- Email: [email protected]
One-way avoidance acquisition and cellular density in the basolateral amygdala: Strain differences in Roman high- and low-avoidance rats
Neuroscience Letters 450 (2009) 317–320
Contents lists available at ScienceDirect
Neuroscience Letters journal homepage: www.elsevier.com/locate/neulet
One-way avoidance acquisition and cellular density in the basolateral amygdala: Strain differences in Roman high- and low-avoidance rats Ma José Gómez a , Ignacio Morón b , Carmen Torres a , Francisco José Esteban c , Lourdes de la Torre a , ˜ e, Antonio Cándido d , Antonio Maldonado d , Alberto Fernández-Teruel e , Adolf Tobena a a,∗ M Dolores Escarabajal a
Department of Psychology, Faculty of Humanities, University of Jaén, Paraje de Las Lagunillas s/n, 23071, Jaén, Spain F. Oloriz Institute of Neurosciences, University of Granada, Granada, Spain c Department of Experimental Biology, Faculty of Experimental Sciences, University of Jaén, Paraje de Las Lagunillas s/n, 23071, Jaén, Spain d Department of Experimental Psychology and Physiology of Behavior, Faculty of Psychology, University of Granada, 18071, Granada, Spain e Medical Psychology Unit, Department of Psychiatry and Forensic Medicine, Institute of Neurosciences, School of Medicine, Autonomous University of Barcelona, Campus Bellaterra, 08193, Barcelona, Spain b
a r t i c l e
i n f o
Article history: Received 29 July 2008 Received in revised form 28 October 2008 Accepted 28 October 2008 Keywords: Anxiety Avoidance Basolateral amygdala Cellular density Roman high- and low-avoidance rats
a b s t r a c t The goal of the present experiment was to study the performance of inbred Roman high- (RHA-I) and low- (RLA-I) avoidance rats in one-way avoidance learning and to relate the behaviour of the animals to cellular density in the basolateral amygdala (BLA), a brain region related to fear and anxiety. Thus, females from both strains were exposed either to 30 s or 1 s in the safe place as a function of experimental condition, until they reached five consecutive avoidance responses. Thereafter, the rats were perfused, and their brains sectioned in 40 m coronal sections, stained with cresyl violet. The area (percentage of field) corresponding to the BLA structures was quantified by computerized-assisted image analysis. The results indicated that RLA-I showed a significantly poorer acquisition of the one-way avoidance task than did RHA-I rats, but only when safe time was the shortest (1 s). In addition, the number of trials needed to reach the behavioural acquisition criterion was negatively correlated with BLA cellular density in RLA-I rats. These data suggest the possibility of relating behavioural and neuro-anatomical indexes, enabling exploration of the biological basis of fear/anxiety behaviours. © 2008 Published by Elsevier Ireland Ltd.
The Swiss sublines of Roman low-avoidance (RLA/Verh) and Roman high-avoidance (RHA/Verh) rats were initially selected and bred for extremely poor vs. rapid acquisition of active two-way avoidance behaviour in the shuttlebox [4], using stock from the original RLA and RHA rats [1]. Inbred strains (RLA-I and RHA-I) derived from the two Swiss (outbred) sublines have been bred and maintained at the Autonomous University of Barcelona since 1997 [5,6]. As a result of the selection for divergent avoidance acquisition, clear behavioural differences have been found between RHA and RLA rats in a variety of tasks related to conflict, innate fear stimuli and frustrative non-reward exposure [8,11,13,17]. Similarly, strain/linebased divergences have also been observed in neuroendocrine indexes of stress, such as a higher activation of the HPA axis in RLA than RHA rats [see [2,15] for review], as well as neurochemical and neuroanatomical differences in brain structures related to fear and anxiety, including the amygdala [2,21,12]. Additionally, RHA-I/RLA-
∗ Corresponding author. Tel.: +34 953 21 26 64; fax: +34 953 21 18 81. E-mail address: [email protected] (M.D. Escarabajal). 0304-3940/$ – see front matter © 2008 Published by Elsevier Ireland Ltd. doi:10.1016/j.neulet.2008.10.112
I differences have been found in anxiogenic CRF projections of the central amygdala (CA), as well as in the number of neuropeptide Y neurons located in the basolateral amygdala (BLA), suggesting that the Roman rat lines/strains can be used as a model to study the neural mechanisms of fear/anxiety [22]. Recent research from our laboratory has demonstrated that the observed strain differences in two-way avoidance seem to generalize to one-way avoidance training when it is carried out under particularly difficult conditions. In this task, the rats learn to run from an aversive/danger compartment (where a warning signal is followed by an electric foot shock), to a safe one (where these stimuli never appear). We found that female, inbred RLA-I rats showed a poorer performance than their RHA-I counterparts when time spent in the safe compartment was shorter (1 s), but not when animals remained 30 s in the safe place. These strain differences were abolished by the IP administration of diazepam [18]. Considering the importance of the amygdala in the regulation of fear/anxiety behaviour, and that the BLA complex is a part of the amygdala where fear conditioning (i.e. the association between the conditioned stimulus and the aversive unconditioned stimulus) takes
318
M.J. Gómez et al. / Neuroscience Letters 450 (2009) 317–320
place [e.g. [19]], the present experiment was carried out with the purpose of (i) confirming the divergent performance of RHA-I and RLA-I rats in one-way avoidance acquisition, and (ii) conducting an histological study in the BLA to look for differences in cellular density which could be related to the behavioural divergencies obtained. Seventeen RHA-I and 16 RLA-I female rats from the colony of inbred Roman rats maintained at the Autonomous University of Barcelona were used. Animals were about 120 days old at the beginning of the experiment (weights between 280 and 330 g) and they were individually housed with food and tapwater ad lib. Room temperature was kept at about 20 ◦ C with a 12 h light/dark cycle. Training took place during the light phase. The day and the time of testing were counterbalanced across groups in order to avoid the influence of the oestrus cycle. The experiment was conducted following European Union (EU) guidelines on the use of animals for research (86/609/EEC). A Letica one-way avoidance chamber of two equal compartments, 27 cm long × 25 cm wide × 28 cm high, made of Plexiglas was used. The compartments were separated by a 0.5 cm thick partition 25 cm wide × 28 cm high, with a square 9 cm × 9 cm hole