Enkephalin knockout male mice are resistant to chronic mild stress

Authors


Abstract

Enhanced stress reactivity or sensitivity to chronic stress increases the susceptibility to mood pathologies such as major depression. The opioid peptide enkephalin is an important modulator of the stress response. Previous studies using preproenkephalin knockout (PENK KO) mice showed that these animals exhibit abnormal stress reactivity and show increased anxiety behavior in acute stress situations. However, the consequence of enkephalin deficiency in the reactivity to chronic stress conditions is not known. In this study, we therefore submitted wild-type (WT) and PENK KO male mice to chronic stress conditions, using the chronic mild stress (CMS) protocol. Subsequently, we studied the CMS effects on the behavioral and hormonal level and also performed gene expression analyses. In WT animals, CMS increased the expression of the enkephalin gene in the paraventricular nucleus (PVN) of the hypothalamus and elevated the corticosterone levels. In addition, WT mice exhibited enhanced anxiety in the zero-maze test and depression-related behaviors in the sucrose preference and forced swim tests. Surprisingly, in PENK KO mice, we did not detect anxiety and depression-related behavioral changes after the CMS procedure, and even measured a decreased hormonal stress response. These results indicate that PENK KO mice are resistant to the CMS effects, suggesting that enkephalin enhances the reactivity to chronic stress.

Stress occurs when internal or external stimuli are perceived as a threat to the organism's homoeostasis and survival. It elicits endocrine, autonomic and behavioral alterations that allow adaption to the new conditions (Strohle & Holsboer 2003; Tafet & Bernardini 2003). However, chronic and/or severe stress, together with a genetic predisposition, may lead to maladaptive physiological responses. Consequently, pathological conditions such as post-traumatic stress disorder or major depression (MD) can develop (Strohle & Holsboer 2003; Tafet & Bernardini 2003).

The paraventricular nucleus (PVN) of the hypothalamus is a central structure in the integration of stress response. The PVN receives inputs from several brain nuclei involved in metabolic and pain control. It also receives inputs from the amygdala, hippocampus and prefrontal cortex (PFC), which are involved in cognitive and emotional processing of stress stimuli (Herman et al. 2003; Tafet & Bernardini 2003). Several distinct functional subdivisions and neuronal populations are present in the PVN. Corticotrophin-releasing hormone (CRH)-expressing neurons, localized in the medial parvocellular subdivision of the PVN, together with the anterior pituitary and the adrenal glands constitute the hypothalamus–pituitary–adrenal (HPA) axis. Stress-induced CRH released by the parvocellular neurons stimulates the production and release of the adrenocorticotropic hormone (ACTH) by the pituitary. This in turn leads to the production and secretion of glucocorticoids by the adrenal cortex (Benarroch 2005; Herman et al. 2002).

Several studies showed the importance of enkephalin in stress responses and in the susceptibility to stress-related pathologies. Different stressors induce alterations in enkephalin mRNA expression and peptide levels (Bertrand et al. 1997; Christiansen et al. 2011; Dziedzicka-Wasylewska & Papp 1996; Lucas et al. 2007; Mansi et al. 2000; Yamada & Nabeshima 1995) and also modulate the number and function of its receptors, the mu opioid receptor (MOR) and delta opioid receptor (DOR) (Stein et al. 1992). Furthermore, systemic administration of exogenous analogs of enkephalin, inhibitors of enkephalin degradation and DOR-specific agonists leads to anxiolytic (Randall-Thompson et al. 2010) and anti-depressant effects (Baamonde et al. 1992; Broom et al. 2002; Tejedor-Real et al. 1995, 1998). In addition, mice deficient in enkephalin preproenkephalin knockout (PENK KO), DOR (DOR KO) or MOR (MOR KO) showed altered stress reactivity. DOR KO mice exhibited exacerbated stress responses (Filliol et al. 2000), whereas MOR KO mice exhibited a higher resilience to stress effects (Filliol et al. 2000; Ide et al. 2010; Komatsu et al. 2011). PENK KO mice exhibited increased anxiety levels under basal conditions (Bilkei-Gorzo et al. 2004b) (Bilkei-Gorzo et al. 2008a,2008b; Konig et al. 1996) and exacerbated anxiety and depression-related phenotypes induced by acute stress (Kung et al. 2010; Ragnauth et al. 2001).

