Expression of BDNF and trkB in the hippocampus of a rat genetic model of vulnerability (Roman low‐avoidance) and resistance (Roman high‐avoidance) to stress‐induced depression

Abstract Introduction The selective breeding of Roman High‐ (RHA) and Low‐Avoidance (RLA) rats for, respectively, rapid versus poor acquisition of the active avoidance response has generated two distinct phenotypes differing in many behavioral traits, including coping strategies to aversive conditions. Thus, RLA rats are considered as a genetic model of vulnerability to stress‐induced depression whereas RHA rats are a model of resilience to that trait. Besides the monoamine hypothesis of depression, there is evidence that alterations in neuronal plasticity in the hippocampus and other brain areas are critically involved in the pathophysiology of mood disorders. Materials and Methods Western blot (WB) and immunohistochemistry were used to investigate the basal immunochemical occurrence of brain‐derived neurotrophic factor (BDNF) and its high‐affinity tyrosine‐kinase receptor trkB in the dorsal and ventral hippocampus of adult RHA and RLA rats. Results WB analysis indicated that the optical density of BDNF‐ and trkB‐positive bands in the dorsal hippocampus is, respectively, 48% and 25% lower in RLA versus RHA rats. Densitometric analysis of BDNF‐ and trkB‐like immunoreactivity (LI) in brain sections showed that BDNF‐LI is 24% to 34% lower in the different sectors of the Ammon's horn of RLA versus RHA rats, whereas line‐related differences are observed in the dentate gyrus (DG) only in the ventral hippocampus. As for trkB‐LI, significant differences are observed only in the dorsal hippocampus, where density is 23% lower in the DG of RLA versus RHA rats, while no differences across lines occur in the Ammon's horn. Conclusion These findings support the hypothesis that a reduced BDNF/trkB signaling in the hippocampus of RLA versus RHA rats may contribute to their more pronounced vulnerability to stress‐induced depression.

The hippocampus is functionally parcelled along its longitudinal septotemporal axis into the dorsal and ventral subregions: the dorsal hippocampus is preferentially involved in spatial learning and memory, while the ventral hippocampus plays a central role in the control of the stress response and anxiety (Tanti & Belzung, 2013).
Despite the numerous studies reporting on the involvement of BDNF in both depression and the therapeutic effects of antidepressants, there is a relative paucity of data regarding the localization of BDNF in the hippocampus of genetic animal models of stress-induced depression. However, previous studies have shown that, in rats, the bilateral infusion of BDNF into the dentate gyrus of the dorsal hippocampus produces antidepressant-like effects in behavioral depression models (Shirayama, Chen, Nakagawa, Russell, & Duman, 2002); moreover, in the dorsal hippocampus, chronic electroconvulsive treatment increases the acute electroconvulsive induction and prolongs the expression of BDNF and trkB mRNA (Nibuya et al., 1995), and also upregulates the expression of the gene encoding BDNF in the dentate gyrus granule cell layer (Ploski, Newton, & Duman, 2006). Hence, the K E Y W O R D S brain-derived neurotrophic factor, genetic model of depression, hippocampus, immunohistochemistry, Roman rat lines, trkB, western blot present study was designed to characterize the hippocampal distribution of BDNF and its receptor trkB in RHA and RLA rats under baseline conditions, with special focus on the dorsal hippocampus, using Western blot (WB) and immunohistochemistry techniques.

| Animals
Sixteen outbred male adult rats from each Roman line (weighing 300-380 g), aged 4 months, were obtained from the colony established in 1998 at the University of Cagliari, Italy (Giorgi et al., 2005). Rats were housed in groups of four per cage and maintained under temperatureand humidity-controlled environmental conditions (23°C ± 1°C and 60% ± 10%, respectively), under a 12-hr light-dark cycle, with lights turned on at 8:00 a.m., and with standard laboratory food and water available ad libitum. To avoid stressful stimuli resulting from manipulation, each rat was gently handled once daily for 3 days before sacrifice and the maintenance activities in the animal house were carried out by a single attendant. All procedures were carried out in compliance with the European Union (Directive 2010/63/EU) and the Italian national guidelines and protocols (D.L. 04/04/2014, n. 26) and approved by the Ethical Committee for Animal Care and Use of the University of Cagliari (No. 18/2014). Every possible effort was made to minimize animal pain and discomfort and to reduce the number of experimental subjects.

