Introduction

Glutamate has been implicated in the pathogenesis of depressive disorders (Sanacora et al. 2008; Skolnick 1999). A post-mortem study found increased levels of glutamate in the frontal cortex of patients with major depression (Hashimoto et al. 2007). Evidence indicates that the N-methyl-D-aspartate receptor (NMDAR) complex is particularly involved in the pathophysiology of depression, since preclinical and clinical studies have indicated that compounds that reduce transmission at NMDAR exhibit antidepressant properties (Berman et al. 2000; Cunha et al. 2008; Zomkowski et al. 2010, 2012). Changes in NMDAR have been also demonstrated in the frontal cortex of suicide victims (Nowak et al. 1995). Noteworthy, the NMDAR antagonist ketamine produces a rapid and sustained antidepressant effect in animal models and in patients that suffer from treatment-resistant major depression (Berman et al. 2000; Li et al. 2011), in consonance with the proposition that NMDAR antagonists may be novel strategies for the treatment of depression (Skolnick 1999). The faster antidepressant-like effect of ketamine has been attributed to NMDA and AMPA receptor modulation (Duman et al. 2012; Zhou et al. 2014).

In response to activation of glutamate receptors (NMDA, AMPA or Kainate) or voltage-gated Ca2+ channels, nitric oxide (NO) is synthesized from l-arginine by nitric oxide synthase (NOS) (Contestabile 2000; Esplugues 2002; Yamamoto et al. 2004). Several studies indicate that l-arginine/NO pathway is also involved in the pathophysiology of depression. Plasma nitrate concentrations and expression of neuronal nitric oxide synthase in the hippocampus were reported to be significantly higher in depressed patients, suggesting that NO production is increased in depression (Oliveira et al. 2008; Suzuki et al. 2001). In line with this, several studies have shown that NOS inhibitors exert antidepressant-like effects in animal models (da Silva et al. 2000; Harkin et al. 1999, 2003; Heiberg et al. 2002; Joca and Guimarães 2006; Volke et al. 2003). However, an antidepressant-like action was also observed with the administration of the substrate for NOS, l-arginine (da Silva et al. 2000; Inan et al. 2004; Spiacci et al. 2008), suggesting that NO may have a dual role in the modulation of depression.

A growing number of reports have provided evidence for the importance of the creatine kinase/phosphocreatine system in the pathophysiology of depression: (1) brain phosphocreatine, detected by phosphorus-31 magnetic resonance spectroscopy, was shown to be decreased in severely depressed patients (Kato et al. 1992); (2) an inverse correlation between Hamilton Depression Rating Scale scores and white matter creatine levels was shown (Dager et al. 2004); (3) single prolonged stress and forced swimming stress decreased creatine concentrations in the rat prefrontal cortex (Herring et al. 2008; Kim et al. 2010; Knox et al. 2010); (4) learned helplessness, a well-validated animal model of depression, decreased the expression of hippocampal creatine transporter (Lugenbiel et al. 2010); (5) the acute administration of the fast-acting antidepressant ketamine increased creatine kinase activity in rats (Assis et al. 2009); (6) clinical trials show that creatine augmentation in antidepressant treatment-resistant patients may be a promising therapeutic approach that exhibits more rapid and efficacious responses (Kondo et al. 2011; Lyoo et al. 2012); (7) creatine administration in mice produced an anti-immobility effect in the tail suspension test (TST) and forced swimming test (FST), two widely used tests for screening antidepressants (Allen et al. 2010, Cunha et al. 2012, 2013a, b).

The mechanisms underlying the neuroprotective and antidepressant actions may be different, although antidepressant agents commonly exhibit neuroprotective properties. The link between these two effects may be the inhibition of NMDAR and the modulation of NO production.

Taking into account the involvement of NMDAR and l-arginine-NO pathway in the pathogenesis of depression and the importance of these molecular targets for the efficacy of antidepressants (Dhir and Kulkarni 2007; Krass et al. 2011; Zomkowski et al. 2010, 2012), this study investigated whether the anti-immobility effect of creatine in the TST is mediated by a modulation of NMDAR and NO pathway and also assessed the ability of creatine to enhance cellular viability and to counteract glutamate excitotoxicity.

Materials and methods

Animals

Female Swiss mice (30–40 g) and female Wistar rats (210–230 g) were housed in groups of fourteen (mice) or five (rats) per plastic cage under controlled conditions of light (from 07:00 to 19:00 h) and temperature (21 ± 1 °C). Animals were allowed free access to standard laboratory food and tap water, and to adapt to the laboratory environment for at least 1-week before the behavioral studies. For behavioral analysis, each experimental group consisted of 8–11 animals. For the biochemical measurements, each experimental group consisted of 4–6 samples. Animals were randomly distributed into the experimental groups. All manipulations were carried out between 14:00 and 17:00 h. All procedures in this study were performed in accordance with the National Institute of Health Guide for the Care and Use of Laboratory Animals and approved by the Ethics Committee of the Institution (CEUA/UFSC). All efforts were made to minimize animals suffering and to reduce the number of animals used in the experiments.

Drugs

The following drugs were used: 6,7-dinitroquinoxaline-2,3-dione (DNQX, 2.5 µg/mouse, icv), 7-nitroindazole (50 mg/kg, ip), arcaine (1 mg/kg, ip), creatine monohydrate (0.01 or 10 mg/kg, po), d-serine (30 µg/mouse, icv), fluoxetine (10 mg/kg, po), ketamine (0.1–1 mg/kg, ip), l-arginine (750 mg/kg, ip), NG-nitro-l-arginine methyl ester (L-NAME, 175 mg/kg, ip), MK-801 (0.001 mg/kg, po), N-methyl-D-aspartate (NMDA, 0.1 pmol/mouse, icv), S-nitroso-N-acetyl-penicillamine (SNAP, 25 µg/mouse, icv), (Sigma Chemical Co, St Louis, MO, USA). Table 1 summarizes the drugs used in the experimental protocols.

Table 1 Pharmacological agents used in the experimental protocols
Full size table

All drugs were dissolved in saline, except 7-nitroindazole that was dissolved in saline with 1 % Tween 80. Drugs were administered by intraperitoneal (ip) route in a constant volume of 10 ml/kg body weight, except NMDA and SNAP which were administered by intracerebroventricular (icv) route and creatine and MK-801 that was administered per os (po) by gavage.

