Chronic social stress during adolescence: Interplay of paroxetine treatment and ageing

Neuropharmacology 72 (2013) 38e46

Contents lists available at SciVerse ScienceDirect

Neuropharmacology journal homepage: www.elsevier.com/locate/neuropharm

Chronic social stress during adolescence: Interplay of paroxetine treatment and ageing Sebastian H. Scharf a, *, Vera Sterlemann a, Claudia Liebl a, Marianne B. Müller a, Mathias V. Schmidt b a b

RG Molecular Stress Physiology, Max Planck Institute of Psychiatry, Kraepelinstr. 2-10, 80804 Munich, Germany RG Neurobiology of Stress, Max Planck Institute of Psychiatry, Kraepelinstr. 2-10, 80804 Munich, Germany

a r t i c l e i n f o

a b s t r a c t

Article history: Received 28 November 2012 Received in revised form 18 February 2013 Accepted 5 March 2013

Exposure to chronic stress during developmental periods is a risk factor for a number of psychiatric disorders. While the direct effects of stress exposure have been studied extensively, little is known about the long-lasting effects and the interaction with ageing. The same holds true for the treatment with selective serotonin reuptake inhibitors (SSRIs), which have been shown to prevent or reverse some stress-induced effects. Here, we studied the direct and long-lasting impact of chronic social stress during adolescence and the impact of chronic treatment with the SSRI paroxetine in adulthood and aged animals. Therefore, male CD1 mice at the age of 28 days were subjected to 7 weeks of chronic social stress. Treatment with paroxetine was performed per os with a dosage of 20 mg/g BW. We were able to reverse most of the effects of chronic social stress in adult mice (4 months old) and to some extend in aged animals (15 months old) with the SSRI treatment. Especially the regulation of the HPA axis seems to be affected in aged mice with a shift to the use of vasopressin. Our results demonstrate that chronic stress exposure and antidepressant treatment at the end of the developmental period can have a significant and long-lasting impact, highly relevant for healthy ageing. Ó 2013 Elsevier Ltd. All rights reserved.

Keywords: Stress HPA SSRI Paroxetine Ageing Long-term

1. Introduction Stress is a part of everyday life and in general describes a state which differs from homeostasis, the ideal situation of an organism (McEwen, 1998). Events that trigger these dissociations are referred to as stressors and can be divided into physical and psychogenic stressors (Dayas et al., 2001). If a stressor is perceived, the reaction is guided by the hypothalamicepituitaryeadrenal (HPA) axis, one of the main physiological systems involved in the stress response (de Kloet et al., 2005; Joels and Baram, 2009; Lupien et al., 2009). While the occasional activation of the HPA axis can have beneficial effects, for example by improving learning and memory (Joels et al., 2006; Lupien and Lepage, 2001), chronic activation can lead to maladaption of the whole system and cause deleterious consequences (de Kloet et al., 2005). Chronic stress exposure has been associated with an increased risk for numerous diseases, including cardiovascular problems (Rosengren et al., 2004), metabolic

* Corresponding author. Tel.: þ49 (0) 89 30622 576; fax: þ49 (0) 89 30622 610. E-mail addresses: [email protected] (S.H. Scharf), liebl@ mpipsykl.mpg.de (C. Liebl), [email protected] (M.B. Müller), [email protected] (M.V. Schmidt). 0028-3908/$ e see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.neuropharm.2013.03.035

disturbances (Abraham et al., 2007) or affective disorders (Charney and Manji, 2004; de Kloet et al., 2005). The influence of stress is also dependent on the phase of exposure: during critical time frames in the maturation of the stress system, in which the vulnerability to environmental challenges is enhanced, stress effects can be especially devastating (Lupien et al., 2009; Marco et al., 2011). One of these time windows is the adolescent period. Studies in rodents have shown that animals exposed to a stressful environment during their adolescence, showed a maladapted phenotype in adulthood (McCormick et al., 2008). Important modulators of HPA axis function were found to be differently expressed in adult animals following adolescent stress, for example corticotrophin-releasing hormone (CRH) in the paraventricular nucleus of the hypothalamus (PVN) or mineralocorticoid receptor (MR) and glucocorticoid receptor (GR) mRNA levels in the hippocampus (Schmidt et al., 2007). It also seems that synaptic plasticity is decreased in previously stressed animals, as decreases in cell proliferation, BDNF levels and synaptophysin levels have been found in the hippocampus (Sterlemann et al., 2008). These differences on the molecular level are also reflected on the physiological as well as on the behavioural levels. For example, chronically elevated levels of corticosterone, concomitant with enlarged adrenal glands and involuted thymus glands are

S.H. Scharf et al. / Neuropharmacology 72 (2013) 38e46

found in most of the studies. The behavioural differences include an increase in anxiety-like behaviour in stressed animals or increased anhedonia in the sucrose preference test. Another finding has been the decline in cognitive function, which nicely fits the molecular data on impaired synaptic plasticity and disrupted LTP formation in the hippocampus (Sterlemann et al., 2008). These data are in line with the human literature on the effects of early trauma, not so much focussing on molecular effects in the brain, but on parameters more easily accessible in humans like risk of disease, cognitive functions and hormone levels (Caspi et al., 2003; Mueller et al., 2010). Interestingly, stress-induced alterations can be attenuated via antidepressant treatment. Antidepressant drugs from the selective serotonin reuptake inhibitor (SSRI) class have been e at least partly e successful in either blocking the effects of early life stress during the stressful episode (Schmidt et al., 2007) or reversing the stress effects by treatment after the stress phase (Carboni et al., 2010; El Khoury et al., 2006; Huot et al., 2001; Navailles et al., 2010). While the consequences of chronic stress exposure have been described in detail for adult animals, information about the longterm effects in aged individuals remain sparse. This holds especially true for stress during the adolescence period. Some studies investigated the HPA axis in aged individuals (Dalm et al., 2005; Otte et al., 2005), even in interplay with early trauma, but the literature remains controversial (Gerritsen et al., 2010). A similar situation is found in the literature about antidepressant treatment: Information about acute and chronic treatment has accumulated over the years, but studies on the long-term effects are limited. In the present study, we therefore investigated the long-term effects of chronic social stress during the adolescence of male mice in adult and aged animals. Furthermore, we assessed the modulation of these effects by antidepressant treatment to gain insight about the interaction of stress, treatment and ageing.

