- Email: [email protected]
Murine taste-immune associative learning
Brain, Behavior, and Immunity 20 (2006) 527–531 www.elsevier.com/locate/ybrbi
Short communication
Murine taste-immune associative learning Maj-Britt Niemi a,1, Gustavo Pacheco-López a,¤,1, Wei Kou b, Margarete Härting b, Adriana del Rey c, Hugo O. Besedovsky c, Manfred Schedlowski a a
Chair of Psychology and Behavioral Immunobiology, Institute for Behavioral Sciences, ETH Zurich, Zurich 8092, Switzerland b Department of Medical Psychology, University of Duisburg-Essen, Essen 45122, Germany c Division of Immunophysiology, Institute of Physiology, Philipps University of Marburg, Marburg 35037, Germany Received 26 September 2005; received in revised form 24 February 2006; accepted 24 February 2006 Available online 21 April 2006
Abstract Taste-immune associative learning can result from contingent pairings of an immune-competent unconditioned stimulus (US) with a gustative conditioned stimulus (CS). Recalling such an association may induce a set of physiological responses aVecting behavior, endocrine, and immune functions. We have established a model of behaviorally conditioned immunosuppression employing the immunosuppressant drug cyclosporine A (CsA) as the US and saccharin as the CS in rats and humans. In order to investigate the inter-species generalization of this neuro-immune interaction, we tested the feasibility of this paradigm in mice. In a single-bottle scheme, male BALB/ c mice (n D 5) were conditioned by conducting three association trials and a single recall trial. Control groups (n D 5/group) were designed to assure associative learning, pharmacological eVects of the US, and placebo eVect. Results show that CsA-conditioned animals displayed signiWcant immunosuppression in the spleen after recall, measured by in vitro T-lymphocyte proliferation, and IL-2 production. However, the same animals did not show evidence of avoidance behavior to the CS. In contrast, evoking the association of saccharin–lithium chloride (inducing gastric malaise) in another set of animals (n D 4/group) resulted in signiWcant and pronounced avoidance of the taste (CS). These animals also displayed signiWcant suppression of splenic T-lymphocyte responsiveness after the recall phase. The present results indicate that mice seem to be capable of associating a gustative stimulus with CsA, resulting in behaviorally conditioned immunosuppression without aVecting appetitive behavior. © 2006 Elsevier Inc. All rights reserved. Keywords: Conditioning; Cyclosporine A; Lithium chloride; Conditioned taste avoidance; Co13nditioned immunosuppression; BALB/c mice
1. Introduction Classical conditioning can be understood as learning about the temporal or causal relationship between external and internal stimuli to allow for the appropriate preparatory set of responses before biologically signiWcant events occur. In this regard, the capacity to associate a certain immune response or status (e.g., allergens, toxins, and anaphylaxis) with a speciWc exteroceptive stimulus (e.g., environments or food Xavors) seems to be of high adaptive value. We consider that this was most likely acquired *
1
Corresponding author. Fax: +41 44 6321355. E-mail address: [email protected] (G. Pacheco-López). These authors contributed equally to this work.
0889-1591/$ - see front matter © 2006 Elsevier Inc. All rights reserved. doi:10.1016/j.bbi.2006.02.004
during evolution as an adaptive strategy in order to protect the organism and/or prepare it for danger. Reduced ingestive behavior may be only part of a complex and diverse repertory of physiological responses that the individual evokes to avoid, reject and/or prepare the organism to counteract, the unconditioned stimulus (US) eVects. It has been reported that mice, as many other mammals, display associative learning (Fanselow and Poulos, 2005) and are capable of developing taste-visceral association, reducing ingestive behavior as a conditioned response (Welzl et al., 2001). In addition, taste-immune associative learning has been reported in mice, aVecting several immune parameters at recall time (Blom et al., 1995; Cohen et al., 1979; Grota et al., 1987; Roudebush and Bryant, 1991).
528
M.-B. Niemi et al. / Brain, Behavior, and Immunity 20 (2006) 527–531
We have established a conditioning paradigm in rats employing the saccharin taste as the conditioned stimulus (CS), paired with immunosuppressant drug cyclosporine A (CsA) as the US. Among sweet tastes, saccharin (SAC) has often been selected as a gustative CS due to its low-caloric properties and innate positive hedonic value (Yamamoto, 1984). Regarding the US, CsA is widely used within immunosuppressive therapies in clinical settings (Graeb et al., 2004). The immunopharmacological mechanism of CsA involves its binding to cyclophilins, which leads to intracellular phosphatase calcineurin inhibition, then selectively reducing the expression of certain cytokines (e.g., interleukin-2: IL-2 and interferon-: IFN-), which in the end results in speciWc T-cell inhibition (Bukrinsky, 2002). Importantly, it has been shown that after taste/CsA association takes place, the mere exposure to the taste (CS) induces an immunosuppressive status, which has been consistently demonstrated in various studies (reviewed in: Exton et al., 2001). Experimental evidence has revealed important mechanisms indicating that such immunosuppressive status is dependent of sympathetic innervations to the spleen, whereas it is not related to hypothalamus–pituitary–adrenal axis activation (Exton et al., 1998). Additionally, residual immunopharmacologic eVects of CsA are not behind this suppressed responsiveness of the immune system, since CsA blood levels at recall phase are undetectable. Supporting this last evidence, it has been documented that CsA in vivo eVects (e.g., calcineurin activity inhibition) correlates with CsA levels in blood and tissue (Halloran et al., 1999). After a single CsA (25 mg/kg) i.p. administration in mice, Halloran et al. reported a peak of calcineurin activity inhibition 1 h post-administration, returning to basal levels after 24 h. In addition to the immunosuppressive status, rats develop avoidance behavior to the taste as a further conditioned response. Importantly, comparable immunological and behaviorally conditioned eVects have been documented in humans (Goebel et al., 2002). It is possible that species-speciWc associations may have evolved depending on speciWc environments; therefore the inter-species generalization of this particular behavioral conditioning model cannot be taken for granted, and thus needs to be explored. Moreover, due to the complexity of neuro-immune interactions it is obviously of great value to
extend the research to mice, in order to proWt from the advanced possibilities compared with rats in immune/genomic approaches. Thus the present study was designed to generate a murine model of behavioral conditioning, pairing saccharin taste as CS with CsA as US. 2. Methods 2.1. Animals Male BALB/c mice (Harlan Laboratories, Borchen, Germany), weighing 25 g at the start of the experiments, were housed individually on an inverted 12 h light/dark schedule (lights oV at 7:00 am) with food available ad libitum. Water was available ad libitum, except during the water deprivation regimen. All experimental protocols were performed according to the guidelines of the Institutional Animal Care and the Ethics Committees of the Medical Faculty of the University of Duisburg-Essen, Germany.
