Conceived and designed the experiments: MH NK KM BL. Performed the experiments: MH NK. Analyzed the data: MH NK KM. Contributed reagents/materials/analysis tools: MH NK KM BL. Wrote the paper: MH NK KM BL.
The authors have declared that no competing interests exist.
Well balanced novelty seeking and exploration are fundamental behaviours for survival and are found to be dysfunctional in several psychiatric disorders. Recent studies suggest that the endocannabinoid (eCB) system is an important control system for investigatory drive. Pharmacological treatment of rodents with cannabinergic drugs results in altered social and object investigation. Interestingly, contradictory results have been obtained, depending on the treatment, drug concentration and experimental conditions. The cannabinoid type 1 (CB1) receptor, a central component of the eCB system, is predominantly found at the synapses of two opposing neuronal populations, i.e. on inhibitory GABAergic and excitatory glutamatergic neurons. In the present study, using different transgenic mouse lines, we aimed at investigating the impact of CB1 receptor inactivation in glutamatergic or GABAergic neurons on investigatory behaviour. We evaluated animate (interaction partner) and inanimate (object) exploratory behaviour in three different paradigms. We show that exploration was increased when CB1 receptor was deleted from cortical and striatal GABAergic neurons. No effect was observed when CB1 receptor was deleted specifically from dopamine receptor D1-expressing striatal GABAergic medium spiny neurons. In contrast, deletion of CB1 receptor from cortical glutamatergic neurons resulted in a decreased exploration. Thus, our results indicate that exploratory behaviour is accurately balanced in both, the social and non-social context, by the eCB system via CB1 receptor activation on cortical glutamatergic and GABAergic neurons. In addition, the results could explain the contradictory findings of previous pharmacological studies and could further suggest a possibility to readjust an imbalance in exploratory behaviour observed in psychiatric disorders.
Adequate novelty seeking and exploration are fundamental behaviours for survival. Dysfunctional exploratory profiles have been found in several distinct neuronal disorders, such as attention deficit disorder and schizophrenia-like diseases, expressed by modulated social behaviour and novelty seeking
One important factor in exploratory behaviour is how a respective situation is evaluated. Brain regions involved in these evaluation processes, such as amygdala, hippocampus, and prefrontal cortex, show high levels of CB1 receptor mRNA and protein
Anxiety plays a critical role in exploratory and investigatory behaviour, and several pharmacological studies have shown the importance of the eCB system in social behaviour
By using several conditional CB1 receptor knock-out mice, we aimed at investigating whether CB1 receptor on different neuronal cell types might explain the contradictory findings in social interaction and object exploration mentioned above. In order to address this question, we applied different behavioural paradigms to analyze inanimate (object) exploration and animate (interaction partner) exploration. Evaluating the results, we could detect a decreased exploratory drive in mice lacking CB1 receptor in cortical glutamatergic neurons. Mice lacking CB1 receptor in GABAergic neurons, including the striatum, displayed opposite results, namely, an increased exploratory drive. No changes in exploration were observed for mice lacking CB1 receptor specifically in striatal dopamine receptor D1-positive GABAergic medium spiny neurons. Thus, we hypothesize that cortical GABAergic interneurons are important for the increased exploratory drive. Altogether, our results suggest that exploratory behaviour (animate and inanimate) is balanced by the eCB system via CB1 receptor activation on the two opposing neuronal subpopulations.
This study was performed on adult (5–7 months old) male mutant mice and their respective wild-type littermates. Animals were housed in a temperature- and humidity-controlled room (22°C±1; 50%±1) with a 12 h light-dark cycle (lights on at 1 am) and had access to food and water
Animals were group-housed (3–5 animals per cage type 2 (26.5×20.5×14.0 cm), EBECO Germany) until one week before behavioural testing. Animals were then separated and single- housed to avoid behavioural differences between dominant and subordinate animals. The same animals were used in each paradigm. Between each experimental paradigm, animals were allowed to rest for one week. All experiments were performed one hour after turning off the lights (2 pm), in the active phase of the animals, with only a minimal red light source in the room (0 lux).
