Eight-month-old male APP/PS1 Tg mice and littermate wild type (WT) mice were obtained from the Nanjing Biochemical Research Institute of Nanjing University and housed in environmentally controlled conditions (room temperature at 22 ± 1°С, 12-h light/dark cycle) with access to standard food and water ad libitum. All experiments were approved and conducted in accordance with the guidelines of the Care and Use of Laboratory Animals approved by the Ethics Committee for Animal Care and Treatment at Beijing University of Chinese Medicine (BUCM-4-2018080101-3007).

All the 8-month-old Tg mice were randomly divided into three groups (n = 9 per group): Tg model group, rhubarb administration group (TgR), and positive drug group (TgP). Littermate wild type mice (n = 9) were used as the wild type control group (WT). Mice in the TgR group were given a rhubarb decoction by gavage at 0.91 g kg⁻¹ every day for 60 consecutive days. Mice in the TgP group were given donepezil solution at 1.5 mg kg⁻¹ daily for 60 days. The Tg model group and the WT control group were given 1 ml normal saline every day. The fecal samples of all mice were collected from 8-month-old mice after 3 days of acclimation. After 30-day-treatment, fecal samples of all mice were collected as 9-month-old samples. And after feeding for 60 days, fecal were collected as 10-month-old samples.

The brain tissues of mice in various treatment groups were subjected to Congo Red staining to visualize amyloid plaques. Paraffin-embedded tissue sections were routinely dewaxed into water. The slices were immersed in alkaline solution for 20 min and washed with water. Alkaline Congo Red solution was soaked for 20 min and washed with water. The sections were immersed in alkaline solution for differentiation, observed under a microscope until the tissue staining contrast was clear, then washed with water to stop staining. Harris hematoxylin was applied for 5 min and washed off. Afterward, 1% hydrochloric acid ethanol solution was added for differentiation for a few seconds and the sections were washed with water for 15 min. The sections were dehydrated with gradient ethanol, hyalinized with xylene, and sealed with neutral gum.

The brain tissues of mice were fixed with neutral formalin, dehydrated with ethanol, and removed with xylene. The brain tissues were embedded in paraffin and sectioned at 5 μm. The paraffin sections were immersed in 100% xylene twice for 10 min each time and soaked in 95, 80, and 70% alcohol for 3 min in turn. Finally, phosphate-buffered saline (PBS) and distilled water were used to rinse the sections three times, for 2 min each time. The sections were soaked in hematoxylin solution, stained for 10 min, and washed with running water until the water was clear with no purple color. The cells were differentiated with 75% hydrochloric acid ethanol for 3 s (three times) and washed with running water. The nuclei were confirmed to have turned blue under the microscope. Eosin staining was performed for 5 min. The stained sections were dehydrated with pure alcohol and washed with xylene until the cut was transparent and sealed with neutral gum.

Tissue sections of the whole brain were embedded with paraffin and prepared for immunohistochemical staining. After deparaffinization and rehydration in graded alcohols, antigen retrieval was performed with citrate buffer for 10 min at 90°С. Next, the sections were gradually cooled at room temperature to block endogenous peroxidase activity. The sections were then washed thrice with PBS and blocked with goat serum for 10 min at room temperature. After the goat serum was removed, the sections were incubated with mouse-anti Iba-1 at 4°С overnight, followed by intubation with biotin-labelled goat anti-mouse secondary antibody for 10 min and streptomycin anti-biotin peroxidase for 10 min at room temperature. Then, the sections were labelled with 3,3′-diaminobenzidine, followed by hematoxylin counterstaining. Lastly, after washing, the sections were dehydrated through gradients of ethanol and xylene. PBS displacement of the primary antibody was used as the negative criterion.