2.2. Drugs CsA (LC Laboratories, Woburn, MA) was dissolved in a solution of 4.2% v/v absolute ethanol (Sigma, Munich, Germany) in miglyol 812 (Caesar & Loretz GmbH, Hilden, Germany). Stock solution of CsA (100 mg/ ml) was freshly emulsiWed in sterile PBS (pH 7.2) to reach a dose of 20 mg/ kg, i.p. injected in a volume of 250 l. LiCl (Sigma, Munich, Germany) was freshly dissolved in sterile saline (NaCl 0.9 %) to reach a dose of 63.6 mg/ kg, i.p. injected in a volume of 250 l.
2.3. Behavioral conditioning Animals were randomly divided into four groups (CsA conditioning, Experiment 1) and two groups (LiCl conditioning, Experiment 2) and submitted to a 10-day regimen of progressive water deprivation, consisting of a daily (8:00 am) drinking session with 2 h water access for 4 days, 1 h access for 3 days, and 30 min Xuid access for the rest of the experiment. Injections were administered within a maximum of 10 min of Xuid access. Experiment 1: conditioned (CsA-CS) animals received saccharin solution 0.2% (SAC) instead of water as a conditioned stimulus (CS), paired with an i.p. injection of 20 mg/kg CsA as an unconditioned stimulus (US). During a single recall trial (day 20), animals were exposed to the CS (taste + vehicle injection). CsA-non-contingent-conditioned animals (CsA-NC) received the same stimuli as CsA-CS animals, but during association trials these mice were exposed to the CS 24 h before US, with the aim of avoiding associative learning. The group conditioned and further exposed to CsA (CsA-US) received the same stimuli as CsA-CS animals during association trials, but at recall were exposed to water and additionally injected with CsA to control for the pharmacological eVect of the drug. The placebo group received SAC and vehicle injections (Table 1).
Table 1 Conditioned taste aversion, resulting from saccharin-taste conditioning at association time with two diVerent US (CsA or LiCl i.p.)
CsA-CS CsA-NCa CsA-US Placebo LiCl-CS LiCl-CSo
n
Association p.o./i.p.
Recall p.o./i.p.
CTA at day 20 of % water baseline
5 5 5 5 4 4
SAC/CsA SAC(¡24 h)Wat/CsA SAC/CsA SAC/PBS SAC/LiCl SAC/LiCl
SAC/PBS SAC/PBS Wat/CsA SAC/PBS SAC/Sal Wat
106.1 § 4.5 113.8 § 13.5 105.7 § 6.0 103.7 § 5.5 ¤¤ 10.8 § 4.2 111.8 § 9.5
CsA-CS, CsA-conditioned; CsA-NC, CsA-non-contingent-conditioned; CsA-US, CsA conditioned and re-exposed to CsA; LiCl-CS, LiCl-conditioned; LiCl-CSo, LiCl-conditioned not evoked; SAC, 0.2% saccharin solution; CsA, 20 mg/kg cyclosporin; LiCl, 63.6 mg/kg; PBS, phosphate buVered saline; Sal, 0.9% NaCl solution; CTA, conditioned taste avoidance. Results are expressed as means § SEM. a In each association trial, mice were exposed to SAC p.o. 24 h before CsA i.p. administration. ¤¤ p 6 .01.
M.-B. Niemi et al. / Brain, Behavior, and Immunity 20 (2006) 527–531 Experiment 2: during association trials (days 11, 14, and 17), LiCl-conditioned animals (LiCl-CS) received saccharin solution 0.2% (SAC) instead of water as CS, paired with an i.p. injection of 63.6 mg/kg LiCl as US. During a single recall trial (day 20), animals were exposed to the CS (taste + vehicle injection). LiCl-conditioned not-evoked (LiCl-CSo) animals received identical treatment to the LiCl-CS group, except that at recall they were exposed to water only (Table 1). One hour post-recall, all animals (Experiments 1 and 2) were sacriWced by decapitation, peripheral blood was collected for corticosterone determination (Experiment 1) and spleens were removed for analysis of lymphocyte proliferation (Experiments 1 and 2) and cytokine production (Experiment 1).
2.4. Splenic lymphocyte proliferation Spleen lymphocytes were released from tissue by injecting cell culture medium (Ready-mix RPMI, PAA laboratories GmbH, Austria) into the spleen. Cells were washed several times in PBS (pH 7.2) and adjusted to a concentration of 1 £ 106 cells/ml. The cells were then cultured for 72 h with or without mitogenic stimulation (1.25, 2.5, and 5 g/ml concanavalin A; Sigma, Munich Germany) at 37 °C in a 5% CO2 atmosphere. After 48 h of incubation, cells were pulsed with 20 l/well [3H] thymidine (0.5 Ci) and harvested 24 h later. Radioactivity was measured using a scintillator-beta counter.
2.5. Cytokine determination
529
and borderline signiWcance for 2.5 g/ml of the same mitogen (df D 3, F D 3.119 p D .0601, data not shown). Fisher’s post hoc test indicated that CsA-conditioned animals (group CS) displayed a signiWcantly (p 6 .05) lower proliferative capacity in response to in vitro mitogenic stimulation (concanavalin A 1.25 g/ml) than CsA-non-contingent conditioned animals (group NC) (Fig. 1). Furthermore, such conditioned eVects on the immune system were comparable in magnitude to the pharmacological eVects of CsA (group US). In addition, ANOVA of IL-2 cytokine concentration in cell culture supernatant conWrmed the immunosuppressive status (df D 3, F D 5.230, p D .0105). Fisher’s post hoc test indicated that IL- 2 concentration in CsAconditioned animals was signiWcantly (p 6 .05) lower than in CsA-non-contingent-conditioned animals (Fig. 1). IFN- concentration followed a pattern similar to that of IL-2; however, diVerences did not reach statistical signiWcance (data not shown). Importantly, plasma corticosterone levels remained low in the CsA-CS group (5.4 § 0.9 pg/dl) compared with CsA-NC group (6.6 § 1.1 pg/dl). Interestingly, CsA-conditioned animals did not display signs of conditioned taste avoidance to the saccharin
Spleen lymphocytes were incubated 24 h with mitogenic stimulation (1.25 g/ml concanavalin A; Sigma, Munich Germany) at 37 °C in a 5% CO2 atmosphere. Supernatants from the cell cultures were collected and stored until determination. IL-2 and IFN- were assayed using commercial ELISA kits for the detection of IL-2 (Biosource, Camarillo, CA, USA) and IFN- (U-CyTech, Utrecht, The Netherlands).
2.6. Corticosterone determination Corticosterone levels in plasma were determined by radioimmunoassay as previously described (Pacheco-López et al., 2003).
2.7. Taste avoidance behavior assessment Taste avoidance behavior was assessed in a forced choice test using a single bottle. Fluid consumption was monitored throughout the entire experiment by weighing drinking bottles before and after each drinking session. Consumption was calculated as the diVerence in bottle weight pre/ post each session. Water consumption from days 8, 9, and 10 served as consumption baseline (100% water baseline). Deviations for each animal were calculated as a percentage of 100% water baseline.
