Introduction

The first report of ketamine’s antidepressant effect after intravenous infusion comes from Berman and his colleagues (Berman et al. 2000). The single administration of ketamine has therapeutic effect and improves depressive symptoms lasting up to 7 days (Zarate et al. 2006, 2012; Diazgranados et al. 2010; Murrough et al. 2013a), possibly up to 4 weeks (Ibrahim et al. 2012). The significant improvement was also reported with repeated infusions in treatment-resistant depressive patients (Murrough et al. 2013b; Shiroma et al. 2014). These findings have triggered the immense interest in research of ketamine’s antidepressant effect and its mechanism of action.

Ketamine is a non-competitive NMDA receptor antagonist (Muir and Lees 1995). However, its antidepressant effect cannot be explained solely by inhibition of NMDA receptors. Ketamine administration sets off signaling cascades resulting in rapid antidepressant effect persisting long after the drug had been eliminated from the body (Hijazi et al. 2003). One of the proposed mechanisms of action responsible for ketamine’s antidepressant effect is activation of mammalian target of rapamycin (mTOR) signaling pathway (Abdallah et al. 2015).

mTOR is serine/threonine protein kinase regulating cell growth, proliferation, transcription, and proteosynthesis, thus directly participating in synaptic plasticity (Swiech et al. 2008). mTOR signaling pathway is activated as early as 30 min after ketamine administration via elevated level of Akt and ERK protein kinases. Two hours later, the level of phosphorylated proteins returns to normal. This effect is dose-dependent; lower doses of ketamine (5–10 mg/kg) have stimulating effect while high anesthetic doses (80 mg/kg) have no effect on mTOR signaling pathway. When rapamycin (inhibitor of mTORC1) is administered prior to ketamine administration, ketamine’s antidepressant effect is eliminated. It indicates that ketamine-induced synaptogenesis is dependent on mTORC1 complex activation (Li et al. 2010; Duman et al. 2012).

Two hours after ketamine administration, the levels of synaptic proteins such as synapsin I, PSD95, and GluR1 increase. This effect is sustained for 7 days. Twenty-four hours after ketamine application, the number of mushroom spines increases (Li et al. 2010; Duman et al. 2012), which indicates the maturing and strengthening of synaptic connections (Yoshihara et al. 2009). The increased affectivity of synaptic transmission is apparent from the increased amplitude of EPSCs facilitated by insertion of AMPA receptors into postsynaptic membrane (Li et al. 2010). Ketamine stimulates signaling pathways involved in protein synthesis and formation of dendritic spines. Recent studies confirmed the negative effect of stress upon number, density, and function of dendritic spines and decreased synaptic plasticity in depressive patients (Marsden 2013). Single ketamine application reverses the deficit in number and function of dendritic spines caused by exposure to chronic stress (Liu and Aghajanian 2008; Li et al. 2011). In addition, stress and depression decrease the expression and release of brain-derived neurotrophic factor (BDNF), resulting in disruption of neurotrophic support leading to atrophy and loss of neurons (Duman 2014). Autry and colleagues showed that ketamine administration increases BDNF translation that mediates rapid behavioral antidepressant responses (Autry et al. 2011).

Understanding of ketamine’s mechanism of action can help us to target mTOR signaling pathway more effectively and come up with strategies leading to the discovery of novel, fast-acting antidepressants with fewer adverse effects. The goal of present study is to examine antidepressant effect of ketamine in the animal model of depression. The bilateral olfactory bulbectomy is an animal model of depression that disrupts the integrity of limbic system known to be affected in patients suffering from major depressive disorder. Bilateral olfactory bulbectomy (OBX) leads to cellular, structural, biochemical, and behavioral changes that can be reversed by chronic treatment with antidepressants (Song and Leonard 2005).

The behavioral tests we chose served to assess locomotor activity (open field), anxiety (elevated plus maze), and cognitive functions (passive avoidance, Morris water maze, and Carousel maze), because hyperactivity, decreased anxiety, and deficits in learning and memory are some of the behavioral changes detected after olfactory bulbectomy. In addition, hippocampi and prefrontal cortex were removed for BDNF analysis and blood samples were collected for analysis of phosphorylated mTOR level. The antidepressant effect of ketamine was assessed 24–144 h after application in order to inspect its prolonged effect on behavior. Since the antidepressant effect of ketamine was reported to last 7 days, all the tests were terminated within 1 week postinjection. Rapamycin was administered in order to investigate if the ketamine’s antidepressant effect was mediated by mTOR signaling pathway.

Materials and methods

Animals

Male Wistar rats (300–350 g) used in the experiments were obtained from the Institute of Physiology; accredited breeding colony originated from Charles River Laboratories, Inc. Animals were housed in transparent plastic cages (30 × 30 × 40 cm) in an air-conditioned animal facility with constant temperature, humidity, and 12/12 light/dark cycle. Water and food were available ad libitum. Experiments were carried on during daylight hours. All procedures were in accordance with Czech and European legislation regarding treatment of laboratory animals (Directive 86/609/EEC).

Experimental design

Figure 1 depicts the time scheme of experiments and the sample collections. After the surgery, the animals had 3 weeks for recovery before they were administered with drugs. Blood samples were collected before and 30 min after drug application, and the experiments started 24 h later. In the first day of behavioral procedures, the animals were tested in open field (OF) and elevated plus maze (EPM). The cognitive tests were pursued on subsequent days. The animals were tested either in passive avoidance (PA) and Morris water maze (MWM) or in active place avoidance on a rotating arena (Carousel maze (CM)). Animals were not tested in all cognitive tests in order to complete the planned experiments within 1 week. Numbers of animals in behavioral tests were as follows: in OF, EPM, and CM (OBX saline n = 14, OBX ketamine n = 14, OBX saline + rapamycine n = 12, OBX ketamine + rapamycine n = 12, sham saline n = 9, sham ketamine n = 11, sham saline + rapamycine n = 11, and sham ketamine + rapamycin n = 15); in PA (OBX saline n = 6, OBX ketamine n = 7, OBX saline + rapamycine n = 8, OBX ketamine + rapamycine n = 8, sham saline n = 7, sham ketamine n = 8, sham saline + rapamycine n = 8, and sham ketamine + rapamycin n = 9); and in MWM (OBX saline n = 8, OBX ketamine n = 9, OBX saline + rapamycine n = 8, OBX ketamine + rapamycine n = 8, sham saline n = 9, sham ketamine n = 11, sham saline + rapamycine n = 9, and sham ketamine + rapamycin n = 9). For mTOR analysis, the group sizes were as follows: OBX saline n = 10, OBX ketamine n = 10, OBX saline + rapamycin n = 10, OBX ketamine + rapamycin n = 8, sham saline n = 12, sham ketamine n = 8, sham saline + rapamycin n = 8, and sham ketamine + rapamycin n = 14 for 30-min analysis and OBX n = 12 and sham n = 10 for 0-min analysis. On the seventh day after drug application, the brains were removed for analysis of BDNF. The numbers of animals for BDNF analysis were OBX saline n = 7, OBX ketamine n = 8, OBX saline + rapamycin n = 11, OBX ketamine + rapamycine n = 8, sham saline n = 9, sham ketamine n = 9, sham saline + rapamycine n = 7, and sham ketamine + rapamycin n = 9.

