Possible Involvement of Standardized Bacopa monniera Extract (CDRI-08) in Epigenetic Regulation of reelin and Brain-Derived Neurotrophic Factor to Enhance Memory

Bacopa monniera extract (CDRI-08; BME) has been known to improve learning and memory, and understanding the molecular mechanisms may help to know its specificity. We investigated whether the BME treatment alters the methylation status of reelin and brain-derived neurotropic factor (BDNF) to enhance the memory through the interaction of N-methyl-D-aspartate receptor (NMDAR) with synaptic proteins. Rat pups were subjected to novel object recognition test following daily oral administration of BME (80 mg/kg) in 0.5% gum acacia (per-orally, p.o.; PND 15–29)/three doses of 5-azacytidine (5-azaC; 3.2 mg/kg) in 0.9% saline (intraperitoneally, i.p.) on PND-30. After the behavioral test, methylation status of reelin, BDNF and activation of NMDAR, and its interactions with synaptic proteins were tested. Rat pups treated with BME/5-azaC showed higher discrimination towards novel objects than with old objects during testing. Further, we observed an elevated level of unmethylated DNA in reelin and BDNF promoter region. Up-regulated reelin along with the splice variant of apolipoprotein E receptor 2 (ApoER 2, ex 19) form a cluster and activate NMDAR through disabled adopter protein-1 (DAB1) to enhance BDNF. Observed results suggest that BME regulate reelin epigenetically, which might enhance NMDAR interactions with synaptic proteins and induction of BDNF. These changes may be linked with improved novel object recognition memory.


INTRODUCTION
Long-term memory formation requires fine tuned cellular signaling coordination with transcriptional and translational regulations of gene expression (Hawk and Abel, 2010). Epigenetic mechanisms specifically control transcription through many ways, among which DNA methylation being common (Levenson et al., 2006;Blaze and Roth, 2013;. This process involves DNA methyltransferases (DNMTs) that mediates de novo methylation within DNA and alters the chromatin structure (Shilatifard, 2006;Lubin, 2011;Nabel and Kohli, 2011); which directly controls transcription. Several studies have reported that DNA methylation/demethylation in a specific promoter region of reelin and brain-derived neurotrophic factor (BDNF) dictates the transcriptional activity thereby critically regulating synaptic plasticity, learning, and memory (Miller and Sweatt, 2007;Levenson et al., 2008;Lubin et al., 2008;Mizuno et al., 2012;Sui et al., 2012). Reelin is a large secreted glycoprotein richly expressed at hippocampus by a subset of γ-amino butyric acid (GABA)-ergic interneurons (Pesold et al., 1998;Abraham and Meyer, 2003;Ramos-Moreno et al., 2006).
Bacopa monniera extract has been traditionally used in Ayurvedic medicine to improve learning and memory. B. monniera extract contains triterpene saponins, which have been named bacosides and bacopasaponins. The major chemical entity shown responsible for neuropharmacological effects of B. monniera are bacoside A (bacogenins A1, A2, A3, and A4) and bacoside B (Pubmed C ID: 53398644); the latter differs only in optical rotation and may probably be an artifact produced during the process of isolating bacoside A (Chatterji et al., 1963(Chatterji et al., , 1965Basu et al., 1967). CDRI-08 mentioned as BME in this article is a bacoside-enriched standardized extract of B. monniera (CDRI-08, contains 55 ± 5% bacosides). Earlier studies demonstrate that CDRI-08 significantly improves the cognitive performance in healthy human participants (Downey et al., 2013;Benson et al., 2014); and in elders and patients with neurodegenerative disorder (Barbhaiya et al., 2008;Calabrese et al., 2008;Stough et al., 2013). Bacosides present in the CDRI-08 are non-polar glycosides, which possibly cross the blood-brain barrier (BBB) by lipid mediated passive diffusion (Pardridge, 1999); and its biodistribution in brain has been confirmed by radiopharmaceuticals (De et al., 2008). Earlier, in rat model, we found that oral treatment of BME elevates the level of serotonin (5-hydroxytryptamine, 5-HT), activates 5-HT 3A receptor (Rajan et al., 2011), and regulates cyclic adenosine monophosphate (cAMP) response element binding (CREB) protein through microRNA (miR)-124 (Preethi et al., 2012). Subsequently, we reported that BME regulates extracellular signal-regulated kinase (ERK)/CREB cascade and BDNF through regulation of histone acetylation and protein phosphatases to improve hippocampal memory (Preethi et al., 2014). In the present study, first we tested whether the treatment of BME/DNMT inhibitor (5-azacytidine and 5-azaC) improves novel object recognition. Second, we examined whether BME/5-azaC treatment alters methylation status of reelin and BDNF, and its effect on NMDA (NR2A) receptor interaction with synaptic proteins.

