Several studies have demonstrated that inappropriate adaptations to stressful stimuli may cause temporary or even permanent alterations in the psychophysiological responses (Del Giudice et al., 2011; Russo et al., 2012), generating affective disorders (Hoffman et al., 2008; Shin and Liberzon, 2010; Leuner and Shors, 2013; Pérez-Cano et al., 2020). Moreover, chronic stress conditions can lead also to various types of cognitive impairment in humans and rodents, such as object recognition, and spatial and declarative memory deficits (Lupien et al., 1994; McEwen and Sapolsky, 1995; Seeman et al., 1997; Rimmerman et al., 2017; Calpe-López et al., 2022).

It should be noted that one of the most common psychopathologies associated with chronic stress is depression disorder (Slavich et al., 2020; Banasr et al., 2021; Chojnowska et al., 2021). Since depression disorder is characterized by heterogeneity of symptoms in humans, several findings have demonstrated that animal models should also involve a set of tests for trying to point out biological features in search for underlying clinical signs (Cathomas et al., 2015; Planchez et al., 2019; Calpe-López et al., 2022).

Still from the experimental point of view, the social defeat stress (SDS) test was developed to investigate some behavioral, physiological, and neurochemical consequences induced by intra-species social confrontations (Toyoda, 2017), generating anxiogenic-like responses in rodents (Rodgers and Cole, 1993; Fuchs and Flügge, 2002). Subsequently, this paradigm was optimized and standardized to investigate various behavioral effects produced by chronic social defeats in mice (Golden et al., 2011). Complementarily, our group has used the SDS to investigate the neuronal basis of defensive behaviors in the attacked mouse (Faria et al., 2020; Victoriano et al., 2020; Santos-Costa et al., 2021).

Concerning the neurobiological modulation of stress, the medial prefrontal Cortex (mPFC) has gained special interest among the brain areas implicated in the modulation of defensive responses (Sullivan and Gratton, 2002; Cerqueira et al., 2008). Relevant findings have characterized the mPFC functional lateralization (Lee et al., 2015; Costa et al., 2016), specifically, Santos-Costa et al. (2021) demonstrated that the changing of anxiety-like behavior depends on hemisphere manipulation.

In mammals, stress can be modulated by top-down inhibitory processes, in which mPFC modulates subcortical structures, e.g., bed nucleus of stria terminalis (BNST), amygdala, and hippocampus (Maier, 2015). Concerning these areas, the BNST is one of the main structures with neural circuits in response to stress, modulation of affective behavior, and stress-reward integration (de Oliveira et al., 2000; de Kloet et al., 2005; Faria et al., 2020). Moreover, studies in rodents suggest that the BNST is also associated with contextual fear and sustained anxiety-like responses (Davis et al., 2010; Avery et al., 2016; Miles and Maren, 2019).

On the other hand, the amygdaloid complex has been commonly studied due to its role in emotional responses, being hyperactive under stressful stimuli (Sah et al., 2003; Yan et al., 2021). Interestingly, a recent study showed that the amygdala integrates information from memory systems, and sensory areas and reciprocally projects to cortical and other subcortical regions (Chen et al., 2021).

Moreover, the hippocampus is also an important key in the modulation of anxiety and mood responses (Kim and Diamond, 2002), and its function is directly affected by stressful stimuli (Moser et al., 1995). Initially, the dorsal hippocampus (DH) was mainly related to location and spatial memory processes, whereas the ventral hippocampus (VH) was more related to playing a role in stress and emotional responses (Bannerman et al., 2004; Cernotova et al., 2021). However, recent studies have emphasized that both DH and VH play an overlapping function on location and spatial memory processes, as well as on affective and emotional behaviors (Riaz et al., 2017; Hartmann et al., 2019; Sant’Ana et al., 2019; Almeida et al., 2020).

Furthermore, the exposure of animals to stress may induce morphological deficits in the hippocampus (Nasca et al., 2019), and animals susceptible to the stress showed smaller hippocampal volume (Rosene and Van Hoesen, 1977; Lee et al., 2009; Rahman et al., 2016). Similar to the BNST and amygdaloid complex, the mPFC and hippocampus also have reciprocal projections that regulate the stress response (Goldman-Rakic et al., 1984; Padilla-Coreano et al., 2016), forming a dual pathway system.

