Is L-methionine a trigger factor for Alzheimer’s-like neurodegeneration?: Changes in Aβ oligomers, tau phosphorylation, synaptic proteins, Wnt signaling and behavioral impairment in wild-type mice

Previous studies have shown that metabolites from the methionine-homocysteine cycle can enhance the activity of γ-secretase, a protease that cleaves APP [36]. Increased γ-secretase activity favors the amyloidogenic processing of APP, causing an increase of Aβ₄₀ and Aβ₄₂ in APP transgenic mice, as measured by the ELISA assay [34]. To test whether this effect occurred in our wild-type mice experimental groups, we measured the Aβ peptide accumulation. Hippocampal lysates from both groups (methionine-treated and non-treated) were used to detect both Aβ₄₂ and Aβ₄₀ levels by two different ELISA assays, obtained by Millipore [37] and a mouse-specific Aβ ELISA previously described [38]. The levels of both peptides increased significantly in methionine-treated mice relative to control mice (Fig. 2a, b, c and d). Additionally, using immunohistochemistry, we examined different sections of brain tissues from control and L-methionine mice using the WO2 antibody. Brain sections from L-methionine mice showed higher immunoreactivity than control mice did in the hippocampus, in agreement with the biochemical analysis (Fig. 2e), however we cannot discard that immunoreactivy observed may reflect also Aβ, APP, APP fragments, βCTF, or Aβ fragments. To determine the apparition of oligomeric Aβ species with the treatment, we performed immunoprecipitation assay of oligomers with the A11 antibody, which binds to a specific oligomeric conformation, and confirm that these oligomers were of Aβ using the 4G8 antibody in tris-Tricine gel. The western blot detects a 56 kDa band only in treated mice, suggesting the presence of amyloid oligomers in the hippocampus of methionine treated animals, which has previously been described as a toxic assembly of Aβ that contributes to cognitive decline [39]. As control, we did a western blot using the Aβ-unrelated antibody (anti-c-jun) to detect IgG and discard a possible reactivity of heavy and light chains in the experiment (Fig. 2f-i). Also, we performed an exploratory indirect immunofluorescence using the 4G8 antibody noticing the presence of fluorescent marks in the treated mice brain and not so in the control mice, suggesting the presence of non-fibrillar aggregates of Aβ peptides or senile plaques in methionine treated mice. (Fig. 2f-ii). According to these results, we decided to evaluate the presence of amyloid-β plaques in L-methionine and control mouse brains. Tissues were analyzed with Thioflavine-S staining in both the cortex and hippocampal sections CA1 and CA3 and in the dentate gyrus (DG). The staining did not show the presence of plaques in treated or control brain tissues (data not shown). Taken together, these results showed an increase in the levels of both Aβ peptides and Aβ oligomers, without presenting detectable plaques, as a direct or indirect consequence of chronic L-methionine overconsumption.

Diverse studies have shown a link between methionine and the immuno-inflammatory activation, which is characteristic of neurodegenerative disease such as AD [40, 41]. To unveil whether a high level of L-methionine increases the brain inflammatory response, hippocampal slices were subjected to immunoblot and immune-cytochemical analysis. The astrogliosis marker GFAP [40, 41] and the microgliosis marker CD11b [40, 41] were evaluated in hippocampal lysates of the control and L-methionine groups (Fig. 3a). The levels of both markers were increased, as shown by densitometric analysis (Fig. 3b). Figure 3c shows representative photographs of the evaluated slices. The top row shows GFAP immunoreactivity in the hippocampi of control mice, and the bottom row shows the immunoreactivity in L-methionine-treated mice at different magnifications (10x, 20x and 40x). In every case, the white boxes indicate the area magnified from the photograph on the left. As shown in Fig. 3d, the fraction area of GFAP-positive hippocampal cells was significantly increased in L-methionine-treated mice. Additionally, the soma area of GFAP-positive cells was also significantly elevated (Fig. 3e). The immunohistochemical analyses were coincident with the data obtained from immunoblot analysis. These data show that an L-methionine-rich diet generates neuroinflammation and an increased activity of astrocytes and microglia in the hippocampus.

High levels of methionine and its metabolites could cause the deregulation of oxidative stress [42‐44]. Therefore we evaluated the total levels of 3-nitrotyrosine (3-NT) in proteins. Immunoblot analyses of control and L-methionine mice showed that the levels of total proteins with 3-NT was increased in L-methionine-treated mouse hippocampi. Lanes 1–4 show control hippocampus lysates, and lanes 5-8 show L-methionine hippocampus extracts, with each lane extracted from different animals (Fig. 4a). Densitometric quantifications of the 170, 70 and 35 kDa proteins are shown in Fig. 4b. In every case, proteins of L-methionine-treated mice had more 3-NT than control animals did. Brain slices were subjected to immunohistochemical analysis using 3-NT antibody, confirming our previous findings. Figure 4c shows representative photographs at two magnifications (20x and 40x) of evaluated slices; the top shows the control, and the bottom shows the L-methionine mouse cortex. Quantification showed that 3-NT was significantly increased in the L-methionine mouse cortex compared with the control mouse cortex (Fig. 4d). These data support the idea that oxidative stress species are increased in the cortex and hippocampus of mice treated with an L-methionine rich diet.

