Introduction

It is well established that chronic stress is aetiologically related to the pathophysiology of affective disorders (Young and Korszun 2002; Bale 2006; Goel and Bale 2009; Oitzl et al. 2010), while gender-specific response in stress and antidepressants attract particular attention since women are more vulnerable to depression (Kessler 2003; Alonso et al. 2004). To date, a growing body of evidence indicates that mitogen-activated protein kinase (MAPK) signalling pathway participate in various stress-induced responses and is considered to be one of the pathophysiological mechanisms in depression (Dwivedi et al. 2006; Li et al. 2009; Duric et al. 2010). However, the question regarding the effects of stress and antidepressants on MAPK signalling from perspective of gender remains elusive.

The term MAPK is used to describe a family of signal transduction mediators, extensively distributed throughout the central nervous system (Ortiz et al. 1995; Flood et al. 1998), which regulate a diverse array of cellular functions via phosphorylation (Grewal et al. 1999; Schaeffer and Weber 1999; Houslay and Kolch 2000; Sweatt 2001). The most common MAPK are the extracellular signal-regulated kinases, ERK1 and 2, which primarily regulate cellular growth and differentiation (Tunquist and Maller 2003; Huang et al. 2004), and the p38 MAPK and c-Jun N-terminal kinases (JNK1, 2 and 3)/stress-activated protein kinases (JNK/SAPK), which mainly function as mediators of cellular stress such as inflammation and apoptosis (Pearson et al. 2001; Kumar et al. 2003; Roux and Blenis 2004).

The supporting evidence for the necessity of appropriate MAPK signalling in stress response can be found in its linkage with two molecules that are of great importance in aetiology of depression, glucocorticoid receptor (GR) and brain-derived neurotrophic factor (BDNF). It is well known that MAPK alter GR transcriptional activity and modulate GR functioning through alteration of its phosphorylation (Rogatsky et al. 1998; Irusen et al. 2002). In addition, activated MAPK affects BDNF expression which regulates plasticity in the brain, known as one of the hallmarks in pathogenesis of depression (Duman et al. 2000; Calabrese et al. 2009). Considering multiple effects of MAPK in the regulation of molecular mediators of stress response, it is not unreasonable to speculate that MAPK signal transduction pathways underlie aberrant brain function. Indeed, decreased ERK activity and expression found in post-mortem hippocampus and cerebral cortex of depressed suicide victims (Duric et al. 2010) indicate the impairment in ERK functioning in depression (Fumagalli et al. 2005; Duman et al. 2007). Moreover, the levels of JNK, ERK and p38 kinases are found to be altered in animal models of depression (Battiwalla et al. 2006; Adzic et al. 2009; Todorovic et al. 2009; Duric et al. 2010; Bruchas et al. 2011; Andreetta et al. 2013) that additionally support importance of MAPK in the aetiology of stress-related disorders.

The present study complements the previous one (Mitic et al. 2013) and aims to investigate the potential role of cytoplasm-nuclear distribution of whole MAPK family—i.e. JNK1, 2 and 3; ERK1 and 2; and p38 kinases, as well as their phosphoisoforms in the hippocampus of chronically stressed female and male rats. We also analysed if potential changes in their levels could be attenuated or reversed with fluoxetine hydrochloride (FLU) treatment. Surprisingly, the effect of antidepressant drugs on the level of MAPKs is almost completely unexplored, particularly from the perspective of different sexes.

Materials and Methods

Preparation of Fluoxetine Hydrochloride Solution

The capsules of flunirin were emptied and dissolved in distilled, sterile water with the aid of ultra-sonication, and the solution was filtered through Whatman No. 42 filter paper. The FLU concentration in the solution was determined using ultra-performance liquid chromatography (UPLC) (Djordjevic et al. 2005).

Animals and Treatment

The experiments were performed on adult 3-month-old female (body mass 250–300 g) and male (body mass 330–400 g) Wistar rats. All animals were housed at 20 ± 2 °C, with a 12 h light/dark cycle (lights on at 07:00 hours), with food (commercial rat pellets) and drinking water available ad libitum.

