INTRODUCTION

Individuals are frequently challenged by stressful events that can trigger the activation of hormonal pathways such as the hypothalamic–pituitary–adrenal (HPA) axis (Chrousos, 2009). Prolonged activation of these systems by chronic stress results in persistently elevated cortisol levels that, in turn, can lead to maladaptive consequences in the organism and may ultimately contribute to the development of psychiatric disorders such as depression (de Kloet et al, 2005; McEwen, 2004). Animal models of chronic stress exposure are a valuable tool to further our understanding of the molecular underpinnings of stress-induced psychopathology (Savignac et al, 2011; Cryan and Holmes, 2005; Joëls and Baram, 2009), as well as providing a paradigm to assess and validate current and novel treatment strategies for depression (Wagner et al, 2012; Mutlu et al, 2012; Scharf et al, 2013).

Most present treatment options for depression are based on the monoamine hypothesis, and aim to increase the availability of monoamines, such as serotonin, in the synaptic cleft (Prins et al, 2011; Rush et al, 2006). However, the late onset of therapeutic effects as well as unsatisfactory relapse rates and side effects illustrates the need for improved therapeutics (Thase, 2006). Recent studies have provided convincing evidence that dysregulation of glutamate signaling, mainly via its postsynaptic receptors α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA), N-methyl-D-aspartate (NMDA), and metabotropic glutamate receptors (mGluRs), contributes to the emergence of psychiatric disorders (Kendell et al, 2005; Sanacora et al, 2012; Mathews et al, 2012; Yim et al, 2012; Tronson et al, 2010). Modulation of glutamate receptor function has therefore been proposed as a promising target for antidepressant, anxiolytic, and antipsychotic drug development (Popoli et al, 2012). Positive and negative modulators of mGluR subtype 5 (mGluR5), in particular, have been suggested as novel agents for the treatment of depression (Palucha et al, 2005; Krystal et al, 2010; Pilc et al, 2008), but the exact molecular mechanisms that mediate their potential therapeutic effects are yet to be fully understood.

In this context, Homer1 has emerged as a potential target protein in depression. Homer1 is a postsynaptic scaffolding protein that links mGluR5 to downstream targets such as inositol triphosphate receptors (Tu et al, 1998). Homer1 also acts as a moderator of the NMDA/mGluR5 complex (Tu et al, 1999; Bertaso et al, 2010) that is highly implicated in stress-induced neuropsychiatric pathologies. Interestingly, the two main splice variants Homer1a and Homer1b/c appear to have opposite molecular effects (Brakeman et al, 1997). Homer1b/c links the mGluR5 to the intracellular signaling machinery and mediates ligand-dependent activity of mGluR5. In contrast, the early immediate gene Homer1a is thought to act as dominant negative isoform disrupting mGluR5/Homer1b/c coupling and predominantly modulates ligand-independent mGluR5 signaling (Ango et al, 2001). Clinical studies provided first evidence that Homer1 is involved in the development of major depressive disorders (Rietschel et al, 2010), whereas preclinical studies describe its importance in anxiety- and depression-related behavior (Szumlinski et al, 2005; Lominac et al, 2005), memory formation (Lominac et al, 2005), fear (Tronson et al, 2010), and reward-related behaviors (Jaubert et al, 2007; Szumlinski et al, 2004). Furthermore, the activity-induced splice variant Homer1a (Brakeman et al, 1997) has been shown to be crucially involved in behavioral alterations that are related to depression (Celikel et al, 2007; Mahan et al, 2012) and anxiety (Lominac et al, 2005). Moreover, prenatal stress was shown to alter Homer1a and Homer1b/c expression in several corticolimbic structures, including the hippocampus (Ary et al, 2007). However, the impact of Homer1 and its modulatory effects on glutamate signaling, particularly via mGluR5, in chronic stress situations is largely unknown.

In this study, we therefore aimed to investigate the role of Homer1 and mGluR5 in the context of chronic social defeat stress (CSDS) that has been shown to model relevant endophenotypes of depression by us and others (Wang et al, 2011; Hartmann et al, 2012; Nestler and Hyman, 2010; Berton et al, 2006). We hypothesized that stress-induced modulation of Homer1 expression and Homer1/mGluR5 interaction may shape the behavioral and neuroendocrine responses as well as vulnerability to chronic social stress. To test this hypothesis, we used conventional Homer1 knockout (KO) mice, virus-induced overexpression of the immediate early gene Homer1a in the murine hippocampus, as well as a targeted pharmacological approach by treatment with 2-chloro-4-((2,5-dimethyl-1-(4-(trifluoromethoxy)phenyl)-1H-imidazol-4-yl)ethynyl)pyridine (CTEP), a novel mGluR5 inverse agonist (Lindemann et al, 2011), in combination with CSDS.

MATERIALS AND METHODS

Further detailed methods descriptions can be found in the Supplementary Material.