Fig. 1
figure 1

Experimental design

Full size image

Surgery

The rats were anesthetized with isoflurane (no. B306, Abbot Laboratories, Queen-Borought, UK) in a preparation chamber (3 % of isoflurane) and then placed in a stereotactic apparatus (no. 430005-GR-GP/K, TSE systems, Germany). Level of isoflurane anesthesia during the operation was held on 2 %. For local anesthesia, 0.5 ml of mesocaine 1 % (Zentiva) was applied before an incision. The eyes were treated with Vidisic ocular gel (Bausch + Lomb). The incision was made in the scalp above the olfactory bulbs. Two 2-mm burr holes were made by a microdrill (Dremel), −8 mm AP and ±2 mm ML from bregma. The olfactory bulbs were removed by a blunt hypodermic needle connected to a water pump (performed according to van der Stelt et al. 2005). The hemorrhage was stopped and the incision was sutured using absorbable material. The sutures were treated with local antibiotic Framykoin (Zentiva). Sham-operated rats underwent the same procedure without removal of the olfactory bulbs. After surgery, 2.5 ml of sterile water was administered intraperitoneally (i.p.) and analgesic Nurofen was added into drinking water. The experiments were performed after 3 weeks of postoperative period. After the termination of behavioral procedures, the brains were extracted and the lesions were inspected. Animals with remains of olfactory tissue were excluded from the analyses.

Drug administration

Within OBX or sham-operated group, animals received either a single i.p. injection of ketamine (10 mg/kg) or saline or were i.p. co-injected with ketamine (10 mg/kg) and rapamycin (1 mg/kg) or saline and rapamycin (1 mg/kg). Rapamycin (Sirolimus) was purchased from ApexBio and dissolved in a vehicle solution containing 20 % DMSO and saline. An injection of rapamycin was applied 30 min prior to ketamine or saline administration. Ketamine was obtained as a ketamine hydrochloride from Vétoquinol Ltd. All animals received the same volume of liquid per 1 kg of body weight (1 ml/kg of b.w.). Blood samples from the rat tail vein were collected at 0 and 30 min after the drug application for later analysis of phosphorylated mTOR level. Rats were subjected to behavioral testing 24 h after drug administration. After behavioral tests were completed, rats were decapitated and hippocampi and prefrontal cortex were removed and stored in −80 °C for later BDNF analysis.

Immunoassay

ELISA for mTOR analysis

Phosphorylated mTor activity was detected in cell lysates by K-LISA™ mTor Activity Kit (Merck KGaA, Darmstadt, Germany). All procedures were done according to the manufacturer’s instructions. Briefly, cell’s pellet was washed twice with Tris-buffered saline (TBS). Lysate was prepared by adding 1 ml of lysis buffer and incubated on ice for 20 min. Cell’s lysate was precleared by adding protein A/protein G-plus agarose beads (Merck KGaA, Darmstadt, Germany) according to the manufacturer’s instructions twice. Anti-mTOR/FRAP mouse monoclonal antibody (Merck KGaA, Darmstadt, Germany) needed to be added to precleared lysates to efficiently immunoprecipitate mTOR from crude biological samples. Protein agarose beads (now bound to immunoprecipitated mTOR protein) were added 2× Kinase Assay Buffer wash buffer and mTor substrate wash buffer to prepare reaction mixture for the assay. Before use, the mixture was centrifuged to pellet agarose beads. Supernatant was transferred to clear tube and stored at −70 °C before use. Standards as well as samples were pipetted to designated wells on a glutathione-coated plate. The contents of the wells were aspirated and washed three times by adding 200 μl of TBS. Particular substances were added to the assay in following order: 100 μl of the p70S6K–GST fusion protein, 100 μl of HRP conjugate, and 100 μl of tetramethylbenzidine (TMB) substrate. Each substance was incubated for 1 h on the plate. After incubation, all substances were washed out. The TMB reaction was stopped by adding 100 μl of stop solution. The plate was read at a wavelength of 450 nm with a reference wavelength of 595 nm using a 96-well plate spectrophotometer (Victor, Work Out version 2.5, Perkin Elmer, USA).

ELISA for BDNF analysis

Excised brain parts (prefrontal cortexes and hippocampi) were extracted according to “modified extraction procedure” of Szapacs et al. (2004), in “lysis buffer” (with slightly modified concentrations of antiproteases, 100 mM PIPES pH 7.0, 500 mM NaCl, 0.2 % Triton X-100, 0.1 % NaN3, 2 % BSA, 2 mM EDTA, 1 mM PMSF, 2.5 μM leupeptin, 0.5 μM pepstatin, and 0.75 μM aprotinin). Of the homogenates, 10 % w/v was prepared by sonication (Bioblock Scientific 72442, France) with microtip (2 mm) on ice by four pulses (setting 60) for 5 s. The homogenates were acidified (25 μl 1 N HCl/0.2 ml homogenate in Eppendorf microtube) and left overnight at refrigerator (ca. 4 °C). The next day, extracts were centrifuged (12 600 rpm/15 min/4 °C in Eppendorf 5415R centrifuge) and to 0.15 ml of supernatants, 1 μl of Neutral Red (10 mg/ml 50 % ethanol) were added, and the samples were neutralized by 1 N NaOH under visual control (careful titration until the color started to change from red to yellow). Neutralized samples were frozen and kept at −40 °C until assayed by ELISA. Content of BDNF was assayed with R&D Systems DuoSet ELISA BDNF Kit DY248 (USA) according to the manufacturer’s instructions.

Behavioral procedures

OF

Rat’s spontaneous activity in new environment was observed in the open field test for 10 min. OF (50 × 50 cm) was dimly illuminated by red light considered to be less distressing for rodents. Activity was detected by evenly spaced infrared light beams. Beam interruptions caused by movements of the animal were registered by the software (Multi Conditioning System, TSE Systems, Germany). Software analyzed the number of rearings and total distance travelled. In addition, the hyperactivity (defined as movement exceeding 20 cm/s) was examined as well. The analysis was carried out twice, the first 2 min and the last 2 min of OF activity separately, since the activity is supposed to be the most intense shortly after exposure to the new environment.