In order to study the influence of enkephalin in chronic stress conditions, we submitted wild-type (WT) and PENK KO mice to a chronic mild stress (CMS) protocol. We analyzed the effects of CMS on brain enkephalin gene expression and compared the hormonal and behavioral responses between WT and PENK KO mice exposed to CMS.

Materials and methods

Animals

Male WT and PENK KO homozygous mice on a C57BL/6J genetic background were used in this study. Mice with a deletion of the preproenkephalin (Penk) gene (Konig et al. 1996) were crossed with C57BL/6J (WT) mice for more than 10 generations in order to obtain PENK KO homozygous mice on a pure C57BL/6J background (Bilkei-Gorzo et al. 2004b). To avoid genetic drift, we routinely renewed our inbred KO colonies by crossing back them with WT C57BL/6J animals after three generations of homozygous breeding and restarted the breeding of the KOs from the F2 generation of the back-crossing. WT mice were originally obtained from a commercial breeder (Janvier, St. Berthevin, France) and bred at our animal facility. At postnatal day 21, pups were weaned and housed in groups of 4–5 animals. Starting at the age of 6 weeks, WT and PENK KO mice were individually housed until the end of the experimental procedure. Rooms were maintained at 23°C, under a 12:12 h inverted light cycle with ad libitum access to water and food, except when animals were food or water restricted during CMS. All experiments were carried out using individually housed animals (except during the brief period of social stress submission, during which four animals were housed together) during the dark phase of the light cycle. To avoid possible maternal care effects, homozygous WT and PENK KO gestating females were housed together from the first week of gestation until the end of the lactation period. The females gave birth and reared the offspring from both genotypes together. Our breeding system fulfills the published requirements (Crusio et al. 2009) of the journal. All experiments followed the guidelines of the German Animal Protection Law.

Genotyping

DNA from tail biopsies was extracted by the hot shot lysis method (Truett et al. 2000). Amplification of DNA for genotyping was performed by adding a master mix solution [Green Taq (Promega, Madison, WI, USA), RNA-free water, 0.5 µl E31 Primer (Metabion, Martinsried, Germany) -GCATCCAGGTAATTGGCAGGAA-, 0.5 µl neoRL Primer (Metabion) -CAGCAGCCTCT GTTCCACATACACTTCAT- and 0.5 µl E1R Primer (Metabion) -TCCTTCACATTCCAGTGTGC-], to the DNA samples, followed by 40 cycles of amplification. The products were separated by electrophoresis. A band of 700 bp was amplified for WT and of 550 bp for PENK KO mice.

Experimental design and CMS procedure

We submitted WT and PENK KO mice to the CMS protocol, which is a validated animal model to study stress-related pathologies, such as MD (Hill et al. 2012; Willner 2005). The experiment was divided into two consecutive series of CMS exposure and behavioral testing because of the large number of animals in the two experimental groups. The first cohort consisted of eight WT control, eight WT CMS, eight PENK KO control and eight PENK KO CMS animals. The second cohort consisted of 15 WT control, 14 WT CMS, 11 PENK KO control and 10 PENK KO CMS animals. After 2 weeks of individual housing, each cohort was submitted to 5 weeks of CMS comprising the following stressors: 1 h restraint stress, 1 h social stress (four animals/cage), 1–3 h of stroboscopic lights, 4–8 h tilted cages, 8–12 h wet bedding, 8–12 h cage without bedding, 24 h light or dark periods, 18–24 h food deprivation followed by 30 min of inaccessible food and 18–24 h of water deprivation followed by 30-min exposure to an empty bottle. The different stressors were applied randomly in order to avoid adaptation to expected stress conditions. WT and PENK KO mice from control groups were handled twice a week during the 5 weeks of the CMS protocol. Animals were left undisturbed for a period of 1 day/week while the sucrose preference was assessed. At the end of the CMS procedure, animals were tested in a battery of behavioral tests to assess anxiety and depression-related phenotypes. At the first day after the end of the CMS protocol, sucrose preference was assessed for the last time (Fig. 1).