2.8. Statistical analyses One-way analysis of variance (ANOVA) and Fisher’s post hoc test were used to examine statistical diVerences between the groups, or unpaired t test analysis to investigate speciWc statistical diVerences between two groups. Data are expressed as means § SEM. Results were considered as signiWcantly diVerent when p values were 6.05.
3. Results One hour after recall of the CsA-conditioning experiment, the immune status was evaluated by in vitro T-lymphocyte proliferation responsiveness, as well as by IL-2 and IFN- production. ANOVA indicated a signiWcant treatment eVect for the mitogenic stimulation induced by 1.25 g/ml of concanavalin A (df D 3, F D 3.9, p D .0304),
Fig. 1. Conditioned eVects observed in the immune system, induced by pairing SAC (0.2%) with CsA i.p. (20 mg/kg). (A) The conditioned eVects on the proliferative capacity of splenic lymphocytes. Production of IL-2 is presented in (B). CS, CsA-conditioned animals; NC, CsA-non-contingentconditioned; US, CsA-associated and re-exposed to CsA; P, saccharin exposure eVect. Results are expressed as means § SEM. ¤p 6 .05, (n D 5 animals per group).
530
M.-B. Niemi et al. / Brain, Behavior, and Immunity 20 (2006) 527–531
during the 2nd or 3rd association trial (data not shown), nor in the recall trial (Table 1). To clarify whether this behavior was due to gustatory/olfactory variation in mice (Kotlus and Blizard, 1998), a second experiment was performed applying an identical conditioning protocol in which LiCl was used as the US. Here, taste avoidance behavior and lymphocyte proliferation were assessed. The LiCl-conditioned group developed pronounced and signiWcant conditioned taste avoidance during the 2nd and 3rd association trials compared to the CsA-CS group (data not shown). At the recall trial, the LiCl-CS group exhibited signiWcant conditioned avoidance behavior in comparison with the CsA-CS and LiCl-CSo groups (df D 7, t D ¡9.297, p 6 .01). Additionally, LiCl-CS animals (22120 § 1285 cpm) displayed signiWcantly reduced splenocyte proliferative capacity to the in vitro mitogenic stimulation induced by 1.25 mg/ml of concanavalin A (df D 6, t D ¡2.579, p 6 .05) compared to the control group (LiCl-Cso; 35970 § 5213 cpm). 4. Discussion The present results show murine capacity to develop taste-immune associative learning. When the saccharin taste precedes the administration of the immunosuppressive drug CsA, mice (n D 5) associate such stimuli, and are able to induce an immunosuppressive status on demand that resembles the CsA-unconditioned eVects. Importantly, these eVects, attributable to behavioral conditioning, do not seem to be related to hypothalamus–pituitary–adrenal axis activation and its subsequent eVects on the immune system. Despite the limited number of animals used in the present experiments, our results indicate that mice, like rats and humans, seem to be able to develop a conditioned immunosuppressive response after recalling taste-CsA association; however, diVerent behaviorally conditioned responses result from such associative learning. In the case of mice, this speciWc taste-immune association does not result in conditioned taste avoidance behavior, whereas rats display reduced appetitive behavior (Exton et al., 2001) and for humans the palatability of the conditioned taste is aVected (Goebel et al., 2002). These variations could be related to the diVerences in the conditioning paradigm employed for each species or to the diVerent mechanisms underlying each of these behavioral responses. In this regard, growing evidence indicates that there are two diVerent neural processes behind appetitive and consummatory feeding behaviors (Parker, 2003). An additional alternative hypothesis is that, depending on the speciWc environmental challenges, species-speciWc associative patterns developed during evolution. Based on conditioning theory and empirical evidence from numerous studies with other drugs that impact various physiological systems, the development of multiple independent conditioned responses could be expected (Eikelboom and Stewart, 1982). In this regard, the present results support the hypothesis that conditioned feeding behaviors per se are not the direct source of the conditioned
eVects on immune function (Bovbjerg, 2003). Furthermore, several studies in rats indicate that there are diVerent neural substrates behind each of these conditioned responses (Chen et al., 2004; Pacheco-López et al., 2005; RamírezAmaya et al., 1996). The absence of conditioned avoidance behavior in CsAconditioned mice might be related to deWcits in CS and/or US stimulation intensity. Thus, it could be possible that the applied CsA dose did not reach the threshold which is necessary to produce avoidance behavior, as has already been shown with other immune-competent unconditioned stimuli. For instance, taste-antigen association results at recall time in enhanced antibody production without inducing conditioned avoidance to the taste in mice (Ader et al., 1993) and rats (Alvarez-Borda et al., 1995). However, animals sensitized to certain antigen, might have stronger immune-signaling to the brain after the second presentation of the same antigen, and when such secondary immune response is associated with a gustative stimulus, then strong avoidance behavior to that taste is displayed in the recall phase (Djuric et al., 1988). Interestingly, it has been reported in rats that avoidance behavior is displayed as a result of taste-immune association, whereas aversive behavior is not aVected (Cross-Mellor et al., 2004), indicating that appetitive behavior might be conditioned independently of consummatory feeding behavior. Therefore, it would be interesting to establish whether CsA can induce conditioned consummatory behavior in mice. Regarding the CS, taste sensory variation in mice has already been documented (Kotlus and Blizard, 1998), therefore another experiment was conducted to conWrm that BALB/c mice could detect, and thus associate, the gustative stimulus employed. Here, animals (n D 4/group) were submitted to a conditioning protocol in which the same taste was paired with gastric malaise induced by i.p. injection of LiCl. At recall time, LiCl-conditioned animals displayed signiWcant avoidance to the taste, indicating that the CS intensity employed in both experiments was suYcient to be sensed and associated by this mouse strain. Additionally, splenic T-lymphocyte responsiveness was signiWcantly inhibited in LiCl-conditioned animals. These results conWrm previous reports in rats (Gauci et al., 1992) and mice (Kelley et al., 1985), indicating that taste-LiCl association results in an immunosuppressive status at recall time. Although the preponderant role of corticosterone at recall time has been delineated with regard to LiCl-conditioned immunosuppressive eVects in mice (Kelley et al., 1985), it should be taken into account that lithium itself aVects T-cell mitogenic activity and IL-2 production in vitro (Wilson et al., 1989), and these LiClunconditioned eVects on the immune system could have been conditioned at association time. In any case, it is necessary to investigate further whether similar conditioned immunomodulatory principles rule for taste-CsA and taste-LiCl associations. Several reports have documented that the conditioned eVects on immune functions are independent of the