EPM

The apparatus consisted of two open arms (45 × 10 cm) crossed at right angles with two opposed arms of the same length. Two of the opposed arms were enclosed by walls 40 cm high, except for the central platform where the arms crossed. The whole apparatus was elevated 50 cm above the floor. At the beginning of each experiment, the rat was placed on the central platform facing the closed arm. Time spent in open arms, closed arms, and central platform and distance moved were recorded during a 10-min test session by a camera positioned above the maze and analyzed by the software (Tracker, Biosignal Group, USA).

Step-through PA

Arena (50 × 50 cm) consisted of two equally sized compartments divided by sliding doors (Multi Conditioning System, TSE Systems, Germany). Animal was always placed into intensively illuminated compartment. During habituation session, rats moved freely between two compartments for 5 min; no shock was delivered. Thirty minutes after habituation, rats were placed into arena again for training session. After the animal crossed to the dark compartment, the sliding door dropped down. The rat received the mild electric shock (2.5 mA, 8 s) through the stainless steel grid floor. The actual testing followed 1 and 24 h after training session. During the 5-min test session, the shocks were not delivered. The latency to cross to the dark chamber was evaluated by the software.

MWM

MWM is used to study spatial learning and memory (Morris 1984; Stuchlik et al. 2007). Rat was placed into circular pool (180 cm in diameter) filled with water to a height of 40 cm. A transparent Plexiglas platform was submerged 1 cm under the water surface. The animal navigated using distal cues as points of reference. Escape latency was assessed as a measure of learning. The rat was detected by infrared camera attached to the ceiling, and trajectory was digitalized by the software (Tracker, Biosignal Group, USA and DT-3155 Card, Data Translation, USA). Escape latencies were measured by a stopwatch. Each rat underwent eight trials per day for four subsequent days. At the beginning of every trial, the rats were released from the same location positioned in the “east” of the pool. The platform was located in the middle of “southwest” quadrant.

Active place avoidance on CM

CM is an apparatus for assessment of active avoidance behavior consisting of rotating circular arena. CM can be used to test cognitive coordination and spatial avoidance navigation. It consisted of a metallic circular arena (82 cm in diameter) placed 1 m above the floor with transparent Plexiglas wall (40 cm high). The rat was tracked by an infrared camera located on the ceiling directly above the arena, and trajectory was digitized by the software (Tracker, Biosignal Group, USA). On the arena surface, a stable unmarked 60° to-be-avoided sector was defined by the room-frame coordinates. The arena rotated with the constant speed of one rotation per minute. Upon the entrance to the sector, animal received a mild electric footshock. The animal had to distinguish between two conflicting spatial representations (arena vs. room) and navigate according to relevant stimuli. The apparatus is described in detail elsewhere (Stuchlik et al. 2013).

Behavioral training consisted of four sessions lasting 20 min. The animals underwent two sessions per day. The performance was evaluated using cognitive parameters—the number of entrances into the to-be-avoided sector and the maximum time avoided within a session. These parameters assess the overall cumulative performance and an ability to maintain avoidance throughout a session, reflecting both between-session and within-session learning. The escape response was measured by the number of shocks rats received. The locomotor activity expressed as the total distance actively walked by the animal (after filtering out passive movement of the arena) was evaluated as well. For statistical analysis of cognitive performance, data from the last session only were taken into account. Since hyperactivity of OBX animals is the most apparent in a novel environment, the distance walked and shocks received were evaluated during the first 20-min session.

Statistical analysis

Statistical analyses were performed by the program GraphPad Prism 5.0 (San Diego, CA, USA) and IBM SPSS Statistics 20 (MWM). The statistical significance for OF test, EPM, and PA tests was detected by a two-way ANOVA with treatment effect (four-factor lever) and the bulbectomy effect (two-factor level) serving as independent variables. In MWM, the effect of treatment and effect of bulbectomy on the performance of animals were assessed by two-way repeated measure ANOVA. Bonferroni’s post hoc test was used when appropriate. Outliers identified by Grubbs test were removed from the analysis. If normality was not assumed, data were transformed. Data are presented as the group means ± standard error of mean (SEM). The significant level was set at <0.05.

Results

OF

Statistical analysis of spontaneous activity during the first 2 min revealed significant effect of the factor bulbectomy, F (1, 90) = 14.61, p = 0.0002, on the locomotion. No significance was detected for the effect of treatment and interaction. Post hoc test did not detect any differences (Fig. 2a). Analysis of the last 2 min did not show any effect of either variable or their interaction (Fig. 3a).

Fig. 2
figure 2

The first 2 min in the open field. The graphs illustrate spontaneous activity in open field during the first 2 min of 10-min session. In all parameters assessed, OBX animals significantly differed from sham controls. a Distance walked by OBX rats was significantly increased when compared to controls. b OBX rats displayed significantly elevated hyperactivity as well. c The most profound difference between OBX and controls was seen in vertical activity. Number of rearings was significantly greater in OBX animals. All values represent group means ± SEM

Full size image
Fig. 3
figure 3

The last 2 min in the open field. The graphs illustrate activity during last 9 and 10 min of a 10-min session. The increased ambulation expressed in terms of distance walked (a) and time spent hyperactive (b) normalized as rats became habituated to the new environment. The number of rearing remained significantly increased in OBX rats. c The habituation rate of vertical activity was slower indicating increased emotionality of OBX rats. All values represent group means ± SEM. *p < 0.05

Full size image

Similarly; the effect of bulbectomy on cumulative time spent by hyperactivity (˃20 cm/s) came out as significant, F (1, 90) = 11.66, p = 0.0010, for first 2-min analysis. No effect of treatment and interaction on behavior was detected (Fig. 2b). Analysis of last 2 min found no significant difference between groups (Fig. 3b).

The effect of bulbectomy on number of rearings during first 2 min was considerable, F (1, 81) = 35.86, p < 0.0001.The effect of treatment and interaction between two effects was not significant (Fig. 2c). The analysis of the first 2 min copied the results from last 2-min analysis, F (1, 81) = 8.709, p = 0.0041, for the effect of bulbectomy. No difference in the effect of interaction and treatment was revealed. Bonferroni’s post hoc test revealed significant difference between OBX rats treated with ketamine vs. OBX rats treated with saline + rapamycin (p < 0.05; Fig. 3c).

EPM

Two-way ANOVA detected significant effect of the treatment F (3, 90) = 3.044, p = 0.0329 and bulbectomy F (1, 90) = 7.877, p = 0.0061 on the distance travelled in EPM. The interaction between two effects did not come out as significant. The OBX animals treated with saline significantly differed from OBX animals treated with ketamine + rapamycin (p < 0.001) or ketamine alone (p < 0.01; as post hoc test showed) with saline-treated animals being more active (Fig. 4a).

