Germinated Brown Rice Alters Aβ(1-42) Aggregation and Modulates Alzheimer's Disease-Related Genes in Differentiated Human SH-SY5Y Cells

The pathogenesis of Alzheimer's disease involves complex etiological factors, of which the deposition of beta-amyloid (Aβ) protein and oxidative stress have been strongly implicated. We explored the effects of H2O2, which is a precursor for highly reactive hydroxyl radicals, on neurotoxicity and genes related to AD on neuronal cells. Candidate bioactive compounds responsible for the effects were quantified using HPLC-DAD. Additionally, the effects of germinated brown rice (GBR) on the morphology of Aβ(1-42) were assessed by Transmission Electron Microscopy and its regulatory effects on gene expressions were explored. The results showed that GBR extract had several phenolic compounds and γ-oryzanol and altered the structure of Aβ(1-42) suggesting an antiamyloidogenic effect. GBR was also able to attenuate the oxidative effects of H2O2 as implied by reduced LDH release and intracellular ROS generation. Furthermore, gene expression analyses showed that the neuroprotective effects of GBR were partly mediated through transcriptional regulation of multiple genes including Presenilins, APP, BACE1, BACE2, ADAM10, Neprilysin, and LRP1. Our findings showed that GBR exhibited neuroprotective properties via transcriptional regulation of APP metabolism with potential impact on Aβ aggregation. These findings can have important implications for the management of neurodegenerative diseases like AD and are worth exploring further.


Background
Alzheimer's disease (AD) is a progressive disorder of the central nervous system, with huge economic burden to individuals and their societies [1]. AD progression is accompanied by gradual loss of neurons responsible for learning and memory. Beta-amyloid (A ) plaques are often found deposited in the brain of AD patients and may be responsible for the loss of synaptic transmission and neuronal death that are characteristic of AD [2]. Even though AD has a complex pathogenesis and its etiology is not precisely understood, major findings have suggested that dysregulation of amyloid precursor protein (APP) metabolism plays a major role in the generation of neurotoxic A [2]. Accordingly, differential regulation of APP metabolic processing involving -, -, or -secretases will result in either amyloidogenic or nonamyloidogenic products with huge implications for disease outcomes [3].
In normal physiologic condition, APP is cleaved bysecretase to produce the nonamyloidogenic soluble neurotrophic fragment (sAPP ), thereby precluding the formation of the toxic A [4]. On the other hand, cleavage of APP by 2 Evidence-Based Complementary and Alternative Medicine -secretase and -secretase follows the amyloidogenic pathway, thus releasing A peptides into the extracellular matrix [5]. The -secretase components include A cleaving enzymes 1 and 2 (BACE1 and BACE2, resp.) while -secretase is formed by Presenilin, nicastrin, Aph-1, and Pen-2 [5][6][7]. Despite clear evidence that inherited mutations of APP and Presenilin genes are causal factors for early onset familial AD [8,9], the exact mechanism underlying overproduction of A in late onset sporadic AD is not known, even though sporadic AD accounts for the majority of AD cases.
As can be recalled, oxidative stress in the brain often results from inability of endogenous antioxidant defense to counter the reactive oxygen species (ROS) produced by the brain cells. Experimental models had demonstrated the JNK-dependent upregulation of BACE1 and -secretase activity as a result of oxidative stress [10] in neuroblastoma cells. Similarly, PC12 cells induced by hydrogen peroxide (H 2 O 2 ) significantly increased BACE1-promoter activity and augmented A production from APP carrying the human Swedish mutation [11]. Furthermore, disruption of APP metabolism has been demonstrated in oxidative stress [10,11], while increased levels of BACE1 have been reported in the brain region of sporadic AD patients [12]. These findings are further supported by the links between oxidative stress, altered mitochondrial function, and dysregulation of calcium hemostasis that precede A plaques development with AD neuropathology [13,14].
It has been proposed that antioxidants have huge potentials in preventing and/or delaying the development and progression of neurodegenerative diseases including AD [15]. This has driven the search for potent antioxidants with neuroprotective potentials. Germinated brown rice (GBR) has been reported to possess important biological activities including antioxidative and neuroprotective properties [16,17]. Additionally, it was reported to protect neuronal cells from oxidative stress-induced cytotoxicity, alleviate depressionlike behaviors as well as enhance learning ability, and improve memory impairment resulting from A administration [18][19][20][21]. These properties of GBR are thought to be mediated by its bioactive-rich constituents [22]. At present, there is limited knowledge on the effects of oxidative stress on normally expressed APP. Nonetheless, it is suggested to represent a near-physiological model of neuronal A generation [10]. Since H 2 O 2 is the main source of the highly reactive hydroxyl radical in the brain, in the present work we studied its effects on the levels of and processing of human APP by -, -, and -secretase and the role of the low density lipoprotein receptor-related protein (LRP) and Neprilysin, as well as the potential of GBR to attenuate such processes. This study is also the first to explore the modulation of APP metabolism in differentiated human neuroblastoma cells by GBR as well as its effect on the aggregation of A in vitro.

