- Email: [email protected]
Sociability trait and serotonin metabolism in the rat social interaction test
Neuroscience Letters 367 (2004) 309–312
Sociability trait and serotonin metabolism in the rat social interaction test Margus Tõnissaar, Mari-Anne Philips1 , Marika Eller, Jaanus Harro∗ Department of Psychology, Centre of Behavioural and Health Sciences, University of Tartu, EE-50410 Tartu, Estonia Received 19 April 2004; received in revised form 5 June 2004; accepted 8 June 2004
Abstract Social behaviour is the basis of one of the most generally accepted independent dimensions of personality. The purpose of the present study was to find out whether the social activity of individual rats, expressed in the social interaction test of anxiety, is consistent, and associated with monoamine levels. Four social interaction tests with 10 days intervals were carried out in 20 rats, and the animals were decapitated 4 days after the last test. There was no consistent correlation between performances in single tests, but the social interaction time in each test correlated strongly with the mean values of social activity in all or the other three tests. Social interaction time of rats correlated moderately but significantly with their partner’s social activity in the test. The average social interaction time correlated strongly with 5-HIAA levels in the frontal cortex (r = −0.67, P < 0.01). Neither exposure of rats singly to the social interaction test box nor the test procedure had any effect on monoamine levels. When animals were decapitated immediately after a single social interaction test, there was a negative correlation between the social interaction time and 5-HIAA and 5-HT levels in the septum, but not in the frontal cortex or hippocampus. Thus, social behaviour is a stable trait, expression of which depends in part upon the partner’s social behaviour. This trait is negatively associated with 5-HT metabolism in the frontal cortex. Social activity of rats in a particular test situation may rather be related to 5-HT metabolism in the septum. © 2004 Elsevier Ireland Ltd. All rights reserved. Keywords: Sociability; Social interaction test; 5-Hydroxyindole acetic acid (5-HIAA); Frontal cortex; Septum; Rats; Traits
Social behaviour is the basis of one of the most generally accepted independent dimensions of personality [3], and has many important roles in the survival of the individual and species. Many clinical disorders include difficulties in creating or maintaining social contacts, such as social anxiety disorder, several personality disorders, autism, etc. The phenotypical expression of social behaviour is regulated by many different neurochemical systems, but a coherent picture is yet to emerge [18]. Recent studies on humans highlight the role of 5-HT in the expression of personality dimensions, mostly in these which are associated with affective and motivational processes [16,21]. To study the neurobiological mechanisms of social behaviour and related disorders in sufficient detail, animal models would be helpful. ∗
Corresponding author. Tel.: +372 7 375911; fax: +372 7 375900. E-mail address: [email protected] (J. Harro). 1 Present address: Institute of Cellular and Molecular Biology, University of Tartu, Riia 23, Tartu, Estonia.
Studies of social behaviour in rodents have mostly been carried out within the constraints of rather specific paradigms, such as aggressive or maternal behaviour. However, there is a frequently used simple animal model that was developed to measure anxiogenic and anxiolytic drug effects [7] but which is based on social behaviour: the social interaction test, in which the time spent in active social interaction between two unfamiliar rats in a neutral arena is measured. Behaviour of rats in the social interaction test does not correlate well with their performance in other animal models of anxiety [20]. This suggests that the model has other important underlying mechanisms than just general anxiety, and could be used for studying neurobiology of social behaviour provided that social behaviour of an animal would be a consistently expressed trait in this test. The purpose of the present study was to address this question and to investigate whether the trait of social behaviour, if measurable, would correlate with monoamine levels in the frontal cortex, septum and hippocampus. Pharmacological interventions in these brain regions have been found
0304-3940/$ – see front matter © 2004 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.neulet.2004.06.023
310
M. Tõnissaar et al. / Neuroscience Letters 367 (2004) 309–312