All experiments on female and male rats were conducted in parallel. As depicted in Fig. 1, both female and male animals were divided into four experimental groups, each containing 12 animals housed, 4 animals per cage: control + vehicle, control + fluoxetine, stress + vehicle and stress + fluoxetine groups. The experiment consisted of two phases and lasted for 6 weeks (42 days). The first experimental phase (stress) lasted 21 days, during which animals of stress + vehicle and the stress + fluoxetine groups were submitted to chronic psychosocial isolation (CPSI). The CPSI procedure was carried out according to our standard protocol (Adzic et al. 2009). The second experimental phase consisted of the vehicle (VEH) or fluoxetine (FLU) treatment for 21 days, while animals remained in the CPSI. The 21 days of CPSI was used as a model of chronic stress during which animals had normal auditory and olfactory experiences but were deprived of any visual or tactile contact with other rats.

Fig. 1
figure 1

Experimental groups and design. For details, see “Materials and Methods” section

Full size image

FLU was injected intraperitoneally to control + FLU group and to CPSI + FLU group at a daily dose of 5 mg/kg of body mass at 09:00 hours during a 21-day period. Both vehicle groups (control and CPSI) received distilled water under the same conditions as matching FLU-treated groups. CPSI groups remained isolated during treatment. For testing the behaviour and molecular parameters, different set of animals were used and experiments were repeated twice. Vaginal smears were microscopically analysed for determination of the oestrous cycle stage 1 week before the start of the experiment and immediately after the sacrifice. Only females with normal 4–5-day cycles were included in the experiment. Neither of treatments (CPSI or FLU or injections) altered the oestrous cycle phase distribution.

All animal procedures were approved by the Ethical Committee for the Use of Laboratory Animals of the VINCA Institute of Nuclear Sciences, according to the guidelines of the EU-registered Serbian Laboratory Animal Science Association (SLASA).

Preparation of Brain Tissues

All animals were sacrificed between 10:00–11:00 a.m., 24 h after the last injection by rapid decapitation with a guillotine (Harvard Apparatus, USA). The hippocampus was removed and frozen in liquid nitrogen until further preparation.

Preparation of Cytosolic and Nuclear Extracts

Different subcellular fractions were prepared from the hippocampus. Frozen hippocampus was weighed and homogenized (1:2w/v) in ice-cold 20-mM Tris–HCl (pH 7.2) buffer containing 10 % glycerol, 50 mM NaCl, 1 mM Na2EDTA, 1 mM Na2EGTA, 2 mM DTT, and cocktail of protease (Sigma-Aldrich) and phosphatase inhibitors (Sigma-Aldrich) with 20 strokes of Potter-Elvehjem Teflon-glass homogenizer. All operations were conducted at 0–4 °C. Samples were centrifuged 10 min at 2,000 × g to give a supernatant and a pellet (containing nuclear and cell debris which was further processed for nuclear extracts). Supernatants were ultra-centrifuged for 1 h at 105,000 × g, and the final supernatants were used as cytosolic fraction. Nuclear pellets were washed (three times) with 0.5 ml of homogenization buffer and centrifuged for10 min at 2,000 × g at 4 °C to obtain nuclear fraction. Final pellets were weighed, resuspended (1:1w/v) in the same buffer supplied with 0.5 M KCl, incubated for 1 h in ice bath (with frequent vortexing) and centrifuged for 10 min at 8,000 × g at 4 °C (Spencer et al. 2000; Moutsatsou et al. 2001; Trzeciak et al. 2004). Protein concentration in cytosolic and nuclear fraction was determined by method of Lowry et al. (1951).

Western Blot Analysis

Protein analysis was performed on cytosolic and nuclear fractions. Equal amounts of protein (60 μg) were separated by SDS-PAGE on 10 % polyacrylamide gels and then electrophoretically transferred onto PVDF membrane (Immobilon-P membrane, Millipore) using a blot system (Transblot, BioRad). The membranes were blocked with phosphate-buffered saline (PBS) buffer containing 5 % milk for 1 h at room temperature and thereafter incubated overnight at 4 °C with the following primary antibodies: Anti-p38 and anti-pp38 antibodies were used to detect total p38 and phosphorylated p38 (Cell Signaling), respectively; anti-human JNK1/JNK2 monoclonal antibody (BD, Biosciences, Oxford, UK) and anti-phospho-SAPK/JNK (Thr183/Tyr185) polyclonal antibody (Cell Signaling) were used to detect total or phosphorylated JNK; and anti-ERK1/2 and anti phosho-ERK1/2 were used to detect total or phosphorylated ERK1/2 (Cell Signaling). Rabbit polyclonal anti-β-actin (Abcam, Cambridge, UK) was used to detect β-actin as a loading control. Membranes were then incubated with respective secondary antibody for 2 h, and immune complexes were visualized by chemiluminescence using ECL Rabbit IgG, horseradish peroxidase (HRP)-linked whole antibody and ECL Mouse IgG, HRP-linked whole antibody (Amersham). Densitometry of protein bands on X-ray film was performed by ImageJ analysis PC software.