Figure 1.

Timeline of the experimental procedure. WT and PENK KO animals were submitted to 5 weeks of CMS (see Materials and methods) or to handling (twice a week). Within CMS exposure, the SPT was performed once a week. Subsequently, all animals were analyzed in a battery of behavioral test as indicated in the scheme. Twenty-four hours after the last behavioral test, mice were killed and the brains removed for gene expression analyses. d, day.

Corticosterone metabolites in feces

Two days after the end of the CMS protocol, fecal samples were collected for corticosterone measurements. For this purpose, we housed the animals in new clean home cages and collected the feces after 24 h. The samples were frozen at −80°C until further analysis. For corticosterone metabolite analysis, feces were unfrozen, placed on open Petri dishes and dried in an oven at 37°C for 1 h. Subsequently, samples were ground into powder with the help of a pestle, weighted and used for the extraction of corticosterone metabolites followed by quantification of corticosterone by an ELISA assay. Corticosterone metabolites were extracted with 1 ml of ethanol per 100 mg of fecal powder, followed by 30 min of vigorous shaking at room temperature (RT) and 30 min of centrifugation at 5000 rpm at RT. We collected 450 µl of the supernatant and dried it in a Speedvac (SpeedVac Savant, Thermo Scientific, Waltham, MA, USA) for 1 h at 35°C. Pellets were frozen at −20°C inside a desiccator to avoid hydration. On the following day, corticosterone ELISA assays were performed following the manufacturer's instructions (Arbor assays: catalog number KO14-H5).

Open field test

Animals were tested for exploratory and locomotor activity in an open field test. The test apparatus consists of an arena of 45 × 45 × 22 cm, dimly illuminated at 20 lux. Animals were placed in one of the corners and allowed to explore the arena for 10 min. The activity of the animals was recorded with infrared beam breaks placed on the outer sides of the arena. The total distance moved in the arena as well as the distance moved and time spent in the central part of the arena were analyzed with the Actimot 2 software (TSE-Systems GmbH, Bad Homburg, Germany).

Zero-maze test

Anxiety-related behavior was tested in the zero-maze test. The test was conducted in a round arena, elevated 38 cm above the ground. The arena was divided into four equal quadrants, with non-transparent walls enclosing the two opposite quadrants. Each mouse was placed into the open area of the maze and the animals' behavior was videotaped using a camera fixed above the maze and analyzed with a video-tracking system. Animals were tested for 5 min with 400 lux illumination. Time spent in the open areas of the maze was evaluated with the Ethovision software, version XT (Noldus, Wageningen, the Netherlands) as a parameter of anxiety.

Sucrose preference test

During individual housing, the animals were provided with two water bottles in their cages. For the sucrose preference test (SPT), one of the water bottles was replaced by a bottle containing 1% sucrose solution for 24 h. Before the beginning of the CMS protocol and the basal SPT measurements, two consecutive SPT tests were made for sucrose taste habituation. The SPT was performed once a week during the 5 weeks of CMS. Sucrose preference was calculated as follows: Sucrose preference = [sucrose intake/(sucrose intake + water intake)] × 100. The amount of sucrose solution and water intake was calculated by subtracting the final weight (after 24 h) from the initial weight of the bottles.

Forced swim test

Animals were tested for despair behavior in the forced swim test. Animals were placed in a Plexiglas cylinder (10 cm internal diameter, 50 cm height) filled with water of 26–28°C for 6 min. Immobility time in the last 4 min was measured using a stopwatch. Animals were judged to be immobile when they remained floating in water, making only movements necessary to keep the head above the water.

