Pharmacology, Biochemistry and Behavior 128 (2015) 50–61

Contents lists available at ScienceDirect

Pharmacology, Biochemistry and Behavior

j ourna l homepage: www.e lsev ie r .com/ locate /pharmbiochembeh

Repeated forced swimming impairs prepulse inhibition and altersbrain-derived neurotrophic factor and astroglial parameters in rats

Milene Borsoi a,⁎, Camila Boque Antonio b, Liz Girardi Müller b, Alice Fialho Viana b, Vivian Hertzfeldt b,Paula Santana Lunardi c, Caroline Zanotto c, Patrícia Nardin c, Ana Paula Ravazzolo d,Stela Maris Kuze Rates a,b, Carlos-Alberto Gonçalves a,c

a Graduate Studies Program in Neurosciences, Federal University of Rio Grande do Sul, Porto Alegre, Brazilb Graduate Studies Program in Pharmaceutical Sciences, Federal University of Rio Grande do Sul, Porto Alegre, Brazilc Graduate Studies Program in Biochemistry, Federal University of Rio Grande do Sul, Porto Alegre, Brazild College of Veterinary Medicine, Federal University of Rio Grande do Sul, Porto Alegre, RS, Brazil

⁎ Corresponding author at: Graduate Studies ProgrUniversity of Rio Grande do Sul (UFRGS), 500 Sarment170, Porto Alegre, RS, Brazil. Tel.: +55 51 3308 5455; fax

E-mail address: mileborsoi@gmail.com (M. Borsoi).

http://dx.doi.org/10.1016/j.pbb.2014.11.0120091-3057/© 2014 Published by Elsevier Inc.

a b s t r a c t

a r t i c l e i n f o

Article history:Received 19 July 2014Received in revised form 12 November 2014Accepted 15 November 2014Available online 20 November 2014

Keywords:Repeated forced swimmingprepulse inhibitionbrain-derived neurotrophic factorglutamate uptakefrontal cortexhippocampus

Glutamate perturbations and altered neurotrophin levels have been strongly associatedwith the neurobiology ofneuropsychiatric disorders. Environmental stress is a risk factor for mood disorders, disrupting glutamatergicactivity in astrocytes in addition to cognitive behaviours. Despite the negative impact of stress-inducedneuropsy-chiatric disorders on public health, the molecular mechanisms underlying the response of the brain to stress hasyet to be fully elucidated. Exposure to repeated swimming has proven useful for evaluating the loss of cognitivefunction after pharmacological and behavioural interventions, but its effect on glutamate function has yet to befully explored. In the present study, rats previously exposed to repeated forced swimming were evaluatedusing the novel object recognition test, object location test and prepulse inhibition (PPI) test. In addition, quan-tification of brain-derived neurotrophic factor (BDNF) mRNA expression and protein levels, glutamate uptake,glutathione, S100B, GluN1 subunit of N-methyl-D-aspartate receptor and calmodulinwere evaluated in the fron-tal cortex and hippocampus after various swimming time points. We found that swimming stress selectivelyimpaired PPI but did not affect memory recognition. Swimming stress altered the frontal cortical and hippocam-pal BDNF expression and the activity of hippocampal astrocytes by reducing hippocampal glutamate uptake andenhancing glutathione content in a time-dependent manner. In conclusion, these data support the assumptionthat astrocytes may regulate the activity of brain structures related to cognition in a manner that alters complexbehaviours. Moreover, they provide new insight regarding the dynamics immediately after an aversive experi-ence, such as after behavioural despair induction, and suggest that forced swimming can be employed to studyaltered glutamatergic activity and PPI disruption in rodents.

© 2014 Published by Elsevier Inc.

1. Introduction

Altered glutamate levels and proteins associated with cellularsurvival participate in the cognitive impairment observed in patientswith various neuropsychiatric disorders, including major depression(Duman, 2014; Krystal et al., 2002; Nudmamud-Thanoi and Reynolds,2004; Tokita et al., 2012). Supporting this assumption, studies havebeen reported changes in several subunits of glutamate receptors indepressed patients (Nudmamud-Thanoi and Reynolds, 2004; Beneytoet al., 2007) and rapid antidepressant-like effects of various N-methyl-D-aspartate receptor (NMDAR) antagonists (Browne and Lucki, 2013;

am in Neurosciences, Federalo Leite Street, zip code 90050-: +55 51 3308 5243.

Tokita et al., 2012). In addition, astrocyte alterations have beenassociated with depression because altered parameters of astrocytefunction are observed in depressed patients (Czéh and Di Benedetto,2013; Popoli et al., 2012; Sanacora and Banasr, 2013), and rodentsshow depressive-like behaviours after pharmacological blockade ofastrocytic glutamate uptake in the amygdala (Lee et al., 2007), afterglial ablation (Banasr and Duman, 2008) and in response to glutaminedeficiency (Lee et al., 2013) in the prefrontal cortex. Along with theseglutamate perturbations, brain-derived neurotrophic factor (BDNF),which is among several signalling pathways underlying synaptic trans-mission and plasticity, has been strongly implicated in depressionneurobiology (Brunoni et al., 2008) and in the cognitive impairmentobserved in depressed patients (Castrén and Rantamäki, 2010; Dumanand Monteggia, 2006).

It is widely accepted that environmental stress is a risk factor formood disorders and that glutamatergic synapses are highly responsive

51M. Borsoi et al. / Pharmacology, Biochemistry and Behavior 128 (2015) 50–61

to environmental stimuli (Popoli et al., 2012). Both glutamate transmis-sion and BDNF levels can be significantly impacted by exposure to stress(Bath et al., 2013; Duman and Monteggia, 2006, Popoli et al., 2012).Despite the negative impact of stress-induced neuropsychiatric disor-ders on public health, the molecular mechanisms underlying theresponse of the brain to stress have not yet been fully elucidated. Expo-sure to repeated swimming has been proven as an effective tool toevaluate the loss of cognitive function after various pharmacologicaland behavioural interventions (Abel and Hannigan, 1992; Naudon andJay, 2005; Porsolt et al., 2010). This test has been widely used to predictthe antidepressant effects of several compounds (Porsolt et al., 1978a;Porsolt et al., 1978b; Cryan et al., 2005) and can induce various neuro-chemical changes observed both in neuropsychiatric diseases(Briones-Aranda et al., 2005; Rada et al., 2003; Sequeira-Cordero et al.,2014; Zucker et al., 2005) and in animal models of neuropsychiatricdiseases based on environmental stress exposure (Anisman andMatheson, 2005; Krishnan and Nestler, 2011; Andolina et al., 2013).However, the effect of repeated swimming exposure on glutamatergicfunction has not been fully explored.

In the present study, we investigate the effect of repeated forcedswimming on cognitive behavioural tests and its role in glutamatergicregulation and neurotrophin levels in rats. Ratswere exposed to repeat-ed forced swimming before being subjected to behavioural testsputatively associated with neuropsychiatric disorders: novel objectrecognition, object location test, and prepulse inhibition (PPI). Wemeasured BDNF and the GluN1 subunit of NMDAR in the frontal cortexand hippocampus of rats at various time points after the swimmingsession, in addition to glutamate uptake. Moreover, we evaluatedother parameters associated with astrocyte function (glutathionecontent and S100B) in the aforementioned brain regions.

2. Results

2.1. Behavioural tests

2.1.1. Immobility behaviour increases during repeated forced swimming,but spontaneous locomotion is not affected

Fig. 1A shows the immobility duration of repeated forced swimmingperformed for three consecutive days. One-way repeated measuresANOVA showed a significant day effect (F9,29 = 14.561, p b 0.001).Post hoc Student-Newman-Keuls analysis revealed an increasedimmobility duration according to the day (day 1 vs. day 2, p = 0.017;day 1 vs. day 3, p b 0.001 and day 2 vs. day 3, p = 0.013).

To verify whether repeated forced swimming affected exploratoryparameters, the number of crossings and rearings were evaluatedduring the habituation phases (absence of objects) of both the novel ob-ject recognition (NOR) task and object location test (OLT), which were

Fig. 1. Immobility acquisition and spontaneous locomotion during repeated forced swimming. (Ameasures ANOVA followed by Student-Newman-Keuls analysis: *p b 0.05; #p b 0.05, and ⁎⁎⁎pevaluated during the habituation phase of memory tasks. Data were grouped by condition (exptest: n.s. (C) Number of rearings. Student’s t test: n.s. Data are expressed as the mean ± SEM (

administered 24 h after the last swimming exposure. Because the habit-uation phase of both the NOR task and OLT were performed identically,the data were grouped by condition (rats submitted to repeated forcedswimming and those that were not). No changes in the number ofcrossings (Student’s t test, t = 1.565, df = 46, p = 0.124) or rearings(Student’s t test, t = 0.799, df = 46, p = 0.428) (Fig. 1B and C, respec-tively) were detected.

2.1.2. Novel object recognition test results are not altered by repeated forcedswimming

Two-way repeated measures ANOVA showed neither a group effect(F1,39 = 0; p = 1.00) nor interaction (F1,39 = 1.010, p = 0.328) butdid show a significant object effect (F1,39 = 18.221, p b 0.001) onshort-term memory (STM). Post hoc analysis revealed that both thenon-stressed (p = 0.002) and repeated forced swimming (p = 0.033)groups spent more time exploring the novel object than the familiarone during the STM retention trial (Fig. 2A), indicating an absence ofcognitive impairment. The investigation ratio for the novel object wasalso not affected by repeated forced swimming (Student’s t test: t =0.013, df= 17, n.s.) (Fig. 2B). Likewise, the STMand long-termmemory(LTM) were unaffected by repeated forced swimming. Two-wayrepeated measures ANOVA showed no group effect (F1,39 = 0; p =1.000) and a significant object effect (F1,39 = 36.880; p b 0.001).Student-Newman-Keuls test indicated that both the non-stressed(p = 0.004) and repeated forced swimming (p b 0.001) groups spentmore time investigating the novel object (Fig. 2C). No significant differ-encewas observed in the investigation ratio of LTM (Student’s t test: t=1.403, df = 18; n.s.) (Fig. 2D). No significant difference in time spentexploring the sample objects during the acquisition phase of STM andLTM in either group was detected (data not shown).