Animals

For all experiments, male C57Bl/6N mice (Charles River Laboratories, Maastricht, The Netherlands) at the age of 12 weeks were used unless noted otherwise. Conventional Homer1 KO and wild-type (WT) littermates were bred from heterozygous breeding pairs on a C57BL/6N background in the animal facilities of the Max Planck Institute of Psychiatry in Munich, Germany. Generation and genotyping of Homer1 KO mice has been reported previously (Yuan et al, 2003) and Homer1 knockout was verified by PCR. All mice were held under standard conditions (12 h light/12 h dark cycle, lights on at 08:00 h, temperature 23±2 °C) and were single-housed and acclimatized to the experimental room for 2 weeks before the beginning of the experiments. Male CD1 mice (16–18 weeks of age) served as resident mice and were held under the conditions described above. They were allowed to habituate to the social defeat cage for 2 weeks before the experiment. Tap water and food (Altromin 1324, Altromin GmbH, Germany) was available ad libitum. The experiments were carried out in accordance with the European Communities’ Council Directive 2010/63/EU. The protocols were approved by the committee for the Care and Use of Laboratory animals of the Government of Upper Bavaria, Germany.

Experimental Design

For all chronic social defeat experiments, a separate batch of animals was used (n=8–12 per group). Male Homer1 WT (n=26) and KO (n=22) animals were randomly distributed across control and stress conditions according to their genotype. For all other experiments including a specific treatment, a total of 48 mice were randomly split into 2 × 2 groups (control vs stress, control vs treatment). The CSDS paradigm lasted for 21 days and was conducted as previously described (Wagner et al, 2011). See Supplementary Material for details.

Experiment 1

To assess the impact of stress on the Homer1 system, we subjected animals to CSDS and measured the expression of the splice variants Homer1a and Homer1b/c in the hippocampus in two separate cohorts of mice. We also investigated possible effects of the stress exposure on Homer1 protein turnover and Homer1/mGluR5 coupling.

Experiment 2

We proceeded to investigate the overall role of Homer1—independent of the specific splice variants and the affected brain region—by submitting conventional Homer1 KO mice to CSDS. We assessed the effects of the stress procedure on physiological, neuroendocrine, and brain gene expression parameters. In addition, the animals were tested for locomotor activity, social behavior, hedonic behavior, and stress coping.

Experiment 3

Next, we asked whether we could block the effects of chronic stress exposure by pharmacological modulation of the mGluR5 signaling pathway. We therefore applied the novel mGluR5 inverse agonist CTEP (Lindemann et al, 2011) during the stress procedure that should counteract the ligand-dependent activity of mGluR5.

Experiment 4

Finally, we explored the possibility that the stress-induced neuroendocrine and behavioral effects could be enhanced by increasing the availability of the splice variant Homer1a in the hippocampus that has previously been shown to trigger ligand-independent mGluR5 activity (Ango et al, 2001).

Drug Treatment

Oral administration of the inverse mGluR5 agonist CTEP (F. Hoffmann-La Roche, Basel, Switzerland) commenced 7 days before the start of the CSDS paradigm to establish stable baseline receptor occupancy and was continued until the end of the experiment. Treatment by CTEP was performed as described previously (Lindemann et al, 2011; Michalon et al, 2012). CTEP was formulated as a microsuspension in vehicle (0.9% NaCl, 0.3% Tween-80). Chronic treatment consisted of one dose per 48 h of 2 mg/kg body weight per os (p.o.) in a volume of 10 ml/kg body weight. Gavaging took place immediately before the daily defeat or handling procedure to minimize confounding effects of oral drug administration.

Viral Overexpression of Homer1a

Viral overexpression was performed as previously described (Wagner et al, 2013). A detailed description of the procedure can be found in the Supplementary Materials section. Successful targeting and quantification of Homer1a overexpression was achieved by in situ hybridization using the riboprobe described below (Supplementary Figure S1). Animals that were not infected bilaterally in both the CA1 and DG regions were excluded from the analysis (n=3).

Behavioral Testing

The following behavioral tests were performed between 0800 and 1200 h in the same room where the animals were housed: open field test (OF), social avoidance test (SA), female urine sniffing test (FUST), and forced swim test (FST). All tests were described and validated previously (Wagner et al, 2011; Wagner et al, 2012; Malkesman et al, 2010). Tests were recorded and analyzed using the video tracking software ANY-maze (ANY-maze 4.20, Stoelting, IL). A detailed description of all behavioral tests performed can be found in the Supplementary Material.

In Situ Hybridization

Frozen brains were sectioned at −20 °C in a cryostat microtome at 18 or 20 μm (for fixated brains), thaw mounted on Super Frost Plus slides, dried, and stored at −80 °C. In situ hybridization using 35S UTP=labeled ribonucleotide probes (Homer1a, Homer1b/c, Homer2a/b, mGluR5, and CRH) was performed as described previously (Wagner et al, 2013; Schmidt et al, 2007). A more detailed description of the procedure can be found in the Supplementary Materials.

