Long-term cognitive enrichment affects opioid receptor expression in the amygdala of domestic pigs

Authors


Birger Puppe, Research Unit Behavioural Physiology, Research Institute for the Biology of Farm Animals (FBN), Wilhelm-Stahl-Allee 2, D-18196 Dummerstorf, Germany. E-mail:puppe@fbn-dummerstorf.de

Abstract

Enriching the housing environment by stimuli that challenge both reward and cognitive mechanisms may enhance behavioural experiences and can improve animal welfare, particularly in farm animals. A newly developed experimental feeding system for domestic pigs using food-rewarded learning of discriminatory and instrumental tasks enabled the animals to successfully master a cognitive challenge and to be rewarded ca. 30 times per day with small food portions. Reward-related behaviour is expected to be modulated by endogenous opioid systems. Furthermore, recent evidence supports a role for the amygdala in processing positive affects by stimulus-reward learning. Hence, the present study investigates mRNA expression of cerebral receptors, which are involved in these processes. In an initial step, reverse transcription–polymerase chain reaction (RT-PCR) provided the first evidence that transcripts of three different opioid receptors (MOR, DOR, KOR), as well as the neuropeptide Y 5 receptor (NPY5R), leptin receptor (LEPR) and proopiomelanocortin (POMC), are expressed in both the porcine amygdala and hypothalamus. Using real-time PCR we could show that the expression of two receptors of the opioid system (amygdala: KOR, DOR), in addition to the expression of NPY5R (hypothalamus) in eight enriched housed pigs was markedly downregulated compared to that of conventionally housed and fed pigs. Focusing on opioid receptors in the amygdala, the present study shows that long-term cognitive enrichment acts as a biologically relevant stimulus that causes modifications of gene expression of reward-sensitive cerebral receptors in domestic pigs.

Assessment of the welfare of animals living under various conditions of captivity (e.g. laboratory, farm or zoo) is of increasing ethical and scientific interest. Hence, various forms of enrichment have been developed to improve the biological relevance of environments in captivity (Newberry 1995). It has been suggested that the integration of technical facilities for rewarded operant learning into the housing environment is a promising approach for a sustained challenge of the farm animals' cognitive abilities (Manteuffel et al. 2009). Linking the appetitive behaviour with successful learning has the potential to improve animal welfare by activating reward evaluating mechanisms in the brain (Spruijt et al. 2001) and, consequently, by positive stimulations of emotional processes (Burgdorf & Panksepp 2006; Désiréet al. 2002; Paul et al. 2005). Opioids are thought to be neuroethological mediators in psychophysiological concepts of motivation and positive reinforcement (van Ree et al. 1999) as well as in the control of rewarding ingestive behaviour (Glass et al. 1999; Levine & Billington 2004). Behaviours in which reward plays an important role may be controlled or at least be modulated by endogenous opioid systems (van Ree et al. 2000). Traditionally, the amygdala is centrally involved in fear conditioning and emotionally aversive situations (Davis et al. 1994; LeDoux 1995). In view of evidence that the amygdala also contributes to attentional, cognitive and appraisal functions or the processing of rewards and emotional states (Baxter & Murray 2002; LeDoux 2008; Sander et al. 2003), we hypothesized that this brain region participates in the modulation of opioid-mediated reward behaviour.