39

(condition: control vs stress; treatment: vehicle vs paroxetine; time point: 5 weeks after stress vs 12 months after stress). 2.2.1. Experiment 1 e long term effects in adult animals To assess the long term effects of the stress exposure performed during the adolescent and early adulthood as well as to analyse the following paroxetine treatment on the adult phenotype, we investigated the animals on multiple levels. To study physiological changes, the body weight of the animals was measured at critical time points of the experiment: before the stress exposure, directly after the stress exposure and 5 weeks after the stress exposure. On the endocrine level, corticosterone (CORT) levels in the plasma of the animals were assessed directly after as well as 5 weeks after the stress exposure. Mice were screened for changes at the behavioural level in the elevated plus maze test, the open field (OF) test, the social interaction test and the tail suspension test. To investigate changes in the CNS and HPA axis regulation, the mRNA levels of specific marker genes were measured, including GR, MR, CRH and AVP mRNA expression. In addition, the weight of the adrenal and thymus glands was assessed. 2.2.2. Experiment 2 e long term effects in aged animals The second cohort of animals was allowed to age under standard conditions while being kept in single housing. These animals were investigated at 12 months after the stress exposure (15 months of age) in a broadly similar fashion as experiment 1. Body weight and adrenal gland weight were measured, together with plasma corticosterone levels. The elevated plus maze test was performed to examine alterations in anxiety-related behaviour. To shed additional light on HPA axis regulation and the efficacy of the feedback mechanisms, we performed two stress challenge tests, the acute stress response test as well as the combined DEX/CRH challenge test in separate batches of animals. In the brain, the same genes as in experiment 1 were investigated. 2.3. Chronic social stress paradigm The chronic social stress paradigm was performed as described in detail elsewhere (Schmidt et al., 2007). In brief, animals were housed in groups of 4 and the composition of the group was changed twice a week for 7 weeks in total. The schedule was laid out in a way that no mouse meets a known mouse during the whole 7 weeks, with the exception of the last change of cage mates. The time point directly after stress was 4 days after the last cage mate change. Clean cages were provided with each change. Control animals were also housed in groups of 4 for 7 weeks, but stayed in the same group the whole time. Cages from the control animals were changed on a weekly basis. Animals with major bite wounds were excluded (3%).

2. Materials and methods 2.4. Treatment and drugs 2.1. Animals The experiments were carried out with male CD1 mice from Charles River Laboratories (Maastricht, the Netherlands). The animals were 28 days old at the beginning of the experiment. Housing consisted of Plexiglas cages (45  25  20 cm) with bedding and additional nesting material under controlled conditions with a regular 12L:12D light cycle (lights on at 6:00 am) as well as constant temperature (23  2  C) and humidity (55  5%). Food (Altromin 1324, Altromin GmbH, Germany) and water were provided ad libitum. Experiments were carried out in the animal facilities of the Max-Planck-Institute of Psychiatry in Munich, Germany. The experiments were carried out in accordance with the European Communities’ Council Directive 2010/63/EU. All efforts were made to minimize animal suffering during the experiments. The protocols were approved by the committee for the Care and Use of Laboratory animals of the Government of Upper Bavaria, Germany. 2.2. Experimental design After arrival, a total of 164 animals were randomly split into two condition groups. The first group was subjected to 7 consecutive weeks of the chronic social stress paradigm (n ¼ 80; described below) and therefore housed in groups of four that were changed regularly during the paradigm. The animals from the control condition (n ¼ 84) were also housed in groups of four, but the cage mates stayed the same during the 7 week procedure. Following the stress phase, mice were single housed and each condition group was split into two treatment groups. One treatment group received the antidepressant drug paroxetine (n ¼ 34 CON; n ¼ 34 STR; see Treatment and drugs) via the drinking water for four weeks, followed by one week of washout phase. The vehicle group (n ¼ 50 CON; n ¼ 46 STR) received normal drinking water. It should be noted here that although single housing represents a stressor itself in many species including rats, in male mice it has been shown that single housing does not affect relevant parameters (Arndt et al., 2009; Bartolomucci et al., 2003). After the treatment and washout phase, the animals were extensively tested and one cohort of animals was sacrificed (Experiment 1), while the other cohort was allowed to age under standard conditions in single housing and was tested again 12 months after the stress exposure (Experiment 2). This resulted in a final 2  2  2 design

Animals were treated with paroxetine per os via the drinking water for 4 weeks. Paroxetine (Seroxat; GlaxoSmithKline, Munich, Germany), an antidepressant drug from the class of selective serotonin reuptake inhibitors (SSRIs), was given at a concentration of 20 mg/kg BW, as this dose was sufficient to elicit antidepressantlike effects in previous experiments (Keck et al., 2003; Schmidt et al., 2007). Paroxetine was diluted in tap water to a concentration of 0.11 mg/ml as the mice drank approximately 7 ml per day. Paroxetine consumption was monitored on a weekly basis and showed a mean discrepancy of the desired dose of 12e24%. The vehicle control group received normal tap water. In the combined DEX/CRH challenge test (Heuser et al., 1994), agents were given via intraperitoneal injection (i.p.). Dexamethasone (Dexa-ratiopharm 100 mg; Ratiopharm) was given at 50 mg/kg BW with an injection volume of 100 ml. The injection was performed 6 h before the blood sampling (injection at 1100 h, blood sampling at 1700 h). Directly after blood sampling, animals were injected with CRH (C3042-.5MG; SigmaeAldrich) at 0.15 mg/kg BW also dissolved in 100 ml injection volume and corticosterone was measured again 30 min after CRH injection (CRH blood sampling: 1730 h). Basal levels were taken at 1700 h after one week of recovery. 2.5. Stress response test As another marker for HPA axis activation and feedback, the stress response test was used. In this test, blood was taken from the tail vein and animals were subsequently restrained in a plastic tube for 30 min. At the end of the restrain, a second sample was taken and the animals were placed back in their home cages to recover. The last sample was taken 90 min after stress onset (60 min after offset). The stress response test was performed between 0800 h and 1200 h. Total sample volume did not exceed 30 ml. 2.6. Behavioural tests Behavioural testing was performed in a separate room with similar conditions to the housing room. All animals were allowed to acclimate to the testing room for at least 6 days before the start of the actual test. Mouse behaviour was tracked, recorded and analysed via the AnyMaze software (Stoelting Co., USA). Detailed descriptions of the performed tests can be found in the supplemental information.