Statistical Analysis

Data are presented as mean ± SEM from three independent experiments in which measurements of behavioural and molecular parameters were repeated three times. Values obtained for protein levels were presented relative to female control group. Within each gender, all results were analysed using two-way ANOVA with CPSI and FLU treatment as independent factors. Gender differences were analysed using three-way ANOVA with gender, CPSI and FLU as independent factors. Statistically significant differences between groups were analysed by the post hoc Tukey’s test. In order to simplify the presentation of data, all statistically significant differences are given as p < 0.05.

Results

Effects of CPSI and FLU on MAPK Signalling in Cytosol and Nucleus of Hippocampus

In order to determine the potential links between FLU normalization of disrupted behaviour previously induced by CPSI and MAPK signalling role, we analysed the cytosol/nuclear levels of all three MAPK levels in hippocampus of both female and male animals. In particular, we analysed the hippocampal levels of JNK1, 2 and 3; ERK1 and 2; and p38 protein, as well as their phosphorylation status (pJNK1, 2 and 3; pERK1 and 2; and pp38) in both cytosol and nucleus. In this moment, it is important to emphasize that results obtained from nuclear levels of JNKs and their phosphoisoforms were already published in our previous study (Mitic et al. 2013) but because of the clarity and easier perception of obtained results, we show them again alongside with their cytosolic levels.

Effects of Stress and FLU on Cytosolic and Nuclear Levels of p38 and pp38

In both genders, CPSI significantly decreased pp38 levels in both cell compartments (females F = 6.751, F = 4.142, p < 0.05; males F = 11,817, F = 8.900, p < 0.05) (Fig. 2), while it had no effect on the levels of total p38, except in female nucleus where stress reduced it (F = 4.654, p < 0.05) (Fig. 2). Overall results from cytosol and nucleus suggest that CPSI down-activated p38 signalling in both genders. Three-way ANOVA did not revealed any significant gender differences at the levels of pp38 and p38 in stressed animals (Fig. 2).

Fig. 2
figure 2

Western blot of phospho-p38 and total p38 in cytosol and nucleus. Representative Western blots demonstrating phospho-p38 (pp38) and total p38 protein levels in the cytosol (a) and nucleus (b) of hippocampus of control and stressed (CPSI) female (♀) and male (♂) Wistar rats treated with vehicle (VEH) or fluoxetine (FLU; 5 mg/kg). Quantitative data represent the levels of analysed proteins, pp38 and p38, expressed as a percentage of female control treated with VEH (female control + VEH group, set at 100 %). β-actin was used as an internal loading control. Data are presented as the mean ± SEM and were analysed with two-way ANOVA in each gender. To determine statistically significant gender differences, we used three-way ANOVA followed by post hoc Turkey’s test employing stress, fluoxetine and gender as the factors. All statistically significant differences are given as p < 0.05 (*vs CPSI, #vs FLU treatment, $female vs male)

Full size image

FLU treatment in control animals induced opposite effects at the levels of cytosolic pp38. In control females, FLU decreased pp38 levels (F = 3.810, p < 0.05), while in males, it increased its levels (F = 9.102, p < 0.05) (Fig. 2). In CPSI females, FLU increased the levels of pp38 in both cell compartments (F = 46.500, F = 25.522, p < 0.05) (Fig. 2) and increased nuclear p38 levels (F = 17.455, p < 0.05) (Fig. 2), while in males, FLU did not cause any changes in cytosolic and nuclear levels of pp38 and p38 (Fig. 2). Our results imply that FLU increased activation of p38 signalling in stressed females, while in males, it had no effect. Three-way ANOVA revealed significant gender differences only at the level of nuclear pp38 upon FLU in stressed animals (gender × CSPI × FLU interaction F = 25.273, p < 0.05) (Fig. 2).

