Oolong tea attenuates neuroinflammation by modulating the microbiota-gut-brain axis in a rat model of autism

The experiments were approved by the Animal Use and Care Committee at South China Agricultural University (Approval No. 2021B144). A total of 30 specific pathogen-free 12-week-old Sprague–Dawley rats (20 female rats, 10 male rats, 250–300 g) were acquired from Guangdong Medical Laboratory Animal Center. Rats were housed under controlled conditions (temperature: 23 ± 1 °C, humidity: 55 ± 5%, 12 h light/dark cycle) with ad libitum access to standard chow and sterile water for 1 week of acclimatization before experiments. Female and male rats were mated at a 2:1 ratio. Embryonic day 0.5 (E0.5) was designated as the day of vaginal plug detection, while birth was considered postnatal day (PND) 0.

On E12.5, pregnant rats in the VPA group received 500 mg/kg VPA (Sigma, P4543; 250 mg/mL saline solution) by intraperitoneal injection, while sham group rats were injected with an equivalent volume of saline.

The tea extracts were prepared as follows. First, 10 g of the OT sample was soaked in 100 mL of boiling water for 10 min, after which it was extracted three times. All the extracts were subsequently mixed, filtered, and concentrated under vacuum. Finally, the concentrated filtrate of tea extracts was freeze-dried into powder and stored at −80 °C.

Component analysis of OT was performed according to a previous study (29). For caffeine, 0.45 g magnesium oxide and 0.1 g freeze-dried tea powder were added to 30 mL of ultrapure water at 100 °C and ultrasonically agitated for 30 min. Then, 1 mL of the extract was filtered and analyzed using high-performance liquid chromatography (HPLC). The analysis utilized a C18 SB column (4.6 × 250 mm, 5 mm) at 35 ± 1 °C with 100% ultrapure water as solvent A and 100% methanol as solvent B. The compounds were eluted for 14 min at an isocratic pressure of 70% A and 30% B at a flow rate of 0.9 mL/min with a detection wavelength of 280 nm.

For theanine, 0.1 g freeze-dried tea powder was extracted in 10 mL of ultrapure water at 100 °C for 30 min. Then, 1 mL of the extract was filtered and analyzed using HPLC. The mobile phases were 100% ultrapure water (A) and 100% acetonitrile (B). The gradient was set at 100% A for the first 12 min, gradually shifted to 20% A from 12 to 14 min, and held at 20% for 5 min. Finally, the gradient was changed to 100% A from 19 to 21 min and then was held for 4 min. Theanine was detected at 210 nm.

For catechins, 0.2 g freeze-dried tea powder was extracted in 8 mL of 70% methanol. Then, 1 mL of supernatant was filtered and analyzed by HPLC using a C18 column (4.6 × 250 mm, 5 mm). The catechin monomer was eluted using a gradient elution procedure with 0.1% aqueous formic acid (v/v) (A) and 100% acetonitrile (B) as the mobile phases. The gradient elution started at 8% B for 5 min, was increased to 25% from 5 to 14 min, and decreased to 8% from 14 to 30 min. Detection was at 280 nm.

Rats were randomly assigned to six groups with 12 rats in each group. In the sham group, the sham rats received distilled water (400 mg/kg/day) through gavage for 4 weeks (from PND21 to PND49). In the VPA group, VPA-treated rats were administered distilled water (400 mg/kg/day) through gavage during this period. In the OT-L, OT-M, and OT-H groups, VPA-treated rats were administered three doses of OT, viz. 100 mg/kg/day, 200 mg/kg/day, and 400 mg/kg/day through gavage for 4 weeks (from PND21 to PND49), respectively (30–32). In the OT + antibiotic cocktail (ABX) group, VPA-treated rats were administered OT (400 mg/kg/day) through gavage and with antibiotics (ampicillin 1 g/L, vancomycin 0.25 g/L, neomycin 1 g/L, and metronidazole 1 g/L) added into their drinking water for 4 weeks (from PND21 to PND49) (33).