Recently, our research group has developed an experimental feeding system for domestic pigs (Sus scrofa) using the food-rewarded learning of discriminatory and instrumental cognitive tasks (Ernst et al. 2005). Indeed, we found that the presented cognitive enrichment did not induce an elevated stress response (Ernst et al. 2006), but caused successful learning behaviour, contributed to the alleviation of behaviour problems (e.g. repetitive belly-nosing), and altered the pigs' emotional reactivity in behavioural tests towards a reduction of fear and excitement (Puppe et al. 2007). Focusing on the involvement of the endogenous opioid system of the amygdala, the question arises whether these welfare-enhancing behavioural effects are reflected at the molecular level by alterations of gene expression. Therefore, an initial task of the present study was to provide first evidence that transcripts of the three different opioid receptors (MOR, DOR, KOR) found in the mammalian brain (Mansour et al. 1994), as well as the neuropeptide Y 5 receptor (NPY5R), leptin receptor (LEPR) and proopiomelanocortin (POMC), are expressed in both the porcine amygdala and hypothalamus. The main aim, however, was to investigate the effects of a long-term food-rewarded cognitive enrichment on opioid receptor expression in the amygdala of domestic pigs. Moreover, we examined the impact of the food-rewarded operant system on the mRNA expression of POMC, NPY5R and LEPR in the hypothalamus because they are known to be involved in the central nervous control of food intake (Schwartz et al. 2000).

Materials and methods

Subjects, housing and general procedures

Sixteen castrated male pigs (German Landrace) from eight different litters (two siblings from each) were assigned to an experimental or control group (n = 8 individuals each) directly after weaning at the age of 4 weeks and housed in a 2.00 × 3.00 m pen to accustom them to their group and to the feed pellets, supplied by a dispenser ad libitum. After 3 weeks of post-weaning adaptation and social habituation, both groups were transferred to experimental pens (postnatal weeks 7–20), each measuring 3.00 × 4.25 m (partially slatted floor, 3.00 × 2.08 m concrete/3.00 × 2.17 m plastic slat). The pens were in the same room but visually separated by a wall.

The control group was fed twice daily by a conventional food dispenser (101 × 80 × 41 cm, Jyden-Dantec, Süderlügum, Germany, accessible to at least four pigs at the same time) as is common in commercial pig husbandry. The experimental group, however, learned to use a computer-controlled ‘call-feeding-station’ (CFS) which was recently developed by our research group to enrich the housing environment by repeated cognitive challenges (Ernst et al. 2005; Puppe et al. 2007). The food consisted of standard food-pellets and the food allowance was adjusted to common recommendations for growing pigs under commercial conditions (see Ernst et al. 2005). The daily amount was adapted to the technical options of the CFS and the respective learning performance of the pigs. Hence, the CFSs offered a weekly increasing amount of food from 0.66 kg/day/animal at the beginning (7-week-old pigs) to 3.25 kg/day/animal at the end of the experiment (20-week-old pigs). In order to achieve equal food supply, the control group was fed the amount of food that the experimental group had consumed in total the day before. Consequently, both pig groups showed the same weight development. In both groups, the animal to feeding place ratio was 2:1, while water was available ad libitum. All pigs were slaughtered at the age of 21 weeks in the experimental slaughterhouse of the Research Institute for the Biology of Farm Animals, Dummerstorf, Germany. The animals were randomly selected and killed by electronarcosis immediately followed by exsanguinations. Nevertheless, the slaughtering order had no effect on the expression of any gene investigated in the present study (P > 0.1).

Cognitive enrichment design and learning behaviour

The entire cognitive enrichment procedure lasted 14 weeks (postnatal weeks 7–20) and consisted of three different phases with changing learning challenges raised by the CFS (see Table 1). The system and the methodology applied have been described in detail by Ernst et al. (2005, 2006) and Puppe et al. (2007). Briefly, the installation for learning and feeding included four CFSs and a controlling computer. A CFS consisted of a wooden chamber (63 × 40 × 103 cm) with an entrance of adjustable width (depending on the size of the animals) and a stainless steel feeding trough in the back. In the feeding area, a plastic button (Ø 6 cm) was located above the trough on the right side and could be operated by the animal. During the phase of association (week 7), the experimental pigs were trained to enter and feed from all four CFSs by classical conditioning over 24 h/day. The pigs were electronically recognized using a commercial transponder (Texas Trading, Windach, Germany) and whenever an animal had voluntarily entered an arbitrary CFS an individual tone (basic triad compositions with loudness of 58 dB) was played by the CFS before a portion of food rewarded the pigs (2-s delay). There were no apparent preferences for particular CFSs.