and a removable gate to allow movement between compartments. The floor in both compartments was hinged to operate a microswitch when depressed; this allowed the apparatus, procedure and responses to be controlled by a PC-XT microcomputer. A speaker was placed in the middle of the lateral wall so that half was oriented to the danger compartment and the other half to the safe compartment. The warning signal was a 2000 Hz tone of 88 dB. The danger compartment was fitted with a grid floor of 19 stainless steel rods 4 mm in diameter and spaced 2 cm apart center to center, connected in series to a Letica LI-2900 module capable of delivering a continuous scrambled shock. The roof of the danger compartment consisted of a black glass panel, which was removed only to put the rat into the chamber. A rigid, non-transparent white plastic carrying box 24 cm long × 14 cm wide × 19 cm high was placed in the safe compartment in contact with the communication hole. This box was used as the safe compartment and to move the rat when the safe time was completed. The carrying box had a handle on top and no wall on the side in contact with the partition of the avoidance chamber and, therefore, with the communication hole and gate. The floor, ceiling and walls of this box were made of the same Plexiglas material. An air extractor installed outside the avoidance chamber produced a background noise of 70 dB. Rats were removed from their homecages and put into the avoidance box, placed in an adjacent quiet room, where they were allowed to freely explore both compartments for 5 min. Thereafter, the communication gate was closed shutting the rat in the danger compartment, and the trials began. Each trial consisted of a warning signal (conditioned stimulus, CS) followed after 5 s by an electric shock of 1 mA (unconditioned stimulus, US). Both the warning signal and the shock were continued until the animal moved into the safe compartment, or until 30 s had elapsed. The gate between the two compartments was opened in the presence of the warning signal and closed when the rat entered the safe compartment. Time in the danger compartment before the onset of the warning signal was the same for all experimental conditions (15 s). Once the 30 s (when safety took longer becoming the task easier) vs. 1 s (when safety was shorter becoming the task more difficult) safe time elapsed, the transportation box was lifted over the apparatus and the rat was returned to the danger compartment (the roof of this compartment was opened briefly for about 2 s, and then closed). The box was then replaced in the safe compartment of the avoidance chamber. The rats were randomly assigned to one of the four groups (see Table 1).
Table 1 Time spent in the safe compartment of the one-way avoidance chamber for RHA-I and RLA-I groups. Groups
N
Time in the safe compartment (s)
Strain
30/RHA-I 1/RHA-I 30/RLA-I 1/RLA-I
8 9 8 8
30 1 30 1
Roman high-avoidance Roman high-avoidance Roman low-avoidance Roman low-avoidance
At the end of the behavioural procedure, each rat was deeply anaesthetized and intracardially perfused with a 0.9% saline followed by a 10% formaline solution. The brain was removed and post-fixed for a further 3 h in the same fixative solution at room temperature, after which it was rinsed by immersion in 30% sucrose, 0.1 M PB at 4 ◦ C overnight. After this, the brain was covered with an OCT (optimal cutting temperature) compound and frozen in isopentane prechilled in liquid nitrogen. Serial 40 m coronal sections were prepared using a cryostat (2800 Frigocut E, Reicher-Jung) and stained with cresyl violet in order to analyze the differential neural density in the basolateral amygdala. Five rat brains from each experimental group were randomly selected for the histological study. The stained sections were quantified by computerized-assisted image analysis using ImageJ (an NIH image analysis and processing software downloaded free from http://rsweb.nih.gov/ij/) connected to a light microscope (Olympus, Hamburg, Germany) as previously described [3]. Five different field areas of the BLA (regardless of the hemisphere) were measured: field at −2.56 from Bregma (Zone A), field at −2.80 from Bregma (Zone B), field at −3.14 from Bregma (Zone C), field at −3.30 from Bregma (Zone D) and field at −3.60 from Bregma (Zone E). One random 0.1 mm2 field (objective 40×) on each section and one to five sections of the BLA for each rat were digitally captured and analyzed after background subtraction (minimal particular size 50 pixels). The field areas of the BLA were chosen both according to the extension of the BLA and to the staining intensity in order to avoid repeated measurements. As a means of avoiding the usual biased segmentation of the stained structures, the stained areas of each field were calculated as a function of the optical density following Sternbergerˇıs method [16], where the line segment corresponding to staining was mathematically extrapolated to the axis representing the percentage of area per field. Number of trials to reach five consecutive avoidance responses (5CARS) was used as dependent variable. When an animal did not reach this criterion after 100 trials had elapsed, it was removed from the avoidance box and a value of 100 was assigned for that criterion. The criterion was considered to be met at the first of the sequence of consecutive responses. The data were submitted to a Kruskal–Wallis test for global significance. Consequently, Mann–Whitney U-tests between-groups comparisons were performed. For all statistical analysis, alpha was set at 0.05. Cellular density values obtained in five different BLA areas were used as dependent variable. The initial visual analysis of the data showed a variability in the density values of the different zones observed in the RLA-I, as compared to the distribution of the values from the RHA-I rats. That perception led us to determine if the values of both strains deviated from the normal distribution. With this objective, a Kolmogorov–Smirnov Test was applied, showing that both strain of rats had a different distribution of their values in the zones A, B, C and E (p ≤ 0.05). After this result, a non-parametric test comparing two independent samples (Mann-Whitney U-Test) was applied. Non-parametric Spearman Rank correlation coefficients were applied to behavioural and cellular measures (density in the 5 BLA
M.J. Gómez et al. / Neuroscience Letters 450 (2009) 317–320
Fig. 1. Mean number of trials to reach five consecutive avoidance responses (5CARS). Bars denote standard errors of the mean. (*): 30/RHA-I vs. 1/RHA-I, p < 0.05. (**): 30/RLA-I vs. 1/RLA-I, p < 0.05. (***): 1/RHA-I vs. 1/RLA-I, p < 0.05.