Hippocampus samples from the brain were homogenized and the concentrations of Aβ, Aβ₄₂, IL-1β, IL-18, and TNF-α were measured using an ELISA kit (BlueGene Biotech, Shanghai, China) according to the instructions. The standard curve was established and used to calculate the levels of Aβ, Aβ₄₂, IL-1β, IL-18, or TNF-α in the tissues. The obtained values were corrected for the wet weight of the brain sample and expressed as µg/mg.

Hundred-milligram fecal samples from each mouse were weighed and placed in a 2 ml centrifuge tube. Three times the volume of 50% methanol aqueous solution was added for 5 min, and centrifuged at 13709 × g for 10 min. The supernatant was dried with nitrogen and 100 μl of 50% methanol aqueous solution was added for re-dissolution. After shaking for 30 s and standing for 10 min, the supernatant was centrifuged at 13709 × g at 4°C for 10 min and the supernatant was placed in an injection vial.

UPLC analysis was carried out on a Waters Acquity™ UPLC system (Waters Corporation, Milford, MA, United States) consisting of a binary solvent system, an autosampler, and a column temperature controller. Chromatographic separation was carried out on an ACQUITY UPLC^® HSS T3 column (2.1 mm × 100 mm, 1.8 μm, Waters, United Kingdom). The mobile phase was composed of eluent A (0.1% formic acid in water) and eluent B (acetonitrile). The line gradient program was optimized as follows: 0–0.5 min, 1% B; 0.5–4 min, 1–20% B; 4–8 min, 20–100% B; 8–9 min, 100% B; 9–9.5 min, 100–1% B; 9.5–11 min 1% B. The column temperature was set at 45°C and the sample chamber temperature was set at 4°C. The mobile phase flow rate was set at 0.3 ml/min and the injection volume was 3 μl for each run.

MS analysis was carried out with a Waters SYNAPT G2-SI MS system (Waters, United States) equipped with an electrospray ionization source. The analysis was performed in the positive and negative ion electrospray modes. The source parameters were set as follows: capillary voltage, 3.0 kV; cone voltage, 28 V; source temperature, 100°C; desolvation temperature, 400°C; the cone gas flow, 35 L/h; desolvation gas flow, 800 L/h. The low collision energy was 6 eV and the high collision energy was 10–65 eV. Mass spectra were recorded across the m/z range of 50–1200 and 3D data were collected in the continuum mode. The mass spectrometry data were acquired and processed with Waters MassLynx V4.1 software.

The UPLC-quadrupole time-of-flight mass spectrometry (UPLC-Q-TOF/MS) data for the fecal samples were imported into Progenesis QI v1.0 (Nolinear Dynamics, Newcastle, United Kingdom) for peak selection and alignment. After normalizing the data using the total ion intensity, a data matrix of interesting features containing the retention time, m/z value, and normalized peak intensity was imported into Metaboanalysis 4.0 for principal component analysis (PCA) and partial least-squares discriminant analysis (PLS-DA) and further confirmed using analysis of variance (ANOVA). The differences in the trends were processed with an unsupervised PCA method. Supervised PLS-DA was used to search for interesting biomarkers. Then, the peak height intensities of the differential metabolites were compared with t-tests using statistical software to confirm the biomarker alterations between the WT and Tg groups at the age of 8, 9, and 10 months. A p value < 0.05 was set as the threshold. As a small sample set, a pooled quality control (QC) sample containing equal aliquots of all the samples was run at the beginning of the sample queue for column conditions and injected at regular intervals and the end of the run (Want et al., 2010). Mass data acquisition was performed for the evaluation of the sensitivity and stability of instrument performance with regard to mass accuracy, retention time stability, and the coefficient of variation (CV). All samples were kept at 4°C during the analysis.