M.-B. Niemi et al. / Brain, Behavior, and Immunity 20 (2006) 527–531
elevation in corticosterone levels (Ader and Cohen, 1993). The present data indicate that recalling the taste-CsA association does not seem to activate the hypothalamic–pituitary–adrenal axis in mice, which is similar to what occurs in rats (Exton et al., 2001). However, the kinetics of the conditioned response on the immune, endocrine, and nervous systems has not been systematically assessed, nor has the development of paradoxical conditioned responses. These important issues need to be addressed in order to understand the principles behind this special kind of associative learning, and to propose possible clinical applications of behaviorally conditioned immunosuppression. In summary, we have described a murine taste-immune associative learning process in which saccharin and CsA stimuli were associated. When such an association was recalled, an immunosuppressive status was observed in the spleen without aVecting appetitive behavior. The present data indicate the independence of conditioned responses aVecting behavior and immune functions in mice. Acknowledgment This work was supported by a grant from the German Research Foundation to M.S. (Sche 341/9-1 and Sche 341/9-2). References Ader, R., Cohen, N., 1993. Psychoneuroimmunology: conditioning and stress. Annu. Rev. Psychol. 44, 53–85. Ader, R., Kelly, K., Moynihan, J., Grota, L., Cohen, N., 1993. Conditioned enhancement of antibody production using antigen as the unconditioned stimulus. Brain Behav. Immun. 7, 334–343. Alvarez-Borda, B., Ramírez-Amaya, V., Pérez-Montfort, R., BermúdezRattoni, F., 1995. Enhancement of antibody production by a learning paradigm. Neurobiol. Learn. Mem. 64, 103–105. Blom, J., Tamarkin, L., Shiber, J., Nelson, R., 1995. Learned immunosuppression is associated with an increased risk of chemically-induced tumors. Neuroimmunomodulation 2, 92–99. Bovbjerg, D.H., 2003. Conditioning, cancer, and immune regulation. Brain Behav. Immun. 17 (Suppl. 1), S58–S61. Bukrinsky, M., 2002. Cyclophilins: unexpected messengers in intercellular communications. Trends Immunol. 23, 323–325. Chen, J., Lin, W., Wang, W., Shao, F., Yang, J., Wang, B., Kuang, F., Duan, X., Ju, G., 2004. Enhancement of antibody production and expression of c-Fos in the insular cortex in response to a conditioned stimulus after a single-trial learning paradigm. Behav. Brain Res. 154, 557–565. Cohen, N., Ader, R., Green, N., Bovbjerg, D., 1979. Conditioned suppression of a thymus-independent antibody response. Psychosom. Med. 41, 487–491. Cross-Mellor, S., Kavaliers, M., Ossenkopp, K., 2004. Comparing immune activation (lipopolysaccharide) and toxin (lithium chloride)-induced gustatory conditioning: lipopolysaccharide produces conditioned taste avoidance but not aversion. Behav. Brain Res. 148, 11–19.
531
Djuric, V., Markovic, B., Lazarevic, M., Jankovic, B., 1988. Anaphylactic shock-induced conditioned taste aversion. II. Correlation between taste aversion and indicators of anaphylactic shock. Brain Behav. Immun. 2, 24–31. Eikelboom, R., Stewart, J., 1982. Conditioning of drug-induced physiological responses. Psychol. Rev. 89, 507–528. Exton, M., Herklotz, J., Westermann, J., Schedlowski, M., 2001. Conditioning in the rat: an in vivo model to investigate the molecular mechanisms and clinical implications of brain-immune communication. Immunol. Rev. 184, 226–235. Exton, M., von Hörsten, S., Schult, M., Vöge, J., Strubel, T., Donath, S., Steinmüller, C., Seeliger, H., Nagel, E., Westermann, J., Schedlowski, M., 1998. Behaviorally conditioned immunosuppression using cyclosporine A: central nervous system reduces IL-2 production via splenic innervation. J. Neuroimmunol. 88, 182–191. Fanselow, M., Poulos, A., 2005. The neuroscience of mammalian associative learning. Annu. Rev. Psychol. 56, 207–234. Gauci, M., Bull, D., Schedlowski, M., Husband, A., King, M., 1992. Lithium chloride and immunomodulation in taste aversion conditioning. Physiol. Behav. 51, 207–210. Goebel, M., Trebst, A., Steiner, J., Xie, Y., Exton, M., Frede, S., Canbay, A., Michel, M., Heemann, U., Schedlowski, M., 2002. Behavioral conditioning of immunosuppression is possible in humans. FASEB J. 16, 1869–1873. Graeb, C., Arbogast, H., Guba, M., Jauch, K., Land, W., 2004. Cyclosporine: 20 years of experience at the University of Munich. Transplant. Proc. 36, 125S–129S. Grota, L., Ader, R., Cohen, N., 1987. Taste aversion learning in autoimmune Mrl-lpr/lpr and Mrl +/+ mice. Brain Behav. Immun. 1, 238–250. Halloran, PF., Helms, LM., Kung, L., Noujaim, J., 1999. The temporal proWle of calcineurin inhibition by cyclosporine in vivo. Transplant 68, 1356–1361. Kelley, K., Dantzer, R., Mormede, P., Salmon, H., Aynaud, J., 1985. Conditioned taste aversion suppresses induction of delayed-type hypersensitivity immune reactions. Physiol. Behav. 34, 189–193. Kotlus, B., Blizard, D., 1998. Measuring gustatory variation in mice: a short-term Xuid-intake test. Physiol. Behav. 64, 37–47. Pacheco-López, G., Niemi, M., Kou, W., Bildhäuser, A., Gross, C., Goebel, M., del Rey, A., Besedovsky, H., Schedlowski, M., 2003. Central catecholamine depletion inhibits peripheral lymphocyte responsiveness in spleen and blood. J. Neurochem. 86, 1024–1031. Pacheco-López, G., Niemi, M., Kou, W., Härting, M., Fandrey, J., Schedlowski, M., 2005. Neural substrates for behaviorally conditioned immunosuppression in the rat. J. Neurosci. 25, 2330–2337. Parker, L.A., 2003. Taste avoidance and taste aversion: evidence for two diVerent processes. Learn. Behav. 31, 165–172. Ramírez-Amaya, V., Alvarez-Borda, B., Ormsby, C., Martínez, R., PérezMontfort, R., Bermúdez-Rattoni, F., 1996. Insular cortex lesions impair the acquisition of conditioned immunosuppression. Brain Behav. Immun. 10, 103–114. Roudebush, R., Bryant, H., 1991. Conditioned immunosuppression of a murine delayed type hypersensitivity response: dissociation from corticosterone elevation. Brain Behav. Immun. 5, 308–317. Welzl, H., D’Adamo, P., Lipp, H., 2001. Conditioned taste aversion as a learning and memory paradigm. Behav. Brain Res. 125, 205–213. Wilson, R., Fraser, W., McKillop, J., Smith, J., O’Reilly, D., Thomson, J., 1989. The “in vitro” eVects of lithium on the immune system. Autoimmunity 4, 109–114. Yamamoto, T., 1984. Taste responses of cortical neurons. Prog. Neurobiol. 23, 273–315.