Tissue samples

Twenty-four hours after the last behavioral test, animals were killed by cervical dislocation. Adrenal glands were removed, cleaned and their weight was measured on an analytical scale. Brains were also removed, snap-frozen in ice-cold isopentane and stored at −80°C. To isolate the different brain areas, brains were warmed up to −20°C and cut into 1 mm coronal sections using a metal matrix for mouse brains (Zivic Instruments, Pittsburgh, PA, USA). The PFC, PVN, amygdala, bed nucleus of stria terminalis (BNST) and hippocampus were isolated from the sections using a 12G punching needle. Eppendorf tubes containing the isolated brain areas were stored and kept at −80°C until further molecular analysis.

RNA extraction

To extract RNA, 1 ml of Trizol reagent per hippocampus sample and 800 µl for PVN, PFC and amygdala were added. The samples were then transferred to MagNA Lyser tubes and homogenized with a Precellys machine (Peqlab, Erlangen, Germany). After homogenization, samples were incubated for 5 min at RT. Subsequently, 200 µl of chloroform was added and samples were mixed by vortexing, incubated for 3 min at RT and centrifuged for 15 min at 12 000 rpm at 4°C. The upper aqueous phase was then transferred to a new tube and RNA was precipitated by adding 500 µl of isopropanol. Samples were incubated for 10 min at RT and centrifuged for 15–20 min at 12 000 rpm at 4°C. The supernatant was disposed, the pellet was washed twice in 1 ml of 75% ethanol followed by 10 min of 8000 rpm centrifugation at 4°C. Afterwards, the supernatant was disposed and the RNA pellet was dried on a thermal plate for 5 min at 55°C. Dried pellets were dissolved in RNA-free water (hippocampus, 15 µl; amygdala and PFC, 6 µl; PVN, 5 µl). Purity and RNA concentration was evaluated by optical density measurements at 260 and 280 nm in a spectrophotometer (NanoDrop Instruments, Wilmington, DE, USA).

Gene expression analysis

Initially, cDNA was synthesized using the SuperScript First-Strand Synthesis System for RT-PCR Kit (Invitrogen Corp., Carlsbad, CA, USA) with oligo-deoxynucleotide-T (dT) primers, according to the manufacturer's instructions. Total RNA (30 ng for BNST, 20 ng for hippocampus and 15 ng for PFC, PVN and amygdala) was used as primary material. mRNA expression of the target genes was determined in triplicates by custom TaqMan® Gene Expression Assays (Applied Biosystems, Darmstadt, Germany) using the following primers: PENK: Mm01212875_m1; GAPDH: Mm99999915_g1 (Applied Biosystems). Each TaqMan® assay reaction consisted of 4 µl cDNA, 5 µl TaqMan® universal PCR Master Mix (Applied Biosystems), 0.5 µl Custom TaqMan® Gene Expression Assay and 0.5 µl of RNA-free water. Samples were processed in a 7500 Real-Time PCR Detection System (Applied Biosystems) with the following cycling parameters: 95°C for 10 min (hot start), 40 cycles at 95°C for 15 seconds (melting) and 60°C for 1 min (annealing and extension). Analysis was performed using the 7500 Sequence Detection Software version 2.2.2 (Applied Biosystems), and data were obtained as a function of threshold cycle (CT). For relative quantification (RQ) of gene expression, the mean CT values of the triplicates of the PENK gene and of the housekeeping gene GAPDH were calculated, followed by subtraction of the mean CT values of the GAPDH gene to the mean CT values of the PENK gene (ΔCT values). Then, the power of all ΔCT values was calculated based on the formula power = 2−ΔCT (Livak & Schmittgen 2001). Finally, RQ values, presented as fold change of the WT control group, were obtained by the division of the power value of each sample by the mean power value of the WT control group.

Statistical analysis

We did not detect interaction effects between cohort number and CMS or genotype; therefore, data from the two cohorts were pooled except for the forced swim test. Here, we present data only from the first cohort of animals because of data recording problems with the second cohort.

Statistical analyses for the corticosterone measurements, adrenal weight and the behavioral data were analyzed by two-way analysis of variance (anova), with CMS and genotype as between-subjects factors. SPT data were analyzed with three-way anova with repeated measures, with CMS, genotype as between-subjects factor and week as within-subjects factor. The sucrose preference values at week 5 were additionally analyzed using two-way anova (CMS and genotype as main factors). Post hoc analyses were performed using the Bonferroni test.