Metabolite peaks were assigned based on MS/MS analysis using the MassFragment™ application manager (Waters Corp., Milford, United States). After applying chemically intelligent peak-matching algorithms, the molecular composition of each metabolite and its fragments were structure-matched using available biochemical databases such as HMDB (http://www.hmdb.ca/), Kyoto Encyclopedia of Genes and Genomes (KEGG; http://www.genome.jp/kegg/), LIPIDMAPS (http://www.lipidmaps.org/), and Chemspider (http://www.chemspider.com). The identified metabolites were subjected to KEGG pathway mapping for metabolic pathway analysis. The KEGG database contains a collection of manually curated pathway maps from genomics, transcriptomics, proteomics, and metabolomics and provides molecular interactions and reaction networks.

The total bacterial DNA was extracted from the fecal samples of mice in each group with the E.Z.N.A ™. Stool DNA kit (D4015, Omega Biotek, United States), and the procedures were carried out according to the manufacturer’s instructions. First, fecal samples were removed from a −80°C refrigerator, 200 mg of feces from each sample was measured into a 1.5 ml sterile centrifuge tube, and genomic DNA was detected with 2% agarose gel electrophoresis. For detection of the bacteria Marvinbryantia (Desai et al., 2016) and Erysipelatoclostridium (Zakham et al., 2019), extracted bacterial DNA was subjected to RT-PCR using the CFX Connect real-time PCR system (Bio-rad Laboratories, Hercules California, United States). PCR was performed with a 10 μl sample containing 5 μl SG Green qPCR Mix (with ROX Q1002, SinoGene), 0.2 μl 20 μm upstream and downstream primers, 1 μl bacterial genomic DNA template, and 3.6 μl deionized water. In the blank control, the template DNA was replaced with deionized water. After the PCR reaction, the melting curve temperature was set at 60–95°C and increased by 0.5°C/s. For detection of the bacteria Bacteroides (Haugland et al., 2010) and norank_f_Ruminococcaceae (Jiang et al., 2017), extracted bacterial DNA was subjected to RT-PCR using the LineGene 9600 Plus real-time PCR system (Bioer Technology, Hangzhou, China). PCR was performed with a 20 μl sample containing 10 μl ChamQ SYBR Color qPCR Master MiX (Vazyme Biotech Co., Ltd, Nanjing China), 0.4 μl 5 μm upstream and downstream primers, 2 μl bacterial genomic DNA template, and 7.2 μl deionized water. In the blank control, the template DNA was replaced with deionized water. After PCR reaction, the melting curve temperature was set at 60–95°C and increased by 0.5°C/s. Table 1 shows the sequences for each primer set, which targeted the 16S rRNA genes for each bacteria group. Each standard curve was prepared based on the cell numbers measured using a bacterial counting chamber with each strain indicated. DNA from each standard strain was extracted as described above and used for RT-PCR. All RT-PCR experiments were performed with duplicates for each sample. The detailed PCR reaction information is listed in the Supplementary Material.

The murine microglial cell line BV-2 was cultured in high-glucose Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% fetal bovine serum (FBS) and 1% penicillin/streptomycin (P/S) in a humidified incubator with 5% CO₂ at 37°C. BV-2 cells were plated into 96-well plates (1.0 × 10⁵ cells/well) and incubated overnight for subsequent experiments.

The 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay is based on the protocol described by Mosmann (1983). The assay was optimized for the cells used in the experiments. Briefly, after the cells were incubated with different concentrations of o-tyrosine (o-tyr) (0.5, 1, and 2 mM) for 48 h at 37°С (Pennathur et al., 1999), the culture medium was removed, then the cells were incubated for 4 h with 0.5 mg ml⁻¹ of MTT and dissolved in serum free medium. Washing with PBS was followed by the addition of 100 µl DMSO and gentle shaking for 10 min to facilitate complete dissolution. After the formazan crystals had dissolved, the absorbance was determined spectrophotometrically at 490 nm on an ELX800 UV universal microplate reader. The results were analyzed with the Soft max pro software (version 2.2.2) and are presented as a percentage of the control values.