Short communication
Murine taste-immune associative learning Maj-Britt Niemi a,1, Gustavo Pacheco-López a,¤,1, Wei Kou b, Margarete Härting b, Adriana del Rey c, Hugo O. Besedovsky c, Manfred Schedlowski a a
Chair of Psychology and Behavioral Immunobiology, Institute for Behavioral Sciences, ETH Zurich, Zurich 8092, Switzerland b Department of Medical Psychology, University of Duisburg-Essen, Essen 45122, Germany c Division of Immunophysiology, Institute of Physiology, Philipps University of Marburg, Marburg 35037, Germany Received 26 September 2005; received in revised form 24 February 2006; accepted 24 February 2006 Available online 21 April 2006
Abstract Taste-immune associative learning can result from contingent pairings of an immune-competent unconditioned stimulus (US) with a gustative conditioned stimulus (CS). Recalling such an association may induce a set of physiological responses aVecting behavior, endocrine, and immune functions. We have established a model of behaviorally conditioned immunosuppression employing the immunosuppressant drug cyclosporine A (CsA) as the US and saccharin as the CS in rats and humans. In order to investigate the inter-species generalization of this neuro-immune interaction, we tested the feasibility of this paradigm in mice. In a single-bottle scheme, male BALB/ c mice (n D 5) were conditioned by conducting three association trials and a single recall trial. Control groups (n D 5/group) were designed to assure associative learning, pharmacological eVects of the US, and placebo eVect. Results show that CsA-conditioned animals displayed signiWcant immunosuppression in the spleen after recall, measured by in vitro T-lymphocyte proliferation, and IL-2 production. However, the same animals did not show evidence of avoidance behavior to the CS. In contrast, evoking the association of saccharin–lithium chloride (inducing gastric malaise) in another set of animals (n D 4/group) resulted in signiWcant and pronounced avoidance of the taste (CS). These animals also displayed signiWcant suppression of splenic T-lymphocyte responsiveness after the recall phase. The present results indicate that mice seem to be capable of associating a gustative stimulus with CsA, resulting in behaviorally conditioned immunosuppression without aVecting appetitive behavior. © 2006 Elsevier Inc. All rights reserved. Keywords: Conditioning; Cyclosporine A; Lithium chloride; Conditioned taste avoidance; Co13nditioned immunosuppression; BALB/c mice
1. Introduction Classical conditioning can be understood as learning about the temporal or causal relationship between external and internal stimuli to allow for the appropriate preparatory set of responses before biologically signiWcant events occur. In this regard, the capacity to associate a certain immune response or status (e.g., allergens, toxins, and anaphylaxis) with a speciWc exteroceptive stimulus (e.g., environments or food Xavors) seems to be of high adaptive value. We consider that this was most likely acquired *
1
Corresponding author. Fax: +41 44 6321355. E-mail address: [email protected] (G. Pacheco-López). These authors contributed equally to this work.
0889-1591/$ - see front matter © 2006 Elsevier Inc. All rights reserved. doi:10.1016/j.bbi.2006.02.004
during evolution as an adaptive strategy in order to protect the organism and/or prepare it for danger. Reduced ingestive behavior may be only part of a complex and diverse repertory of physiological responses that the individual evokes to avoid, reject and/or prepare the organism to counteract, the unconditioned stimulus (US) eVects. It has been reported that mice, as many other mammals, display associative learning (Fanselow and Poulos, 2005) and are capable of developing taste-visceral association, reducing ingestive behavior as a conditioned response (Welzl et al., 2001). In addition, taste-immune associative learning has been reported in mice, aVecting several immune parameters at recall time (Blom et al., 1995; Cohen et al., 1979; Grota et al., 1987; Roudebush and Bryant, 1991).
528
M.-B. Niemi et al. / Brain, Behavior, and Immunity 20 (2006) 527–531
We have established a conditioning paradigm in rats employing the saccharin taste as the conditioned stimulus (CS), paired with immunosuppressant drug cyclosporine A (CsA) as the US. Among sweet tastes, saccharin (SAC) has often been selected as a gustative CS due to its low-caloric properties and innate positive hedonic value (Yamamoto, 1984). Regarding the US, CsA is widely used within immunosuppressive therapies in clinical settings (Graeb et al., 2004). The immunopharmacological mechanism of CsA involves its binding to cyclophilins, which leads to intracellular phosphatase calcineurin inhibition, then selectively reducing the expression of certain cytokines (e.g., interleukin-2: IL-2 and interferon-: IFN-), which in the end results in speciWc T-cell inhibition (Bukrinsky, 2002). Importantly, it has been shown that after taste/CsA association takes place, the mere exposure to the taste (CS) induces an immunosuppressive status, which has been consistently demonstrated in various studies (reviewed in: Exton et al., 2001). Experimental evidence has revealed important mechanisms indicating that such immunosuppressive status is dependent of sympathetic innervations to the spleen, whereas it is not related to hypothalamus–pituitary–adrenal axis activation (Exton et al., 1998). Additionally, residual immunopharmacologic eVects of CsA are not behind this suppressed responsiveness of the immune system, since CsA blood levels at recall phase are undetectable. Supporting this last evidence, it has been documented that CsA in vivo eVects (e.g., calcineurin activity inhibition) correlates with CsA levels in blood and tissue (Halloran et al., 1999). After a single CsA (25 mg/kg) i.p. administration in mice, Halloran et al. reported a peak of calcineurin activity inhibition 1 h post-administration, returning to basal levels after 24 h. In addition to the immunosuppressive status, rats develop avoidance behavior to the taste as a further conditioned response. Importantly, comparable immunological and behaviorally conditioned eVects have been documented in humans (Goebel et al., 2002). It is possible that species-speciWc associations may have evolved depending on speciWc environments; therefore the inter-species generalization of this particular behavioral conditioning model cannot be taken for granted, and thus needs to be explored. Moreover, due to the complexity of neuro-immune interactions it is obviously of great value to
extend the research to mice, in order to proWt from the advanced possibilities compared with rats in immune/genomic approaches. Thus the present study was designed to generate a murine model of behavioral conditioning, pairing saccharin taste as CS with CsA as US. 2. Methods 2.1. Animals Male BALB/c mice (Harlan Laboratories, Borchen, Germany), weighing 25 g at the start of the experiments, were housed individually on an inverted 12 h light/dark schedule (lights oV at 7:00 am) with food available ad libitum. Water was available ad libitum, except during the water deprivation regimen. All experimental protocols were performed according to the guidelines of the Institutional Animal Care and the Ethics Committees of the Medical Faculty of the University of Duisburg-Essen, Germany.