The mRNA expression levels were analyzed with the Mann–Whitney U-test. Statistical significance was set at P < 0.05.

Results

CMS increases the expression of the PENK gene in the PVN

PENK expression levels were measured in WT control and CMS animals. In the PFC, we found no significant differences in PENK gene expression (Fig. 2a; U = 18.00, n.s.). In the PVN, we observed an increase in the PENK mRNA levels after CMS (Fig. 2b; U = 3.000; P < 0.05). In the amygdala (U = 12.00, n.s.), in the BNST (U = 10.00, n.s.) or in the hippocampus (U = 13.00, n.s.), no significant changes were observed (Fig. 2c–e).

Figure 2.

Analyses of PENK gene expression. (a) PFC, (b) PVN, (c) amygdala, (d) hippocampus and (e) BNST in WT control and WT CMS mice. CMS led to a significant increase of mRNA expression in the PVN. No significant changes were observed in the other brain regions analyzed. Data are expressed as fold change compared to WT controls. *P < 0.05 by Mann–Whitney U-test.

CMS increases corticosterone levels and elevates adrenal gland weights in WT but not in PENK KO mice

We measured corticosterone levels in feces produced within 24 h at the end of the CMS protocol in WT and PENK KO mice (Fig. 3a). We found significant effects neither for the factor CMS (F1,66 = 2.15, n.s.) nor for the genotype (F1,66 = 0.98, n.s.). However, the analysis revealed a significant CMS × genotype interaction (F1,66 = 20.47, P < 0.0001). Exposure to CMS increased the corticosterone levels in WT animals (WT control vs. WT CMS: P < 0.001), but did not significantly change the levels in PENK KO animals. Moreover, post hoc analyses also revealed differences between WT and PENK KO animals within the control and CMS groups (WT control vs. PENK KO control: P < 0.05; WT CMS vs. PENK KO CMS: P < 0.001). Corticosterone levels were higher in PENK KO mice under baseline conditions but lower after CMS compared with those in WT mice. The weight of the adrenal glands was significantly increased after CMS (F1,43 = 4.690; P < 0.05), whereas we did not detect the genotypic effect (F1,43 = 0.0256, n.s.) or CMS × genotype interaction (F1,43 = 0.1086, n.s.). However, post hoc analysis of the data revealed that the CMS-induced increase in adrenal weights was significant in the WT but not in the PENK KO animals (WT control vs. WT CMS: P < 0.05) (Fig. 3b).

Figure 3.

Analysis of hormonal stress reactivity. (a) Measurement of corticosterone levels in feces, produced within 24 h, after 5 weeks of CMS (n = 17–18/group). CMS led to significantly increased corticosterone levels in WT but not in PENK KO mice. Corticosterone levels in the control group were higher in PENK KO compared with those in WT animals. After CMS, corticosterone levels were higher in WT than in PENK KO mice. (b) Adrenal weights are higher in the WT but not in the PENK KO group after CMS. Data are expressed as mean ± SEM. #P < 0.05, ###P < 0.001, difference between WT and PENK KO mice. *P < 0.05, ***P < 0.001, difference between control and CMS WT mice. Two-way anova followed by Bonferroni test.