Fig. 4
figure 4

Elevated plus maze. In both parameters tested, OBX animals exhibited significantly altered behavior when compared to sham controls. The effect of treatment was significant as well. a Locomotor activity was significantly increased in OBX rats as a response to novelty exposure. Ketamine treatment resulted in suppression of hyperlocomotion exhibited by saline-treated OBX rats. However, co-application of ketamine + rapamycin had similar effect on locomotion as ketamine administration alone. b Open arm avoidance was significantly reduced in OBX rats compared to controls. Time spent in open arms was increased in OBX rats treated with ketamine. Rapamycin eliminated the anxiolytic effect of ketamine in ketamine + rapamycin-treated OBX animals. All values represent group means ± SEM. *p < 0.05, **p < 0.01, ***p < 0.001

Full size image

Two-way ANOVA revealed significant effect of treatment F (3, 90) = 4.327, p = 0.0042 and bulbectomy F (1, 90) = 8.609, p = 0.0068 on time spent in open arms. No effect of the interaction between two factors was recognized. The significant difference between OBX animals treated with ketamine and OBX animals treated with ketamine + rapamycin (p < 0.001) or saline + rapamycin (p < 0.05) was detected by post hoc test. Ketamine-treated animals exhibited lowered anxiety (Fig. 4b). As the data did not pass normality test, the square root transformation was conducted to attain Gaussian distribution. However, for better illustration, original units and scales were used in the graph.

Step-through PA

Two-way ANOVA detected significant effect of treatment F (3, 53) = 3.089, p = 0.0348 on memory and learning in PA test while interaction and bulbectomy had no effect. Within a group of OBX rats, the post hoc test revealed significant difference between saline- vs. ketamine-treated animals (p < 0.05), saline- vs. saline +rapamycin-treated animals (p < 0.01), and saline vs. ketamine + rapamycin animals (p < 0.05; Fig. 5). Data were transformed using common logarithm in order to pass normality test. The graph is portrayed with original units (seconds). The treatment that was applied has positively influenced performance in the test and enhanced the latency to cross into dark compartment.

Fig. 5
figure 5

Step-through passive avoidance. No effect of bulbectomy was revealed, but the effect of treatment came out significant. Ketamine, ketamine + rapamycin, and saline + rapamycin treatments prolonged the latency to enter dark compartment after previous shock exposure. All values represent group means ± SEM. *p < 0.05, **p < 0.01 compared to the saline-treated OBX group

Full size image

MWM

For evaluation of spatial memory and learning in MWM, two-way repeated measure ANOVA was carried out. The performance during 4 days (escape latencies, first to fourth day) was set as within-subject effect and the bulbectomy and treatment used as between-subject effects. The within-subject effect came out significant, F (3, 177) = 117.023, p < 0.0001, showing that animals from all groups improved with each subsequent day. The interaction between the day and bulbectomy was significant as well, F (3, 177) = 5.521, p = 0.004. It indicates that the rate of learning was influenced by bulbectomy regardless of the treatment applied. The only significant between-subject effect was bulbectomy F (1, 59) = 84.827, p < 0.0001 (Fig. 6).

Fig. 6
figure 6

Morris water maze. OBX group significantly differed from controls in spatial navigation evaluated during 4 days when animals were released from the same starting position in the pool. Each rat was subjected to eight trials a day. Escape latencies of OBX rats were substantially increased indicating cognitive deficit caused by bulbectomy. Treatment used had no effect. Both groups improved with each subsequent day; however, the rate of learning in controls was faster. All values represent group means. For better clarity, error bars are not present

Full size image

CM

Numbers of shocks received during the first and the last sessions were assessed by two-way ANOVA. The effect of bulbectomy was significant even during the first session, F (1, 90) = 13.55, p = 0.0004. No effect of treatment and interaction was detected. Post hoc test revealed significant difference between saline-treated and ketamine-treated OBX animals (Fig. 7a). Ketamine application resulted in reduced number of shocks delivered, indicating enhanced escape response. In the last session, the effect of bulbectomy on shock number was more prominent, F (1, 90) = 37.32, p < 0.0001. The treatment effect reached the level of significance, as well F (3, 90) = 2.83, p = 0.0428. Ketamine +rapamycin-treated OBX rats significantly differed from all other OBX groups (post hoc analysis, p < 0.05 from saline and saline + rapamycin and p < 0.01 from ketamine; Fig. 7b).

Fig. 7
figure 7

Carousel maze I. Significant difference between OBX animals and controls was found in non-cognitive parameters. a The number of shocks OBX animals received during the first session was significantly lower than in controls. Sham controls lacked the escape reaction. Ketamine-treated OBX rats received significantly less shocks than saline-treated OBX animals. b Between-session decline in number of shocks received by sham controls was not as prominent as in OBX rats. OBX rats learned to avoid the sector more effectively as seen from dropping number of shocks. Ketamine + rapamycin treatment of OBX rats resulted in increased number of shocks delivered compared to OBX rats treated solely with ketamine. c The distance walked by OBX animals was significantly longer than in controls, indicating the loss of motivation to solve the task by sham controls, resulting in behavior resembling learned helplessness. Ketamine + rapamycin co-application led to decreased distance walked in OBX rats when compared to ketamine-treated OBX rats. All values represent group means ± SEM. *p < 0.05

Full size image

Two-way ANOVA analysis showed significant effect of bulbectomy in the distance walked between sham controls and OBX rats, F (1, 89) = 37.77, p < 0.0001, even during the first session. Post hoc test detected a significant difference between ketamine-treated OBX group and ketamine + rapamycin-treated OBX group (p < 0.05; Fig. 7c). Sham-operated controls had reduced ambulation, corresponding to diminished escape response after shock delivery.

The numbers of entrances and maximum time avoided during the last session were analyzed by two-way ANOVA. In parameter the number of entrances into forbidden sector, the significant effect of interaction between the effects of bulbectomy and treatment was discovered, F (3, 90) = 9.912, p = 0.0367, as well as the effect of bulbectomy, F (1, 90) = 2.008, p = 0.0022. The post hoc test detected significant difference between ketamine + rapamycin-treated OBX rats and all other OBX groups (p < 0.01 from saline and ketamine and p < 0.05 from saline + rapamycin; Fig. 8a). The next cognitive parameter evaluated was the maximum time avoided within one session. The significant effect of both bulbectomy and treatment was detected, F (1, 90) = 24.50, p < 0.0001 for bulbectomy and F (3, 90) = 4.053, p = 0.0094 for the treatment. The significant difference was again found between ketamine + rapamycin-treated OBX rats and all other OBX groups (p < 0.05 from all other treatment groups; Fig. 8b). Data of cognitive parameters and of number of shocks from fourth session were transformed using common logarithm.

