Anxiolytic-like effects observed in rats exposed to the elevated zero-maze following treatment with 5-HT2/5-HT3/5-HT4 ligands

The present study examined the effects of administering selective 5-HT antagonists and agonists to rats tested in the elevated zero-maze (EZM) model of anxiety. The EZM paradigm has advantages over the elevated plus-maze (EPM) paradigm with respect to measuring anxiety, yet has been utilized less frequently. Three experiments were conducted each with a diazepam control (0.25, 0.5 and 0.75 mg/kg). In the first experiment, we administered the 5-HT2C antagonist RS 102221 (0.5, 1.0, and 2.0 mg/kg) and 5-HT2C agonist MK-212 (0.25, 0.5 and 0.75 mg/kg); in the second experiment, we administered the 5-HT3 antagonist Y-25130 (0.1, 1.0 and 3.0 mg/kg) and 5-HT3 agonist SR 57227A (0.1, 1.0 and 3.0 mg/kg), and in the third experiment, we administered the 5-HT4 antagonist RS 39604 (0.01, 0.1, 1.0 mg/kg) and 5-HT4 agonist RS 67333 (0.01, 0.1 and 0.5 mg/kg). The administration of 5-HT2/3/4 subtype antagonists all generated behavioral profiles indicative of anxiolytic-like effects in the EZM, which was apparent from examination of both traditional and ethological measures. While little effect was observed from 5-HT2 and 5-HT3 agonists, the 5-HT4 agonist RS 67333 was found to produce a paradoxical anxiolytic-like effect similar to that produced by the 5-HT4 antagonist RS 39604. We conclude by discussing the implications of these findings.


Results
Combined descriptive statistics for the study variables including time spent on open arms, open arm entries, head dips, SAP, risk assessment, and rearing duration/frequency are presented in Table 1 along with the correlations between aggregated study variables. Tables 2-4 respectively include the specific results for the three experiments. Finally, Table 5 includes a summary of the direction of statistically significant findings across all three experiments.    Mann-Whitney U analysis revealed that the moderate dose of SR 57227A (1.0 mg/kg) was associated with a significant increase in risk assessment (U 5 16.0, p , 0.01). No other statistically significant effects were observed across the 0.1 mg/kg and 3.0 mg/kg doses of SR 57227A.    other reports of anxiolytic-like action of 5-HT 2C blockade in animal models of anxiety 28 . An examination of traditional measures of anxiety revealed a general lack of effect for the selective 5-HT 2C agonist MK-212. However, MK-212 was found to significantly reduce the duration of rearing at all doses and to significantly reduce the frequency of rearing at both the lowest and highest doses tested. Factor analysis has found rearing to load on a 'motor activity' factor in the EPM 29 suggesting the possibility that the general lack of effects observed after MK-212 administration may have been due to behavioral inhibition/sedation. Overall, these findings are in agreement with a number of previous studies which also report a general lack of effect of 5-HT 2 agonism in animal models of anxiety 3,30,31 .
The finding that the 1.0 mg/kg dose of Y-25130 increased both time spent on the open arms and total arm entries is in agreement with previous studies examining the effects of 5-HT 3 antagonism on traditional measures in both the EPM and EZM 32,33 and is compatible with an anxiolytic-like profile comparable to that obtained with diazepam. Consistent evidence for this interpretation was also observed from ethological measures with the 1.0 mg/kg dose leading to decreased SAP and risk assessment in conjunction with increased head dips. This latter finding is also in agreement with previous studies including work with the EPM 34 , social interaction test 35 , and light-dark test 17 , in suggesting an anxiolytic effect of 5-HT 3 antagonists in animals. However, it should be noted that these effects were not observed for the smaller (0.1 mg/kg) and larger (3.0 mg/kg) doses suggesting the possibility of a hormetic dose-response curve.
While the 5-HT 3 agonist SR 57227A was found to significantly modify most anxiety-related behaviors as measured in the EZM, examination of the effects of specific doses revealed a general lack of significant effects when compared to the saline control condition. The 0.1 mg/kg dose of this drug failed to modify any behavioral measures, although the general pattern of results, which include non-significant increases in behavior associated with both anxiogenesis and anxiolysis, would suggest non-specific effects.