The novel object recognition task combines a general exploration test with a visual recognition memory paradigm. Therefore, it is used to evaluate object exploration and object recognition. The test was performed in a white plastic open field chamber (H40 cm×W40 cm×L40 cm). The protocol used was modified from Ennaceur and Delacour , Tang et al., and Tordera et al.
For habituation, the animals were placed into the empty open field and allowed to explore the box for 10 min once a day for two days. The first habituation session was analyzed according to a standard open field paradigm, hence, total distance moved and time spent in the center (defined as 20 cm×20 cm) was evaluated using SMART software (PanLab, Spain). On day 3, two identical objects (O1 left, and O1 right; two metal cubes with H4 cm×W3 cm×L5 cm) were placed symmetrically 6–7 cm from the walls and separated 16–18 cm from each other. The mouse was placed into the box at an equal distance from both objects and video-recorded for 10 min. After this first exposure to the object, the mouse was returned to its home cage. 2 h and 24 h later, the mouse was placed again into the open field and exposed to the familiar object (O1) and to a novel object (O2 for the 2 h time point, and O3 for the 24 h time point, respectively) each time for 10 min (retention tests). The novel object O2 was a plastic billiard ball (5.72 cm in diameter) fixed on a metal plate (0.2 cm) and O3 was a round glass flask (H6 cm×W3 cm), filled with sand and closed with a black rubber plug. The familiar object was always positioned on the left side, while the new object was on the right side. Box and objects were cleaned with 70% ethanol after each trial to avoid olfactory cues. Experiment was video-recorded and the total time that the animal spent exploring each of the two objects in training and retention phase was evaluated by an experimenter blind to the genotype. Object exploration was defined as the orientation of the nose directly to the object at a distance <2 cm and/or touching the object with the nose and whiskers. Time spent climbing and sitting on the object were not regarded as exploration, and was therefore excluded from measurement
A modified sociability test was performed, based on a published protocol
The test animal was placed into the middle compartment for 5 min with entries to the side compartments blocked.
After the habituation phase, blockades of the entries were removed, allowing free access to the side compartments for 10 min. By doing this, the animal tested was exposed to a novel C57BL/6N interaction partner and a novel object (round cage described below), positioned in the two side compartments. The position of the interaction partner (left vs. right compartment) was alternated between trials to avoid any bias. The interaction partner itself was enclosed in a round cage (10 cm in diameter; 30 cm high [upper 20 cm Plexiglass, lower 10 cm covered by metal bars 1 cm apart to allow interaction but prevents fighting]). To minimize stress levels of the animals used as interaction partners, they were habituated to the cages four times for 10 min distributed over two days prior to the actual test days. To counterbalance individual differences of these interaction partners they were equally used for wild-type and mutant test mice. The novel object control (empty cage, no animal) was always positioned in the opposite compartment to the cage with the interaction partner. The discrimination index (DI) was calculated as the difference between the time spent exploring the novel object (nO) and the novel animal (nA), divided by the total time exploring both [(nO−nA)/(nO+nA)]. A positive DI is considered to reflect increased preference for the social interaction partner.
2 h after the sociability phase, an additional, unknown interaction partner (novel) was introduced. The interaction partner from the sociability phase (familiar) was again placed into the same cage and same compartment as before. The novel animal was placed into the former empty cage and positioned at the respective side compartment. Openings were unblocked. The test animal was placed into the middle compartment, and the test animal was allowed to freely explore for 10 min. The DI was calculated as the difference between the time spent exploring the new (N) and the familiar (F) animal, divided by the total time exploring both [(N−F)/(N+F)]. A positive DI is considered to reflect increased memory retention for the familiar animal.