Fig. 8
figure 8

Carousel maze II. In cognitive parameters, the substantial difference was found between OBX and sham controls as well. a Number of entrances (errors) made by sham controls in last session reflects cognitive deficit. Administration of ketamine + rapamycin resulted in increased number of entrances, approaching the performance of sham controls, when compared to all other OBX groups. b During the last session, OBX animals spent significantly greater portion of time outside the sector than sham controls, as indicated by the maximum time avoided. The maximum time avoided of ketamine + rapamycin-treated OBX rats was significantly reduced compared to all other OBX groups. All values represent group means ± SEM. *p < 0.05, **p < 0.01

Full size image

BDNF

Two-way ANOVA analysis showed significant difference in BDNF content between sham controls and OBX animals in prefrontal cortex, F (1, 60) = 16.85, p = 0.0001, for effect of bulbectomy. Neither treatment effect nor the effect of interaction was detected. Post hoc test did not reveal difference between groups (Fig. 9a). Hippocampal content of BDNF was comparable between control and experimental groups; neither bulbectomy effect nor treatment effect was significant (Fig. 9b).

Fig. 9
figure 9

BDNF. a BDNF content in PFC was significantly decreased in OBX animals in comparison to sham controls regardless of treatment used. b BDNF level in hippocampus did not differ between OBX and control groups. All values represent group means ± SEM

Full size image

mTOR

Measurements of absorbance at 450 nm were expressed in percentage of baseline absorbance at time 0 (before drug application), defined as 100 % absorbance. Two-way ANOVA detected significant effects of treatment F (3, 72) = 4.36, p = 0.007 and bulbectomy F (1, 72) = 15.7, p = 0.0002 as well as their interaction F (3, 72) = 3.454, p = 0.0209. Post hoc test revealed significant difference in OBX group between saline-treated and ketamine-treated rats (p < 0.05), ketamine-treated and saline +rapamycin-treated rats (p < 0.001), and saline + rapamycin- and ketamine + rapamycin-treated animals (p < 0.05; Fig. 10).

Fig. 10
figure 10

mTOR. Treatment applied affected level of phosphorylated mTOR in OBX rats that where more responsive to the treatment. The level of mTOR was significantly increased in OBX rats when compared to controls. Ketamine treatment increased the mTOR level in OBX animals when compared to saline-treated or saline + rapamycin-treated OBX rats. However, ketamine + rapamycin co-application did not reduce the mTOR level and did not eliminate the effect of ketamine. All values represent group means ± SEM. *p < 0.05, ***p < 0.001

Full size image

Discussion

The objective of the present study was to evaluate an antidepressant effect of a single ketamine injection in a variety of behavioral tests involving various degrees of stress and to investigate the role of mTOR signaling pathway in mediating its antidepressant effect. Since antidepressant effect of ketamine was reported to last up to 7 days, all tests were terminated within 1 week postinjection. In addition, we examined the changes in BDNF and mTOR levels triggered by different treatments applied.

The main results of our research may be summarized as follows: (1) OBX rats significantly differed from sham-operated controls in behavioral tests (except for PA see below), (2) treatment used had significant effect on depressive phenotype only—OBX animals were more responsive to the treatment, (3) rapamycin blocked the effect of ketamine in task-dependent manner; and (4) intensity of stressor (mild vs. severe) used in the behavioral tests had opposite effect on controls and on OBX animals.

Open field test

Activity of OBX rats in the OF arena was significantly increased compared to sham-operated animals during first 2 min. This increase in locomotion (Fig. 2a) was accompanied with increased hyperactivity (Fig. 2b) and increased vertical activity (Fig. 2c) measured by a number of rearings. Enhanced locomotor activity reflects hyper-arousal of OBX animals as a response to a novel environment. Therefore, the difference between control and experimental groups was the most prominent during the first 2 min and gradually decreased as animals habituated to the OF apparatus. The horizontal activity of OBX animals normalized to that of sham-operated controls in the final 2 min of the experiment regardless of treatment used (Fig. 3a, b). However, vertical activity of OBX animals remained substantially elevated compared to controls even during the last 2 min (Fig. 3c). Treatment did not normalize the increased activity of OBX rats in the OF expressed in terms of distance walked, hyperactivity, and rearing. However, there was a tendency of ketamine-treated OBX animals to reduce their activity when compared with OBX animals treated with saline or saline + rapamycin with the latter being significant in rearing activity during the last 2 min.

Increased exploration in open field following olfactory bulbectomy is one of the first reported behavioral changes (Sieck and Baumbach 1974) evident as soon as 1 week after surgery (Hendriksen et al. 2015). Hyperactivity, as a consequence of novelty exposure, is probably caused by increased release of glutamate in the striatum (Ho et al. 2000) and normalized by chronic treatment with antidepressants. The rearing and locomotion can be dissociated by pharmacological manipulation (Balcells-Olivero and Vezina 1997), indicating that the two behavioral parameters might differ in the neural substrates regulating their activity. It was suggested that unlike locomotion, rearing reflects not only exploratory activity but also emotionality of animals (Gironi Carnevale et al. 1990). If there is sufficient time for OBX animals to habituate to the new environment, they are able to reduce their hyperactivity to the level of sham controls. It corresponds with gradual decrease of elevated glutamate levels in striatum after habituation to the open field lasting at least 10 min (Ho et al. 2000). Hyperactivity is the most frequently detected alternation in behavior, and our data are consistent with literature findings (Table 1). Interestingly, it was documented that animals with higher locomotor and rearing responses to novelty showed higher rate of habituation with repeated testing (Thiel et al. 1999).

Table 1 The table describes bulbectomy-induced alternations in open field and elevated plus maze behavior reported in literature
Full size table

Elevated plus maze

The EPM is a standard test for evaluation of anxiety or anxiolytic effect of compounds administered (Dawson and Tricklebank 1995; Carobrez and Bertoglio 2005), where the time spent in open arms is indicative of lowered anxiety (Hogg 1996). In the EPM test, OBX and sham controls significantly differed in both parameters tested, distance walked and time spent in open arms. Ketamine administration resulted in reduction of novelty-induced hyperlocomotion when compared with OBX animals injected with saline (Fig. 4a). In the parameter time spent in open arms, ketamine exerted strong anxiolytic effect that was completely blocked by prior rapamycin administration (Fig. 4b). In spite of increased ambulation of saline-treated OBX rats, the animals were able to partially adjust their behavior as a reaction to aversive stimuli represented by open arms and avoid them.