Icv administration

The icv injections were performed by a ‘‘free hand’’ method performed and validated in our research group (Moretti et al. 2012; Kaster et al. 2012). Icv administration was performed using a microsyringe (25 μl, Hamilton) connected to a 26-gauge stainless-steel needle that was inserted perpendicularly 2 mm deep through the skull according to the procedure described by Laursen and Belknap (1986). Briefly, mice were lightly anesthetized with ether (i.e. just that necessary for loss of the postural reflex) and then gently restrained by hand for icv injections. The sterilization of the injection site was carried out using gauze embedded in 70 % ethanol. The needle was inserted unilaterally 1 mm into the midline point equidistant from each eye, at an equal distance between the eyes and the ears and perpendicular to the plane of the skull. A volume of 5 μl of sterile saline (vehicle) or drugs (dissolved in sterile saline) was injected directly into the left lateral ventricle, at the following coordinates from bregma taken from the atlas of Franklin and Paxinos (1997): anterioposterior (AP) = −0.1 mm; mediolateral (ML) = 1 mm; and dorsoventral (DV) = −3 mm. Mice exhibited normal behavior within 1 min after injection. After completion of the experiments, all animals were decapitated and their brains were examined freshly (Moretti et al. 2012). Mice presenting misplacement of the injection site or any sign of cerebral hemorrhage were excluded from the statistical analysis (overall less than 5 % of the total animals used).

Experimental design

Involvement of NMDA receptors in the anti-immobility effect of creatine treatment in mice submitted to TST

Taking into account that the minimum effective dose of creatine that produced a reduction in the immobility time in the TST (at the significance level of p < 0.01) was 10 mg/kg, po and the sub-effective dose was 0.01 mg/kg, po (Cunha et al. 2012), these doses were used in the present study. Notably, these doses of creatine did not exert psychostimulant effects in the open-field test (Cunha et al. 2012). In the experiments designed to verify the involvement of the NMDAR modulation in the mechanism underlying the anti-immobility effect of creatine in the TST, mice were treated with creatine (10 mg/kg, po) or vehicle. After 45 min, NMDA (0.1 pmol/mouse, icv), d-serine (30 µg/mouse, icv, agonist of the glycine site on the NMDAR) or vehicle was administered. Fifteen minutes later, mice were submitted to the TST or open-field test. The experimental protocols for NMDA and d-serine administrations were selected on the basis of previous results from our laboratory (Bettio et al. 2012; Ludka et al. 2013; Kaster et al. 2012; Moretti et al. 2011; Rieger et al. 2014; Zeni et al. 2011; Zomkowski et al. 2010, 2012) and others (Wlaz et al. 2011).

To investigate the possible involvement of the polyamine site at NMDAR in the anti-immobility effect of creatine in the TST, mice were treated with arcaine (1 mg/kg, ip, an antagonist at the polyamine site of NMDAR). After 30 min, they received either creatine (10 mg/kg) or vehicle. Mice were submitted to behavioral testing 30 min later. The dose and protocol of administration of arcaine used were selected based on a study by Zomkowski et al. (2006).

To investigate the possible synergistic effect in the TST obtained by the combined administration of sub-effective doses of creatine and MK-801, mice were treated with a sub-effective dose of MK-801 (0.001 mg/kg, po, NMDAR antagonist) and, immediately after, a sub-effective dose of either creatine (0.01 mg/kg, po) or vehicle was administered. Mice were submitted to behavioral testing 60 min later. The dose of MK-801 was chosen based on previous studies (Bettio et al. 2012; Kaster et al. 2012; Ludka et al. 2013; Moretti et al. 2012; Zeni et al. 2011; Zomkowski et al. 2012). In another set of experiments, mice were administered with creatine (0.01 mg/kg, po) or vehicle. After 30 min, ketamine (0.1 mg/kg, ip, NMDAR antagonist) or vehicle was administered. Mice were submitted to behavioral testing 30 min later. The dose and time point of ketamine administration were selected on the basis of previous results from our laboratory (Mantovani et al. 2003; Bettio et al. 2012; Ludka et al. 2013) and others (Cruz et al. 2009; Iijima et al. 2012).

Involvement of AMPA receptors in the anti-immobility effect of creatine administration in mouse TST

The influence of AMPA receptors in the anti-immobility effect of creatine in the TST was investigated. Forty five minutes after creatine (10 mg/kg, po), fluoxetine (10 mg/kg, po, conventional antidepressant), or vehicle administrations, mice received DNQX (2.5 µg/mouse, icv, AMPA receptor antagonist). After 15 min, mice were submitted to behavioral testing. In another set of experiments, 15 min after the administration of ketamine (1 mg/kg, ip, positive control), mice received DNQX (2.5 µg/mouse, icv) and a further 15 min was elapsed to the behavioral testing. The doses of the fluoxetine and DNQX used were selected based on previous studies (Cunha et al. 2008; Sharp et al. 1995).

Involvement of l-arginine/nitric oxide pathway in anti-immobility effect of creatine treatment in the TST

To measure hippocampal and cerebrocortical NO x levels, creatine (10 mg/kg, po, an effective dose in the TST) or vehicle was administered to mice 60 min before decapitation. Ketamine (1 mg/kg, ip, effective dose in the TST), administered 30 min before decapitation, was used as a positive control. To investigate the possible involvement of the l-arginine-NO pathway in the anti-immobility effect of creatine in TST, 30 min after creatine administration (10 mg/kg, po), mice received l-arginine (750 mg/kg, ip, a precursor of NO), L-NAME (175 mg/kg, ip, non-selective NOS inhibitor), 7-nitroindazole (50 mg/kg, ip, neuronal NOS inhibitor), or vehicle. Mice were submitted to behavioral testing 30 min later. In another set of experiments, 45 min after the administration of creatine (10 mg/kg, po), mice were injected with SNAP or vehicle, and a further 15 min elapsed before animals was submitted to behavioral testing. In an independent experiment, mice were administered with the nitrergic modulators and creatine using the same administration protocols described above and, thereafter, the animals were decapitated for measurement of NO x levels. The doses and time points for l-arginine (Bettio et al. 2012; Ludka et al. 2013; Zeni et al. 2011; Zomkowski et al. 2012), SNAP (Brocardo et al. 2008; Kaster et al. 2005; Zomkowski et al. 2012), 7-nitroindazole and L-NAME (Bettio et al. 2012; da Silva et al. 2000; Harkin et al. 1999) administrations were selected on the basis of previous studies.