2.1.3. Object location test results are not affected by repeated forcedswimming

The results of the object location test were not affected by repeatedforced swimming for any retention memory evaluated. Two-wayrepeated measures ANOVA revealed no significant group effect in eitherSTM (F1,27 = 0, p = 1.00) or LTM (F1,31 = 0, p = 1.00) (Fig. 4A and 4C,respectively). Student-Newman-Keuls analysis revealed that for bothSTM (p b 0.001) (Fig. 3A) and LTM (p = 0.003), rats exposed to forcedswimming spentmore time exploring the novel object than the familiarone, indicating an absence of cognitive impairment (Fig. 3C). Thepreference index for the relocated object was not affected during boththe STM (t = 0.718, df = 12, p = 0.486) and LTM (t = 0.526, df =14, p= 0.607) acquisition trials (Fig. 3B and D, respectively). As expect-ed, similar amounts of time were spent exploring the sample objectsduring the acquisition trial in both the STM and LTM trials of the non-stressed and repeated forced swimming groups (data not shown).

) Ratswere exposed to 10min of swimming on three consecutive days. One-way repeatedb 0.001. Twenty-four hours after the last swimming session, spontaneous locomotion wasosure to repeated forced swimming or no exposure). (B) Number of crossings. Student’s tn= 10 per group).

Fig. 2.Effect of repeated forced swimming onnovel object recognition test. The novel object recognitionwas initiated 24h after the last swimming session. STMand LTMwere evaluated indifferent animals. (A) STM accessed 90min after the acquisition trial. Two-way repeatedmeasures ANOVA followed by Student-Newman-Keuls analysis: *p b 0.05. (B) Preference index ofnovel object in the STM. Student’s t test: n.s. (C) LTMaccessed 24 h after the acquisition trial. Two-way repeatedmeasures ANOVA followed by Student-Newman-Keuls analysis: **p b 0.05and ***p b 0.001. (D) Preference index for the novel object in the LTM test. Student’s t test: n.s. Data are expressed as themean± SEM. (n=10per group). STM, short-termmemory; LTM,long-term memory.

52 M. Borsoi et al. / Pharmacology, Biochemistry and Behavior 128 (2015) 50–61

2.1.4. Prepulse inhibition is reduced after repeated forced swimmingThe PPI response was evaluated twice: before and after repeated

forced swimming. According to two-way repeated measures ANOVA,both the non-stressed and repeated forced swimming groups showedsimilar performance in the first PPI evaluation because no group effectwas found (F1,38= 0.046, p= 0.834). During the second PPI evaluation,a significant group effect was observed (F1,38 = 10.669, p = 0.001), and

Fig. 3. Effect of repeated forced swimming on the results of the object location test. The object loin different animals. (A) STM accessed 90 min after the acquisition trial. Two-way repeated meindex for the object in a novel position in the STM test. Student’s t test: n.s. (C) LTM accessed 2Newman-Keuls analysis: **p b 0.01. (D) Preference index for the object in a novel position ingroup). STM, short-term memory; LTM, long-term memory.

Student-Newman-Keuls analysis revealed a reduced inhibition of startleresponse at 68 dB (p= 0.046), 73 dB (p b 0.001) and 77 dB (p=0.004)prepulse intensities in the repeated forced swimming group comparedto non-stressed animals (Fig. 4A). Analysis of ASR, which represents areaction to the pulse alone, revealed no difference between groups inboth first (Student’s t test: t = 1.355, df = 11, p = 0.202) and second(Student’s t test: t = 1.221, df = 11, p = 0.247) evaluations (Fig. 4B).

cation testwas initiated 24 h after the last swimming session. STM and LTMwere evaluatedasures ANOVA followed by Student-Newman-Keuls analysis: ***p b 0.001. (B) Preference4 h after the acquisition trial. Two-way repeated measures ANOVA followed by Student-the LTM test. Student’s t test: n.s. Data are expressed as the mean ± SEM. (n = 10 per

Fig. 4. Effect of repeated forced swimming on prepulse inhibition of the acoustic startle response. Rats were submitted to PPI twice andwere exposed to repeated forced swimming or notexposed (non-stressed) between the two PPI evaluations. (A) Percentage of PPImeasured during the second evaluation. Two-way repeatedmeasures ANOVA followed Student-Newman-Keuls analysis: *pb 0.05, **p b 0.01 and ***pb 0.001. (B) ASRmagnitudemeasuredduring the secondevaluation. Student’s t test: n.s. The results are expressed as themean±SEM(n=6-8per group). PPI, prepulse inhibition; ASR, acoustic startle response.

53M. Borsoi et al. / Pharmacology, Biochemistry and Behavior 128 (2015) 50–61

2.2. Neurochemical analysis

2.2.1. mRNA expression and protein levels of BDNF are differentiallyaffected by repeated forced swimming.

Taking account that BDNF levels are involved in the cognitiveimpairment found in stress-related disorders (Bath et al., 2013), themRNA expression and protein levels of this neurotrophin were evaluat-ed in the frontal cortex and hippocampus. Fig. 5 shows the effects ofrepeated forced swimming on the mRNA and protein expression ofBDNF measured using ELISA at various time points after the lastswimming session. Kruskal-Wallis one-way ANOVA revealed a signifi-cant change in BDNF expression in the frontal cortex (H = 8.647,df = 3, p = 0.034) but not the hippocampus (H = 2.449, df = 3, p =0.485). Dunn's Multiple Comparison test revealed that this BDNFreduction in the frontal cortex was significant in rats killed 24 h afterthe last swimming session compared with those of the non-stressedgroup (p b 0.05) (Fig. 5B). Although we found reduced mRNA expres-sion of BDNF in the frontal cortex, BDNF synthesis was not changed

Fig. 5.Effect of repeated forced swimming onmRNAexpression andprotein levels of BDNF. Ratswasmeasured in the frontal cortex (A) and hippocampus (B). Dunn'sMultiple Comparison test:as the lower and upper edges of the boxes, respectively; thewhiskers represent the range of theand hippocampus (D) using ELISA; One-way ANOVA: *p b 0.05; **p b 0.01; ***p b 0.001 and ##pneurotrophic factor.

because no difference was observed in the protein levels according toone-way ANOVA (F3,28 = 1.510, p = 0.236) (Fig. 5C). In contrast, anincrease in the hippocampal levels of BDNF protein was observed(one-way ANOVA, F3,28 = 9.995, p b 0.001). Student-Newman-Keulsanalysis revealed that this increase occurred at 24 h (p = 0.002), 48 h(p b 0.001) and 72 h (p = 0.038) after the last swimming sessioncompared with levels observed in non-stressed animals, and the levelsbegan to decline at 72 h (48 h vs. 72 h, p = 0.006) (Fig. 5D).

2.2.2. Protein levels of the GluN1 subunit of NMDAR and calmodulin are notaffected by repeated forced swimming

Considering that calcium entering through synaptic NMDARsmediates the activity-dependent neural survival and plasticity byinducing BDNF expression and that BDNF can also modulate NDMAR ex-pression in return (Caldeira et al., 2007a, 2007b), we evaluated the GluN1subunit protein levels of NMDAR using western blot. According to one-wayANOVA, protein levels of GluN1were not altered in the frontal cortex(F3,18 = 0.492, p = 0.693) or hippocampus (F3,19 = 0.796, p = 0.514)

werekilled24, 48, and 72 h after the last swimming session. ThemRNA expression of BDNF*p b 0.05. Boxplots represent themedian as themiddle line and the 25% and 75% quartilesfull data set (n=4 per group). BDNFprotein levelsweremeasured in the frontal cortex (C)b 0.01. Data are expressed as themean± S.E.M (n=7-10 per group). BDNF, brain-derived

54 M. Borsoi et al. / Pharmacology, Biochemistry and Behavior 128 (2015) 50–61

(Fig. 6A andB, respectively). Additionally, the protein levels of the calciumsensor calmodulin, which may modulate NMDAR activity (Rycroft andGibb, 2004), were also evaluated. However, no significant change wasfound in any of analysed structures (frontal cortex: F3,19 = 1.072, p =0.389; hippocampus: F3,19 = 0.388, p = 0.763) (Fig. 6C and D,respectively).

2.2.3. Astrocytic function: glutamate uptake is reduced and GSH levels areincreased in the hippocampus 48 h after the last swimming session

Astrocytes are themajor cell responsible for glutamate transport andregulation and are also essential for maintaining GSH levels, the mainantioxidant of the brain (Danbolt, 2001). Moreover, the S100B proteinis a calcium-binding protein involved in cognitive processes andproduced predominantly by astrocytes in the central nervous system(Donato, 2003). Consequently, the effects of repeated forced swimmingon astrocytic function were evaluated by evaluating glutamate uptakeand GSH and S100B levels.