Several technical tools have been used to characterize the involvement of brain areas in the modulation of behavioral, physiological, or pharmacological responses. In this context, the quantification of the Fos family protein expression in brain areas has widely been used for this purpose (Nestler and Carlezon, 2006; Robison and Nestler, 2011), with an emphasis on the ΔFosB that has great molecular stability, being a cumulative product of a sustained stimulus (e.g., chronic stress) (Citri and Malenka, 2008).

In addition to identifying how active a brain region is, it is worth checking neurons’ phenotypes involved in stress modulation. Considering the higher predominance of glutamatergic neurons in the BNST (Kim and Kim, 2021), amygdala (Reznikov et al., 2007), and hippocampus (Sanacora et al., 2012), the glutamate neurotransmission seems to be a strong candidate in the modulation of SDS neurofunctional consequences. On this wise, Ca²⁺/calmodulin-dependent protein kinase II (CaMKII) is the most abundant kinase present in excitatory synapses in the mammalian brain, phosphorylating a large number of synaptic proteins (Jin et al., 2013).

In this study, we hypothesized that exposure to SDS can induce depressive-like behaviors and cognitive impairment in addition to social avoidance and anxiogenesis responses. Furthermore, these behavioral changes evoked by SDS would be regulated via glutamatergic neurotransmission to the BNST, amygdaloid complex, and hippocampus. Aiming to confirm this hypothesis, we investigated the influence of social defeat stress protocol in mice on (i) the avoidance behavior assessed in the social interaction test and (ii) the anxiety-like behavior in the elevated plus-maze, and the open field tests (iii) the short-term memory in the object recognition test, (iv) depressive-like behaviors in the: body weight gain, coat state test, sucrose splash test, nesting building, novel object exploration test, and sweet drive test, (v) the presence of projections from the mPFC to the BNST, amygdala, and dorsal and ventral hippocampus, and (vi) ΔFosB, CaMKII as well as ΔFosB + CaMKII labeling in neurons located in the BNST, amygdaloid complex, dorsal and the ventral hippocampus of mice.

Corroborating previous findings (Golden et al., 2011; Dias et al., 2014; Carnevali et al., 2020; Faria et al., 2020; Victoriano et al., 2020; Lee et al., 2021; Nishiguchi et al., 2021; Santos-Costa et al., 2021), the behavioral results shown here demonstrated that SDS mice displayed increased defensive and anxiety-like behaviors, characterized by enhancement in social avoidance, reduced EPM open arm and OF center exploration time. Importantly, the exposure of mice to a 10-day SDS protocol did not influence general activity, since both closed-arm entries in the EPM and the number of crossings in the OF were not statistically different in comparison with NAI mice. Moreover, our findings reveal that the stressed mice displayed short-term memory impairment assessed in the OR test. Since the SDS mice exhibited reduced exploration in the novel object test on day 10, indicating that the reduced exploration of the novel object observed in the OR test could be influenced by object novelty rather than short-term memory impairment, it is essential to emphasize that novel object and operating room tests differ in their procedures. The novel object test measures the exploration time of a new object placed in the mouse’s cage, whereas the OR test involves a 10-min habituation phase in an arena, followed by a training phase carried out 24 h later with two identical objects (T1) and 2 h later, one of the familiar objects is replaced by a new object (T2). Furthermore, no significant differences were found between groups at T1 (data not shown). There is wide literature evidence indicating the potential mechanisms by which stress situations lead to memory impairments, and most of them have been specially related to the role of the hippocampus in this cognitive process (Moreira et al., 2016). In this sense, the reasons for changes in memory formation involve a decrease in both apical dendritic branching and total dendritic length (Watanabe et al., 1992; Schwabe et al., 2012), which alter the consolidation of the memory of the axis mPFC-Hippocampus to the dorsal striatum (Kim and Cho, 2017). It is important to highlight that this type of memory impairment is an adaptive process wherein only stress-related memories are acquired and consolidated (Kim and Diamond, 2002; de Kloet et al., 2005; Schwabe et al., 2012).