Effect of creatine treatment on hippocampal and cerebrocortical cellular viability

To evaluate the effect of creatine on ex vivo hippocampal and cerebrocortical cellular viability, creatine (10 mg/kg, po) or vehicle was administered to mice and 60 min later they were killed. After that, forebrains were rapidly removed from the skull and placed into ice-cold Krebs bicarbonate buffer (KRB), containing (in mM): NaCl 122, KCl 3, CaCl2 1.3, NaHCO3 25, MgSO4 1.2, KH2PO4 0.4, d-glucose 10, pre-bubbled with 95 % O2/5 % CO2 up to pH 7.4. The hippocampi or cerebral cortices were quickly dissected and sectioned into transverse slices of 400 µm thick using a tissue chopper (McIlwain Tissue Chopper, The Micle Laboratory Engineering Company). After the preincubation time (30 min in KRB), the medium was withdrawn and replaced by a nutritive culture medium composed of 50 % KRB, 50 % Dulbecco’s modified Eagle’s medium (DMEM, Gibco), 20 mM HEPES and 100 μg/ml gentamicin, at 37 °C in a CO2 atmosphere (Molz et al. 2008) and slices were maintained for an additional 4 h period to evaluate cellular viability.

Effect of creatine treatment on glutamate-induced neurotoxicity in hippocampal slices

To assess the ability of creatine treatment to counteract the excitotoxicity elicited by an in vitro challenge with glutamate, creatine (10 mg/kg, po) or vehicle was administered to mice and 60 min later the hippocampi were sectioned into 400 µm slices. After the stabilization period (30 min), slices were incubated with glutamate (Sigma, St. Louis, MO; 10 mM, in KRB buffer) or KRB buffer for 60 min. After this period, the medium was withdrawn and replaced by a nutritive culture medium, described above, at 37 °C in a CO2 atmosphere and slices were maintained for an additional 4-h period to evaluate cellular viability.

Effect of in vitro creatine incubation on hippocampal slices of mice or rats challenged with glutamate

To assess the ability of creatine (in vitro) to counteract glutamate toxicity, hippocampi from mice or rats were dissected and sectioned into transverse slices of 400 µm thick. After the stabilization period (30 min), the slices were incubated with glutamate (10 mM) or KRB buffer for 60 min. Creatine at concentrations 2.5, 5 or 10 mM was added to the slices in the presence or absence of glutamate. After this period, the medium was withdrawn and replaced by the nutritive culture medium described above and slices were maintained for an additional 4-h period to evaluate cellular viability.

Behavioral tests

Tail suspension test (TST)

The total duration of immobility induced by tail suspension was measured according to the method described by Steru et al. (1985). Briefly, mice both acoustically and visually isolated were suspended 50 cm above the floor by adhesive tape placed approximately 1 cm from the tip of the tail. Immobility time was recorded in a 6-min period by an observer blind to the drug treatment. Effective antidepressant treatments decrease immobility time of mice (Steru et al. 1985).

Open-field test

To assess the possible effects of creatine and/or the pharmacologic agents on locomotor activity, mice were subjected to the open-field paradigm (Cunha et al. 2008). Mice were individually placed in a wooden box (40 × 60 × 50 cm) with the floor divided into 12 equal rectangles (13.3 × 15 cm). The number of rectangles crossed by the animal with its four paws (crossing) was registered in a 6-min session and was considered as indicative of locomotor activity. Although the evaluation in the open-field paradigm is simple, non-automated method for assessing locomotor activity, it is able to detect either hyperlocomotion (Machado et al. 2012; Rodrigues et al. 2005) or hypolocomotion (Cunha et al. 2008; Rosa et al. 2003).

Biochemical measurements

NO x analysis

NO is unstable and quickly oxidized to nitrate and nitrite after production. Then, NO metabolites were determined using the NO x analysis, a modified method described by Hevel and Marletta (1994). Briefly, homogenates from cerebral cortex or hippocampus (pool of three hippocampi) were mixed with 25 % trichloroacetic acid and centrifuged at 1,800g for 10 min. The supernatant was immediately neutralized with 2 M potassium bicarbonate. Nitrate was measured as nitrite after enzymatic conversion by nitrate reductase by a colorimetric assay based on the Griess color reaction (Ludka et al. 2013; Zomkowski et al. 2012). A standard curve was performed using sodium nitrite (0–80 μM). Results were expressed as percentage of control (100 %).

Evaluation of cellular viability

In hippocampal or cerebral cortex slices, cellular viability was determined through the ability of cells to reduce the formazan salt 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) (Mosmann 1983). Hippocampal or cortical slices were incubated with MTT (0.5 mg/ml) in KRB for 30 min at 37 °C. The tetrazolium ring of MTT can be cleaved by active dehydrogenases to produce a precipitated formazan. The formazan produced was solubilized by adding 200 µl dimethylsulfoxide (DMSO), resulting in a colored compound whose optical density was measured in an ELISA reader (540 nm).

Protein measurement

Protein content was evaluated according to the Lowry method (Lowry et al. 1951) using bovine serum albumin (Sigma Chemical Co.) as standard.

Statistical analysis

Kolmogorov–Smirnov test was used to evaluate the normality assumption of behavioral and biochemical data. All variables in the present study showed a normal distribution. Comparisons between experimental and control groups were performed by Student’s t test (effect of creatine or ketamine on hippocampal and cerebrocortical NO x levels and effect of creatine on cellular viability of hippocampal and cerebrocortical slices) or two-way ANOVA (behavioral analysis). Post-hoc Tukey’s HSD test was used when the F values of ANOVA were significant. A value of p < 0.05 was considered to be significant.

Results

Effect of NMDA, d-serine or arcaine on anti-immobility effect of creatine in TST

Figure 1d shows that the treatment of mice with NMDA (0.1 pmol/mouse, icv) was able to abolish the anti-immobility effect of creatine (10 mg/kg, po) in the TST in mice. Two-way ANOVA revealed significant differences for creatine treatment [F(1,30) = 5.82, p < 0.01], NMDA treatment [F(1,30) = 8.38, p < 0.01] and creatine × NMDA interaction [F(1,30) = 15.85, p < 0.01]. NMDA administration alone or in combination with creatine did not significantly affect ambulation in the open-field test (p > 0.05, data not shown).