Postmortem brains from patients with schizophrenia are characterized by a decrease in the dendritic spine density in the neurons of the prefrontal cortex [45]. Interestingly, the schizophrenic symptoms are exacerbated with L-methionine treatment [46] because the amino acid is apparently responsible for the decrease in dendritic spine density in the mouse frontal cortex [47]. In addition, in chronic schizophrenic cerebellar samples, a loss of synaptic but not cytoskeletal proteins was observed [48]. We have previously shown that in a double transgenic mice model of AD, a decrease in the levels of several synaptic proteins was part of the synaptic failure observed in the hippocampus and cortex of AD mice [49, 50]. Therefore, we were interested in evaluating the effect of chronic L-methionine treatment on these synaptic proteins. First we evaluated the following pre-synaptic proteins: vesicle glutamate transporter 1 (VGluT1), synaptophysin (SYP), and synapsin (SYN) (Fig. 5a). The levels of VGluT1 and SYN, but not SYP, were significantly decreased in animals treated with L-methionine (Fig. 5b). Then, we evaluated the post-synaptic proteins corresponding to AMPA receptor subunit 2 (GluA2), NMDA receptor subunit 2B (GluN2B) and the scaffold protein of the post-synaptic density, PSD-95 (Fig. 5c). Mice treated with high doses of L-methionine exhibited decreased levels of GluA2 and PSD-95 compared with control mice (Fig. 5d). These results indicate that chronic L-methionine treatment induces the loss of several pre-and postsynaptic proteins in the hippocampi of wild-type mice.

The progressive decline of cognitive abilities is a defining characteristic of AD [31]. It is not clear whether methionine and/or its metabolites is/are partly responsible for the cognitive impairment or are predictors of cognitive decline, as previously suggested [28, 51, 52]. Furthermore, several studies have shown that high intake of methionine could cause a loss of memory [28, 51]. To evaluate whether a methionine-enriched diet has an effect on behavior, particularly in learning and memory processes, we subjected control and L-methionine-treated animals to different tasks, including the memory flexibility (MF) test, a variant of the Morris water maze, and the Novel Object Recognition (NOR) test. Figure 6a (i) shows that the animals from the control group continuously decreased the number of trials required to complete the task and reach the criterion, whereas animals from the L-methionine group did not show the same behavior as the days passed. This difference was significant on the fourth day of training; Control mice needed an average of 5 trials to reach the criterion, when L-methionine animals required an average of 10, needing twice as control trials to reach the criterion of the test. This difference indicates that L-methionine mice had significant memory impairment. The average velocity in the maze is shown in Fig. 6a (ii). No difference between the groups was observed, so the treatments did not affect the motor activity of animals, and the data obtained exclusively reflect memory impairment. After two days of rest, animals were subjected to the NOR test. Figure 6c (i) shows that the animals from the L-methionine group spent ~35 % less time exploring the new object compared with control mice. The cognition index was measured as the percentage (%) of time that the mice spent exploring the new object versus the time that the mice spent exploring both objects. Then, to discard any anxious behavior that could have been developed during the diet change, we performed an open field (OF) test. Figure 6b (i) shows no change in the % of time between control and L-methionine mice. Representative OF tracks are shown in Fig. 6b (ii). Other parameters of the OF were quantified; L-methionine animals did not show any significant difference in the traveled distance, % of time in the center, % of time in the borders or the number of sections crossed compared with control animals (Additional file 4: Table S2). These data are important to validate any posterior behavioral assessments because anxiety could have altered the animals’ conduct. Overall, L-methionine-treated mice showed memory impairment in two different tests (MF and NOR), which indicated that a diet high in L-methionine triggered a decline in the cognitive capacities of wild-type mice.