40

S.H. Scharf et al. / Neuropharmacology 72 (2013) 38e46

2.7. Sampling procedure Trunk or tail-cut blood was collected using 1.5 ml EDTA-coated microcentrifuge tubes (Kabe Labortechnik, Germany). Tail blood was collected without anaesthesia as described previously (Fluttert et al., 2000) by a small incision in the dorsal tail vein using a razorblade. For trunk blood collection, animals were deeply anesthetised using isoflurane and decapitated. All blood samples were immediately put on ice and centrifuged for 15 min at 8000 rpm at 4  C. Plasma was transferred to clean 1.5 ml microcentrifuge tubes and stored at 20  C. Animals were always sacrificed between 0800 h and 1200 h (during the light phase). The time between the first disturbance of the animals and the blood sampling or sacrifice was less than 3 min in all cases. After sacrifice, adrenal glands, thymus glands and the whole brain were dissected by trained personnel. Adrenals and thymus glands were stored in isotonic NaCl solution, freed from remaining fat tissue under magnification and the weight was determined. Whole brains were snap-frozen in pre-cooled 2-Methylbutane immediately after dissection and stored at 80  C until further procession. 2.8. Radioimmunoassay To determine the concentrations of corticosterone in the plasma of the animals, a radioimmunoassay (RIA) was performed using a commercially available kit (ImmunoChemTM Double Antibody Corticosterone 125I RIA Kit, MP Biomedicals, LLC, Orangeburg, NY) with a detection limit of 7.7 ng/ml following the standard protocol. To extend the sensitivity of the assay, basal samples were diluted less to stay within the linear range and an additional low standard was used for standard curve generation. 2.9. In situ hybridization Frozen brains were sectioned at 16  C in a cryostat microtome at 20 mm in the coronal plane through the level of the hypothalamic PVN and the dorsal hippocampus. The sections were thaw-mounted on superfrost slides, dried and kept at 80  C. In situ hybridization using 35S-UTP-labelled ribonucleotide probes were performed as described previously (Schmidt et al., 2007). Briefly, sections were fixed in 4% paraformaldehyde and acetylated in 0.25% acetic anhydride in 0.1 M triethanolamine/HCl. Subsequently, brain sections were dehydrated in increasing concentrations of ethanol. Tissue sections were saturated with 100 ml of hybridization buffer containing approximately 1.5  106 cpm 35S-labelled riboprobe. Brain sections were coverslipped and incubated overnight at 55  C. The following day, the sections were rinsed in 2  SSC (standard saline citrate), treated with RNAse A (20 mg/l) and washed in increasingly stringent SSC solutions at room temperature. Finally, sections were washed in 0.1  SSC for 1 h at 65  C and dehydrated through increasing concentrations of alcohol. Radioactively labelled slides were apposed to Kodak Biomax MR films (Eastman Kodak Co., Rochester, NY) and were developed using an automated developing machine. Films were digitized and relative expression was measured by optical densitometry using the ImageJ software (available at http://rsb.info.nih.gov/ij; developed by Wayne Rasband, National Institutes of Health, Bethesda, MD). For each animal, the mean of the bilateral structures of two sections was calculated if applicable, deducting the background from the value. Background signal was measured in a structure not expressing the gene of interest, which was the stratum radiatum in the case of the hippocampus and the peri-PVN region in case of the PVN. 2.10. Statistical analysis For statistical analyses, the statistical program R was used. Outliers were detected using the sequential Grubbs method (Rosner, 1983). In case of multiple testing, outliers were calculated for the single time points and added up for the final analysis. For all statistical calculations, number of outliers as well as corrected sample sizes are given always in the form of (VEH CON/VEH STR/PAR CON/PAR STR) if not stated otherwise. Comparisons between two groups were done by student Ttests (independent samples, two-tailed). For the multifactorial datasets (2  2 design), ANOVA was performed using an interaction model, followed by Tukey HSD post hoc testing. Significance levels were set at p < 0.05 and a trend was recognized at p < 0.1. Figures were created using the SigmaPlot software (SigmaPlot 10.0) with help of Adobe Photoshop CS3 and Adobe Illustrator CS3.

3. Results 3.1. Experiment 1 e long term effects in adult animals 3.1.1. Body weight The body weight of the animals was investigated at key time points in the experiments, as depicted in Table 1. We could not observe any effect of the stress exposure on body weight of the animals during the stress exposure (2 outliers, CON: 83/STR: 79).

Table 1 Body weight of the animals. Data are given in mean  SEM. Body weight [g]

Vehicle Control

Paroxetine Stress

Absolute weight 19.61  0.22 20.18 before stress [g] Weight gain during 97.28  2.17 92.66 stress [%] Weight gain during 4.00  0.72 1.48 recovery [%] Weight gain during 19.42  1.96 23.38 ageing [%]

Control

Stress

 0.21 20.22  0.23

20.34  0.24

 2.07 89.02  3.07

90.10  2.64

 0.49 11.37  1.03* 12.43  1.18*  2.40

9.75  1.11* 12.32  1.76*

*Significantly different to respective vehicle-treated animals p < 0.05.

When investigating body weight gain during the recovery period (3 outliers, 48/45/36/32), we found a significant effect of treatment (F1,157 ¼ 126.238, p < 0.05). Paroxetine-treated animals showed a higher weight gain within controls as well as within the stress group. Furthermore, vehicle-treated control animals differed from paroxetine-treated stress animals and vehicle-treated stress animals differed from paroxetine-treated controls. Results are included in Table 1. 3.1.2. Basal plasma corticosterone Corticosterone in the plasma of the mice was measured at the end of the chronic social stress paradigm as well as at the end of the washout phase following the treatment phase. Plasma corticosterone levels in the morning were significantly elevated in the stress animals directly after the stress paradigm (7 outliers, CON: 77, STR: 77, T114.665 ¼ 2.6908, p < 0.05), as seen in Fig. 1A. At the end of the treatment period followed by 1 week of washout phase (1 outlier, 23/24/16/12), corticosterone levels were investigated again and ANOVA revealed an effect of condition  treatment interaction effect (F1,71 ¼ 4.025, p < 0.05), as depicted in Fig. 1B. Post hoc testing confirmed that within the vehicle-treated group, stress animals show significantly higher levels of CORT. In addition, paroxetine treatment reduced CORT levels in the previously stressed animals, but not in control animals. 3.1.3. Adrenal and thymus weights In concurrence with the sacrifice of the animals 5 weeks after stress, adrenal and thymus glands were removed and analysed. For the weight of the thymus glands (1 outlier, 23/20/11/12), ANOVA showed a significant condition effect (F1,62 ¼ 21.756, p < 0.05). The weight of thymus glands was significantly decreased in stressed animals within the vehicle-treated animals (see Fig. 1C). Regarding the adrenal gland weights (1 outlier, 23/22/11/12), we found a significant effect of treatment (F1,64 ¼ 14.224, p < 0.05). Post hoc testing showed that vehicle-treated stress animals had a higher adrenal gland weight than paroxetine-treated control stress animals (see Fig. 1D). 3.1.4. Behavioural effects The open field test in our setup provides information about explorative behaviour as well as habituation to a novel environment (no outliers, 24/22/24/22). Along with the habituation time effect (F4,352 ¼ 52.953, p < 0.05), repeated measures ANOVA revealed a time  condition interaction effect (F4,352 ¼ 3.825, p < 0.05), as seen in Fig. 1J. No significant post hoc effects were found. For the other behavioural tests, no significant condition, treatment or interaction effects were observed (Supplemental Figure S1). 3.1.5. Gene expression of marker genes To test the effects of the stress exposure on gene expression in the brain, several HPA-axis relevant genes were investigated 5

