Therapeutic and preventive effects of methylene blue on Alzheimer's disease pathology in a transgenic mouse model

Neuropharmacology 76 (2014) 68e79

Contents lists available at ScienceDirect

Neuropharmacology journal homepage: www.elsevier.com/locate/neuropharm

Therapeutic and preventive effects of methylene blue on Alzheimer’s disease pathology in a transgenic mouse model V. Paban a, *, C. Manrique b, M. Filali c, S. Maunoir-Regimbal d, F. Fauvelle d, B. Alescio-Lautier a a

Aix-Marseille Université, UMR 7260, FR3C, Lab. Neurosciences Intégratives et Adaptatives, 13331 Marseille, France Aix-Marseille Université et CNRS, Fédération de recherche comportement-cerveau-cognition (3C), FR3512 Marseille, France Neurobehavioral Phenotyping Platform, Quebec Genomics Center T2-50, 2705 Boulevard Laurier, Sainte-Foy, Québec, Canada d IRBA-CRSSA/EBR/RNI/Laboratoire de RMN, 24 av. des Maquis du Grésivaudan, 38700 La Tronche, France b c

a r t i c l e i n f o

a b s t r a c t

Article history: Received 21 February 2013 Received in revised form 14 May 2013 Accepted 19 June 2013

Methylene blue (MB) belongs to the phenothiazinium family. It has been used to treat a variety of human conditions and has beneficial effects on the central nervous system in rodents with and without brain alteration. The present study was designed to test whether chronic MB treatment taken after (therapeutic effect) or before (preventive effect) the onset of beta-amyloid pathology influences cognition in a transgenic mouse model (APP/PS1). In addition, the present study aims at revealing whether these behavioral effects might be related to brain alteration in beta-amyloid deposition. To this end, we conducted an in vivo study and compared two routes of drug administration, drinking water versus intraperitoneal injection. Results showed that transgenic mice treated with MB orally or following intraperitoneal injection were protected from cognitive impairments in a variety of social, learning, and exploratory tasks. Immunoreactive beta-amyloid deposition was significantly reduced in the hippocampus and adjacent cortex in MB-treated transgenic mice. Interestingly, these beneficial effects were observed independently of beta-amyloid load at the time of MB treatment. This suggests that MB treatment is beneficial at both therapeutic and preventive levels. Using solid-state High Resolution Magic Angle Spinning Nuclear Magnetic Resonance (HRMAS-NMR), we showed that MB administration after the onset of amyloid pathology significantly restored the concentration of two metabolites related to mitochondrial metabolism, namely alanine and lactate. We conclude that MB might be useful for the therapy and prevention of Alzheimer’s disease. This article is part of the Special Issue entitled ‘The Synaptic Basis of Neurodegenerative Disorders’. Ó 2013 Elsevier Ltd. All rights reserved.

Keywords: APP/PS1 mice Memory Beta-amyloid HRMAS NMR jMRUI Hippocampus Cortex

1. Introduction Alzheimer’s disease (AD) is the most common cause of dementia among elderly people. Clinically, AD results in progressive memory loss and other cognitive impairments, such as aphasia, apraxia, and personality changes. Despite the global nature of the cognitive dysfunction in AD, memory disorder is clearly the most prevalent and prominent feature of the early stages of the disease. The principal hallmarks of AD include extracellular accumulation of beta-amyloid peptide in the core of the neuritic plaque and intracellular accumulation of Tau proteins as neurofibrillary tangles and neuropil threads. Both of them are mandatory to diagnose of AD (Jellinger, 1998). Others histopathological features, such as

* Corresponding author. Tel.: þ33 413550887. E-mail address: [email protected] (V. Paban). 0028-3908/$ e see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.neuropharm.2013.06.033

mitochondrial deficits, do not belong to the diagnosis criteria, but have been also considered important pathological components. Methylene blue (MB) belongs to the phenothiazinium family. It has been used for more than a century in a wide range of fields, including biology, chemistry, and medicine. In clinical medicine, MB is used in a wide range of indications such as methemoglobinemia, ifosfamide-induced encephalopathy, and thyroid surgery (see Oz et al., 2011 for review). MB has also been reported to have therapeutic effects in psychosis and mania (Deutsch et al., 1997; Narsapur and Naylor, 1983; Naylor et al., 1988). In rodents, Rojas et al. (2012) reviewed the role of MB on memory in normal brain as well as various animal model of neurodegenerative disease including Huntington’s disease (Sontag et al., 2012) and AD (Medina et al., 2010; O’Leary et al., 2010). In vitro studies support the premise that MB inhibits the formation of beta-amyloid oligomers by promoting fibril formation (Atamna and Kumar, 2010; Necula et al., 2007; Oz et al., 2009) and Huntingtin protein

V. Paban et al. / Neuropharmacology 76 (2014) 68e79

aggregation (Sontag et al., 2012). It exerts an anti-tau aggregation effect (Oz et al., 2009; Wischik et al., 2008; Taniguchi et al., 2005). MB crosses the bloodebrain barrier and has high bioavailability in the brain (Peter et al., 2000). It has also been demonstrated to penetrate selectively neuronal cell types after systemic administration, in particular hippocampal cells (Müller, 1998). MB has a wide range of activities that is due to its multiple cellular and molecular targets (Atamna et al., 2008; Martijn and Wiklund, 2010; Miclescu et al., 2010; Vutskits et al., 2008; Zhang et al., 2010). The aim of the present study is to evaluate the therapeutic and preventive effects of MB on AD, using a transgenic mouse model. Different mouse models featuring a range of aspects of AD are available (for review, see Lithner et al., 2011). The doubletransgenic mice used in the present study incorporate a chimeric human/mice amyloid precursor protein (APP) construct bearing the Swedish double mutation and the exon-9-deleted PS1 mutation (APP/PS1). Previous studies from our group and from other investigators have showed that these mice develop age-related aggregation of amyloid plaques in the hippocampus and the cortex as well as progressive cognitive impairment (Filali and Lalonde, 2009; Filali et al., 2011; Liu et al., 2002; Wang et al., 2003). To investigate whether MB treatment could have beneficial effects on cognitive impairment, APP/PS1 mice received MB administration. Low doses of MB were administered after (therapeutic effect) or before (preventive effect) the onset of beta-amyloid pathology. To further gain insights into the mechanisms related to betaamyloid deposition and the molecular effects of in vivo treatment with MB, solid-state High Resolution Magic Angle Spinning Nuclear Magnetic Resonance (HRMAS-NMR) was used to identify metabolic changes. HRMAS-NMR uses intact tissue samples and does not require soluble extraction (Martínez-Bisbal et al., 2004; Sitter et al., 2002). Therefore, it offers the advantages of retaining tissue morphology and eliminating contamination associated with extraction procedures. Multivariate statistical analyses can then be performed with these data, leading to the “metabolic profiling” used today for classification and prediction. It has been successfully applied in various tissues and diseases, including Alzheimer pathology (Mao et al., 2007; Paban et al., 2010; Sitter et al., 2002; Tzika et al., 2002; Woo et al., 2010). Two brain regions were examined, the hippocampus and cortical adjacent areas, which are highly involved in memory and higher cognitive functions and in