Germination of Brown Rice and Extraction.
BR was germinated as reported in an earlier publication [16]. Briefly, BR was sterilized in 0.1% NaOCl (1 : 5, w/v) for 30 min and incubated in 0.5% H 2 O 2 (1 : 5, w/v) for 6 h. The solvents were discarded and the BR was further incubated at 37 ∘ C for another 18 h in anoxic condition. Germination was evidenced by the growth of sprouts. After drying at 50 ∘ C, the final moisture content was between 8 and 11%. Dried GBR was ground using a grinder and extracted using ethyl acetate as previously described [23]. Briefly, GBR powder was dissolved in ethyl acetate (1 : 4 w/v). The mixtures were sonicated for 1 h. The extract was filtered through Whatman filter paper number 1 and the entire extraction process was repeated twice on the residue obtained from the filtration process. The filtrates were pooled and solvent was removed from the filtrates under reduced pressure (Rotavapor R210, Buchi, Postfach, Flawil, Switzerland). Finally, the extracts were cooled in a desiccator and kept at −80 ∘ C until further analyses.

High Performance Liquid Chromatography (HPLC-DAD).
HPLC analysis for phenolic compounds was performed using Agilent G1310A pumps (Agilent, Santa Clara, CA, USA) with a diode array detector set at wavelengths of 280 nm and 320 nm. Chromatographic separations were performed on a LUNA C-18 column (5 mm, 250 × 4.6 mm) (Phenomenex, Torrance, CA, USA). The solvent composition and gradient elution conditions used were the same as those described previously [17]. The mobile phase was composed of solvent (A) water-acetic acid (94 : 6, v/v, pH 2.27) and solvent (B) acetonitrile. The solvent gradient was 0-15% B for 40 min, 15-45% B for 40 min, and 45-100% B for 10 min. A flow rate of 0.5 mL/min was used and 20 L of sample was injected. Samples and mobile phases were filtered through a 0.22 m Millipore filter, type GV (Millipore, Bedford, MA, USA), prior to HPLC injection. Samples were analyzed in triplicate. Determination and quantification of phenolic compounds were done by comparing their retention times and UV-Vis spectral data to known, previously injected standards.
Oryzanol content was analyzed according to Azlan et al. [24] using reverse-phase HPLC. The HPLC system consisted of Hewlett-Packard (Germany) Model G1311A High Performance Liquid Chromatography connected to ALS Autoinjector Series 1100 (Hewlett-Packard, Germany). The mobile phase consisted of acetonitrile/methanol/isopropanol (50 : 45 : 5 by volume). Oryzanol components were detected at 325 nm with a Hewlett-Packard Model 1100 Series Photodiode Array Detector. Oryzanols were separated using a 250 × 4 mm id Hewlett-Packard column packed with 5 mm ODS-C18 Hypersil silica.

Transmission Electron Microscopy (TEM).
Twenty microliters (20 L) of 20 M A (1-42) was incubated with or without the presence of 1 ppm and 10 ppm GBR extract for 48 hours at 37 ∘ C. The concentrations of GBR extract (1 and 10 ppm) were chosen based on previous determination [23]. The samples were prepared on continuous carbon support films, followed by glow discharging and negative staining of 2% aqueous uranyl acetate (pH 4.5) by the single droplet procedure. The A  in the presence and absence of GBR extract was adsorbed onto the carbon film for 5 min, dried on a filter paper, and incubated with uranyl acetate solution individually and was viewed under TEM Philips HMG 400 at 10,000x and 20,000x magnifications. A representative assessment of A fibril formation was made at several positions across each EM grid, to avoid any biased or subjective data selection.

Cell
Culture. The human neuroblastoma SH-SY5Y cells were grown in complete culture medium containing mixture of Minimum Essential Eagle's Medium and Ham's Nutrient F-12 (1 : 1), which was supplemented with 10% fetal bovine serum, 1% MEM nonessential amino acids, and 50 g/mL gentamicin. Cells were maintained at 37 ∘ C under 5% CO 2 and 95% air.