After olfactory bulbectomy, a degeneration of inhibitory projection from olfactory lobes to amygdala occurs, resulting in reduced inhibitory input and subsequent disinihibition of the amygdala. Prolonged disinhibition of amygdala was reported also after chronic stress, which is relevant to depressive disorders (Liu et al. 2014). Amygdala plays a crucial role in processing of emotional stimuli, memory, and decision making. Therefore, it is not surprising that disinhibition of amygdala may result in sensitization to stress (McNish and Davis 1997) and may account for altered behavioral changes in OBX animals. Hyperexcitability of amygdala in OBX animals correlates with increased 2-deoxyglucose uptake indicating an increased demand for energy by axon terminals (Shibata and Watanabe 1994). It is compatible with the hypothesis that limbic hyperactivity in the amygdala, during evaluation of emotional stimuli, combined with prefrontal hypoactivity leading to a failure of cortical control, might cause negative emotional biases in depressed patients (Drevets 1999; Phillips et al. 2003; DeRubeis et al. 2008). OBX rats appear to be sensitive to increasingly aversive environments (Primeaux and Holmes 1999). This may explain partial avoidance of open arms in EPM by OBX animals treated with saline and differences between three cognitive tests used. The results from literature reported increased anxiety, decreased anxiety, and no change in anxiety scores in OBX animals (Table 1). However, anxiety phenotype of sham controls detected in EPM in our study is consistent with results obtained from cognitive test.

Cognitive tests

The most widely used test for assessment of memory function in OBX rats is PA learning based on the association formed between an aversive stimulus and a specific environmental context. The PA is the only test from whole behavioral assay used where the difference between OBX animals and sham controls was not found, whereas the treatment used affected the performance. The saline-treated OBX rats displayed significantly shorter latency to enter dark compartment compared to all other experimental groups (Fig. 5). Twenty-four hours later, no difference was found (data not shown). These results are in opposition to the results from the MWM. However, the locomotor, cognitive, and motivational variables are confounded in the passive avoidance test (Bizot and Thiébot 1996). Therefore, interpretation of obtained results as a cognitive deficit in OBX rats injected with saline may be misleading. Our results are in accordance with literature. Memory deficit in passive avoidance task after bulbectomy is frequently detected (Table 2).

Table 2 The table describes bulbectomy-induced memory deficits in passive avoidance and Morris water maze reported in literature
Full size table

In spatial learning assessed in MWM, OBX animals displayed significantly longer escape latencies when compared to sham-operated controls regardless of treatment applied. Both groups improved with each subsequent day; however, the rate of acquisition was faster in control group (Fig. 6). The results are in agreement with findings from other studies assessing cognitive functions after bulbectomy (Kelly et al. 1997; Song and Leonard 2005). Nevertheless, several studies reported no change in acquisition between control and OBX groups (Table 2). But, there are some evidence that deficit in MWM is transient and with prolonged time from surgery it spontaneously recovers (van Rijzingen et al. 1995).

The most intriguing results were obtained from active place avoidance in CM. The OBX phenotype was again significantly distinct from controls, but this time controls exhibited cognitive deficits. Cognitive parameters, number of entrances into the forbidden sector and the maximum time avoided, were assessed in the last fourth session. Both parameters were affected by the treatment administered. Prior rapamycin administration suppressed the effect of ketamine, leading to impaired cognitive coordination, memory, and learning expressed in terms of increased number of entrances (Fig. 8a) and decreased maximum time outside the forbidden sector (Fig. 8b) exhibited by ketamine + rapamycin-treated OBX rats. Their level of performance approximated that of sham controls. Number of shocks received correlates with the number of entrances; however, it illustrates well that sham animals lacked the escape reaction. Between-session decline in number of shocks received by sham controls was not as prominent as in OBX rats (Fig. 7b). On the other side, the OBX animals exhibited increased sensitivity to shocks even during the first session (Fig. 7a) and it intensified with subsequent sessions (Fig. 7b). Ketamine-treated OBX animals reacted to the shocks to a greater extent than all other experimental groups in first session (Fig. 7a). In the last session, prior rapamycin administration to ketamine-treated OBX resulted in lack of escape reaction (Fig. 7b). To the best of our knowledge, there are no data available about bulbectomy experiments conducted in CM test to compare it with our results.

Afore-mentioned results indicate that OBX animals can adjust their behavior to increasingly aversive environments. The number of shocks received during the first session had opposing impact on control and experimental groups. OBX rats learned to avoid the sector more effectively. The sham-operated controls resigned to solving the task, and they let themselves to be dragged passively to the forbidden sector. In some, it even resulted in behavior resembling learned helplessness, marked not only by cognitive deficits but also by shorter distance walked (Fig. 8c). The active place avoidance in CM is very demanding and challenging task which requires constant attention. The design of the experiment was very stressful for the animals that underwent two 20-min sessions a day, 2 days in a row. For successful solving of stressful cognitive tasks, the amygdala plays a crucial role and may account for the differences between cognitive tests.

The amygdala is crucial for mediating behavioral response to stress (Regev et al. 2012) and therefore influences learning under stressful conditions (Kim et al. 2001). Although the amygdala-lesioned animals showed deteriorated performances in PA, they executed spatial learning tasks as well as the control animals (Takashina et al. 1995). It was demonstrated that stress and glucocorticoids cause changes in learning strategy. Stress impaired hippocampal-dependent learning and led to the switch from spatial learning to the stimulus–response learning. Amygdala mediates stress effects on hippocampal LTP and learning, resulting in impaired performance in the MWM (Kim et al. 2001). These findings are in accordance with our results obtained from MWM. Hyperactive amygdala negatively affects spatial learning. In line with this reasoning, the performance of OBX rats in PA test should not be affected. However, the results from passive avoidance might not express cognitive deficits solely (Bizot and Thiébot 1996). The decreased latency displayed by saline-treated OBX rats may reflect lack of impulse control (Kamei et al. 2007) and/or defensive behavior (Stock et al. 2001) that were observed in OBX rats.

CM is highly hippocampal-dependent spatial test requiring hippocampal integrity since even unilateral inactivation leads to deficits in cognitive functions (Cimadevilla et al. 2001). Therefore, the amygdala hyperexcitability coupled with extremely stressful task should deteriorate performance in CM in OBX animals. During episode of stress, corticotropin-releasing factor (CRF) is released into the amygdala where activation of local CRF receptor-containing neurons results in stress-induced alternations in behavior. Stress-induced activation of basolateral amygdala by CRF could shift the balance between glutamate-mediated excitation and GABA-mediated inhibition toward greater excitability and override the effect of prefrontal inhibition. The ability to cope with the aversive stimulus is reduced (Shekhar et al. 2005). Habituation to repeated stress exposure is an adaptive response, wherein repeated exposure reduces hypothalamic–pituitary–adrenal (HPA) axis activation. Stress of sufficient intensity may be able to block the habituation process and lead to maladaptive consequences (Herman 2013). This might account for cognitive impairment of sham controls in active place avoidance task. OBX phenotype is marked by disinhibited amygdala. Therefore, excessive exposure to severe stress might blunt the effect of CRF release on already hyperactivated amygdala and consequently on the behavior in OBX animals. Complementary explanation is that the arena-bound olfactory stimuli, owing to bulb ablation, were not perceived by OBX rats and the absence of conflicting spatial representations partly rescued the cognition.