Effects of CPSI and FLU on Cytosolic and Nuclear Levels on JNK and pJNK Levels

In females, CPSI decreased pJNK1, 2 and 3 levels in both cell compartments (cytosol F = 12.691, p < 0.05; nucleus F = 8.726, F = 7.557, p < 0.05) (Figs. 3 and 4), although this decrease for cytosolic pJNK2 and 3 was not statistically significant. Furthermore, CPSI increased cytosolic levels of total JNK1 in females (F = 5.563, p < 0.05) (Fig. 3), while it reduced nuclear levels of total JNK1, 2 and 3 (F = 13.5, F = 11.8, p < 0.05) (Fig. 4). Overall, these results from cytosol and nucleus suggest that CPSI down-activated whole JNK pathway in females. In males, CPSI decreased cytosolic levels of pJNK1, 2 and 3 (F = 12.690, F = 6.593, p < 0.05) (Fig. 3), while it increased their nuclear levels (F = 16.068, F = 14.568, p < 0.05) (Fig. 4). At the levels of total JNK levels in males, CPSI increased only cytosolic levels of total JNK1 (F = 6.982, p < 0.05) (Fig. 3). Overall, our results from cytosol and nucleus imply that CPSI increased nuclear translocation of pJNK1, 2 and 3 in males. Three-way ANOVA revealed significant gender differences at the level of nuclear pJNK2 and 3 under CPSI (gender × CPSI interaction F = 22.647, p < 0.05) (Fig. 4).

Fig. 3
figure 3

Western blot of phospho-JNK and total JNK in cytosol. a Representative Western blots demonstrating phospho-JNK (pJNK1, 2 and 3) and total JNK (JNK1, 2 and 3) protein levels in the cytosol of hippocampus of control and stressed (CPSI) female (♀) and male (♂) Wistar rats treated with vehicle (VEH) or fluoxetine (FLU; 5 mg/kg). b Quantitative data represent the levels of analysed proteins, pJNKs and JNKs, expressed as a percentage of female control treated with VEH (female control + VEH group, set at 100 %). β-actin was used as an internal loading control. Data are presented as the mean ± SEM and were analysed with two-way ANOVA in each gender. To determine statistically significant gender differences, we used three-way ANOVA followed by post hoc Turkey’s test employing stress, fluoxetine and gender as the factors. All statistically significant differences are given as p < 0.05 (*vs CPSI, #vs FLU treatment, $female vs male)

Full size image
Fig. 4
figure 4

Western blot of phospho-JNK and total JNK in nucleus. a Representative Western blots demonstrating phospho-JNK (pJNK1, 2 and 3) and total JNK (JNK1, 2 and 3) protein levels in the nucleus of hippocampus of control and stressed (CPSI) female (♀) and male (♂) Wistar rats treated with vehicle (VEH) or fluoxetine (FLU; 5 mg/kg). b Quantitative data represent the levels of analysed proteins, pJNKs and JNKs, expressed as a percentage of female control treated with VEH (female control + VEH group, set at 100 %). β-actin was used as an internal loading control. Data are presented as the mean ± SEM and were analysed with two-way ANOVA in each gender. To determine statistically significant gender differences, we used three-way ANOVA followed by post hoc Turkey’s test employing stress, fluoxetine and gender as the factors. All statistically significant differences are given as p < 0.05 (*vs CPSI, #vs FLU treatment, $female vs male)

Full size image

FLU treatment in CPSI females increased pJNK1, 2 and 3 levels in both cell compartments (cytosol F = 16.046, F = 4.449, p < 0.05; nucleus F = 12.6, F = 6.6, p < 0.05) (Figs. 3 and 4). In addition, FLU treatment decreased all three total JNK levels in female cytosol (F = 53.816, F = 16.292, p < 0.05) (Fig. 3) and increased their nuclear levels (F = 21.562, F = 13.112, p < 0.05) (Fig. 4). Overall, these results suggest that FLU treatment in CPSI females increased pJNK signalling in both cell compartments. In CPSI males, FLU treatment increased pJNK1, 2 and 3 levels in cytosol (F = 7.904, F = 4.449, p < 0.05) (Fig. 3), while it had no effect in nucleus (Fig. 4). At the levels of total JNK, FLU did not show any significant effect neither in cytosol nor in nucleus (Figs. 3 and 4). FLU treatment of control males increased the total nuclear JNK1 levels (F = 12.765, p < 0.05) (Fig. 4). Our results imply that in CPSI males, FLU treatment increased cytosolic pJNK signalling. Three-way ANOVA revealed significant gender differences only at the levels of total JNK2 and 3 in both cytosol and nucleus upon FLU treatment in stressed animals (gender × CSPI × FLU interaction F = 18.632, F = 24.678, p < 0.05) (Figs. 3 and 4).