Based on preliminary data, a minimum sample size of six animals per group was required to detect significant behavioral differences (34). Six to eight rats per group were used and analyzed. Animals were randomly assigned to groups, and investigators remained blinded to group assignments throughout the study and analysis. Behavioral testings were conducted at fixed daily times to avoid circadian effects.

Exclusion criteria: Rats were excluded if they died during the experiment, showed >20% weight loss, or failed to complete behavioral tests (e.g., persistent immobility or aggressive behavior).

At PND49, fresh feces samples were collected. The 16S rRNA gene amplicon sequencing and analysis were performed by OE Biotech Co., Ltd. (Shanghai, China). Total genomic DNA was isolated using a DNA Extraction Kit. Quality and quantity of DNA were assessed using NanoDrop and agarose gel. The diluted DNA served as a template for polymerase chain reaction amplification of bacterial 16S rRNA genes, with the barcoded primers and Takara Ex Taq (Takara) employed in the reaction. For analyzing bacterial diversity, V3–V4 variable regions of 16S rRNA genes were amplified using universal primers 343F and 798R.

Raw sequencing data were in FASTQ format. Paired-end reads underwent preprocessing with the cutadapt software to identify and remove adapter sequences. After trimming, paired-end reads were subjected to filtering of low-quality sequences, denoising, merging, and chimera detection and removal using DADA2 with QIIME2 (2020.11). At last, the software outputs the representative reads and the amplicon sequence variant (ASV) abundance table. The representative read of each ASV was selected using the QIIME2 package. All representative reads were annotated and blasted against Silva138_Eukaryotausing q2-feature-classifier with the default parameters. 0001

All bioinformatics analyses of amplicon reads were conducted based on ASVs. α diversity was computed using R software (version 3.5.1). Principal coordinates analysis (PCoA) based on weighted UniFrac distances was visualized with the ggplot2 package (version 3.4.0) in R software. LEfSe, a tool commonly applied to identify biomarkers, is capable of uncovering metagenomic features (39). Sequence data have been deposited in the NCBI SRA database (Accession Number: PRJNA1298520).

The rats were euthanized at PND50. The plasma samples were obtained for enzyme-linked immunosorbent assay (ELISA) analysis. The intestine and brain were stored at −80 °C and used for Western blot analysis. The intestine and brain were placed in 4% paraformaldehyde, followed by Nissl staining and immunofluorescence analysis.

Nissl staining was employed to identify neuronal damage. Paraffin sections were cleared three times (10 min each), with excess liquid gently shaken off between steps. They were then dehydrated through an ethanol series (anhydrous, 95, 85, 75%; 5 min each), rinsed in distilled water (3 times), and stained with preheated 1% toluidine blue (50 °C, then 56 °C for 20 min). After thorough distilled water washes, sections were differentiated under microscopic control (using 95% ethanol or 0.1% glacial acetic acid) until Nissl bodies were distinct. Rapid dehydration in anhydrous ethanol was followed by clearing and mounting with neutral balsam. Photos of cortical neurons were taken with a conventional optical microscope to view their morphology.

The expression levels of lipopolysaccharide (LPS), interleukin-6 (IL-6), and tumor necrosis factor-α (TNF-α) in the plasma, intestine, cortex, and hippocampus were evaluated using rat ELISA kits (Ek-Bioscience, Shanghai, China). Before initiating the assay, all reagents were prepared. Wells for standards and test samples were designated; 50 μL of the standard was dispensed into each standard well. For test sample wells, 10 μL of the sample was added, followed by 40 μL of sample diluent. Then, 100 μL of HRP-conjugate reagent was added to every well, and the plate was sealed with an adhesive cover and incubated at 37 °C for 60 min. After incubation, the contents of each well were aspirated, and the wells were washed. This process was repeated four times, for a total of five washes. Following the final wash, any residual wash solution was removed by aspiration or decantation. The plate was then inverted and blotted on clean paper towels. Next, 50 μL of chromogen solution A and 50 μL of chromogen solution B were added to each well. The plate was mixed gently and incubated at 37 °C for 15 min. Subsequently, 50 μL of stop solution was added to each well. Within 15 min, the optical density was measured at 450 nm using a microplate reader.