to influence behaviour in the social interaction test [8,15], and even though serotonergic mechanisms are most consistently implicated in animal and human studies on social behaviour [4,16], catecholaminegic mechanisms have also been suggested to contribute [3]. We also studied whether acute performance in the social interaction test itself has any acute effect on monoamines, or would correlate with ex vivo monoamines. Male Wistar rats from the National Laboratory Animal Centre, Kuopio, Finland were group-housed after arrival in plastic cages with food (Lactamin R35, Sweden) and water ad libitum. Room temperature was maintained at 21 ± 2 ◦ C and 12-h light:12-h dark (lights on at 07:00 h) was applied. Animals weighing 278–382 g at the beginning of the experiment were used. The test developed by File and Hyde [7] was used as described earlier [12]. Rats were housed individually at least 10 days before testing. Two unfamiliar, weight-matched rats were placed for 10 min into a brightly-lit chamber (30 cm × 30 cm × 60 cm) with floor covered with wood shavings. The total time spent in active social behaviour (allogrooming, sniffing the partner, crawling under and over, following) was recorded by two observers. Inter-rater reliability was always found to be good (r > 0.90). Experiment I measured whether the social behaviour of a given rat is stable in the social interaction test. Social interaction test was carried out in 20 animals in four separate sessions with intervals of 10 days. Rats were paired on the basis of their body weight before the test and every test was carried out with an unfamiliar partner. Animals were decapitated 4 days after the fourth test and brains immediately dissected on ice, with the rat brain atlas of Paxinos and Watson as a guide. Samples of frontal cortex also contain the area 1 of the cingulate cortex. Septum represents the septal complex between the lateral ventricles in the striatal slice cut approximately between the bregma level and 1.4 mm ahead of it. Hippocampi were separated from overlaying cerebral cortex as a whole after removal of the basal forebrain and other brain parts. The purpose of Experiment II was to clarify whether exposure to the social interaction test itself has any effect on monoamine levels. After 10 days of single housing, thirty animals were divided into three groups (n = 10). One group of animals was directly transported from home cages to a
separate room and immediately decapitated; animals of the second group were singly placed into the social interaction test box for 10 min and then decapitated. The third group of animals was decapitated immediately after the social interaction test had been carried out as usual. Monoamines were assayed essentially as described in [12] with minor modifications in the method. All calculations were performed using StatView 4.5 software (Abacus Concepts. Cary, NC, USA). Group comparisons were made by repeated measures or one-way ANOVA as appropriate. Correlations shown are Pearson correlation coefficients. Statistical significance was set at P < 0.05. In Experiment I, the mean duration of social interaction over all four tests was 94 s. According to repeated measures ANOVA, mean social interaction time did not differ significantly between the tests. In Table 1, pairwise correlations between social interaction times of individual animals in different tests are presented. Correlations were stronger between tests, which followed each other, but the overall picture suggests that there is no consistent correlation between performances in single tests. Nevertheless, social interaction time in each single test correlated significantly with the mean values of social activity as calculated on the basis of performance in all tests or in the other three tests (Table 1). It should be noted that inclusion of the data from a particular test into the average social activity measure artificially increases the correlation, but these correlations have also been shown because the small number of independent observations is also likely to introduce a bias. Social interaction time of rats correlated moderately but significantly with the partner’s social activity across all tests (r = 0.35, P < 0.001) which is thus a source of variation. We carried out only four tests because it was not possible to find more weight-matched unfamiliar partners in the group of animals used in this experiment. In the fourth test the mean difference in body weight was larger than considered optimal (7 g in test I; 11 g in test II; 14 g in test III and 20 g in test IV), and there was a tendency of weaker correlation between the fourth test and other tests. Differences between body weights of partners in the social interaction test did not correlate with social interaction time, however. We carried out linear regression analyses of rat mean social interaction time across four tests as dependent of tissue levels of NA, DA, DOPAC, HVA, 5-HT, and 5-HIAA in
Table 1 Correlations between the social interaction (SI) tests or between each test and the mean of the other three tests or with the mean of all tests SI test I SI SI SI SI
test test test test ∗
I II III IV
P < 0.05. P < 0.01. ∗∗∗ P < 0.001. ∗∗∗∗ P < 0.0001. ∗∗
0.64∗ 0.28 0.34
SI test II
0.41 0.14
SI test III
0.60∗
Mean of the other three tests
Mean of all tests
0.58∗∗
0.73∗∗∗ 0.77∗∗∗∗ 0.79∗∗∗∗ 0.68∗∗∗
0.55∗∗ 0.65∗∗∗ 0.50∗
M. Tõnissaar et al. / Neuroscience Letters 367 (2004) 309–312
311
Fig. 1. Experiment I: Regression analysis of 5-HIAA levels in the frontal cortex and social interaction time in the social interaction test (mean of four tests) [r 2 = 0.44; F(1; 16) = 12.82, P < 0.01].