Reactive oxygen species (ROS) production in neuronal and BV-2 cells was assessed with a 2′,7′-dichlorodihydrofluorescein diacetate (DCFH-DA) probe. Different concentrations of o-tyr (0.5, 1, and 2 mM) were incubated with the cells for 48 h at 37°С. After the culture medium was removed, the cells were treated with 5 µM DCFH-DA at 37°С for 30 min. Washing with PBS three times was followed by the addition of 100 µl PBS and the intracellular ROS levels in the cells were viewed with a fluorescence microscope (Nikon Eclipse).

Nitric oxide (NO) production in neuronal and BV-2 cells was performed using 2′,7′-dichlorodihydrofluorescein diacetate (DAF-FM DA). Different concentrations of o-tyr (0.5, 1, and 2 mM) were incubated with the cells for 48 h at 37°С. After the culture medium was removed, DAF-FM DA (5 µM) was added to the wells. After incubating at 37°С for 20 min, the chemical was removed and the cells were washed three times with PBS followed by measurement with a fluorescence microscope.

Statistical analyses were performed using R 2.15.0 and GraphPad Prism software v 8.0. All the data were presented as the mean ± SEM. The significance of the differences between two groups was analyzed using the Student’s unpaired t-test and multiple comparisons were analyzed using one-way ANOVA followed by Dunnett’s post hoc test. The differential abundances of genera and metabolites were determined using non-parametric tests including the Wilcoxon rank sum test and Mann–Whitney U test. The correlations among fecal metabolites, 16S levels, and physiological and biochemical indexes were tested with both the Pearson Correlation Coefficient and Spearman rank correlation. p values were corrected for multiple comparisons using the Benjamini–Hochberg false discovery rate (FDR) and p < 0.05 was statistically significant.

Our previous study revealed that Rheum tanguticum ameliorated cognitive impairment in AD rat model, possibly due to the regulation of the gut microbiota (Zhao et al., 2019). In this study, to further verify the therapeutic effects for the treatment of AD, rhubarb extraction with a uniform and stable quality (Supplementary Figure S1A) and donepezil as positive drug were administrated to APP/PS1 transgenic mice for 2 months, named the TgR and TgP group, respectively. The main content of this study was shown in Figure 1. The effect of rhubarb on the learning and memory abilities of the Tg mice was tested by the Morris water maze. During the training period, the escape latencies of the mice in each group were similar on the first day, and gradually decreased in the next 5 days (Figure 2A). On the sixth day, the escape latency of Tg mice was significantly longer than that of wild-type (WT) mice. After the treatment of rhubarb or donepezil, the escape latency of Tg mice was significantly shortened and there was no significant difference between them. In the probe trial, Tg mice exhibited a lower platform passing times (Figure 2B) and shorter swimming time in target zone (Figure 2C). However, rhubarb or donepezil treatment significantly reversed these defects in Tg mice. These results indicated that rhubarb could attenuate spatial learning and memory deficits in APP/PS1 mice (Figure 2D).

The step-down test was performed to investigate the passive avoidance ability of Tg mice. Compared with the WT mice, Tg mice showed more errors times during the training period (Figure 2E) and shorter step-down delay (Figure 2F) accompanied by a greater number of errors (Figure 2G) during the experimental period. Fortunately, rhubarb or donepezil treatment reduced the errors times during the training period, increased the step-down delay, and reduced the number of errors during the experimental period in Tg mice with no significant difference. These results suggested that rhubarb extract could alleviate the stimulation avoidance response and cognitive impairment in APP/PS1 mice.

