Antidepressant-like activity of the fatty acid amide hydrolase inhibitor URB597 in a rat model of chronic mild stress.

Background: Theendocannabinoidanandamidemaybeinvolvedintheregulationofemotionalreactivity.Inparticular,ithasbeenshown that pharmacological inhibition of the enzyme fatty acid amide hydrolase (FAAH), which catalyzes the intracellular hydrolysis of anandamide, elicits anxiolytic-like and antidepressant-like effects in rodents. Methods: We investigated the impact of chronic treatment with the selective FAAH inhibitor, URB597 (also termed KDS-4103), on the outcomes of the chronic mild stress (CMS) in rats, a behavioral model with high isomorphism to human depression. Results: Daily administration of URB597 (.3 mg·kg (cid:1) 1 , intraperitoneal [IP]) for 5 weeks corrected the reduction in body weight gain and sucrose intake induced by CMS. The antidepressant imipramine (20 mg·kg (cid:1) 1 , once daily, IP) produced a similar response, whereas lower doses of URB597 were either marginally effective (.1 mg·kg (cid:1) 1 ) or ineffective (.03 mg·kg (cid:1) 1 ). Treatment with URB597 (.3 mg·kg (cid:1) 1 ) resulted in a profound inhibition of brain FAAH activity in both CMS-exposed and control rats. Furthermore, the drug regimen increased anandamide levels in midbrain, striatum, and thalamus. Conclusions: URB597 exerts antidepressant-like effects in a highly speciﬁc and predictive animal model of depression. These effects may depend on the ability of URB597 to enhance anandamide signaling in select regions of the brain.

Various classes of FAAH inhibitors have been developed (Piomelli 2005). Among them is a group of O-aryl-carbamates that inhibit FAAH activity with high potency and selectivity (Fegley et al. 2005;Kathuria et al. 2003;Mor et al. 2004;Tarzia et al. 2003). A well-characterized member of this class, the compound URB597 (also termed KDS-4103), blocks rat brain FAAH activity in vivo with a half-maximal inhibitory dose (ID 50 ) of .15 mg·kg Ϫ1 (intraperitoneal [IP]), increases brain anandamide levels, and enhances the pharmacological effects of exogenous anandamide (Fegley et al. 2005;Kathuria et al. 2003). URB597 does not elicit a typical spectrum of cannabinoid responses (Kathuria et al. 2003)-for example, it does not produce catalepsy, hypothermia, or hyperphagia (Fegley et al. 2005)-but does exert significant analgesic, anxiolytic-like, and antidepressant-like effects. For example, it reduces nocifensive behavior in rodent models of inflammatory pain (Holt et al. 2005;Jayamanne et al. 2006), increases the amount of time spent by rats and mice in the open arms of an elevated maze (Kathuria et al. 2003;Patel and Hillard 2006), lowers isolation-induced vocalizations in rat pups (Kathuria et al. 2003), and decreases plasma corticosterone levels in restrained mice (Patel et al. 2004). In addition, URB597 enhances stress-coping responses in the rat forced swim test and the mouse tail suspension test (Gobbi et al. 2005). Notably, these effects are not accompanied by overt rewarding actions, as suggested by the inability of URB597 to induce conditioned place preference or substitute for ⌬ 9 -THC in a drug-discrimination paradigm (Gobbi et al. 2005). Thus, unlike ⌬ 9 -THC and other directacting cannabinoid receptor agonists (Lichtman and Martin 2005), URB597 appears to be devoid of significant abuse potential.
The effects of URB597 on stress-coping behaviors are consistent with evidence indicating the involvement of endocannabinoids in the pathophysiology of depression. Thus, pharmacological blockade of endocannabinoid transport elicits antidepressant-like responses in the rat forced swim test (Hill and Gorzalka 2005). Moreover, cannabinoid-1 (CB1) deficient mice display a series of behavioral alterations related to human depressive symptoms, including anhedonia, anxiety, enhanced stress responsiveness, and feeding dysregulation (Cota et al. 2003;Martin et al. 2002;Sanchis-Segura et al. 2004).