Initially, sections were dewaxed for 10 min. They were then rinsed for 5 min each in anhydrous ethanol, 95% ethanol, and 75% ethanol. The slices were then placed in an EDTA antigen repair solution and boiled. After being removed from the retrieval solution, the sections were immersed in TBST three times. Next, 10% donkey serum was added and incubated at 37 °C for 30 min. The sections were then incubated with ZO-1 antibody (1:50, Thermo Fisher, 61-7300), TLR-4 antibody (1:200, Proteintech, 19811-1-AP), Neun (1:500, Oasis Biofarm, OB-PGP006-01), Iba-1 antibody (1:200, Oasis Biofarm, OB-PGP049-01), and GFAP antibody (1:200, Oasis Biofarm, OB-PGP055) overnight at 4 °C. The next day, the sections were rinsed three times with TBST. A 100 μL aliquot of the secondary antibody working solution was added to each section and incubated at 37 °C for 45 min. Subsequently, 100 μL of the DAPI working solution was added to each section to stain the nuclei. After a 5-min light exposure, the sections were rinsed with TBST. Finally, the sections were visualized under a fluorescence microscope.

The Western blot analysis was performed according to previously reported procedures (40). Total protein content from the intestine and cortex of the rat was extracted using ristocetin-induced platelet aggregation buffer supplemented with protease inhibitors. Nuclear protein and cytoplasmic protein were extracted using a nuclear and cytoplasmic extraction kit (Thermo Fisher, 78833). Protein concentrations were quantified using the BCA protein concentration kit (Beyotime, P0010S). The protein underwent electrophoresis using a 10% SDS-PAGE gel and was subsequently transferred to a 0.45 mm PVDF membrane. The membrane was then incubated in a protein blocking solution for 1 h. The membranes underwent incubation with the following primary antibodies: claudin-1 (1:250, Invitrogen, 51-9000), occludin (1:1,000, CST, 91131), ZO-1 (1:250, Invitrogen, 61-7300), claudin-5 (1:250, Sigma, SAB4502981), TLR-4 (1:500, Invitrogen, PA5-23124), IκB-α (1:1,000, CST, 9242), NF-κB (1:1,000, CST, 8242), GAPDH (1:1,000, CST, 2118s), and Histone H3 (1:1,000, ab1791). On the following day, the membranes were rinsed three times with TBST solution, then incubated with a second antibody (1:5,000, Boster, BA1054) for 1 h at room temperature. After incubation, the membrane was washed three times with TBST solution to eliminate unbound secondary antibodies. Finally, the protein bands were observed using enhanced chemiluminescence. The protein expression levels were quantified using ImageJ software.

Data analysis was performed using SPSS 17.0 software. The results are expressed as the mean ± standard error of the mean (SEM). Statistical evaluations were carried out using one-way analysis of variance, followed by Fisher's post hoc test for pairwise comparisons. A p-value threshold of <0.05 was adopted to define statistical significance.

The experimental timeline is displayed in Figure 1A. VPA was administered at E12.5, and OT was orally administered from PND21 to PND49. The self-grooming test was conducted on PND45, the marble burying test was measured on PND46, the three-chamber test was performed on PND48, feces were collected on PND49, and the rats were euthanized on PND50.

The contents of caffeine, theanine, and catechins in the extract of OT are shown in Table 1.

To evaluate the neuroprotective effects of OT on VPA-treated rats, behavioral tests were carried out. To assess repetitive behaviors in rats, the self-grooming test and marble burying test are frequently utilized. In the self-grooming test, the VPA group spent significantly more time self-grooming than the sham group (Figure 1B, p < 0.01). In contrast, the OT-H group exhibited shorter self-grooming time than the VPA group (Figure 1B, p < 0.05). No statistically significant differences were observed in the self-grooming test for the OT-L group and the OT-M group compared to the VPA group (Figure 1B, p > 0.05).

Marble burying test indicated that the VPA group buried more marbles than the sham group (Figure 1C, p < 0.01). The OT-H group buried fewer marbles than the VPA group (Figure 1C, p < 0.05). No statistically significant differences were observed in the marble burying test for the OT-L group and the OT-M group compared to the VPA group (Figure 1C, p > 0.05).