One-way ANOVA revealed a significant change in glutamate uptakein acute hippocampal slices (F3,43 = 4.039, p = 0.013). Student-Newman-Keuls analysis revealed reduced glutamate uptake 48 h afterthe last swimming session compared with that in non-stressed animals(p = 0.007) (Fig. 7A). This alteration was transient because ratsdecapitated 72 h after the last swimming session presented similarlevels of glutamate uptake compared with those of the non-stressedgroup (p = 0.285). In contrast to the reduction in glutamate uptake,an increase in GSH levels was observed in the hippocampus (F3,29 =3.365, p = 0.033) (Fig. 7C), whereas no changes were found in thefrontal cortex (F3,29 = 0.745, p = 0.534) (Figure7B). Student-Newman-Keuls analysis revealed that GSH levels were significantlyincreased in the 48 h post-swimming group compared with the non-stressed (p = 0.043) and 24 h post-swimming (p = 0.038) groups.Similar to the glutamate uptake alteration, the change in GSH levels

Fig. 6. Effect of repeated forced swimming on the protein levels of subunit GluN1 of N-methyl-last swimming session. The GluN1 subunit protein level wasmeasured in the frontal cortex (A)and hippocampus (D); One-way ANOVA: n.s. Data are expressed as the mean ± S.E.M (n = 5-

was transient, and the 72 h post-swimming group did not show alteredGSH content. Regarding the S100B immunocontent, no significantdifferences were observed in either the frontal cortex (F3,35 = 0.237,p = 0.870) or the hippocampus (F3,35 = 0.688, p = 0.566) accordingto one-way ANOVA (Fig. 7D and E, respectively).

2.3. Correlations analysis

The immobility duration during the last swimming session showed asignificant negative correlation with BDNF measured 24 h after swim-ming for both mRNA expression (Pearson’s test: r = -0.930, p =0.034) and protein levels (Pearson’s test: r = -0.767, p = 0.026). Theother neurochemical and behavioural parameters analysed were notsignificantly correlated with the immobility duration.

3. Discussion

In the present study, we investigated the effects of repeated forcedswimming in tests that evaluate various cognitive components. Thisstress significantly reduced PPI at the three prepulse intensities investi-gated. Taking into account that the results of the novel object recogni-tion and object location tests were unaffected, our results indicate thatrepeated forced swimming selectively impairs pre-attentive process.Moreover, we demonstrated that mRNA expression of BDNF wasreduced in the frontal cortex 24 h after the last swimming exposure,and this change was negatively correlated with immobility. However,the change in protein levels was observed exclusively in the hippocam-pus. In addition, a reduction in glutamate uptake and increase in GSHcontent were observed in the hippocampus 48 h after the lastswimming session. These results indicate that swimming stress caninduce cognitive and transient glutamatergic changes, which in part

D-aspartate receptor (NMDAR) and calmodulin. Rats were killed 24, 48, and 72 h after theand hippocampus (B). The calmodulin protein level wasmeasured in the frontal cortex (C)7 per group). Experiments were performed in duplicate.

Fig. 7. Effects of repeated forced swimming on astrocytic function. Rats were killed 24, 48, and 72 h after the last swimming session. Glutamate uptakewas evaluated in acute hippocampalslices (A); One-way ANOVA followed by Student-Newman-Keuls analysis: *p b 0.05 (n = 5-7 per group). Glutathione (GSH) content was measured in the frontal cortex (B) and hippo-campus (C); One-way ANOVA followed by Student-Newman-Keuls analysis: *p b 0.05 and #p b 0.05. Data are expressed as the mean ± S.E.M (n = 5-7 per group). Experiments wereperformed in duplicate. The S100B immunocontent was measured in the frontal cortex (D) and hippocampus (E). One-way ANOVA: n.s. (n = 5-7 per group). Data are expressed asthe mean ± S.E.M. Experiments were performed in duplicate. GSH, glutathione.

55M. Borsoi et al. / Pharmacology, Biochemistry and Behavior 128 (2015) 50–61

resemble the changes observed during stress-related depressivedisorders.

The lack of a significant impairment in NOR performance among ratsexposed to repeated swimming has been demonstrated in other studies.Naudon and Jay (2005) classified rats into high- or low-immobilitygroups according to their immobility profile in the forced swimmingtest and then compared their performance in the NOR task. Both high-and low-immobility rats presented a similar discrimination index inthe NOR task. A study using repeated swimming demonstrated thatthe effect of this stress onmemory recognition seems to be time depen-dent because mice exposed to repeated swimming sessions presentNOR impairment 1 or 4 h after the last swimming exposure, but thiseffect is no longer presentwhen they are tested 24 h after the last swim-ming (Carey et al., 2009). Thereby, our results are in line with literaturesuggesting that repeated forced swimming does not impair NORperformance.

In addition, repeated forced swimming did not impair performancein the OLT, a task used to assess spatial memory. This result contradictsthose of other studies that have demonstrated deficits in spatial memo-ry after forced swimming using various tests. Abel andHannigan (1992)reported that exposure to swimming for ten days reduced the escaperesponse in the Morris water maze. Similarly, intermittent swim stress

impairs Morris water maze performance (Warner and Drugan, 2012).Naudon and Jay (2005) demonstrated that high- and low-immobilityrats, classified according to their performance in the forced swimmingtest, exhibit more errors in the classical radial maze elimination task.It is important to note that during both the Morris water maze andradial maze tests, the rodents must learn task rules to perform well onthe tests. This condition canmake it difficult to differentiate their spatialmemory ability from their ability to acquire the rules of the task. Theadvantage of the OLT used in the current study is that rats show aninnate tendency to spend more time exploring objects in a novel envi-ronment without the need to apply positive or negative reinforcement(Ennaceur and Delacour, 1988). Moreover, stress intensity, duration,controllability, and predictability all differentially affect spatial memory(Cazakoff et al., 2010); we also cannot exclude differences in the forcedswimming exposure protocols used in the mentioned studies and thepossibility that methodological issues associated with these memorytests influenced the apparently conflicting results.

PPI was significantly reduced after exposure to repeated forcedswimming, whereas the ASR magnitude was unaffected, indicatingthat the observed PPI deficit was not due to an altered perception ofthe startle stimulus itself. PPI is an important element of healthyinformation processing and is impaired in several psychiatric disorders

56 M. Borsoi et al. / Pharmacology, Biochemistry and Behavior 128 (2015) 50–61

(Kohl et al., 2013). Although no clear links have been implicated to thePPI elicited by depression (Quednow et al., 2006; Perry et al., 2004),animal models of stress exposure, which are often used to modeldepressive-like symptoms, have been demonstrated to differentiallydisturb PPI (Kohl et al., 2013; Pijlman et al., 2003; Sutherland et al.,2010). Dysfunctional PPI is reported in obsessive-compulsive disorder,post-traumatic stress disorder (Kohl et al., 2013) and schizophrenia(Braff et al., 2001; Koch, 2013; Weike et al., 2000). Interestingly, theonset or exacerbation of symptoms in these disorders can be triggeredby exposure to stressful events (Walker andDiforio, 1997), demonstrat-ing that the neural mechanisms involved in the stress response maymodulate systems involved in PPI regulation. The primary circuitrymediating the PPI response mainly comprises the brainstem, buthigher-order cognitive process can also modulate this response (Liet al., 2009). Studies have demonstrated that both the hippocampusand frontal cortex are involved in regulating the PPI response (Bastand Feldon, 2003; Swerdlow et al., 2001a), suggesting that neurochem-ical changes in the brain regions implicated in cognition may also affectPPI. Interestingly, there are some recent data implicating BDNF activityin mechanisms of sensorimotor behaviour (Naumenko et al., 2013;Papaleo et al., 2011).

Herein we demonstrated a reduction in the mRNA expression ofBDNF in the frontal cortex 24 h after the last swimming exposure, butnot in the hippocampus. Several studies have reported a range ofconflicting results regarding the mRNA expression of BDNF. Sequeira-Cordero et al. (2014) showed reduced BDNF levels in the frontal cortexand nucleus accumbens 1 and 6 h after the last swimming exposure butnot 24 h after. Shishkina et al. (2010) reported no changes in BDNFmRNA expression in the frontal cortex but a significant increase in thehippocampus 24 h after swimming. Altogether, these data indicatethat the effects of swimming stress on BDNF expression exhibitregion- and time-dependent patterns. However, BDNF synthesis wasaugmented in the hippocampus throughout the period evaluated afterthe last swimming session without any change in mRNA expression.Considering that BDNF protein levels do not necessarily change inparallel with a change in BDNF mRNA expression, we also evaluatedthe protein levels of this neurotrophin in the frontal cortex and foundthem to be unchanged, whereas hippocampal levels were increased.Therefore, assuming that glutamate can indirectly stimulate BDNFrelease (Nakajima et al., 2008), it is possible to conceive that the puta-tive increase on hippocampal glutamate levels (due to the reduceduptake) may have triggered release of stored BDNF.

Furthermore, the reduced frontal cortical mRNA expression of BDNFwas negatively correlated with immobility, indicating an associationbetween BDNF expression and depressive-like behaviour. This relation-ship was also observed with BDNF protein levels in the frontal cortex:rats with the highest immobility displayed reduced levels of BDNFprotein 24 h after swimming. Importantly, we showed that the numberof crossings and rearings were not altered after swimming exposure,indicating that the observed increase in immobility did not reflect areduction in locomotor activity. These results reinforce the findings ofother studies indicating that individual differences in the systematicvariations of specific behaviours may be involved in the neurochemicalchanges observed after swimming exposure (Estanislau et al., 2011;Naudon and Jay, 2005Sequeira-Cordero et al., 2013, 2014; Shishkinaet al., 2010).