MTT Assay.
For determination of neuroprotective effects of GBR extract on A (1-42) toxicity, MTT analysis was carried out. SH-SY5Y cells were seeded at a density of 2 × 10 5 cells/mL in a 96-well plate. Two days after seeding, cells were differentiated with 10 M retinoic acid for 6 days prior to treatment. Human A (1-42) was diluted to 200 M and incubated at 37 ∘ C for 48 h to produce aggregated A in order to cause toxicity to the cells. Cells were pretreated with GBR extract at 1 ppm and 10 ppm for 24 h before being exposed to aggregated A (1-42) at final concentration of 20 M for another 24 h prior to MTT analysis.

Lactate Dehydrogenase (LDH) Release
Assay. LDH release was determined using CytoTox96 Non-Radioactive Cytotoxicity Assay (Promega, Southampton, UK). In this colorimetric assay, LDH converts lactate and nicotinamide adenine dinucleotide (NAD+) to pyruvate and NADH, respectively. This initial reaction is coupled to a second reaction where diaphorase converts iodonitrotetrazolium salt and NADH to a red colored formazan compound and NAD+, respectively. Cells were seeded at a density of 2 × 10 5 cells/mL in a 96-well plate. After 2 days of seeding, cells were differentiated with 10 M retinoic acid for 6 days prior to treatment. Cells were pretreated with GBR extract (1 ppm and 10 ppm) for 24 h prior to treatment with 250 M H 2 O 2 for another 2 h [23]. After cells were exposed to H 2 O 2 with or without pretreatment with GBR, 50 L of the cell culture medium was transferred from each well of the assay plate to the corresponding well of flat bottom 96-well assay plate. Maximum LDH release was achieved by adding Lysis Buffer to the corresponding wells and 50 L of the medium was used. Substrate mix was reconstituted using Assay Buffer, from which 50 L was added to each well of the plate. The plate was covered and incubated at room temperature and protected from light for 30 minutes, after which 50 L of stop solution was added to each well of plate. Absorbance was read at 490 nm using on BioTek Synergy H1 Hybrid Reader (BioTek Instruments Inc., Winooski, VT, USA). The results were calculated using the formula % LDH release = ( LDH release in the medium max. LDH release ) * 100.
(1)  All experiments were carried out in triplicate. The results were expressed as mean ± SD. Analyses of variances (ANOVA) were carried out to evaluate the significant difference, and Tukey tests were used to determine the significance of the difference. Linear correlation was determined using Pearson correlation, and < 0.05 was considered significant.

HPLC Determination of GBR Bioactive Compounds.
The germination process of BR is known to affect its composition and functional properties [26]. The improved concentration of several bioactive compounds present in GBR is suggested to be responsible for its health beneficial effects. In the present study, the bioactive compounds shown in Table 2  Although other bioactive compounds have been reported in GBR [26], the choice of ethyl acetate in this study may have precluded their extraction. However, the presence of multiple bioactive compounds in the ethyl acetate extract may contribute to synergistic improvements in bioactivity. Moreover, the bioactivity of GBR is reported to be due to the presence of multiple bioactive compounds through the process of food synergy, and as the data here demonstrates, any bioactivity observed in the present study would be attributed to the presence of multiple compounds, as we have demonstrated severally [27]. As can be recalled, oxidative stress is strongly implicated in neuronal cell death in AD [14]. Furthermore, the bioactive compounds in the ethyl acetate extract of GBR are mostly phenolics, which have been shown to have potent antioxidant effects, and as such can counter the effects of oxidative stress [28].