levels) to evaluate possible trends of association among these variables. Significance was considered at or below the 95% critical value (p ≤ 0.05). Fig. 1 presents the mean numbers of trials needed to reach five consecutive avoidance responses (5CARS) in RHA-I and RLAI groups exposed for 30 vs. 1 s in the safe compartment. Overall Kruskal-Wallis analysis of the results showed statistically significant differences among groups, H(3) = 19,930, p < 0.0001. Significant differences were found between RHA-I groups (30/RHA-I vs. 1/RHAI; U = 4.00, p < 0.002) and RLA-I groups (30/RLA-I vs. 1/RLA-I; U = 3.00, p < 0.002). Significant differences were also found between both strains when were exposed to 1 s in safe compartment (1/RHAI vs. 1/RLA-I; U = 12.00, p < 0.021) but not when were exposed to 30 s. Fig. 2 presents, for both rat strains, the mean neural density values in the different zones that were registered from the BLA areas. The comparison between both strains of rats showed a significantly higher neural density in Zone A (U = 30.00, p < 0.001), Zone C (U = 32.00, p < 0.001) and Zone E (U = 53.00, p < 0.02) in RLA-I rats, as well as a marginally significant trend in Zone B (p = 0.055). Finally, when the number of trials needed to reach 5CARS was correlated with the cellular density values obtained in the five BLA areas examined, the values for RLA-I rats (n = 10, collapsing both avoidance −1/RLA-I and 30/RLA-I- conditions) were “rho”= −0.67 (p < 0.04; BLA area A), −0.73 (p < 0.02; BLA area B), −0.71 (p = 0.03; BLA area C), −0.84 (p < 0.005; BLA area D) and −0.81 (p < 0.01; BLA area E). Such a consistent correlation pattern was not observed in RHA-I rats, where no correlation was significant (all “rho” between 0.08 and −0.39, non significant). Thus, only in RLA-I rats was a higher number of trials needed to reach the behavioural acquisition criterion associated with a lower value of cellular density.
319
The present findings indicate that the RHA-I/RLA-I differences observed in one-way avoidance learning are dependent on the time spent in the safe compartment. Thus, RLA-I rats showed a poorer avoidance performance than RHA-I rats when they remained 1 s in the safe compartment, but not when time in safety was increased to 30 s. The former experimental condition can be considered as highly aversive and more similar in this regard (i.e. aversiveness) to two-way avoidance, because animals have only a limited access to the reinforcing safe cues associated to the absence of foot shock. These data replicate previous findings obtained in our lab [17,18] and suggest that differences between strains of rats selectively bred for high and low rates of two-way active avoidance behaviour can generalize to one-way active avoidance learning when it is carried out under particularly difficult conditions, i.e. low reward/safety and higher fear/aversiveness. Secondly, we analysed the BLA, a brain structure related to fear and anxiety, obtaining significant strain differences in the mean values of cellular density in several regions of this limbic structure. The present data are in line with previous studies showing neuro-anatomical and functional differences in the amygdaloid complex between the Roman rat lines/strains [12,21,22]. Compared to the RHA line/strain, RLA rats have consistently shown higher facility for fear conditioning in a variety of situations including fear-potentiated startle [7,8,11,12,22]. Such evidence appears to be congruent with RLA-I rats also having an increased cellular density in the BLA, as shown in the present work, as this amygdaloid complex underlies memory traces for fear learning [19]. In this regard, the consistent pattern of negative correlations between avoidance performance and neuronal BLA density (in the five areas examined) in RLA-I rats (regardless of the “avoidance training condition”) is striking, as it suggests that the more the number of trials needed to reach 5CARS, the less the cellular density values found in the BLA. Although it has to be considered with caution (because of the low “n”), that consistent correlation trend makes it tempting to speculate that (at least) in the more fearful RLA-I rats (but not in the less fearful RHA-Is) a lower BLA neuronal density would be associated with a higher difficulty to solve the fear-driven one-way avoidance task (i.e. higher 5CARS scores) possibly because the strength of the CS-US association (in which BLA neurons are involved) is also weakened [19]. These data seem to be in keeping with the results obtained in previous studies in which it was observed that lesions or pharmacological inactivation of specific areas of the amygdala, including the BLA, significantly impaired one-way and two-way avoidance acquisition [9,10,14,20]. Although the histological between-strain differences observed in this study could hypothetically underlie the RHA-I/RLA-I divergencies obtained in one-way avoidance learning, more studies are needed in order to ascertain whether divergencies in amygdala morphology and/or function are causally related to differences in anxious behaviour and fear in those rat strains. Acknowledgements This research was supported by MCYT, Ministerio de Ciencia y Tecnología, Spanish grants to C. Torres (SEJ2004-03231/PSIC), A. Maldonado (SEJ2006-11906/PSIC), A. Fernández-Teruel and ˜ (SAF2003-03480), by the DGR (2005SGR-00885) and A. Tobena through EURATools European project (European Commission Contract n◦ LSHG-CT-2005-019015). References
Fig. 2. Mean neural density in the different zones of the BLA area in RHA-I and RLA-I rats. Bars denote standard errors of the mean. (*): RHA-I vs. RLA-I, p < 0.05.
[1] P.L. Broadhurst, G. Bignani, Correlative effects of psychogenetic selection. A study of the Roman high and low avoidance strains of rats, Behav. Res. Ther. 3 (1965) 273–280.