The resident-intruder test was performed by placing a novel, group-housed intruder into the home cage of the test animal for 10 min. This paradigm allows evaluating social exploration and aggressive behaviour
Data are presented as mean ± standard error of the mean (SEM) of individual data points. Results were considered to be significant at p<0.05. All behavioural endpoints of the novel object recognition task were initially analyzed using two-way ANOVA, using genotype and object as variables and Bonferroni post-tests to correct for multiple comparisons. In some cases, to analyze the locomotion effects in the open field, the sociability in the sociability test and the aggression in the resident-intruder paradigm for each genotype, data were analysed using an unpaired Student's t-test or Kruskal-Wallis statistic. Additionally, in order to evaluate whether the DI of the genotypes deviated significantly from zero, we used the unpaired t-test with Welch's correction. Graphs and statistics were generated by GraphPad Prism 4.03 (GraphPad Software;
The evaluation of the locomotor activity in the open field revealed that only the GABA-CB1−/− mice showed an alteration (T(18) = 3.213, p = 0.0048;
Glu-CB1 | GABA-CB1 | D1-CB1 | ||||
Paradigm | +/+ | −/− | +/+ | −/− | +/+ | −/− |
|
||||||
|
2824±209 | 2324±182 | 2621±306 | 3801±202 |
4456±91 | 4368±131 |
|
||||||
Habituation | 2012±113 | 1569±162 |
1597±59 | 1730±64 | 1669±81 | 1850±118 |
Sociability | 5171±205 | 4891±312 | 4945±127 | 5083±175 | 5055±162 | 5079±157 |
Social Novelty | 3973±211 | 3125±227 |
4023±141 | 4623±175 |
3863±292 | 3942±160 |
|
||||||
|
63.3 ± 12 | 80.9±20 | 156.0±48 | 82.0±21 | 65.4±11 | 82.4±15 |
|
||||||
|
||||||
Training | 0.01±0.01 | −0.08±0.08 | −0.03±0.03 | −0.03±0.02 | −0.08±0.05 | 0.03±0.02 |
Retention 2 h | 0.16±0.03 |
0.06±0.05 | 0.00±0.06 | −0.03±0.03 | 0.15±0.06 |
0.08±0.03 |
Retention 24 h | 0.25±0.06 |
0.06±0.13 | 0.18±0.06 |
0.15±0.05 |
0.27±0.08 |
0.18±0.05 |
|
||||||
Sociability | 0.29±0.03 |
0.12±0.07 |
0.27±0.03 |
0.35±0.04 |
0.20±0.04 |
0.30±0.04 |
Social Novelty | 0.05±0.03 | 0.08±0.09 | −0.01±0.05 | 0.09±0.03 |
0.03±0.06 | 0.03±0.06 |
Evaluation of locomotion (distance moved), anxiety (time in center) and memory (discrimination index) for all mutant lines; +/+ (wild-type), −/− (mutant); t-test analysis:
*p<0.05;
**p<0.01 (significance between genotype);
p<0.05 (significant from 0; positive recognition of novel object).
The analysis of the novel object recognition task (referred to as NORT in
(A–C) Total time of exploration of two identical objects (O1, both on left and right side) during the training session for three conditional CB1 receptor mutant lines (Glu-CB1 [n = 23+13], GABA-CB1 [n = 18+23], D1-CB1 [12+12]) and their wild-type control littermates. (D–F) Total time of exploration of familiar object (O1) and novel object (O2 or O3) during the retention session after 2 h or 24 h (G–I). Glu-CB1−/− mice displayed a reduced exploration, while GABA-CB1−/− mice showed an increased exploration both in the training and retention session when compared to their wild-type littermate controls. No significant genotype differences were observed in the D1-CB1 mutant line. 2-way ANOVA (genotype differences) *p<0.05, ***p<0.001; t-test (discrimination index DI) #p<0.05.