69

neurodegenerative disorders. The effect of in vivo MB treatment was also investigated at the level of beta-amyloid present in APP/ PS1 mice brain. 2. Material and methods 2.1. Animals Male transgenic mice harbouring the chimeric human/mouse APP gene with the Swedish mutation and the human presenilin I (A246E variant) (strain B6C3Tg(APP695)3Dbo Tg(PSEN1)5Dbo/J; Jackson Laboratories) were used. The mice were first on a B6C3 background and backcrossed for at least 10 generations to C57BL/6J. All newborn pups were genotyped and included in the different experimental groups. The housing conditions were controlled (temperature 21  C; light from 7:00 to 19:00; humidity 50e60%), and food and water were freely available. All protocols were conducted according to the Canadian Council on Animal Care guidelines, as administered by the Laval University Animal Welfare Committee. 2.2. Treatment and experimental design Methylene blue grade was provided by Provepharm S.A.S (Marseille, France) and administered either orally (via drinking water) or intraperitoneally (i.p.) 2 times per week. Because MB is bitter, sucrose was added in the drinking water. We chose a dose of 2 mg/10 ml in drinking water. The daily MB intake was estimated to be about 0.80 mg/day per mouse, on the assumption that the average daily consumption of water is 4 ml for a 35 g adult mouse. For i.p. injection, we chose a dose of 2 mg/kg diluted in 0.2 ml of distilled water. These doses were selected based on positive findings with 0.025% w/w supplemented diet for 16 weeks (Medina et al., 2010), and 1e4 mg/kg i.p. for 5 days (Deiana et al., 2009). The experimental design is summarized in Fig. 1. To analyze the therapeutic effects of MB (Fig. 1A), treatment was delivered after the onset of beta-amyloid pathology, i.e., in mice of 6 month-old, which developed amyloid deposits in the hippocampus and cortical adjacent areas. Two routes of administration were analysed: orally or intraperitoneally. Note that mice were submitted two times at the behavioral tests battery, i.e., at 6 (before treatment) and 9 (after treatment) months of age. Indeed, we recently reported that APP/PS1 mice had significant impaired performance starting from 6 months old and performed worsened up to 9 months of age (Filali and Lalonde, 2009). So, the question was to investigate whether 3 months period of MB treatment could stabilized these behavioral decline. Group-housed APP/PS1 bigenic (N ¼ 47) and littermate wildtype (Wt) mice (N ¼ 34) were separated into 6 groups, allowing us to investigate MB’s effects in both Wt and Alzheimer’s mice model: APP/PS1 sucrose water or saline (N ¼ 15, w/o MB), APP/PS1 MB in drinking water (N ¼ 15, w/MB Oral), APP/PS1 MB following i.p. injection (N ¼ 15, w/MB I.P.), Wt sucrose water or saline (N ¼ 11), Wt MB in drinking water (N ¼ 11), and Wt MB following i.p. injection (N ¼ 11). To analyze the preventive effects of MB (Fig. 1B), treatment was administered before the onset of beta-amyloid pathology, i.e., in mice of 2 months old. Based on the therapeutic effects of MB showing comparable results no matter what the administration route was, only one way of delivery was studied, we have chosen to

Fig. 1. Timeline of therapeutic (A) and preventive (B) effects of MB in APP/PS1 mice. Behavioral testing consisted in nesting task, lefteright discrimination learning, passive avoidance learning, and open field test administered in order. Indications are provided on the analysis done after sacrifices. IHC: immunohistochemistry.

70

V. Paban et al. / Neuropharmacology 76 (2014) 68e79

administer MB in the drinking water. As depicted in the literature, APP/PS1 mice have normal performance at 2 months of age, behavioral impairments were present later when amyloid deposit appeared, i.e., at 6 months of age. Thus, the question was to determine whether MB could prevent the apparition of these behavioral deficits. To do so, mice with or without oral MB administration were only tested at that time, i.e., at 6 months of age. Group-housed APP/PS1 (N ¼ 22) and littermate Wt mice (N ¼ 24) were separated into 4 groups: APP/PS1 sucrose water (N ¼ 11, w/o MB), APP/PS1 MB in drinking water (N ¼ 11, w/MB Oral), Wt sucrose water (N ¼ 12), Wt MB in drinking water (N ¼ 12). 2.3. Behavioral tests The behavioral testing was conducted in a room separate from the vivarium. Three AD-related behavioral functions were assessed: social behavior, learning, and exploratory activity. The nesting test was given first, followed by left-right discrimination learning, passive avoidance learning, and an open field test. 2.3.1. Nesting test The nesting behavior was used to measure changes in social behavior. Mice were individually housed in a cage containing sawdust for one week. On the first day of testing, two pieces of cotton (5  5 cm, Nestlets, Ancare, Bellmore, NY, USA) were introduced inside the home cage to allow nesting behavior. One day later, the quality of the nest was determined according to a five-point scale as described by Deacon (2006): 1 ¼ nestlet not noticeably touched, 2 ¼ nestlet partially torn up, 3 ¼ mostly shredded but often no identifiable nest site, 4 ¼ an identifiable but flat nest, and 5 ¼ a (near) perfect nest. 2.3.2. Lefteright discrimination learning The spatial learning abilities of mice were assessed with the T-water maze task. The T maze consisted of a stem and two arms (length of stem: 64 cm, length of arms: 30 cm, width: 12 cm, height of walls: 16 cm) made of clear Plexiglass. It was filled with water (23  1  C) at a height of 12 cm. An escape platform (11 cm  11 cm) was placed at the end of the target arm and was submerged 1 cm below the surface. The acquisition phase allows to evaluate animals for lefteright spatial learning. During the first two trials, platforms were placed on both arms to test the spontaneous turning preference of the mice. Afterwards, only the least chosen arm was reinforced by the escape platform, with approximately the same number of mice being reinforced on either side. The mice were placed in the stem of the T-maze and chose to swim either left or right until they found the submerged platform and escaped to it, to a maximum of 60 s. If the animals did not find the platform within this time limit, they were gently guided onto it by the experimenter. After reaching the platform, the mice remained on it for 20 s and then were immediately put back in the maze for up to a maximum of 48 trials. A rest period of at least 10e15 min was allowed between each block of 10 trials. A mouse was considered to have achieved criterion after 5 consecutive errorless trials. The reversal-learning phase was conducted 2 days later. The same protocol was repeated, except that the mice were trained to find the escape platform on the side opposite to that on which they had learned during the acquisition phase. Two measures were taken: number of trials to reach the criterion (5 of 5 correct choices made on consecutive trials) and escape latency. 2.3.3. Passive avoidance learning Based on the mice natural tendency to prefer dark environments, they were evaluated on their retention of nonspatial memory using a one-trial passive avoidance task. The passive avoidance apparatus (Ugo Basile, Italy, 47 cm  18 cm, wall height: 26) was divided into two compartments, one illuminated (start compartment) and the other dark (escape compartment). The floor of each compartment contained a grid, but only the grid of the dark compartment was electrified by a generator. On the training day, mice were placed into the lighted (start) compartment for a 60 s acclimation period. The guillotine door was then opened, and the latency to enter the dark side was recorded. After they entirely entered the dark compartment (4-paw criterion), mice received a mild footshock (0.5 mA, 2 s). Ten seconds after the shock they were removed from the dark compartment and returned to the home cage. On the next day, the mice were placed in the lit compartment and the latency taken to enter the dark compartment was measured as test latency, limited to a maximum of 300 s. 2.3.4. Open-field test The open-field test analyzes spontaneous locomotor activity, exploratory behavior, and anxiety. The apparatus is a circular open field made of Plexiglas (80 cm diameter, 40 cm height). Each animal was placed in the middle of the open field and allowed to explore the field for 5 min under continuous video observation. Paths were tracked digitally (Any-maze, Stoelting) and the distance traveled, velocity, time, and distance in the different zones (center, corners, outer) were calculated. The arena was thoroughly cleaned with 70% ethanol solution after each trial.