2.2. Drugs CsA (LC Laboratories, Woburn, MA) was dissolved in a solution of 4.2% v/v absolute ethanol (Sigma, Munich, Germany) in miglyol 812 (Caesar & Loretz GmbH, Hilden, Germany). Stock solution of CsA (100 mg/ ml) was freshly emulsiWed in sterile PBS (pH 7.2) to reach a dose of 20 mg/ kg, i.p. injected in a volume of 250 l. LiCl (Sigma, Munich, Germany) was freshly dissolved in sterile saline (NaCl 0.9 %) to reach a dose of 63.6 mg/ kg, i.p. injected in a volume of 250 l.
2.3. Behavioral conditioning Animals were randomly divided into four groups (CsA conditioning, Experiment 1) and two groups (LiCl conditioning, Experiment 2) and submitted to a 10-day regimen of progressive water deprivation, consisting of a daily (8:00 am) drinking session with 2 h water access for 4 days, 1 h access for 3 days, and 30 min Xuid access for the rest of the experiment. Injections were administered within a maximum of 10 min of Xuid access. Experiment 1: conditioned (CsA-CS) animals received saccharin solution 0.2% (SAC) instead of water as a conditioned stimulus (CS), paired with an i.p. injection of 20 mg/kg CsA as an unconditioned stimulus (US). During a single recall trial (day 20), animals were exposed to the CS (taste + vehicle injection). CsA-non-contingent-conditioned animals (CsA-NC) received the same stimuli as CsA-CS animals, but during association trials these mice were exposed to the CS 24 h before US, with the aim of avoiding associative learning. The group conditioned and further exposed to CsA (CsA-US) received the same stimuli as CsA-CS animals during association trials, but at recall were exposed to water and additionally injected with CsA to control for the pharmacological eVect of the drug. The placebo group received SAC and vehicle injections (Table 1).
Table 1 Conditioned taste aversion, resulting from saccharin-taste conditioning at association time with two diVerent US (CsA or LiCl i.p.)
CsA-CS CsA-NCa CsA-US Placebo LiCl-CS LiCl-CSo
n
Association p.o./i.p.
Recall p.o./i.p.
CTA at day 20 of % water baseline
5 5 5 5 4 4
SAC/CsA SAC(¡24 h)Wat/CsA SAC/CsA SAC/PBS SAC/LiCl SAC/LiCl
SAC/PBS SAC/PBS Wat/CsA SAC/PBS SAC/Sal Wat
106.1 § 4.5 113.8 § 13.5 105.7 § 6.0 103.7 § 5.5 ¤¤ 10.8 § 4.2 111.8 § 9.5
CsA-CS, CsA-conditioned; CsA-NC, CsA-non-contingent-conditioned; CsA-US, CsA conditioned and re-exposed to CsA; LiCl-CS, LiCl-conditioned; LiCl-CSo, LiCl-conditioned not evoked; SAC, 0.2% saccharin solution; CsA, 20 mg/kg cyclosporin; LiCl, 63.6 mg/kg; PBS, phosphate buVered saline; Sal, 0.9% NaCl solution; CTA, conditioned taste avoidance. Results are expressed as means § SEM. a In each association trial, mice were exposed to SAC p.o. 24 h before CsA i.p. administration. ¤¤ p 6 .01.
M.-B. Niemi et al. / Brain, Behavior, and Immunity 20 (2006) 527–531 Experiment 2: during association trials (days 11, 14, and 17), LiCl-conditioned animals (LiCl-CS) received saccharin solution 0.2% (SAC) instead of water as CS, paired with an i.p. injection of 63.6 mg/kg LiCl as US. During a single recall trial (day 20), animals were exposed to the CS (taste + vehicle injection). LiCl-conditioned not-evoked (LiCl-CSo) animals received identical treatment to the LiCl-CS group, except that at recall they were exposed to water only (Table 1). One hour post-recall, all animals (Experiments 1 and 2) were sacriWced by decapitation, peripheral blood was collected for corticosterone determination (Experiment 1) and spleens were removed for analysis of lymphocyte proliferation (Experiments 1 and 2) and cytokine production (Experiment 1).
2.4. Splenic lymphocyte proliferation Spleen lymphocytes were released from tissue by injecting cell culture medium (Ready-mix RPMI, PAA laboratories GmbH, Austria) into the spleen. Cells were washed several times in PBS (pH 7.2) and adjusted to a concentration of 1 £ 106 cells/ml. The cells were then cultured for 72 h with or without mitogenic stimulation (1.25, 2.5, and 5 g/ml concanavalin A; Sigma, Munich Germany) at 37 °C in a 5% CO2 atmosphere. After 48 h of incubation, cells were pulsed with 20 l/well [3H] thymidine (0.5 Ci) and harvested 24 h later. Radioactivity was measured using a scintillator-beta counter.
2.5. Cytokine determination
529
and borderline signiWcance for 2.5 g/ml of the same mitogen (df D 3, F D 3.119 p D .0601, data not shown). Fisher’s post hoc test indicated that CsA-conditioned animals (group CS) displayed a signiWcantly (p 6 .05) lower proliferative capacity in response to in vitro mitogenic stimulation (concanavalin A 1.25 g/ml) than CsA-non-contingent conditioned animals (group NC) (Fig. 1). Furthermore, such conditioned eVects on the immune system were comparable in magnitude to the pharmacological eVects of CsA (group US). In addition, ANOVA of IL-2 cytokine concentration in cell culture supernatant conWrmed the immunosuppressive status (df D 3, F D 5.230, p D .0105). Fisher’s post hoc test indicated that IL- 2 concentration in CsAconditioned animals was signiWcantly (p 6 .05) lower than in CsA-non-contingent-conditioned animals (Fig. 1). IFN- concentration followed a pattern similar to that of IL-2; however, diVerences did not reach statistical signiWcance (data not shown). Importantly, plasma corticosterone levels remained low in the CsA-CS group (5.4 § 0.9 pg/dl) compared with CsA-NC group (6.6 § 1.1 pg/dl). Interestingly, CsA-conditioned animals did not display signs of conditioned taste avoidance to the saccharin
Spleen lymphocytes were incubated 24 h with mitogenic stimulation (1.25 g/ml concanavalin A; Sigma, Munich Germany) at 37 °C in a 5% CO2 atmosphere. Supernatants from the cell cultures were collected and stored until determination. IL-2 and IFN- were assayed using commercial ELISA kits for the detection of IL-2 (Biosource, Camarillo, CA, USA) and IFN- (U-CyTech, Utrecht, The Netherlands).
2.6. Corticosterone determination Corticosterone levels in plasma were determined by radioimmunoassay as previously described (Pacheco-López et al., 2003).
2.7. Taste avoidance behavior assessment Taste avoidance behavior was assessed in a forced choice test using a single bottle. Fluid consumption was monitored throughout the entire experiment by weighing drinking bottles before and after each drinking session. Consumption was calculated as the diVerence in bottle weight pre/ post each session. Water consumption from days 8, 9, and 10 served as consumption baseline (100% water baseline). Deviations for each animal were calculated as a percentage of 100% water baseline.