320
M.J. Gómez et al. / Neuroscience Letters 450 (2009) 317–320
˜ A. Fernández-Teruel, A. Armario, [2] J. Carrasco, C. Márquez, R. Nadal, A. Tobena, Characterization of central and peripheral components of the hypothalamuspituitary-adrenal axis in the inbred Roman rat strains, Psychoneuroendocrinology 33 (4) (2008) 437–445. [3] M.L. Del Moral, F.J. Esteban, R. Hernández, S. Blanco, F.J. Molina, E. Martínez-Lara, E. Siles, G. Viedma, A. Ruiz, J.A. Pedrosa, M.A. Peinado, Immunohistochemistry of neuronal nitric oxide synthase and protein nitration in the striatum of the aged rat, Microsc. Res. Techniq. 64 (4) (2004) 304–311. [4] P. Driscoll, K. Bättig, Behavioral, emotional and neurochemical profiles of rats selected for extreme differences in active, two-way avoidance performance, in: I. Lieblich (Ed.), Genetics of the Brain, Elsevier, Amsterdam, 1982, pp. 95–123. [5] P. Driscoll, R.M. Escorihuela, A. Fernández-Teruel, O. Giorgi, H. Schwegler, Th. Steimer, A. Wiersma, M.G. Corda, J. Flint, J.M. Koolhaas, W. Langhans, P.E. Schulz, ˜ Genetic selection and differential stress responses: the J. Siegel, A. Tobena, Roman lines/strains of rats, in: P. Csermely (Ed.), Stress of Life: From Molecules to Man, New York Academy of Sciences, New York, NY, US, 1998, pp. 501–510. ˜ P. Driscoll, [6] R.M. Escorihuela, A. Fernández-Teruel, L. Gil, R. Aguilar, A. Tobena, Inbred Roman high- and low- avoidance rats: differences in anxiety, noveltyseeking, and shuttlebox behaviors, Physiol. Behav. 67 (1) (1999) 19–26. ˜ W. Langhans, K. Bättig, P. [7] R.M. Escorihuela, A. Fernández-Teruel, A. Tobena, Driscoll, Labyrinth exploration, emotional reactivity, and conditioned fear in young Roman/Verh inbred rats, Behav. Genet. 27 (6) (1997) 573–578. [8] P. Ferré, A. Fernández-Teruel, R.M. Escorihuela, P. Driscoll, M.G. Corda, O. Giorgi, ˜ Behavior of the Roman/Verh high- and low-avoidance rat lines in A. Tobena, anxiety tests: relationship with defecation and self-grooming, Physiol. Behav. 58 (6) (1995) 1209–1213. [9] S.P. Grossman, L. Grossman, L. Walsh, Functional organization of the rat amydala with respect to avoidance behavior, J. Comp. Physiol. Psychol. 88 (2) (1975) 829–850. [10] S. Killcross, T.W. Robbins, B.J. Everitt, Different types of fear-conditioned behaviour mediated by separate nuclei within amygdala, Nature 388 (1997) 377–380. ˜ [11] R. López-Aumatell, G. Blázquez, L. Gil, R. Aguilar, T. Canete, L. Giménez-Llort, A. ˜ A. Fernández-Teruel, The Roman High- and Low-avoidance rat strains Tobena, differ in fear-potentiated startle and classical aversive conditioning, Psicothema 21 (1) (2009) 27–32. [12] B. Roozendaal, A. Wiersma, P. Driscoll, J.M. Koolhaas, B. Bohus, Vasopressinergic modulation of stress responses in the central amygdala of the
[13]
[14]
[15]
[16]
[17]
[18]
[19]
[20]
[21]
[22]
Roman high-avoidance and low-avoidance rat, Brain Res. 596 (1–2) (1992) 35–40. J.M. Rosas, J.E. Callejas-Aguilera, M.D. Escarabajal, M.J. Gómez, L. de la Torre, A. ˜ A. Fernández-Teruel, C. Torres, Successive negative contrast Agüero, A. Tobena, effect in instrumental runway behaviour: a study with Roman high- (RHA) and Roman low- (RLA) avoidance rats, Behav. Brain Res. 185 (1) (2007) 1–8. ˜ A. Savonenko, T. Werka, E. Nikolaev, K. Zielinski, L. Kaczmarek, Complex effects of NMDA receptor antagonist APV in the basolateral amygdala on acquisition of two-way avoidance reaction and long-term fear memory, Learn. Mem. 10 (4) (2003) 293–303. T. Steimer, P. Driscoll, Divergent stress responses and coping styles in psychogenetically selected Roman high-(RHA) and low-(RLA) avoidance rats: behavioural, neuroendocrine and developmental aspects, Stress 6 (2) (2003) 87–100. L.A. Sternberger, N.H. Sternberger, The unlabeled antibody method: comparison of peroxidase-antiperoxidase with avidin-biotin complex by a new method of quantification, J. Histochem. Cytochem. 34 (5) (1986) 599–605. ˜ M.C. Torres, A. Cándido, M.D. Escarabajal, L. de la Torre, A. Maldonado, A. Tobena, A. Fernández-Teruel, Successive negative contrast effect in one-way avoidance learning in female roman rats, Physiol. Behav. 85 (4) (2005) 377–382. M.C. Torres, M.D. Escarabajal, A. Cándido, L. de la Torre, M.J. Gómez, A. Maldon˜ A. Fernández-Teruel, One-way avoidance learning and diazepam ado, A. Tobena, in female roman high-avoidance and low-avoidance rats, Behav. Pharmacol. 18 (3) (2007) 251–253. D.L. Walker, M. Davis, The role of amygdala glutamate receptors in fear learning, fear-potentiated startle, and extinction, Pharmacol. Biochem. Behav. 71 (3) (2002) 379–392. T. Werka, J. Skår, H. Ursin, Exploration and avoidance in rats with lesions in amygdala and piriform cortex, J. Comp. Physiol. Psychol. 92 (4) (1978) 672–681. A. Wiersma, J.P. Konsman, S. Knollema, B. Bohus, J.M. Koolhaas, Differential effects of CRH infusion into the central nucleus of the amygdala in the Roman high-avoidance and low-avoidance rats, Psychoneuroendocrinology 23 (3) (1998) 261–274. D.M. Yilmazer-Hanke, H. Faber-Zuschratter, R. Linke, H. Schwegler, Contribution of amygdala neurons containing peptides and calcium-binding proteins to fearpotentiated startle and exploration-related anxiety in inbred Roman high- and low-avoidance rats, Eur. J. Neurosci. 15 (7) (2002) 1206–1218.