RNA Processing

RNA was isolated from whole hippocampi using the TRIZOL reagent (Invitrogen) as previously described (Schmidt et al, 2010). The quality of the RNA was assessed using an Agilent 2100 Bioanalyzer (Agilent Technologies, Palo Alto, CA). The concentration and purity of total RNA was also assessed by 260 nm UV absorption and by 260/280 ratios, respectively (Nanophotometer, Implen, Munich, Germany). All samples had an RNA integrity number ≥7 (range 7.0–8.9; mean 8.0±SD 0.4).

Quantitative Reverse Transcriptase-PCR

RNA samples were transcribed into cDNA applying a High-Capacity cDNA Reverse Transcription Kit (Applied Biosystems) following the manufacturer’s protocol. The qPCR of 100 ng cDNA per sample was performed using the Quantifast SYBR Green PCR Kit (Qiagen) and the Lightcycler 2.0 (Roche) according to the standard protocols given in the manufacturers’ manuals. All samples were normalized to the housekeeping gene glyceraldehyde 3-phosphate dehydrogenase (GAPDH).

Co-Immunoprecipitation

For co-immunoprecipitation (Co-IP), hippocampal tissues from two animals of the same experimental group were pooled. Preparation of membrane fractions and Co-IP analysis was performed as described previously (Wagner et al, 2013; Wagner et al, 2012). In short, membrane fractions were isolated using the Calbiochem Proteoextract kit (EMD Biosciences), protein concentration was determined, and 1.2 mg of lysate was incubated with 2.5 μg mGluR5 antibody (Millipore) overnight at 4 °C. Then, 20 μl of BSA-blocked Protein G Dynabeads (Invitrogen, 100-03D) was added to the lysate–antibody mix followed by 3 h of incubation at 4 °C. The beads were washed 3 times with PBS and protein–antibody complexes were eluted with 100 μg/ml mGluR5-peptide solution (Millipore) in Co-IP buffer for 30 min at 4 °C. Then, 15 μg of the cell lysates and 10 μl of the immunoprecipitates were further processed by western blot analysis (Wang et al, 2011). Antibodies used were rabbit anti-Homer1 (1 : 1000, Synaptic Systems), rabbit-anti mGluR5, and goat anti-actin (1 : 2000, Santa Cruz Biotechnology) for primary as well as horseradish peroxidase-conjugated secondary antibodies (1 : 2000, DAKO).

Quantitative Protein Turnover Analysis

Protein turnover was analyzed as described previously (Webhofer et al, 2013). For details, see Supplementary Methods.

Statistical Analysis

All data presented are shown as means+SEM, and analyzed by SPSS 18.0. Two-tailed Student’s t-test was employed for comparison of two independent groups (control vs CSDS). Two-factorial ANOVA was employed when appropriate. Significant interaction effects were followed by Fisher’s LSD post hoc analysis when appropriate. As nominal level of significance, P<0.05 was accepted.

RESULTS

Hippocampal Homer1 Is Regulated by Chronic Stress

We measured Homer1b/c mRNA levels in response to CSDS by in situ hybridization and found an increase in both the CA1 (T18=−3.275, p<0.01) and CA3 (T18=−4.556, p<0.001), but not the dentate gyrus (DG) (T18=−1.791, p=0.090), regions of the dorsal hippocampus (Figure 1a and c). On the other hand, Homer1a was not significantly regulated in any of the investigated regions of the hippocampus (Figure 1b and c), and this is in line with the idea that this splice variant is an immediate early gene predominantly activated after an acute challenge (Ango et al, 2001). The Homer1b/c upregulation was replicated in an independent batch of animals (Supplementary Figure S2A). The physiological and behavioral parameters of these experiments have been reported before (Wang et al, 2011; Wagner et al, 2011). To further validate these transcriptional alterations, we performed RT-PCR in a third batch of animals that underwent the same CSDS paradigm, and confirmed our findings (Supplementary Figure S2B). Next, we also investigated the protein turnover rate of Homer1 in the synaptosomal fraction of the hippocampus that could indirectly affect Homer1 protein availability. Although 30% of Homer1 protein was metabolized over the course of 7 days (Supplementary Figure S2C), no difference in protein turnover between CSDS and control animals was found. Finally, we tested whether CSDS might affect physical interaction of Homer1 with mGluR5. Interestingly, we observed that chronic stress reduced coupling of Homer1 to mGluR5, without altering absolute protein levels (Figure 1d).