2.4. HRMAS NMR analysis 2.4.1. Sample preparation At the end of behavioral testing period, APP/PS1 bigenic (N ¼ 27) and littermate wildtype mice (N ¼ 27) were anesthetized with isoflurane gas, decapitated, and their brains were removed. The hippocampus and the adjacent cortex from one hemisphere were dissected out, immediately frozen in liquid nitrogen, and stored at 80  C to prevent biochemical and structural degradation. For HRMAS-NMR analysis, 30 mL of a cold 1 mM D2O solution of 3-(trimethylsilyl) propionic2,2,3,3-d4 acid (TSP) was added to 15 mg of frozen biopsy in a 50 mL zirconium rotor. 2.4.2. Data acquisition Spectra were recorded on a Bruker Avance 400 spectrometer (Bruker Biospin, Wissembourg, France) at a proton frequency of 400.13 MHz, using a Carr-PurcellMeiboom-Gill (CPMG) pulse sequence synchronized with the spinning rate of 4 kHz (interpulse delay 250 ms, total spin echo time 30 ms) (Wieruszeski et al., 2001) and presaturation of residual water signal, for a total acquisition time of 16 min. Resonance assignment was performed as previously described (Rabeson et al., 2008). 2.4.3. Metabolite quantification Quantification was performed with the jMRUI software package ((http://www. mrui.uab.es/mrui/) using the “subtract-QUEST” procedure (Ratiney et al., 2005)). The procedure involves a simulated metabolite database set including acetate (Ace), alanine (Ala), aspartate (Asp), creatine and phosphocreatine (Cr), choline (Cho), gamma-amino-butyric acid (GABA), glutamate (Glu), glutamine (Gln), glutathione (GSH), glycerophosphocholine (GPC), glycine (Gly), lactate (Lac), myo-inositol (Mins), N-acetylaspartate (NAA), phosphoethanolamine (PE), phosphocholine (PC), scyllo-inositol (s-Ins), and taurine (Tau). The amplitude of metabolites calculated by QUEST was normalized to the total spectrum signal, so that only relative concentrations were produced. 2.4.4. Multivariate statistics Quantified data were loaded in the SIMCA-P software version 13 (Umetrics, Umea, Sweden) as variables and scaled to unit variance before analysis. Principal component analysis (PCA) was first carried out with this X data matrix to check the homogeneity of data and eventually to exclude outliers, for each brain structure studied, hippocampus and cortex. To this aim, data were visualized by score plots, where each point represents a NMR spectrum and thus a sample. PLS-DA analysis (partial least square-discriminant analysis) was run to discriminate populations of rats by adding the strain as an additional Y matrix, i.e. wild or transgenic APP/PS1 mice (Wold et al., 1984). Then orthogonal partial least square-discriminant analysis (OPLS-DA) was run to improve the models (Bylesjö et al., 2006). OPLS-DA is identical to PLS-DA analysis, but this method allows to separate systemic variations of X that are unrelated to Y (orthogonal) from linear variations of X that are related to Y. The results were visualized by plotting the predictive component versus the orthogonal one. All OPLS-DA models were cross-validated to allow evaluation of the statistical significance of the model. During cross-validation the model is iteratively rebuilt using only 6/7 of the data as training set. The model is then used to predict the class of the remaining 1/7 data which serves as a test set. Cross-validation leads to the calculation of the Q2 and R2Y factors. R2Y is a quality factor while Q2 is a predictive factor, i.e., a good Q2 allows the model to be used for prediction. Typically, a robust model has a Q2 > 0.5 and a R2 > 0.5. The reliability of our OPLS-DA models was assessed by a CV-ANOVA test (analysis of variance test of cross-validated predictive residuals). The metabolites also received a Variable Influence on Projection (VIP) which allows ranking them according to their contribution to the separation. Metabolites with VIP < 0.5, which are less important for separation, were excluded and new OPLS-DA models were built. The new model was retained if R2Y and Q2 values were increased. Finally, we compared the output of the model with the actual results obtained following MB treatment. Treated animals, independently of the administration route, were imported in the model with no a priori class assignment, and the score scatter plots drawn for graphical inspection. 2.5. Immunohistochemical and microscopic analyses APP/PS1 bigenic (N ¼ 20) and littermate wildtype mice (N ¼ 20) were deeply anesthetized via an intraperitoneal injection of a mixture of ketamine hydrochloride and xylazine and then perfused transcardially with 4% paraformaldehyde. The brains were removed and post-fixed. They were quickly frozen and cut into entire series of 40-mm-thick coronal sections. Sections were processed for immunohistochemistry using a mouse anti-beta-amyloid (Covance, USA, 6E10, 1:1000). Immunohistochemical stainings were performed following the manufacturer’s protocol using a Vectastain ABC Elite kit (Vector Laboratories, France) coupled with the diaminobenzidine reaction as described previously (Chambon et al., 2007). The sections were digitized using a Nikon high-resolution digital camera, and the images were converted to grayscale using Image software (Lucia, Nikon).

Fig. 2. Therapeutic effects of MB on social behavior (A), spatial learning (B), passive avoidance learning (C), and exploratory activity (D). Transgenic APP/PS1 and wildtype (Wt) mice were submitted two times at the behavioral tests battery, i.e., at 6 (before treatment) and 9 (after treatment) months of age. MB was delivered either orally (w/MB Oral) or intraperitoneally (w/MB I.P.), or not administered (w/o MB). *p< 0.05; ***p < 0.001.

72

V. Paban et al. / Neuropharmacology 76 (2014) 68e79

Quantification of beta-amyloid staining was processed in the hippocampus and adjacent cortices in two coronal tissue sections at the rostral (1.00 mm to bregma) and caudal (2.75 mm to bregma) level (Paxinos and Watson, 1998). Data was expressed as percent plaque load, corresponding to the total amount of area covered with plaques relative to the total brain area. On 6E10 immunostained sections (3/ animal) frequencies of small (<6 mm2), medium (6.5e9.5 mm2), and large (>10 mm2) plaques were determined as percent of all plaques.

treatment period, Bonferroni multiple comparisons revealed shorter retention latencies for the APP/PS1 mice without treatment than for the APP/PS1 mice with MB treatment (p < 0.02). However, APP/PS1 mice treated with MB by I.P. injection had w/o MB w/ MB Oral

2.6. Data analysis

4 months treatment

3

2

1

0 Wt

APP/PS1

3. Results

w/o MB w/ MB Oral

3.1. Behavioral changes

*

Means of trials to criterion

4 months treatment

*

12 10 8 6 4 2 0 Wt

APP/PS1

w/o MB w/ MB Oral

4 months treatment

Latency (sec)

250 200 150 100 50 0 Wt

APP/PS1 w/o MB w/ MB Oral

35 30 25 Meter

The results of the therapeutic effects of MB on behavior are shown in Fig. 2. In the social behavior test (Fig. 2A), Student’s t-test computed on the nest scores of the APP/PS1 and Wt groups at 6 months of age, i.e., before MB treatment yielded a significant effect for the genotype factor (t(79) ¼ 4,72; p < 0,0001), indicating that APP/PS1 mice had poorer nesting activity than their wildtype littermates. Following 3 months of MB treatment, ANOVA revealed significant effects of genotype (F(1, 73) ¼ 23,70; p < 0,0001), treatment (F(1, 73) ¼ 19,45; p < 0,0001), and interaction (F(1, 73) ¼ 8,82; p < 0.004), but no administration effects of route (F(1, 73) ¼ 1.84; p ¼ 0.18) and interaction (F(1, 73) ¼ 0.004; p ¼ 0.95). Nest-building ability of the APP/PS1 mice at 9 months of age decreased when compared to that measured in 6 months-old mice, varying from incomplete versions to a complete absence of any nest at all. In contrast, APP/PS1 mice treated with MB increased nesting activity relative to APP/PS1 without treatment (p < 0.0001), independent of the administration route (p ¼ 0.86). MB treatment had no effect in Wt mice. In the T water maze, during the acquisition phase of lefteright discrimination learning, all groups exhibited comparable performance (data not shown). By contrast, during the reversal-learning phase (Fig. 2B), analysis performed on the means of trials to criterion yielded a significant genotype effect at both 6 months (t(79) ¼ 5.99; p < 0.0001) and 9 months (F(1, 73) ¼ 48.13; p < 0.0001), as the number of trials to criterion of APP/PS1 mice without treatment was higher than the one of Wt. Following MB administration, ANOVA of trials to criterion revealed a significant effect of the treatment (F(1, 73) ¼ 5.15; p < 0.03), but no interaction of treatment  administration route (F(1, 73) ¼ 0.10; p ¼ 0.75). This suggests that the effects of MB did not depend on the administration route. MB improved performances in APP/PS1 mice following drinking water and I.P. delivery. Bonferroni multiple comparisons revealed APP/PS1 mice with MB treatment had lesser trials to criterion than APP/PS1 mice without treatment (p < 0.005). Note, however, that APP/PS1 mice with MB treatment committed more errors than Wt mice. MB treatment had no effect in Wt mice. In the passive avoidance learning, there was no intergroup difference in initial latencies before entering the dark compartment at 6 and 9 months, i.e., before and after the treatment (data not shown). By contrast, ANOVA of retention latencies revealed a significant effect of genotype (Fig. 2C), as APP/PS1 mice exhibited poorer retention than Wt at both 6 (t(79) ¼ 3.32; p < 0.001) and 9 months (F(1,73) ¼ 22.75; p < 0.0001). Following the 3 months

*

4

Nesting score

All statistical analyses were performed using XLSTAT software. The effects of genotype, treatment, administration route, and their interaction with the behavioral and immunohistochemical were analysed with ANOVA followed by Bonferroni’s post hoc tests. Student’s t-test was also used to analyse intergroup differences on the genotype factor. Results were expressed as mean  SEM. HRMAS-NMR data were tested for the normality of distributions (ShapiroeWilk test, p > 0.05). For each metabolite the differences between the 2 genotypes groups were analyzed by Welch t-test, and p < 0.05 were considered significant. On these metabolites, the effects of genotypes, treatment, administration route, and their interaction were analysed by ANOVA followed by Bonferroni’s post hoc tests.