S.H. Scharf et al. / Neuropharmacology 72 (2013) 38e46

41

Fig. 1. Characteristics of the adult animals. All data (except A) refer to the time point of 5 weeks after stress. (A) Corticosterone directly following the stress exposure. Stress animals showed elevated corticosterone levels compared to controls. (B) Corticosterone levels were still increased 5 weeks after stress. Paroxetine treatment was able to reverse the stress effect and significantly decrease the corticosterone levels. (C) Normalised weight of the thymus gland. Stress exposure caused involution of the thymus, which was not reversible via SSRI treatment. (D) Normalised adrenal gland weight. Adrenal gland weight was significantly decreased by paroxetine treatment in stressed animals only. (E) GR mRNA expression in the CA1 subfield of the hippocampus. (F) AVP mRNA expression in the PVN. (G) CRH mRNA expression in the PVN. (H) Representative images from (E). (I) Representative images from (G). (J) Adaption to a novel environment in the OF test. Previously stressed animals show impaired adaption, which was reversed via paroxetine treatment. CON ¼ control, STR ¼ stress, VEH ¼ vehicle-treated animals, PAR ¼ paroxetine-treated animals. Data are given in mean þ SEM. *p < 0.05 vs control, yp < 0.05 vs vehicle-treated animals.

weeks after cessation of the stress experience. In the hippocampus, the GR and MR mRNA levels were investigated, while in the PVN the GR, CRH and AVP were investigated. In the hippocampus, ANOVA revealed a trend towards a condition effect (F1,39 ¼ 3.546, p ¼ 0.067) for GR expression in the CA1 region (2 outliers, 12/10/11/ 10), which could not be attributed to a single contrast via post hoc testing (see Fig. 1E). No effect was found for the MR in the

hippocampus or the GR in the PVN. For AVP in the PVN (1 outlier, 10/8/7/7), ANOVA revealed a trend towards a treatment effect (F1,28 ¼ 3.183, p ¼ 0.085), with no significant post hoc comparisons (see Fig. 1F). When investigating CRH in the PVN (no outliers, 10/8/ 8/8), we found a significant condition  treatment interaction effect (F1,30 ¼ 4.229, p < 0.05), with no significant post hoc tests (see Fig. 1G).

42

S.H. Scharf et al. / Neuropharmacology 72 (2013) 38e46

3.2. Experiment 2 e long term effects in aged animals 3.2.1. Body weight We also investigated the body weight gain of the animals from the end of recovery until the experiments in aged animals (no outliers, 23/21/22/21). Here, ANOVA showed a trend towards a condition effect (F1,83 ¼ 3.089, p ¼ 0.083) as well as a significant treatment effect (F1,83 ¼ 31.074, p < 0.05). Post hoc test showed that paroxetine treatment significantly reduced BW gain within the control as well as within the stress condition. Data are summarised in Table 1. 3.2.2. Stress response test To test the influence of long-term effects of stress exposure and pharmacological treatment on HPA-axis dynamics under activation, the stress response test was performed 12 months after cessation of the original chronic stress phase (7 cumulative outliers, 9/8/11/6). Repeated measures ANOVA confirmed an effect of time (F2,56 ¼ 39.962, p < 0.05) and a time  condition interaction (F2,56 ¼ 4.918, p < 0.05) as well as a trend towards a time  condition  treatment interaction (F2,56 ¼ 2.424, p ¼ 0.098). Post hoc tests were then calculated for the different time points. At the basal level, vehicle-treated stress animals showed a trend towards lower levels of CORT compared to vehicle-treated controls. In the response as well as in the recovery samples, only paroxetine-treated stress animals showed lower values compared to paroxetine-treated controls. To better dissect the differences in stress-induced elevation of corticosterone levels (1 outlier, 10/10/11/9) and feedback (2 outliers, 10/8/11/10) after cessation of the stressor, values are displayed separately as difference between basal and 30 min or 30 min and 90 min in Table 2. For the stress-induced increase, ANOVA revealed a trend towards an effect of condition (F1,36 ¼ 3.956, p ¼ 0.054) and a condition  treatment interaction (F1,36 ¼ 8.564, p < 0.05). Post hoc testing showed a decrease in response for the stress animals compared to controls within the paroxetine treated group. No significant effects were found for the feedback comparison. 3.2.3. DEX/CRH test Along with the stress response test, the combined DEX/CRH challenge test (5 cumulative outliers, 10/11/8/11) was used to investigate changes in activation and cessation of the HPA-axis in aged animals (see Fig. 2B). Repeated measures ANOVA was able to show an overall effect of time (F2,72 ¼ 60.940, p < 0.05) and time  condition  treatment interaction (F2,72 ¼ 9.544, p < 0.05). No significant differences were found in basal evening CORT levels as well as in the CORT levels after DEX treatment. For the CRH stimulation, post hoc testing showed a trend for lower levels of CORT in stress animals compared to controls within the vehicle