Table 1.  Synopsis of the cognitive enrichment design (for details see text)
PhaseChallenges and learning proceduresDuration
1Individual tone—feeding association in all four ‘call-feeding-stations' (CFS) by classical conditioning1 week
2Individual tone discrimination and approach to the calling CFS by operant conditioning3 weeks
3Additional working-load by operant learning of button-pushing with fixed ratios10 weeks

In the following discrimination phase (week 8–10) the individual tone was used as a summons to enter the specific CFS that was calling. During an active phase of 12 h (between 08:00 and 20:00) the signal calls were replayed randomly with respect to CFS and time. A biphasic day was simulated, including a break of approximately 1.5 h without any call around midday. Taking into account the rising food requirements of the growing pigs the number of signal calls was increased from 24 to 31 calls per day, each reinforced by a portion of 40 g. The number of summonses for each animal combined with the amount of dispensed food ensured that 80% of correct responses to the learning task covered 100% of the required food for the respective age.

During the working phase (week 11–20) a button had to be additionally pushed by the pigs after entering the right CFS to maintain a high learning challenge. Moreover, the rate of button-pushing (fixed ratio) was increased when an experimental group had fulfilled the previous task with an average success of more than 90% over 7 days and no group member was more than three times below a 70% success rate. Consequently, the animals could work on a fixed ratio schedule of 1, 3 and 6. Similar to the discrimination phase the growing food requirements of the animals were satisfied by an increasing call rate from 21 to 31, reinforced, in addition, by an increasing portion of food from 70 to 105 g. Learning data were registered as per cent of correct responses (mean ± SEM), rewarded by a portion of feed (success rate). To fulfil the criteria of a correct response, the called pig had to enter the calling CFS within a time interval of 180 s (discrimination phase) and 60 s (working phase), respectively. When another, non-called pig entered this CFS, the pig was recognized but not rewarded (incorrect response) and the call for the right pig was repeated to maintain its chance for a correct response within the time budget. At the end of the experiment, the non-called pigs showed no reaction on the summons.

RNA isolation

The amygdala, including its sub-nuclei (LeDoux 2008), as well as the hypothalamic area surrounding the third ventricle were dissected from the brains of the pigs immediately after slaughter [all areas were determined using the stereotaxic atlas of the pig brain (Félix et al. 1999) and according to experiences from previous studies (Kanitz et al. 2004; Manteuffel et al. 2007)] and were stored in RNAlater (Qiagen, Hilden, Germany) at −70°C before further processing. RNA from amygdalar or hypothalamic tissues was isolated with RNeasy Lipid Tissue Mini Kit (Qiagen), as recommended by the supplier. RNA was quantified in a NanoDrop instrument (PEQLAB, Erlangen, Germany). Quality of RNA was monitored from randomly selected samples by denaturing agarose (1%) gel electrophoresis.

Reverse transcription

RT was carried out with 2 μg of total RNA preparation from amygdala or hypothalamus, 500 nM final concentration of antisense primer (RT primer, Table 2), and Moloney mouse leukemia virus reverse transcriptase (M-MLV RT RNase H Minus Point Mutant, Promega, Mannheim, Germany) in 25 μl of the incubation buffer provided by the supplier, supplemented with deoxy-NTPs (Roche, Mannheim, Germany) and RNasin (Promega), for 60 min at 42°C (previously described by Kalbe et al. 2007). The freshly synthesized cDNA samples were cleaned with the High Pure PCR Product Purification Kit (Roche) and eluted in 50 μl elution buffer.