*

5

20 15

4 months treatment

10 5 0 Wt

APP/PS1

Fig. 3. Preventive effects of MB on social behavior (A), spatial learning (B), passive avoidance learning (C), and exploratory activity (D). Transgenic APP/PS1 and wildtype (Wt) mice were submitted at the behavioral tests battery only at the age of 6 monthsold, i.e., after MB treatment. MB was delivered in drinking water (w/MB Oral), or not administered (w/o MB). *p < 0.05.

V. Paban et al. / Neuropharmacology 76 (2014) 68e79

better performance than bigenic mice without treatment (p < 0.03), whereas following MB oral administration, this did not reach significance level (p ¼ 0.24). Note that retention latencies in APP/PS1 mice treated (by drinking water or by I.P.) did not reach those measured in Wt mice. MB treatment had no effect in Wt mice. In the open field (Fig. 2D), there was no genotype effect (t(79) ¼ 0.88; p ¼ 0.38) before treatment (data not shown). Three months later, ANOVA on the distance traveled, velocity, time, and distance in the different zones (center, corners, outer) yielded no significant effect (F(1, 73)  2.90; p  0.09), mice exhibited similar behavior independent of the genotype and treatment. Thus, MB had no effect on general activity in mice. The results of the preventive effects of MB on behavior are shown in Fig. 3. In the social behavior test (Fig. 3A), 6-month-old APP/PS1 mice without treatment had a lower quality of nest building compared with Wt (p < 0.001). Following 4 months of MB treatment, the nest score was up to 3.6  0.31 in APP/PS1, i.e., similar to the one of Wt mice. In the T water maze, during the learning phase, Wt and transgenic mice exhibited comparable performance, independent of the treatment (data not shown). By contrast, during the reversallearning phase, APP/PS1 mice without treatment had poorer performance than Wt as the number of trials to criterion increased in bigenic mice (p < 0.01) (Fig. 3B). Long-term MB treatment reduced

73

the means of trials to criterion in APP/PS1 mice significantly (p < 0.01). MB treatment had no effect in Wt mice. In the passive avoidance learning, there was no difference during the acquisition phase between the four groups of mice (data not shown). During retention test, ANOVA of the latencies yielded a significant effect of genotype (F(1,45) ¼ 3.95; p < 0.05), indicating that APP/PS1 mice exhibited poorer retention than Wt (Fig. 3C). Following the 4 months treatment period, APP/PS1 treated mice showed better performance than non-treated mice, but this did not reach statistical significance. ANOVA yielded no treatment effect (F(1, 45) ¼ 0.42; p > 0.52), and no interaction (F(1, 45) ¼ 0.69; p > 0.41). ANOVA on the distance traveled, velocity, time, and distance in the different zones (center, corners, outer) of the open field yielded no genotype effect (F(1, 45)  1.82; p  0.18), no treatment effect (F(1, 45)  3.67; p  0.06), and no interaction (F(1, 45)  3.31; p  0.07), indicating that mice exhibited similar behavior independently of the genotype and treatment (Fig. 3D). Note that the general health of the mice was regularly checked with a modified SHIRPA primary screen (Rogers et al., 1999). No difference between groups was observed (data not shown). The mice were assayed daily for body weight. Repeated-measures ANOVA on body weight of APP/PS1 mice with or without MB administration yielded a time effect (p  0.04), but no treatment and administration route effects (p  0.22).

Fig. 4. Therapeutic effects of MB on amyloid deposition. (A) Beta-amyloid burden (percent of total area immunostained for beta-amyloid deposition) in hippocampus and cortex of APP/PS1 mice without MB treatment (w/o MB), and with MB treatment delivered either orally (w/MB Oral) or intraperitoneally (w/MB I.P.). (B) Photomicrograph of beta-amyloid plaques in section of APP/PS1 mice cortex without MB (left) and with oral MB treatment (right). Arrows indicate 3 plaques size. (C) Total plaque number in hippocampus and cortex in APP/PS1 mice with and without MB treatment. (D) Numbers of small (<6 mm2), medium (6.5e9.5 mm2), and large (>10 mm2) plaques in hippocampus and cortex. Plaques were visualized using 6E10 immunohistochemistry. All group data are presented as mean  S.E.M. *p < 0.05; **p < 0.01; ***p < 0.001.

74

V. Paban et al. / Neuropharmacology 76 (2014) 68e79

3.2. Beta-amyloid deposition changes MB treatment administered after the onset of beta-amyloid pathology decreases beta-amyloid deposition, as indexed by betaamyloid burden, in both the hippocampus and adjacent cortex (Fig. 4). Quantification of beta-amyloid deposition revealed substantial reductions of around 25% following drinking water and I.P. MB delivery. ANOVA revealed a significant decrease in the total number of plaques (F(1, 113)  7.12; p  0.008), and in the number of small, medium, and large plaques (F(1, 113)  5.89; p  0.02) after MB treatment. MB treatment administered before the onset of beta-amyloid pathology decreased beta-amyloid deposition in both the hippocampus and the cortex (t(94)  1.98; p  0.02; Fig. 5). This decrease of amyloid staining was of around 15%. Interestingly, no significant decrease was observed in the number of medium and large plaque (Fig. 5C). 3.3. Metabolic profile changes To further gain insights into the molecular effects of in vivo treatment with MB, the HRMAS-NMR method was used to identify metabolic changes in the hippocampus and the cortex of mice with and without MB treatment. Note that the metabolic profiles were assessed only when MB was administered after the onset of amyloid pathology (therapeutic effect). 3.3.1. Metabolic profile changes in the hippocampus From the PCA and OPLS-DA analyses generated with nontreated mice and the 18 quantified metabolites as X variables, 2

samples and 7 metabolites with VIP < 0.5 were excluded (Fig. 6A). The score plot of the resulting OPLS-DA model showed separation of samples according to genotype (Fig. 6B). The model generated with one significant OPLS-predictive component had a cumulative R2Y of 0.88 and a cumulative Q2 of 0.54. The metabolites that predominantly contributed to the separation of wildtype and transgenic mice were Lac, GABA, NAA, Ala, Cho, Glu, GPC, M-ins, Gly, PE and Asp. The Welch t-test performed on individual metabolite concentrations yielded significant differences for Lac, GABA, NAA, Ala and Glu, (t(12)  2.21; p  0.05) but not for Cho, GPC, M-ins, Gly, PE and Asp (t(12)  1.51; p  0.16). Fig. 6D shows that the relative concentration of Lac and Ala increased whereas those of GABA, NAA, and Glu decreased in APP/PS1 mice compared to wildtype mice. Correlation matrix is reported in Table 1A. Ala showed a negative correlation with GABA and NAA. NAA was negatively correlated with Lac. Interestingly, beta-amyloid staining showed negative correlation with Glu (R ¼ 0.6; p < 0.04). The others metabolites did not show any trend with plaque load (p  0.06). As the statistical model described above presented a relatively good predictive power (Q2 ¼ 0.54), it was used to predict the efficacy of MB treatment, considering that mice treated with the most efficient paradigm should be classified in the wildtype group. Score plot projections (Fig. 6C) synthesized the individual metabolic profiles and the weak effect of treatment. Treated mice are distributed in the two groups, with no clear shift of transgenic mouse metabolic profile toward the wild group. Welch t-test showed that the concentration of Lac and Ala were significantly restored in transgenic mice (ANOVA, genotype  treatment interaction F(1, 25)  9.51; p  0.05) (Fig. 6D). This was not the case for the level of the GABA, NAA, and Glu. However, following MB the

Fig. 5. Preventive effects of MB on amyloid deposition. (A) Beta-amyloid burden in hippocampus and cortex of APP/PS1 mice without MB treatment (w/o MB), and with oral MB treatment (w/MB Oral). (B) Photomicrograph of beta-amyloid plaques in section of APP/PS1 mice hippocampus without MB (left) and with oral MB treatment (right). (C) Total plaque number in hippocampus and cortex in APP/PS1 mice with and without MB treatment. (D) Numbers of small (<6 mm2), medium (6.5e9.5 mm2), and large (>10 mm2) plaques in hippocampus and cortex. Plaques were visualized using 6E10 immunohistochemistry. All group data are presented as mean  S.E.M. *p < 0.05; ***p < 0.001.