the frontal cortex, hippocampus, and septum. A significant correlation (r = −0.67, P < 0.01) was found between the mean social interaction time and 5-HIAA levels in the frontal cortex (Fig. 1). (Data of two rats were excluded because of anomalously high 5-HIAA values, several fold the group mean, probably due to inadequate sample handling.) Mean social interaction time did not correlate with any other monoamine measure, including the 5-HIAA/5-HT ratio. In Experiment II, ANOVA did not reveal any difference between the three groups in monoamine levels in any of the three brain regions (data not shown). In the animals of the group that underwent the social interaction test, the mean social interaction time was 73 s when one rat with exceptionally high social activity (227 s) was omitted from the analysis. There was a negative correlation between the social interaction time and 5-HIAA (r = −0.90, P < 0.01) and 5-HT levels (r = −0.79, P < 0.05) in the septum (Fig. 2A and B, respectively), but not in the frontal cortex or hippocampus. Evolutionary continuity between humans and other animals suggests that some dimensions of personality may be common across a wide range of species [6,9] and animals may also have persistent structures of behaviours, which may be stable across multiple years [6]. In the present study, activity in each social interaction test did not correlate consistently with other tests pairwise, but the individual average social behaviour correlated well with performance in each occasion, suggesting that in the rat, there is a sociability trait which is measurable by repeated testing in the social interaction test. In comparison to the physical environment, social environment is less predictable and very reactive to the behaviour of the individual itself. Social behaviour of rats was found to depend on the social activity of partners, and thus a conclusion can be drawn that social activity of an individual rat is a trait, manifestation of which depends on the partner’s social behaviour. We have reproduced the finding that mean social interaction time correlates consistently with performance in all tests in three subsequent independent experiments (unpublished).
Fig. 2. (A) Experiment II: Regression analysis of 5-HIAA levels in the septum and social interaction time in the social interaction test [r 2 = 0.81; F(1; 7) = 30.26, P < 0.001]. (B) Experiment II: Regression analysis of 5-HT levels in the septum and social interaction time in the social interaction test [r 2 = 0.62; F(1; 7) = 11.56, P < 0.05].
The sociability trait was found to be strongly negatively associated with the levels of 5-HIAA in the frontal cortex which were measured 4 days after the last social interaction test. Even though ex vivo 5-HIAA levels cannot be interpreted unambigously, lower levels of a metabolite are suggestive of lower basal release. Thus this result was not expected because several well-known CSF studies have found a high positive correlation between serotonin metabolism and sociability [13,17,19]. Nevertheless, there is also evidence for negative association between 5-HT metabolism and social competence [22]. Even in the simple social interaction test, both increased [4,10] and reduced [1,8,14,15] social behaviour has been suggested to be mediated via an increase in serotonergic function. In our experiments, the negative association between septal 5-HT measures and social activity in the experiment in which the animals were decapitated immediately after social interaction test further supports the notions that at least in certain conditions 5-HT metabolism is inhibitory to social activity, but that different brain regions contribute to sociability trait and ongoing social activity. Differences in methodological detail in sample collection and processing could account for some of the variance across laboratories, but these could be more easy explana-
312
M. Tõnissaar et al. / Neuroscience Letters 367 (2004) 309–312
tions for data not confirming positive results, whereas in the case of association of 5-HT and social behaviour we are rather confronting data which show significant correlations in opposite directions. Serotonergic measures are probably related to sociability-associated mechanisms in a complex manner, and this can cause differences between laboratories in several ways. For example, low 5-HIAA levels in the CSF have been associated with high impulsivity, which may mediate social incompetence [13,17]. By using in vivo microdialysis, Dalley et al. [2] recently demonstrated that higher extracellular levels of 5-HT in the prefrontal cortex are associated with higher impulsivity. It is possible that 5-HIAA levels in the CSF reflect better serotonergic activity in the brainstem, which is a negative mirror image of forebrain mechanisms due to somatodendritic inhibition of 5-HT projection neurons. However, it is also possible that, when the environmental signals become more complex, the 5-HT behaviour link is not linear. As an example, we have recently demonstrated in a longitudinal study that platelet MAO activity, a peripheral marker of central serotonergic activity [5], is predictive of regular smoking in adolescents, but the relationship is very clearly U-shaped [11]. Apparently the different contribution of a variety of 5-HT receptor subtypes in different brain regions should be considered. In summary, these experiments demonstrate that a sociability trait underlies the performance in the rat social interaction test, and that this trait is negatively associated with frontal 5-HT metabolism.
Acknowledgements This work was supported by the Estonian Science Foundation Grant No. 4531, the Estonian Ministry of Education and Science project No. 2643, and the EU Framework 6 Integrated Project NEWMOOD (LSHM-CT-2004-503474).
References [1] G. Bagdy, M. Graf, Z.E. Anheuer, E.A. Modos, S. Kantor, Anxiety-like effects induced by acute fluoxetine, sertraline or m-CPP treatment are reversed by pretreatment with the 5-HT2C receptor antagonist SB-242084 but not the 5-HT1A receptor antagonist WAY-100635, Int. J. Neuropsychopharmacol. 4 (2001) 399–408. [2] W.J. Dalley, D.E. Theobald, D.M. Eagle, F. Passetti, T.W. Robbins, Deficits in impulse control associated with tonically-elevated serotonergic function in rat prefrontal cortex, Neuropsychopharmacology 26 (2002) 716–728. [3] R.A. Depue, P.F. Collins, Neurobiology of the structure of personality: dopamine, facilitation of incentive motivation, and extraversion, Behav. Brain Sci. 22 (1999) 491–569. [4] M.S. Duxon, K.R. Starr, N. Upton, Latency to paroxetine-induced anxiolysis in the rat is reduced by co-administration of the 5-HT1A receptor antagonist WAY 100635, Br. J. Pharmacol. 130 (2000) 1713–1719.