Animals
Wistar rat pups both male and female (Rattus norvegicus; B.wt: 19.6 ± 0.6 g on PND-15) were housed in a standard laboratory cage (43 × 27 × 15 cm) with paddy husk as a bedding material. Rat pups were maintained at animal house under controlled environmental condition [12-h light/dark cycle (7:00-19.00); temperature: 22 ± 2 • C; humidity: 50 ± 5%)] with ad libitum access to food and water. Protocols for animal use were performed following the guidelines of Institutional Animal Ethics Committee (BDU/IAEC/2014/OE/08/Dt. 18.03.2014), ensuring that number of animals and pain was kept to a minimum.

Novel-Object Recognition Test
The apparatus used for novel-object recognition (NOR) test was conducted in circular open-field board (82 cm diameter) made of wood, coated black surrounded by a wall of 32 cm height (Barbosa et al., 2013). Two sets of objects (Cuboids and Spherical objects) were used. The cuboids (13 × 10 × 10 cm) FIGURE 1 | Effect of BME/5-azaC on training at novel object recognition test. (A) Con/BME/5-azaC groups spent equal time and showed no discrimination for objects during training 1, (B) and training 2. Data were shown as mean ± SEM, there was no significant difference between comparisons (Con verses BME; Con verses 5-azaC; BME verses 5-azaC).
were made of plaster of paris and painted with blue color (nontoxic children water color). The plastic balls (radius: 26 cm) were green in color and filled with sand to ensure that animals cannot displace them. The apparatus and objects were cleaned with 75% ethanol after each behavior session. The sessions were recorded by a video camera (Sony megapixel 12.1, Full HD 1080) placed at a distance of 150 cm above the apparatus. In novel object recognition test, rodents' innate habit of exploring novel object in their environment has been utilized to evaluate their memory (Ennaceur and Delacour, 1988). Exploration time was scored as when the rat nose was touching the object or orienting its head towards the object within a distance of 1 cm. The relative exploration time was recorded and expressed by a discrimination index [D.I. = (t novel − t old )/(t novel + t old ) × 100%] as reported earlier (Stefanko et al., 2009). Mean exploration times were calculated and the discrimination indexes between groups were FIGURE 2 | Effect of BME/5-azaC on novel object recognition test. (A) Con/BME/5-azaC group rats spent more time(s) around novel object than old objects. Whereas, BME/5-azaC groups spent significantly more time around novel object compared to control group. (B) BME/5-azaC treated groups displayed significantly higher DI (%) for novel object compared to control group. Data were shown as mean ± SEM, asterisk indicates significant difference ( * * P < 0.01; * * * P < 0.001). Comparisons between groups are represented as a = Con verses BME; b = Con verses 5-azaC; and c = BME verses 5-azaC; $ = Con old verses novel; # = BME old verses novel; @ = 5-azaC old verses novel.
compared. Rat was allowed to freely explore the experimental area in groups (n = 3) on PND-28, 29 and individually on PND-30 for 10 min. On PND-31, all animals were trained for two sessions (5 min each session) with an interval of 1 h. During training, rats were exposed to four identical objects. On PND-32 during retention session, two new objects were placed in the position of any two old objects and the exploration behavior of each rat was recorded for 5 min (Barbosa et al., 2013).

Tissue Collection, RNA and Protein Isolation
After the NOR test, animals representing each group (n = 6) were euthanized, and the hippocampus tissue was dissected as described by Glowinski and Iversen (1996) for the preparation of RNA and protein. Total RNA was isolated from hippocampus tissue samples using TRIzol (Merck Specialties Pvt. Ltd., India), according to the manufacturer's instructions and stored with RNase inhibitor (Merck Specialties Pvt. Ltd, India) at −80 • C. The hippocampus tissue samples were homogenized in icecold lysis buffer (150 mM NaCl, 50 mM Tris-HCl pH 7.5, 5 mM EDTA, 0.1% v/v NP-40, 1 mM DTT, 0.2 mM sodium orthovanadate, 0.23 mM PMSF) and 10 µl/ml protease inhibitor cocktail (Sigma-Aldrich, USA). The homogenates were kept on ice for 30 min and then centrifuged at 10,000 × g for 30 min at 4 • C. The clear supernatants were collected in a fresh tube and centrifuged again at 12,000 × g for 15 min at 4 • C. The final supernatants were collected and stored as aliquots, in order to avoid repeated freeze-thaw of the samples and were stored at −80 • C. The concentration of each protein sample was quantified by measuring the absorbance at 595 nm using Biophotometer plus (Eppendorf Inc., Germany).