V. Paban et al. / Neuropharmacology 76 (2014) 68e79

75

Fig. 6. Hippocampal metabonomic changes following MB treatment after the onset of amyloid pathology. (A) Variable Influence on Projection (VIP) of OPLS-DA model built with 18 metabolites in wildtype (Wt) and APP/PS1 mice hippocampus obtained from 1H HRMAS NMR data and quantified by jMRUI. (B) Score plot of the OPLS-DA model with 11 metabolites (after exclusion of the metabolites with VIP<0.5) Q2 ¼ 0.54 and R2Y ¼ 0.88. (C) Projection of APP/PS1 mice with MB treatment in the OPLS-DA model. (D) Variation rate of relative concentration of the 5 selected metabolites between transgenic and Wt mice with or without MB treatment in the hippocampus. MB was provided either in drinking water (w/MB Oral) or by I.P. injection (w/MB I.P.). A positive score indicates an increased and a negative score indicates a decreased of relative concentration in the transgenic mice compared to Wt mice. ¤p < 0.05; ¤¤p < 0.01; ANOVA; genotype  treatment interaction.

variation rate was smaller than the one measured without treatment (about 5e10%). In wildtype mice, MB altered the level of 2 metabolites. Lac (F(1, 25) ¼ 4.32; p < 0.05) was decreased and M-ins (F(2, 25) ¼ 6.71; p < 0.02) was increased following MB administration, in particular when the drug was delivered through drinking water (data not shown). 3.3.2. Metabolic profile changes in the cortex The OPLS-DA model built from non-treated wildtype and APP/ PS1 mice was improved after exclusion of 5 metabolites with

VIP < 0.5 (Fig. 7A). The resulting model showed a marked separation between samples according to genotype (Fig. 7B). This model generated with one predictive OPLS-DA component had a cumulative R2Y of 0.77 and a cumulative Q2 of 0.48. The separation was attributed to GPC, Lac, PC, Cho, Ala, GABA, PE, Tau, M-ins, Gln, Ace, NAA, and Glu. Welch t-test for each metabolite revealed significant differences between the two groups for GPC, Lac, Cho, Ala, and GABA (t(12)  2.19; p  0.05). Fig. 7D shows that GPC, Lac, and Ala increased whereas Cho and GABA decreased significantly in transgenic compared to wildtype mice. Correlation matrix is

76

V. Paban et al. / Neuropharmacology 76 (2014) 68e79

Table 1 Correlation matrix between metabolites identified by HRMAS NMR and betaamyloid burden in the hippocampus (A) and adjacent cortex (B) of APP/PS1 mice without treatment.

A Ala GABA Glu Lac NAA

Ala

GABA

Glu

Lac

1 -0,6 * -0,2 0,4 -0,6 *

1 0,2 -0,2 0,2

1 -0,3 0,2

1 -0,6 *

1

0,3

-0,4

-0,6 *

0,5

-0,5

Lac

beta-amyloid

NAA

beta-amyloid

1

B Ala 1 -0,3 -0,4 0,5 0,6 *

Cho

GABA

GPC

Ala Cho GABA GPC Lac

1 0,8 * -0,4 -0,3

1 -0,3 -0,2

1 0,1

1

beta-amyloid

0,6 *

-0,2

-0,5

0,6 *

0,2

beta-amyloid

1

reported in Table 1B. A positive correlation was observed between GABA and Cho, and Lac and Ala. Beta-amyloid staining showed positive correlation with Ala and GPC (R  0.6; p  0.01). The others metabolites did not show any trend with plaque load (p  0.07). Like for hippocampus, no clear recovery was observed on the score scatter plot of treated mice projected in the OPLS-DA model. However metabolic profiles of treated animals seem to be intermediate between transgenic and wild group. Welch t-test indicated that the concentration of Lac and Ala were significantly restored in transgenic mice (F(1, 24)  4.11; p  0.05) (Fig. 7D). No significant effect of MB in APP/PS1 mice was noted on GPC, Cho, and GABA. The concentration of these metabolites was still altered following MB treatment; only a tendency to restore could be noted for Cho and GABA following I.P. MB treatment. In wildtype mice, MB administration induced a decreased in Lac and Tau (F(1, 24)  4.75; p  0.04) (data not shown). 4. Discussion Without any treatment, APP/PS1 mice exhibit age-related deficits in learning and memory associated with accumulation of betaamyloid and metabolic dysregulation. Following 3 months of oral or intraperitoneal administration of MB after the onset of beta-amyloid pathology, APP/PS1 mice performed as well as wildtype mice in their nest-building behavior, suggesting a better social interaction and a decrease in apathy. MBtreated APP/PS1 mice performed better than non-treated transgenic mice during the reversal phase of the T-water maze and the retention of passive avoidance task, indicating that the progressive decline observed in non-treated APP/PS1 mice was not measured following MB treatment during that time. MB alleviated the cognitive deficit in spatial and non-spatial memory. Interestingly, the present study showed that these beneficial effects were measured no matter what the route of administration was, oral versus intraperitoneal. Note, however, that transgenic mice did not reach similar score than wildtype mice. Future optimization of MB dosing and/or treatment time may yield stronger beneficial effects from MB treatment. Indeed, as reviewed by Bruchey and GonzalezLima (2008) and Rojas et al. (2012), MB has hormetic dosee response behavioral effects in normal rodents. When used at doses over 10 mg/kg, MB is behaviorally ineffective and might even have adverse effects on general locomotor behavior. By contrast, when MB is used at low doses ranging between 1 and 4 mg/kg, MB

improves memory in various tasks, including inhibitory avoidance, spatial learning, object recognition, discrimination memory, and habituation to a familiar environment. To date, only a single study has investigated the cognitive effects of MB in mammal model of AD. In this initial study, Medina et al. (2010) reported that 16 weeks of MB dietary after the onset of pathology reversed the learning impairment in the spatial references version of the Morris water maze in the 3xTg-Alzheimer mice model. Interestingly, they found that soluble beta-amyloid levels were reduced in the brains of the MB-treated 3xTg-AD mice compared with the 3xTg-AD mice on the control diet. By contrast, the treatment did not affect abnormal tau phosphorylation. This latter result is in agreement with in vivo experiments indicating that MB does not reduce tau pathology (van Bebber et al., 2010; O’Leary et al., 2010). APP/PS1 mice show an aged dependent increased in betaamyloid deposit starting at 6 months (Liu et al., 2002; Wang et al., 2003). This amyloid burden correlates with the number of plaques. Accordingly, we showed that transgenic mice at 9 months of age exhibited a higher amount of small plaques in the hippocampus and cortical areas. Interestingly, following MB administration, the number of amyloid plaques and the area occupied by such deposits were lower in transgenic mice treated with MB. Since the treatment was begun at the early stage of plaque deposition, this suggests that MB prevents the formation of plaques. We cannot exclude a clearance effect of MB on existing plaques since the treatment was administered after the onset of beta-amyloid pathology, i.e., in mice already harbouring amyloid deposits. If so, MB cleared beta-amyloid deposits no matter what plaque size was small, medium and large. Our results are consistent with in vitro studies showing that MB inhibits the aggregation of amyloid oligomers by enhancing fibril formation (Necula et al., 2007; Taniguchi et al., 2005). Oral MB treatment was not only beneficial at the therapeutic level, but also at the level of the prevention. MB administration before the onset of beta-amyloid pathology helped preventing the cognitive decline. APP/PS1 mice treated with MB did not show deterioration of their performances over the 6 months period. These positive effects were measured in multiple behavioral domains, notably social behavior and learning abilities. Concomitantly, the number of beta-amyloid plaques was lower in the brain of MB-treated bigenic mice. Since, MB was delivered before the onset of the pathology, these data suggest that MB either prevented the synthesis of beta-amyloid or favored the immediate clearance of the peptide. It should be noted that amyloid removal following MB administration was limited to 15e25%. One may suggest that the cognitive beneficial effects of MB measured in the present study are the result of MB action, not only on amyloid burden, but also on its multiple cellular and molecular targets. In particular, the memoryimproving effects of MB have been shown to be mediated by an enhancement of the neuronal oxidative metabolic capacity, known to support cellular processes with high-energy demand involved in learning and memory. In order to gain further understanding of the biochemical mechanisms behind the beneficial effects of MB in APP/PS1 mice, we conducted a HRMAS-NMR study of the changes in metabolite concentration in transgenic mice once the disease was well established. Moreover, the use of multivariate statistical analysis like OPLS-DA allowed to built characteristic metabolic profiles of APP/PS1 mice, that synthesize all individual metabolic variations that occur in hippocampus and cortex. The models built could then be used for prediction of therapeutic effect of MB. Abnormalities in cerebral metabolism in neurodegenerative diseases including AD have been widely documented. Consistent with previous in vivo and ex vivo MRS studies, brain regions showed specific patterns of