[5] C. Fahlke, H. Garpenstrand, L. Oreland, S.J. Suomi, J.D. Higley, Platelet monoamine oxidase activity in a nohuman primate model of type 2 excessive alcohol consumption, Am. J. Psychiatry 159 (2002) 2107–2109. [6] A.J. Figueredo, R.L. Cox, R.J. Rhine, A generalizability analysis of subjective personality assessments in the stumptail macaque and the zebra finch, Multivariate Behav. Res. 30 (1995) 167–197. [7] S.E. File, J.R. Hyde, Can social interaction be used to measure anxiety? Br. J. Pharmacol. 62 (1978) 19–24. [8] S.E. File, H. Zangrossi, N. Andrews, Social interaction and elevated plus-maze tests: changes in release and uptake of 5-HT and GABA, Neuropharmacology 32 (1993) 217–221. [9] S.D. Gosling, O.P. John, Personality dimension in non-human animals: a cross-species review, Curr. Directions Psychol. Sci. 8 (1999) 69–75. [10] M. Hamon, E. Doucet, K. Lefevre, M.-C. Miquel, L. Lanfumey, R. Insausti, D. Frechilla, J. Del Rio, D. Verge, Antibodies and antisense oligonucleotide for probing the distribution and putative functions of central 5-HT6 receptors, Neuropsychopharmacology 21 (1999) 68S– 76S. [11] J. Harro, K. Fischer, S. Vansteelandt, M. Harro, Both low and high platelet monoamine oxidase increase the probability of becoming a smoker, Eur. Neuropsychopharmacol. 14 (2004) 65–69. [12] J. Harro, M. Tõnissaar, M. Eller, A. Kask, L. Oreland, Chronic variable stress and partial 5-HT denervation by parachloroamphetamine treatment in the rat: effects on behavior and monoamine neurochemistry, Brain Res. 899 (2001) 227–239. [13] J.D. Higley, P.T. Mehlman, S.B. Higley, B. Fernald, J. Vickers, S.G. Lindell, D.M. Taub, S.J. Suomi, M. Linnoila, Excessive mortality in young free-ranging male nonhuman primates with low cerebrospinal fluid 5-hydroxyindoleacetic acid concentrations, Arch. Gen. Psychiatry 53 (1996) 537–543. [14] A.J. Kennedy, E.L. Gibson, M.T. O’Connell, G. Curzon, Effects of housing, restraint and chronic treatments with mCPP and sertraline on behavioural responses to mCPP, Psychopharmacology 113 (1993) 262–268. [15] P.J. Kenny, S. Cheeta, S.E. File, Anxiogenic effects of nicotine in the dorsal hippocampus are mediated by 5-HT1A and not muscarinic M1 receptors, Neuropharmacology 39 (2000) 300–307. [16] B. Knutson, O.M. Wolkowitz, S.W. Cole, T. Chan, E.A. Moore, R.C. Johnson, J. Terpstra, R.A. Turner, V.I. Reus, Selective alteration of personality and social behaviour by serotonergic intervention, Am. J. Psychiatry 155 (1998) 373–379. [17] P.T. Mehlman, J.D. Higley, I. Faucher, A.A. Lilly, D.M. Taub, J. Vickers, S.J. Suomi, M. Linnoila, Correlation of CSF 5-HIAA concentration with sociality and the timing of emigration in free-ranging primates, Am. J. Psychiatry 152 (1995) 907–913. [18] J. Panksepp, Affective Neuroscience: The Foundations of Human and Animal Emotions, Oxford University Press, 1998, 480 pp. [19] G.P. Placidi, M.A. Oquendo, K.M. Malone, Y.Y. Huang, S.P. Ellis, J.J. Mann, Aggressivity, suicide attempts, and depression: relationship to cerebrospinal fluid monoamine metabolite levels, Biol. Psychiatry 50 (2001) 783–791. [20] A. Ramos, O. Berton, P. Mormede, F. Chaouloff, A multiple-test study of anxiety-related behaviours in six inbred rat strains, Behav. Brain Res. 85 (1997) 57–69. [21] W.S. Tse, A.J. Bond, Serotonergic intervention affects both social dominance and affiliative behaviour, Psychopharmacology 161 (2002) 324–330. [22] U. Yodyingyuad, C. de la Riva, D.H. Abbott, J. Herbert, E.B. Keverne, Relationship between dominance hierarchy, cerebrospinal fluid levels of amine transmitter metabolites (5-hydroxyindole acetic acid and homovanillic acid) and plasma cortisol in monkeys, Neuroscience 16 (1985) 851–858.