Evaluation of the discrimination index (DI) revealed that all groups, independent of the line, showed no differences within the training session regarding the exploration of the left and the right object O1, respectively. (Glu-CB1+/+ [T(20) = 0.8230, p = 0.4202]; Glu-CB1−/− [T(11) = 0.9582, p = 0.3585]; GABA-CB1+/+ [T(15) = 1.118, p = 0.2812]; GABA-CB1−/− [T(22) = 1.959, p = 0.0630]; D1-CB1+/+ [T(11) = 1.447, p = 0.1758]; D1-CB1−/− [T(11) = 1.679, p = 0.1213];
In the 2 h retention phase, several groups lacked a significant discrimination between the familiar and the novel object. Only Glu-CB1+/+, D1-CB1+/+ and D1-CB1−/− animals displayed a significant preference towards the novel stimulus (Glu-CB1+/+ [T(21) = 4.806, p<0.0001]; Glu-CB1−/− [T(12) = 1.220, p = 0.2458]; GABA-CB1+/+ [T(15) = 0.07097, p = 0.9444]; GABA-CB1−/− [T(22) = 1.366, p = 0.1858]; D1-CB1+/+ [T(10) = 2.502, p = 0.0313]; D1-CB1−/− [T(10) = 2.238, p = 0.0492];
In the 24 h retention phase, independently of the genotype, all groups showed a significant preference towards the novel object, with the only exception of the Glu-CB1−/− animals (Glu-CB1+/+ [T(21) = 4.472, p = 0.0002]; Glu-CB1−/− [T(12) = 0.4328, p = 0.6729]; GABA-CB1+/+ [T(15) = 2.818, p = 0.0129]; GABA-CB1−/− [T(22) = 3.072, p = 0.0056]; D1-CB1+/+ [T(11) = 3.601, p = 0.0042]; D1-CB1−/− [T(11) = 3.540, p = 0.0046];
The evaluation of object specific exploration (O1 left or O1-3 right) over the three sessions (training, 2 h retention and 24 h retention), revealed a significant difference for the Glu-CB1−/− as compared to their littermate controls. Thus, the Glu-CB1−/− mutants showed a steadily decreasing investigatory behaviour for both, the left object (increasing familiarity) and the right object (always novel) (Glu-CB1−/− interaction [object/time]: F(2,48) = 0.1537, p = 0.8580; Bonferroni post-test: training p>0.05, 2 h p>0.05, 24 h p>0.05;
During the sociability phase, the Glu-CB1−/− animals showed a significant increase in time spent in the middle compartment (T(33) = 2.247, p = 0.0314;
(A–C) Comparison of animate (mouse) and inanimate (object, “empty”) exploration for the three mutants lines (Glu-CB1 [n = 22+13], GABA-CB1 [n = 18+23], D1-CB1 [16+16]) and their wild-type littermate controls during the sociability phase. (D–F) Exploration of the familiar and the novel interaction partner for during the social novelty phase. Glu-CB1−/− mice displayed no significant change in the exploration session, where there was a choice between the object and the interaction partner. In the social novelty phase, however, the interaction with a novel interaction partner was decreased when compared with their wild-type littermate controls. GABA-CB1−/− mice showed an increased social interaction in both sessions. In the D1-CB1 mutant line, no genotype differences were observed neither in the sociability nor in the social novelty phase. n = 11–20 animals; t-test *p<0.05, **p<0.01.