After the behavioral tests, the mice were sacrificed to further investigate the effect of rhubarb on the pathological changes in the brain of Tg mice, including Aβ deposition, activated microglia, neuronal damage and neurotransmitter disorders. The main pathological feature of AD is senile plaques composed of extracellular Aβ, which mainly deposited in the brain parenchyma and cerebral vessels (Pistollato et al., 2016; Friedland and Chapman, 2017). Here, Aβ deposition in the brain of Tg mice was stained by Congo Red, and the amount of total Aβ and Aβ₄₂ were quantified by ELISA. As expected, cerebral Aβ deposition (Figure 3A), and the levels of hippocampal Aβ and Aβ₄₂ (Figures 3B,C) in Tg mice was significantly higher than that of WT mice. After the treatment of rhubarb or donepezil, Aβ deposition in the brain of Tg mice was significantly reduced (Figure 3A). Moreover, although there was no significant difference in the level of total Aβ between the TgR and Tg group, the level of Aβ₄₂ was markedly reduced in the TgR group (Figures 3B,C). Donepezil treatment significantly reduced both the levels of total Aβ and Aβ₄₂.

Recent studies have shown that Aβ deposition in the brain can activate microglia and release inflammatory factors (Wu Y. et al., 2019). Therefore, the activation of microglia in the brain and the content of inflammatory factors in the hippocampus were evaluated by Iba-1 immunohistochemistry and ELISA, respectively. Compared with the WT mice, both the density and the number of activated microglia in the hippocampus of Tg mice were larger (Figure 3A). In addition, the levels of IL-1β and IL-18 in Tg mice were significantly higher than that of WT mice, and the level of and TNF-α was slightly higher (Figures 3D–F). As shown in Figure 3, rhubarb treatment exhibited a significant decrease in the number of activated microglia, and the levels of IL-18 and IL-1β (Figure 3) alone with a slight decrease of TNF-α similar to the donepezil treatment.

Activated microglia release inflammatory factors and induce neuroinflammation reaction, and eventually lead to neuronal apoptosis, resulting in the decline of learning and memory abilities (Rothhammer et al., 2018; Wu Y. et al., 2019). Therefore, the morphology of neurons in the brain tissue of mice was observed by HE staining. The results showed the disordered arrangement of neurons, condensed cytoplasm and karyopyknosis, and decreased number of neurons were observed in Tg mice (Figure 3A). Rhubarb or donepezil treatment reversed the above histopathological changes.

Neuronal injury can lead to the neurotransmitter disorders in the brain. Different kinds of neurotransmitters participate in neural activities in multiple brain regions. Therefore, the levels of 5-HT, DA, NE, Ach, Glu, and GABA in the brain of mice were detected by UPLC-TQ/MS. Compared with the WT mice, the levels of NE, Ach, 5-HT, and DA in the brain of Tg mice were significantly decreased, and the levels of Glu and GABA were slightly reduced (Figures 3G–L). As illustrated in Figure 3A, rhubarb and donepezil showed a favourable regulation in neurotransmitters in brain of Tg mice. Altogether, these results provided evidence that rhubarb could alleviate the pathological changes in Tg mice. Moreover, after 60 days treatment with rhubarb, there were no obvious abnormalities in body weight (Supplementary Figure S2A) and also no tissue damage or any other adverse effect in brain tissues (Supplementary Figure S2B).

In order to explore the changes in intestinal microbiota during the progression of AD, gut microbial dysbiosis in Tg mice in an age-dependent alteration (six samples per time point, n = 36) was evaluated, and distributed intestinal microbiota was identified by 16S rRNA gene sequencing. As a result, a total of 1237622 valid sequences were obtained from 36 samples (average 34378 ± 9394 reads per sample). These sequences were classified into 521 OTUs with a 97% similarity level. The α-diversity index includes the Chao1, Shannon, and Simpson indexes, which were used to determine the ecological diversity of the microbial community. The results showed that the α-diversity of gut microbiota in Tg mice decreased to a certain extent compared with the value of WT mice (Figures 4A–D). The dominant microbiota at the phylum level in the WT mice changed regularly with age (Figure 4E).

The β diversity of all bacteria in the six groups was calculated by the unweighted UniFrac metric and visualized by principal coordinate analysis, respectively. We found that the gut microbiota composition between WT and Tg mice changed similarly with time, but the variation in Tg mice was more significant (Figure 4F).