Figure 1
figure 1

Hippocampal Homer1 mRNA is regulated by chronic stress. (a) Chronic social defeat stress (CSDS) led to increased Homer1b/c mRNA levels in the CA1 and CA3 regions of the dorsal hippocampus. (b) The mRNA levels of the immediate early gene Homer1a were not significantly altered at the time of killing 24 h after the last defeat session of the CSDS paradigm. (c) Representative radiograph pictures of Homer1b/c and Homer1a expression. (d) Binding of Homer1 to its primary interaction partner mGluR5 is reduced in stressed animals (T9=9.429, p<0.001). *Significant from control, p<0.05.

PowerPoint slide

Deletion of Homer1 Leads to Disturbed HPA Axis Function and Behavioral Hyperactivity

To further elucidate the role of Homer1 in chronic stress-induced behavioral and neuroendocrine alterations, we exposed Homer1 KO animals to the CSDS paradigm. We identified a large increase in the corticosterone response to a novel acute stressor (FST) in KO compared with WT animals, which was apparent in both stressed and nonstressed mice (Figure 2a). This was further reflected by significantly enlarged adrenal glands in these animals (Figure 2b), but not in basal circulating corticosterone levels (Supplementary Figure S3A and B). Interestingly, this neuroendocrine phenotype was accompanied by gene transcription alterations in the paraventricular nucleus (PVN) of the hypothalamus. Here, CSDS induced an increase in CRH mRNA in the PVN irrespective of the genotype (Figure 2c). Concurrently, KO animals showed significantly reduced CRH mRNA levels compared with their WT littermates. We also checked for potential compensatory changes in the Homer1 KO mice on the level of mGluR5 and Homer2a/b expression in the hippocampus. For mGluR5 (Supplementary Figure S3C), CSDS led to a mild increase in expression in the CA1 region in WT, but not in KO animals. However, there was no main effect of genotype in any of the hippocampal regions, arguing against a compensatory effect due to the lack of Homer1 expression. Regarding the Homer2a/b expression (Supplementary Figure S3D) we detected a modest decrease in Homer1 KO mice that was restricted to the CA1 region. In addition, chronic stress exposure tended to reduce Homer2a/b expression, an effect that was significant in the CA3 region.

Figure 2
figure 2

Neuroendocrine and behavioral profile of Homer1 KO exposed to CSDS. (a) The corticosterone response to a novel acute stressor was increased in mice that were exposed to CSDS. At the same time, KO mice showed a strongly enhanced response to the stressor irrespective of the condition (ANOVA main effects: condition: F1, 44=43.232, p<0.001; genotype: F1, 44=15.796, p<0.001). (b) Analogous to the corticosterone response, CSDS induced an increase in adrenal glands weight in both genotypes (main condition effect: F1, 45=82.163, p<0.001). However, KO mice already possessed enlarged adrenal glands under basal condition and this effect was also present in the stressed group (main genotype effect: F1, 45=81.040, p<0.001). (c) In the paraventricular nucleus of the hypothalamus (PVN), CRH expression was increased in response to CSDS (main condition effect: F1, 43=11.264, p<0.01). At the same time, KO animals showed significantly lower CRH levels (main genotype effect: F1, 43=10.744, p<0.01). See representative radiograph pictures in the panel below. (d) Although CSDS resulted in a reduction of locomotion in both groups (main condition effect: F1, 43=13.129), a marked hyperactivity was detected in KO mice under basal and stress conditions (main genotype effect: F1, 43=12.630, p<0.01). (e) In the social avoidance (SA) test, CSDS resulted in a reduction of interaction in WT but not in KO mice (ANOVA genotype main effect: F1, 37=5.337, p<0.05; interaction effect: F1, 37=4.644, p<0.05), showing a significantly increased interaction ratio compared with their stressed WT littermates. (f) In the water trial of the female urine sniffing test (FUST), animals with a Homer1 deletion showed a reduced interest in the presented cotton tip (main genotype effect: F1, 45=9.269, p<0.01). In the urine trial, CSDS had a strong effect on sniffing time (main condition effect: F1, 45=9.185, p<0.01) where sniffing time was reduced irrespective of the genotype. (g) Independent of CSDS, KO mice exhibited their hyperactive phenotype in the forced swim test (FST) as indicated by reduced floating times (main genotype effect: F1, 45=32.662, p<0.001). *Significant main condition effect, p<0.05; #significant main genotype effect, p<0.05; +significant from control of same genotype, p<0.05; §significant from WT of same condition.

PowerPoint slide

Deletion of Homer1 also resulted in considerable changes in the animals’ behavior. In the OF test, CSDS reduced total locomotion in both genotypes, whereas KO animals displayed greatly increased locomotor activity irrespective of the stress condition (Figure 2d). In the SA test, WT animals displayed a reduced social interaction ratio when under the effects of CSDS (p<0.05), whereas interaction ratios of stressed KO mice were not significantly reduced compared with control conditions (p=0.256; Figure 2e). A strong genotype difference was also apparent in the FUST, where KO mice spent significantly less time with the presented cotton tip in the water trial (Figure 2f). In the urine trial, we detected a reduction in sniffing time of stressed mice compared with their nonstressed littermates, irrespective of the genotype. The above-mentioned hyperactive phenotype was also visible in the FST, where KO mice spent significantly less time floating (Figure 2g).