Fig. 1
figure 1

Activation of NMDA receptors abolishes the anti-immobility effect of creatine. Timeline of NMDA (a), d-serine (b) and arcaine (c) administrations to reverse the anti-immobility effect of creatine. Effect of treatments of mice with NMDA (0.1 pmol/mouse, icv), d-serine (30 µg/mouse, icv) or arcaine (1 mg/kg, ip) on the anti-immobility effect of creatine (10 mg/kg, po) in the TST (d, e and f, respectively). Each column represents the mean + SEM. *p < 0.05, **p < 0.01 compared with the vehicle-treated control; # p < 0.05, ## p < 0.01, as compared to same group treated with vehicle (creatine/vehicle)

Full size image

Similarly Fig. 1e shows that the treatment of mice with d-serine (30 µg/mouse, icv) abolished the anti-immobility effect of creatine (10 mg/kg, po) in the TST. Two-way ANOVA revealed significant differences for the creatine treatment [F(1,33) = 11.92, p < 0.01], d-serine [F(1,33) = 5.19, p < 0.05] and creatine × d-serine interaction [F(1,33) = 4.65, p < 0.05]. d-serine administration alone or in combination with creatine did not affect the ambulation in the open-field test (p > 0.05, data not shown).

Arcaine administration (1 mg/kg, ip) reversed the anti-immobility effect of creatine in the TST (Fig. 1f). The two-way ANOVA revealed significant differences for the creatine treatment [F(1,26) = 6.27, p < 0.05] and arcaine × creatine interaction [F(1,33) = 4.69, p < 0.05], but not a main effect of arcaine treatment [F(1,33) = 1.63, p = 0.21]. Arcaine administration alone or in combination with creatine did not affect ambulation in the open-field test (p > 0.05, data not shown).

Effect of combined treatment with sub-effective doses of the NMDA antagonists MK-801 or ketamine and creatine in the TST.

Figure 2c shows that combined administration of sub-effective doses of MK-801 (0.001 mg/kg, po) and creatine (0.01 mg/kg, po) caused an anti-immobility effect in the TST, but the administration of creatine or MK-801 alone did not reduce immobility time in the TST. Two-way ANOVA revealed significant differences for the creatine treatment [F(1,32) = 4.82, p < 0.01], MK-801 treatment [F(1,32) = 8.68, p < 0.01] and creatine × MK-801 interaction [F(1,32) = 6.07, p < 0.01]. Creatine administration alone or in combination with MK-801 did not affect ambulation in the open-field test (p > 0.05, data not shown).

Fig. 2
figure 2

A sub-effective dose of creatine produces an anti-immobility effect in the TST when combined with sub-effective doses of NMDA receptor antagonists. Timeline of experimental protocol of mice treated with creatine in combination with MK-801 (a) or ketamine (b). Effects of sub-effective doses of creatine (0.01 mg/kg, po) in combination with MK-801 (0.001 mg/kg, po, an NMDAR antagonist) or ketamine (0.1 mg/kg, ip, an NMDAR antagonist) on the immobility time in the TST (c and d, respectively). Each column represents the mean + SEM. **p < 0.01 compared with the vehicle-treated control

Full size image

Combined administration of sub-effective doses of ketamine (0.1 mg/kg, ip) and creatine (0.01 mg/kg, po) produced synergistic anti-immobility effect in the TST (Fig. 2d). Two-way ANOVA revealed significant differences for creatine treatment [F(1,33) = 18.40, p < 0.01] and creatine × ketamine interaction [F(1,33) = 4.81, p < 0.05], but not a main effect of ketamine treatment [F(1,33) = 1.56, p = 0.22]. Ketamine administration alone or in combination with creatine did not significantly alter the ambulation in the open-field test (p > 0.05, data not shown).

Effect of the DNQX on the anti-immobility effect of creatine in the TST

Results depicted in Fig. 3 show that the anti-immobility effect of creatine (10 mg/kg, po) or fluoxetine (10 mg/kg, po) was not prevented by pretreatment of mice with the AMPA receptor antagonist DNQX (2.5 µg/mouse, icv) in the TST (p > 0.05, Fig. 3d, e, respectively). The number of crossings in open-field test was not altered by DNQX, creatine or fluoxetine treatments (p > 0.05, data not shown). Conversely, administration of DNQX abolished the anti-immobility effect of ketamine in the TST (Fig. 3f). Two-way ANOVA revealed significant differences for ketamine treatment [F(1,24) = 4.70, p < 0.05] and ketamine × DNQX interaction [F(1,24) = 4.95, p < 0.05], but not a main effect of DNQX treatment [F(1,24) = 1.64, p = 0.21]. Ketamine administration alone or in combination with DNQX did not affect ambulation in the open-field test (p > 0.05, data not shown).

Fig. 3
figure 3

DNQX, an AMPA receptor antagonist, abolishes the anti-immobility effect of ketamine, but not of creatine or fluoxetine. Timeline of reversal protocol of the anti-immobility effect of creatine, fluoxetine or ketamine by DNQX (a, b and c, respectively). Effects of treatments of mice with DNQX (2.5 µg/mouse, icv) and creatine (10 mg/kg, po), fluoxetine (10 mg/kg, po) or ketamine (1 mg/kg, ip) on the immobility time in the TST (d, e and f, respectively). Each column represents the mean + SEM. *p < 0.05, **p < 0.01 compared with the vehicle-treated control; # p < 0.05, ## p < 0.01, as compared to same group treated with vehicle (creatine/vehicle)

Full size image

Effect of creatine treatment on NO x levels in hippocampus and cerebral cortex

NO x measurement was performed in hippocampus and cerebral cortex of mice treated with creatine at dose of 10 mg/kg (po) or vehicle 60 before decapitation (Fig. 4). Creatine at dose of 10 mg/kg significantly increased NO x levels in hippocampus and cerebral cortex of mice (46.3 and 43.6 %, respectively), as compared with control group [T Test: t(10) = −2.92; p < 0.05; t(14) = −3.63; p < 0.01, respectively]. Ketamine, used as a positive control, also increased NO x levels in the hippocampus and cerebral cortex [116.2 and 93.2 %, respectively, t test: t(9) = 2.30; p < 0.05 and t(21) = 2.30; p < 0.05, respectively]. Considering that the effect of creatine on NO x levels was similar in both brain structures, the next experiments that investigate the influence of NO modulators on the creatine-induced increase in NO x levels were performed in cerebral cortex due to the higher amount of tissue available to the assays.