In our previous study, the effect of elevated levels of CRF and corticosterone on cognitive functions in rats was studied in our laboratory (Rezacova et al. 2011). Rats were exposed to elevated CRF levels applied by i.c.v. osmotic mini-pump for a period of 28 days. This manipulation resulted in increased release of corticosterone. Activation of glucocorticoid receptors by corticosterone constitutes the negative feedback loop on the secretion of CRF from hypothalamus. However, after prolonged stress exposure, hypothalamus becomes insensitive to high levels of glucocorticoids (GCs), resulting in feedback resistance. The dysregulation of the HPA axis results in impaired HPA axis habituation, subsequently leading to cognitive impairment. Elevated levels of CRF and GC resulted in permanent cognitive deficit in CM, while the performance in MWM assessing working memory was not affected (Rezacova et al. 2011). CRF regulates adaptive mechanisms during stress response, but its high levels lead to anxiety (Heinrichs and Koob 2004). These previous results are in accordance with results obtained from the present study and support the hypothesis that impairment in cognitive coordination displayed by sham controls may be caused by stress-induced activation of amygdala and impaired stress response. However, we did not evaluate any physiological marker of stress; therefore, we made assumption based solely on the nature of and phenomenology of the task.

MWM and CM spatial tests differ in their complexity, severity of stressor applied, and spatiotemporal design. The mild and severe stresses had opposite effects on control and OBX animals. In addition, the distinct features of the spatial tasks used might explain these discrepancies. MWM requires very precise navigation since the hidden platform constitutes less than 8 % of MWM surface. On the other hand, to-be-avoided sector in CM makes more than 16 % of the maze surface. OBX rats conducted to MWM were losing motivation to search the pool if they were not to find the platform within the first few seconds, marked by gradually decreasing velocity and shifting to random search of the pool (personal observation). In CM, passive strategy would result in excessive number of shocks received, and therefore, it might increase motivation to avoid the sector. The time for acquisition of the task may play a role, too. In MWM, the rats spent in the maze the maximum of 32 min (often a lot less) spread between 4 days. In CM, the rats underwent two 20-min sessions a day. During the first session in CM, no difference in cognitive parameters was found (data not shown) between two phenotypes tested contrary to difference in number of shocks received, indicating the escape deficits in control animals (Fig. 7a, b). It suggests that if the aversive enough stressor is applied to OBX animals for extended period of time, the OBX animals are capable to cope with the task. On the other hand, severity of stressor hindered control animals from proper information processing required for task solving. However, the resistance of OBX rats to severe stress experienced during CM testing remains to be investigated. The feedback control of HPA axis activity by prelimbic region of medial prefrontal cortex (mPFC) (Radley et al. 2006) via GC-mediated inhibition (Akana et al. 2001), together with its crucial role in cognitive functions, (Granon and Poucet 2000) might play a role. With numerous connections between mPFC and amygdala (Vertes 2004), amygdala is recruited in stress response pathway (Akana et al. 2001). In the present study, we did not examine any molecular changes in amygdala to support our conclusions. Analysis of molecular changes in amygdala will be the subject of the future study.

Treatment effects

It was suggested that activation of mTOR signaling pathway promotes therapeutic effect of rapid-acting antidepressants (Dutta et al. 2015; Krystal et al. 2013). Assumption was based on the discovery that inhibition of this pathway diminishes the antidepressant effect of ketamine in behavioral tasks such as novelty-suppressed feeding (Li et al. 2011), sucrose consumption test (Li et al. 2011), and reverses ketamine-induced dendritic outgrowth and synaptogenesis (Li et al. 2010, 2011). The effect of rapamycin-induced inhibition of protein synthesis is observed 24 h but not 1 h after its application (Iijima et al. 2012). Therefore, rapamycin has profound effect on long-term memory formation and inhibits fear/spatial memory consolidation and retrieval, leaving acquisition unaffected (Blundell et al. 2008; Deli et al. 2012; Fifield et al. 2015).

Experimental testing started 24 h after drug application. With increasing severity of stressors applied in the tests, the treatment effect was becoming more significant. In OF, the treatment did not affect the performance. In EPM, ketamine showed antidepressant and anxiolytic properties when applied to OBX rats; ketamine normalized the distance walked and increased the time spent in open arms. Pretreatment by rapamycin eliminated the anxiolytic effect of ketamine in EPM. In cognitive tests, the ketamine treatment had positive effect in PA and CM tests, whereas in MWM, no effect of treatment was detected. This inconsistency may arise from task-specific demands, amygdala’s hyperactivity, and from spatiotemporal design (as described above).

However, it should be noted that OBX animals treated with saline + rapamycin and ketamine + rapamycin in PA achieved the same results as ketamine-injected OBX rats. The treatment in passive avoidance might have suppressed the impulsivity and enhanced defensive behavior. In CM, ketamine + rapamycin administration led to cognitive impairment not observed in saline-, ketamine-, or saline + rapamycin-treated OBX rats. The performance of ketamine + rapamycin-injected OBX animals approached that of controls. It suggests that ketamine + rapamycin application resulted in impaired stress response manifested by cognitive deficits in CM.

Rapamycin application exhibits task specificity, and in some tasks, it eliminates ketamine antidepressant effect, while in others, it does not. It appears that rapamycin affects the anxiety by itself. Indeed, Hadamitzky et al. showed that systemic rapamycin injection enhanced electrical activity in amygdala accompanied by increased anxiety-like behavior in EPM (Hadamitzky et al. 2014). Prenatal rapamycin administration increased anxiety-like behavior during postnatal development and adulthood (Tsai et al. 2013). This may account for avoidance of open arms in EPM and avoidance of shocks in saline + rapamycin-treated OBX animals. Balance between mTOR inhibition and activation is essential to maintain neuronal health and function in order to adequately respond to stress challenge (Polman et al. 2012). In CM, co-application of ketamine, that has anxiolytic effect and enhances mTOR activity, together with rapamycin, which has diametrically opposite effect on anxiety and mTOR pathway, may distort this delicate balance and result in impaired stress response. High-intensity stress may suppress mTOR so that it cannot further stimulate some stress responses (Aramburu et al. 2014) and it may be manifested as learned helplessness-like behavior seen in sham controls and ketamine + rapamycine-treated OBX rats.