CMS increases anxiety levels in the zero-maze test in WT but not in PENK KO mice

Anxiety-related behaviors were assessed in the zero-maze test (Fig. 4a), where time in the open areas is inversely correlated to the level of anxiety. We found a significant CMS effect (F1,77 = 8.008, P < 0.01), and post hoc analyses showed a decrease in the time spent in the open areas in WT CMS compared with the WT control mice (WT control vs. WT CMS: P < 0.01). No genotype (F1,77 = 0.1147, n.s.) or CMS × genotype (F1,77 = 1.817, n.s.) effect was observed. We also performed an open field test to control for potential differences in the general activity or exploratory behavior (Fig. 4b). We found a significant main effect for CMS (F1,78 = 5.110, P < 0.05), due to an increase in the distance traveled by the CMS animals. Nevertheless, post hoc analyses were not significant. We did not detect significant differences for genotype (F1,78 = 0.03712, n.s.) or CMS × genotype interaction (F1,78 = 0.6320, n.s.). Time or activity in the central part of the arena, often used as anxiety markers, did not differ between the genotypes (genotype effect on time: F1,75 = 0.1700, n.s.; genotype effect on activity: F1,77 = 0.1451, n.s.). In addition, we found neither CMS effects (effect of CMS on time: F1,75 = 0.5893, n.s.; effect of CMS on activity: F1,77 = 1.453, n.s.) nor interaction between genotype and CMS (time: F1,75 = 1.351, n.s.; activity: F1,77 = 0.0134, n.s.) (data not shown).

Figure 4.

Assessment of anxiety-related behavior in the zero-maze test and of locomotor activity in the open field test. In the zero-maze test (a), CMS exposure significantly increased anxiety levels in WT mice, as indicated by a decrease in the time spent in the open areas of the maze. CMS had no effect in PENK KO animals (n = 18–22/group). In the open field test (b), a main effect of CMS was observed. CMS led to an increased activity in WT and PENK KO mice (n = 18–23/group). Data are expressed as mean ± SEM. **P < 0.01, difference between control and CMS WT mice using two-way anova followed by Bonferroni test.

CMS induces anhedonia in WT but not in PENK KO mice

A state of anhedonia (reduced responsiveness to pleasurable stimuli) is indicated by a decreased sucrose preference over water. We assessed the sucrose preference at weekly intervals during the CMS protocol. Repeated measures three-way anova was significant for the CMS effect (F1,72 = 6.704, P < 0.05) and week×CMS (F5,360 = 2.660, P < 0.05) interaction. The CMS test procedure significantly influenced the sucrose preference, shown by a significant CMS effect at week 5 according to two-way anova (CMS effect: F1,72 = 21.60; P < 0.01). Moreover, CMS × genotype interaction effect was significant (F1,72 = 7.374; P < 0.01), which shows that CMS effects differed between the genotypes. Post hoc analysis of the data using Bonferroni test revealed that CMS significantly decreased sucrose preference in WT but not in PENK KO mice (WT control vs. WT CMS: P < 0.05) (Fig. 5a).

Figure 5.

Depression-related behaviours. Anhedonia was measured in the SPT (a). Sucrose preference, measured after 5 weeks of CMS revealed that CMS led to a decreased preference for the sucrose solution (anhedonia) in WT but not in PENK KO mice (n = 17–22/group). (b) Despair-like behavior was measured in the forced swim test. CMS led to an increase in the immobility time in WT but not in PENK KO mice (n = 8/group). Data are expressed as mean ± SEM. *P < 0.05, ***P < 0.001 using two-way anova followed by Bonferroni test.

CMS increases immobility in the forced swim test in WT but not in PENK KO mice

Effect of CMS on despair-like behavior was measured in the forced swim test (Fig. 5b) as an increase in the immobility time. We found no significant effect for genotype (F1,28 = 0.01431, n.s.) or CMS × genotype interaction (F1,28 = 3.709, n.s.), but a significant CMS effect (F1,28 = 5.497, P < 0.05). The post hoc test showed that the immobility time significantly increased in WT but not in PENK KO mice after CMS (WT control vs. WT CMS: P < 0.05).

Discussion

In this study, we demonstrate that male WT mice exposed to 5 weeks of CMS have elevated corticosterone levels and show increased anxiety in the zero-maze but not in the open field test and depression-related behaviors in SPT and FST. In contrast, PENK KO mice showed none of these effects, indicating that mice lacking enkephalins are resilient to CMS.