Antiamyloidogenic Properties of GBR.
Due to the limited capability of neuron cells to regenerate, it is vital to prevent cell death through regulation of oxidative stress andamyloid production, formation, fibrillogenesis, and deposition in order to prevent their damaging effects on brain cells. As such, compounds or drugs that are able to regulate these processes have huge potentials in the management of AD. Accordingly, anti-inflammatory drugs [29] and antioxidative compounds [30] have been shown to prevent and slow down the onset of AD, through possessing antiamyloidogenic properties. Additionally, recent studies have shown that epigallocatechin gallate (EGCG) [31] and cannabinoids [32] can reduce the formation of A fibrils and aggregates formation, while a hybrid of the AD drug tacrine, tacrine-6-ferulic acid dimer, was reported to alter the conformation of A and block A -induced cell death as its possible neuroprotective action [33]. In this study, ethyl acetate extract of GBR also altered A formation (Figure 1). In the absence of GBR extract, A (1-42) appeared negatively stained and showed characteristic fibrillary amyloid morphology. However, in the presence of GBR (1 and 10 ppm), A (1-42) showed amorphous positively stained deposits, resembling remnants of amyloid fibril morphology [34]. In fact, at 10 ppm of GBR, A (1-42) formed distinctly amorphous deposits that appeared very different from those present when GBR was not added. This is consistent with findings on the ability of natural product compounds to inhibit A aggregation [15]. A (1-42)-Induced Cell Death. GBR extract ameliorated the aggregation of A (1-42); thus we further assessed its protective effects on A (1-42) toxicity in SH-SY5Y cells. In this study, we incubated aggregated A (1-42) with or without the presence of 1 ppm and 10 ppm of GBR extract. Figure 2 shows that incubation of SH-SY5Y cells with 20 M of A (1-42) decreased the cell viability to 55.5% as compared to untreated control. Pretreatment of the cells with GBR extract at 1 ppm and 10 ppm was found to substantially increase the cell survival to 59.82% and 77.02%, respectively. There have been numerous reports on natural compounds with neuroprotective effects against A -induced cell death, many of which exhibited antioxidant properties   used in this study has been previously reported to regulate cell survival and apoptosis pathway through the modulation of MAPKs (JNK, ERK, and p38), p53, AKT, and NF- [23].

GBR Attenuated H 2 O 2 -Induced Neurotoxicity.
To study the neuroprotective effects of GBR on oxidative stressinduced toxicity in SH-SY5Y cells, the cells were pretreated with GBR at 1 and 10 ppm for 24 h prior to H 2 O 2 insult. Figure 3 shows that GBR reduced the release of LDH from the cells, which may have indicated the amelioration of oxidative stress-induced neurotoxicity in differentiated SH-SY5Y neuroblastoma cells in comparison to nontreated cells. LDH is a stable cytoplasmic enzyme which is released from the cells due to loss of plasma membrane integrity. In an in vitro model of cell apoptosis, an increase in LDH activity in the culture supernatant signifies an increase in dead cells or disintegration of plasma membrane. Thus, the measurement of cytosolic LDH release from the cells is commonly used to indicate the loss of membrane integrity especially in apoptotic cells [37].
From the determination of ROS production using DCFH-DA as represented by Figure 4, it was observed that H 2 O 2 increased the level of intracellular ROS throughout the incubation time in comparison to the untreated normal control cells. Cells pretreated with GBR prior to the addition of H 2 O 2 , nevertheless, exhibited lower levels of ROS, in corroboration of the GBR-induced attenuation of neurotoxicity as evidenced by reduced LDH release. This is in agreement with our previous findings by which an ethyl acetate extract of GBR possessed high antioxidative capacity and regulated several antioxidant genes such as superoxide dismutases (SODs) and catalase [23]. Oxidative stress results from imbalance between the levels of antioxidants and prooxidants and has been the target in the management of many chronic diseases. In the brain, increased levels of oxidative damage are implicated in aging as well as the processes leading up to neurodegenerative diseases like AD. The attenuation of ROS by GBR extract suggests its potential for attenuating oxidative stress-induced neurotoxicity.