Wnt signaling plays an important role in the development of the nervous system and participates in the adult brain’s mediation and regulation of the structure and function of synapses [53, 54]. Over the years, several studies from our laboratory have strongly suggested a relationship between the loss of Wnt signaling and the neurotoxicity of Aβ, which is the main cause of the neuro-degeneration observed in AD [55‐57]. Additionally, canonical Wnt signaling is necessary for object recognition memory [58]. Therefore, we evaluated whether an L-methionine-enriched diet affects canonical Wnt signaling. Interestingly, methionine causes a decrease in the active form of β-catenin levels (Fig. 7a) and an increase in the activity of GSK-3β, as indicated by the reduction in the levels of its inhibited enzymatic form (phospho-Ser-9) (Fig. 7a). Figure 7b shows the densitometric analysis of the above results. The activation of Wnt signaling leads to the transcriptional activation of Wnt target genes [59‐61]. In contrast, in mice with an L-methionine-enriched diet, we observed a decrease in the levels of cyclin-D1 and c-jun. Interestingly, the decrease was almost 60 % compared with the control diet (Fig. 7c and d). These results suggest a loss of function of Wnt signaling in animals treated with L-methionine. Overall, this study shows that an L-methionine-enriched diet contributes to the appearance of Alzheimer’s-like neurodegeneration.

Primary antibodies used: mouse monoclonal anti-PSD95 clone K28/43, obtained from Antibodies Inc (UC Davis/NIH NeuroMab Facility), mouse anti-GluN2B (clone N59/36; UC Davis/NIH NeuroMab Facility), mouse anti-GluA2 (clone L21/32; UC Davis/NIH NeuroMab Facility), goat anti-synaptophysin (sc-7568 Santa Cruz Biotechnology, Inc.), mouse anti-Aβ 4G8 (Covance, Princeton, USA), mouse anti-Aβ WO2 (MABN10, Anti-Amyloid β Antibody, clone W0-2), conformational anti-amyloid oligomer A11 (AB9234, Millipore), Anti-β-Actin clone AC-15 (A1978, Sigma Aldrich), rabbit anti-GFAP (M0761, Dako, Cytomation) and anti-3-nitrotyrosine (A21285, Invitrogen).

A total of 12 wild-type (C57/BL6) female animals were treated with 8.2 g/kg L-methionine placed in their drinking water for 12 weeks. L-methionine concentration was determined based on the total water intake of mice to achieve high levels of the amino acid in the plasma without reaching toxic levels, which could lead to organ failure [30]. In fact, once a week throughout the course of treatment, we monitored the animals’ body weight and water intake. These results are shown in Additional file 1: Figure S1. To further analyze the health of the mice, lipid and hepatic parameters were measured at the end of the treatment period. No significant differences were observed between the control and L-methionine-treated groups (Additional file 4: Table S1). In parallel, a second group of control animals (C57/BL6) were treated for the same time period with regular tap water (n = 12). Water supplies with or without methionine were changed three times a week and consumption was measured.

The hippocampi from treated or control mice were dissected on ice and immediately frozen at–150 °C or processed as detailed previously [49, 102]. Briefly, slices were homogenized in RIPA buffer (50 mMTris-Cl, pH 7.5, 150 mMNaCl, 1 % NP-40, 0.5 % sodium deoxycholate, and 1 % SDS) supplemented with a protease inhibitor cocktail (Sigma-Aldrich P8340) and phosphatase inhibitors (50 mMNaF, 1 mM Na₃VO₄ and 30 μM Na₄P₂0₇) using a Potter homogenizer and were then passed sequentially through different caliber syringes. Protein samples were centrifuged at 14,000 rpm at 4 °C twice for 15 min. The protein concentration was determined using the BCA Protein Assay Kit (Pierce Biotechnology, Rockford, IL). Twenty- and forty-microgram samples were resolved by 10 % SDS-PAGE and transferred to a PVDF membrane. The reactions were followed by incubation with anti-mouse, anti-goat or anti-rabbit IgG peroxidase-conjugated antibody (Pierce, Rockford, IL) and developed using an ECL kit (Western Lightning Plus ECL, PerkinElmer).

To determine the concentrations of Aβ peptides, two Sandwich Enzyme-linked Immunosorbent Assay (ELISA) specific for Aβ_1-40/Aβ_1-42 were used as previously described. (EZBRAIN40, EZBRAIN42; EMD Millipore Corporation, Billerica, MA and ELISA kit (Catalog Number: 294-62501) Wako Human/Rat(Mouse) β-Amyloid (40) ELISA Kit (Catalog Number: 294-62501), and Wako Human/Rat(Mouse) β-Amyloid (42) [37, 38]. Briefly, hippocampal homogenates of control or methionine-treated animals were diluted to 2 μg/μl in homogenization buffer containing protease inhibitors. Approximately 50 μl of crude homogenate was prepared to measure Aβ_1-40/Aβ_1-42 levels according to the manufacturer’s instructions. Plates were read at the respective light wave on a Metertech 960 ELISA Analyzer. To detect soluble Aβ peptides for western blot, samples from the hippocampus were centrifuged at 20,0000 g for 1 h and the protein concentration from each supernatant was determined using the BCA Protein Assay Kit (Pierce Biotechnology, Rockford, IL).