2.8. Statistical analyses One-way analysis of variance (ANOVA) and Fisher’s post hoc test were used to examine statistical diVerences between the groups, or unpaired t test analysis to investigate speciWc statistical diVerences between two groups. Data are expressed as means § SEM. Results were considered as signiWcantly diVerent when p values were 6.05.
3. Results One hour after recall of the CsA-conditioning experiment, the immune status was evaluated by in vitro T-lymphocyte proliferation responsiveness, as well as by IL-2 and IFN- production. ANOVA indicated a signiWcant treatment eVect for the mitogenic stimulation induced by 1.25 g/ml of concanavalin A (df D 3, F D 3.9, p D .0304),
Fig. 1. Conditioned eVects observed in the immune system, induced by pairing SAC (0.2%) with CsA i.p. (20 mg/kg). (A) The conditioned eVects on the proliferative capacity of splenic lymphocytes. Production of IL-2 is presented in (B). CS, CsA-conditioned animals; NC, CsA-non-contingentconditioned; US, CsA-associated and re-exposed to CsA; P, saccharin exposure eVect. Results are expressed as means § SEM. ¤p 6 .05, (n D 5 animals per group).
530
M.-B. Niemi et al. / Brain, Behavior, and Immunity 20 (2006) 527–531
during the 2nd or 3rd association trial (data not shown), nor in the recall trial (Table 1). To clarify whether this behavior was due to gustatory/olfactory variation in mice (Kotlus and Blizard, 1998), a second experiment was performed applying an identical conditioning protocol in which LiCl was used as the US. Here, taste avoidance behavior and lymphocyte proliferation were assessed. The LiCl-conditioned group developed pronounced and signiWcant conditioned taste avoidance during the 2nd and 3rd association trials compared to the CsA-CS group (data not shown). At the recall trial, the LiCl-CS group exhibited signiWcant conditioned avoidance behavior in comparison with the CsA-CS and LiCl-CSo groups (df D 7, t D ¡9.297, p 6 .01). Additionally, LiCl-CS animals (22120 § 1285 cpm) displayed signiWcantly reduced splenocyte proliferative capacity to the in vitro mitogenic stimulation induced by 1.25 mg/ml of concanavalin A (df D 6, t D ¡2.579, p 6 .05) compared to the control group (LiCl-Cso; 35970 § 5213 cpm). 4. Discussion The present results show murine capacity to develop taste-immune associative learning. When the saccharin taste precedes the administration of the immunosuppressive drug CsA, mice (n D 5) associate such stimuli, and are able to induce an immunosuppressive status on demand that resembles the CsA-unconditioned eVects. Importantly, these eVects, attributable to behavioral conditioning, do not seem to be related to hypothalamus–pituitary–adrenal axis activation and its subsequent eVects on the immune system. Despite the limited number of animals used in the present experiments, our results indicate that mice, like rats and humans, seem to be able to develop a conditioned immunosuppressive response after recalling taste-CsA association; however, diVerent behaviorally conditioned responses result from such associative learning. In the case of mice, this speciWc taste-immune association does not result in conditioned taste avoidance behavior, whereas rats display reduced appetitive behavior (Exton et al., 2001) and for humans the palatability of the conditioned taste is aVected (Goebel et al., 2002). These variations could be related to the diVerences in the conditioning paradigm employed for each species or to the diVerent mechanisms underlying each of these behavioral responses. In this regard, growing evidence indicates that there are two diVerent neural processes behind appetitive and consummatory feeding behaviors (Parker, 2003). An additional alternative hypothesis is that, depending on the speciWc environmental challenges, species-speciWc associative patterns developed during evolution. Based on conditioning theory and empirical evidence from numerous studies with other drugs that impact various physiological systems, the development of multiple independent conditioned responses could be expected (Eikelboom and Stewart, 1982). In this regard, the present results support the hypothesis that conditioned feeding behaviors per se are not the direct source of the conditioned
eVects on immune function (Bovbjerg, 2003). Furthermore, several studies in rats indicate that there are diVerent neural substrates behind each of these conditioned responses (Chen et al., 2004; Pacheco-López et al., 2005; RamírezAmaya et al., 1996). The absence of conditioned avoidance behavior in CsAconditioned mice might be related to deWcits in CS and/or US stimulation intensity. Thus, it could be possible that the applied CsA dose did not reach the threshold which is necessary to produce avoidance behavior, as has already been shown with other immune-competent unconditioned stimuli. For instance, taste-antigen association results at recall time in enhanced antibody production without inducing conditioned avoidance to the taste in mice (Ader et al., 1993) and rats (Alvarez-Borda et al., 1995). However, animals sensitized to certain antigen, might have stronger immune-signaling to the brain after the second presentation of the same antigen, and when such secondary immune response is associated with a gustative stimulus, then strong avoidance behavior to that taste is displayed in the recall phase (Djuric et al., 1988). Interestingly, it has been reported in rats that avoidance behavior is displayed as a result of taste-immune association, whereas aversive behavior is not aVected (Cross-Mellor et al., 2004), indicating that appetitive behavior might be conditioned independently of consummatory feeding behavior. Therefore, it would be interesting to establish whether CsA can induce conditioned consummatory behavior in mice. Regarding the CS, taste sensory variation in mice has already been documented (Kotlus and Blizard, 1998), therefore another experiment was conducted to conWrm that BALB/c mice could detect, and thus associate, the gustative stimulus employed. Here, animals (n D 4/group) were submitted to a conditioning protocol in which the same taste was paired with gastric malaise induced by i.p. injection of LiCl. At recall time, LiCl-conditioned animals displayed signiWcant avoidance to the taste, indicating that the CS intensity employed in both experiments was suYcient to be sensed and associated by this mouse strain. Additionally, splenic T-lymphocyte responsiveness was signiWcantly inhibited in LiCl-conditioned animals. These results conWrm previous reports in rats (Gauci et al., 1992) and mice (Kelley et al., 1985), indicating that taste-LiCl association results in an immunosuppressive status at recall time. Although the preponderant role of corticosterone at recall time has been delineated with regard to LiCl-conditioned immunosuppressive eVects in mice (Kelley et al., 1985), it should be taken into account that lithium itself aVects T-cell mitogenic activity and IL-2 production in vitro (Wilson et al., 1989), and these LiClunconditioned eVects on the immune system could have been conditioned at association time. In any case, it is necessary to investigate further whether similar conditioned immunomodulatory principles rule for taste-CsA and taste-LiCl associations. Several reports have documented that the conditioned eVects on immune functions are independent of the