The three-chamber social interaction test was used to assess sociability and social preference. As shown in Figures 1D,E, the VPA group exhibited a lower sociability index and social preference index than the sham group (Figures 1D,E, p < 0.01), whereas OT-H supplementation reversed the results (Figures 1D,E, p < 0.05). No statistically significant differences were observed in the sociability index and social preference index for the OT-L group and the OT-M group compared to the VPA group (Figures 1D,E, p > 0.05). Therefore, we administered 400 mg/kg/day of OT to VPA-treated rats for all subsequent assays.

Moreover, Nissl staining was used to assess the neuroprotective effects of OT on VPA-treated rats. The VPA group showed significantly fewer Nissl-stained cells in the cortex than the sham group (Figure 1G, p < 0.001). However, OT supplementation significantly attenuated the neuronal cell death in the OT group compared with the VPA group (Figure 1G, p < 0.05). Taken together, these findings suggest that OT has neuroprotective effects on VPA-treated rats.

To characterize the gut microbiota in VPA-treated rats, fecal samples were subjected to 16S rRNA gene sequencing analysis. The Chao1 index indicates sample richness (Figure 2A). The analysis of α diversity using the abundance-based coverage (ACE) estimator revealed no significant differences (Figure 2B, p > 0.05), indicating comparable gut microbiota diversity across the three groups. Conversely, PCoA revealed significant variations in gut microbiota composition among the groups (Figure 2C, p = 0.001). Moreover, compared with the sham group, the VPA group showed a marked increase in the family Ruminococcaceae and the genera Bacteroides and Ruminococcus (Figures 2D–F, p < 0.05 or p < 0.01). Furthermore, the OT group had lower abundances of these gut microbiota than the VPA group (Figures 2D–F, p < 0.05 or p < 0.01).

At the family level, heatmap analysis (Figure 2G) showed that the VPA group had a higher abundance of Ruminococcaceae than the sham and OT groups. At the genus level, heatmap analysis (Figure 2H) showed that the VPA group had a higher abundance of Bacteroides and Ruminococcus than the sham and OT groups.

LEfSe is a new way to identify biomarkers through category comparisons. The gut microbiota exhibiting the most substantial differences (with an LDA score >3.0) among the three groups were detected (Figure 2I). Ruminococcaceae, Bacteroides, Bacteroidaceae, Ruminococcus, Marvinbryantia, Acholeplasmatales, Acholeplasmataceae, and Anaeroplasma were more abundant in the VPA group. Taken together, these findings imply that OT has the potential to alter the microbiota composition in the intestines of VPA-treated rats.

To investigate whether OT affects LPS and proinflammatory cytokines in VPA-treated rats. ELISA was employed to assess the levels of LPS, IL-6, and TNF-α in the plasma, intestine, cortex, and hippocampus. ELISA revealed that compared with the sham groups, the VPA group exhibited significantly elevated levels of LPS, IL-6, and TNF-α in the plasma (Figures 3A–C, p < 0.05), intestine (Figures 3D–F, p < 0.05 or p < 0.01), cortex (Figures 3G–I, p < 0.05), and hippocampus (Figures 3J–L, p < 0.05). Specifically, OT supplementation significantly reduced these levels in the plasma (Figures 3A–C, p < 0.05), intestine (Figures 3D–F, p < 0.05), and cortex (Figures 3G–I, p < 0.05) in the VPA-treated rats. There were no statistically significant differences in the levels of LPS, IL-6, and TNF-α in the hippocampus between the VPA group and the OT group (Figures 3J–L, p > 0.05). Therefore, we focused on the cortex in the following assays. These results indicate that OT inhibits LPS production and alleviates inflammation in VPA-treated rats.

To investigate intestinal barrier dysfunction in VPA-treated rats, Western blotting was employed to analyze tight junction proteins in the intestine. In comparison to the sham group, the VPA group exhibited notably reduced expression of the intestinal claudin-1, occludin, and ZO-1 (Figure 4A, p < 0.01). In contrast, the OT group demonstrated significantly elevated expression of these proteins relative to the VPA group (Figure 4A, p < 0.05). Immunofluorescence staining further revealed that the OT group had more intestinal ZO-1-positive cells than the VPA group (Figure 4B, p < 0.05). Collectively, these findings suggest that OT attenuates intestinal barrier dysfunction.