The 5-HT 4 antagonist RS 39604 significantly increased both time spent on the open arms and total arm entries in a manner consistent with anxiolysis. These findings are in agreement with those of Silvestre and colleagues 11 as well as Kennett and colleagues 28,36 who previously reported an anxiolytic-like effect of the 5-HT 4 antagonists GR 113808, SB 204070, SB 204070A, and SB 207266A on traditional measures in the EPM. With respect to ethological measures, RS 39604 was found to increase head dips and reduce risk assessment, but failed to modify SAP. Overall, RS 39604 was found to modify less anxiety-related behaviors than diazepam in the current experiment. This finding is again in agreement with the work of Silvestre et al. 11 and Kennett et al. 28,36 , both of whom reported that the effects of 5-HT 4 antagonists were smaller than those of either diazepam or chlordiazepoxide in the EPM, leading both authors to conclude that 5-HT 4 antagonists have modest anxiolytic-like activity when compared with the benzodiazepines.
The 5-HT 4 agonist RS 67333 was found to increase both time on the open arm of the EZM and to increase the number of entries onto open arms, a profile consistent with anxiolysis. This interpretation is further supported by ethological measures (i.e., all doses resulted in a reduction in risk assessment with the 0.01 mg/kg and 0.1 mg/kg doses were associated with an increase in head dips). While this finding is in agreement with previous findings in suggesting that 5-HT 4 agonism is associated with a decrease in anxiety-induced behavioral inhibition 37 , the result is surprising in light of the finding that   5-HT 4 antagonism led to a similar anxiolytic profile. One possible explanation for this seeming paradox is a non-rectilinear relation between 5-HT 4 activity and anxiety. An inverted-U shaped anxiolytic profile would be sensitive to 5-HT 4 disruptions whether they were in the form of increased or decreased 5-HT 4 activation. However, further research will be necessary to develop this possibility.
In conclusion, results from this study indicate a potential role for 5-HT 2 , and 5-HT 3 antagonists and an apparently paradoxical role for both 5-HT 4 agonists and 5-HT 4 antagonists in producing anxiolyticlike effects on rats tested in the EZM paradigm. Such results from the use of the EZM can be compared to data from the EPM that is considered comparable following diazepam challenge, although the more direct measure of time spent in the open may represent an advantage of the elevated zero-maze 7 .
The results for the 5-HT 2 , 5-HT 3 and 5-HT 4 antagonists employed in our experiments are consistent with the original hypothesis for the role of serotonin in the pathogenesis of anxiety that was predicated on an association between a reduction in 5-HT turnover and the anxiolytic effects of benzodiazepines suggesting that a reduction of 5-HT neurotransmission results in an anxiolytic-like effect, whereas increased activity produces an anxiogenic-like effect 38 . Notwithstanding this conclusion, the behavioral pharmacology of 5-HT receptor ligands is often more inconsistent than the effects of standard anxiolytics and not all results are explainable in terms of the classic hypothesis, as evinced in our third experiment that demonstrated an anxiolytic-like effect with a compound that increased serotonergic activity.
Moreover, it should be noted standard anxiolytics (e.g., diazepam) sometimes produce counter-intuitive results as well such as the finding in Experiment 2 (see Table 3) that the low dose of diazepam was associated with less time spent on the open arms of the EZM than the saline control condition. Given that data for these groups were collected over a time span of 18 months, we postulated that such variations may, in part, be due to circadian rhythm effects. Benzodiazepines, such as diazepam, are thought to act by potentiating the action of the neurotransmitter c-aminobutyric acid, which in turn, has been linked to regulation of the sleep-wake cycle 39 . Hence, the variations in responses to diazepam may be the result of the benzodiazepine acting at varying levels of GABA that mediate the generation of circadian rhythms.
A dual role for 5-HT in the mediation of different types of fear has been posited by Graeff and Zangrossi 40 . Specifically, serotonin may either enhance or reduce anxiety-like behavior depending upon the receptor subtype involved. Indeed, the fact that selective serotonin reuptake inhibitors are efficacious in the treatment of generalized anxiety disorder and in panic disorder indicates that there are conditions in which increased 5-HT activity can reduce anxiety.