The evaluation of the DI showed only minimal differences between the genotypes. In the sociability phase, the Glu-CB1−/− animals showed an impaired preference towards the interaction partner as compared to their controls (T(33) = 2.537, p<0.0161;
For all lines and genotypes, except for the Glu-CB1−/− mice, we observed a strong preference towards the social interaction partner over the object in the sociability phase (Glu-CB1+/+ [T(21) = 10.47, p<0.0001]; Glu-CB1−/− [T(12) = 1.559, p = 0.1450]; GABA-CB1+/+ [T(27) = 8.309, p<0.0001]; GABA-CB1−/− [T(30) = 8.187, p<0.0001]; D1-CB1+/+ [T(16) = 5.017, p = 0.0002]; D1-CB1−/− [T(13) = 7.458, p<0.0001];
The evaluation of the locomotor activity revealed no significant changes in the habituation phase of the sociability test, except for the Glu-CB1−/− mice, which showed a decrease in locomotion (Glu-CB1 line [T(33) = 2.312, p = 0.0271]; GABA-CB1 line [T(60) = 1.506, p = 0.1374]; D1-CB1 line [T(29) = 1.571, p = 0.1270]). In the sociability phase, no alteration in the distance moved was observed in any of the lines (Glu-CB1 line [T(29) = 0.7833, p = 0.4398]; GABA-CB1 line [T(62) = 0.6159, p = 0.5402]; D1-CB1 line [T(30) = 0.1082, p = 0.9145]). However, a significant decrease and increase in the distance moved was detected in the social novelty phase for the Glu-CB1−/− mice and the GABA-CB1−/− mice, respectively (Glu-CB1 line [T(33) = 2.575, p = 0.0146]; GABA-CB1 line [T(38) = 2.591, p = 0.0135]). The D1-CB1−/− mice again showed no change in the distance moved as compared to their respective wild-type littermates (T(30) = 0.2386, p = 0.8130).
Glu-CB1−/− mice displayed a significant decrease interacting with the intruder animals for the 10 min interaction phase as compared with wild-types (T(35) = 2.297, p = 0.0277). Splitting the 10 min period into two 5 min bins revealed that the difference in interaction was mainly visible for the first 5 min bin (T(35) = 3.106, p = 0.0038) (
(A–C) Social interaction with an unknown, younger intruder for all three mutant lines (Glu-CB1 [n = 23+13], GABA-CB1 [n = 18+23], D1-CB1 [n = 16+16]). (D–E) Number of fights induced by the resident is shown for all three mutant lines. Glu-CB1−/− mice showed a significantly reduced exploration during the first 5 min observation period and an increased aggression towards the intruder when compared to wild-type littermate controls. GABA-CB1−/− mice displayed an increased interaction with the intruder, but no difference in aggressive behaviour. D1-CB1−/− mice showed no behavioural changes as compared to their wild-type littermate controls. t-test *p<0.05, **p<0.01.
Additional analysis revealed that Glu-CB1+/+ animals displayed a significant increase in aggression as compared to the other control groups, GABA-CB1+/+ and D1-CB1+/+. Thus, differences were detected in number of fights (Kruskal-Wallis statistic = 7.478, p = 0.0238; Dunn's Multiple Comparison Post-Test: Glu-CB1+/+ vs GABA-CB1+/+ p<0.05, Glu-CB1+/+ vs D1-CB1+/+ p>0.05, GABA-CB1+/+ vs D1-CB1+/+ p>0.05), as well as % of time fighting (Kruskal-Wallis statistic = 7.584, p = 0.0226; Dunn's Multiple Comparison Post-Test: Glu-CB1+/+ vs GABA-CB1+/+ p<0.05, Glu-CB1+/+ vs D1-CB1+/+ p>0.05, GABA-CB1+/+ vs D1-CB1+/+ p>0.05).
Using different conditional CB1 receptor mutant mice, we were able to show that the deletion of the CB1 receptor from forebrain GABAergic or cortical glutamatergic neurons resulted in an opposite behavioural outcome regarding animate and inanimate exploration. On the other hand, deletion of the CB1 receptor from dopamine receptor D1-expressing GABAergic striatal medium spiny neurons did not result in any significant changes. These findings suggest a regulatory function of the eCB system in cortical GABAergic and glutamatergic circuits to prevent neuronal and behavioural imbalance.