Increased BDNF levels after acute stress is proposed to be partlyresponsible for the effect of this neurotrophin on adaptive responsesto stress by facilitating synaptic plasticity during learning and memory,especially in the hippocampus (Ozawa et al., 2014). This is supported bythe lack of a significant impairment in NOR performance found in thecurrent work. However, we observed an impairment of cognitive func-tion via a reduced PPI response despite a persistent increase in thehippocampal BDNF protein levels. Although it is unclear how BDNFsignalling participates in the disrupted behaviours observed inneuropsychiatric disorders (Carlino et al., 2013), BDNF activity may be

indirectly involved in the modulation of PPI. It is known that gluta-matergic activity modulates cortical and subcortical circuits implicatedin PPI (Swerdlow et al., 2001b; Gururajan et al., 2010). It is possiblethat BDNF affects this process by enhancing quantal glutamate release(Tyler and Pozzo-Miller, 2001) and regulating AMPAR and NMDAR(Caldeira et al., 2007a; Caldeira et al., 2007b). Although our results donot directly identify a role for changes in BDNF in PPI impairment, it ispossible that the altered expression of this neurotrophin coulddysregulate the glutamatergic system in a nonspecific way, which inturn could affect prefrontal cortical activity.

Another point to be considered is the effect of forced swimming onGABAergic activity, as immobility leads to an increase in extracellulargamma-amino butyric acid (GABA) levels in the ventral pallidum (VP)(Skirzewski et al., 2011). The VP is the principal region connecting theforebrain to the brainstem circuits involved in the startle response,and tonic inhibition of the VP is essential for regulating normal PPIresponses (Forcelli et al., 2012). Moreover, forced swimming increasesglutamate levels in the nucleus accumbens (Sequeira-Cordero et al.,2014; Rada et al., 2003), which is the brain structure that provides aprincipal source of GABAergic input to the VP (Johnson et al., 1994).Therefore, evaluating GABAergic activity or the glutamate/GABAbalance in further studies would help explain the underlying mecha-nisms related to the impairment of PPI and altered glutamatergicactivity.

We found reduced glutamate uptake in the hippocampal slices ofrats killed 48 h after the last swimming session. This result suggeststhat the clearance of extracellular glutamate from the synaptic cleftwas impaired after swimming stress. Upon release, glutamate activatesa variety of receptors before its removal by astrocytes via the high-affinity glutamate transporters GLAST/EAAT1 and/or GLT1/EAAT2(Danbolt, 2001). When the ability to remove extracellular glutamate isimpaired, an altered ratio of synaptic/extrasynaptic glutamate canresult, whichmay alter NMDAR activation and allow an excessive calci-um influx, which in turn may interact with reactive oxygen species andtrigger apoptosis (Lozovaya et al., 1999). Increased BDNF protein levelsmay confer protection against putative apoptosis triggered by stress(Melo et al., 2013; Wang and Green, 2011). Indeed, stress appears toincrease glutamatergic activity in the hippocampus (Fontella et al.,2004), which should promote calcium signalling and BDNF expression.

Reduced NMDAR expression has been suggested to involve anadaptive response to excessive receptor activation, and post-mortemstudies have reported altered expression of NMDAR in depressed andschizophrenic patients (Beneyto et al., 2007; Feyissa et al., 2009;Nudmamud-Thanoi and Reynolds, 2004). TheGluN1 subunit is essentialfor functional NMDAR, and changes in its expression may alter thegeneral activity of NMDAR. Moreover, GluN1 is among the differentclasses of proteins that were shown to be synthesized at the synapsefollowing stimulation with BDNF (Schratt et al., 2004). Nevertheless,in the current study, exposure to repeated swimming was not able toalter GluN1 levels, neither in the frontal cortex nor in the hippocampus.Calmodulin (CaM) has been shown to regulate NMDAR activity (Rycroftand Gibb, 2004). CaM is a ubiquitous calcium sensor protein that bindsCa+2 with high affinity in response to several extracellular signals thatalter cellular Ca+2 levels (Tidow and Nissen, 2013); this protein alsointeracts with Ca2+/calmodulin-dependent protein kinase II (CaMKII),which is known to activate BDNF (Alonso et al., 2002). However, similarto results for the GluN1 subunit, repeated swimming did not provokechanges in CaM levels. These findings indicate that the mechanismsunderlying BNDF elevation after repeated forced swimming are likelynot due to changes in GluN1 or CaM levels.

Changes in glutamate uptake are mainly a function of astrocytes.Reduced glutamate uptake reflects impaired astrocytic activity. GSH issynthetised and secreted by astrocytes for modulating antioxidantdefence (Dringen, 2000). GSH dysfunction seems to be involved inpsychiatric diseases associated with stressful conditions, as post-mortem studies revealed reduced GSH levels in patients with major

57M. Borsoi et al. / Pharmacology, Biochemistry and Behavior 128 (2015) 50–61

depression, bipolar disorder and schizophrenia (Gawryluk et al.,2011). Repeated stress, including immobilisation and cold exposure,decreases GSH content in the rodent brain (Madrigal et al., 2001;Sahin and Gümüşlü, 2004). Centrally administered GSH producesantidepressant-like effects in mice in both the forced swimmingand tail suspension tests (Rosa et al., 2013). Given these results, weexpected to observe a reduction in GSH content after the swimmingstress in the present study. However, GSH content was increased inthe hippocampus 48 h after swimming. This finding implies analtered cellular redox state in the hippocampus. Glutamate trans-porters participate in a redox regulatory mechanism (Trotti et al.,1997), and although we did not evaluate the levels of glutamatetransporters, we cannot exclude the possibility that an alteredredox state regulated the glutamate transporters in the currentstudy. As a result, we may suppose that forced swimming potentiallyaltered the redox state of the glutamate transporters, impairing theirfunctionality and ultimately reducing uptake. Considering thepresent data, we cannot identify the precise mechanisms involvedin the observed GSH elevations. To elucidate the effect of our stressprotocol on GSH levels, it will be interesting to evaluate putativechanges in the activity of the glutamate-cysteine ligase enzyme(the main enzyme responsible for GSH synthesis) and/or investigatechanges in the mechanism of glutathione export.

S100B, a calcium-binding protein produced predominantly in thebrain by astrocytes, exerts neurotrophic effects on neurons and glialcells (Donato, 2003). S100B is involved in several neuropsychiatricdisorders, and altered levels of this protein have been observed inpatients with major depression (Rajkowska and Stockmeier, 2013). Inaddition, antidepressant drugs are able to alter hippocampal S100Blevels (Rong et al., 2010) and secretion (Tramontina et al., 2008). Micesubmitted to chronic unpredictable mild stress, an animal model ofdepression, exhibit increased hippocampal S100B, and BDNF adminis-tration is able to partially restore S100B levels in this model (Ye et al.,2011). Nonetheless, despite enhanced hippocampal BDNF, repeatedforced swimming did not alter S100B levels in the hippocampus andfrontal cortex. Our result partially agree with those of Luo et al. (2010)who demonstrated decreased S100B in the frontal cortex but no changein the hippocampus of rats submitted to two distinct animal models ofdepression: chronic unpredictable stress and olfactory bulbectomy.However, increased hippocampal S100B has been reported withlectin-induced depressive-like behaviour (Gonçalves et al., 2013). Theconflicting findings in these studies could be due to differences in stressparadigms, species, age of animals, and the time course of depressivedisorder. Altogether, these findings support the notion that S100Bmay be responsible for regionally specific regulation in an animalmodel related to depression.

In summary, this study demonstrated for the first time that repeatedforced swimming stress selectively impairs pre-attentive processes inrats. Repeated forced swimming did not impair behavioural tests ofrecognition and spatial memory, but it did negatively affect the PPIresponse, a cognitive deficit widely described in schizophrenia andanxiety related disorders, as well as in animal models based on stressexposure used to model depressive-like symptoms. Along with thisbehavioural impairment, the applied stress regimen caused a significantreduction in frontal cortical mRNA BDNF expression and increase inBDNF protein levels in the hippocampus. Changes in the activity ofhippocampal astrocytes mediated by altered glutamate uptake andglutathione content were observed, reinforcing the hypothesis thatstress-related disorders encompass abnormal glutamatergic signallinginvolved in cognitive processes. Our data are in line with the supposi-tion that astrocytes may regulate the activity of the prefrontal cortexand related structures in a manner that alters complex behaviours.Moreover, it provides new insight into the dynamics immediatelyafter an aversive experience, such as those following the induction ofbehavioural despair. In conclusion, this work highlights the possibilityof using repeated forced swimming to study cognitive impairment

involving the prefrontal cortex and hippocampal glutamatergic activityin rats.

4. Experimental Procedure

4.1. Subjects

Adult male Wistar rats (250-300 g) purchased from the FundaçãoEstadual de Produção e Pesquisa em Saúde – RS (FEPPS) colony wereused. The animals were housed in groups of 5 animals in standardplastic cages and allowed at least oneweek to acclimatise to the housingconditions before beginning the experiments. All animals were keptunder a 12-h light/dark cycle (lights on 7:00 a.m.) at constant tempera-ture (23° ± 1 °C) with free access to standard rodent diet (Nuvilab®)and tapwater. Animals were adapted to the room laboratory conditions1 hour before the tests, which were performed in a controlled-temperature (22 ± 1 °C) environment. Behavioural experiments werecarried out between 11:00 a.m. and 3:00 p.m. All procedures werepreviously approved by The Animal Care Local Ethical Committee(CEUA-UFRGS; projects approval number 20406) and performedaccording to the EuropeanCommunities Council Directive of 24Novem-ber 1986 (86/609/EEC).

4.2. Repeated forced swimming

Repeated forced swimming was carried out in an acrylic box withfour sections that measured 30 cm × 30 cm × 40 cm. The externalwalls and covering were transparent, but the inside sections weredark to ensure isolation of each of the four quadrants. Each swimmingsession (10 min duration) was performed in water at 22 ± 1 °C and ata height of 30 cm. Ratswere exposed to three consecutive days of swim-ming; each session was separated by a 24 h interval. The immobilityduration was measured in each session by an expert observer. Ratswere considered immobile when they ceased to struggle and floatedmotionless in the water, making only the movements necessary tokeep their head above water. At the end of each swimming session,the rats were removed from the water and gently dried.