Transcriptional Regulation of AD-Related Genes by GBR.
Oxidative stress plays vital roles in the development of A peptides through the regulation of APP metabolism [10,14]. The amyloid-cascade theory hypothesizes that proteolytic cleavage of APP may result in neurotoxic -amyloid peptides, which are the hallmark of AD [2]. However, as can be recalled, depending on the site of cleavage of the APP molecule by -, -, or -secretases, the metabolic processing of APP may be in favor of amyloidogenic or nonamyloidogenic pathways [3]. Unlike in the nonamyloidogenic pathway which precludes the formation of neurotoxic A , the amyloidogenic pathway produces aggregated, fibrillary complexes termed senile plaques, which were found to be abundant in AD [38].
In the present study, we analyzed the effects of H 2 O 2 on AD-related genes as well as the ability of GBR to modulate their transcription ( Figure 5). Expression levels of Presenilin 1, Presenilin 2, A cleaving enzymes (BACE1, BACE2), ADAM10, LRP, Neprilysin, and APP were determined. Presenilin 1, Presenilin 2, BACE1, and BACE2 were found to be upregulated as a result of H 2 O 2 -induced oxidative stress in comparison with the untreated normal cells. Furthermore, pretreatment with GBR attenuated the H 2 O 2 -induced transcriptional changes on Presenilin 1, BACE1, and BACE2, but not that of Presenilin 2. Presenilin 1 and Presenilin 2 are important -secretase enzymes with protease activities that cleave APP to produce A fragments, and their deficiency or inhibition may reduce A production [39]. This has been suggested as a potential target for antiamyloidogenic approach in AD research [40]. On the other hand, BACE1 and its homolog BACE2 form part of the -secretases that are involved in the processing of APP to produce A [41]. BACE2 was reported to be responsible for the release of A in mutant transfected cells [42] and was able to cleave APP at the secretase cleavage site, in an in vitro model [43]. BACE1, on the other hand, has been established as a more dominant inducer of APP cleavage leading to A formation [42,44].
Moreover, in a transgenic mice model bearing APP mutation, overexpression of BACE increased the processing of APP and contributed to high levels of A [45]. This is consistent with findings of elevated BACE activity in the brains of AD patients with high load of A plaques [46].  Although accumulated A can be toxic, the sAPP is protective towards neuronal cells. In the present study, H 2 O 2induced oxidative stress did not promote APP expression in comparison to the normal control ( > 0.05). Interestingly, we found that pretreatment with GBR (10 ppm) upregulated APP expression. There are reports of the neuroprotective effects of soluble APP mediated by -secretase activity in PC12 cells [47]. Our study also showed that the expression of ADAM10, a part of -secretase, was downregulated under oxidative stress condition, while GBR increased its expression similar to what was observed in the normal control. This is in agreement with previous findings that support the neuroprotective effects of ADAM10 [48], thus suggesting the potential role of -secretases, particularly ADAM10 in the therapeutic development of AD.
Similarly, H 2 O 2 treatment resulted in reduced expression of Neprilysin and LRP genes when compared to the control ( < 0.05). Treating the cells with GBR prior to the H 2 O 2 insult, nevertheless, resulted in an increased expression of Neprilysin and LRP genes. Moreover, the expression of LRP in GBR-treated cells was found to be dose-dependent, implying that GBR is able to increase the transport of A from the brain into the bloodstream. Aside from its involvement in cholesterol metabolism and transport, LRP binds several ligands including ApoE, APP, and A [49] and has been demonstrated to regulate the clearance of A from the brain into the bloodstream [50]. Accordingly, A transport clearance across the blood-brain barrier (BBB) in mice was reported to decrease with aging [51], the basis of which may be the loss of LRP transport and hence the accumulation of A in blood vessels and perivascular brain tissue, as demonstrated in aged individuals [52]. Similarly, brain vascular dysfunction, reduced cerebral blood flow, BBB leakage, and accumulation of multiple blood-derived vasculotoxins and neurotoxins following injury to brain endothelial cells have recently been shown to collectively contribute to AD pathogenesis [53]. Moreover, A binds to brain endothelial LRP1 at the abluminal side of the BBB and clears into the peripheral circulation [54] and has been demonstrated to induce oxidative stress and mediates BBB changes and brain capillary dysfunction [55]. Conversely, antioxidants in GBR extract may protect brain vasculature and improve LRP1 mediated clearance of A , thereby attenuating the effects of A -induced toxicity. On the other hand, Neprilysin is reported to contribute towards the degradation of A [56]. In the hippocampal region of AD patients, it was demonstrated that Neprilysin mRNA and protein levels were significantly lower than in control subjects [57]. Accordingly, deficiency in Neprilysin may result in limited degradation of both endogenously produced A and exogenously administered A [58] and could eventually promote the accumulation of A in AD. Thus, regulation of LRP and Neprilysin genes by GBR suggested that the neuroprotective effects of GBR in reducing the accumulation of A are partly mediated by increased degradation and clearance of A from the brain tissue.

Evidence-Based Complementary and Alternative Medicine
Based on the transcriptional regulation of AD-related genes by GBR, in neuronal cells that have been exposed to H 2 O 2 , we propose that the mechanistic basis for the neuroprotective effects of GBR may be through multiple pathways ( Figure 6) including the regulation of APP metabolism genes as well as apoptosis pathway involving JNK, ERK, and p53, as we have previously demonstrated [23]. The antiamyloidogenic effects of GBR may be due to its bioactive-rich content with potent antioxidant effects that are able to counter H 2 O 2induced neurotoxicity.

Conclusion
In the present study, we demonstrated that GBR modified the structure of A (1-42) when incubated together and also protected neuronal cells from H 2 O 2 -induced neurotoxicity (reduced LDH release) partly through reduced oxidative stress (lowered ROS production). We also showed that transcriptional regulation of AD-related genes by GBR tended towards neuroprotection. As such, we propose that GBR not only possesses the potential to manage oxidative stress and oxidative stress-induced A formation, but also can reduce the aggregation of A as well as increase its clearance and degradation. Further work involving in vivo model to assess the effectiveness of GBR at physiological level is of importance and is therefore indicated.