Acute stress based models, such as the forced swim and the tail suspension tests, are inadequate to investigate the neurobiological relationships between chronic stress and the pathogenesis of mood disorders and fail to reproduce the constellation of symptoms observed in depression. An experimental paradigm that mimics more closely several of such symptoms is chronic mild stress (CMS) (Willner et al. 1987). This model is based on the finding that rodents exposed to an unpredictable, relatively continuous sequence of mild stressors develop a series of abnormal behavioral and physiological responses that are reminiscent of those observed in depressed patients (DSM-IV 296.2X-296.3X [American Psychiatric Association 1994]): decreased responsiveness to rewarding stimuli such as palatable foods, alterations in feeding and body weight (BW), enhanced fearfulness, impaired sleep architecture, and inadequate self-care (Willner 2005). Importantly, these abnormalities are corrected by chronic, but not acute, administration of antidepressant drugs. For these reasons, CMS is currently regarded by many investigators as one of the most naturalistic and predictive animal models of depression (Henn and Vollmayr 2005;Willner 1997;Willner 2005).
In the present study, we examined the behavioral impact of URB597 in rats exposed to CMS. We monitored changes in body weight and sucrose intake as indices of stress-induced responses, since both parameters are consistently reduced by CMS in an antidepressant-sensitive manner (Moreau 1997;Willner 1997).

Subjects
One hundred twenty male Wistar rats (Charles River, Raleigh, North Carolina), weighing approximately 200 g upon arrival at the animal facilities, served as subjects in the present study. They were individually housed in rooms maintained on a 12-hour light-dark cycle (lights on at 7:00 AM and off at 7:00 PM) under equivalent conditions of temperature (22 Ϯ 2°C) and relative humidity (55% to 60%) Food and water were available ad libitum, except during the periods of food and water deprivation required by the stress protocol or before assessment of sucrose consumption. All experimental procedures were approved by the local ethics committee and carried out in strict accordance with the National Institutes of Health guidelines for care and use of experimental animals.

Drugs
URB597 was obtained from Kadmus Pharmaceuticals (Irvine, California) and the Institute of Medicinal Chemistry, University of Urbino "Carlo Bo" (Italy). The compound was prepared for administration as follows: 3 mg was suspended in .5 mL polyethylene glycol (PEG-400) (Sigma-Aldrich, St. Louis, Missouri), vigorously stirring the mixture with a spatula; .5 mL Tween-80 (Sigma-Aldrich) was added, and the mixture was sonicated for 5 min, obtaining a clear solution. Nine milliliters of .9% saline were added and the resulting suspension was sonicated for 10 min at room temperature. Thus, the final drug concentration was .3 mg/mL in a vehicle of PEG-400, Tween-80, and saline (5:5:90, vol/vol). Imipramine hydrochloride was purchased from Sigma-Aldrich and dissolved in saline with a few drops of .1 mol/L sodium hydroxide (NaOH) to normalize pH.

Experimental Procedures
Sucrose Consumption. Upon arrival, the rats were randomly assigned to one of two separate rooms (60 rats per room), where they were housed in individual cages. After an initial habituation period of 10 days, all animals were tested for baseline sucrose consumption. They were deprived of food and water for 15 hours prior to the test, starting 1 hour before the onset of the dark phase. Each animal was given access to one preweighed bottle containing a 1% sucrose solution in tap water. One hour later, the bottle was removed and weighed again, and food and water were placed back in the cage. Sucrose consumption tests were repeated every 3 to 4 days for the following 2 weeks. After stabilization of sucrose consumption, baseline sucrose preference was assessed. The preference test was similar to the consumption test, except that the animals were given the choice between a 1% sucrose solution and plain water. After preference assessment, each group was divided into five matched subgroups (n ϭ 12 each) with sucrose consumption and preference equally distributed across subgroups. The preference test was run only once, immediately before the beginning of the CMS protocol.