Protein extracts were obtained from hippocampal tissue lysed in RIPA buffer supplemented with protease and phosphatase inhibitors. Aβ immunoprecipitation was performed using rabbit anti-A11 antibody using the protein G-Agarose. Immunocomplexes were analyzed by Tris-tricine gels, transferred to PVDF membranes and immunoblotted with a mouse anti-4G8 antibody.

The procedure was performed as previously described [102]. Animals were fixed with 4 % paraformaldehyde, and their brains were dissected and maintained in 4 % paraformaldehyde for 2 weeks to ensure proper fixation before being placed in a PBS-20 % sucrose solution for 48 h. Brain slices (30 μm) were obtained using a Leica CM1850 Cryostat and maintained in Olmos solution at 4 °C. Free-floating cytochemistry procedures were performed as follows: slices of tissue were treated with H₂O₂ in PBS to eliminate endogenous peroxidase activity, then washed with PBS and blocked with PBS/BSA (2 mg/ml) at room temperature for 60 min to avoid non-specific binding. Slices were incubated with glial fibrillar acidic protein, 3-nitrotyrosine, AT8 or WO2 antibodies in PBS + 0.2% Triton X-100, for 12 h at 4 °C. After washing with PBS + 0.2% Triton X-100, the sliceswere incubated with secondary antibodies coupled to HRP for 60 min at room temperature. Slices were incubated with 0.06 % diaminobenzidine in PBS for 15 min, and 0.01 % H₂O₂ was added. For all assays, n ≥ 4 animals were used, and three slices for each mouse were analyzed. Stained brain sections were photographed using an Olympus BX51 microscope (Tokyo, Japan) coupled to a Micro-publisher 3.3 RTV camera (QImaging). The luminance of the incident light and the time of exposure were calibrated to assign pixel values ranging from 0 to 255 in the RGB image (no-light to full-light transmission), which were used along all preparations. The photographs were captured with 20x and 40x objectives. The images were analyzed using theImage-J software (NIH) to measure the total area occupied by the cortex or hippocampus.

Immunofluorescence in brain slices was performed as described previously [103, 104]. Slices were washed three times in ice-cold PBS and then permeabilized for 30 min with 0.2 % Triton X-100 in PBS. After several rinses in ice-cold PBS, the slices were incubated in blocking solution (3 % bovine serum albumin in PBS) for 1 h at room temperature followed by an overnight incubation at 4 °C with mouse anti-4G8 antibody (Covance, Princeton, USA). After primary antibody incubation, the slices were extensively washed with PBS and then incubated with GFP-conjugated secondary antibodies (Molecular Probes, Carlsbad, USA) for 2 h at 37 °C. Then, slices were mounted with mounting medium on gelatin-coated slides and analyzed by fluorescence microscopy. The images were analyzed using NIH Image J software.

Memory flexibility (MF) tests were performed as previously described [50]. Briefly, a circular white pool was prepared with non-toxic white paint plus a hidden platform (diameter: 9 cm) in four quadrants; the animals were pre-trained in this pool for 60 s (s), one day before the actual testing began. The water temperature was kept between 18 and 20 °C. To acclimate the animals to the room and the swimming strategy, the pool was used without the platform. Then, animals were subjected to testing for 4 consecutive days with a maximum of 15 trials per day. Every day, the platform position in a quadrant was changed. Testing stopped when the animal reached the platform on three consecutive trials with an average of 20 s (s) or less. Data are presented as the number of trials after which animals reached the criteria.

Novel object recognition (NOR) tests were performed a day after the OF test in a 38 x 38 x 32 cm white acrylic box [105]. Animals were pre-trained to habituate to the box for two consecutive days, without objects present. For testing, animals were placed individually at the center of the box in the presence of two identical objects (old objects) for 10 min. After that period, the box and objects were cleaned with 50 % methanol solution. The animal was later (after 2 h) exposed to one of the old objects and a new object of a different shape and color than the old object, and the box and objects were cleaned again to continue with the next animal. The recognition index was calculated as the time spent exploring the new object divided by the time exploring both objects.

Anticoagulant-treated serum samples were obtained from the blood of treated and control mice. Plasma was separated through centrifugation and sent to Barnafi-Krause Laboratories (Santiago, Chile) and analyzed.

Data analysis was carried out using the Prism software (GraphPad Software Inc.). The results were expressed as the means ± standard error. For statistical analysis, normally distributed data were analyzed by one-way ANOVA with post hoc tests performed using Tukey’s test. Non-normally distributed data were analyzed by the Kruskal-Wallis test and post hoc tests were performed using Dunn’s test. Behavioral data were analyzed using the non-parametric t-test.