V. Paban et al. / Neuropharmacology 76 (2014) 68e79

77

Fig. 7. Cortical metabonomic changes following MB treatment after the onset of amyloid pathology. (A) Variable Influence on Projection (VIP) of OPLS-DA model built with 18 metabolites concentrations in wildtype (Wt) and APP/PS1 mice cortex, obtained from 1H HRMAS NMR data and quantified by jMRUI. (B) Score plot of the OPLS-DA model with 13 metabolites (after exclusion of the metabolites with VIP<0.5) Q2 ¼ 0.48 and R2Y ¼ 0.77. (C) Projection of APP/PS1 mice with MB treatment in the OPLS-DA model. (D) Variation rate of relative concentration of the 5 selected metabolites between Wt and transgenic mice with or without MB treatment in the cortex. MB was provided either in drinking water (w/ MB Oral) or by I.P. injection (w/MB I.P.). A positive score indicates an increased and a negative score indicates a decreased of relative concentration in the transgenic mice compared to Wt mice. ¤p < 0.05; ¤¤¤p < 0.001; ANOVA; genotype  treatment interaction.

metabolic changes following amyloid deposition. A decrease in the level of NAA is the most consistent data reported in APP mice model compared to wildtype (Chen et al., 2009; Dedeoglu et al., 2004; Marjanska et al., 2005; Oberg et al., 2008; von Kienlin et al., 2005; Woo et al., 2010; Xu et al., 2010). Although the biological function of NAA is not fully understood, it has been suggested that

NAA may reflect the energy status of neurons (Salek et al., 2010). Hence, its decrease can be considered as a marker of both neuronal loss and mitochondrial dysfunction (Angelie et al., 2001; Demougeot et al., 2001). The present data add to our knowledge showing a concomitantly reduction in Glu and GABA, both playing an important role in the regulation of neuronal excitability

78

V. Paban et al. / Neuropharmacology 76 (2014) 68e79

throughout the nervous system. In the hippocampus, we found a negative correlation between Glu and plaque load. There is considerable evidence that beta-amyloid oligomers interact with both ionotropic and metabotropic Glu receptors that have an important role in beta-amyloid neurotoxicity (Patel and Jhamandas, 2012). Interestingly, this decrease in Glu level was associated with Lac accumulation. Glu and Lac metabolism in astrocytes has been studied extensively (Alves et al., 1995; Brown and Ransom, 2007; Cerdán et al., 2006). Additionally, we showed that Ala level was increased in APP/PS1 mice. Elevations in Lac and Ala are classic markers for mitochondrial dysfunction. Hence, the high level of Lac and Ala measured in the hippocampus and cortex, associated with decrease in NAA, emphasized the existence of a mitochondrial impairment in mice harboring amyloid deposition. Interestingly, mitochondrial dysfunction has been shown to be a fundamental feature of AD (Anandatheerthavarada and Devi, 2007; Beal, 2005; Esposito et al., 2006; Schon and Area-Gomez, 2012). Our metabolic approach also highlighted a dysregulation of Cho and GPC in the cortex of APP/PS1 mice. None of the gliosis metabolites, Tau, Mins or S-ins, were significantly altered in our APP/PS1 mice model. Conflicting results have been reported concerning the level of inositol and other indicators of gliosis in APP transgenic mice, some showing increases (Chen et al., 2009; Dedeoglu et al., 2004; Marjanska et al., 2005 Woo et al., 2010), and others decreases (Labak et al., 2010; von Kienlin et al., 2005; Yenkoyan et al., 2011). Such discrepancies might be related, for example, to the plaque load at the time of sacrifice. Indeed, studies in which increases are reported mice were over 20 months old and yielded a high plaque load. In the present study, amyloid burden did not exceed 10%, and this would not be enough to generate metabolite changes related to gliosis. Alternatively, difference in genetic background, or technical approaches might explain these contradictory findings. Following MB administration after the onset of amyloid pathology, among the seven metabolites affected in transgenic mice, only those related to mitochondrial metabolism, namely Ala and Lac were significantly restored. We interpret these decreases as indicating that mitochondrial dysfunction was reduced by MB treatment. Such findings are in accordance with studies showing the protective role of MB on mitochondria in various in vitro cell models (Atamna et al., 2008; Callaway et al., 2004; Bruchey and Gonzalez-Lima, 2008; Zhang et al., 2010; Wrubel et al., 2007). MB is known as a redox indicator with a low redox potential, which allows it to cycle in mitochondria. The review of Atamna and Kumar (2010) is of particular interest, illustrating the molecular mechanism by which MB increases mitochondrial complex IV, enhances cellular oxygen consumption, and increases heme synthesis. We showed that MB administration decreased the level of Lac metabolite in transgenic mice, but also in wildtype animal. These findings suggest that the mitochondrial effects of MB did not depend on the pathogenic state of the animal. Given the multiple molecular and cellular targets of MB, the present study provides evidence that the most significant effect of MB treatment, in vivo, was on mitochondrial metabolic pathways. Such effects were correlated with decrease in the beta-amyloid deposition and cognitive improvement in APP/PS1 mice. To date, the challenge is to identify the cellular mechanisms underlying these relationships and the way - direct and/or indirect by which they may interact. Metabonomic is a particularly powerful tool to assess brain changes, such as those occurring in mitochondria following in vivo MB treatment; however it does not allow to cover the large spectrum of biological activities belonging to MB. Alternatively, one may argued that MB acts on others cellular targets than mitochondria. Further studies should be conducted along this line to determine whether the beneficial effects of MB are mediated by non-mitochondrial targets.