CMS Protocol. In one of the two rooms, rats were subjected to a random sequence of mild stressors (Willner et al. 1987). These included cage soiling with water, group housing in a confined space, water and/or food deprivation, intermittent lighting, reversal of light/dark cycle, cage tilting to 45°, exposure to loud white noise, and strobe lights (300 flashes/min). Stressors were individually applied in succession and lasted 8 to 12 hours each. Control rats housed in the second room were not subjected to the stress protocol but were food and water deprived for a period of 15 hours preceding sucrose consumption tests like CMS-exposed rats. Sucrose intake and body weight were measured weekly for the duration of the study.
Drug Treatments. Starting 5 weeks after the beginning of the CMS protocol, control and CMS-exposed rats received daily injections of vehicle, URB597 (.03, .1, .3 mg·kg Ϫ1 , IP), or imipramine (20 mg·kg Ϫ1 , IP). All drugs were freshly prepared and administered in 1 mL·kg Ϫ1 of vehicle. All the injections were made between 4:00 PM and 6:00 PM, irrespective of the stress schedule.
Lipid Analyses. At the end of the study, the rats were anesthetized with halothane and decapitated during the application of the last stressor (2 hours after the last drug or vehicle injection). Brains were removed within approximately 30 sec of decapitation and select regions were dissected using a rat brain atlas (Paxinos and Watson 1986) as a guide, frozen in dry ice, and stored at -80°C until analysis. Endocannabinoids and related lipids were extracted with methanol-chloroform, fractionated by open-bed silica gel chromatography, and quantified by isotopedilution liquid chromatography/mass spectrometry (LC/MS), as described by Fegley et al. (2005).

Data Analysis
Results were analyzed using multifactor analysis of variance (MANOVA), followed by Spjotvoll-Stoline correction for Tukey's test for post hoc comparisons. Multivariate comparisons for repeated measures were analyzed with the Dunn-Sidak correction for Bonferroni test. Throughout the study, the impact of CMS was studied calculating absolute sucrose intake (ASI) and body weight variations. To test whether changes in sucrose intake could be affected by body weight alterations, we also analyzed differences in relative sucrose consumption, defined as the ratio between absolute sucrose intake and body weight (Bekris et al. 2005;Willner et al. 1996). We used the Pearson product-moment correlation coefficient to examine the correlation between individual differences in absolute sucrose intake and body weight between the onset of the CMS protocol and treatment initiation. Figure 1 illustrates the effects of CMS on body weight, absolute sucrose intake, and relative sucrose intake (RSI) in the two groups of animals used in the present study (n ϭ 60 each).

Effects of URB597 on Body Weight Gain
The effects of vehicle, URB597 (.3 mg·kg Ϫ1 ), and imipramine (20 mg·kg Ϫ1 ) on BW gain are shown in Figure 2. Daily vehicle administration did not affect the BW trajectories of control and CMS-exposed rats (Figure 2A) [interaction treatment ϫ time in CMS-exposed rats: F (40,490) ϭ 2.85 p Ͻ .0001, MANOVA; p Ͻ .05 for comparisons between control-vehicle and CMS-vehicle for weeks 3-10, Dunn-Sidak]. By contrast, URB597 administration normalized BW gain in CMS-exposed rats, bringing it to control levels within 4 weeks of treatment ( Figure 2B) (p Ͻ .05  , and (C) imipramine (20 mg·kg Ϫ1 , IP) on BW of CMS-exposed and control rats; (D) comparison of effects of vehicle, URB597, and imipramine in CMS-exposed and control rats at the 10th week. Values are expressed as mean Ϯ SEM , control rats; OE, CMS-exposed rats. *p Ͻ .05 versus control rats; **p Ͻ .01 versus control rats; #p Ͻ .05 versus CMS-exposed, vehicle-treated rats. IP, intraperitoneal; BW, body weight; CMS, chronic mild stress.