The treatment applied had minimum effect on performance of sham controls. Even more surprising is the finding that treatment had no effect upon mTOR level in sham animals. It seems as if ketamine influenced mTOR level preferably in OBX group. From the results of EPM test, it seems that both phenotypes (OBX and sham) differ in baseline lever of anxiety—saline-treated sham controls exhibiting higher anxiety scores than saline-treated OBX rats. The drug application therefore was more efficient in phenotype with lowered anxiety level. Increased anxiety may negatively influence performance under stressful condition as in CM test. The reason for increased anxiety in sham controls needs to be thoroughly investigated.

The depression and anxiety disorders share some common features beside treatment medication (SSRI, tricyclics, and MAOIs for the treatment of both diseases). Several studies pointed to the relationship between anxiety and greater amygdala reactivity, indicative of its defective function in anxiety disorders (Brühl et al. 2011; Hyde et al. 2011; Shah et al. 2009). Anxiolytics enhancing GABAergic transmission (Sanders and Shekhar 1995; Nuss 2015) or/and reducing NMDAR-mediated glutamate transmission (Sajdyk and Shekhar 1997) influence excitability of amygdala and restore its inhibitory tone and normal activity (Paulus et al. 2005). Increased anxiety-like behaviors may play a key role in onset of affective disorders. Most behavioral tests for assessment of antidepressant and anxiolytic properties of various compounds use some kinds of stressor (open spaces, high illumination, shocks, cold water, social defeat, etc.). If animals are capable of coping with stressor-induced anxiety, they probably do better in other tasks as well, e.g., cognitive tests. This prompts a question as to what extent is antidepressant effect of ketamine attained by its anxiolytic properties. Can anxiolytic and antidepressant effects of ketamine be separated from each other? Or can administration of anxiogenic drugs suppress the effect of antidepressant treatment? Of course, brain circuits and role of neurotransmitters involved in regulation of anxiety are much more complex and not yet fully understood.

BDNF

The BDNF level in the prefrontal cortex of OBX animals was significantly reduced when compared to sham-operated controls (Fig. 9a). This finding is in agreement with our previous results showing substantial difference between OBX groups and sham controls in behavioral tasks assessing executive functions relaying on prefrontal cortex. Interestingly, no difference in hippocampal BDNF content between experimental and control groups was detected (Fig. 9b). It suggests that the difference between two phenotypes in behavioral tests was at least partly associated with imbalance of BDNF in prefrontal cortex. However, Hellweg et al. reported an increase in BDNF level in hippocampus and frontal cortex after bulbectomy in mice (Hellweg et al. 2007). Possible explanation for this controversy may lay in differences in experimental design. In the present study, the rats underwent several stressful behavioral tasks, while in Hellweg’s study, mice were subjected to open field before BDNF analysis. It was documented that increased level of GC down-regulates BDNF expression (Dwivedi et al. 2006) and it hints at differential impact of behavioral procedures on BDNF levels between two studies. Lack of treatment effect on BDNF level was not surprising since BDNF level was analyzed a week after treatment administration when the effect of a single i.p. application of ketamine is very unlikely. Therefore, we may not draw any conclusions about ketamine effect on BDNF expression.

In animal models of stress, it was repeatedly shown that acute and chronic stresses decrease the expression of BDNF in hippocampus (Roceri et al. 2002; Grønli et al. 2006). In clinical and postmortem studies, BDNF levels were found to be lower in depressed patients than healthy volunteers in hippocampus and PFC (Brunoni et al. 2008; Yu and Chen 2011). Chronic treatment with antidepressant as well as application of ketamine increase level of BDNF in the brain (Garcia et al. 2008a; Autry et al. 2011; Yang et al. 2013). However, the BDNF knockout did not induce depressive-like behavior in male mice (Groves 2007; Monteggia et al. 2007), and few studies reported no change in BDNF level in depressed patients (Fernandes et al. 2009; Gustafsson et al. 2009). Similarly, ketamine’s antidepressant action was not always positively correlated with increased BDNF levels (Garcia et al. 2008b; Machado-Vieira et al. 2009; Rybakowski et al. 2013). Even if not being an etiological factor for onset of depression, imbalance in BDNF expression is co-existing circumstance accompanying the illness.

mTOR

OBX animals were more responsive to the treatment than sham controls; their peripheral level of phosphorylated mTOR was significantly elevated (Fig. 10). The difference between OBX and sham animals in mTOR level was not detected before drug administration (0 min, data not shown), indicating that bulbectomy resulted in increased susceptibility to treatment, outcome corroborated by behavioral analyses. Ketamine administration led to increases in mTOR level in OBX animals. This ketamine-induced activation of mTOR signaling pathway was reported previously in preclinical (Yang et al. 2013) and clinical studies (Denk et al. 2011).

It is rather surprising that ketamine + rapamycin treatment did not suppress the rise in mTOR level (Fig. 10). However, its co-application resulted in impaired behavior in some behavioral tests used—especially those where animals were exposed to severe stress. The stress might play a pivotal role in regulation of mTOR signaling pathway and in energy metabolism, balancing catabolic and anabolic cell processes. Indeed, alternations in mTOR signaling were observed after chronic stress. Chronic exposure to stress decreases phosphorylation level of mTOR in amygdala (Chandran et al. 2013) and ERK in frontal cortex and hippocampus of rats (First et al. 2011). Postmortem studies revealed a significant reduction in mTOR protein expression (Jernigan et al. 2011) and altered kinase activity of Akt (Karege et al. 2007; Dwivedi et al. 2010) in depressive patients. The imbalance in on-off switch to cell metabolism induced by ketamine + rapamycin application might have been intensified by stress encountered during behavioral testing resulting in cognitive deficits. It prompts a question if sole dysfunction of mTOR pathway is responsible for manifestation of depressive symptoms.

Conclusions

The olfactory bulbectomy induced changes in phenotype, which corresponded to depressive-like behavior. OBX led to increased horizontal and vertical activities and decreased anxiety when assessed in OF and EPM. The severity of stress applied in cognitive tests had differential effects upon OBX animals and sham controls. Neuronal mechanisms responsible for this dichotomy remain to be elucidated. The activation and/or inhibition of mTOR pathway by ketamine and rapamycin, respectively, did not anticipate effect upon behavior of animals. It suggests that ketamine antidepressant effect is not exclusively mediated by this signaling pathway and/or that dysfunction of this pathway is not the only causal factor for manifestation of depressive symptoms. The present study has shown the depression as a complex phenomenon affecting a whole variety of behaviors and neural structures, indicating that there probably is not one single mechanism and cause accountable for the onset and progression of illness.