| Sampling
Immediately after sacrifice by guillotine, the brains were rapidly dissected and processed for either WB or immunohistochemistry. For WB, brains were cooled in dry ice for 15 s, placed in a brain matrix and cut in 2 mm thick coronal slices using the stereotaxic coordinates of the rat brain atlas of Paxinos and Watson (1998) as a reference.
F I G U R E 1 Schematic representation of a rat brain coronal section ( Figure 33 from Paxinos & Watson, 1998). The red circle denotes the area of dorsal hippocampus taken for western blot analysis by means of a 2.5 mm punch

| Western blot
Total protein concentrations were determined as described by Lowry, Rosebrough, Farr, & Randall (1951)  licor.com/bio/products/software/image_studio_lite/). Columns and bars denote the mean ± SEM of eight rats in each experimental group. *: p < .05; ***: p < .001 versus the RHA group (two-tailed Student's t test for independent samples) P, for BDNF and trkB respectively, Santa Cruz Biotechnology, Santa Cruz, CA, USA). Slides were observed with an Olympus BX61 microscope and digital images were captured with a Leica DFC450C camera.

| Statistical analysis
WB and immunohistochemical data were statistically evaluated using the Student's t test for independent samples.

| Western blot
The anti-BDNF antibody recognizes a protein band with a relative mw of about 13 kDa (Figure 2), in agreement with the reported mw of the monomeric form of the protein (Rosenthal et al., 1991). As shown in Figure 2, the molecular mass of the hippocampal BDNF protein is very T A B L E 1 Statistical analysis of BDNF-and trkB-like immunoreactivity by western blot and immunohistochemistry in the hippocampus of RLA and RHA rat lines close to that of the rhBDNF (Leibrock et al., 1989). In tissue homogenates from the dorsal hippocampus the relative levels of BDNF protein (RLA = 0.39 ± 0.05 vs. RHA = 0.74 ± 0.08) were 48% lower in RLA versus RHA line (p < .001; Table 1).
The anti-trkB antibody recognized a protein band with a relative mw ≅ 140 kDa (Figure 2), consistent with the mw of the full-length, biologically active form of the receptor protein (Klein, Parada, Coulier, & Barbacid, 1989
Both the BDNF-and trkB-immunostainings mostly labeled neuronal proximal processes and nerve fibers distributed within the Ammon's horn and the dentate gyrus. A number of BDNF-and trkB-positive cell bodies were also observed. Although BDNF-and trkB-like immunoreactivities shared a similar distribution, they showed a differential density in the hippocampal subregions in both dorsal and ventral hippocampus. sectors. Some BDNF-positive perikarya were observed in the pyramidal, molecular and oriens layers (Figures 3e,f). In the CA3/ CA2 subfield, a band of BDNF-like immunoreactive punctate structures, that appeared to be denser in RHA than RLA rats, was observed in the stratum lucidum (Figures 3g,h) (Figure 4g,h). Rare trkBimmunolabeled neuronal cell bodies and/or their proximal processes were observed in the thickness of the Ammon's horn pyramidal layer ( Figure 4e) and in the hilus of the dentate gyrus (Figures 4g,h; 6e,f).
Densitometric analysis in the CA sectors of the hippocampus proper and in the dentate gyrus (Figures 5 and 7) revealed that line differences are limited to the dorsal hippocampus and particularly to the dentate gyrus where the density of trkB-like immunoreactive structures was 23% lower in RLA than RHA rats (RLA 169.1 ± 10.12 vs. RHA 259.2 ± 11.3; p < .0001), while no significant differences were found in the Ammon's horn (see Table 1).