Interestingly, although SDS mice did not present body weight alteration during the 10-days of SDS, they gained more body weight than the control group during the social isolation period, a finding that corroborates the previous study, which observed an increase in food intake and body weight gain after chronic psychosocial stress in male mice (Moles et al., 2006). A similar result was demonstrated when mice were exposed to the social defeat/overcrowding protocol, where stressed mice showed weight gain during and after the agonistic interactions (Keenan et al., 2018). Interestingly, previous findings demonstrated that chronic stress protocol can induce changes in the ghrelin system, the hormone directly involved in eating and metabolism (Cummings et al., 2002; Patterson et al., 2013). Additionally, recently Dulabi et al. (2020) showed that the mice in social isolation condition have the ghrelin plasma levels accentuated compared to the animals living in groups. It is noteworthy that the NAI group exhibited a decrease in body weight gain during social isolation, indicating that this housing condition only affects weight gain in animals that were not subjected to SDS. Moreover, corroborating previous reports (Nollet et al., 2013b; Nazir et al., 2022), SDS mice displayed a worsening coat state compared to NAI animals, a result that emphasizes that a 10-day SDS exposure and a 10-day social isolation induces some apathy degree.

In contrast, this type of chronic stress did not change mice’s response in the splash test, i.e., no differences between stressed and non-stressed animals were observed in the latency, total, and frequency of grooming of the dorsal coat splashed with sucrose. The splash test is commonly used to measure stress-induced apathy/anhedonia deficit in rodents (Isingrini et al., 2010; Baptista-de-Souza et al., 2020). In this context, previous findings have demonstrated that SDS induces lower sucrose preference (Iñiguez et al., 2014). In addition, classical studies using unpredictable chronic mild stress have demonstrated apathy/anhedonia-like response when evaluated in the sucrose preference test (Willner, 2017; Planchez et al., 2019). Thus, although we expected that the exposure of mice to a 10-day SDS, followed by 10-day social isolation, would reduce grooming response in the sucrose splash test, it is important to highlight that inconsistent results have also been reported elsewhere. For instance, Mouri et al. (2018) have demonstrated no differences in the sucrose preference test using young mice exposed to a modified chronic SDS protocol (10 days with 10 min of physical interaction only, without any other sensory threat). Although the present work indicates that depression-like behaviors such as anhedonia (lack of response to a natural reward) (Kalueff et al., 2016; Zhou et al., 2022) do not change following a 10-day SDS and 10-day social isolation, it is important to consider that the type of experimental protocol may be crucial to induce apathy/anhedonia-like responses.

Furthermore, in the analysis of the nest-building test, we did not find differences between groups. Although the decrease in the nest building quality has been well described after acute stress protocols (Thodberg et al., 2002; Nollet et al., 2013a), chronic and subchronic stress tests may require a more complex interpretation. For instance, Otabi et al. (2016, 2020) have demonstrated that although mice can build nests under chronic stress, they spend more time on the building as well an increased latency to start building compared to control mice. Furthermore, those authors observed that the time elapsed between the stress session and the nest evaluation also may influence the quality of the nest.

A potential relationship between the behavioral effects induced by SDS with brain area activities was investigated in some animals that were exposed to the social interaction test (SIT) only (i.e., SDS animals that were not exposed to the other behavioral tests). We have recently demonstrated the specificities of mPFC subareas modulating the SDS responses, particularly the role of the prelimbic (PrL) area (Santos-Costa et al., 2021). As previously described by Vertes (2004), this region projects sparse efferents to BNST. As shown in the supplementary results (Supplementary Figure 1) both left and right PrL mPFC project to BNST. Interestingly, Faria et al. (2020) also demonstrated the interplay between mPFC (specifically the right side) and BNST wherein the blockade of NMDA receptors in the BNST impairs the anxiogenic-like effect provoked by intra-right mPFC injection of NOC-9, an nitric oxide donor (Seccia et al., 1996; Del Carlo and Loeser, 2002). Furthermore, the results of immunofluorescence analysis in the BNST of stressed mice showed increased bilateral expression of ΔFosB and ΔFosB + CaMKII double labeling without alterations in CaMKII measures. Considering that CaMKII is a relevant marker of excitatory neurons in several areas of the forebrain (Liu and Murray, 2012), it is not unreasonable to suggest that even as we did not observe an increase in the CaMKII in stressed mice, our double-labeling analysis suggests that the neuronal activation induced by SDS protocol reached glutamatergic neurons in the BNST.