4.3. Behavioural tests

The behavioural tests were performed during the light cycle andinitiated 24 h after the last swimming session using three distinctgroups of animals. The novel object recognition (NOR) task and objectlocation test (OLT)were carried out in the same apparatus. Spontaneouslocomotion was evaluated during the habituation phase of both theNOR task and OLT. Prepulse inhibition (PPI) of the acoustic startleresponse (ASR) was evaluated twice: 24 hours before the first forcedswimming session and 24 hours after the last session. Naïve rats thathad not been exposed to repeated forced swimming (non-stressed)were evaluated as a control.

4.3.1. Novel object recognitionThe novel object recognition (NOR) task is based on the principle

that laboratory rodents are attracted to novelty: rats (or mice) spendmore time exploring a novel object than an object that is familiar(Ennaceur and Delacour, 1988). In the present study, we tested ratson NOR, which comprised three phases: habituation, training (acquisi-tion trial), and test (retention trial); each session lasted 5 minutes andwas performed in an open field apparatus (80 cm diameter; 12quadrants). During the habituation phase, the ratwas placed in the cen-tre of the apparatus and allowed to freely explore it in without anyobjects present. The number of crossings and rearings were countedto evaluate the spontaneous locomotion.

Twenty-four hours after the habituation phase, the animals weresubmitted to the acquisition trial; the rat was returned to the apparatus,which now contained two identical sample objects (A + A) placed at a

58 M. Borsoi et al. / Pharmacology, Biochemistry and Behavior 128 (2015) 50–61

distance of 10 cm from the wall. Ninety minutes or 24 h later, the ratswere returned to the apparatus for retention trials, in which short-term (STM) and long-term (LTM) memories were evaluated, respec-tively. Different animals were used to evaluate STM and LTM. Duringthe retention trials, the rat was returned to the apparatus, which nowcontained two objects; one of them was identical to the object used inthe acquisition trial, and the other was novel (A + B). The relativeexploration index in each retention session was calculated as 100 x(time exploring the familiar or novel object / time exploring bothobjects). The investigation ratio (IR) was also calculated: time exploringthe novel object / time exploring both objects. Exploration was scoredwhen the animal was observed sniffing or touching the object withthe nose and/or forepaws. Sitting on objects was not considered to indi-cate exploratory behaviour. The apparatus and the objects were cleanedthoroughly with 50% ethanol between trials to ensure the absence ofolfactory cues.

4.3.2. Object location testThe object location test (OLT) takes advantage of the innate tenden-

cy of rats to spontaneously explore objects in a novel place, and this testis used to assess spatial memory (Ennaceur et al., 1997). Here, rats weretested on OLT in three phases: habituation, training (acquisition trial)and test (retention trial); each session lasted 5 minutes and wasperformed in the same apparatus used during the novel object recogni-tion task. The habituation and acquisition trials were performed asdescribed for NOR (see Section 4.3.1). During the acquisition trial, thenumber of crossings and rearings were counted to evaluate the sponta-neous locomotion. Retention trials were performed 90min or 24 h afterthe training phase to access STM and LTM, respectively. STM and LTMwere evaluated in different animals. The same two objects used duringthe acquisition trial were presented in both retention trials; however,one of themwas positioned in a novel location. The relative explorationindex during each retention session was calculated as 100 x (timeexploring the object in familiar or novel position / time exploring bothobjects). The investigation ratio (IR) was also calculated: time exploringthe object in a novel position / time exploring both objects. Explorationwas scored when the animal was observed sniffing or touching theobjectwith the nose and/or forepaws. Sitting on objects was not consid-ered exploratory behaviour. The apparatus and theobjectswere cleanedthoroughly with 50% ethanol between trials to ensure the absence ofolfactory cues.

4.3.3. Prepulse inhibitionPrepulse inhibition (PPI) refers to a reduction in the magnitude of

the acoustic startle response (ASR) when a distinctive non-startlingstimulus is presented before a startling stimulus. The startle response(or startle reflex) refers to a fast defensive response to a suddenstimulus. PPI is a consistent phenomenon across species and is widelyused tomeasure sensorimotor gatingmechanisms essential for the inte-gration of cognitive and sensory information (Li et al., 2009; Valsamisand Schmid, 2011). Both the PPI and ARS magnitudes were evaluatedusing a startle chamber (Insight®, São Paulo, Brazil), in which a loud-speaker produced a continuous background noise (60 dB) along withan acoustic startle pulse and prepulse. A white noise pulse was usedas the startle stimulus, which had an intensity of 105 dB and durationof 20 ms. The prepulse intensities used were 68, 73 and 77 dB. Thetest trial started with an apparatus acclimation period of 5 min. Toexamine the magnitude of the acoustic startle response (ASR), the firstand last five trials consisting of presenting the 20 ms pulse alone with105 dB stimuli. The middle 60 trials consisted of blocks of randomisedtrials with pulse-alone and prepulse-pulse combinations to measurethe prepulse inhibition. Prepulse-pulse trials involved a single 105 dBpulse preceded 100 ms earlier by a prepulse of either 68, 73 or 77 dB(20 ms of duration). The percent PPI of the startle response wascalculated as follows: 100 – [(response to acoustic prepulse plus startlestimulus trials/startle response alone trials) x 100]. The PPI test was

performed twice: before and after the repeated forced swimming. Therepeated forced swimming was initiated 24 h after the first PPIevaluation, and the second exposure to PPI was carried out 24 h afterthe last swimming session.

4.4. Neurochemical analysis

Neurochemical analyses were performed in separate groups ofanimals submitted to the repeated forced swimming and decapitated24, 48 or 72 h after the last swimming session in a separate series ofexperiments. Except for glutamate uptake, which was performed inhippocampal slices only, the other neurochemical assays were carriedout using homogenates of frontal cortex and hippocampus from bothhemispheres. Rats not exposed to repeated forced swimming (non-stressed) were used as a control and were sacrificed in parallel withthe animals exposed to repeated forced swimming.

4.4.1. Real-time PCR quantification of BDNF mRNAStructures were homogenised using TRIzol reagent (Invitrogen) for

total RNA extraction according to the manufacturer's instructions. Theconcentration and purity of the RNA were assessed at 260 nm and260/280 nm, respectively. Additionally, the integrity was evaluatedusing agarose gel electrophoresis stained with ethidium bromide stain-ing. To obtain the reverse transcript (cDNA), 2 μg of total RNA werereverse transcribed using theHigh Capacity cDNA Transcription kit (Ap-plied Biosystems Inc., Foster City, CA) in a 10 μL reaction. After the cDNAwas obtained, samples were stored at -20 °C.

BDNF mRNA expression was carried out using fluorescence-based real-time PCR after amplifying 100 ng of cDNA in duplicateusing TaqMan-based reactions with specific primers, FAM-labelledprobes for rat BDNF (Assays-by-Demand, Life Technologies, cat. #Rn02531967_s1) and VIC-labelled glyceraldehyde-3-phosphate dehy-drogenase (GAPDH; Assays-by-Demand, Life Technologies, cat. #Rn01775763_g1) as the endogenous control for normalisation. Amplifi-cations were carried out in a thermocycler (StepOne Plus, AppliedBiosystems) for 70 cycles. The fluorescence was evaluated at eachamplification cycle, and the datawere analysed using the 2-ΔΔCtmethodfor relative quantification. Expression of the target genes was calibratedagainst conditions found in age-matched naïve rats.

4.4.2. BDNF immunocontentBDNF immunocontent was measured using ELISA (Cechetti et al.,

2008). Briefly, brain regions were individually homogenised in lysisbuffer containing (in mM) 137 NaCl, 20 Tris–HCl (pH 8.0), Igepal (1%),glycerol (10%), 1 PMSF, 0.5 sodium vanadate, 0.1 EDTA and 0.1 EGTA.Homogenates were incubated in 96-well flat-bottom plates previouslycoated with anti-BDNF monoclonal antibody (E-Max ELISA kit,Promega, USA). After blocking, plates were incubated with polyclonalanti-human antibody for 2 h and horseradish peroxidase for 1 h. Then,colour reaction with tetramethyl benzidine was quantified in a platereader at 450 nm; the standard BDNF curve ranged from 0 to500 pg/mL. Final concentrations were calculated and expressed as pg/mg protein.

4.4.3. GluN1 and calmodulin protein determinationGluN1 and calmodulin were assessed via western blotting. Equal

amounts of protein (20 μg) from each sample were boiled in samplebuffer (0.0625 M Tris–HCl, pH 6.8, 2% (w/v) SDS, 5% (w/v) β-mercaptoethanol, 10% (v/v) glycerol, 0.002% (w/v) bromophenol blue)and electrophoresed in 10% (w/v) SDS-polyacrylamide gel. The separat-ed proteins were transferred onto a nitrocellulose membrane. Equalloading of each sample was confirmed with Ponceau S staining(Sigma). Anti-NMDAR1 (Millipore) or anti-calmodulin (Sigma) anti-bodies were used at a dilution of 1:1000 to determine GluN1 andcalmodulin levels, respectively. After incubating with the primary anti-body for 1 h at room temperature, filters were washed and incubated

59M. Borsoi et al. / Pharmacology, Biochemistry and Behavior 128 (2015) 50–61

with peroxidase-conjugated anti-rabbit immunoglobulin (IgG) orperoxidase-conjugated anti-goat immunoglobulin (IgG) at a dilutionof 1:2000. The chemiluminescence signal was detected using an ECLkit (Amersham).