It is interesting to point out that the control levels of all three nuclear JNKs and their phosphoisoforms were higher in females than those in males, particularly those of total JNK1 and all three pJNKs (Fig. 4).

Effects of CPSI and FLU on Cytosolic and Nuclear Levels on ERK and pERK Levels

In females, CPSI increased cytosolic levels of pERK1 and 2 (F = 11.67, F = 22.31, p < 0.05) (Fig. 5), while it decreased their nuclear levels (F = 4.171, F = 4.852, p < 0.05) (Fig. 6), suggesting decreased nuclear translocation of pERK1 and 2, i.e. induced cytosolic retention of pERK1 and 2. Furthermore, CPSI decreased total ERK1 and 2 levels in the nucleus (F = 6.151, F = 8.120, p < 0.05) (Fig. 6), while it had no effect of cytosolic total ERK proteins (Fig. 5). In males, CPSI decreased cytosolic levels of pERK1 and 2 (F = 13.720, F = 14.130, p < 0.05) (Fig. 5), while it increased their nuclear levels (F = 53.911, F = 36.122, p < 0.05) (Fig. 6), suggesting elevated nuclear translocation of pERK1 and 2 under CPSI. At the levels of total ERK1 and 2, CPSI increased their cytosolic levels (F = 8.250, F = 13.800, p < 0.05) (Fig. 5), while it had no effect on their total nuclear levels (Fig. 6). Three-way ANOVA revealed significant gender differences upon CPSI at the levels of all ERKs and pERKs in both cell compartments (gender × CPSI interaction cytosol F = 3.621, F = 3.542; nucleus F = 2.121, F = 3.923, p < 0.05) (Figs. 5 and 6).

Fig. 5
figure 5

Western blot of phospho-ERK and total ERK in cytosol. a Representative Western blots demonstrating phospho-ERK (pERK1 and 2) and total ERK (ERK1 and 2) protein levels in the cytosol of hippocampus of control and stressed (CPSI) female (♀) and male (♂) Wistar rats treated with vehicle (VEH) or fluoxetine (FLU; 5 mg/kg). b Quantitative data represent the levels of analysed proteins, pERKs and ERKs, expressed as a percentage of female control treated with VEH (female control + VEH group, set at 100 %). β-actin was used as an internal loading control. Data are presented as the mean ± SEM and were analysed with two-way ANOVA in each gender. To determine statistically significant gender differences, we used three-way ANOVA followed by post hoc Turkey’s test employing stress, fluoxetine and gender as the factors. All statistically significant differences are given as p < 0.05 (*vs CPSI, #vs FLU treatment, $female vs male)

Full size image
Fig. 6
figure 6

Western blot of phospho-ERK and total ERK in nucleus. a Representative Western blots demonstrating phospho-ERK (pERK1 and 2) and total ERK (ERK1 and 2) protein levels in the nucleus of hippocampus of control and stressed (CPSI) female (♀) and male (♂) Wistar rats treated with vehicle (VEH) or fluoxetine (FLU; 5 mg/kg). b Quantitative data represent the levels of analysed proteins, pERKs and ERKs, expressed as a percentage of female control treated with VEH (female control + VEH group, set at 100 %). β-actin was used as an internal loading control. Data are presented as the mean ± SEM and were analysed with two-way ANOVA in each gender. To determine statistically significant gender differences, we used three-way ANOVA followed by post hoc Turkey’s test employing stress, fluoxetine and gender as the factors. All statistically significant differences are given as p < 0.05 (*vs CPSI, #vs FLU treatment, $female vs male)

Full size image

FLU treatment in CPSI females increased cytosolic pERK1 levels (cytosol F = 21.233, p < 0.05) (Fig. 5) and both nuclear pERK1 and 2 levels (nucleus F = 16.722, F = 13.071, p < 0.05) (Fig. 6). At the levels of total ERK1 and 2, FLU treatment increased their nuclear levels (F = 17.681, F = 12.676, p < 0.05) (Fig. 6), while it had no effect on their cytosolic levels (Fig. 5). Overall, these results of total and pERKs in both cell compartments imply that FLU treatment in CPSI females predominantly upregulated nuclear ERK signalling (Figs. 5 and 6). In CPSI males, FLU treatment showed a trend of increase in the nuclear levels of both total and pERKs (Fig. 6) but without statistical significance. Three-way ANOVA failed to show any significant effects between genders either on total ERKs or pERKs upon FLU treatment in CPSI animals (Figs. 5 and 6).