Table 2.  Primers used for RT, PCR and real-time PCR
Gene (GenBank accession numbers)Primers5′→ 3′ sequenceSize (bp)TA (°C)TFA (°C)
  1. RT, reverse transcription; TA, annealing temperature; TFA, fluorescence acquisition temperature;

  2. *Primers according to Lin et al. (2000)

B2MForwardACGGAAAGCCAAATTACCTGAACTG2616082
(NM_213978)ReverseTCTGTGATGCCGGTTAGTGGTCT   
 RTTACGATTTCTGATTTATGTGCATAAAG   
POMCForwardTCCGAGAAGAGCCAGACG1266085
(X00135)ReverseGGCTTTGGGGTCGGCTTC   
 RTTAACTACAGGTAAGAGGCGGGATAT   
NPY5RForwardTGGTTTCATGGGGAATCTC3275279
(AF106083)ReverseACTGTCGCAATCAGGAAGTAG   
 RTGGAGAGCAAATGGCAAGG   
LEPRForward*TCGGAAGATATCAGTGTTGA3135480
(AF036908)Reverse*TTGGGATGCTGATCTGATAA   
 RTTTCATAAGTTCATCCAATCCT   
MORForwardTGTATCCCAACCTCCTCCACCA2355883
(NM_001001538)ReverseTGCCTCCCACACATTCTCAT   
 RTCTCAACGCACATACCACAGCAT   
DORForwardCCGGTCAAAGCCCTGGACT2096087
(U71149)ReverseAAGGCGAAGAGGAACACACAGA   
 RTAGCCCAGCACACTACGAAG   
KORForwardGGAGGAACCAAAGTCAGGGAAG2966085
(L49437)ReverseAGGATGAAAATGTGAATGGGGG   
 RTAGTCCCTGAAACACCGCTTGAA   

Polymerase chain reaction

Table 2 summarizes the information on the genes detected by reverse transcription–polymerase chain reactionRT-PCR in porcine amygdala and hypothalamus (GenBank accession numbers, primers, and so on). PCR was performed with 2.5 μl of purified cDNA samples from three pigs of the control group and Taq DNA polymerase (Qbiogene, Heidelberg, Germany) in 25 μl of the incubation buffer provided by the supplier, supplemented with deoxy-NTPs and primers (Table 2, 500 nM each). For amplification, the following cycling conditions were performed: pre-incubation at 94°C for 10 min, followed by 35 cycles of denaturation at 94°C for 30 s, annealing for 30 s (see Table 2 for annealing temperatures), extension at 72°C for 10 s and a final extension at 72°C for 5 min. The knowledge about the exon/intron organization of the porcine chromosomal genes used is currently lacking with the exception of the reference gene β-2-microglobulin (B2M). According to published human genomic organization, intron-spanning primers for the opioid receptor genes were designed to avoid amplification of residual genomic DNA. All primers were purchased by Sigma-Genosys (Steinheim, Germany).

After amplification, 10 μl of each PCR product were analysed by agarose gel electrophoresis (3%). The amplicons were cloned into the pGEM-T vector system from Promega, sequenced and used as external standards for real-time PCR. Sequencing was performed with the automated sequencing system ABI PRISM 310 genetic analyzer using the ABI PRISMBig Dye kit (both from PE Applied Biosystems, Weiterstadt, Germany).

Real-time PCR

Transcript expression of μ- (MOR), δ- (DOR) and κ- (KOR) opioid receptors as well as B2M was measured for cDNAs from amygdala, whereas, for cDNAs from the hypothalamus, POMC, NPY5R, LEPR and B2M expression was monitored using real-time PCR. Therefore, 0.6 μl of each purified cDNA sample was amplified in duplicate with the LightCycler-FastStart DNA MasterPLUS SYBR Green I kit (Roche) in 10 μl total reaction volume. Amplification and quantification of generated products were performed in a LightCycler instrument 2.0 (Roche) under the following cycling conditions: pre-incubation at 95°C for 10 min, followed by 40 cycles denaturation at 95°C for 15 s, annealing for 10 s, extension at 72°C for 10 s and single point fluorescence acquisition for 6 s in order to avoid quantification of primer artefacts (Kalbe et al. 2008). The annealing and the fluorescence acquisition temperatures specific for the investigated genes are shown in Table 2. The melting peaks of all samples were routinely determined by melting curve analysis in order to ascertain that only the expected products had been generated. Additionally, molecular sizes of PCR products were monitored by agarose gel electrophoresis analysis (not shown). It should be noted that MOR, DOR and KOR were referred to as mu, delta and kappa opioid receptors by previous opioid receptor binding studies in the pig brain (e.g. Loijens et al. 2002; Zanella et al. 1996).