Methods
Animals. All work was licensed by the Home Office via the Northern Ireland Department of Health, Social Services and Public Safety in accordance with the UK Animals (Scientific Procedures) Act 1986 and was in full compliance with Queen's University Belfast's Policy on the Use of Animals in Research and Teaching. In each of the three experiments, one hundred male Sprague-Dawley rats were randomly assigned as follows: saline control (n 5 10), diazepam (n 5 30), 5-HT antagonist (n 5 30), and 5-HT agonist (n 5 30). Animals were supplied from a breeding stock at Queen's University and weighing between 280 and 350 grams were group-housed (5 per cage; cage size 51 3 39 3 19 cm) in a temperature-controlled environment 22 6 1uC under a 12-hour reverse light-dark cycle (lights off at 0800 hours) for four weeks prior to testing and prior to experimentation has been handled only for routine husbandry. Food and water were available ad libitum. All animals were both drug and experimentally naive.  Apparatus. The design of the maze employed was based on that originally proposed by Shepherd et al. 9 and consists of a black Perspex annular platform, 105 cm in diameter and 10 cm wide, elevated 65 cm above the ground. The maze was divided into four equal quadrants, two of which were enclosed with black Perspex walls, 27 cm in height. The walls were on both the inner and outer sides of the platform so as to provide the 'closed' areas of the maze. The 'open' arms had no walls, but did have a 1 cm lip which acts as a tactile aid to animals when on the open areas of the maze. Lighting was provided by two 60 watt red ceiling lights at a height of 2 m positioned directly over the maze and one 60 watt red bulb positioned closer to the maze. Two cameras poised at 45-degree angles were positioned 1 m above the ground and 1 m from either side of the maze. The cameras were connected to a multiplexer in an adjacent laboratory in order to prevent distractions and provided continuous recording of behavior for subsequent analysis. Procedure. All testing was conducted between 0800 and 1400 hours. On the day of testing one animal was selected and transferred in a covered box (to maintain darkness) measuring 32 3 16 3 12.5 cm to a room adjacent to the experimental room. Here a subcutaneous (s.c) injection was administered under red light and the animal was placed in a cage measuring 40.5 3 28 3 12 cm containing only sawdust bedding (i.e., no food and water). The animal remained in this staging cage for 30 minutes in order to allow adequate absorption of drug, after which time the animal was transferred to the experimental room (again in a darkened box) and placed on the zero-maze at a junction between an open and closed arm, facing into the closed area. Each animal was placed on an alternative side of the maze so as to avoid any bias. All testing was conducted under red-light. After 5 minutes, the animal was removed from the maze and the maze was washed using an ethanol/water solution so as to avoid olfactory cues transferring from one experimental session to the next. Once testing was complete, computer-based event recording and ethological analysis software (Hindsight ver. 1.5) was used to register the relevant experimental variables. Videotapes were scored blindly by a highly trained observer (intra-observer reliability . 0.8). Behavioral parameters comprised traditional spatio-temporal, locomotion, maintenance and ethological measures [21][22][23] , which are outlined below. Ethological measures. Stretched Attend Posture was defined as the frequency with which the animal makes a forward elongation of the head and shoulders followed by a retraction to the original position when the animal is on an open arm 23 . Head dips measured the frequency of the animal protruding its head over the ledge of an open arm and down towards the floor 25 . Finally, Risk Assessment was scored as the frequency of an animal exiting a closed arm of the maze with forepaws and head only, and investigating the surroundings. Risk assessment was scored independently of stretch attend posture, which often accompanied risk assessment 21 .

Measures
Locomotion and maintenance measures. Rearing Frequency was defined as the number of vertical movements against the side or end of walls of the closed arms of the maze or raising up its hind paws sans support in the open area of the maze 26 . Rearing Duration measured the time each animal was engaged in rearing behavior.
Analysis. Data for each behavioral element were grouped according to treatment and were initially checked for normality and homogeneity of variance. Levene's tests were statistically significant (i.e. assumptions that distributions were normal or variances were equal had to be rejected); therefore, non-parametric tests were conducted (as opposed to data transformation). Data were analyzed using Kruskal-Wallis nonparametric one-way analysis of variance across treatment groups. Where significant variations in the data were identified, post-hoc Mann-Whitney U comparisons with control condition were performed. In all cases, a standard significance level a was set at p 5 0.05.