To further explore the microbiota that was related to the aggravation of the disease, the gut microbiota in WT and Tg mice was compared based on a PLS-DA model. According to the VIP > 0.7, 72 differential bacterial genera were screened at 8 months of age (Figure 5A), 76 at 9 months of age (Figure 5B), and 69 at 10 months of age (Figure 5C), respectively. However, only nine genera exhibited the same trends at the three time points. In the Tg group, the relative abundances of Tyzzerella, Ruminococcaceae_UCG_009, Bacteroides, Escherichia-Shigella, and Marvinbryantia, norank_f_Ruminococcaceae and Erysipelatoclostridium increased gradually, while Odoribacter and Akkermansia decreased gradually (Figure 5D).

After that, drug-responsive bacteria between the Tg and TgR group were identified by RT-PCR analysis. As shown in Figures 5E–H, rhubarb treatment increased the number of norank_f_Ruminococcaceae, Erysipelatoclostridium and Bacteroides and reduced the number of Marvinbryantia compared with the Tg group. To confirm the changes in the gut microbiota in Tg mice with age, co-network analysis was conducted on all OTUs detected in fecal samples from Tg mice at three time points (Figure 5I, Supplementary Figure S3). The results showed that there were close interactions between the intestinal microbiota, which increased or decreased together.

To clarify the effects of gut microbiota changes on AD-related pathological changes in mice, the metabolic profiles of fecal metabolites from the Tg and WT groups at 8, 9, and 10 months of age were analysed by UPLC-Q-TOF/MS in positive and negative ion modes. BPIs of the metabolomic profiles of Tg sample in the negative ion modes is shown in Figure 6A. A total of 9941 features in the positive ion mode and 5139 features in the negative ion mode were detected in all samples from the WT and Tg groups.

The fecal metabolic profiles of the WT and Tg mice were analysed with PCA. Different metabolites were observed among the groups at 8, 9, and 10 months of age. The PCA scores for the WT and Tg at three time points are shown in Figure 6B (positive ion mode) and Figure 6C (negative ion mode). It can be observed that the profile for the fecal metabolites showed age-dependent changes. The obvious difference between the WT and Tg groups indicated that the fecal metabolites of Tg mice had changed significantly. To explore potential biomarkers related to the pathogenesis of AD, the fecal metabolomics data for WT and Tg mice was compared by a PLS-DA model (Supplementary Figure S4). According to PLS-DA model with a VIP value greater than 1 and a p value less than 0.05 for the ANOVA t-test, differential metabolites were identified based on the fragment information and accurate mass number. Through a search and comparison of the HMDB, Metlin, and Chemspider databases, 50 differential metabolites with similar trends were found in the feces of Tg mice at three time points (Figure 6D). We speculate that some of these 50 differential metabolites were directly related to the pathological process in AD.

Drug-specific responsive microbial metabolites in the TgR group were screened by metabolomics analysis. After the administration of rhubarb extract for 30 and 60 days, the fecal metabolic profile for Tg mice showed a gradual recovery to a profile similar to that of WT mice (Figures 6E,F, Supplementary Figures S1B,C). Further, 27 of the 50 metabolites (Figure 6D) present at both time points in the TgR group showed a rhubarb intervention-dependent trend.

To gain insight into the mechanism underlying the rhubarb-induced improvement of cognitive impairment in AD mice, the metabolic pathways for 50 age-dependent metabolites with differences between the WT and Tg groups and 27 biomarkers for the long-term effects of rhubarb on AD cognitive impairment were evaluated. The results showed that the main metabolic pathways for rhubarb were histidine metabolism, D-glutamine, and D-glutamic acid metabolism, aminoacyl tRNA biosynthesis, alanine, aspartic acid, and glutamic acid metabolism (Figure 6G).