M.-B. Niemi et al. / Brain, Behavior, and Immunity 20 (2006) 527–531
elevation in corticosterone levels (Ader and Cohen, 1993). The present data indicate that recalling the taste-CsA association does not seem to activate the hypothalamic–pituitary–adrenal axis in mice, which is similar to what occurs in rats (Exton et al., 2001). However, the kinetics of the conditioned response on the immune, endocrine, and nervous systems has not been systematically assessed, nor has the development of paradoxical conditioned responses. These important issues need to be addressed in order to understand the principles behind this special kind of associative learning, and to propose possible clinical applications of behaviorally conditioned immunosuppression. In summary, we have described a murine taste-immune associative learning process in which saccharin and CsA stimuli were associated. When such an association was recalled, an immunosuppressive status was observed in the spleen without aVecting appetitive behavior. The present data indicate the independence of conditioned responses aVecting behavior and immune functions in mice. Acknowledgment This work was supported by a grant from the German Research Foundation to M.S. (Sche 341/9-1 and Sche 341/9-2). References Ader, R., Cohen, N., 1993. Psychoneuroimmunology: conditioning and stress. Annu. Rev. Psychol. 44, 53–85. Ader, R., Kelly, K., Moynihan, J., Grota, L., Cohen, N., 1993. Conditioned enhancement of antibody production using antigen as the unconditioned stimulus. Brain Behav. Immun. 7, 334–343. Alvarez-Borda, B., Ramírez-Amaya, V., Pérez-Montfort, R., BermúdezRattoni, F., 1995. Enhancement of antibody production by a learning paradigm. Neurobiol. Learn. Mem. 64, 103–105. Blom, J., Tamarkin, L., Shiber, J., Nelson, R., 1995. Learned immunosuppression is associated with an increased risk of chemically-induced tumors. Neuroimmunomodulation 2, 92–99. Bovbjerg, D.H., 2003. Conditioning, cancer, and immune regulation. Brain Behav. Immun. 17 (Suppl. 1), S58–S61. Bukrinsky, M., 2002. Cyclophilins: unexpected messengers in intercellular communications. Trends Immunol. 23, 323–325. Chen, J., Lin, W., Wang, W., Shao, F., Yang, J., Wang, B., Kuang, F., Duan, X., Ju, G., 2004. Enhancement of antibody production and expression of c-Fos in the insular cortex in response to a conditioned stimulus after a single-trial learning paradigm. Behav. Brain Res. 154, 557–565. Cohen, N., Ader, R., Green, N., Bovbjerg, D., 1979. Conditioned suppression of a thymus-independent antibody response. Psychosom. Med. 41, 487–491. Cross-Mellor, S., Kavaliers, M., Ossenkopp, K., 2004. Comparing immune activation (lipopolysaccharide) and toxin (lithium chloride)-induced gustatory conditioning: lipopolysaccharide produces conditioned taste avoidance but not aversion. Behav. Brain Res. 148, 11–19.
531
Djuric, V., Markovic, B., Lazarevic, M., Jankovic, B., 1988. Anaphylactic shock-induced conditioned taste aversion. II. Correlation between taste aversion and indicators of anaphylactic shock. Brain Behav. Immun. 2, 24–31. Eikelboom, R., Stewart, J., 1982. Conditioning of drug-induced physiological responses. Psychol. Rev. 89, 507–528. Exton, M., Herklotz, J., Westermann, J., Schedlowski, M., 2001. Conditioning in the rat: an in vivo model to investigate the molecular mechanisms and clinical implications of brain-immune communication. Immunol. Rev. 184, 226–235. Exton, M., von Hörsten, S., Schult, M., Vöge, J., Strubel, T., Donath, S., Steinmüller, C., Seeliger, H., Nagel, E., Westermann, J., Schedlowski, M., 1998. Behaviorally conditioned immunosuppression using cyclosporine A: central nervous system reduces IL-2 production via splenic innervation. J. Neuroimmunol. 88, 182–191. Fanselow, M., Poulos, A., 2005. The neuroscience of mammalian associative learning. Annu. Rev. Psychol. 56, 207–234. Gauci, M., Bull, D., Schedlowski, M., Husband, A., King, M., 1992. Lithium chloride and immunomodulation in taste aversion conditioning. Physiol. Behav. 51, 207–210. Goebel, M., Trebst, A., Steiner, J., Xie, Y., Exton, M., Frede, S., Canbay, A., Michel, M., Heemann, U., Schedlowski, M., 2002. Behavioral conditioning of immunosuppression is possible in humans. FASEB J. 16, 1869–1873. Graeb, C., Arbogast, H., Guba, M., Jauch, K., Land, W., 2004. Cyclosporine: 20 years of experience at the University of Munich. Transplant. Proc. 36, 125S–129S. Grota, L., Ader, R., Cohen, N., 1987. Taste aversion learning in autoimmune Mrl-lpr/lpr and Mrl +/+ mice. Brain Behav. Immun. 1, 238–250. Halloran, PF., Helms, LM., Kung, L., Noujaim, J., 1999. The temporal proWle of calcineurin inhibition by cyclosporine in vivo. Transplant 68, 1356–1361. Kelley, K., Dantzer, R., Mormede, P., Salmon, H., Aynaud, J., 1985. Conditioned taste aversion suppresses induction of delayed-type hypersensitivity immune reactions. Physiol. Behav. 34, 189–193. Kotlus, B., Blizard, D., 1998. Measuring gustatory variation in mice: a short-term Xuid-intake test. Physiol. Behav. 64, 37–47. Pacheco-López, G., Niemi, M., Kou, W., Bildhäuser, A., Gross, C., Goebel, M., del Rey, A., Besedovsky, H., Schedlowski, M., 2003. Central catecholamine depletion inhibits peripheral lymphocyte responsiveness in spleen and blood. J. Neurochem. 86, 1024–1031. Pacheco-López, G., Niemi, M., Kou, W., Härting, M., Fandrey, J., Schedlowski, M., 2005. Neural substrates for behaviorally conditioned immunosuppression in the rat. J. Neurosci. 25, 2330–2337. Parker, L.A., 2003. Taste avoidance and taste aversion: evidence for two diVerent processes. Learn. Behav. 31, 165–172. Ramírez-Amaya, V., Alvarez-Borda, B., Ormsby, C., Martínez, R., PérezMontfort, R., Bermúdez-Rattoni, F., 1996. Insular cortex lesions impair the acquisition of conditioned immunosuppression. Brain Behav. Immun. 10, 103–114. Roudebush, R., Bryant, H., 1991. Conditioned immunosuppression of a murine delayed type hypersensitivity response: dissociation from corticosterone elevation. Brain Behav. Immun. 5, 308–317. Welzl, H., D’Adamo, P., Lipp, H., 2001. Conditioned taste aversion as a learning and memory paradigm. Behav. Brain Res. 125, 205–213. Wilson, R., Fraser, W., McKillop, J., Smith, J., O’Reilly, D., Thomson, J., 1989. The “in vitro” eVects of lithium on the immune system. Autoimmunity 4, 109–114. Yamamoto, T., 1984. Taste responses of cortical neurons. Prog. Neurobiol. 23, 273–315.