4.4.4. Glutamate uptake assayGlutamate uptake was evaluated in hippocampal acute slices.

Hippocampi were sliced in transverse sections of 0.3 mm using aMcIlwain tissue chopper. Slices were transferred to 24-well plates con-taining 0.3 mL of HEPES-buffered saline solution (120 mM NaCl, 2 mMKCl, 1 mM CaCl2, 1 mM MgSO4, 1 mM KH2PO4, 10 mM glucose and25 mM HEPES, pH 7.4) per well. The medium was changed every15 min for 2 h at room temperature. After this stabilisation period, theslices were incubated for 1 h at 30 °C.

After the incubation period, the medium was replaced by Hank’sbalanced salt solution (HBSS) containing (in mM) 137 NaCl, 0.63Na2HPO4, 4.17 NaOHCO3, 5.36 KCl, 0.44 KH2PO4, 1.26 CaCl2, 0.41MgSO4, 0.49MgCl2 and 5.55 glucose, pH 7.4. The slices weremaintainedat 35 °C, and the assay was started by adding 0.1 mM l-glutamate and0.66 μCi/mL l-[2,3-3H]-glutamate. Incubation was stopped after 5 minby removing the medium and rinsing the slices three times with ice-cold HBSS. Slices were then lysed in a solution containing 0.5 N NaOH.Sodium-independent (nonspecific) uptake was determined usinga solution with N-methyl-d-glutamine instead of NaCl. Sodium-dependent glutamate uptake was obtained by subtracting the nonspe-cific uptake. Radioactivity was measured with a scintillation counter(2800TR TriCarb Liquid Scintillation Analyzer, Perkin Elmer, Waltham,MA, USA). Final glutamate uptake was expressed as nmol/mg protein/min.

4.4.5. Glutathione contentTotal glutathione (GSH) content was determined as described

previously (Browne and Armstrong, 1998). Briefly, tissues werehomogenised in sodium phosphate buffer (0.1 M, pH 8.0) containing5 mM EDTA, and the protein was precipitated with 1.7% meta-phosphoric acid. The supernatant was assayed with o-phthaldialdehyde(1 mg/mL of methanol) at room temperature for 15 min. Fluorescencewas measured using excitation and emission wavelengths of 350 and420 nm, respectively. A calibration curve was performed with standardglutathione solutions (0-500 μM). Final concentrations were calculatedand expressed as nmol/mg protein.

4.4.6. S100B immunocontentThe level of S100B was measured using ELISA (Leite et al., 2008).

Briefly, 50 μL of samples plus 50 μL of 1.5 mM Tris-HCl buffer, pH 8.8,were incubated for 2 h on a microtiter plate previously coated withmonoclonal anti-S100B (SH-B1, Sigma). Polyclonal anti-S100 (DAKO)was incubated for 30 min and then peroxidase-conjugated anti-rabbitantibody was added for an additional 30 min. The colour reaction witho-phenylenediamine was measured at 492 nm. A standard curve ofS100B ranged from 0.025 to 2.5 ng/mL. Final concentrations werecalculated and expressed as ng/mg protein.

4.4.7. Protein determinationTotal protein concentration was measured using Lowry's method

with bovine serum albumin as a standard (Lowry et al., 1951).

4.5. Statistical analysis

Normally distributed data were presented as the mean ± standarderror of the mean (SEM), whereas non-normally distributed data wereexpressed as median and 25th and 75th percentile. The value ncorresponds to the number of animals used in each experimentalgroup. One-way repeated measures analysis of variance (ANOVA) wasused to evaluate differences in the immobility duration betweendays of swimming exposure. Two-way repeated measures analysis of

variance (ANOVA) was used to make multiple comparisons betweengroups in the memory tests (novel object recognition and object loca-tion test) and prepulse inhibition test. The number of crossings, numberof rearings, and preference index of both memory tests, as well as themagnitude of the acoustic startle response, were examined usingStudent’s t test. Kruskal-Wallis ANOVA on ranks was used to comparethe mRNA expression of BDNF. Western blot results (GluN1 andcalmodulin protein levels), glutamate uptake, glutathione, S100B andBDNF protein determinations were examined using one-way analysisof variance (ANOVA). Pearson’s correlation test was used to verifycorrelations between the immobility duration observed in the lastswimming session and neurochemical parameters or behavioural testresults. When appropriate, the post hoc test used was Student-Newman-Keuls analysis or Dunn’s Multiple Comparison test. Statisticalanalysis was performed with SigmaStat 3.5 Software (Jandel ScientificCorporation). Statistical significance was set at p b 0.05.

Author contributions

MB, CAG and SMKR conceived and designed the experiments. MB,CBA, LGM, AFV, VH, PSL, CZ and PN performed the experiments. MBdrafted the manuscript. MB, APR, CAG and SMKR made significantcontributions to revising themanuscript. All authors read and approvedthe final manuscript.

Acknowledgments

The authors gratefully acknowledge to Graduate Studies Program inNeurosciences, Graduate Studies Program in Pharmaceutical Sciences,Graduate Studies Program in Biochemistry, and College of VeterinaryMedicine from Federal University of Rio Grande do Sul. This work wassupported by a grant from Improvement of Higher Education Personnel(CAPES).

References

Abel EL, Hannigan JH. Effects of chronic forced swimming and exposure to alarmsubstance: physiological and behavioral consequences. Physiol Behav 1992;52:781–5.

Alonso M, Vianna MRM, Depino AM, Mello e Souza T, Pereira P, Szapiro G, et al. BDNF-triggered events in the rat hippocampus are required for both short- and long-termmemory formation. Hippocampus 2002;12:551–60. http://dx.doi.org/10.1002/hipo.10035.

Andolina D, Maran D, Valzania A, Conversi D, Puglisi-Allegra S. Prefrontal/amygdalarsystem determines stress coping behavior through 5-HT/GABA connection.Neuropsychopharmacology 2013;38(10):2057–67. http://dx.doi.org/10.1038/npp.2013.107.

Anisman H, Matheson K. Stress, depression, and anhedonia: caveats concerning animalmodels. Neurosci Biobehav Rev 2005;29:525–46. http://dx.doi.org/10.1016/j.neubiorev.2005.03.007.

Banasr M, Duman RS. Glial loss in the prefrontal cortex is sufficient to induce depressive-like behaviors. Biol Psychiatry 2008;64:863–70. http://dx.doi.org/10.1016/j.biopsych.2008.06.008.

Bast T, Feldon J. Hippocampal modulation of sensorimotor processes. Prog Neurobiol2003;70:319–45.

Bath KG, Schilit A, Lee FS. Stress effects on BDNF expression: effects of age, sex, and formof stress. Neuroscience 2013;239:149–56. http://dx.doi.org/10.1016/j.neuroscience.2013.01.074.

Beneyto M, Kristiansen LV, Oni-Orisan A, McCullumsmith RE, Meador-Woodruff JH.Abnormal Glutamate Receptor Expression in the Medial Temporal Lobe in Schizo-phrenia and Mood Disorders. Neuropsychopharmacology 2007;32:1888–902.http://dx.doi.org/10.1038/sj.npp.1301312.

Braff DL, GeyerMA, SwerdlowNR. Human studies of prepulse inhibition of startle: normalsubjects, patient groups, and pharmacological studies. Psychopharmacology (Berl)2001;156:234–58.

Briones-Aranda A, Rocha L, Picazo O. Alterations in GABAergic function following forcedswimming stress. Pharmacol Biochem Behav 2005;80:463–70. http://dx.doi.org/10.1016/j.pbb.2005.01.002.

Browne RW, Armstrong D. Reduced glutathione and glutathione disulfide. Methods MolBiol Clifton NJ 1998;108:347–52. http://dx.doi.org/10.1385/0-89603-472-0:347.

Browne CA, Lucki I. Antidepressant effects of ketamine: mechanisms underlying fast-acting novel antidepressants. Neuropharmacology 2013;4:161. http://dx.doi.org/10.3389/fphar.2013.00161.

Brunoni AR, Lopes M, Fregni F. A systematic review and meta-analysis of clinical studieson major depression and BDNF levels: implications for the role of neuroplasticity in

60 M. Borsoi et al. / Pharmacology, Biochemistry and Behavior 128 (2015) 50–61

depression. Int J Neuropsychopharmacol Off Sci J Coll Int NeuropsychopharmacolCINP 2008;11:1169–80. http://dx.doi.org/10.1017/S1461145708009309.

Caldeira MV, Melo CV, Pereira DB, Carvalho R, Correia SS, Backos DS, et al. Brain-derivedneurotrophic factor regulates the expression and synaptic delivery of alpha-amino-3-hydroxy-5-methyl-4-isoxazole propionic acid receptor subunits in hippocampalneurons. J Biol Chem 2007a;282:12619–28. http://dx.doi.org/10.1074/jbc.M700607200.

Caldeira MV, Melo CV, Pereira DB, Carvalho RF, Carvalho AL, Duarte CB. BDNF regu-lates the expression and traffic of NMDA receptors in cultured hippocampalneurons. Mol Cell Neurosci 2007b;35:208–19. http://dx.doi.org/10.1016/j.mcn.2007.02.019.

Carey AN, Lyons AM, Shay CF, Dunton O, McLaughlin JP. Endogenous kappa opioidactivation mediates stress-induced deficits in learning and memory. J Neurosci Off JSoc Neurosci 2009;29:4293–300. http://dx.doi.org/10.1523/JNEUROSCI. 6146-08.2009.