The relative quantification was performed with the LightCycler software version 4.5 using the quantification module: Relative Quantification–Monocolour. Thereby, the relative expression ratio of the target gene is calculated based on the PCR efficiencies and the crossing point deviation of an unknown sample versus a control, and expressed in comparison to an endogenous reference gene as described by Pfaffl (2001). To calculate the PCR efficiency, routine dilutions of the gene-specific external standard of known concentrations covering five orders of magnitude (5 × 10−16 to 5 × 10−12 g DNA) were co-amplified during each run. Data for mRNA expression of genes investigated are presented as relative expression ratio normalized to B2M as endogenous reference gene, because the B2M expression was unaffected between control and experimental group in hypothalamic (U = 23.0, P = 0.382) and amygdalar (U = 24.0, P = 0.442) tissues.

Statistical analysis

Our study focused on individually perceived differences in a cognitive enrichment paradigm affecting individual gene expression. As in similar studies (Langbein et al. 2006) animals were separated from the group while acting at the learning device. Such an design should genuinely limit the risk of substantial dependencies between individuals (Iason & Elston 2002; Phillips 2002), and hence, the present study treated individuals as independent replicates. Statistical analyses were performed using GraphPad InStat (version 3.00 for Windows 95, GraphPad Software, San Diego, California, USA). A non-parametric Mann–Whitney test was used to calculate differences in transcript concentrations between experimental and control pigs. Differences were considered as significant if P < 0.05 and as tending to be significant if P < 0.1.

Results

Learning behaviour

After the association phase, during which the pigs of the experimental group were shaped for the CFSs, the animals started their active learning behaviour with the discrimination phase, in which they were called with their individual sound to an arbitrary CFS, followed by the working phase where they had to press a button with increasing fixed ratio in order to obtain the reward. During the discrimination phase, all pigs reached the given learning criterion of 80% after 3 days (85.7 ± 1.9%). According to our purpose, they maintained a stable average success rate of over 80% during the following days (discrimination phase: 89.4 ± 3.4%, working phase: 84.1 ± 1.4%). Detailed results of behaviour and some physiological and immunological traits of the animals have been described in detail elsewhere (Ernst et al. 2006; Puppe et al. 2007).

Detection of POMC and brain receptors in amygdala and hypothalamus

RNA samples from porcine amygdala and hypothalamus were screened for the expression of POMC, the opioid receptors KOR, DOR and MOR and the receptors NPY5R and LEPR with reverse transcription–polymerase chain reaction (RT-PCR) and the results using one randomly selected control pig are shown in Fig. 1. The agarose gel indicates the presence of the POMC specific transcript in both amygdala and hypothalamus. The transcript expression of all three opioid receptors could be detected in amygdalar and hypothalamic tissues. Similarly, NPY5R and LEPR mRNA expression was found in both brain regions investigated. These findings were reproduced with RNA isolated from three randomly selected control pigs. The amplicons were cloned, subsequently sequenced and used as external standards for real-time PCR.

Figure 1.