Fig. 4
figure 4

Creatine, similar to ketamine, increases NO x levels in hippocampus and cerebral cortex. Effect of treatment with creatine (10 mg/kg, po) or ketamine (1 mg/kg, ip, an NMDAR antagonist) on the NOx levels in the hippocampus and cerebral cortex of mice. Each column represents the mean + SEM. *p < 0.05 compared with the vehicle-treated control. The NO x levels in the control group (100 %) was 13.20 µM nitrite/mg protein (hippocampus) and 22.3 µM nitrite/mg protein (cerebral cortex)

Full size image

Effect of l-arginine, SNAP or 7-nitroindazole on the anti-immobility effect of creatine in the TST, and on cerebrocortical NO x levels

Taking into account that NO modulation may affect either behavioral responses of creatine in the TST or its ability to enhance NO x levels, the next set of the experiments evaluated the influence of the NO enhancers l-arginine and SNAP and the NOS inhibitors L-NAME and 7-nitroindazole on the effect of creatine in the TST and on NO x levels in the cerebral cortex. Figure 5c shows that treatment of mice with l-arginine (750 mg/kg, ip) abolished the anti-immobility effect of creatine (10 mg/kg, po) in TST. Two-way ANOVA revealed significant differences for creatine treatment [F(1,23) = 23.48, p < 0.01], l-arginine treatment [F(1,23) = 5.17, p < 0.05] and creatine × l-arginine interaction [F(1,23) = 5.11, p < 0.05]. l-arginine administration alone or in combination with creatine did not affect the ambulation in the open-field (p > 0.05, data not shown). Additionally, Fig. 5e shows that the treatment of mice with l-arginine (750 mg/kg, ip), at dose that per se produced an enhancement on NO x levels, was able to potentiate the creatine-induced increase in NO x levels. Two-way ANOVA analysis revealed significant differences for the creatine treatment [F(1,24) = 43.02, p < 0.01] and l-arginine treatment [F(1,24) = 17.17, p < 0.01], but not creatine × l-arginine interaction [F(1,24) = 0.37, p = 0.55].

Fig. 5
figure 5

Influence of treatment with l-arginine or SNAP on the effect of creatine in the TST, open-field test and NO x levels. Timeline of reversal protocol of the anti-immobility effect of creatine by l-arginine (a) and SNAP (b). Effect of treatment of mice with l-arginine (750 mg/kg, ip, a precursor of NO) or SNAP (25 µg/mouse, icv, a NO donor) on the immobility time in the TST (c and d, respectively) and on NO x levels in the cerebral cortex (e and f, respectively). Each column represents the mean + SEM. **p < 0.01 compared with the vehicle-treated control; # p < 0.05, ## p < 0.01 as compared to group pretreated with creatine and treated with vehicle; $$ p < 0.01 as compared to group pretreated with vehicle and treated with l-arginine. The NO x level in the control group (100 %) was 7.64 µM nitrite/mg protein (g) and 7.82 µM nitrite/mg protein (h)

Full size image

Figure 5d shows that treatment of mice with the NO donor SNAP (25 µg/mouse, icv) was able to abolish the anti-immobility effect of creatine (10 mg/kg, po) in the TST. Two-way ANOVA revealed significant differences for creatine treatment [F(1,24) = 5.47, p < 0.05], SNAP treatment [F(1,24) = 7.60, p < 0.05] and creatine × SNAP interaction [F(1,24) = 5.81, p < 0.05]. SNAP administration alone or in combination with creatine did not affect ambulation in the open-field (p > 0.05, data not shown). Moreover, Fig. 5f shows that treatment of mice with SNAP, which per se produced an enhancement on NO x levels, enhanced the effect of creatine (10 mg/kg, po) on NO x levels. Two-way ANOVA revealed significant differences for SNAP treatment [F(1,26) = 13.90, p < 0.01] and creatine × SNAP interaction [F(1,26) = 7.47, p < 0.05], but not a main effect for creatine treatment [F(1,26) = 0.50, p = 0.49].

Administration of a non-selective NOS inhibitor, namely L-NAME (175 mg/kg, ip), prevented the anti-immobility effect of creatine in the TST (Fig. 6b). Two-way ANOVA revealed significant differences for creatine treatment [F(1,32) = 4.44, p < 0.05], L-NAME treatment [F(1,32) = 20.22, p < 0.01] and L-NAME × creatine interaction [F(1,32) = 18.24, p < 0.01]. L-NAME administration alone or in combination with creatine did not affect ambulation in the open-field (p > 0.05, data not shown). Figure 6d shows that treatment of mice with L-NAME abolished the increase in cerebrocortical NO x levels induced by creatine administration (10 mg/kg, po). Two-way ANOVA revealed significant differences for creatine treatment [F(1,28) = 16.86, p < 0.01] and L-NAME treatment [F(1,28) = 30.31, p < 0.01], but not for creatine × L-NAME interaction [F(1,28) = 0.15, p = 0.70].