Chronic Decrease of mGluR5 Activity Reverses Behavioral But Not Neuroendocrine Consequences of Stress and Normalizes Homer1 Expression Levels

Next, we assessed whether decrease of mGluR5 signaling could be protective to the depression-like phenotype of chronically stressed mice. We detected robust CSDS effects, both in basal circulating corticosterone levels (Supplementary Figure S4A) and in the corticosterone response to a novel stressor (Figure 3a). Treatment with the selective mGluR5 inverse agonist CTEP did not affect these parameters under either basal or stress conditions. Recovery from the novel stressor was also disturbed in animals that underwent CSDS, but this was not influenced by CTEP treatment (Supplementary Figure S4B). Accordingly, on the physiological level, CSDS resulted in increased adrenal gland weight with no effect of CTEP treatment (Figure 3b). We observed increased CRH mRNA levels in the PVN in response to CSDS, irrespective of treatment condition (Figure 3c). In this experiment we also investigated expression changes of the two Homer1 isoforms in the hippocampus. Interestingly, the stress-induced elevation of Homer1b/c mRNA in the CA1 region of the dorsal hippocampus was reversed by CTEP treatment (Supplementary Figure S4C). Full expression data of Homer1 isoforms in the hippocampus can be found in Supplementary Table S1.

Figure 3
figure 3

Treatment with CTEP does not alter HPA axis function but reverses stress-induced behavioral effects. (a) CSDS resulted in a stronger activation of the HPA axis in response to a novel stressor (main condition effect: F1, 44=30.317, p<0.001). CTEP did not affect this phenotype. (b) CSDS induced an increase in adrenal size that was not affected by CTEP treatment (main condition effect: F1, 44=140.316, p<0.001). (c) In stressed vehicle-treated animals, CRH mRNA levels in the PVN were elevated compared with their respective control group (main condition effect: F1, 44=8.894, p<0.01). CTEP-treated animals did not show a significant alteration in CRH levels. See representative radiograph pictures in the panel below. (d) Although vehicle-treated animals displayed a strong decrease in locomotion while being stressed (p<0.001), CTEP treatment was able to counteract this phenotype and significantly enhanced locomotion in the CSDS group (p<0.01) (main condition effect: F1, 43=32.861, p<0.001; main treatment effect: F1, 43=4.981, p<0.05; interaction effect: F1, 43=5.971, p<0.05). (e) There was no effect of CSDS or CTEP in the SA test. (f) In the FUST, the water trial did not reveal any differences between treatment and condition groups, but stressed animals sniffed significantly less on the urine-dipped cotton tip (main condition effect: F1, 43=8.349, p<0.01; interaction effect: F1, 43=17.281, p<0.001). Further post hoc tests indicated that CTEP reduces the interest in female urine under basal conditions (p<0.01). Yet, although vehicle-treated mice that underwent CSDS showed a strong reduction in sniffing time (p<0.001), this effect was reversed by the CTEP treatment in the same condition group (p<0.05), indicating a protective effect of CTEP. (g) Although CSDS led to a decrease in active stress coping behavior in the FST, CTEP did not influence this behavioral parameter (main condition effect: F1, 44=14.109, p<0.01). *Significant main condition effect, p<0.05; #significant main treatment effect, p<0.05; §significant to corresponding vehicle group; +significant to corresponding nonstressed control group.

PowerPoint slide

On the behavioral level, a stress-induced reduction of locomotion in the OF test was reversed by chronic CTEP treatment (Figure 3d), without affecting basal locomotion in the control groups. We did not detect significant effects in the SA test (Figure 3e), but found that CTEP treatment was able to reverse the anhedonic phenotype induced by CSDS in the FUST (Figure 3f). However, CTEP-treated animals also expressed less interest in the urine-dipped cotton tip under control conditions. In the FST, chronic treatment of CTEP did not exert beneficial effects on the animals (Figure 3g).

Overexpression of Homer1a in the Hippocampus Promotes Vulnerability to Stress

Finally, we tested whether overexpression of the Homer1a splice variant, which inhibits mGluR5/Homer1b/c interaction and promotes ligand-independent mGluR5 signaling, would alter stress susceptibility. The injection site as well as qualitative and quantitative analyses of the viral overexpression can be found in Supplementary Figure S1. Basal corticosterone levels were elevated in response to CSDS, but overexpression of Homer1a in the hippocampus did not have an effect (Supplementary Figure S5A). After a novel acute stressor, Homer1a OE animals of the nonstressed condition showed an increased response compared with Empty animals, whereas CSDS enhanced circulating corticosterone equally in both treatment groups (Figure 4a). After 90 min of recovery, both Homer1a OE and Empty animals showed a disturbed HPA axis recovery indicated by significantly increased corticosterone levels in the CSDS groups (Supplementary Figure S5B). Homer1a OE mice had significantly bigger adrenal glands when exposed to CSDS compared with Empty mice that were stressed, indicating an increased HPA axis activity over the course of the stress period (Figure 4b). These HPA axis alterations were not accompanied by gene expression differences of CRH in the PVN (Figure 4c).