for comparisons between control-URB597 .3 and CMS-URB597 .3 for weeks 3-8, Dunn-Sidak). Imipramine exerted an effect similar to, albeit somewhat faster than, URB597 ( Figure 2C) (p Ͻ .05 for comparisons between control-imipramine and CMSimipramine for weeks 3-6, Dunn-Sidak). Lower doses of URB597 (.03 and .1 mg·kg Ϫ1 ) were ineffective ( Figure 2D). In control rats, URB597 did not produce BW changes ( Figure 2B, 2D) [interaction treatment ϫ time in control rats: F (40,530) ϭ 1.37, ns, two-way analysis of variance (ANOVA) for repeated measures], while imipramine caused a significant BW decrease ( Figure 2D) [F (1,21) ϭ 4.40, p Ͻ .05]. These findings suggest that URB597 corrects the loss in BW gain caused by CMS without affecting BW gain in nonstressed rats. Lower doses of URB597 showed a similar trend but did not reach statistical significance ( Figure 3C). Imipramine exerted an effect quantitatively comparable with that of URB597 (p Ͻ .05 for comparisons between vehicle and URB597, Spjotvoll-Stoline), which was significant starting from the fourth week of treatment ( Figure  . We interpret these results to indicate that URB597 normalizes sucrose consumption in CMS-exposed rats but does not influence this parameter in nonstressed rats.

Discussion
The main finding of the present study is that chronic treatment with the potent and selective FAAH inhibitor URB597 normalizes BW gain and sucrose intake in rats exposed to CMS, a highly specific behavioral model of depression. These findings confirm and extend previous results, showing that URB597 enhances acute stress-coping responses in the mouse tail suspension test and the rat forced swim test (Gobbi et al. 2005;Kathuria et al. 2003;Patel et al. 2004;Patel and Hillard 2006).
Chronic mild stress was first formalized as an animal model of depression in 1987 (Willner et al. 1987) and, despite some initial criticisms (Forbes et al. 1996), has been since validated by numerous independent studies (Willner 2005). This paradigm offers important advantages over other models of depression: first, it induces in rodents an array of abnormal behaviors that are isomorphic with key features of human depression (Willner 2005); second, CMS-exposed rats-like depressed patients-respond to chronic, but not acute, treatment with antidepressant drugs (Willner 1997); third, psychostimulant agents, which can give false positives in acute stress-based models (De Pablo et al. 1989;Duncan et al. 1985;Steru et al. 1985), have no such effects in CMS .   The present results show that BW gain and sucrose intake are significantly reduced by CMS and are corrected by chronic treatment with URB597. Changes in sucrose intake are thought to reflect variations in operant appetitive behaviors and, therefore, to be a reliable index of anhedonic-like responses (Moreau 1997;Willner 1997). The alternative possibility that variations in sucrose intake might be secondary to BW changes (Forbes et al. 1996;Matthews et al. 1995) was ruled out by the lack of statistical correlation between CMS-induced variations in BW and ASI. Moreover, in agreement with previous findings (Willner 1997(Willner , 2005, CMS markedly affected RSI, a composite measure of intake normalized to BW. URB597 selectively normalized BW gain, ASI, and RSI in CMS-exposed but not control rats, suggesting that the FAAH inhibitor specifically targets neural processes altered by stress without affecting normal appetitive responses. Consistent with this idea, prior reports have shown that URB597 is devoid of orexigenic or rewarding properties (Fegley et al. 2005;Gobbi et al. 2005).
Previous research has shown that the CB1 antagonist rimonabant prevents the mood-enhancing actions of URB597, suggesting that such effects are based on the drug's ability to magnify endogenous anandamide signaling at CB1 receptors (Gobbi et al. 2005;Kathuria et al. 2003). In the present study, we could not provide a similar demonstration because of the intrinsic effects of rimonabant on BW, appetite, and hedonic drive to food (Arnone et al. 1997;Colombo et al. 1998), as well as on other behavioral end points affected by CMS (Griebel et al. 2005). Nevertheless, the idea that anandamide may be responsible for the effects of URB597 in CMS is supported by two sets of results. Firstly, direct cannabinoid agonists produce antidepressant-like responses in both CMS (Onaivi et al. 2006) and forced swim tests (Jiang et al. 2005). Secondly, as documented in the present report, chronic treatment with URB597 elevates anandamide levels in brain regions, such as midbrain and striatum, which are thought to participate in the induction of CMS-related behaviors (Bekris et al. 2005;Dziedzicka-Wasylewska et al. 1997;Papp et al. 1994).