Carlino D, De Vanna M, Tongiorgi E. Is altered BDNF biosynthesis a general feature inpatients with cognitive dysfunctions? Neurosci Rev J Bringing Neurobiol NeurolPsychiatry 2013;19:345–53. http://dx.doi.org/10.1177/1073858412469444.

Castrén E, Rantamäki T. Role of brain-derived neurotrophic factor in the aetiology ofdepression: implications for pharmacological treatment. CNS Drugs 2010;24:1–7.http://dx.doi.org/10.2165/11530010-000000000-00000.

Cazakoff BN, Johnson KJ, Howland JG. Converging effects of acute stress on spatial andrecognition memory in rodents: a review of recent behavioural and pharmacologicalfindings. Prog Neuropsychopharmacol Biol Psychiatry 2010;34:733–41. http://dx.doi.org/10.1016/j.pnpbp.2010.04.002.

Cechetti F, Fochesatto C, Scopel D, Nardin P, Gonçalves CA, Netto CA, et al. Effect of aneuroprotective exercise protocol on oxidative state and BDNF levels in the rathippocampus. Brain Res 2008;10(1188):182–8.

Cryan JF, Valentino RJ, Lucki I. Assessing substrates underlying the behavioral effects ofantidepressants using the modified rat forced swimming test. Neurosci BiobehavRev 2005;29:547–69. http://dx.doi.org/10.1016/j.neubiorev.2005.03.008.

Czéh B, Di Benedetto B. Antidepressants act directly on astrocytes: evidences and func-tional consequences. Eur Neuropsychopharmacol J Eur Coll Neuropsychopharmacol2013;23:171–85. http://dx.doi.org/10.1016/j.euroneuro.2012.04.017.

Danbolt NC. Glutamate uptake. Prog Neurobiol 2001;65:1–105.Donato R. Intracellular and extracellular roles of S100 proteins. Microsc Res Tech 2003;60:

540–51. http://dx.doi.org/10.1002/jemt.10296.Dringen R. Metabolism and functions of glutathione in brain. Prog Neurobiol 2000;62:

649–71.Duman RS. Neurobiology of stress, depression, and rapid acting antidepressants: remod-

eling synaptic connections. Depress Anxiety 2014;31:291–6. http://dx.doi.org/10.1002/da.22227.

Duman RS, Monteggia LM. A neurotrophic model for stress-related mood disorders. BiolPsychiatry 2006;59:1116–27. http://dx.doi.org/10.1016/j.biopsych.2006.02.013.

Ennaceur A, Delacour J. A new one-trial test for neurobiological studies of memory in rats.1: Behavioral data. Behav. Brain Res 1988;31:47–59.

Ennaceur A, Neave N, Aggleton JP. Spontaneous object recognition and object locationmemory in rats: the effects of lesions in the cingulate cortices, the medial prefrontalcortex, the cingulum bundle and the fornix. Exp Brain Res 1997;113:509–19.

Estanislau C, Ramos AC, Ferraresi PD, Costa NF, de Carvalho HMCP, Batistela S. Individualdifferences in the elevated plus-maze and the forced swim test. Behav Process 2011;86:46–51. http://dx.doi.org/10.1016/j.beproc.2010.08.008.

Feyissa AM, Chandran A, Stockmeier CA, Karolewicz B. Reduced levels of NR2A and NR2Bsubunits of NMDA receptor and PSD-95 in the prefrontal cortex in major depression.Prog Neuropsychopharmacol Biol Psychiatry 2009;33(1):70–5. http://dx.doi.org/10.1016/j.pnpbp.2008.10.005. (Epub 2008 Oct 25).

Fontella FU, Vendite DA, Tabajara AS, Porciúncula LO, da Silva Torres IL, Jardim FM, et al.Repeated restraint stress alters hippocampal glutamate uptake and release in the rat.Neurochem Res 2004;29(9):1703–9.

Forcelli PA,West EA, Murnen AT, Malkova L. Ventral pallidummediates amygdala-evokeddeficits in prepulse inhibition. Behav Neurosci 2012;126:290–300.

Gawryluk JW, Wang JF, Andreazza AC, Shao L, Young LT. Decreased levels of glutathione,the major brain antioxidant, in post-mortem prefrontal cortex from patients withpsychiatric disorders. Int J Neuropsychopharmacol 2011;14(1):123–30. http://dx.doi.org/10.1017/S1461145710000805.

Gonçalves FM, Freitas AE, Peres TV, Rieger DK, Ben J, Maestri M, et al. Vatairea macrocarpalectin (VML) induces depressive-like behavior and expression of neuroinflammatorymarkers in mice. Neurochem Res 2013;38:2375–84. http://dx.doi.org/10.1007/s11064-013-1150-9.

Gururajan A, Taylor DA, Malone DT. Effect of testing conditions on the propsychotic actionof MK-801 on prepulse inhibition, social behaviour and locomotor activity. PhysiolBehav 2010;99:131–8. http://dx.doi.org/10.1016/j.physbeh.2009.11.005.

Johnson LR, Aylward RL, Hussain Z, Totterdell S. Input from the amygdala to the rat nucle-us accumbens: its relationship with tyrosine hydroxylase immunoreactivity andidentified neurons. Neuroscience 1994;61:851–65.

Koch M. Clinical relevance of animal models of schizophrenia. Suppl Clin Neurophysiol2013;62:113–20.

Kohl S, Heekeren K, Klosterkötter J, Kuhn J. Prepulse inhibition in psychiatric disorders–apart from schizophrenia. J Psychiatr Res 2013;47:445–52. http://dx.doi.org/10.1016/j.jpsychires.2012.11.018.

Krishnan V, Nestler EJ. Animal models of depression: molecular perspectives. Curr TopBehav Neurosci 2011;7:121–47. http://dx.doi.org/10.1007/7854_2010_108.

Krystal JH, Sanacora G, Blumberg H, Anand A, Charney DS, Marek G, et al. Glutamateand GABA systems as targets for novel antidepressant and mood-stabilizingtreatments. Mol Psychiatry 2002;7(Suppl. 1):S71–80. http://dx.doi.org/10.1038/sj.mp.4001021.

Lee Y, Gaskins D, Anand A, Shekhar A. Glia mechanisms in mood regulation: a novelmodel of mood disorders. Psychopharmacology (Berl) 2007;191:55–65. http://dx.doi.org/10.1007/s00213-006-0652-4.

Lee Y, Son H, Kim G, Kim S, Lee DH, Roh GS, et al. Glutamine deficiency in the prefrontalcortex increases depressive-like behaviours in male mice. J Psychiatry Neurosci JPN2013;38:183–91. http://dx.doi.org/10.1503/jpn.120024.

Leite MC, Galland F, Brolese G, Guerra MC, Bortolotto JW, Freitas R, et al. A simple,sensitive and widely applicable ELISA for S100B: Methodological features of themea-surement of this glial protein. J Neurosci Methods 2008;169:93–9. http://dx.doi.org/10.1016/j.jneumeth.2007.11.021.

Li L, Du Y, Li N, Wu X, Wu Y. Top-down modulation of prepulse inhibition of the startlereflex in humans and rats. Neurosci Biobehav Rev 2009;33:1157–67. http://dx.doi.org/10.1016/j.neubiorev.2009.02.001.

Lowry OH, Rosebrough NJ, Farr AL, Randall RJ. Proteinmeasurement with the Folin phenolreagent. J Biol Chem 1951;193:265–75.

Lozovaya NA, Kopanitsa MV, Boychuk YA, Krishtal OA. Enhancement of glutamate releaseuncovers spillover-mediated transmission by N-methyl-D-aspartate receptors in therat hippocampus. Neuroscience 1999;91:1321–30.

Luo K-R, Hong C-J, Liou Y-J, Hou S-J, Huang Y-H, Tsai S-J. Differential regulation ofneurotrophin S100B and BDNF in two rat models of depression. ProgNeuropsychopharmacol Biol Psychiatry 2010;34:1433–9. http://dx.doi.org/10.1016/j.pnpbp.2010.07.033.

Madrigal JL, Olivenza R, Moro MA, Lizasoain I, Lorenzo P, Rodrigo J, et al. Glutathionedepletion, lipid peroxidation and mitochondrial dysfunction are induced by chronicstress in rat brain. Neuropsychopharmacol Off Publ Am Coll Neuropsychopharmacol2001;24:420–9. http://dx.doi.org/10.1016/S0893-133X(00)00208-6.

Melo CV, Mele M, Curcio M, Comprido D, Silva CG, Duarte CB. BDNF regulates theexpression and distribution of vesicular glutamate transporters in cultured hippo-campal neurons. PLoS One 2013;8:e53793. http://dx.doi.org/10.1371/journal.pone.0053793.

Nakajima T, Sato M, Akaza N, Umezawa Y. Cell-based fluorescent indicator to visualizebrain-derived neurotrophic factor secreted from living neurons. ACS Chem Biol2008;3(6):352–8. http://dx.doi.org/10.1021/cb800052v.

Naudon L, Jay TM. Opposite behaviours in the forced swimming test are linked todifferences in spatial working memory performances in the rat. Neuroscience 2005;130:285–93. http://dx.doi.org/10.1016/j.neuroscience.2004.09.009.

Naumenko VS, Bazovkina DV, MorozovaMV, Popova NK. Effects of brain-derived and glialcell line-derived neurotrophic factors on startle response and disrupted prepulseinhibition in mice of DBA/2 J inbred strain. Neurosci Lett 2013;550:115–8. http://dx.doi.org/10.1016/j.neulet.2013.06.056.

Nudmamud-Thanoi S, Reynolds GP. The NR1 subunit of the glutamate/NMDA receptor inthe superior temporal cortex in schizophrenia and affective disorders. Neurosci Lett2004;372:173–7. http://dx.doi.org/10.1016/j.neulet.2004.09.035.