Sociability trait and serotonin metabolism in the rat social interaction test Margus Tõnissaar, Mari-Anne Philips1 , Marika Eller, Jaanus Harro∗ Department of Psychology, Centre of Behavioural and Health Sciences, University of Tartu, EE-50410 Tartu, Estonia Received 19 April 2004; received in revised form 5 June 2004; accepted 8 June 2004
Abstract Social behaviour is the basis of one of the most generally accepted independent dimensions of personality. The purpose of the present study was to find out whether the social activity of individual rats, expressed in the social interaction test of anxiety, is consistent, and associated with monoamine levels. Four social interaction tests with 10 days intervals were carried out in 20 rats, and the animals were decapitated 4 days after the last test. There was no consistent correlation between performances in single tests, but the social interaction time in each test correlated strongly with the mean values of social activity in all or the other three tests. Social interaction time of rats correlated moderately but significantly with their partner’s social activity in the test. The average social interaction time correlated strongly with 5-HIAA levels in the frontal cortex (r = −0.67, P < 0.01). Neither exposure of rats singly to the social interaction test box nor the test procedure had any effect on monoamine levels. When animals were decapitated immediately after a single social interaction test, there was a negative correlation between the social interaction time and 5-HIAA and 5-HT levels in the septum, but not in the frontal cortex or hippocampus. Thus, social behaviour is a stable trait, expression of which depends in part upon the partner’s social behaviour. This trait is negatively associated with 5-HT metabolism in the frontal cortex. Social activity of rats in a particular test situation may rather be related to 5-HT metabolism in the septum. © 2004 Elsevier Ireland Ltd. All rights reserved. Keywords: Sociability; Social interaction test; 5-Hydroxyindole acetic acid (5-HIAA); Frontal cortex; Septum; Rats; Traits
Social behaviour is the basis of one of the most generally accepted independent dimensions of personality [3], and has many important roles in the survival of the individual and species. Many clinical disorders include difficulties in creating or maintaining social contacts, such as social anxiety disorder, several personality disorders, autism, etc. The phenotypical expression of social behaviour is regulated by many different neurochemical systems, but a coherent picture is yet to emerge [18]. Recent studies on humans highlight the role of 5-HT in the expression of personality dimensions, mostly in these which are associated with affective and motivational processes [16,21]. To study the neurobiological mechanisms of social behaviour and related disorders in sufficient detail, animal models would be helpful. ∗
Corresponding author. Tel.: +372 7 375911; fax: +372 7 375900. E-mail address: [email protected] (J. Harro). 1 Present address: Institute of Cellular and Molecular Biology, University of Tartu, Riia 23, Tartu, Estonia.
Studies of social behaviour in rodents have mostly been carried out within the constraints of rather specific paradigms, such as aggressive or maternal behaviour. However, there is a frequently used simple animal model that was developed to measure anxiogenic and anxiolytic drug effects [7] but which is based on social behaviour: the social interaction test, in which the time spent in active social interaction between two unfamiliar rats in a neutral arena is measured. Behaviour of rats in the social interaction test does not correlate well with their performance in other animal models of anxiety [20]. This suggests that the model has other important underlying mechanisms than just general anxiety, and could be used for studying neurobiology of social behaviour provided that social behaviour of an animal would be a consistently expressed trait in this test. The purpose of the present study was to address this question and to investigate whether the trait of social behaviour, if measurable, would correlate with monoamine levels in the frontal cortex, septum and hippocampus. Pharmacological interventions in these brain regions have been found
0304-3940/$ – see front matter © 2004 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.neulet.2004.06.023
310
M. Tõnissaar et al. / Neuroscience Letters 367 (2004) 309–312
to influence behaviour in the social interaction test [8,15], and even though serotonergic mechanisms are most consistently implicated in animal and human studies on social behaviour [4,16], catecholaminegic mechanisms have also been suggested to contribute [3]. We also studied whether acute performance in the social interaction test itself has any acute effect on monoamines, or would correlate with ex vivo monoamines. Male Wistar rats from the National Laboratory Animal Centre, Kuopio, Finland were group-housed after arrival in plastic cages with food (Lactamin R35, Sweden) and water ad libitum. Room temperature was maintained at 21 ± 2 ◦ C and 12-h light:12-h dark (lights on at 07:00 h) was applied. Animals weighing 278–382 g at the beginning of the experiment were used. The test developed by File and Hyde [7] was used as described earlier [12]. Rats were housed individually at least 10 days before testing. Two unfamiliar, weight-matched rats were placed for 10 min into a brightly-lit chamber (30 cm × 30 cm × 60 cm) with floor covered with wood shavings. The total time spent in active social behaviour (allogrooming, sniffing the partner, crawling under and over, following) was recorded by two observers. Inter-rater reliability was always found to be good (r > 0.90). Experiment I measured whether the social behaviour of a given rat is stable in the social interaction test. Social interaction test was carried out in 20 animals in four separate sessions with intervals of 10 days. Rats were paired on the basis of their body weight before the test and every test was carried out with an unfamiliar partner. Animals were decapitated 4 days after the fourth test and brains immediately dissected on ice, with the rat brain atlas of Paxinos and Watson as a guide. Samples of frontal cortex also contain the area 1 of the cingulate cortex. Septum represents the septal complex between the lateral ventricles in the striatal slice cut approximately between the bregma level and 1.4 mm ahead of it. Hippocampi were separated from overlaying cerebral cortex as a whole after removal of the basal forebrain and other brain parts. The purpose of Experiment II was to clarify whether exposure to the social interaction test itself has any effect on monoamine levels. After 10 days of single housing, thirty animals were divided into three groups (n = 10). One group of animals was directly transported from home cages to a