Western Blot Analysis
Equal concentrations (30 µg) of proteins were resolved on 9% SDS-polyacrylamide gel. The separated proteins were transferred electrophoretically onto polyvinylidene difluoride (PVDF) membrane (Merck Millipore, India) using a semi-dry western apparatus (SD 20; Cleaver Scientific Ltd, UK). The membranes were blocked and incubated at 4 • C for 9 h with one of the following specific primary antibodies (Santa-Cruz  3 | Effect of BME/5-azaC on regulation of DNA methylation in the reelin promoter and reelin expression. BME/5-azaC treatment significantly (A) increased unmethylated DNA, (B) and decreased methylated DNA in reelin promoter region relative to control group. (C) reelin mRNA was significantly higher in BME/5-azaC treated groups compared to control group. Data were shown as mean ± SEM, asterisk indicates significant difference ( * P < 0.05; * * P < 0.01; * * * P < 0.001). Comparisons between groups are represented as a = Con verses BME; b = Con verses 5-azaC; and c = BME verses 5-azaC.
ChemiDoc XR + System (Bio-Rad Laboratories, Inc., USA), and optical density of trace quantity for each band was measured using Image Lab 2 software (Bio-Rad Laboratories, Inc., USA). The trace quantity of each bands were normalized with β-actin bands, following that fold changes were calculated by dividing normalized values of BME/5-azaC groups by control groups.

Co-immunoprecipitation
Spin columns were packed with AminoLink Plus Coupling resin slurry (50 µl) and washed with provided buffer, then 20 µg of anti-rabbit NR2A antibody (Cat #4205, adjusted to 200 µl volume with 1× coupling buffer) was added in the column and incubated on a rotator at room temperature for 2 h to immobilize antibodies. For Co-IP, 200 µg of protein lysates from each groups Con/BME/5-azaC were immunoprecipitated with NR2A antibody in immobilized spin columns overnight at 4 • C. Immunoprecipitated protein was eluted using elution buffer. The complete Co-IP experiment performed following the instructions provided with the kit (Catlogue no. 26149, Pierce Co-Immunoprecipitation Kit, Thermofisher, IL, USA). Equal volume of NR2A precipitated protein from each sample elutes was analyzed by immunoblotting using a SFK (SC-7020, 1:300) and PSD-95 (SC-28941, 1:200) antibody. Immunoblotting was performed following the procedure reported in the western blots section.

Statistics
Data were presented as a mean ± standard error of the mean (SEM) and plotted with KyPlot (ver 1.0) for graphical representation. For behavioral analysis (NOR), multivariate ANOVA was performed to test the effect of factors (training sessions, objects) and their interactions during training. Oneway analysis of variance (ANOVA) and Two-way analysis of variance (Two-way ANOVA) were used to assess the significance between groups and groups × objects interactions during testing respectively. Post hoc Bonferroni test was used to examine the difference between groups. Association between the recognition memory (DI) and methylation/unmethylation DNA was estimated using Pearson correlations (SPSS, ver.15). For expression data, One-way analysis of variance (ANOVA) was used to observe the significance between groups. Differences were considered significant if P < 0.05.

Effect of BME on Reelin Mediated NMDA in Phosphorylation of DAB1
Reelin binding to ApoER 2 (ex 19) and VLDLR results in activation of DAB1. To expand our understanding, we tested FIGURE 8 | Effect of BME/ 5-azaC on DNA methylation of BDNF promoter region. (A) BME/ 5-azaC treatment increased unmethylated DNA in BDNF promoter region, (B) but did not alter the methylated BDNF. Data were shown as mean ± SEM, asterisk indicates significant difference ( * * * P < 0.001). Comparisons between groups are represented as a = Con verses BME; b = Con verses 5-azaC; and c = BME verses 5-azaC.