Fig. 6
figure 6

Influence of NOS inhibitors on the effect of creatine in the TST, open-field test and NOx levels. Timeline of experimental protocol for drug administrations and behavioral tests (a). Effect of treatment of mice with 7-nitroindazole (50 mg/kg, ip, a neuronal nitric oxide synthase inhibitor) or L-NAME (175 mg/kg, ip, a nitric oxide synthase inhibitor) on the anti-immobility effect of creatine (10 mg/kg, po) in the TST (b and c, respectively) and on cerebrocortical NO x levels (d and e, respectively). Each column represents the mean + SEM. *p < 0.05, **p < 0.01 compared with the vehicle-treated control; # p < 0.05, ## p < 0.01 as compared to group pretreated with creatine and treated with vehicle. The NO x level in the control group (100 %) was 7.96 µM nitrite/mg protein (d) and 7.11 µM nitrite/mg protein (e)

Full size image

Administration of a nNOS inhibitor 7-nitroindazole (50 mg/kg, ip) prevented the anti-immobility effect of creatine in the TST (Fig. 6c). Two-way ANOVA revealed significant differences for creatine treatment [F(1,26) = 10.67, p < 0.01] and 7-nitroindazole × creatine interaction [F(1,26) = 5.62, p < 0.05], but not a main effect of 7-nitroindazole treatment [F(1,26) = 4.09, p = 0.05]. The administration of 7-nitroindazole alone or in combination with creatine did not affect ambulation in the open-field test (p > 0.05, data not shown). Moreover, Fig. 6e shows that treatment of mice with 7-nitroindazole abolished creatine-induced increase of NO x levels in cerebral cortex. Two-way ANOVA revealed significant differences for creatine treatment [F(1,28) = 11.74, p < 0.01] and creatine × 7-nitroindazole interaction [F(1,28) = 5.31, p < 0.05], but not for 7-nitroindazole treatment [F(1,28) = 0.44, p = 0.51].

Effect of creatine on cellular viability of hippocampal and cerebrocortical slices

Cellular viability analysis was performed in mice administered with creatine at dose of 10 mg/kg or vehicle 60 min before decapitation. Figure 7 shows increased cellular viability (19.6 %) in hippocampal slices obtained from mice treated with creatine as compared with slices obtained from control group (mice administered with vehicle), as revealed by t test: t(16) = −2.40; p < 0.05. Moreover, creatine administration significantly increased cellular viability (11.8 %) in the cerebral cortex as compared with the control group, as revealed by t test: t(15) = 3.84; p < 0.01 (Fig. 7). Considering that the effect of creatine was slightly higher in the hippocampus than in the cerebral cortex, the next experiments that investigated whether creatine could counteract glutamate toxicity were performed in hippocampal slices.

Fig. 7
figure 7

Creatine increases cellular viability in hippocampal and cerebrocortical slices of mice. Effect of the treatment of mice with creatine (10 mg/kg, po) on cellular viability of hippocampal and cerebrocortical slices. Each column represents the mean + SEM. *p < 0.05 as compared to control group treated with vehicle

Full size image

Effect of creatine on the glutamate-induced neurotoxicity in hippocampal slices

Hippocampal slices obtained from mice administered with creatine at dose of 10 mg/kg or vehicle 60 min before decapitation were submitted to glutamate challenge (10 mM, in vitro) for 1 h. Slices incubated with glutamate presented a reduced cellular viability when compared to control slices. Creatine administration in mice was not able to reverse the glutamate-induced decrease in hippocampal cellular viability (Fig. 8a). The two-way ANOVA revealed significant differences for creatine treatment [F(1,16) = 8.36, p < 0.05], glutamate incubation [F(1,16) = 50.76, p < 0.01] and creatine × glutamate interaction [F(1,16) = 4.67, p < 0.05].

Fig. 8
figure 8

In vitro, but not ex vivo, creatine protects against glutamate-induced decrease in hippocampal cellular viability. Effect of administration of mice with creatine (10 mg/kg, po) on glutamate-induced neurotoxicity in hippocampal slices of mice (a) and the creatine incubation (2.5–10 mM) on the cellular viability of mouse or rat hippocampal slices exposed to glutamate (10 mM) (b and c, respectively). Each column represents the mean + SEM. *p < 0.05 as compared to control group treated with vehicle. **p < 0.01, ***p < 0.001 as compared to control group treated with vehicle. # p < 0.05, ## p < 0.01 as compared to group treated with glutamate

Full size image

We also investigated the protective effect of in vitro creatine incubation (2.5, 5 and 10 mM) against neurotoxicity induced by glutamate (10 mM) in mice and rats hippocampal slices (Fig. 8b, c, respectively). In vitro incubation of mice hippocampal slices with 5–10 mM creatine was effective to abolish glutamate-induced decrease in cellular viability (Fig. 8b), as revealed by two-way ANOVA (creatine incubation [F(3,34) = 4.09, p < 0.05], glutamate incubation [F(1,34) = 113.20, p < 0.01] and creatine × glutamate interaction [F(3,34) = 6.55, p < 0.01]. Reinforcing the neuroprotective properties of creatine, the results presented in Fig. 8c show that the in vitro incubation of rats hippocampal slices with 2.5, 5 and 10 mM creatine was also effective to abolish glutamate toxicity, as revealed by one-way ANOVA [F(1,4) = 6.06; P < 0.01].

Discussion

This study shows that oral administration of creatine, a substrate of creatine kinase and a precursor of phosphocreatine, has an anti-immobility effect in TST, not influenced by any unspecific locomotor effect, by a mechanism that involves NMDAR and NO modulation.

The NMDAR antagonist ketamine has been shown to elicit a rapid antidepressant action in preclinical and clinical reports (Autry et al. 2011; Berman et al. 2000; Li et al. 2011). Interestingly, a study by Assis et al. (2009) showed that the acute administration of ketamine, besides producing an anti-immobility effect, increased the creatine kinase activity in striatum, cerebral cortex and cerebellum. Memantine, another NMDAR antagonist with anti-immobility effect, also increased creatine kinase activity in prefrontal cortex and hippocampus of rats (Reus et al. 2012). In vitro and in vivo studies have reported that ketamine and MK-801 facilitate recovery of phosphocreatine levels after ischemic or anoxic brain injury, clinical conditions associated with glutamatergic dysfunction (Bielenberg et al. 1987; Haraldseth et al. 1990; Raley and Lipton 1990; Spandou et al. 1999). Interestingly, creatine has been proposed to modulate the glutamatergic system (Andreassen et al. 2001; Bender et al. 2005; Almeida et al. 2006; Royes et al. 2008), but the involvement of the glutamatergic system in the anti-immobility effect of this compound was not previously explored. Therefore, this study investigated the involvement of the NMDAR in the anti-immobility effect of creatine in the TST.