Contents lists available at ScienceDirect
Neuroscience Letters journal homepage: www.elsevier.com/locate/neulet
One-way avoidance acquisition and cellular density in the basolateral amygdala: Strain differences in Roman high- and low-avoidance rats Ma José Gómez a , Ignacio Morón b , Carmen Torres a , Francisco José Esteban c , Lourdes de la Torre a , ˜ e, Antonio Cándido d , Antonio Maldonado d , Alberto Fernández-Teruel e , Adolf Tobena a a,∗ M Dolores Escarabajal a
Department of Psychology, Faculty of Humanities, University of Jaén, Paraje de Las Lagunillas s/n, 23071, Jaén, Spain F. Oloriz Institute of Neurosciences, University of Granada, Granada, Spain c Department of Experimental Biology, Faculty of Experimental Sciences, University of Jaén, Paraje de Las Lagunillas s/n, 23071, Jaén, Spain d Department of Experimental Psychology and Physiology of Behavior, Faculty of Psychology, University of Granada, 18071, Granada, Spain e Medical Psychology Unit, Department of Psychiatry and Forensic Medicine, Institute of Neurosciences, School of Medicine, Autonomous University of Barcelona, Campus Bellaterra, 08193, Barcelona, Spain b
a r t i c l e
i n f o
Article history: Received 29 July 2008 Received in revised form 28 October 2008 Accepted 28 October 2008 Keywords: Anxiety Avoidance Basolateral amygdala Cellular density Roman high- and low-avoidance rats
a b s t r a c t The goal of the present experiment was to study the performance of inbred Roman high- (RHA-I) and low- (RLA-I) avoidance rats in one-way avoidance learning and to relate the behaviour of the animals to cellular density in the basolateral amygdala (BLA), a brain region related to fear and anxiety. Thus, females from both strains were exposed either to 30 s or 1 s in the safe place as a function of experimental condition, until they reached five consecutive avoidance responses. Thereafter, the rats were perfused, and their brains sectioned in 40 m coronal sections, stained with cresyl violet. The area (percentage of field) corresponding to the BLA structures was quantified by computerized-assisted image analysis. The results indicated that RLA-I showed a significantly poorer acquisition of the one-way avoidance task than did RHA-I rats, but only when safe time was the shortest (1 s). In addition, the number of trials needed to reach the behavioural acquisition criterion was negatively correlated with BLA cellular density in RLA-I rats. These data suggest the possibility of relating behavioural and neuro-anatomical indexes, enabling exploration of the biological basis of fear/anxiety behaviours. © 2008 Published by Elsevier Ireland Ltd.
The Swiss sublines of Roman low-avoidance (RLA/Verh) and Roman high-avoidance (RHA/Verh) rats were initially selected and bred for extremely poor vs. rapid acquisition of active two-way avoidance behaviour in the shuttlebox [4], using stock from the original RLA and RHA rats [1]. Inbred strains (RLA-I and RHA-I) derived from the two Swiss (outbred) sublines have been bred and maintained at the Autonomous University of Barcelona since 1997 [5,6]. As a result of the selection for divergent avoidance acquisition, clear behavioural differences have been found between RHA and RLA rats in a variety of tasks related to conflict, innate fear stimuli and frustrative non-reward exposure [8,11,13,17]. Similarly, strain/linebased divergences have also been observed in neuroendocrine indexes of stress, such as a higher activation of the HPA axis in RLA than RHA rats [see [2,15] for review], as well as neurochemical and neuroanatomical differences in brain structures related to fear and anxiety, including the amygdala [2,21,12]. Additionally, RHA-I/RLA-
∗ Corresponding author. Tel.: +34 953 21 26 64; fax: +34 953 21 18 81. E-mail address: [email protected] (M.D. Escarabajal). 0304-3940/$ – see front matter © 2008 Published by Elsevier Ireland Ltd. doi:10.1016/j.neulet.2008.10.112
I differences have been found in anxiogenic CRF projections of the central amygdala (CA), as well as in the number of neuropeptide Y neurons located in the basolateral amygdala (BLA), suggesting that the Roman rat lines/strains can be used as a model to study the neural mechanisms of fear/anxiety [22]. Recent research from our laboratory has demonstrated that the observed strain differences in two-way avoidance seem to generalize to one-way avoidance training when it is carried out under particularly difficult conditions. In this task, the rats learn to run from an aversive/danger compartment (where a warning signal is followed by an electric foot shock), to a safe one (where these stimuli never appear). We found that female, inbred RLA-I rats showed a poorer performance than their RHA-I counterparts when time spent in the safe compartment was shorter (1 s), but not when animals remained 30 s in the safe place. These strain differences were abolished by the IP administration of diazepam [18]. Considering the importance of the amygdala in the regulation of fear/anxiety behaviour, and that the BLA complex is a part of the amygdala where fear conditioning (i.e. the association between the conditioned stimulus and the aversive unconditioned stimulus) takes
318
M.J. Gómez et al. / Neuroscience Letters 450 (2009) 317–320
place [e.g. [19]], the present experiment was carried out with the purpose of (i) confirming the divergent performance of RHA-I and RLA-I rats in one-way avoidance acquisition, and (ii) conducting an histological study in the BLA to look for differences in cellular density which could be related to the behavioural divergencies obtained. Seventeen RHA-I and 16 RLA-I female rats from the colony of inbred Roman rats maintained at the Autonomous University of Barcelona were used. Animals were about 120 days old at the beginning of the experiment (weights between 280 and 330 g) and they were individually housed with food and tapwater ad lib. Room temperature was kept at about 20 ◦ C with a 12 h light/dark cycle. Training took place during the light phase. The day and the time of testing were counterbalanced across groups in order to avoid the influence of the oestrus cycle. The experiment was conducted following European Union (EU) guidelines on the use of animals for research (86/609/EEC). A Letica one-way avoidance chamber of two equal compartments, 27 cm long × 25 cm wide × 28 cm high, made of Plexiglas was used. The compartments were separated by a 0.5 cm thick partition 25 cm wide × 28 cm high, with a square 9 cm × 9 cm hole and a removable gate to allow movement between compartments. The floor in both compartments was hinged