Evidence for POMC and selected brain receptor transcripts using RT-PCR. PCR fragments were separated on a 3% agarose gel stained with ethidium bromide. (a) Transcripts of κ-, δ- and μ-opioid receptors (KOR, DOR and MOR) could be detected in the porcine amygdala (lane 1) and hypothalamus (lane 2). (b) Transcripts of proopiomelanocortin (POMC), neuropeptide Y 5 receptor (NPY5R) and leptin receptor (LEPR) could be detected in the porcine amygdala (lane 1) and hypothalamus (lane 2). Negative controls (lane 3) were carried with the PCR reactions to exclude contaminations of the PCR reagents. DNA molecular weight marker was provided in the first lane.

Quantification of KOR, DOR and MOR transcripts in the porcine amygdala

The long-term cognitive enrichment paradigm affected the transcript expressions of KOR and DOR, but not MOR in the amygdala of domestic pigs (Fig. 2). Compared with the animals of the control group, the KOR mRNA expression in the amygdala of the experimental pigs was decreased (U = 11.0, P = 0.028). Moreover, DOR transcript levels in the amygdala of the experimental pigs tended to be decreased (U = 15.0, P = 0.083). In contrast, the experimental design did not show any effect on the transcript expression of the MOR gene in the amygdala (P > 0.1).

Figure 2.

Effect of cognitive enrichment on mRNA expression of KOR, DOR and MOR in the amygdala of domestic pigs. Real-time PCR analysis of κ-, δ- and μ-opioid receptor (KOR, DOR and MOR) transcripts was carried out with RNA isolated from amygdalae of pigs kept under cognitive enrichment (n = 8) or control conditions (n = 8). Data (means + SEM) are expressed in arbitrary units after normalization to β-2-microglobulin (B2M) mRNA expression. Significant differences are indicated by an asterisk (*P < 0.05); differences in tendency are indicated by a hash (#P < 0.1).

Quantification of POMC, NPY5R and LEPR transcripts in the porcine hypothalamus

In the hypothalamus, the transcript expression of POMC, NPY5R and LEPR genes was measured (Fig. 3). Compared with the animals of the control group, the mRNA expression of NPY5R in the food-rewarded experimental pigs was downregulated (U = 12.5, P = 0.046). In contrast, the application of the cognitive enrichment did not affect the gene expression of POMC and LEPR in the hypothalamus (P > 0.1).

Figure 3.

Effect of cognitive enrichment on mRNA expression of POMC, NPY5R and LEPR in the hypothalamus. Real-time PCR analysis of proopiomelanocortin (POMC), neuropeptide Y 5 receptor (NPY5R) and leptin receptor (LEPR) transcripts was carried out with RNA isolated from hypothalami of pigs kept under cognitive enrichment (n = 8) or control conditions (n = 8). Data (means + SEM) are expressed in arbitrary units after normalization to β-2-microglobulin (B2M) mRNA expression. Significant differences are indicated by an asterisk (*P < 0.05).

Discussion

There is a need to improve the welfare of domestic pigs by allowing them to display more of their natural behavioural repertoire, including cognitive abilities, and to experience various stimuli that can help to prevent boredom and behavioural disorders (Puppe et al. 2008). In the present study, we have found that a long-term cognitive enrichment integrated into the housing environment affected both opioid receptor mRNA expression in the amygdala and NPY5R mRNA expression in the hypothalamus. To our knowledge it is the first attempt to correlate a behavioural paradigm, which has been showed to improve welfare of pigs and probably leads to positive affect (Manteuffel et al. 2009), with underlying modulations in central gene expression.

The present study shows the mRNA expression of NPY5R to the best of our knowledge for the first time in porcine amygdala and hypothalamus. Our results are consistent with findings in the amygdala and hypothalamus of rats (e.g. Durkin et al. 2000; Wolak et al. 2003). Moreover, we confirmed the LEPR mRNA expression in porcine amygdala and hypothalamus (Lin et al. 2000) and the POMC expression (Kineman et al. 1989; Schwerin et al. 2005). Furthermore, the present study also represents the first report showing the mRNA expression of all three opioid receptors in the amygdala and hypothalamus of pigs. These results are in agreement with mRNA distribution of the opioid receptors in model animals such as rat or mouse (reviewed in Mansour et al. 1994; Minami & Satoh 1995). Previous studies revealed appearance of opioid receptors in selected porcine brain regions using RT-PCR or binding assays (Loijens et al. 2002; Pampusch et al. 1998; Zanella et al. 1996).