Previous studies have shown that the mPFC projects to the amygdaloid complex, and the PrL area is one of the mPFC subregions that send efferents to the amygdala (Marek et al., 2019; Song et al., 2019). On this wise, our findings corroborate these previous studies demonstrating that the amygdala receives bilaterally projections from PrL (see Supplementary Figure 1). Interestingly, the results obtained with the immunofluorescence analysis within the amygdaloid complex of the defeated animals showed a bilateral overexpression of ΔFosB, CaMKII, and ΔFosB + CaMKII. The findings concerning the rise of the activation within the amygdaloid nucleus in SDS mice are sustained by previous studies with various social defeat stress protocols (Numa et al., 2019; Dulka et al., 2020). However, as far as we know, no study has yet demonstrated that this increase of functionality in this area (observed through ΔFosB measures) goes along with glutamatergic neuro markers, as well as CaMKII. It is important to highlight that this result not only reveals that the activation occurs in the glutamatergic neurons, but the protocol was able to induce an increase of CaMKII expression in the amygdala. Considering that the classical role of CaMKII involves the regulation of synaptic function, including long-term potentiation, and that occurs in a Ca²⁺ -dependent manner (Lisman, 2003; Okamoto et al., 2009), we suggest that these changes may collaborate with the neuroplasticity of this brain area as a consequence of chronic stress condition. Moreover, considering that chronic stress leads to an increase of glutamatergic release in the basolateral nucleus of the amygdala, and consequently, this effect may cause enhanced brain-derived neurotrophic factor (BDNF) expression and dendritic outgrowth (Vyas et al., 2002; Boyle, 2013), our findings add more evidence about the mechanism involved in the amygdala neurofunctional changes induced by chronic stress. Although attractive, further studies (e.g., using other neurotrophic markers and investigating distinct amygdaloid subnuclei) are needed to clarify this hypothesis.

Additionally, considering that previous studies have shown that activating the mPFC reduces the activity of subcortical structures (Quirk et al., 2003; Cerqueira et al., 2007; Johnstone et al., 2007; Johnson et al., 2019; Yang et al., 2022), it is plausible to suggest that the upregulation of ΔFosB in the BNST and amygdala could be related to the inhibition of left PrL in SDS animals, as previously reported (Santos-Costa et al., 2021).

In the immunofluorescence analyses shown here, we found differences in the hippocampal subregions, both for the expression of ΔFosB and CaMKII and for ΔFosB + CaMKII double staining. In the dorsal hippocampus, there was a bilateral decrease in ΔFosB expression in stressed animals, whereas no differences in CaMKII expression and double staining were recorded, indicating a possible neuronal inhibition of such a region. Although our results contrast with previous findings showing an opposite effect on the hippocampal dorsal dentate gyrus (Cooper et al., 2017), they are consistent with the effect of stress on the hippocampus, such as dendrite atrophy or the neurogenesis suppression in this region (Watanabe et al., 1992; Galea et al., 1997; Nasca et al., 2019). On the other hand, our results also indicate that the affected neurons may be not glutamatergic ones, indicating that other neurotransmissions (e.g., serotonergic, cholinergic, adrenergic, and/or endocannabinoid) might be involved (Amaral and Kurz, 1985; Verney et al., 1985; Hartmann et al., 2019).