separate room and immediately decapitated; animals of the second group were singly placed into the social interaction test box for 10 min and then decapitated. The third group of animals was decapitated immediately after the social interaction test had been carried out as usual. Monoamines were assayed essentially as described in [12] with minor modifications in the method. All calculations were performed using StatView 4.5 software (Abacus Concepts. Cary, NC, USA). Group comparisons were made by repeated measures or one-way ANOVA as appropriate. Correlations shown are Pearson correlation coefficients. Statistical significance was set at P < 0.05. In Experiment I, the mean duration of social interaction over all four tests was 94 s. According to repeated measures ANOVA, mean social interaction time did not differ significantly between the tests. In Table 1, pairwise correlations between social interaction times of individual animals in different tests are presented. Correlations were stronger between tests, which followed each other, but the overall picture suggests that there is no consistent correlation between performances in single tests. Nevertheless, social interaction time in each single test correlated significantly with the mean values of social activity as calculated on the basis of performance in all tests or in the other three tests (Table 1). It should be noted that inclusion of the data from a particular test into the average social activity measure artificially increases the correlation, but these correlations have also been shown because the small number of independent observations is also likely to introduce a bias. Social interaction time of rats correlated moderately but significantly with the partner’s social activity across all tests (r = 0.35, P < 0.001) which is thus a source of variation. We carried out only four tests because it was not possible to find more weight-matched unfamiliar partners in the group of animals used in this experiment. In the fourth test the mean difference in body weight was larger than considered optimal (7 g in test I; 11 g in test II; 14 g in test III and 20 g in test IV), and there was a tendency of weaker correlation between the fourth test and other tests. Differences between body weights of partners in the social interaction test did not correlate with social interaction time, however. We carried out linear regression analyses of rat mean social interaction time across four tests as dependent of tissue levels of NA, DA, DOPAC, HVA, 5-HT, and 5-HIAA in
Table 1 Correlations between the social interaction (SI) tests or between each test and the mean of the other three tests or with the mean of all tests SI test I SI SI SI SI
test test test test ∗
I II III IV
P < 0.05. P < 0.01. ∗∗∗ P < 0.001. ∗∗∗∗ P < 0.0001. ∗∗
0.64∗ 0.28 0.34
SI test II
0.41 0.14
SI test III
0.60∗
Mean of the other three tests
Mean of all tests
0.58∗∗
0.73∗∗∗ 0.77∗∗∗∗ 0.79∗∗∗∗ 0.68∗∗∗
0.55∗∗ 0.65∗∗∗ 0.50∗
M. Tõnissaar et al. / Neuroscience Letters 367 (2004) 309–312
311
Fig. 1. Experiment I: Regression analysis of 5-HIAA levels in the frontal cortex and social interaction time in the social interaction test (mean of four tests) [r 2 = 0.44; F(1; 16) = 12.82, P < 0.01].
the frontal cortex, hippocampus, and septum. A significant correlation (r = −0.67, P < 0.01) was found between the mean social interaction time and 5-HIAA levels in the frontal cortex (Fig. 1). (Data of two rats were excluded because of anomalously high 5-HIAA values, several fold the group mean, probably due to inadequate sample handling.) Mean social interaction time did not correlate with any other monoamine measure, including the 5-HIAA/5-HT ratio. In Experiment II, ANOVA did not reveal any difference between the three groups in monoamine levels in any of the three brain regions (data not shown). In the animals of the group that underwent the social interaction test, the mean social interaction time was 73 s when one rat with exceptionally high social activity (227 s) was omitted from the analysis. There was a negative correlation between the social interaction time and 5-HIAA (r = −0.90, P < 0.01) and 5-HT levels (r = −0.79, P < 0.05) in the septum (Fig. 2A and B, respectively), but not in the frontal cortex or hippocampus. Evolutionary continuity between humans and other animals suggests that some dimensions of personality may be common across a wide range of species [6,9] and animals may also have persistent structures of behaviours, which may be stable across multiple years [6]. In the present study, activity in each social interaction test did not correlate consistently with other tests pairwise, but the individual average social behaviour correlated well with performance in each occasion, suggesting that in the rat, there is a sociability trait which is measurable by repeated testing in the social interaction test. In comparison to the physical environment, social environment is less predictable and very reactive to the behaviour of the individual itself. Social behaviour of rats was found to depend on the social activity of partners, and thus a conclusion can be drawn that social activity of an individual rat is a trait, manifestation of which depends on the partner’s social behaviour. We have reproduced the finding that mean social interaction time correlates consistently with performance in all tests in three subsequent independent experiments (unpublished).