Figure 4
figure 4

Overexpression of Homer1a in the hippocampus promotes stress vulnerability. (a) Under control conditions, Homer1a OE led to a hyperactivation of the HPA axis compared with Empty animals (p<0.05). This effect was not apparent in mice that underwent CSDS, possibly because of a ceiling effect, as CSDS strongly enhanced the corticosterone response to a novel stressor (main condition effect: F1, 44=61.134, p<0.001; interaction effect: F1, 44=4.845, p<0.05). (b) CSDS and overexpression of Homer1a had profound impact on the adrenal gland size of the animals (main condition effect: F1, 44=61.134, p<0.001; main AAV treatment effect: F1, 44=5.365, p<0.05; interaction effect: F1, 44=4.845, p<0.05). The post hoc testing confirmed that CSDS increased adrenal gland sizes in both AAV groups (Empty: p<0.001; Homer1a OE: p<0.001), but in stressed Homer1a OE animals, this increase was significantly bigger compared with stressed Empty animals (p<0.05). (c) The mRNA levels of CRH in the PVN were not significantly altered in this experiment. See representative radiograph pictures in the panel below. (d) Although CSDS did not lead to a reduction in locomotion in Empty animals, overexpression of Homer1a affected the animals’ behavior, indicating a more pronounced susceptibility to CSDS (main condition effect: F1, 44=6.722, p<0.05). (e) This effect was also apparent in the SA test, where CSDS led to a reduced interaction ratio that was more pronounced in Homer1a OE animals (main condition effect: F1, 44=5.171, p<0.05). (f) The FUST revealed a significant stress effect in both the water (F1, 45=5.863, p<0.05) and the urine trials (F1, 44=27.368, p<0.001), with both AAV groups showing significant reductions in sniffing time when exposed to CSDS compared with the respective control groups. (g) In the FST, ANOVA revealed a condition (F1, 43=7.045, p<0.05), an AAV treatment (F1, 43=6.185, p<0.05), and a condition × AAV interaction effect (F1, 43=5.496, p<0.05). Following post hoc analysis, Homer1a OE mice showed significantly increased floating time compared with both their respective controls (p<0.001) and stressed Empty mice (p<0.05). *Significant main condition effect, p<0.05; #significant main AAV treatment effect, p<0.05; +significant from control of same AAV treatment, p<0.05; §significant from Empty AAV of same condition.

PowerPoint slide

While being exposed to CSDS, mice overexpressing Homer1a in the hippocampus showed a significant reduction in locomotion in the OF test (Figure 4d). Similar results were obtained in the SA test, where CSDS led to a reduced interaction ratio (Figure 4e). In the FUST, CSDS induced anhedonic behavior, with no apparent effects of Homer1a overexpression (Figure 4f). Overexpression of Homer1a in the hippocampus led to an increased behavioral despair and less active stress coping behavior upon CSDS exposure as depicted by increased floating time in the FST (Figure 4g), additionally indicating increased susceptibility to CSDS.

DISCUSSION

In this study, we provide evidence for the involvement of the Homer1/mGluR5 complex in mediating the depression-like phenotype induced by chronic stress exposure. First, we showed that hippocampal Homer1b/c mRNA is regulated, and Homer1/mGluR5 interaction is disrupted, by CSDS. A total knockout of Homer1 results in strong hyperactivity and stress susceptibility. Conversely, we were able to rescue some stress-induced behavioral alterations by chronic administration of the novel, orally bioavailable mGluR5 inverse agonist CTEP without interfering with HPA axis function. Furthermore, we demonstrate that overexpressing Homer1a in the hippocampus, thereby modulating the activity of mGluR5, increases the vulnerability to chronic stress on the physiological, neuroendocrines and behavioral levels. These findings suggest that the Homer1/mGluR5 complex may be valuable as a novel target for the treatment of stress-induced psychopathology, such as depression.