treatment group. In the paroxetine treated group, stress animals showed a trend towards higher levels of CORT. In addition, the post hoc test showed a significant decrease of CORT in paroxetine treated animals compared to vehicle treated animals within the control condition. Analogue to the stress response test, the single steps were calculated resulting in dexamethasone-induced suppression (2 outliers, 12/11/9/11) as well as CRH-induced elevation (2 outliers, 12/11/9/11), as depicted in Table 2. While no overall effect was found in dexamethasone suppression, ANOVA revealed an effect of condition  treatment interaction (F1,38 ¼ 10.391, p < 0.05) in the CRH-induced corticosterone increase. In control animals, paroxetine treatment caused a decrease in CORT response compared to vehicle-treated animals. 3.2.4. Adrenal weights When investigating adrenal gland weights in the aged cohort (no outliers, 22/21/22/21), ANOVA revealed an effect of condition (F1,81 ¼ 4.093, p < 0.05), with a trend towards lower weight of stress animals within the vehicle-treated group, depicted in Fig. 2C. 3.2.5. Behavioural effects No significant effects were observed in the elevated plus maze (Supplemental Figure S2). 3.2.6. Gene expression of marker genes For the investigation of the influence of ageing on stressinduced gene expression, the GR (no outliers, 10/9/20/21) and MR mRNA levels (no outliers, 10/9/20/20; for the DG: 1 outlier, 9/9/20/ 20) were investigated again in the hippocampus, together with GR (1 outlier, 10/8/21/20), CRH (1 outlier, 10/8/21/20) and AVP (parvo: 1 outlier, 10/9/20/19; magno: 1 outlier, 9/8/19/20) in the PVN. In the hippocampus, ANOVA revealed a significant condition effect for the MR in the CA1, CA2 and CA3 regions and a trend in the DG region (CA1: F1,55 ¼ 4.694, p < 0.05; CA2: F1,55 ¼ 4.032, p < 0.05; CA3: F1,55 ¼ 5.217, p < 0.05; DG: F1,55 ¼ 2.932, p ¼ 0.093; see Fig. 3AeD) and a trend for a condition effect for the GR in the CA3 region (trend, F1,56 ¼ 3.297, p ¼ 0.075, see Fig. 3E and F). Post hoc testing did not show any significant effects for any of the investigated genes in the hippocampus. In the PVN, we found a condition effect for the GR (F1,55 ¼ 4.111, p < 0.05) as shown in Fig. 4A, with no significant post hoc effects. For CRH, ANOVA revealed a condition  treatment interaction effect (F1,55 ¼ 5.650, p < 0.05). Further post hoc testing showed that both the vehicle e as well as the paroxetine-treated animals showed decreased CRH expression 12 months after the stress exposure (see Fig. 4C). ANOVA also revealed a condition effect for AVP in the parvocellular part (F1,54 ¼ 5.248, p < 0.05) as well as a trend in the magnocellular part (F1,52 ¼ 2.967, p ¼ 0.091) of the PVN (see Fig. 4E). 4. Discussion

Table 2 Corticosterone dynamics in the stress response and the DEX/CRH test. Increase: difference of basal and 30 min values. Feedback: difference of 30 and 90 min values. Dex suppression: Difference between basal and DEX time point. CRH stimulation: difference between DEX and CRH time point. Data are given in mean  SEM. Corticosterone [ng/ml] Vehicle Control

Paroxetine Stress

Control

Stress

Stress response test Increase Feedback

161.7  35.0 203.6  25.3 289.7  51.9 86.6  23.7* 134.0  41.6 16.8  52.5 111.4  73.2 16.5  15.8

DEX/CRH test DEX suppression CRH stimulation

6.5  15.4 15.0  16.9 314.7  76.8 151.0  26.2

1.2  8.0 9.6  31.5 93.2  23.6* 251.1  54.2

*Significantly different to respective vehicle-treated animals p < 0.05.

In the present study, we subjected male adolescent mice to 7 weeks of chronic social stress. The resulting effects 5 weeks after stress in adult animals including elevated basal corticosterone levels, increased adrenal gland, decreased thymus glands and differences in central gene expression and behaviour were reversed to a major extent by 4 weeks of chronic treatment with the SSRI paroxetine. In aged animals that were investigated 12 months after the stress exposure, some of the stress effects were still present, for example CRH expression in the PVN, while other parameters had normalised. Paroxetine treatment was able to counteract the stress effects in some cases like the DEX/CRH test, but a long-term influence of the treatment by itself was present in many investigated parameters.

S.H. Scharf et al. / Neuropharmacology 72 (2013) 38e46

43

Fig. 2. HPA dynamics in aged animals, 12 months after the stress exposure. (A) Results from the stress response test in the morning. Regarding the basal levels, the vehicle-treated stress animals showed lower basal levels of CORT. We saw no group differences for the response (30 min) as well as the recovery (90 min), with the exception of paroxetine-treated stress animals, which showed significantly lower values at both time points. (B) Dex/CRH test performed in the evening. No differences were found for the basal levels as well as for the dex-induced suppression. For the CRH response, stressed animals showed a blunted response, which was also true for the paroxetine-treated control animals. Paroxetinetreated stress animals however were returned to control levels. (C) Normalised adrenal glands weight. Adrenal gland weight was lower in animals stressed during their adolescence. CON ¼ control, STR ¼ stress, VEH ¼ vehicle-treated animals, PAR ¼ paroxetine-treated animals. Data are given in mean þ SEM. *p < 0.05 vs control, #p < 0.1 vs control, yp < 0.05 vs vehicle-treated animals.

We showed consistent alterations following 5 weeks of recovery from a chronic social stress paradigm including elevated basal corticosterone levels, increased adrenal weight, decreased thymus weight, differences in central gene expression and impaired habituation to a novel environment in adult animals, all of which are parameters previously associated with chronic stress exposure (Park et al., 2001; Schmidt et al., 2007; Sterlemann et al., 2008; Strekalova et al., 2004; Ulrich-Lai et al., 2006; Wagner et al., 2011). Other parameters generally found to be altered directly following stress were found to have normalized, such as sociability or stresscoping behaviour. Interestingly, we were able to reverse all of the observed alterations via post hoc treatment with the SSRI paroxetine, except the decrease in thymus weight, which is possibly due to its lack of plasticity (Pearse, 2006). Previous studies with the same paradigm showed that stress-induced alterations can be blocked via treatment during the stress phase (Schmidt et al., 2007). The present data now support the notion that not only the time during the stressful episode is important for the development of (mal) adaptive modulations, but also the time window after the stress exposure, a time characterised by coping strategies and recovery. In addition, this shows that antidepressant drugs can reverse already developed alterations via an active process and are not only efficient in suppressing the development of maladaptions via passive blocking. This strongly suggests that the therapeutic window of opportunity to correct stress-induced pathologies stretches from during the stress exposure itself to the time following the stress exposure and that treatment after the stress exposure e closely mimicking the clinical situation e seems to be equally effective. The majority of effects we found in the present study were consistent with previous studies and the literature in the field. Nevertheless, some aspects found in previous studies, like reduced levels of MR mRNA were absent in the present study (Schmidt et al., 2007; Sterlemann et al., 2008). One possible explanation is the heterogeneous genetic background of the animals (outbred CD1 mice) used in our studies. Recent publications have shown that the genetic background and individual differences can play a strong role in the response to and the recovery from stress, up to the point that stress-resilient animals cannot be distinguished from control animals in multiple parameters (Schmidt et al., 2010a,b). Therefore, the composition of the stress group used for the study and the ratio of stress-resilient to stress-vulnerable animals can significantly influence the results and cause different outcomes in various tests.