Fig. 2. (A) Experiment II: Regression analysis of 5-HIAA levels in the septum and social interaction time in the social interaction test [r 2 = 0.81; F(1; 7) = 30.26, P < 0.001]. (B) Experiment II: Regression analysis of 5-HT levels in the septum and social interaction time in the social interaction test [r 2 = 0.62; F(1; 7) = 11.56, P < 0.05].
The sociability trait was found to be strongly negatively associated with the levels of 5-HIAA in the frontal cortex which were measured 4 days after the last social interaction test. Even though ex vivo 5-HIAA levels cannot be interpreted unambigously, lower levels of a metabolite are suggestive of lower basal release. Thus this result was not expected because several well-known CSF studies have found a high positive correlation between serotonin metabolism and sociability [13,17,19]. Nevertheless, there is also evidence for negative association between 5-HT metabolism and social competence [22]. Even in the simple social interaction test, both increased [4,10] and reduced [1,8,14,15] social behaviour has been suggested to be mediated via an increase in serotonergic function. In our experiments, the negative association between septal 5-HT measures and social activity in the experiment in which the animals were decapitated immediately after social interaction test further supports the notions that at least in certain conditions 5-HT metabolism is inhibitory to social activity, but that different brain regions contribute to sociability trait and ongoing social activity. Differences in methodological detail in sample collection and processing could account for some of the variance across laboratories, but these could be more easy explana-
312
M. Tõnissaar et al. / Neuroscience Letters 367 (2004) 309–312
tions for data not confirming positive results, whereas in the case of association of 5-HT and social behaviour we are rather confronting data which show significant correlations in opposite directions. Serotonergic measures are probably related to sociability-associated mechanisms in a complex manner, and this can cause differences between laboratories in several ways. For example, low 5-HIAA levels in the CSF have been associated with high impulsivity, which may mediate social incompetence [13,17]. By using in vivo microdialysis, Dalley et al. [2] recently demonstrated that higher extracellular levels of 5-HT in the prefrontal cortex are associated with higher impulsivity. It is possible that 5-HIAA levels in the CSF reflect better serotonergic activity in the brainstem, which is a negative mirror image of forebrain mechanisms due to somatodendritic inhibition of 5-HT projection neurons. However, it is also possible that, when the environmental signals become more complex, the 5-HT behaviour link is not linear. As an example, we have recently demonstrated in a longitudinal study that platelet MAO activity, a peripheral marker of central serotonergic activity [5], is predictive of regular smoking in adolescents, but the relationship is very clearly U-shaped [11]. Apparently the different contribution of a variety of 5-HT receptor subtypes in different brain regions should be considered. In summary, these experiments demonstrate that a sociability trait underlies the performance in the rat social interaction test, and that this trait is negatively associated with frontal 5-HT metabolism.
Acknowledgements This work was supported by the Estonian Science Foundation Grant No. 4531, the Estonian Ministry of Education and Science project No. 2643, and the EU Framework 6 Integrated Project NEWMOOD (LSHM-CT-2004-503474).
References [1] G. Bagdy, M. Graf, Z.E. Anheuer, E.A. Modos, S. Kantor, Anxiety-like effects induced by acute fluoxetine, sertraline or m-CPP treatment are reversed by pretreatment with the 5-HT2C receptor antagonist SB-242084 but not the 5-HT1A receptor antagonist WAY-100635, Int. J. Neuropsychopharmacol. 4 (2001) 399–408. [2] W.J. Dalley, D.E. Theobald, D.M. Eagle, F. Passetti, T.W. Robbins, Deficits in impulse control associated with tonically-elevated serotonergic function in rat prefrontal cortex, Neuropsychopharmacology 26 (2002) 716–728. [3] R.A. Depue, P.F. Collins, Neurobiology of the structure of personality: dopamine, facilitation of incentive motivation, and extraversion, Behav. Brain Sci. 22 (1999) 491–569. [4] M.S. Duxon, K.R. Starr, N. Upton, Latency to paroxetine-induced anxiolysis in the rat is reduced by co-administration of the 5-HT1A receptor antagonist WAY 100635, Br. J. Pharmacol. 130 (2000) 1713–1719.