to operate a microswitch when depressed; this allowed the apparatus, procedure and responses to be controlled by a PC-XT microcomputer. A speaker was placed in the middle of the lateral wall so that half was oriented to the danger compartment and the other half to the safe compartment. The warning signal was a 2000 Hz tone of 88 dB. The danger compartment was fitted with a grid floor of 19 stainless steel rods 4 mm in diameter and spaced 2 cm apart center to center, connected in series to a Letica LI-2900 module capable of delivering a continuous scrambled shock. The roof of the danger compartment consisted of a black glass panel, which was removed only to put the rat into the chamber. A rigid, non-transparent white plastic carrying box 24 cm long × 14 cm wide × 19 cm high was placed in the safe compartment in contact with the communication hole. This box was used as the safe compartment and to move the rat when the safe time was completed. The carrying box had a handle on top and no wall on the side in contact with the partition of the avoidance chamber and, therefore, with the communication hole and gate. The floor, ceiling and walls of this box were made of the same Plexiglas material. An air extractor installed outside the avoidance chamber produced a background noise of 70 dB. Rats were removed from their homecages and put into the avoidance box, placed in an adjacent quiet room, where they were allowed to freely explore both compartments for 5 min. Thereafter, the communication gate was closed shutting the rat in the danger compartment, and the trials began. Each trial consisted of a warning signal (conditioned stimulus, CS) followed after 5 s by an electric shock of 1 mA (unconditioned stimulus, US). Both the warning signal and the shock were continued until the animal moved into the safe compartment, or until 30 s had elapsed. The gate between the two compartments was opened in the presence of the warning signal and closed when the rat entered the safe compartment. Time in the danger compartment before the onset of the warning signal was the same for all experimental conditions (15 s). Once the 30 s (when safety took longer becoming the task easier) vs. 1 s (when safety was shorter becoming the task more difficult) safe time elapsed, the transportation box was lifted over the apparatus and the rat was returned to the danger compartment (the roof of this compartment was opened briefly for about 2 s, and then closed). The box was then replaced in the safe compartment of the avoidance chamber. The rats were randomly assigned to one of the four groups (see Table 1).
Table 1 Time spent in the safe compartment of the one-way avoidance chamber for RHA-I and RLA-I groups. Groups
N
Time in the safe compartment (s)
Strain
30/RHA-I 1/RHA-I 30/RLA-I 1/RLA-I
8 9 8 8
30 1 30 1
Roman high-avoidance Roman high-avoidance Roman low-avoidance Roman low-avoidance
At the end of the behavioural procedure, each rat was deeply anaesthetized and intracardially perfused with a 0.9% saline followed by a 10% formaline solution. The brain was removed and post-fixed for a further 3 h in the same fixative solution at room temperature, after which it was rinsed by immersion in 30% sucrose, 0.1 M PB at 4 ◦ C overnight. After this, the brain was covered with an OCT (optimal cutting temperature) compound and frozen in isopentane prechilled in liquid nitrogen. Serial 40 m coronal sections were prepared using a cryostat (2800 Frigocut E, Reicher-Jung) and stained with cresyl violet in order to analyze the differential neural density in the basolateral amygdala. Five rat brains from each experimental group were randomly selected for the histological study. The stained sections were quantified by computerized-assisted image analysis using ImageJ (an NIH image analysis and processing software downloaded free from http://rsweb.nih.gov/ij/) connected to a light microscope (Olympus, Hamburg, Germany) as previously described [3]. Five different field areas of the BLA (regardless of the hemisphere) were measured: field at −2.56 from Bregma (Zone A), field at −2.80 from Bregma (Zone B), field at −3.14 from Bregma (Zone C), field at −3.30 from Bregma (Zone D) and field at −3.60 from Bregma (Zone E). One random 0.1 mm2 field (objective 40×) on each section and one to five sections of the BLA for each rat were digitally captured and analyzed after background subtraction (minimal particular size 50 pixels). The field areas of the BLA were chosen both according to the extension of the BLA and to the staining intensity in order to avoid repeated measurements. As a means of avoiding the usual biased segmentation of the stained structures, the stained areas of each field were calculated as a function of the optical density following Sternbergerˇıs method [16], where the line segment corresponding to staining was mathematically extrapolated to the axis representing the percentage of area per field. Number of trials to reach five consecutive avoidance responses (5CARS) was used as dependent variable. When an animal did not reach this criterion after 100 trials had elapsed, it was removed from the avoidance box and a value of 100 was assigned for that criterion. The criterion was considered to be met at the first of the sequence of consecutive responses. The data were submitted to a Kruskal–Wallis test for global significance. Consequently, Mann–Whitney U-tests between-groups comparisons were performed. For all statistical analysis, alpha was set at 0.05. Cellular density values obtained in five different BLA areas were used as dependent variable. The initial visual analysis of the data showed a variability in the density values of the different zones observed in the RLA-I, as compared to the distribution of the values from the RHA-I rats. That perception led us to determine if the values of both strains deviated from the normal distribution. With this objective, a Kolmogorov–Smirnov Test was applied, showing that both strain of rats had a different distribution of their values in the zones A, B, C and E (p ≤ 0.05). After this result, a non-parametric test comparing two independent samples (Mann-Whitney U-Test) was applied. Non-parametric Spearman Rank correlation coefficients were applied to behavioural and cellular measures (density in the 5 BLA
M.J. Gómez et al. / Neuroscience Letters 450 (2009) 317–320
Fig. 1. Mean number of trials to reach five consecutive avoidance responses (5CARS). Bars denote standard errors of the mean. (*): 30/RHA-I vs. 1/RHA-I, p < 0.05. (**): 30/RLA-I vs. 1/RLA-I, p < 0.05. (***): 1/RHA-I vs. 1/RLA-I, p < 0.05.