In this series of experiments, Homer1 has been shown to be regulated in response to CSDS on both mRNA levels, whereas the turnover rate was not affected. In a previous study applying a slightly different chronic defeat model, Homer1 expression was also reported to be upregulated in the nucleus accumbens (Berton et al, 2006). Furthermore, microarray data from another study in our lab indicated hippocampal Homer1 to be differentially regulated between stress-resilient and -vulnerable animals (Schmidt et al, 2010), further strengthening the evidence of Homer1 being involved in stress-induced psychopathology. At least on the mRNA level, the effects of chronic stress were specific to the Homer1b/c splice variant, whereas the expression of the immediate early gene Homer1a was not affected. This is in contrast to prenatal stress, where in the hippocampus mainly Homer1a expression was found to be increased (Ary et al, 2007). This difference could be because of the developmental time window of stress exposure, and also the time of testing (weanling vs adult) or the sex of the animals (females vs males). We also observed a significant decrease in Homer1/mGluR5 coupling following chronic social defeat. This effect cannot be specified to either Homer1 splice variant at the moment because of the lack of specific and reliable antibodies targeting Homer1a (Tronson et al, 2010). Further studies will be required to unravel the functional consequence of this effect.

We report a major disturbance of HPA axis activity in mice that are deficient in Homer1. This is evidenced on one hand on the physiological level, where Homer1 KO mice show enlarged adrenal glands, in line with previous reports (Grinevich et al, 2011). On the other hand, we also showed that corticosterone release in response to stress is severely altered in these animals. Hyperactive corticosterone responses induced by CSDS are frequently observed in this paradigm (Wang et al, 2011; Wagner et al, 2011; Hartmann et al, 2012) and deletion of Homer1 further increased this effect, indicating a prominent regulatory role of this glutamatergic pathway in the feedback regulation of the HPA axis. Compensatory effects caused by changes in mGluR5 or Homer2a/b expression (Ary et al, 2013) in Homer1 KO mice seem unlikely, as the observed effects in the hippocampus were rather modest.

We further observed a strong hyperactive phenotype because of Homer1 deletion that has been previously reported in studies that employed this mouse model (Szumlinski et al, 2005; Jaubert et al, 2007). These hyperactive behaviors in general led to an apparent reversal of the CSDS-induced phenotype that was mostly visible in locomotive and social behavior. However, we also detected a reduced interest in interacting with novel stimuli in KO animals, such as in the FUST. These behavioral patterns may be ascribed to an attention deficit hyperactivity disorder (ADHD)-like phenotype (Sagvolden et al, 2005) that has previously been linked to altered Homer1 expression profiles in the prefrontal cortex and the hippocampus (Hong et al, 2009; Hong et al, 2011), although the increased activity of Homer1 KO mice may also be a confounding factor in some of the behavioral tests. It is important to note that based on the present data, we are not able to discern immediate effects of Homer1 deletion from developmental effects that originate in earlier stages of the animals’ life. Indeed, Homer1 has been shown to be strongly expressed in developing tissues (Shiraishi-Yamaguchi and Furuichi, 2007) and a total knockout is therefore likely to exert major effects on these animals before the CSDS procedure started. Nonetheless, these findings indicate the importance of Homer1-mediated signaling in the response to stress.

As the deletion of Homer1 results in the loss of both Homer1a and Homer1b/c splice variants, which are hypothesized to have opposing effects (Brakeman et al, 1997), it is important to more specifically modulate mGluR5/Homer1 signaling. We therefore administered the novel, bioavailable mGluR5 inverse agonist CTEP (Lindemann et al, 2011) to mice subjected to chronic stress. Under basal conditions, CTEP did not have any detrimental effects on the physiological or neuroendocrine level. In addition, decrease of mGluR5 activity over the course of the stress exposure did not affect HPA axis function or modulation, as both treatment groups showed similar corticosterone profiles under all measured conditions. Given the results from the KO animals presented above, it is likely that Homer1 influences HPA axis responsiveness independent from mGluR5 signaling. Yet, CTEP did have beneficial effects on the behavioral phenotype of stressed animals. Here, stress-induced anhedonia and reduced locomotion was rescued in animals that received CTEP. Thus, although CTEP did not reverse the stress-induced molecular profile, it showed therapeutic value in behavioral parameters. These results strengthen the idea of combining different antidepressant treatments to maximize therapeutic efficacy (Connolly and Thase, 2011; Palaniyappan et al, 2009). Indeed, CTEP may serve as a basis for future antidepressants that specifically target the glutamate system, as its pharmacokinetic properties are significantly improved from previous mGluR5 antagonists such as MPEP and MTEP (Lindemann et al, 2011; Anderson et al, 2003; Busse et al, 2004).

In contrast to the pharmacological modulation of mGluR5 activity by CTEP, which was also not hippocampus specific and likely altered mGluR5 signaling in many different brain regions, a specific increase in the short Homer1a isoform is expected to disrupt glutamate-stimulated intracellular calcium signaling (Tu et al, 1998; Yuan et al, 2003) and activate ligand-independent mGluR5 pathways (Ango et al, 2001). Indeed, a specific overexpression of Homer1a in the hippocampus led to changes in HPA axis activity under both basal and CSDS conditions, thereby indicating an increase in vulnerability to both acute and chronic stress. This phenotype is not only present on the neuroendocrine and physiological level but also reflected in different behavioral parameters related to stress coping (FST) and locomotor activity, but not social interaction and hedonic behavior. Interestingly, some of the behavioral alterations induced by Homer1a overexpression in the hippocampus, for example in the FST, parallel those previously observed by Lominac et al (2005) following overexpression of Homer1a in the prefrontal cortex.