Opioid receptors are classified as G protein-coupled receptors (Uhl et al. 1994). The endogenous opioid ligands show preferences for different receptors: β-endorphin (generated from POMC) binding to MOR, enkephalins (generated from pro-enkephalin) binding to DOR and dynorphins (generated from pro-dynorphin) binding to KOR (Dhawan et al. 1996). Three time-dependent processes are involved in the response to these agonists, occurring over a time scale ranging from seconds to days: desensitization (seconds to hours), internalization (minutes to hours) and downregulation (hours to days) (cf. Hausdorff et al. 1990). A decrease of receptor expression (downregulation) is mainly caused by long-term exposure to agonists (Liggett et al. 1993) and may persist for many days, even after their removal (Henriksen & Willoch 2008). In fact, downregulation of expression after chronic agonist exposure has been reported for all types of opioid receptors (e.g. Ronnekleiv et al. 1996; Rosin et al. 1999; Trapaidze et al. 2000). Electrophysiological and functional imaging studies (Chieng et al. 2006; Liberzon et al. 2002; Morris et al. 1996) supported evidence for reduced amygdala activity in pleasant emotional states. More precisely, Koepp et al. (2009) showed that a positive shift in mood is associated with decreased cerebral diprenorphine (a non-selective ligand) binding, especially in the amygdala as a brain region with a high density of opioid receptors (Lewis et al. 1981). The reduction in ligand binding is consistent with an increased release of endogenous opioids, resulting in a reduced number of opioid receptors available for binding. It can be explained by either competition for the binding site between injected ligand and endogenous opioids, or agonist-driven receptor-internalization through release of endogenous opioids (Sternini et al. 2000), or a combination of both (Laruelle 2000). These facts are in agreement with the view that opioids reduce amygdalar activity, which is further corroborated by our results showing decreased mRNA expression of two opioid receptors in the amygdala (KOR and DOR). Therefore, suggestions about the reduction in amygdala neuronal activity during positive emotional states (Burgdorf & Panksepp 2006) are supported. Moreover, involvement in reward-related processes has been showed for all types of opioid receptors (van Ree et al. 2000). We have shown that KOR was the opioid receptor most affected by the experience of the cognitive enrichment procedure. Our results may be supported by findings of Smith et al. (2003) who have shown that the KOR system of enriched rats is sensitive to social and environmental manipulations affecting a variety of behaviourally important signals and the emotional experience. DOR tended to be downregulated in the amygdalae of the experimental pigs in our study. Several studies have showed that DOR is involved in reinforcement processes (e.g. Bals-Kubik et al. 1990; Shippenberg et al. 1987). In addition, Skoubis et al. (2005) revealed that the rewarding properties of the opioid system seem to be mediated by endogenous enkephalins, which were able to modulate emotion- and motivation-related behaviours by activating DOR (Mas Nieto et al. 2005). However, we have found no differences in the MOR mRNA expression, and, in agreement, we measured unchanged POMC mRNA expression in the hypothalamus. The latter needs further research because, for example, it is known that MOR seems to be mainly involved in different motivational aspects of addictive behaviour (cf. van Ree et al. 1999).