levels) to evaluate possible trends of association among these variables. Significance was considered at or below the 95% critical value (p ≤ 0.05). Fig. 1 presents the mean numbers of trials needed to reach five consecutive avoidance responses (5CARS) in RHA-I and RLAI groups exposed for 30 vs. 1 s in the safe compartment. Overall Kruskal-Wallis analysis of the results showed statistically significant differences among groups, H(3) = 19,930, p < 0.0001. Significant differences were found between RHA-I groups (30/RHA-I vs. 1/RHAI; U = 4.00, p < 0.002) and RLA-I groups (30/RLA-I vs. 1/RLA-I; U = 3.00, p < 0.002). Significant differences were also found between both strains when were exposed to 1 s in safe compartment (1/RHAI vs. 1/RLA-I; U = 12.00, p < 0.021) but not when were exposed to 30 s. Fig. 2 presents, for both rat strains, the mean neural density values in the different zones that were registered from the BLA areas. The comparison between both strains of rats showed a significantly higher neural density in Zone A (U = 30.00, p < 0.001), Zone C (U = 32.00, p < 0.001) and Zone E (U = 53.00, p < 0.02) in RLA-I rats, as well as a marginally significant trend in Zone B (p = 0.055). Finally, when the number of trials needed to reach 5CARS was correlated with the cellular density values obtained in the five BLA areas examined, the values for RLA-I rats (n = 10, collapsing both avoidance −1/RLA-I and 30/RLA-I- conditions) were “rho”= −0.67 (p < 0.04; BLA area A), −0.73 (p < 0.02; BLA area B), −0.71 (p = 0.03; BLA area C), −0.84 (p < 0.005; BLA area D) and −0.81 (p < 0.01; BLA area E). Such a consistent correlation pattern was not observed in RHA-I rats, where no correlation was significant (all “rho” between 0.08 and −0.39, non significant). Thus, only in RLA-I rats was a higher number of trials needed to reach the behavioural acquisition criterion associated with a lower value of cellular density.
319
The present findings indicate that the RHA-I/RLA-I differences observed in one-way avoidance learning are dependent on the time spent in the safe compartment. Thus, RLA-I rats showed a poorer avoidance performance than RHA-I rats when they remained 1 s in the safe compartment, but not when time in safety was increased to 30 s. The former experimental condition can be considered as highly aversive and more similar in this regard (i.e. aversiveness) to two-way avoidance, because animals have only a limited access to the reinforcing safe cues associated to the absence of foot shock. These data replicate previous findings obtained in our lab [17,18] and suggest that differences between strains of rats selectively bred for high and low rates of two-way active avoidance behaviour can generalize to one-way active avoidance learning when it is carried out under particularly difficult conditions, i.e. low reward/safety and higher fear/aversiveness. Secondly, we analysed the BLA, a brain structure related to fear and anxiety, obtaining significant strain differences in the mean values of cellular density in several regions of this limbic structure. The present data are in line with previous studies showing neuro-anatomical and functional differences in the amygdaloid complex between the Roman rat lines/strains [12,21,22]. Compared to the RHA line/strain, RLA rats have consistently shown higher facility for fear conditioning in a variety of situations including fear-potentiated startle [7,8,11,12,22]. Such evidence appears to be congruent with RLA-I rats also having an increased cellular density in the BLA, as shown in the present work, as this amygdaloid complex underlies memory traces for fear learning [19]. In this regard, the consistent pattern of negative correlations between avoidance performance and neuronal BLA density (in the five areas examined) in RLA-I rats (regardless of the “avoidance training condition”) is striking, as it suggests that the more the number of trials needed to reach 5CARS, the less the cellular density values found in the BLA. Although it has to be considered with caution (because of the low “n”), that consistent correlation trend makes it tempting to speculate that (at least) in the more fearful RLA-I rats (but not in the less fearful RHA-Is) a lower BLA neuronal density would be associated with a higher difficulty to solve the fear-driven one-way avoidance task (i.e. higher 5CARS scores) possibly because the strength of the CS-US association (in which BLA neurons are involved) is also weakened [19]. These data seem to be in keeping with the results obtained in previous studies in which it was observed that lesions or pharmacological inactivation of specific areas of the amygdala, including the BLA, significantly impaired one-way and two-way avoidance acquisition [9,10,14,20]. Although the histological between-strain differences observed in this study could hypothetically underlie the RHA-I/RLA-I divergencies obtained in one-way avoidance learning, more studies are needed in order to ascertain whether divergencies in amygdala morphology and/or function are causally related to differences in anxious behaviour and fear in those rat strains. Acknowledgements This research was supported by MCYT, Ministerio de Ciencia y Tecnología, Spanish grants to C. Torres (SEJ2004-03231/PSIC), A. Maldonado (SEJ2006-11906/PSIC), A. Fernández-Teruel and ˜ (SAF2003-03480), by the DGR (2005SGR-00885) and A. Tobena through EURATools European project (European Commission Contract n◦ LSHG-CT-2005-019015). References
Fig. 2. Mean neural density in the different zones of the BLA area in RHA-I and RLA-I rats. Bars denote standard errors of the mean. (*): RHA-I vs. RLA-I, p < 0.05.
[1] P.L. Broadhurst, G. Bignani, Correlative effects of psychogenetic selection. A study of the Roman high and low avoidance strains of rats, Behav. Res. Ther. 3 (1965) 273–280.