We also investigated aged animals (15 months of age), 12 months after the end of the stress exposure. Here, we found a general suppression of the HPA axis in stressed animals under low CORT conditions, evident by the decreased levels of basal CORT in the morning in combination with lower adrenal gland weights. However, we found no differences in basal CORT levels in the evening. Studies investigating effects of chronic stress exposure show elevated levels of CORT in the majority of cases (Bartolomucci et al., 2001; Rice et al., 2008; Schmidt et al., 2007). However, these studies mostly focus on adult animals and do not integrate the influence of ageing. In addition, the majority of chronic stress studies only show the effects directly after cessation of the stress. What we see here might be a compensatory mechanism caused by the chronic stress exposure, reflected in a consistent fashion in the lower morning basal levels of CORT, the lower expression of CRH in the PVN as well as the lower adrenal gland weights. Although ageing often leads to an increase in basal CORT levels, it has been shown that a subgroup of aged individuals actually showed a decrease in basal CORT (Lupien et al., 1998). Another hypothesis might be accelerated (or sometimes called pathological) ageing. Pioneered by Sapolsky and colleagues (reviewed in Sapolsky, 1999), this hypothesis states that frequent exposure to stress or chronically elevated levels of glucocorticoids in general can speed up the natural deterioration of physiological systems (including the HPA axis). Therefore, we might observe a more strongly deteriorated state of the HPA axis in our stressed animals. Another intriguing finding was the potential change in the interplay of CRH and AVP, which act in synergy on the final production and release of ACTH (Hatzinger et al., 2000; Torpy et al., 1994). This can for example been seen in the different outcomes comparing the stress response as well as the DEX/CRH test. Here, we were able to show that in the PVN of formerly stressed mice, CRH mRNA was strongly downregulated while AVP mRNA was slightly increased. This suggests a shift to the influence of vasopressin in these animals, which can also be observed in the stress challenge tests: In the stress response test, which should activate both AVP and CRH, the CORT response of the stressed animals was comparable to those of control animals. However, in the DEX/CRH test, in which exclusively CRH is used to elicit a response, the downstream secretion of CORT was significantly lower in stressed animals, suggesting the decreased efficacy of CRH signalling in these animals. Multiple studies showed the high importance of AVP for the adaption of the HPA axis following

44

S.H. Scharf et al. / Neuropharmacology 72 (2013) 38e46

Fig. 3. Central gene expression in the hippocampus of aged animals (12 months after stress). (A)e(C) Relative expression of the MR in the CA1, CA2 and CA3 region of the hippocampus respectively. (D) Representative images from the MR. (E) Relative expression of the GR in the CA3 region of the hippocampus. (F) Representative images of GR expression in the hippocampus. CON ¼ control, STR ¼ stress, VEH ¼ vehicletreated animals, PAR ¼ paroxetine-treated animals. Data are given in mean þ SEM.

chronic stress (reviewed in Aguilera and Rabadan-Diehl, 2000), which might also be the case for the lasting effects of chronic stress during ageing. In addition, while vasopressin seems to be more important for CORT induction, CRH seems to influence basal levels. However, this effect is more prominent in aged individuals, as we were not able to observe this coupling in adult animals. Previous studies with the same paradigm showed no significant differences in CRH expression in adult animals (Schmidt et al., 2007) combined with higher levels of CORT, suggesting that in this specific model, the role of CRH in adult animals is limited, at least in regard to basal levels, which might change in interplay with ageing.

Fig. 4. Central gene expression in the PVN of aged animals (12 months after stress). (A) Relative expression of the GR. (B) Representative images from the GR. (C) Relative expression of CRH mRNA in the PVN. Stress animals showed a reduction of CRH mRNA in both the vehicle e as well as the paroxetine-treated group. (D) Representative images of CRH expression. (E) Relative expression of AVP in the parvocellular part of the PVN. (F) Representative colour-reversed dark-field micrographs for AVP. CON ¼ control, STR ¼ stress, VEH ¼ vehicle-treated animals, PAR ¼ paroxetine-treated animals. Data are given in mean þ SEM. *p < 0.05 vs control.

Another intriguing finding of the study is the strong influence of the paroxetine treatment on some parameters in the aged animals. Control animals treated with paroxetine for example show a strongly blunted response to CRH stimulation without any prior stress exposure. It has been shown in aged rats that treatment with an SSRI can reduce the response to an acute stressor (Handa et al., 1993), which fits our observations here, although the effect we see