Activation of Homer1a gene transcription is a rapid and plastic process in response to synaptic activity (de Bartolomeis and Iasevoli, 2003; Brakeman et al, 1997; Kato et al, 1997). It can be hypothesized that repeated transcriptional activation of this immediate early gene in response to the daily defeat sessions induces counter-regulatory changes in the central stress systems, including the upregulation of Homer1b/c (Berton et al, 2006). These disturbances, in turn, may contribute to the vulnerable behavioral and neuroendocrine phenotypes that we observed under the influence of CSDS. Prolonged ligand-independent activation of mGluR5 via abundant Homer1a protein levels also severely affects IP3 receptor activation and subsequent downstream signaling (Ango et al, 2001; Kammermeier, 2008). The continuous presence of Homer1a may therefore profoundly change neuronal signaling pathways that may in turn render the organism more vulnerable to chronic stress. In addition, it has previously been shown that interactions between NMDA and mGluR5 receptors are mediated by the PSD95/Shank/Homer1 complex (Hayashi et al, 2009; Bertaso et al, 2010), and Homer1a was demonstrated to be a key modulator of mGluR5 coupling to effector targets that produce excitatory postsynaptic currents (Kammermeier and Worley, 2007). Given the increasing body of evidence showing that NMDA receptor targeting agents produce rapid-acting antidepressant effects (Krystal et al, 2013; Kavalali and Monteggia, 2012), a Homer1a-mediated overactivation of this signaling pathway may profoundly affect antidepressant treatment efficacy. The development of new drugs that target this system, mainly via mGluR5, is therefore of great value and importance (Sanacora et al, 2012; Krystal et al, 2010).

There are some limitations in the presented data sets that need to be considered when interpreting the results. Although the different experiments were performed analogously, a direct comparison of the results is hampered by differences in the mouse lines, the applied treatment (eg, gavaging), or the history of surgery. Furthermore, although we observed behavioral and neuroendocrine alterations following mGlur5/Homer1 manipulations, a direct mechanistic link of the presumably altered hippocampal function to the respective readouts is still lacking. Another still unresolved question is whether overexpression of Homer1b/c could have altered the response to CSDS. However, as this manipulation would have a similar effect as CSDS exposure per se and a further increase in Homer1b/c might not additionally increase mGluR5 signaling due to a ceiling effect, this experiment was not included.

In summary, our study provides compelling evidence for the involvement of the Homer1/mGluR5 complex in the emergence and regulation of stress-induced behavioral and neuroendocrine phenotypes (see Figure 5 for a schematic overview). We demonstrated that the Homer1/mGluR5 pathway is altered by CSDS, and HPA axis function is strongly disturbed in animals that carry a total knockout of Homer1. We further demonstrated that stimulating ligand-independent mGluR5 activity by increased levels of Homer1a lead to a stress-vulnerable behavioral phenotype. Conversely, decrease of mGluR5 activity by CTEP was able to recover the stress-induced behavioral alterations. With the present data indicating an involvement of the Homer1/mGluR5 pathway in stress-related psychiatric disorder, further research to fully characterize the contributing molecular mechanisms is highly warranted.

Figure 5
figure 5

Schematic model describing the outcome and proposed mechanism of the different manipulations of mGluR5/Homer1 interaction under chronic stress conditions. (a) Under wild-type conditions, chronic stress increases the levels of the long isoform Homer1b/c (H1b/c), thereby strengthening the intracellular link of mGluR5 to calcium stores in the endoplasmic reticulum (ER), but also to NMDA receptor (NMDA-R) signaling. (b) In Homer1 KO animals, all splice variants of Homer1 are missing, and thereby the mGluR5 ligand-dependent and ligand-independent signaling cascades will be disrupted. (c) When applying the inverse agonist CTEP during chronic stress, ligand-dependent signaling via inositol triphosphate (IP3) will be diminished, but the interaction of Homer1b/c with IP3 receptors and with the NMDA-R signaling cascade will remain unaffected. (d) In contrast, overexpression of Homer1a (H1a) mimics a situation similar to an acute stress challenge. Here Homer1a replaces Homer1b/c at the mGluR5, thereby stimulating ligand-independent signaling. At the same time, however, the structural link of mGluR5 to the ER and to NMDA-Rs will be weakened.

PowerPoint slide

FUNDING AND DISCLOSURE

Lothar Lindemann, Georg Jaschke, and Joseph G Wettstein are full-time employees of F. Hoffmann-La Roche AG, Basel, Switzerland. The other authors declare no conflict of interest.