Mice lacking the CB1 receptor on glutamatergic neurons displayed a decreased exploratory behaviour, both in animate interaction (the interaction with a partner) and inanimate interaction (the interaction with an object). A similar decrease in object and social exploration was found in earlier studies, which were related with increased fear
The anxiogenic-like behaviour associated to these mutants can also explain the significantly higher aggression level found in the resident-intruder paradigm (
Taken together, these results suggest an anxiolytic-like function of the CB1 receptor on glutamatergic neurons and an anxiogenic-like function of the CB1 receptor on GABAergic interneurons. However, a generalized conclusion on the involvement CB1 receptor on cortical glutamatergic neurons in anxiety is not yet possible to be drawn, as under our experimental conditions, the open field test was not congruent with this notion. Neither Glu-CB1−/− nor GABA-CB1−/− mutants spend a different period of time in the more aversive center zone as compared with their respective wild-type littermates (
An alternative explanation for the observed differences can be alterations in spontaneous locomotor activity. In fact, we observed for both the Glu-CB1−/− and GABA-CB1−/− changes in the distance moved, namely a decrease and increase, respectively. It seems unlikely that the difference in locomotion was the driving force underlying the exploration phenotypes, as the mutants, in contrast to the variation in animate and social investigation, did not always display the locomotor alterations (
A further explanation for the behavioural differences might be memory alterations in the respective mutant. However, this might only account for the Glu-CB1−/− mutants, as all other animals, independently of line and genotype, displayed a similar memory and recognition performance. Especially after 24 hours, mice recognized and distinguished strongly between familiar and novel objects (
As mentioned above, all groups, independently of the line and the genotype, showed a stronger preference for the social interaction partner as compared to the object in the sociability test (
Taken together the strong differences observed in the GABA-CB1−/− and Glu-CB1−/− animals in respect to their wild-type littermates might be explained by anxiolytic and anxiogenic responses to novelty, respectively. Nevertheless, the eCB system has also been shown to be involved in learning and memory function, which should be kept in mind here
Our results, namely the increase of exploration following the deletion of GABAergic CB1 receptor and the decrease of exploratory behaviour following the deletion of glutamatergic CB1 receptors, may explain the contradictory findings using Δ9-THC, URB597 and VDM11, as described in above. We suggest that increased or decreased exploratory drive, respectively, as response to cannabinoid treatment depends on the predominant modulation of either GABAergic or glutamatergic CB1 receptor, e.g. the activation of GABAergic CB1 receptor decreases exploration, while the activation of glutamatergic CB1 receptor leads to an increased investigatory drive. Thus, the decreased exploration induced by chronic and systemic activation of the eCB system with Δ9-THC might be due to the exogenous activation of the CB1 receptor in GABAergic interneurons
Our results might also be interesting in respect to some disorders, which are associated with inappropriate exploratory drive. Thus, a direct and indirect relation between these disorders and a dysregulation of GABAergic and/or glutamatergic transmission can be proposed. In animal models for autism, modulation of GABAergic transmission seems to be important
In conclusion, our results indicate a major, but opposite role of the eCB system in cortical GABAergic and glutamatergic neurons in the regulation of exploration (
Locomotion | Object Exploration | Social Exploration | Aggression | |
Wild-type | Normal | Normal | Normal | Normal |
Glu-CB1−/− | Decreased | Decreased | Decreased | Increased |
GABA-CB1−/− | Increased | Increased | Increased | Normal |
D1-CB1−/− | Normal | Normal | Normal | Normal |
“Normal” refers to similar to the wild-type behaviour on spontaneous locomotor activity (locomotion), investigation of object (object exploration) or of interaction partner (social exploration) and fights initiated (aggression).
We would like to acknowledge Andrea Conrad, Danka Dorman, Anisa Kosan, and Anne Rohrbacher for genotyping mutant mice, Floortje Remmers for help in the statistics, and Alejandro Aparisi Rey, Norbert Sachser, Gleb Shumyatsky and Raj Kamal Srivastava for critical reading of the manuscript.