Under basal conditions (unstressed), anxiety and depression-related behaviors were similar between PENK KO and WT mice. Although these findings seem to be in contrast to previous data indicating enhanced basal anxiety levels in PENK KO mice using the same paradigms (Bilkei-Gorzo et al. 2004a, 2008a,2008b; Konig et al. 1996), they can be readily explained by the different experimental conditions. In this study, mice were single housed for 8 weeks and tested for anxiety-related behavior (zero-maze test) after they had been analyzed for sucrose preference and in the open field test. In the aforementioned studies, PENK KO animals were group housed and naive to each behavioral test. It is well known that group housing vs. single housing (Voikar et al. 2005) as well as test–test interaction (Mcilwain et al. 2001; Voikar et al. 2004) strongly influence anxiety-related phenotypes. We have shown that the basal phenotype of Penk KO mice is particularly sensitive to environmental effects (Bilkei-Gorzo et al. 2008a). We now reanalyzed our previous result of zero-maze experiments, which were carried out in the same time period, in which single- and group-housed WT and PENK KO mice were tested. We found a significant effect of interaction between genotype and housing on the time spent in the open areas. Post hoc analysis revealed that PENK KO mice presented enhanced anxiety levels compared with WT mice when the animals were group housed but not when they were single housed for 2 weeks. Moreover, the different housing conditions did not alter the behavior reactivity of WT mice (Fig. S1, Supporting Information).

In humans, chronic stress has been shown to be a risk factor for the development of anxiety states and depressive-like symptoms (Kendler et al. 1999; Tafet & Bernardini 2003). Likewise, the CMS induces anhedonia (Willner 1997), anxiety and despair behaviors (Mineur et al. 2006) as well as increased corticosterone levels (Song et al. 2006) in mice. These behavioral and hormonal changes can be reversed by anti-depressant treatment (Willner 2005), suggesting that CMS exposure represents an animal model of depression. In accordance with the literature, our results in WT animals also showed that CMS induced a decrease in the sucrose preference, increase in the immobility time in the forced swim test and reduced time spent in the open parts of the zero-maze, indicative of anhedonia, despair behavior and increased anxiety, respectively. In addition, we also observed increased baseline corticosterone levels and higher adrenal weights (hypertrophy) in CMS-exposed WT animals.

It is well known that corticosterone levels show fluctuations over a period of 24 h, i.e. a circadian pattern and also fast, pulsatile fluctuations, i.e. an ultradian cycle. The frequency and amplitude of corticosterone ultradian fluctuations delineate its circadian pattern (Lightman & Conway-Campbell 2010; Walker et al. 2010). Therefore, in order to avoid misleading results due to mismatching circadian/ultradian time points between individuals and between genotypes, we measured fecal corticosterone levels. These values reflect the mean corticosterone concentration and thus are probably more reliable markers of long-lasting changes under chronic stress conditions because they are less susceptible to fast fluctuations (Millspaugh & Washburn 2004). Furthermore, this non-invasive sampling method is probably a more reliable marker to detect differences in the general hormonal status of the animals due to CMS exposure (Touma et al. 2004).

In strong contrast to our results in WT animals, we observed none of these changes in PENK KO mice, suggesting that KO mice are resilient to CMS. This result thus identifies enkephalin as a key component for the development of CMS-induced physiological and behavioral changes.

The resilience of PENK KO mice to the CMS effects was surprising and the opposite of what we had expected. Several previous studies showed an anti-depressant effect of enkephalin signaling in the forced swim, conditioned suppression of motility and in the learned helplessness models. Thus, WT animals treated with exogenous analogs of enkephalin (Tejedor-Real et al. 1995) or with inhibitors of enkephalin catabolism (Baamonde et al. 1992; Tejedor-Real et al. 1995, 1998) showed attenuated depression-related behaviors, observed by a reduction in the immobility time (Baamonde et al. 1992) and number of escape failures (Tejedor-Real et al. 1995, 1998). Nevertheless, our previous studies did not show a depression-related phenotype in PENK KO mice (Bilkei-Gorzo et al. 2007). We therefore hypothesized that such phenotype may become apparent after exposing the animals to CMS. Clearly, this was not the case.

