AG content has been shown to increase immediately after a 30 minute restraint in the hypothalamus (Evanson et al., 2010) and the hippocampus (Wang et al., 2011) and is increased in the prefrontal cortex 30 minutes after the offset of the restraint (Hill et al., 2011). Based on these data, we hypothesize that a single, acute stress can recruit 2-AG mediated signaling in many brain regions but that the timing of the increase in 2-AG with respect to the stress episode is not the same in all brain regions. From an experimental point of view, detection of changes in bulk tissue 2-AG contents can be difficult when only one time point is examined. While changes to tissue 2-AG contents are inconsistently detected following a single stress exposure, re-exposure of rats and mice to the same stressor 5–10 more times has been shown to robustly increase 2-AG contents in limbic regions of the brain (Patel et al., 2005, Rademacher et al., 2008, Patel et al., 2009). Therefore, the available data are consistent with the hypothesis that subsequent exposures to the same stress induce a larger and more prolonged increase in 2-AG compared to the first exposure to that stressor. Recent data from Patel and colleagues demonstrated that, in amygdala, repeated restraint but not a single restraint, decreases MAGL expression (Sumislawski et al., 2011), which is a plausible mechanism for the enhanced 2-AG response to repeated homotypic stress.
The design of the present study differed from these earlier study in several ways. First, the stressor was a forced swim rather than restraint. Second, the stress was shorter, only 6 minutes compared to 30 minutes of restraint. In spite of the difference in the duration of the stress, mice were killed at the end of the stress. Third, we have compared tissue 2-AG contents in mice exposed to one or two stress episodes. 2-AG contents were not different between stressed and unstressed mice, which is not surprising given the short duration of the stress and the time at which the mice were killed. However, when the concentrations of 2-AG in brain tissue and the concentrations of glucocorticoids in blood are considered in the same mouse, 2-AG did not correlate with circulating corticosterone after the first episode of restraint, but it did after the second exposure to the same restraint 24 hours later. It is our hypothesis that tissue 2-AG concentration is a relatively insensitive measure of the change in 2-AG-mediated signaling when considered in isolation. However, when it is considered in tandem with another stress-sensitive marker, a significant difference between the 2-AG response to the first and the second swim stress episode is apparent.
We have found significant positive correlations between 2-AG in three brain regions and circulating corticosterone. There are three possible mechanisms that could underlie these correlations. First, 2-AG concentrations in brain could regulate the change in circulating corticosterone via effects on the HPA axis. This is not very likely since increased 2-AG in the prefrontal cortex (Hill et al., 2011) and hippocampus (Wang et al., 2011) decrease HPA axis activation and would thus be negatively associated with circulating corticosterone. The second possibility is that circulating corticosterone enters the brain and increases 2-AG contents through GR activation. Support for this mechanism comes from data that GR receptor blockade inhibits stress-induced increases in 2-AG in prefrontal cortex and hippocampus at 30 minutes after the offset of stress (Hill et al., 2011, Wang et al., 2011). On the other hand, previous data from our laboratory demonstrated that exogenous administration of corticosterone did not affect 2-AG contents in prefrontal cortex, hippocampus or amygdala 10 minutes after injection (Hill et al., 2010). Therefore, it is likely that the effect of GR to increase 2-AG in limbic regions (but not hypothalamus (Evanson et al., 2010)) is time-delayed with respect to the increase in corticosterone, possibly because it is a genomic action of GR. Another reason that we do not favor this mechanism is that corticosterone was increased to the same extent after the first and the second swim; but corticosterone and 2-AG only correlate after the second swim. A third possible mechanism is that limbic 2-AG and circulating corticosterone correlate because they are both evoked by the same stimulus. In other words, they increase in parallel, not in series. It is our current hypothesis that this stimulus is one of the early components of the stress response, for example, arousal or CRF. We hypothesize that 2-AG content is not regulated by this factor in response to the first exposure to the swim, but the first exposure results in a biochemical change that links 2-AG concentrations to a repeat exposure. It is not clear from these studies whether the phenomenon observed here is an early event in the process by which several repeats of the same stress ultimately result in a measurable increase in 2-AG concentrations.
There are two interesting findings in this paper. First, 2-OG is present in brain and its concentrations vary considerably among brain regions. In addition, 2-OG concentrations in prefrontal cortex are significantly reduced in all mice exposed to a swim stress 24 hours later. Second, 2-AG tissue concentrations are not different among the four groups of mice examined. However, 2-AG concentrations and circulating glucocorticoid concentrations become significantly correlated after the second swim exposure. They are not correlated after the first swim, although corticosterone increases to the same extent. We hypothesize that exposure to a significant stress alters 2-AG biochemistry such that a subsequent exposure to the same stress impacts 2-AG via a mechanism that we do not yet understand. Although 2-AG concentrations do not increase, we hypothesize that 2-AG contents in three limbic brain regions become linked to the stress response when a stress is repeated. This process could contribute to learning.
These studies were supported by NIDA grants R01 DA09155 and DA026996 and an ARRA supplement to grant DA09155.
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.