5. Conclusion The present data show that MB treatment protects the brain of transgenic mice against cognitive impairment while concomitantly reducing brain amyloid load and mitochondrial dysfunction. Since this phenothiazine has been tested in humans, it is conceivable to propose its use as a treatment for Alzheimer’s disease before the onset of amyloid pathology, but also once the disease is well established. Furthers studies are still needed to better understand the in vivo effects of MB on cognitive functioning, in particular in regards to its hormetic pharmacological effects. Conflict of interest There is no conflict of interest. Acknowledgment This research was supported by a Provence Alpes Côte d’Azur region and European funding. We are grateful for Dr. Lopez Christophe for reading the manuscript. References Alves, P.M., McKenna, M.C., Sonnewald, U., 1995. Lactate metabolism in mouse brain astrocytes studied by [13C]NMR spectroscopy. Neuroreport 6, 2201e2204. Anandatheerthavarada, H.K., Devi, L., 2007. Amyloid precursor protein and mitochondrial dysfunction in Alzheimer’s disease. Neuroscientist 13, 626e638. Angelie, E., Bonmartin, A., Boudraa, A., Gonnaud, P.M., Mallet, J.J., SappeyMarinier, D., 2001. Regional differences and metabolic changes in normal aging of the human brain: proton MR spectroscopic imaging study. A.J.N.R. Am. J. Neuroradiol. 22, 119e127. Atamna, H., Kumar, R., 2010. Protective role of methylene blue in Alzheimer’s disease via mitochondria and cytochrome c oxidase. J. Alzheimers Dis. 20, S439e S452. Atamna, H., Nguyen, A., Schultz, C., Boyle, K., Newberry, J., Kato, H., Ames, B.N., 2008. Methylene blue delays cellular senescence and enhances key mitochondrial biochemical pathways. FASEB J. 22, 703e712. Beal, M.F., 2005. Oxidative damage as an early marker of Alzheimer’s disease and mild cognitive impairment. Neurobiol. Aging 26, 585e586. Brown, A.M., Ransom, B.R., 2007. Astrocyte glycogen and brain energy metabolism. Glia 55, 1263e1271. Bruchey, A.K., Gonzalez-Lima, F., 2008. Behavioral, physiological and biochemical hormetic responses to the autoxidizable dye methylene blue. Am. J. Pharmacol. Toxicol. 3, 72e79. Bylesjö, M., Rantalainen, M., Cloarec, O., Nicholson, J.K., Holmes, E., Tryyg, J., 2006. OPLS discriminant analysis: combining the strengths of PLS-DA and SIMCA classification. J. Chemom. 20, 341e351. Callaway, N.L., Riha, P.D., Bruchey, A.K., Munshi, Z., Gonzalez-Lima, F., 2004. Methylene blue improves brain oxidative metabolism and memory retention in rats. Pharmacol. Biochem. Behav. 77, 175e181. Cerdán, S., Rodrigues, T.B., Sierra, A., Benito, M., Fonseca, L.L., Fonseca, C.P., GarcíaMartín, M.L., 2006. The redox switch/redox coupling hypothesis. Neurochem. Int. 48, 523e530. Chambon, C., Paban, V., Manrique, C., Alescio-Lautier, B., 2007. Behavioral and immunohistological effects of cholinergic damage in immunolesioned rats: alteration of c-Fos and polysialylated neural cell adhesion molecule expression. Neuroscience 147, 893e905. Chen, S.Q., Wang, P.J., Ten, G.J., Zhan, W., Li, M.H., Zang, F.C., 2009. Role of myoinositol by magnetic resonance spectroscopy in early diagnosis of Alzheimer’s disease in APP/PS1 transgenic mice. Dement. Geriatr. Cogn. Disord. 28, 558e566. Deacon, R.M., 2006. Assessing nest building in mice. Nat. Protoc. 1, 1117e1119. Dedeoglu, A., Choi, J.K., Cormier, K., Kowall, N.W., Jenkins, B.G., 2004. Magnetic resonance spectroscopic analysis of Alzheimer’s disease mouse brain that express mutant human APP shows altered neurochemical profile. Brain Res. 1012, 60e65. Deiana, S., Harrington, C.R., Wischik, C.M., Riedel, G., 2009. Methylthioninium chloride reverses cognitive deficits induced by scopolamine: comparison with rivastigmine. Psychopharmacology (Berl) 202, 53e65. Demougeot, C., Garnier, P., Mossiat, C., Bertrand, N., Giroud, M., Beley, A., Marie, C., 2001. N-acetylaspartate, a marker of both cellular dysfunction and neuronal loss: its relevance to studies of acute brain injury. J. Neurochem. 77, 408e415. Deutsch, S.I., Rosse, R.B., Schwartz, B.L., Fay-McCarthy, M., Rosenberg, P.B., Fearing, K., 1997. Methylene blue adjuvant therapy of schizophrenia. Clin. Neuropharmacol 20, 357e363.

V. Paban et al. / Neuropharmacology 76 (2014) 68e79 Esposito, L., Raber, J., Kekonius, L., Yan, F., Yu, G.Q., Bien-Ly, N., Puoliväli, J., ScearceLevie, K., Masliah, E., Mucke, L., 2006. Reduction in mitochondrial superoxide dismutase modulates Alzheimer’s disease-like pathology and accelerates the onset of behavioral changes in human amyloid precursor protein transgenic mice. J. Neurosci. 26, 5167e5179. Filali, M., Lalonde, R., Rivest, S., 2011. Anomalies in social behaviors and exploratory activities in an APPswe/PS1 mouse model of Alzheimer’s disease. Physiol. Behav. 104, 880e885. Filali, M., Lalonde, R., 2009. Age-related cognitive decline and nesting behavior in an APPswe/PS1 bigenic model of Alzheimer’s disease. Brain Res. 1292, 93e99. Jellinger, K.A., 1998. The neuropathological diagnosis of Alzheimer disease. J. Neural Transm. Suppl. 53, 97e118. Labak, M., Foniok, T., Kirk, D., Rushforth, D., Tomanek, B., Jasiñski, A., Grieb, P., 2010. Metabolic changes in rat brain following intracerebroventricular injections of streptozotocin: a model of sporadic Alzheimer’s disease. Acta Neurochir. Suppl. 106, 177e181. Lithner, C.U., Hedberg, M.M., Nordberg, A., 2011. Transgenic mice as a model for Alzheimer’s disease. Curr. Alzheimer Res. 8, 818e831. Liu, L., Ikonen, S., Heikkinen, T., Heikkilä, M., Puoliväli, J., van Groen, T., Tanila, H., 2002. Effects of fimbriaefornix lesion and amyloid pathology on spatial learning and memory in transgenic APP þ PS1 mice. Behav. Brain Res. 134, 433e445. Mao, H., Toufexis, D., Wang, X., Lacreuse, A., Wu, S., 2007. Changes of metabolite profile in kainic acid induced hippocampal injury in rats measured by HRMAS NMR. Exp. Brain Res. 183, 477e485. Marjanska, M., Curran, G.L., Wengenack, T.M., Henry, P.G., Bliss, R.L., Poduslo, J.F., Jack Jr., C.R., Ugurbil, K., Garwood, M., 2005. Monitoring disease progression in transgenic mouse models of Alzheimer’s disease with proton magnetic resonance spectroscopy. Proc. Natl. Acad. Sci. U. S. A. 102, 11906e11910. Martijn, C., Wiklund, L., 2010. Effect of methylene blue on the genomic response to reperfusion injury induced by cardiac arrest and cardiopulmonary resuscitation in porcine brain. BMC Med. Genomics 3, 27e33. Martínez-Bisbal, M.C., Arana, E., Martí-Bonmatí, L., Mollá, E., Celda, B., 2004. Cognitive impairment: classification by 1H magnetic resonance spectroscopy. Eur. J. Neurol. 11, 187e193. Medina, D.X., Caccamo, A., Oddo, S., 2010. Methylene blue reduces ab levels and rescues early cognitive deficit by increasing proteasome activity. Brain Pathol. 21, 140e149. Miclescu, A., Sharma, H.S., Martijn, C., Wiklund, L., 2010. Methylene blue protects the cortical blood-brain barrier against ischemia/reperfusion-induced disruptions. Crit. Care Med. 38, 2199e2206. Müller, T., 1998. Methylene blue supravital staining: an evaluation of its applicability to the mammalian brain and pineal gland. Histopathology 13, 1019e 1026. Narsapur, S.L., Naylor, G.J., 1983. Methylene blue. A possible treatment for manic depressive psychosis. J. Affect. Disord. 5, 155e161. Naylor, G.J., Smith, A.H., Connelly, P., 1988. Methylene blue in mania. Biol. Psychiatry 24, 941e942. Necula, M., Breydo, L., Milton, S., Kayed, R., van der Veer, W.E., Tone, P., Glabe, C.G., 2007. Methylene blue inhibits amyloid Abeta oligomerization by promoting fibrillization. Biochemistry 46, 8850e8860. Oberg, J., Spenger, C., Wang, F.H., Andersson, A., Westman, E., Skoglund, P., Sunnemark, D., Norinder, U., Klason, T., Wahlund, L.O., Lindberg, M., 2008. Age related changes in brain metabolites observed by 1H MRS in APP/PS1 mice. Neurobiol. Aging 29, 1423e1433. O’Leary, J.C., Li, Q., Marinec, P., Blair, L.J., Congdon, E.E., Johnson, A.G., Jinwal, U.K., Koren, J., Jones, J.R., Kraft, C., Peters, M., Abisambra, J.F., Duff, K.E., Weeber, E.J., Gestwicki, J.E., Dickey, C.A., 2010. Phenothiazine-mediated rescue of cognition in tau transgenic mice requires neuroprotection and reduced soluble tau burden. Mol. Neurodegener. 5, 45e53. Oz, M., Lorke, D.E., Hasan, M., Petroianu, G.A., 2011. Cellular and molecular actions of Methylene Blue in the nervous system. Med. Res. Rev. 31, 93e117. Oz, M., Lorke, D.E., Petroianu, G.A., 2009. Methylene blue and Alzheimer’s disease. Biochem. Pharmacol. 78, 927e932. Paban, V., Fauvelle, F., Alescio-Lautier, B., 2010. Age-related changes in metabolic profiles of rat hippocampus and cortices. Eur. J. Neurosci. 31, 1063e1073. Patel, A.N., Jhamandas, J.H., 2012. Neuronal receptors as targets for the action of amyloid-beta protein (Abeta) in the brain. Expert Rev. Mol. Med. 14, e2e12. Paxinos, G., Watson, C., 1998. The Rat Brain in Stereotaxic Coordinates, fourth ed. Academic Press, San Diego. Peter, C., Hongwan, D., Küpfer, A., Lauterburg, B.H., 2000. Pharmacokinetics and organ distribution of intravenous and oral methylene blue. Eur. J. Clin. Pharmacol. 56, 247e250.