S.H. Scharf et al. / Neuropharmacology 72 (2013) 38e46

is much stronger. At the same time, paroxetine treatment did interact with the stress exposure in the stress response test to induce a blunted response, while the vehicle-treated stress animals did not show differences to control animals. This highlights that although paroxetine treatment did reverse most of the stress effect in adult animals, we also see some effects of paroxetine alone and interaction with stress and ageing, which is still poorly understood. This shift and the dynamics of the HPA axis in general can also be used to illustrate the influence of antidepressant treatment in combination with both ageing and adverse experience. Treatment with paroxetine in control animals did not influence the stress response, but also induced a shift to a decreased influence of CRH, quite similar to the effects of the stress exposure. However, when the SSRI was given after a stressful episode, the CRH response was restored to control levels and the stress response was lower. This has two important implications: (1) Treatment with paroxetine after stress is able to prevent long-term effects even in aged individuals and (2) Treatment with SSRIs has long lasting influences of its own. Whether this influence is beneficial or maladaptive strongly depends on the situation of the individual. However our current data demonstrates that the long-term effects of SSRI treatment, especially in interplay with ageing, are still largely unclear and that additional studies are necessary to elucidate the complex long-term effects of these drugs. The present study has several limitations. We cannot rule out the possibility that the single housing of the animals during the recovery period has effects of its own. These however should be the same for all experimental groups as all animals are single-housed for the same periods of time. Further, our study at this point only investigates mRNA levels. Follow-up studies should include the investigation of protein levels to get a clearer picture how changes in mRNA levels are translated into changes in protein expression. Finally, behavioural assessment in the aged population has to be interpreted with caution, as result may be confounded by a lower physical fitness of the animals (e.g. high body weight, potentially lower eye sight, etc.). Taken together, the present study consistently demonstrated the efficacy of paroxetine to actively reverse effects caused by adolescent stress exposure up to a late age. Furthermore, we were able to show that paroxetine treatment by itself effects HPA axis dynamics and can induce a shift to a more prominent role of AVP in the stress response. Disclaimer The authors declare that there are no conflicts of interest. This study was funded by the Max Planck Society (MPS). The MPS had no role in study design, data collection and analysis, decision to publish or preparation of the manuscript. Appendix A. Supplementary information Supplementary material associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j. neuropharm.2013.03.035. References Abraham, N., Brunner, E., Eriksson, J., Robertson, R., 2007. Metabolic syndrome: psychosocial, neuroendocrine, and classical risk factors in type 2 diabetes. Annals of the New York Academy of Sciences 1113, 256e275. Aguilera, G., Rabadan-Diehl, C., 2000. Vasopressinergic regulation of the hypothalamicepituitaryeadrenal axis: implications for stress adaptation. Regulatory Peptides 96, 23e29. Arndt, S.S., Laarakker, M.C., van Lith, H.A., van der Staay, F.J., Gieling, E., Salomons, A.R., van’t Klooster, J., Ohl, F., 2009. Individual housing of mice e impact on behaviour and stress responses. Physiology & Behavior 97, 385e393.

45

Bartolomucci, A., Palanza, P., Gaspani, L., Limiroli, E., Panerai, A.E., Ceresini, G., Poli, M.D., Parmigiani, S., 2001. Social status in mice: behavioral, endocrine and immune changes are context dependent. Physiology & Behavior 73, 401e410. Bartolomucci, A., Palanza, P., Sacerdote, P., Ceresini, G., Chirieleison, A., Panerai, A.E., Parmigiani, S., 2003. Individual housing induces altered immuno-endocrine responses to psychological stress in male mice. Psychoneuroendocrinology 28, 540e558. Carboni, L., Becchi, S., Piubelli, C., Mallei, A., Giambelli, R., Razzoli, M., Mathé, A.A., Popoli, M., Domenici, E., 2010. Early-life stress and antidepressants modulate peripheral biomarkers in a gene-environment rat model of depression. Progress in Neuro-Psychopharmacology & Biological Psychiatry 34, 1037e1048. Caspi, A., Sugden, K., Moffitt, T.E., Taylor, A., Craig, I.W., Harrington, H., McClay, J., Mill, J., Martin, J., Braithwaite, A., Poulton, R., 2003. Influence of life stress on depression: moderation by a polymorphism in the 5-HTT gene. Science 301, 386e389. Charney, D.S., Manji, H.K., 2004. Life stress, genes, and depression: multiple pathways lead to increased risk and new opportunities for intervention. Science Signaling, re5. Dalm, S., Enthoven, L., Meijer, O.C., van der Mark, M.H., Karssen, A.M., de Kloet, E.R., Oitzl, M.S., 2005. Age-related changes in hypothalamicepituitaryeadrenal axis activity of male C57BL/6J mice. Neuroendocrinology 81, 372e380. Dayas, C.V., Buller, K.M., Crane, J.W., Xu, Y., Day, T.A., 2001. Stressor categorization: acute physical and psychological stressors elicit distinctive recruitment patterns in the amygdala and in medullary noradrenergic cell groups. European Journal of Neuroscience 14, 1143e1152. de Kloet, E.R., Joels, M., Holsboer, F., 2005. Stress and the brain: from adaption to disease. Nature Reviews Neuroscience 6, 463e475. El Khoury, A., Gruber, S.H.M., Mork, A., Mathé, A.A., 2006. Adult life behavioral consequences of early maternal separation are alleviated by escitalopram treatment in a rat model of depression. Progress in Neuro-Psychopharmacology & Biological Psychiatry 30, 535e540. Fluttert, M., Dalm, S., Oitzl, M.S., 2000. A refined method for sequential blood sampling by tail incision in rats. Laboratory Animals 34, 372e378. Gerritsen, L., Geerlings, M.I., Beekman, A.T.F., Deeg, D.J.H., Penninx, B.W.J.H., Comijs, H.C., 2010. Early and late life events and salivary cortisol in older persons. Psychological Medicine 40, 1569e1578. Handa, R.J., Cross, M.K., George, M., Gordon, B.H., Burgess, L.H., Cabrera, T.M., Hata, N., Campbell, D.B., Lorens, S.A., 1993. Neuroendocrine and neurochemical responses to novelty stress in young and old male F344 rats: effects of d-fenfluramine treatment. Pharmacology Biochemistry and Behavior 46, 101e109. Hatzinger, M., Wotjak, C.T., Naruo, T., Simchen, R., Keck, M.E., Landgraf, R., Holsboer, F., Neumann, I.D., 2000. Endogenous vasopressin contributes to hypothalamicepituitaryeadrenocortical alterations in aged rats. Journal of Endocrinology 164, 197e205. Heuser, I., Yassouridis, A., Holsboer, F., 1994. The combined dexamethasone/CRH test: a refined laboratory test for psychiatric disorders. Journal of Psychiatric Research 28, 341e356. Huot, R., Thrivikraman, Meaney M., Plotsky, P., 2001. Development of adult ethanol preference and anxiety as a consequence of neonatal maternal separation in Long Evans rats and reversal with antidepressant treatment. Psychopharmacology 158, 366e373. Joels, M., Baram, T.Z., 2009. The neuro-symphony of stress. Nature Reviews Neuroscience 10, 459e466. Joels, M., Pu, Z., Wiegert, O., Oitzl, M.S., Krugers, H.J., 2006. Learning under stress: how does it work? Trends in Cognitive Sciences 10, 152e158. Keck, M.E., Welt, T., Müller, M.B., Uhr, M., Ohl, F., Wigger, A., Toschi, N., Holsboer, F., Landgraf, R., 2003. Reduction of hypothalamic vasopressinergic hyperdrive contributes to clinically relevant behavioral and neuroendocrine effects of chronic paroxetine treatment in a psychopathological rat model. Neuropsychopharmacology 28, 235e243. Lupien, S.J., de, L.M., de, S.S., Convit, A., Tarshish, C., Nair, N.P., Thakur, M., McEwen, B.S., Hauger, R.L., Meaney, M.J., 1998. Cortisol levels during human aging predict hippocampal atrophy and memory deficits. Nature Neuroscience 1, 69e73. Lupien, S.J., Lepage, M., 2001. Stress, memory, and the hippocampus: can’t live with it, can’t live without it. Behavioural Brain Research 127, 137e158. Lupien, S.J., McEwen, B.S., Gunnar, M.R., Heim, C., 2009. Effects of stress throughout the lifespan on the brain, behaviour and cognition. Nature Reviews Neuroscience 10, 434e445. Marco, E.M., Macrì, S., Laviola, G., 2011. Critical age windows for neurodevelopmental psychiatric disorders: evidence from animal models. Neurotoxicity Research 19, 286e307. McCormick, C.M., Smith, C., Mathews, I.Z., 2008. Effects of chronic social stress in adolescence on anxiety and neuroendocrine response to mild stress in male and female rats. Behavioural Brain Research 187, 228e238. McEwen, B.S., 1998. Stress, adaptation, and disease: allostasis and allostatic load. Annals of the New York Academy of Sciences 840, 33e44. Mueller, S.C., Maheu, F.S., Dozier, M., Peloso, E., Mandell, D., Leibenluft, E., Pine, D.S., Ernst, M., 2010. Early-life stress is associated with impairment in cognitive control in adolescence: an fMRI study. Neuropsychologia 48, 3037e3044. Navailles, S., Zimnisky, R., Schmauss, C., 2010. Expression of glucocorticoid receptor and early growth response gene 1 during postnatal development of two inbred strains of mice exposed to early life stress. Developmental Neuroscience 32, 139e148. Otte, C., Hart, S., Neylan, T.C., Marmar, C.R., Yaffe, K., Mohr, D.C., 2005. A metaanalysis of cortisol response to challenge in human aging: importance of gender. Psychoneuroendocrinology 30, 80e91.