[5] C. Fahlke, H. Garpenstrand, L. Oreland, S.J. Suomi, J.D. Higley, Platelet monoamine oxidase activity in a nohuman primate model of type 2 excessive alcohol consumption, Am. J. Psychiatry 159 (2002) 2107–2109. [6] A.J. Figueredo, R.L. Cox, R.J. Rhine, A generalizability analysis of subjective personality assessments in the stumptail macaque and the zebra finch, Multivariate Behav. Res. 30 (1995) 167–197. [7] S.E. File, J.R. Hyde, Can social interaction be used to measure anxiety? Br. J. Pharmacol. 62 (1978) 19–24. [8] S.E. File, H. Zangrossi, N. Andrews, Social interaction and elevated plus-maze tests: changes in release and uptake of 5-HT and GABA, Neuropharmacology 32 (1993) 217–221. [9] S.D. Gosling, O.P. John, Personality dimension in non-human animals: a cross-species review, Curr. Directions Psychol. Sci. 8 (1999) 69–75. [10] M. Hamon, E. Doucet, K. Lefevre, M.-C. Miquel, L. Lanfumey, R. Insausti, D. Frechilla, J. Del Rio, D. Verge, Antibodies and antisense oligonucleotide for probing the distribution and putative functions of central 5-HT6 receptors, Neuropsychopharmacology 21 (1999) 68S– 76S. [11] J. Harro, K. Fischer, S. Vansteelandt, M. Harro, Both low and high platelet monoamine oxidase increase the probability of becoming a smoker, Eur. Neuropsychopharmacol. 14 (2004) 65–69. [12] J. Harro, M. Tõnissaar, M. Eller, A. Kask, L. Oreland, Chronic variable stress and partial 5-HT denervation by parachloroamphetamine treatment in the rat: effects on behavior and monoamine neurochemistry, Brain Res. 899 (2001) 227–239. [13] J.D. Higley, P.T. Mehlman, S.B. Higley, B. Fernald, J. Vickers, S.G. Lindell, D.M. Taub, S.J. Suomi, M. Linnoila, Excessive mortality in young free-ranging male nonhuman primates with low cerebrospinal fluid 5-hydroxyindoleacetic acid concentrations, Arch. Gen. Psychiatry 53 (1996) 537–543. [14] A.J. Kennedy, E.L. Gibson, M.T. O’Connell, G. Curzon, Effects of housing, restraint and chronic treatments with mCPP and sertraline on behavioural responses to mCPP, Psychopharmacology 113 (1993) 262–268. [15] P.J. Kenny, S. Cheeta, S.E. File, Anxiogenic effects of nicotine in the dorsal hippocampus are mediated by 5-HT1A and not muscarinic M1 receptors, Neuropharmacology 39 (2000) 300–307. [16] B. Knutson, O.M. Wolkowitz, S.W. Cole, T. Chan, E.A. Moore, R.C. Johnson, J. Terpstra, R.A. Turner, V.I. Reus, Selective alteration of personality and social behaviour by serotonergic intervention, Am. J. Psychiatry 155 (1998) 373–379. [17] P.T. Mehlman, J.D. Higley, I. Faucher, A.A. Lilly, D.M. Taub, J. Vickers, S.J. Suomi, M. Linnoila, Correlation of CSF 5-HIAA concentration with sociality and the timing of emigration in free-ranging primates, Am. J. Psychiatry 152 (1995) 907–913. [18] J. Panksepp, Affective Neuroscience: The Foundations of Human and Animal Emotions, Oxford University Press, 1998, 480 pp. [19] G.P. Placidi, M.A. Oquendo, K.M. Malone, Y.Y. Huang, S.P. Ellis, J.J. Mann, Aggressivity, suicide attempts, and depression: relationship to cerebrospinal fluid monoamine metabolite levels, Biol. Psychiatry 50 (2001) 783–791. [20] A. Ramos, O. Berton, P. Mormede, F. Chaouloff, A multiple-test study of anxiety-related behaviours in six inbred rat strains, Behav. Brain Res. 85 (1997) 57–69. [21] W.S. Tse, A.J. Bond, Serotonergic intervention affects both social dominance and affiliative behaviour, Psychopharmacology 161 (2002) 324–330. [22] U. Yodyingyuad, C. de la Riva, D.H. Abbott, J. Herbert, E.B. Keverne, Relationship between dominance hierarchy, cerebrospinal fluid levels of amine transmitter metabolites (5-hydroxyindole acetic acid and homovanillic acid) and plasma cortisol in monkeys, Neuroscience 16 (1985) 851–858.