Discussion

In this study, we investigate the cytoplasm-nuclear distribution of whole MAPK family, JNK, ERK and p38 kinases, as well as their phosphoisoforms in the hippocampus of chronically stressed female and male rats, and could the potential changes in their levels be attenuated or reversed with fluoxetine treatment. Also, we wanted to analyse potential crosstalk between changes in MAPK levels with changes in depressive-like behaviour and BDNF that we investigated in this animal model of chronic stress (Mitic et al. 2013).

Although there is a very obscure literature concerning the role of p38 in response to chronic stress and to antidepressant treatment, down-activation of p38 signalling that we found upon chronic stress in both genders was similar with findings of Budziszewska et al. (2010). These findings, together with the fact that p38 affects glucocorticoid receptor signalling (Holsboer 2000; Pariante and Miller 2001; Wang et al. 2004; Miller et al. 2005), suggest that disturbance in p38 kinase could have significant role in pathophysiology of stress-related disorders. Moreover, the fact that fluoxetine reversed the stress-disrupted p38 signalling in females together with normalization of their behaviour (Mitic et al. 2013) underscores the significance of p38 kinase in female stress response. On the other hand, the absence of fluoxetine effect on p38 in stressed males and gender-opposite normalization of behaviour upon fluoxetine lead us to hypothesize that the therapeutic effect of fluoxetine is related to its effect on p38 signalling only in females.

Along with down-activation of p38 signalling in females, chronic stress down-activated JNK signalling in both cytosol and nucleus, while in males, stress potentiated nuclear translocation of pJNKs. Our results from females were in accordance with recent studies that reported that prenatal stress as well as chronic unpredictable stressor exposure during adulthood reduced levels of pJNKs in the hippocampus (Li et al. 2009; Budziszewska et al. 2010). These findings imply that gender-different nuclear levels of pJNKs found in stressed animals are the result of different subcellular distribution of pJNKs, particularly to their nuclear translocation. Alterations in nuclear JNK signalling could be responsible for the sex-specific differences in behaviour. Namely, down-activation of JNK signalling in females was fallowed with decreased immobility (Mitic et al. 2013), while in males, increased immobility was paralleled with increased nuclear propagation of pJNKs. Moreover, following stressor exposure, nuclear pJNKs activate a range of transcription factors, such as c-Jun, ATF-2, Elk-1, c-Myc, p53 and NFAT4 (Widmann et al. 1999), which regulate numerous stress regulatory genes and control multiple cellular processes which overall could be responsible for stress-induced depressive-like behaviour.

These gender differences found in pJNK levels in cytosol and nucleus could also be an outcome from a possible gender-specific upstream regulation of JNK signalling either through activation by MKK3/6 and MMK4/7 (Wada et al. 2001; Brust et al. 2007) or inactivation by MAPK phosphatase 1 (MKP-1), MKP-3/6, MKP-7 and PP2A (Muda et al. 1996; Hirsch and Stork 1997; Willoughby and Collins 2005). However, this assumption is less likely to be true because we did not notice any differences in p38 signalling in our study which is usually targeted with same upstream kinases and phosphatases like JNKs. Additional argument that highlights the role of JNK signalling in stress response was supported by our further finding that fluoxetine affected pJNK signalling in stressed animals. In females, fluoxetine reversed the stress-disrupted changes in both cellular compartments, while in males, its effect was restricted only at the levels of cytosolic pJNKs. Even though there is not a literature background on which we can correlate our results in regard to fluoxetine effect on pJNK signalling, our findings are in agreement with the previous. Namely, it was documented that fluoxetine as well as imipramine fully normalized pJNK levels in stressed animals (Budziszewska et al. 2010; Galeotti and Ghelardini 2012). Still, the precise mechanism how fluoxetine affects nucleo-cytoplasmic trafficking of pJNKs from the perspective of gender remains to be further explored.