Different from the results obtained in the dHPC, the immunofluorescence in the vHPC showed a bilateral increase of ΔFosB and CaMKII expression in the stressed animals, indicating an important role of this region in the modulation of stress-related responses. Given that the exposure to chronic SDS also induced defensive- and anxiety-like behaviors, we suggest that these anxiogenic profiles induced by stress might have the involvement of glutamatergic neurons located in the vHPC. If so, these results corroborate previous findings showing the involvement of the ventral portion of the hippocampus in the modulation of anxiogenic behavior (Henke, 1990; Bannerman et al., 2004; Tovote et al., 2015) as well as the role of the glutamatergic neurons in its control (Robison and Nestler, 2011). Interestingly, present results also showed a bilateral decrease in ΔFosB + CaMKII colocalization in the vHPC of stressed animals, indicating that besides stimulating glutamatergic neurons, chronic SDS also activates other neurotransmission systems (e.g., serotonergic, noradrenergic and/or endocannabinoid) (Köhler et al., 1981; Haring and Davis, 1985; Hudson et al., 2018) that might play an inhibitory role on glutamatergic neurons.

Although the neurotransmission for both hippocampal areas is not widely clear, the opposite ΔFosB expression recorded in the dHPC and vHPC suggests some clues about the role of the hippocampus in the expression of the coping strategy. We suggest that the dHPC, a brain area widely related to the spatial navigation function (Bannerman et al., 2004), is inhibited, while the vHPC, an anxiety-related area (Cernotova et al., 2021), is overactivated under chronic SDS. Together, these results suggest that chronic SDS activates vHPC while inhibiting dHPC, raising a hypothesis that the activation of the vHPC might lead to dHPC inhibition. In this context, corroborating previous findings (Cerqueira et al., 2007, 2008; Lee et al., 2015), we have recently demonstrated that the left PrL seems to exert an inhibitory action on the right PrL (i.e., activation of inhibitory interneurons) (Santos-Costa et al., 2021). If so, a similar process could also occur in the hippocampus, but instead of an interhemispheric inhibition, a ventral-dorsal one. Thus, in non-stressed conditions (control), the vHPC neurons might be inhibited, preventing stress-related memory prioritization, which is unnecessary in basal conditions. Nevertheless, chronic social defeat induces hippocampal changes (Cerqueira et al., 2007, 2008), specifically in the dHPC, and disinhibits the vHPC, changing the stress-related memory prioritization, resulting in declarative memory impairment and anxiogenic-like behaviors as shown in experiment 1.

Taking these results together, we suggest that SDS protocol (i) induces an increase of (a) defensive (i.e., decreasing social interaction) and (b) anxiety-like (i.e., reduction of EPM open arm and central OF exploration time) behaviors, and provoked (c) memory impairment without eliciting a clear correlation with the anhedonic response; (ii) promotes a bilateral increase of BNST and amygdala activation pattern, which seems to be associated to glutamatergic neurotransmission; and (iii) can differentially alter the neuronal expression of the dorsal and ventral hippocampus, having opposite consequences on neuronal activation in each subregion, i.e., while the exposure of mice to chronic SDS seems to activate the vHPC, resulting in increased defensive- and anxiety-related behaviors, this stressful situation would inhibit the dHPC, leading to memory impairment. The hypothesis that these behavioral and neuronal effects induced by chronic SDS are coordinated by neuronal projections from the right and left PrL mPFC to the BNST, amygdala, and hippocampus (Supplementary Figure 1) remains to be determined.

The present study demonstrates that chronic SDS can induce (i) social avoidance, (ii) defensive/anxiogenic-like behaviors, and (iii) memory impairment while provoking (iv) distinct patterns of neuronal activation on BNST, amygdala, and dorsal/ventral hippocampus of mice. On this wise, our study presents new advances for understanding the effects of chronic social stress on emotional, affective, and memory states, while collaborating with the discovery of potential brain targets for treatment options for these types of maladaptive disorders.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

The animal study was reviewed and approved by the All experimental protocols were conducted according to the ethical principles of the Brazilian National Council for the Control of Animal Experimentation (CONCEA) and approved by the local Research Ethics Committee (CEUA/FCF -UNESP: 23/2020 and 25/2020).

VC, JCR, DB-d-S, and LC-d-S performed the first experiment. SR and JA-Z performed the second experiment. JLR and CP contributed to the conceptualization review and writing the original draft. RN-d-S contributed to the conception, and design of the study, and reviewed and edited the manuscript. All authors contributed to the manuscript revision, and read, and approved the submitted version.