Cognitive enrichment integrated into the farm animal environment has the potential to be a suitable way of considerably increasing animal welfare by challenging a variety of behaviourally relevant stimuli (Manteuffel et al. 2009). Based on studies using functional magnetic resonance imaging (fMRI) it was published that the human amygdala responds more vigorously to stimuli that are behaviourally relevant compared to irrelevant ones (Ousdal et al. 2008). Similarly, the pigs in our experiment had to detect environmental acoustic stimuli that were relevant to the individual processing of food-rewarded behaviour. Our study suggests that KOR and DOR may be involved in the integration and reinforcement of this response. However, it was reported that brain opioid receptors are also involved in stereotypic behaviours of domestic pigs (Loijens et al. 2002; Zanella et al. 1996). Further work is needed to clarify the emotion-induced activity of opioid receptors in different brain regions under well-defined experimental conditions.

There is some evidence that the opioid system is affected by social behaviours. For instance, Vanderschuren et al. (1995) showed that the rewarding value of social play is mediated by MOR, whereas KOR is rather associated with the integration of environmental stimuli related to social play. Ploj and Nylander (2003) showed a long-term increase in amygdalar DOR density in adult life as a result of neonatal social separation stress, whereas van den Berg et al. (1999) reported upregulation of MOR and KOR in adult life as a result of a juvenile isolation of rats. Because our study allowed the group housing of the investigated subjects, possible interactions between social-related behaviours and the expression of brain receptors may have occurred. It is important to note that all procedures such as socialization, housing and slaughter were identical between pigs of the experimental and control group with the exception of the food-rewarded enrichment investigated. Previously, we have shown that this enrichment ensured largely stress-free and individual learning (Ernst et al. 2005; Puppe et al. 2007), suggesting an avoidance or at least minimization of social disturbances. Nevertheless, understanding possible relations between affective cognition, social behaviour and modulation of opioid receptors are of further research interest.

A feeding-stimulating effect of NPY has been reported in most vertebrates, including pigs (Parrot et al. 1986). The central administration of NPY increased the number of meals in rats (Ammar et al. 2000) and guinea pigs (Lecklin et al. 2003), reflecting generalized behavioural activation and reward produced by the initiation of a meal. It seems therefore, that our experimental pigs showed a feeding pattern similar to that of NPY-treated animals. This suggestion is supported by the decrease of NPY5R mRNA expression in the hypothalamus of experimental pigs in the present study. The NPY5R (described as ‘feeding receptor’, Gerald et al. 1996) and the NPY1R are both involved in the stimulation of food intake, suggesting that they might modify different phases of feeding behaviour (Lecklin et al. 2003). Additionally, our experiment showed no effect on the mRNA expression of the LEPR, obviously because the experimental and the control pigs showed no differences in weight, weight development or the amount of ingested food (Ernst et al. 2005; Puppe et al. 2007).

In conclusion, first evidence provided that transcripts of opioid receptors (MOR, DOR, KOR) and other receptors (e.g. NPY5R) were expressed in both the porcine amygdala and hypothalamus. Moreover, the present initial study shows that welfare-enhancing effects of a food-rewarded cognitive enrichment in domestic pigs can be supported at molecular level by corresponding investigations of central gene expression. The used cognitive enrichment may be viewed as a complex of behaviourally relevant stimuli challenging the pig's coping behaviour, affects their reward-dependent motivational system and induces changes in the central opioid receptor expression in the amygdala. To verify our results, it will be necessary to analyse the opioid receptors and their endogenous ligands using protein expression methods or binding assays, to measure additional indicators of affective states (e.g. dopamine, serotonin), and to include other brain regions of the reward system (e.g. nucleus accumbens). Apparently, feeding of pigs with repeated small portions of food during the day corresponds to their natural feeding behaviour and should be considered as an alternative to the conventional feeding in pig husbandry.

Acknowledgements

We appreciate the excellent technical assistance of Hilke Brandt. We also thank Margitta Hartung and Martina Pohlmann for tissue sampling. Special thanks go to Gerhard Manteuffel for reviewing the manuscript and helpful comments. Furthermore, the behavioural approach described in this study was supported by Deutsche Forschungsgemeinschaft (DFG).

Ancillary