Stress response involves distinct brain areas and several neurotransmitters (Carrasco & Van De Kar 2003; Dedovic et al. 2009; Herman et al. 2003; Locatelli et al. 2010; Tafet & Bernardini 2003). Activation of the distinct stress-related neuronal pathways is highly dependent on the modality of the stress stimuli. Thus, stress response circuits can differ upon reactive vs. anticipatory stressors, physical vs. psychological stressors, acute vs. chronic stressors or repeated vs. novel stressors (Dedovic et al. 2009; Gaillet et al. 1991; Herman et al. 2003; Pacak et al. 1998; Ulrich-Lai & Herman 2009). Furthermore, enkephalin and its receptors are widely expressed within brain areas involved in stress response, such as hippocampus, hypothalamus and amygdala (Beaulieu et al. 1996; Le Merrer et al. 2009; Poulin et al. 2006), and co-localized with several of the main neurotransmitters, such as CRH, GABA and glutamate (Ceccatelli et al. 1989; Kalyuzhny & Wessendorf 1998; Poulin et al. 2008; Pretel & Piekut 1990; Zhu & Pan 2005). Consequently, the neuronal pathways on which enkephalin can be involved are multiple and may lead to differential modulatory effects in stress response. When we analyzed PENK gene expression in WT mice in order to scrutinize molecular mechanisms, we observed an increase in the PENK gene expression in the PVN (Fig. 2b). Similar results were shown after exposure to other stressors, such as immobilization (Palkovits 2000) and chronic variable stress (Christiansen et al. 2011). Therefore, not excluding the contribution of enkephalin signaling in other brain areas in CMS reactivity, we focused our discussion on enkephalin signaling in the PVN.

The PVN contains several neuronal populations involved in the control of endocrine stress response. Among them are the CRH neurons, part of the HPA axis, the main branch of the endocrine stress response (Herman et al. 2003). Although only 20% of the CRH-expressing neurons co-express enkephalin, enkephalin-expressing neurons within the PVN and in the inhibitory peri-PVN region may have a strong control on the activity of CRH-expressing parvonuclear neurons. Indeed, we observed that deletion of Penk influenced corticosterone levels both in unstressed and in stressed animals (Fig. 3a), suggesting that enkephalin is an important modulator of basal and stress-induced hormone release and that the lack of enkephalin may contribute to the attenuation of the HPA axis activity under chronic stress conditions.

Based on our present data and previous findings, we propose a mechanism for the surprising resilience phenotype of PENK KO mice under CMS conditions. Parvocellular CRH neurons within the PVN are controlled by inhibitory GABAergic neurons (Cullinan et al. 2008; Herman et al. 2003, 2004; Miklos & Kovacs 2002). These GABAergic neurons are known to express and be inhibited by the activation of presynaptic MOR (Wamsteeker Cusulin et al. 2013). Under basal, stress-free situation, the low amount of enkephalin, which the peri-PVN neurons produce, contributes to the inhibitory control of CRH-expressing PVN neurons. Thus, deletion of enkephalin leads to an enhanced basal activity in the PVN. When stress upregulates PENK expression, it leads to a release of enkephalin also in the peri-PVN region. The locally released enkephalin may activate to MOR autoreceptors of the inhibitory GABAergic neurons leading to a decreased GABAergic tone on the CRH-expressing neurons in the PVN. Disinhibition of this neuronal population can lead to hyperactivity of the HPA axis and to an increase in the corticosterone levels under stress conditions in WT animals. In contrast, because of the lack of enkephalin in PENK KO mice, the GABAergic tone on the parvocellular CHR neurons is maintained under stress. As a consequence, the activity of the HPA axis and the secretion of corticosterone remain low in these mice (Fig. 3a) contributing to the resilience of PENK KO mice to CMS. Nevertheless, this model does not exclude the possibility that other brain regions are also involved in the chronic stress-resistant phenotype of PENK KO mice.

Acknowledgment

This research was supported by a research grant from the Deutsche Forschungsgemeinschaft, Sonderforschungsbereich 645. Andreas Zimmer is a member of the DFG Cluster of Excellence ImmunoSensation. The authors declare that they have no conflict of interest.

Ancillary