| DISCUSSION
The main finding of the present study is that, under baseline conditions, the relative protein levels of BDNF and its selective receptor trkB, measured by WB, are significantly lower in the dorsal hippocampus of depression-vulnerable RLA rats versus depression-resistant RHA rats. Moreover, the densitometric analysis of immunostained brain slices revealed that the BDNF-LI is lower in the Ammon's horn whereas trkB-LI is lower in the dentate gyrus of RLA rats as compared to their RHA counterparts. These findings add experimental support to the neurotrophic hypothesis of depression, which postulates that decreased levels of BDNF in the hippocampus play a critical role in the pathophysiology of this mental disorder (Duman & Monteggia, 2006;Kozisek, Middlemas, & Bylund, 2008). Accordingly, by means of the same antibody used in this study, the relative levels of the BDNF mature protein in the hippocampus of the Flinders sensitive line of rats, a genetic model of depression, have been shown to be significantly lower than in their Flinders resistant controls (Elfving et al., 2010).
Moreover, a decrease in the BDNF mature protein has been reported in the ventral hippocampus and frontal cortex of rats with a genetic deletion of the serotonin transporter (Calabrese et al., 2013). A selective deletion of the BDNF-encoding gene within the mouse dentate gyrus also results in a reduction in antidepressant efficacy (Adachi, Barrot, Autry, Theobald, & Monteggia, 2008) and, conversely, the infusion of BDNF into the lateral cerebral ventricle (Hoshaw, Malberg, & Lucki, 2005) or the dorsal hippocampus (Shirayama et al., 2002) elicits antidepressant effects in behavioral models of depression.
There is also growing evidence that antidepressant treatments may exert some of their therapeutic effects by increasing BDNF expression levels and affecting BDNF transcription in the hippocampus.
Thus, (i) electroconvulsive brain stimulation, a treatment of choice for medication-resistant depression, increases the hippocampal contents of BDNF and trkB mRNAs (Altar, Whitehead, Chen, Wörtwein, & Madsen, 2003;Angelucci, Aloe, Jiménez-Vasquez, & Mathé, 2002;Nibuya et al., 1995), particularly in the dorsal hippocampus (Ploski et al., 2006), (ii) antidepressant treatment rapidly elevates the content of BDNF mature protein via posttranscriptional mechanisms, prevents the stress-induced decrease in the hippocampal concentration of BDNF mRNA and counteracts the depression-associated alterations in neural plasticity by normalizing BDNF in the rodent prefrontal cortex and hippocampus (Baj et al., 2012;Kozisek et al., 2008;Musazzi et al., 2009). In keeping with these experimental findings, postmortem samples from clinically depressed patients treated with antidepressant medications show increased BDNF immunoreactivity in the hilus, the dentate gyrus and the supragranular regions of the hippocampus as compared with antidepressant-untreated controls (Chen, Dowlatshahi, MacQueen, Wang, & Young, 2001).
Interestingly, it has been shown that different antidepressant treatments, such as physical activity (Baj et al., 2012;Kozisek et al., 2008), The lower levels of BDNF-like immunoreactivity in the Ammon's horn of RLA versus RHA rats, especially in the CA3 sector, are consistent with the hypothesis that, in the RLA line, CA3 neurons may have a slower synthesis rate of BDNF. In the CA3 sector BDNF is known to be both locally produced and anterogradely transported along the mossy fiber axons of the granule cells in the dentate gyrus.
Thus, the possible slower production of BDNF protein in RLA versus RHA rats, as also suggested by the lower density of BDNF-LI in the dentate gyrus of the ventral hippocampus, may in turn lead to a reduced target-derived support to promote the synaptic contacts with the mossy fibers. Interestingly, a selective increase in both BDNF mRNA and protein in the CA3 and CA2 sectors has been reported upon repeated antidepressant treatment (De Foubert et al., 2004).
The significantly lower density of trkB immunoreactivity observed in the dentate gyrus of the dorsal hippocampus of RLA versus RHA rats further suggests that the functional tone of BDNF-trkB signaling in the dentate gyrus may be less intense in RLA rats as compared to their RHA counterparts.