DISCUSSION
Earlier studies demonstrated that DNA methylation/demethylation involves in the regulation of activity dependent neuronal gene expression (Ming and Song, 2011;Baker-Andresen et al., 2013) and synaptic plasticity (Miller and Sweatt, 2007;Feng et al., 2010). A family of DNMTs (DNMT1, 3a, 3b) can catalyze the process of DNA methylation (Sweatt, 2009). The expression pattern of DNMTs (DNMT1, 3a, 3b) differ in different brain regions following behavioral training and exhibits different behavioral phenotype (Feng et al., 2010;Miller et al., 2010;. In fact, DNMT inhibitors (5-azaC) are known to alter learning and memory/memory consolidation (Miller and Sweatt, 2007). Considering the role of DNA methylation in memory formation, first we showed the effect of BME in novel object recognition by comparing with DNMT inhibitor, 5-azaC. Using NOR test, we found that BME effectively enhanced the NOR memory, this suggested that BME treatment improved the process of recollection and familiarity identification (Squire et al., 2007). It has been shown that methylation status alters rapidly and dynamically within hippocampus (Miller and Sweatt, 2007). The level of DNMT 3a was also significantly elevated following contextual training in NOR test, but its conditional knockout impaired NOR memory . Observed behavioral results suggested that BME/5-azaC treatment might optimally regulate the DNMT 3a level at this age of rats and hence improve NOR memory.
Following the improvement in behavioral phenotype; we sought to correlate this with methylation status. As quantified by RT-PCR, a higher level of unmethylated and lower level of methylated reelin DNA was estimated in BME and 5-azaC treated groups when compared to control. Supporting this, earlier studies have shown that 5-azaC treatment reduces the methylated DNA in reelin promoter (Miller and Sweatt, 2007;Sui et al., 2012). Consistent with the demethylation status, the level of reelin mRNA was significantly elevated in both BME and 5-azaC treated rats. The positive correlation between recognition memory (DI) and unmethylated reelin DNA in BME/5-azaC group suggested that BME possibly regulated methylation to enhance memory. These results are in line with earlier reports on methylation status and expression of reelin during neural plasticity (Levenson et al., 2006;Miller and Sweatt, 2007). In addition, observed behavioral data may be linked with the up-regulated reelin expression (Rogers et al., 2011) and increased dendritic spine density in hippocampus (Vollala et al., 2011).
Reelin exerts its effects through ApoER 2 ex 19 receptor (Weeber et al., 2002;Beffert et al., 2005). Interestingly, BME/ 5-azaC treatment up-regulated the ApoER 2 (exon 19) variant and down-regulated ApoER 2( ) transcript. The up-regulated ApoER 2 exon 19 attaches to postsynaptic density protein PSD-95 and forms a signaling complex with NMDARs Hoe et al., 2006). However, to execute the neuronal transmission, NMDAR should be activated by p-DAB1 (Chen et al., 2005;Beffert et al., 2006). Consistently, BME/5-azaC treatment enhanced the level of p-DAB1. In process p-DAB1 activates Src kinase family (SFK) through phosphorylation, which in turn leads to phosphorylation of NMDAR subunit (NR2) (Ballif et al., 2003;Bock and Herz, 2003;Strasser et al., 2004;Chen et al., 2005;Mota et al., 2014). However, synaptic strength is determined by the ratio of NMDAR subunits (NR2A/NR2B) (Quinlan et al., 2004;Lebel et al., 2006;Brigman et al., 2008;Levenson et al., 2008). In this study, the level of NR2A and NR2B subunits and the estimated NR2A/NR2B ratio were higher in BME and 5-azaC groups compared to control. In addition, the co-immunoprecipitation studies further substantiate our hypothesis and showed enhanced interaction of NR2A subunit with SFK and PSD-95 in BME and 5-azaC groups compared to control group. This complex possibly enhances the Ca 2+ influx through the release of glutamate, and then induces LTP and LTM (Chen et al., 2005;Qiu et al., 2006). Further, the activated NMDAR is in turn found to mediate events associated with methylation of BDNF and its subsequent regulation of transcripts and translation in the hippocampus during LTM. We found that level of unmethylated BDNF DNA was higher in BME and 5-azaC groups compared to control. Subsequently, the level of Bdnf (exon IV) mRNA and protein were found to be higher in BME and 5-azaC groups compared to control. There could be a possible correlation of unmethylated BDNF with its expression. However, we did not find a significant difference in the level of methylated form (Martinowich et al., 2003;Munoz et al., 2010). Interestingly, we observed a positive correlation between DI with unmethylated BDNF DNA in BME/5-azaC group. BME/5-azaC possibly regulate the exon specific expression by controlling methylation process (Sales et al., 2011), and the expression of BDNF in hippocampus enhances memory (Bekinschtein et al., 2007(Bekinschtein et al., , 2008aLubin et al., 2008;Mizuno et al., 2012).
Taken together, we have shown the effect of BME in the regulation of DNA methylation using known reelin-dependent NMDAR-BDNF signaling. Our results show that the BME treatment dynamically controlled the methylation of reelin and subsequent alternative splicing of ApoER 2. Further, it might be implicated in the activation and interaction of NMDAR with synaptic protein (PSD-95) to induce BDNF. This mechanism might contribute to the modulation of synaptic plasticity and thus to enhance learning and memory.