It is also important to point out that the chronic URB597 regimen used in the present study did not increase anandamide levels in hippocampus and prefrontal cortex, even though it fully inhibited FAAH activity throughout the brain. This lack of effect, at variance with that of acutely administered URB597 (Fegley et al. 2005;Gobbi et al. 2005;Kathuria et al. 2003), suggests that prolonged exposure to the FAAH inhibitor may result in a down-regulation of anandamide mobilization in certain brain regions (e.g., hippocampus) but not others (e.g., midbrain). The mechanism underlying this hypothetical down-regulation of anandamide signaling remains undetermined.
Since acute or subchronic treatment with URB597 enhances firing of serotonergic and noradrenergic neurons in the midbrain (Gobbi et al. 2005), it is possible that persistent anandamide accumulation in this region might contribute to the antidepressant-like effects of URB597 in CMS (Haidkind et al. 2003;Liu et al. 2003;Papp et al. 1994). Testing this hypothesis will require, however, further experimentation.
In apparent contrast with our findings, subchronic treatment with rimonabant was found to improve hair coat appearance and reduce CMS-induced immobility in the forced swimming test in mice subjected to CMS (Griebel et al. 2005). We cannot provide an adequate explanation for this discrepancy, but two possibilities should be considered. Rimonabant may interact with an as yet uncharacterized cannabinoid-sensitive receptor distinct from CB1 or with a subpopulation of CB1 receptors distinct from those engaged by neurally released anandamide (Breivogel et al. 2001;Hajos et al. 2001;Haller et al. 2002;Rodgers et al. 2005). Alternatively, mice may differ from rats in their response to cannabinoid-acting drugs. In this context, it should be noted that rimonabant use in obese humans appears to be accompanied by increased, rather than reduced, anxiety and depression symptoms (Gelfand and Cannon 2006;Pi-Sunyer et al. 2006;Van Gaal et al. 2005).
We found that CMS is associated with an increase in CB1 receptor mRNA expression in the prefrontal cortex. This result is consistent with recent data showing an elevated CB1 receptor density in the prefrontal cortex of rats exposed to 21 days of chronic unpredictable stress (Hillard et al., unpublished data, 2006). Such alteration might be related to multiple neurochemical changes described in CMS-exposed rats, including increased dopaminergic activity and decreased serotonergic activity (Bekris et al. 2005). Conversely, our finding that CB1 mRNA is reduced in the midbrain of CMS-exposed rats might parallel other neurotransmitter abnormalities observed in these animals, such as decreased D 2 receptor signaling in the substantia nigra and ventral tegmental area (Dziedzicka-Wasylewska et al. 1997) or desensitization of 5-hydroxy-tryptamine 1A (5-HT 1A ) autoreceptors in the dorsal raphe nuclei (Froger et al. 2004).
Notably, our CMS protocol did not reproduce the changes in hippocampal CB1 protein and endocannabinoid levels previously reported by . These authors showed that 2-AG and CB1 receptor levels are reduced in the hippocampus of rats subjected to a 21-day chronic unpredictable stress protocol. Duration of exposure to stress might well account for the different outcomes obtained in the two experiments. Irrespective  of these speculations, our results show that prolonged treatment with URB597 opposes the changes in CB1 mRNA expression induced by CMS, without affecting CB1 mRNA levels in control animals. A plausible interpretation of these findings is that persistent FAAH inhibition normalizes neuroplastic changes induced by CMS in rats.
In conclusion, our studies confirm the antidepressant-like properties of URB597, documenting the ability of this FAAH inhibitor to correct the alterations associated with CMS, a naturalistic model of depression with high face, construct, and predictive validity. While the mechanism underlying the antidepressant-like effects of URB597 is still unknown, our results highlight the critical role of anandamide in the regulation of affective states and point to FAAH as a promising new target for the therapy of depression and other stress-related disorders.