46

S.H. Scharf et al. / Neuropharmacology 72 (2013) 38e46

Park, C.R., Campbell, A.M., Diamond, D.M., 2001. Chronic psychosocial stress impairs learning and memory and increases sensitivity to yohimbine in adult rats. Biological Psychiatry 50, 994e1004. Pearse, G., 2006. Histopathology of the thymus. Toxicologic Pathology 34, 515e547. Rice, C., Sandman, C.A., Lenjavi, M.R., Baram, T.Z., 2008. A novel mouse model for acute and long-lasting consequences of early life stress. Endocrinology 149, 4892e4900. Rosengren, A., Hawken, S., Ounpuu, S., Sliwa, K., Zubaid, M., Almahmeed, W.A., Blackett, K.N., Sitthi-amorn, C., Sato, H., Yusuf, S., 2004. Association of psychosocial risk factors with risk of acute myocardial infarction in 11119 cases and 13648 controls from 52 countries (the INTERHEART study): caseecontrol study. The Lancet 364, 953e962. Rosner, B., 1983. Percentage points for a generalized ESD many-outlier procedure. Technometrics 25, 165e172. Sapolsky, R.M., 1999. Glucocorticoids, stress, and their adverse neurological effects: relevance to aging. Experimental Gerontology 34, 721e732. Schmidt, M.V., Scharf, S.H., Sterlemann, V., Ganea, K., Liebl, C., Holsboer, F., Müller, M.B., 2010a. High susceptibility to chronic social stress is associated with a depression-like phenotype. Psychoneuroendocrinology 35, 635e643. Schmidt, M.V., Sterlemann, V., Ganea, K., Liebl, C., Alam, S., Harbich, D., Greetfeld, M., Uhr, M., Holsboer, F., Müller, M.B., 2007. Persistent neuroendocrine and behavioral effects of a novel, etiologically relevant mouse paradigm for chronic social stress during adolescence. Psychoneuroendocrinology 32, 417e429.

Schmidt, M.V., Trümbach, D., Weber, P., Wagner, K.V., Scharf, S.H., Liebl, C., Datson, N., Namendorf, C., Gerlach, T., Kühne, C., Uhr, M., Deussing, J.M., Wurst, W., Binder, E.B., Holsboer, F., Müller, M.B., 2010b. Individual stress vulnerability is predicted by short-term memory and AMPA receptor subunit ratio in the hippocampus. The Journal of Neuroscience: The Official Journal of the Society for Neuroscience 30, 16949e16958. Sterlemann, V., Ganea, K., Liebl, C., Harbich, D., Alam, S., Holsboer, F., Müller, M.B., Schmidt, M.V., 2008. Long-term behavioral and neuroendocrine alterations following chronic social stress in mice: Implications for stress-related disorders. Hormones and Behavior 53, 386e394. Strekalova, T., Spanagel, R., Bartsch, D., Henn, F.A., Gass, P., 2004. Stress-induced anhedonia in mice is associated with deficits in forced swimming and exploration. Neuropsychopharmacology 29, 2007e2017. Torpy, D.J., Grice, J.E., Hockings, G.I., Walters, M.M., Crosbie, G.V., Jackson, R.V., 1994. Alprazolam attenuates vasopressin-stimulated adrenocorticotropin and cortisol release: evidence for synergy between vasopressin and corticotropin-releasing hormone in humans. Journal of Clinical Endocrinology & Metabolism 79, 140e144. Ulrich-Lai, Y.M., Figueiredo, H.F., Ostrander, M.M., Choi, D.C., Engeland, W.C., Herman, J.P., 2006. Chronic stress induces adrenal hyperplasia and hypertrophy in a subregion-specific manner. AJP e Endocrinology and Metabolism 291, E965eE973. Wagner, K.V., Wang, X.-D., Liebl, C., Scharf, S.H., Müller, M.B., Schmidt, M.V., 2011. Pituitary glucocorticoid receptor deletion reduces vulnerability to chronic stress. Psychoneuroendocrinology 36, 579e587.