320
M.J. Gómez et al. / Neuroscience Letters 450 (2009) 317–320
˜ A. Fernández-Teruel, A. Armario, [2] J. Carrasco, C. Márquez, R. Nadal, A. Tobena, Characterization of central and peripheral components of the hypothalamuspituitary-adrenal axis in the inbred Roman rat strains, Psychoneuroendocrinology 33 (4) (2008) 437–445. [3] M.L. Del Moral, F.J. Esteban, R. Hernández, S. Blanco, F.J. Molina, E. Martínez-Lara, E. Siles, G. Viedma, A. Ruiz, J.A. Pedrosa, M.A. Peinado, Immunohistochemistry of neuronal nitric oxide synthase and protein nitration in the striatum of the aged rat, Microsc. Res. Techniq. 64 (4) (2004) 304–311. [4] P. Driscoll, K. Bättig, Behavioral, emotional and neurochemical profiles of rats selected for extreme differences in active, two-way avoidance performance, in: I. Lieblich (Ed.), Genetics of the Brain, Elsevier, Amsterdam, 1982, pp. 95–123. [5] P. Driscoll, R.M. Escorihuela, A. Fernández-Teruel, O. Giorgi, H. Schwegler, Th. Steimer, A. Wiersma, M.G. Corda, J. Flint, J.M. Koolhaas, W. Langhans, P.E. Schulz, ˜ Genetic selection and differential stress responses: the J. Siegel, A. Tobena, Roman lines/strains of rats, in: P. Csermely (Ed.), Stress of Life: From Molecules to Man, New York Academy of Sciences, New York, NY, US, 1998, pp. 501–510. ˜ P. Driscoll, [6] R.M. Escorihuela, A. Fernández-Teruel, L. Gil, R. Aguilar, A. Tobena, Inbred Roman high- and low- avoidance rats: differences in anxiety, noveltyseeking, and shuttlebox behaviors, Physiol. Behav. 67 (1) (1999) 19–26. ˜ W. Langhans, K. Bättig, P. [7] R.M. Escorihuela, A. Fernández-Teruel, A. Tobena, Driscoll, Labyrinth exploration, emotional reactivity, and conditioned fear in young Roman/Verh inbred rats, Behav. Genet. 27 (6) (1997) 573–578. [8] P. Ferré, A. Fernández-Teruel, R.M. Escorihuela, P. Driscoll, M.G. Corda, O. Giorgi, ˜ Behavior of the Roman/Verh high- and low-avoidance rat lines in A. Tobena, anxiety tests: relationship with defecation and self-grooming, Physiol. Behav. 58 (6) (1995) 1209–1213. [9] S.P. Grossman, L. Grossman, L. Walsh, Functional organization of the rat amydala with respect to avoidance behavior, J. Comp. Physiol. Psychol. 88 (2) (1975) 829–850. [10] S. Killcross, T.W. Robbins, B.J. Everitt, Different types of fear-conditioned behaviour mediated by separate nuclei within amygdala, Nature 388 (1997) 377–380. ˜ [11] R. López-Aumatell, G. Blázquez, L. Gil, R. Aguilar, T. Canete, L. Giménez-Llort, A. ˜ A. Fernández-Teruel, The Roman High- and Low-avoidance rat strains Tobena, differ in fear-potentiated startle and classical aversive conditioning, Psicothema 21 (1) (2009) 27–32. [12] B. Roozendaal, A. Wiersma, P. Driscoll, J.M. Koolhaas, B. Bohus, Vasopressinergic modulation of stress responses in the central amygdala of the
[13]
[14]
[15]
[16]
[17]
[18]
[19]
[20]
[21]
[22]
Roman high-avoidance and low-avoidance rat, Brain Res. 596 (1–2) (1992) 35–40. J.M. Rosas, J.E. Callejas-Aguilera, M.D. Escarabajal, M.J. Gómez, L. de la Torre, A. ˜ A. Fernández-Teruel, C. Torres, Successive negative contrast Agüero, A. Tobena, effect in instrumental runway behaviour: a study with Roman high- (RHA) and Roman low- (RLA) avoidance rats, Behav. Brain Res. 185 (1) (2007) 1–8. ˜ A. Savonenko, T. Werka, E. Nikolaev, K. Zielinski, L. Kaczmarek, Complex effects of NMDA receptor antagonist APV in the basolateral amygdala on acquisition of two-way avoidance reaction and long-term fear memory, Learn. Mem. 10 (4) (2003) 293–303. T. Steimer, P. Driscoll, Divergent stress responses and coping styles in psychogenetically selected Roman high-(RHA) and low-(RLA) avoidance rats: behavioural, neuroendocrine and developmental aspects, Stress 6 (2) (2003) 87–100. L.A. Sternberger, N.H. Sternberger, The unlabeled antibody method: comparison of peroxidase-antiperoxidase with avidin-biotin complex by a new method of quantification, J. Histochem. Cytochem. 34 (5) (1986) 599–605. ˜ M.C. Torres, A. Cándido, M.D. Escarabajal, L. de la Torre, A. Maldonado, A. Tobena, A. Fernández-Teruel, Successive negative contrast effect in one-way avoidance learning in female roman rats, Physiol. Behav. 85 (4) (2005) 377–382. M.C. Torres, M.D. Escarabajal, A. Cándido, L. de la Torre, M.J. Gómez, A. Maldon˜ A. Fernández-Teruel, One-way avoidance learning and diazepam ado, A. Tobena, in female roman high-avoidance and low-avoidance rats, Behav. Pharmacol. 18 (3) (2007) 251–253. D.L. Walker, M. Davis, The role of amygdala glutamate receptors in fear learning, fear-potentiated startle, and extinction, Pharmacol. Biochem. Behav. 71 (3) (2002) 379–392. T. Werka, J. Skår, H. Ursin, Exploration and avoidance in rats with lesions in amygdala and piriform cortex, J. Comp. Physiol. Psychol. 92 (4) (1978) 672–681. A. Wiersma, J.P. Konsman, S. Knollema, B. Bohus, J.M. Koolhaas, Differential effects of CRH infusion into the central nucleus of the amygdala in the Roman high-avoidance and low-avoidance rats, Psychoneuroendocrinology 23 (3) (1998) 261–274. D.M. Yilmazer-Hanke, H. Faber-Zuschratter, R. Linke, H. Schwegler, Contribution of amygdala neurons containing peptides and calcium-binding proteins to fearpotentiated startle and exploration-related anxiety in inbred Roman high- and low-avoidance rats, Eur. J. Neurosci. 15 (7) (2002) 1206–1218.