Ozawa T, Yamada K, Ichitani Y. Hippocampal BDNF treatment facilitates consolidation ofspatial memory in spontaneous place recognition in rats. Behav Brain Res 2014;263:210–6. http://dx.doi.org/10.1016/j.bbr.2014.01.034.

Papaleo F, Silverman JL, Aney J, Tian Q, Barkan CL, Chadman KK, et al. Working memorydeficits, increased anxiety-like traits, and seizure susceptibility in BDNF overexpress-ing mice. LearnMem Cold Spring Harb N 2011;18:534–44. http://dx.doi.org/10.1101/lm.2213711.

Perry W, Minassian A, Feifel D. Prepulse inhibition in patients with non-psychotic majordepressive disorder. J Affect Disord 2004;81:179–84. http://dx.doi.org/10.1016/S0165-0327(03)00157-5.

Pijlman FTA, Herremans AHJ, van de Kieft J, Kruse CG, van Ree JM. Behavioural changesafter different stress paradigms: prepulse inhibition increased after physical, butnot emotional stress. Eur Neuropsychopharmacol J Eur Coll Neuropsychopharmacol2003;13:369–80.

Popoli M, Yan Z, McEwen BS, Sanacora G. The stressed synapse: the impact of stress andglucocorticoids on glutamate transmission. Nat Rev Neurosci 2012;13:22–37. http://dx.doi.org/10.1038/nrn3138.

Porsolt RD, Anton G, Blavet N, Jalfre M. Behavioural despair in rats: a newmodel sensitiveto antidepressant treatments. Eur J Pharmacol 1978a;47:379–91.

Porsolt RD, Bertin A, Jalfre M. “Behavioural despair” in rats and mice: strain differencesand the effects of imipramine. Eur J Pharmacol 1978b;51:291–4.

Porsolt RD, Moser PC, Castagné V. Behavioral indices in antipsychotic drug discovery.J Pharmacol Exp Ther 2010;333:632–8. http://dx.doi.org/10.1124/jpet.110.166710.

Quednow BB, Westheide J, Kühn K-U, Werner P, Maier W, Hawellek B, et al. Normalprepulse inhibition and habituation of acoustic startle response in suicidal depressivepatients without psychotic symptoms. J Affect Disord 2006;92:299–303. http://dx.doi.org/10.1016/j.jad.2006.01.022.

Rada P, Moreno SA, Tucci S, Gonzalez LE, Harrison T, Chau DT, et al. Glutamate release inthe nucleus accumbens is involved in behavioral depression during the PORSOLTswim test. Neuroscience 2003;119:557–65. (doi).

Rajkowska G, Stockmeier CA. Astrocyte pathology in major depressive disorder: insightsfrom human postmortem brain tissue. Curr Drug Targets 2013;14:1225–36.

Rong H, Wang G, Liu T, Wang H, Wan Q, Weng S. Chronic mild stress induces fluoxetine-reversible decreases in hippocampal and cerebrospinal fluid levels of the neurotroph-ic factor S100B and its specific receptor. Int J Mol Sci 2010;11:5310–22. http://dx.doi.org/10.3390/ijms11125310.

Rosa JM, Dafre AL, Rodrigues ALS. Antidepressant-like responses in the forced swimmingtest elicited by glutathione and redoxmodulation. Behav Brain Res 2013;253:165–72.http://dx.doi.org/10.1016/j.bbr.2013.07.009.

Rycroft BK, Gibb AJ. Inhibitory interactions of calcineurin (phosphatase 2B) and calmod-ulin on rat hippocampal NMDA receptors. Neuropharmacology 2004;47:505–14.http://dx.doi.org/10.1016/j.neuropharm.2004.06.001.

61M. Borsoi et al. / Pharmacology, Biochemistry and Behavior 128 (2015) 50–61

Sahin E, Gümüşlü S. Alterations in brain antioxidant status, protein oxidation and lipidperoxidation in response to different stress models. Behav Brain Res 2004;155:241–8. http://dx.doi.org/10.1016/j.bbr.2004.04.022.

Sanacora G, Banasr M. From pathophysiology to novel antidepressant drugs: glialcontributions to the pathology and treatment of mood disorders. Biol Psychiatry2013;73:1172–9. http://dx.doi.org/10.1016/j.biopsych.2013.03.032.

Schratt GM, Nigh EA, Chen WG, Hu L, Greenberg ME. BDNF regulates the translation of aselect group of mRNAs by a mammalian target of rapamycin-phosphatidylinositol 3-kinase-dependent pathway during neuronal development. J Neurosci 2004;24(33):7366–77. (Aug 18).

Sequeira-Cordero A, Mora-Gallegos A, Cuenca-Berger P, Fornaguera-Trías J. Individualdifferences in the immobility behavior in juvenile and adult rats are associatedwith monoaminergic neurotransmission and with the expression of corticotropin-releasing factor receptor 1 in the nucleus accumbens. Behav Brain Res 2013;252:77–87.

Sequeira-Cordero A, Mora-Gallegos A, Cuenca-Berger P, Fornaguera-Trías J. Individualdifferences in the forced swimming test and neurochemical kinetics in the ratbrain. Physiol Behav 2014;128C:60–9. http://dx.doi.org/10.1016/j.physbeh.2014.01.037.

Shishkina GT, Kalinina TS, Berezova IV, Bulygina VV, Dygalo NN. Resistance to the devel-opment of stress-induced behavioral despair in the forced swim test associated withelevated hippocampal Bcl-xl expression. Behav Brain Res 2010;213:218–24. http://dx.doi.org/10.1016/j.bbr.2010.05.003.

Skirzewski M, López W, Mosquera E, Betancourt L, Catlow B, Chiurillo M, et al. EnhancedGABAergic tone in the ventral pallidum: memory of unpleasant experiences?Neuroscience 2011;196:131–46. http://dx.doi.org/10.1016/j.neuroscience.2011.08.058.

Sutherland JE, Burian LC, Covault J, Conti LH. The effect of restraint stress on prepulseinhibition and on corticotropin-releasing factor (CRF) and CRF receptor gene expres-sion in Wistar-Kyoto and Brown Norway rats. Pharmacol Biochem Behav 2010;97:227–38. http://dx.doi.org/10.1016/j.pbb.2010.08.003.

Swerdlow NR, Geyer MA, Braff DL. Neural circuit regulation of prepulse inhibition ofstartle in the rat: current knowledge and future challenges. Psychopharmacology(Berlin) 2001a;156:194–215.

Swerdlow NR, Hanlon FM, Henning L, Kim YK, Gaudet I, Halim ND. Regulation ofsensorimotor gating in rats by hippocampal NMDA: anatomical localization. BrainRes 2001b;898:195–203.

Tidow H, Nissen P. Structural diversity of calmodulin binding to its target sites. FEBS J2013;280:5551–65. http://dx.doi.org/10.1111/febs.12296.

Tokita K, Yamaji T, Hashimoto K. Roles of glutamate signaling in preclinical and/ormechanistic models of depression. Pharmacol Biochem Behav 2012;100:688–704.http://dx.doi.org/10.1016/j.pbb.2011.04.016.

Tramontina AC, Tramontina F, Bobermin LD, Zanotto C, Souza DF, Leite MC, et al. Secretionof S100B, an astrocyte-derived neurotrophic protein, is stimulated by fluoxetine via amechanism independent of serotonin. Prog Neuropsychopharmacol Biol Psychiatry2008;32:1580–3. http://dx.doi.org/10.1016/j.pnpbp.2008.06.001.

Trotti D, Rizzini BL, Rossi D, Haugeto O, Racagni G, Danbolt NC, et al. Neuronal and glialglutamate transporters possess an SH-based redox regulatory mechanism. Eur JNeurosci 1997;9:1236–43.

Tyler WJ, Pozzo-Miller LD. BDNF enhances quantal neurotransmitter release andincreases the number of docked vesicles at the active zones of hippocampal excitato-ry synapses. J Neurosci Off J Soc Neurosci 2001;21:4249–58.

Valsamis B, Schmid S. Habituation and prepulse inhibition of acoustic startle in rodents. JVis Exp JoVE 2011:e3446. http://dx.doi.org/10.3791/3446.

Walker EF, Diforio D. Schizophrenia: a neural diathesis-stress model. Psychol Rev 1997;104:667–85.

Wang Q, Green SH. Functional role of neurotrophin-3 in synapse regeneration by spiralganglion neurons on inner hair cells after excitotoxic trauma in vitro. J Neurosci OffJ Soc Neurosci 2011;31:7938–49. http://dx.doi.org/10.1523/JNEUROSCI. 1434-10.2011.

Warner TA, Drugan RC. Morris water maze performance deficit produced by intermittentswim stress is partially mediated by norepinephrine. Pharmacol Biochem Behav2012;101:24–34. http://dx.doi.org/10.1016/j.pbb.2011.11.009.

Weike AI, Bauer U, Hamm AO. Effective neuroleptic medication removes prepulseinhibition deficits in schizophrenia patients. Biol Psychiatry 2000;47:61–70.

Ye Y, Wang G, Wang H, Wang X. Brain-derived neurotrophic factor (BDNF) infusionrestored astrocytic plasticity in the hippocampus of a rat model of depression.Neurosci Lett 2011;503:15–9. http://dx.doi.org/10.1016/j.neulet.2011.07.055.

Zucker M, Weizman A, Rehavi M. Repeated swim stress leads to down-regulation ofvesicular monoamine transporter 2 in rat brain nucleus accumbens and striatum.Eur Neuropsychopharmacol J Eur Coll Neuropsychopharmacol 2005;15:199–201.http://dx.doi.org/10.1016/j.euroneuro.2004.08.009.