The immunolabeling of both BDNF and trkB in the dentate gyrus is localized to cell bodies and nerve fiber networks. The occurrence of BDNF-containing granule cells is in line with the reported synthesis of BDNF mRNA in this hippocampal region (Phillips et al., 1991).
However, the very small number of immunoreactive granule cells in our preparations does not allow us to evaluate possible quantitative differences in their occurrence between RLA and RHA rats. According to the neurotrophic hypothesis of depression, the biological substrate at the basis of the effectiveness of antidepressants is the integration of the newly born granule cells to the hippocampal neural circuitry (Baj et al., 2012;Duman & Monteggia, 2006) by establishing the typical synaptic connections of mature granule cells, i.e., to receive projections from the entorhinal cortex through their dendritic arbors extending in the molecular layer and to establish axo-dendritic synapses with the CA3 pyramidal neurons (Amaral, Scharfman, & Lavenex, 2007 in the rat hippocampus after chronic antidepressant treatment (Nibuya et al., 1995;Shirayama et al., 2002), and studies in mice with selective BDNF-encoding gene deletion in specific hippocampal subfields have demonstrated that BDNF expression in the dentate gyrus, but not in the CA1 sector, is essential for the effectiveness of antidepressants due to its possible involvement in supporting survival and differentiation of newborn granule cells (Adachi et al., 2008). Notably, BDNF/ trkB-LI labels other cell types in the dentate gyrus of the Roman rats, such as those located in the subgranular layer and in the hilus, some of which resembling the dentate pyramidal basket cells and the mossy cells described by Amaral et al. (2007), that may further contribute to the modulation of the hippocampal functional connectivity. In our preparations, at the level of the dorsal hippocampus, both BDNF-and trkB-like immunoreactive nerve fiber plexuses appear to be denser in the inner third of the molecular layer, where the axons originating from the hilar mossy cells play a commissural/associational role in the rat dentate gyrus (Amaral et al., 2007 and references therein).
The origin of the nerve fibers in the inner third of the molecular layer as well as the extent to which the possible commissural/associational projections contribute to the line-dependent differences in trkB-LI and, therefore, to the line-related differences in BDNF signaling in the dentate gyrus, are not completely understood.
Studies on the BDNF targeting on hippocampal CA3 dendrites (Baj, Pinhero, Vaghi, & Tongiorgi, 2016;Baj et al., 2012;Righi, Tongiorgi, & Cattaneo, 2000) further suggest the possibility of concurrent mechanisms of BDNF-mediated trophic support. Thus, upon the anterograde transport along the mossy fibers (Altar & DiStefano, 1998;Altar et al., 1997), the endogenous BDNF secreted along the neuronal activity may contribute to local mechanisms of trophic support that drive the accumulation of BDNF and trkB mRNAs in specific subcellular compartments of the CA3 pyramidal neurons (Righi et al., 2000). Further studies are warranted to assess the colocalization of BDNF and the trkB receptor protein and to characterize the hippocampal neural circuitry involved in the trophic activity of BDNF/trkB signaling in the Roman rats.
As mentioned above (see Introduction), RLA and RHA rats represent two divergent phenotypes respectively susceptible and resistant to develop depression-like behavior under aversive environmental conditions. Thus, in the forced swim test (FST), RLA rats exhibit a depression-like behavior characterized by greater immobility and fewer climbing counts when compared to their RHA counterparts, which exhibit proactive coping characterized by a shorter immobility time and more frequent climbing counts (Piras et al., 2010). Notably, subacute as well as chronic treatments with clinically effective antidepressant drugs decrease immobility and increase active behaviors (i.e., climbing and swimming) in RLA rats, but do not modify the performance of RHA rats (Piras et al., 2010(Piras et al., , 2014. Therefore, the Roman rats may be used to identify and characterize neural substrates and mechanisms, such as BDNF/trkB signaling, underlying the vulnerability to stress-induced depression as well as the molecular adaptations that mediate the resistance to such changes and the mechanisms of action of antidepressant drugs.

DISCLOSURE
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.