To further screen the microbial metabolites associated with bacteria in response to rhubarb, an improved co-network analysis was conducted based on the gut microbiota data at the OTU level and 50 differential metabolites detected in mouse fecal samples, with five pathological indicators and six neurotransmitters. As shown in Figure 7, Bacteroides, Marvinbryantia, norank_f_Ruminococcaceae, and Erysipelatoclostridium were significantly correlated with several biomarkers and pathological indicators. Phosphatidylcholine (PC; 15:0/18:2(9Z,12Z)), o-tyr, 3-hydroxyundecanoyl carnitine, L-glutamic acid, LysoPE (14:0/0:0), 3-hydroxytetradecanedioic acid, and pyroglutamic acid showed strong correlations with both the microbiota and indicator nodes. Interestingly, o-tyr showed a significant correlation with both the microbiota and indicator nodes (Supplementary Table S6), suggesting that o-tyr may lay on an important node in this network. We speculated that these four genera and their related metabolites and pathways may influence AD-related pathological indicators and play an important role in the progression of AD. A variety of pathological indicators for AD are related to intestinal microbiota disorders and metabolic disorders (Supplementary Figure S5). It suggested that the decline of cognitive function in AD was related to the microbiota composition and metabolites and further aggravates cognitive impairment.

As shown in Figure 7, some bacteria and metabolites play pivotal roles in the microbiota-metabolite-indicator network and may react specifically to rhubarb. To understand the potential effect of rhubarb on the microbiota-metabolite-indicator network, metabolite nodes with higher node degrees were analysed. Interestingly, the abundance of o-tyr, 3-hydroxyundecanoyl carnitine, L-glutamic acid, LysoPE (14:0/0:0), 3-Hydroxytetradecanedioic acid, and pyroglutamic acid, which connected the microbiota and indicator nodes in the network, were recovered after rhubarb treatment (Figures 7B–H).

As mentioned above, the relative peak area of o-tyr in the feces of the Tg group continued to increase with age. The level of o-tyr was significantly reduced by rhubarb intervention (Figure 7B). In the microbiota-metabolites-indicators network, o-tyr was significantly correlated with two types of bacteria, four pathological indicators, and seven metabolites, suggesting it may play a pivotal role in connecting the microbiota to the pathological indicators in the microbiota-metabolites-indicators network.

It has been reported that o-tyrosine could not only produce intracellular hydroxyl radicals, but other reactive oxygen species (ROS), such as hydrogen peroxide and superoxide radical anion (Ipson and Fisher, 2016). We speculated that in the process of aging, o-tyr from bacteria could enter the brain due to the increased permeability of the blood-brain barrier, induced neuronal oxidative stress and apoptosis. In this study, in vitro experiments on primary neurons and BV-2 cells were performed to validate whether o-tyr could promote the pathology of AD. ROS is considered to participate in the pathogenesis of AD, which can cause oxidative stress and trigger damages to cells in the brain. The intracellular ROS levels in neuronal and BV-2 cells were determined with DCFH-DA and the fluorescence intensity was further monitored by fluorescence microscope. As shown in Figures 8A,C,F,H, as the increase of o-tyr concentration, the fluorescence intensity of both cells significantly enhanced.

NO, as an inflammatory factor, plays an important role in neuronal death. To investigate the effect of o-tyr on the inflammatory response in neuronal and BV-2 cells, DAF-FM DA was used as a probe for NO. As shown in Figure 8, the fluorescence intensity in both neuronal and BV-2 cells significantly enhanced with the increase of the o-tyr concentration.

Studies have shown the potential of o-tyr to induce apoptosis. Therefore, the impact of o-tyr on the viability of neuronal and BV-2 cells was evaluated by the MTT assay. As shown in Figure 8, both of the two cells treated with 2 mM o-tyr reduced the cell viability to approximately 50%. These results demonstrated that o-tyr adversely affected the function and viability of the neuronal and BV-2 cells.