79

Rabeson, H., Fauvelle, F., Testylier, G., Foquin, A., Carpentier, P., Dorandeu, F., van Ormondt, D., Graveron-Demilly, D., 2008. Quantitation with QUEST of brain HRMAS-NMR signals: application to metabolic disorders in experimental epileptic seizures. Magn. Reson. Med. 59, 1266e1273. Ratiney, H., Sdika, M., Coenradie, Y., Cavassila, S., van Ormondt, D., GraveronDemilly, D., 2005. Time-domain semi-parametric estimation based on a metabolite basis set. N.M.R. Biomed. 18, 1e13. Rogers, D.C., Jones, D.N., Nelson, P.R., Jones, C.M., Quilter, C.A., Robinson, T.L., Hagan, J.J., 1999. Use of SHIRPA and discriminant analysis to characterise marked differences in the behavioural phenotype of six inbred mouse strains. Behav. Brain Res. 105, 207e217. Rojas, J.C., Bruchey, A.K., Gonzalez-Lima, F., 2012. Neurometabolic mechanisms for memory enhancement and neuroprotection of methylene blue. Prog. Neurobiol. 96, 32e45. Salek, R.M., Xia, J., Innes, A., Sweatman, B.C., Adalbert, R., Randle, S., McGowan, E., Emson, P.C., Griffin, J.L., 2010. A metabolomic study of the CRND8 transgenic mouse model of Alzheimer’s disease. Neurochem. Int. 56, 937e947. Schon, E.A., Area-Gomez, E., 2012. Mitochondria-associated ER membranes in Alzheimer disease. Mol. Cell. Neurosci. doi:pii: S1044e7431(12)00127-3.10.1016/ j.mcn.2012.07.011. Sitter, B., Sonnewald, U., Spraul, M., Fjösne, H.E., Gribbestad, I.S., 2002. High-resolution magic angle spinning MRS of breast cancer tissue. N.M.R. Biomed. 15, 327e337. Sontag, E.M., Lotz, G.P., Agrawal, N., Tran, A., Aron, R., Yang, G., Necula, M., Lau, A., Finkbeiner, S., Glabe, C., Marsh, J.L., Muchowski, P.J., Thompson, L.M., 2012. Methylene blue modulates huntingtin aggregation intermediates and is protective in Huntington’s disease models. J. Neurosci. 32, 1109e1119. Taniguchi, S., Suzuki, N., Masuda, M., Hisanaga, S., Iwatsubo, T., Goedert, M., Hasegawa, M., 2005. Inhibition of heparin-induced tau filament formation by phenothiazines, polyphenols, and porphyrins. J. Biol. Chem. 280, 7614e7623. Tzika, A.A., Cheng, L.L., Goumnerova, L., Madsen, J.R., Zurakowski, D., Astrakas, L.G., Zarifi, M.K., Scott, R.M., Anthony, D.C., Gonzalez, R.G., Black, P.M., 2002. Biochemical characterization of pediatric brain tumors by using in vivo and ex vivo magnetic resonance spectroscopy. J. Neurosurg. 96, 1023e1031. van Bebber, F., Paquet, D., Hruscha, A., Schmid, B., Haass, C., 2010. Methylene blue fails to inhibit Tau and polyglutamine protein-dependent toxicity in zebrafish. Neurobiol. Dis. 39, 265e271. von Kienlin, M., Künnecke, B., Metzger, F., Steiner, G., Richards, J.G., Ozmen, L., Jacobsen, H., Loetscher, H., 2005. Altered metabolic profile in the frontal cortex of PS2APP transgenic mice, monitored throughout their life span. Neurobiol. Dis. 18, 32e39. Vutskits, L., Briner, A., Klauser, P., Gascon, E., Dayer, A.G., Kiss, J.Z., Muller, D., Licker, M.J., Morel, D.R., 2008. Adverse effects of methylene blue on the central nervous system. Anesthesiology 108, 684e692. Wang, J., Tanila, H., Puoliväli, J., Kadish, I., van Groen, T., 2003. Gender differences in the amount and deposition of amyloidbeta in APPswe and PS1 double transgenic mice. Neurobiol. Dis. 14, 318e327. Wieruszeski, J.M., Montagne, G., Chessari, G., Rousselot-Pailley, P., Lippens, G., 2001. Rotor synchronization of radiofrequency and gradient pulses in high-resolution magic angle spinning NMR. J. Magn. Reson. 152, 95e102. Wischik, C.M., Bentham, P., Wischik, D.J., Seng, K.M., 2008. Tau aggregation inhibitor (TAI) therapy with remberÔ arrests disease progression in mild and moderate Alzheimer’s disease over 50 weeks. Alzheimer’s Dement. 4, T167eT175. Wold, S., Ruhe, A., Wold, H., Dunn, W.I., 1984. The collinearity problem in linear regression. The partial least squares approach to generalized inverses. S.I.A.M. J. Sci. Stat. Comput. 5, 735e743. Woo, D.C., Lee, S.H., Lee, D.W., Kim, S.Y., Kim, G.Y., Rhim, H.S., Choi, C.B., Kim, H.Y., Lee, C.U., Choe, B.Y., 2010. Regional metabolic alteration of Alzheimer’s disease in mouse brain expressing mutant human APP-PS1 by 1H HR-MAS. Behav. Brain Res. 211, 125e131. Wrubel, K.M., Riha, P.D., Maldonado, M.A., McCollum, D., Gonzalez-Lima, F., 2007. The brain metabolic enhancer methylene blue improves discrimination learning in rats. Pharmacol. Biochem. Behav. 86, 712e717. Xu, W., Zhan, Y., Huang, W., Wang, X., Zhang, S., Lei, H., 2010. Reduction of hippocampal N-acetyl aspartate level in aged APP(Swe)/PS1(dE9) transgenic mice is associated with degeneration of CA3 pyramidal neurons. J. Neurosci. Res. 88, 3155e3160. Yenkoyan, K., Safaryan, K., Chavushyan, V., Meliksetyan, I., Navasardyan, G., Sarkissian, J., Galoyan, A., Aghajanov, M., 2011. Neuroprotective action of proline-rich polypeptide-1 in â-amyloid induced neurodegeneration in rats. Brain Res. Bull. 86, 262e271. Zhang, Y., Zhao, J., Zhang, T., Yang, Z., 2010. In vitro assessment of the effect of methylene blue on voltage-gated sodium channels and action potentials in rat hippocampal CA1 pyramidal neurons. Neurotoxicology 31, 724e729.