We show that the behavioral response in the TST induced by creatine treatment involves a direct or indirect inhibition of NMDAR, since the treatment of mice with either NMDA or d-serine, a potent co-agonist of synaptic NMDARs (Papouin et al. 2012), caused complete reversal of the anti-immobility effect elicited by creatine. Similar results have been reported for other compounds that exert anti-immobility effects by a mechanism dependent on NMDAR inhibition (Brocardo et al. 2008; Kaster et al. 2012; Ludka et al. 2013; Moretti et al. 2011; Poleszak et al. 2008; Zeni et al. 2011; Zomkowski et al. 2010). Corroborating the hypothesis that creatine modulates the NMDAR, we observed that arcaine, an antagonist of the polyamine site of the NMDAR, abolished the anti-immobility action of creatine. This result suggests that the anti-immobility effect of creatine could be due to its interaction with this modulatory site at NMDAR. In line with this, arcaine abolished the antidepressant-like effect of putrescine, a polyamine with anti-immobility effects in the TST and FST (Zomkowski et al. 2006). Moreover, it has been suggested that creatine could act as an agonist of the polyamine site at the NMDAR since endogenous polyamine spermidine intensifies, while polyamine antagonist, such as arcaine, blocks enhanced spatial learning induced by creatine (Oliveira et al. 2008).

Another finding that reinforces the notion that an NMDAR modulation of glutamatergic system plays a role in the anti-immobility effect of creatine is the synergistic anti-immobility effect observed when sub-effective doses of creatine and MK-801 or ketamine were administered to mice. Accordingly, it has been reported that MK-801 or ketamine produces similar effects when associated with several compounds that display antidepressant properties (Ludka et al. 2013; Moretti et al. 2011; Zeni et al. 2011; Zomkowski et al. 2012). In line with this, a combination of creatine and MK-801 provided protection in an experimental model of anoxia in hippocampal slices (Carter et al. 1995). Considering that dopamine D1 receptor activation is involved in the anti-immobility effect of creatine (Cunha et al. 2012) and that the dopamine D1 receptor agonist SKF38393 could enhance the antidepressant-like effect of non-competitive NMDAR antagonist MK-801 (Yuan et al. 2011), it is possible that the interaction of NMDAR and dopamine D1 receptors could play a role in the anti-immobility effect of creatine.

The antidepressant response induced by ketamine has been shown to be associated with stimulation of the postsynaptic alpha-amino-3-hydroxy-5-methyl-4-isoxazole propionic acid (AMPA) receptors and BDNF release (Krystal et al. 2013; Moghaddam et al. 1997). We hypothesized that creatine could share with ketamine a similar mechanism of action, activating AMPA receptors in glutamatergic neurons. Interestingly, the AMPA receptor antagonist DNQX abolished the anti-immobility effect of ketamine in the TST, in agreement with literature data that reported a similar effect using NBQX (Koike et al. 2011), but did not affect the anti-immobility effect of creatine or fluoxetine. These results suggest that creatine and fluoxetine produce an anti-immobility effect in the TST by a mechanism independent of AMPA receptor activation, as opposed to ketamine.

Once NMDAR is activated, intracellular calcium levels increase, leading to NO synthesis by nNOS (Esplugues 2002). High concentrations of NO lead to neuronal cell death, whereas lower concentrations are neuroprotective by controlling proliferation and differentiation of mouse neuronal progenitor cells (Baranano and Snyder 2001; Calabrese et al. 2007; Cheng et al. 2003). A dual effect of NO was also suggested regarding mood regulation, since both the NO precursor l-arginine and NOS inhibitors, depending on the doses, may elicit antidepressant-like effects (da Silva et al. 2000; Inan et al. 2004). It is likely that NO exerts beneficial effects on mood, when present at low levels. Interestingly, in the present study, creatine at 10 mg/kg increased the NO x levels in the hippocampus and cerebral cortex of mice. Of note, the administration of ketamine, at a dose that produces an anti-immobility effect in the TST (Ludka et al. 2013), caused a similar increase in the hippocampal and cerebrocortical NO x levels. Therefore, in our study a slight increase in NO levels is associated with antidepressant-like responses. Ketamine also increased NO concentrations in the hippocampus and striatum of rats (Wu et al. 2000) and stimulated the l-arginine/NO pathway via neuronal NOS (Romero et al. 2011). Regarding the modulation of NO, creatine was shown to increase the number of NOS-immunoreactive striatal neurons (Ducray et al. 2006), but was able to abolish the increase in NO levels induced by glutamate in neuronal/glial cells (Juravleva et al. 2003). Future investigations are necessary to understand the detailed mechanisms underlying the modulation of NO levels by creatine.

The increase in NO levels, in turn, may regulate the monoaminergic tonus in the CNS. The experimental treatment with l-arginine or NO donors was reported to induce an augmentation of dopamine release, suggesting that endogenous NO stimulates dopamine activity (Lorrain and Hull 1993). In addition, studies performed in hippocampal slices demonstrate that NO donors increase norepinephrine release (Satoh et al. 1996). Evidence also indicates that NO signaling plays an important role in the mechanism of norepinephrine-induced neuroprotection in both in vivo and in vitro models (Chen and Russo-Neustadt 2007; Lonart et al. 1992). In addition, the NO donor SNAP was reported to increase serotonin levels in the hippocampus, as opposed to the decrease in serotonin induced by NMDA (Segieth et al. 2001). Therefore, it is possible that creatine-induced increase in NO levels with consequent monoaminergic modulation may be associated with its anti-immobility effect in the TST. This hypothesis is consistent with the fact that the behavioral effects in the TST induced by creatine are dependent on the activation of dopamine D1 and D2 receptors, α1-adrenoceptors and 5-HT1A receptors (Cunha et al. 2012, 2013a, b).