We next examined the association between OT and TLR-4/IκB-α/NF-κB signaling pathway in the intestines of VPA-treated rats. In comparison to the sham group, the VPA group exhibited upregulated expression of the intestinal TLR-4, NF-κB, and NF-κB (nuclear), with significant downregulation of IκB-α and NF-κB (cytosol) (Figure 4C, p < 0.05, p < 0.01, or p < 0.001). In contrast, the OT group showed downregulated TLR-4, NF-κB, and NF-κB (nuclear) expression, accompanied by upregulated IκB-α and NF-κB (cytosol) expression, compared to the VPA group (Figure 4C, p < 0.05). Collectively, these findings imply a potential role for the TLR-4 signaling pathway in the protective effect of OT in VPA-treated rats.

To evaluate whether OT affects BBB function, Western blotting was used to analyze tight junction proteins in the cortex. In comparison to the sham group, the VPA group exhibited notably reduced expression of claudin-5, occludin, and ZO-1 in the cortex (Figure 5A, p < 0.05 or p < 0.001). In contrast, the OT group showed increased expression of these proteins in the cortex relative to the VPA group (Figure 5A, p < 0.05). Immunofluorescence staining further revealed that the OT group had more ZO-1-positive cells in the cortex than the VPA group (Figure 5B, p < 0.01).

We next examined whether OT affected the TLR-4/IκB-α/NF-κB signaling pathway in the cortex of VPA-treated rats. In comparison to the sham group, the VPA group exhibited upregulated expression of TLR-4, NF-κB, and NF-κB (nuclear) in the cortex, with significant downregulation of IκB-α and NF-κB (cytosol) (Figure 5C, p < 0.01 or p < 0.001). In contrast, the OT group exhibited downregulated TLR-4, NF-κB, and NF-κB (nuclear) expression, accompanied by upregulated IκB-α and NF-κB (cytosol) expression, compared to the VPA group (Figure 5C, p < 0.05). Furthermore, we performed multiplex immunofluorescence staining on the cortex. TLR-4 was co-stained with Neun, Iba-1, and GFAP, respectively. We found that TLR-4 was co-expressed with Neun, but not with Iba-1 and GFAP. The inhibition of the TLR-4 pathway by OT mainly occurs in neurons. Immunofluorescence staining further revealed that the OT group had fewer TLR-4-positive neurons in the cortex than the VPA group (Figure 5D, p < 0.01). These data imply that OT attenuates BBB disruption and inhibits the TLR-4/IκB-α/NF-κB signaling pathway.

We next examined whether OT attenuated neuroinflammation in the cortex of VPA-treated rats. The VPA group had more Iba-1-positive cells and GFAP-positive cells than the sham group (Figures 5E,F, p < 0.01). In contrast, the OT group demonstrated significantly reduced expression of Iba-1-positive cells and GFAP-positive cells relative to the VPA group (Figures 5E,F, p < 0.01). These findings imply that OT attenuated neuroinflammation in the cortex of VPA-treated rats.

To obtain direct evidence of the neuroprotective effect of OT on gut microbiota, we further administered an antibiotic cocktail (ABX) treatment to deplete gut microbiota, after which behavioral tests were conducted and levels of LPS and proinflammatory cytokines in the cortex were reassessed. In comparison with the OT group, the antibiotic treatment markedly increased self-grooming time and the number of buried marbles and reduced both the sociability and social preference indices in VPA-treated rats (Figures 6A–D, p < 0.05). Additionally, ELISA revealed that in comparison with the OT group, the OT + ABX group exhibited significantly elevated levels of LPS (Figure 6E, p < 0.05), IL-6 (Figure 6F, p < 0.05), and TNF-α (Figure 6G, p < 0.05) in the cortex. In summary, antibiotic treatment eliminated the neuroprotective effects of OT and increased neuroinflammation in VPA-treated rats. These findings indicate that gut microbiota plays an important role in the neuroprotection of OT in VPA-treated rats.