Regarding ERK signalling in the hippocampus, we found gender-specific effect of both chronic stress and fluoxetine. Namely, in females, stress retained pERKs in cytosol, while in males, it induced nuclear translocation of pERKs. In the light of existing literature regarding the activation of ERK pathway (Dwivedi et al. 2001; Feng et al. 2003; Qi et al. 2006; Iio et al. 2011), our results more precisely elucidate the effect of stress on their cytoplasm-nuclear trafficking. Regulatory mechanism of subcellular distribution of ERKs has been elucidated extensively (Torii et al. 2004; Kondoh et al. 2005; Roskoski 2012). In short notes, besides phosphorylation of ERK by upstream MEK and forming ERK/MEK heterodimer in cytoplasm, nuclear translocation is accompanied by the dissociation of ERK1/2/MEK1/2 complex and proper functioning of nuclear anchoring proteins (Fukuda et al. 1997; Lenormand et al. 1998; Adachi et al. 1999; Brunet et al. 1999; Kondoh et al. 2005). In respect to our results, we can hypothesize that cytosolic retention of pERKs in females could be a consequence of disrupted dissociation of ERK1/2/MEK1/2 complex and/or impaired nuclear anchoring or scaffolding under chronic stress. Such cytoplasmic sequestration of ERKs can result in decrease of nuclear ERK activity which overall could diminish ERK signalling. Observed differences in pERK1/2 nuclear levels between stressed females and males can result in gender-specific modulation of transcription factors, such as CREB, ATF-2, c-FOS, ELK-1, transcription suppressors and chromatin remodelling proteins which ultimately could influence the gender-specific behaviour under chronic stress.

Similarly to JNKs, we cannot rule out the possibility that the gender-specific changes in the levels of pERKs that we found upon chronic stress could be also a consequence of gender-different upstream regulation of ERK activation either through their corresponding upstream activators (Pouyssegur et al. 2002) or through corresponding ERK deactivators.

While ERK phosphorylation has been documented as an intracellular signalling mechanism mediating antidepressant efficacy, supporting evidence is largely limited to studies in naive rodents or in vitro models (Hisaoka et al. 2001; Tiraboschi et al. 2004; Fumagalli et al. 2005), despite evidence from post-mortem studies of depressive suicide subjects (Dwivedi et al. 2001; Dwivedi et al. 2006). Our results provide evidence of gender-specific and compartment dependent regulation of ERK signalling in response to chronic fluoxetine treatment of stressed animals. Namely, we found that effects of fluoxetine on nuclear ERKs were more pronounced in females, while its effect in males was similar to those obtained by Qi et al. (2006) and by Taler et al. (2008). This normalization of nuclear ERK signalling by fluoxetine was associated with normalization of stressed female behaviour (Mitic et al. 2013), while in males, convergence between ERK signalling and behaviour was not found. Altogether, these results reinforce the importance of ERK1/2 phosphorylation as potential targets of antidepressant as well as their participation in pathophysiological mechanism of depression particularly in females.

To give full strength to our experimental paradigm, that proper MAPK signalling is highly important and gender-specific in normalization of the impaired behaviour and adequate fluoxetine action in stressed animals, we correlate present findings with hippocampal BDNF gene expression determined in both genders (Mitic et al. 2013). ERK, JNK and p38 could be downstream and upstream targets of BDNF, which is consider to be one of the pivotal molecules in depression (Duman and Monteggia 2006; Calabrese et al. 2009), so the bond between MAPK signalling and BDNF could be crucial in terms for adequate stress response and proper antidepressant action. Indeed, in stressed females we found that decrease in all three nuclear pMAPKs correlated with decreased BDNF expression, which promoted impaired behaviour. Concomitant fluoxetine treatment reversed all changes evoked by stress. In contrast, in stressed males, decreased BDNF expression and depressive-like behaviour were followed with increased nuclear pJNK and pERK signalling, suggesting that regulation of BDNF and induction of depressive-like behaviour were probably independent or at least not crucially affected by pMAPK signalling. Fluoxetine treatment supported this notion, since normalization of depressive-like behaviour was not followed with changes in BDNF expression, while its effect on MAPK signalling was restricted on cytosolic pJNK levels.

Taken together, our results contribute to better understanding of pathophysiological mechanism of stress-related disorders via nuclear pMAPK/BDNF system particularly from sex perspective and may contribute to the development of more selective antidepressant therapies, especially in females.