Considering that the modulation on NO levels has been reported to be involved in the pathophysiology of depression (da Silva et al. 2000; Inan et al. 2004; Suzuki et al. 2001), we investigated the participation of the l-arginine/NO pathway in the anti-immobility effect of creatine. Our results indicate that NO modulation influences the anti-immobility effect of creatine in the TST, because the treatment of mice with l-arginine or SNAP completely abolished its anti-immobility effect, without altering locomotor activity of mice. These results are in line with previous studies which showed that the anti-immobility effects of several compounds were prevented by l-arginine or SNAP (Brocardo et al. 2008; Kaster et al. 2005; Ludka et al. 2013; Moretti et al. 2011; Zomkowski et al. 2010). A study by Krass et al. (2011) also reported that pretreatment with l-arginine abolished the anti-immobility effect of imipramine, venlafaxine and bupropion in the FST. Noteworthy, in the present study, either l-arginine or SNAP administration potentiated the creatine-induced increase in NO x levels. Therefore, we postulate that a slight increase in NO is related with an anti-immobility effect in the TST, as opposed to a higher increase in NO levels. This is in line with the hypothesis that NO exerts a dual role in the modulation of depression, since the administration of l-arginine significantly reduced the duration of immobility in the FST at low doses (250–500 mg/kg, ip), but not at higher doses (750–1,000 mg/kg) (da Silva et al. 2000). The reduction of immobility time induced by l-arginine was reversed by the NOS inhibitor L-NNA. Conversely, L-NNA, depending on its dose is able to cause an antidepressant-like effect, which can be reversed by l-arginine, administered at a dose that produces no effect in the FST (da Silva et al. 2000). Interestingly, exogenous administration of l-arginine (250–500 mg/kg), at same doses that produce anti-immobility effect, increased neuronal NO signal (Heinzen and Pollack 2002) and DETA/NONOate, a NO donor, produces anti-immobility effect associated with enhancement of NO x levels and hippocampal neurogenesis (Hua et al. 2008). Accordingly, a low dose of L-NAME (100 mg/kg, ip) decreased the immobility time in the FST, whereas high doses (175–300 mg/kg, ip) did not produce any anti-immobility effect in the FST (Harkin et al. 1999).

To further investigate the role of NO in the anti-immobility effect of creatine, in another set of experiments, we evaluated the influence of the administration of the non-selective NOS inhibitor L-NAME and the potent and selective neuronal NOS inhibitor 7-nitroindazole on the effect of creatine in the TST. The ability of these NOS inhibitors to abolish either the anti-immobility effect of the creatine or the creatine-induced increase in NO x levels in cerebral cortex reinforces the hypothesis that the slight creatine-induced increase in NO levels may afford an anti-immobility effect. Interestingly, a study reported that creatine supplementation was able to prevent the reduction on ATP and creatine levels caused by L-NAME administration in rats, suggesting a relationship between NO modulation and creatine effects (Constantin-Teodosiu et al. 1995).

In summary, we provide evidence for a neuromodulatory effect of creatine on NMDAR and on NO levels, although the mechanisms by which creatine interacts with NMDAR require further studies. In line with our results, creatine (10 mM) has been shown to decrease the excitability of hippocampal slices of rats submitted to control or anoxia condition (Parodi et al. 2003) and to produce a direct inhibitory action on the hippocampal NMDAR-mediated calcium response (Genius et al. 2012). Notably, creatine binds to the central benzodiazepine receptor (Kawasaki et al. 2001), an effect that may be associated with NMDAR inhibition. Another possibility that may account for the observed results is that NO could inhibit NMDAR function through a direct action on the receptor-channel complex (Lei et al. 1992; Manzoni et al. 1992), in line with the finding that creatine increased the hippocampal and cerebrocortical NO x levels. Nitrosylation of the NR1 and NR2 subunits of the NMDAR results in functional downregulation of the NMDAR activity, therefore, protecting neurons from excitotoxic insults caused by excessive receptor activation (Choi and Lipton 2000; Lei et al. 1992; Lipton et al. 1998).

Considering that antidepressant and neuroprotective effects may be correlated (Young, 2002; Zeni et al. 2011), in another set of experiments we investigated the effects of the administration of creatine at a dose that produced anti-immobility effect in the TST (10 mg/kg) on the ex vivo cellular viability of hippocampal and cerebrocortical slices of mice, as compared to the slices obtained from mice not treated with creatine. Creatine (10 mg/kg, po) increased cellular viability of hippocampal and cerebrocortical slices, suggesting that the administration of this compound exerts a beneficial effect on the signaling mechanisms that maintain cell survival. In line with our results, creatine increased tyrosine hydroxylase positive cells in ventral mesencephalic organotypic tissue culture (Andres et al. 2005). Interestingly, other ex vivo treatments have been reported to cause similar effects, including ketamine that improved the quality (cellular viability) of neuronal brain slice preparations of neonatal mouse (de Oliveira et al. 2010). Accordingly, NO, which was increased by either creatine or ketamine administration, is required for neuronal survival (Contestabile and Ciani 2004).

Interestingly, the glutamate-induced reduction on cellular viability of hippocampal slices of mice was not prevented by ex vivo creatine treatment (10 mg/kg, po), suggesting that the anti-immobility effect of creatine is not associated with a protective effect against hippocampal glutamate toxicity. However, creatine incubated in vitro (0.1–10 mM) was reported to protect against glutamate-induced toxicity in cell cultures (Brewer and Wallimann 2000; Genius et al. 2012; Juravleva et al. 2003, 2005). Considering that hippocampal slices offer certain advantages over cell cultures, as the pattern of synaptic connections within the slice is minimally altered and neuron–astrocyte–microglia interactions are preserved (Somjen et al. 1987), in a next set of experiments the in vitro effect of creatine incubation in hippocampal slices was evaluated. To this end, creatine was incubated at relatively high concentrations in rat and mouse hippocampal slices challenged with glutamate. We demonstrated that high creatine concentrations (2.5–10 mM) incubated either in rat or mouse hippocampal slices prevented the toxicity induced by glutamate (10 mM). Opening NMDAR could reduce the membrane potential and lead to a reduced transport rate of creatine into the cytosol. Since creatine transporter works at half maximal rate under physiological conditions (Guerrero-Ontiveros and Wallimann 1998), the reduced transportation rate of creatine may be compensated by an elevated concentration of creatine in the incubation medium.

In conclusion, the present study indicates that modulation of NMDAR and NO levels plays a significant role in the anti-immobility effect of creatine, and reinforces the notion that these are pivotal targets for antidepressant action. Furthermore, the results presented herein show that creatine may also elicit a protective effect against glutamate-induced reduction on hippocampal cellular viability when present at high concentrations in the incubation medium, a result that warrants future studies for the investigation of creatine to counteract glutamatergic excitotoxicity. Under basal conditions, the higher hippocampal cellular viability of slices obtained from mice-treated ex vivo with a dose of creatine that affords anti-immobility effect reinforces the notion that this compound should be further investigated as a promising antidepressant and neuroprotective agent.