Establishment of P. gingivalis bacteremia model

The P. gingivalis bacteremia model was established according to the methods described previously.22,25,26,27 This study was approved by the Laboratory Animal Welfare and Ethics Committee of China Medical University (Approval number: KT2021150). Eight-week-old healthy SD rats (body weight 180–220 g) were purchased from Changsheng Biotechnology Cable Company (Benxi City, Liaoning Province, China). The P. gingivalis ATCC 33277 strain was obtained from the American Tissue Culture Collection (Maryland, USA). The bacteria were maintained anaerobically at 37 °C on brain-heart-infusion (BHI) agar medium plates, supplemented with 5% sterilized and defibrinated sheep blood, 5 µL·mL−1 hemin, and 1 µL·mL−1 Vitamin K. The P. gingivalis was cultured in a liquid BHI medium for 16–18 h before experiments. All bacterial culture reagents were purchased from Aobo Bio-tech (Beijing, China). The experimental groups were injected intravenously with P. gingivalis through the tail vein three times a week for eight weeks. In the high-intensity group, the rats were injected with 200 µL PBS containing 108 CFU P. gingivalis22 while in the low-intensity group, the rats were injected with 200 µL PBS containing 103 CFU P. gingivalis,26,29 and in the control group the rats were injected with 200 µL PBS (n = 6 in each group).

Transmission electron microscopy

The rats were injected with bacterial for 8 weeks by the tail vein. After deep anesthesia, the cerebral cortex and hippocampus brain tissues were taken out and fixed in 2.5% glutaraldehyde. The bEnd.3 were infected with P. ginigivalis for 24 h. The ultrathin sections were prepared, and the ultrastructure was observed with a transmission electron microscope (TEM, H7650, Hitachi, Tokyo, Japan).

Western blot

Western blot was detected with BIO-RAD protein analysis system (BIO-RAD, USA). The primary antibodies were as follows: Cav-1 (Abcam, Cambridge, UK), Fibrinogen (Proteintech, Wuhan, China), Albumin (ABclonal, Wuhan, China), Occludin (Proteintech, Wuhan, China), anti-P. gingivalis (DSHB, Texas, USA), GAPDH (Proteintech, Wuhan, China). The fluorescent secondary antibody (Proteintech, Wuhan, China). Infrared fluorescence scanning imaging system (Odyssey CLx, LI-COR USA) was used to detect protein bands. ImageJ 1.52v software (NIH Image, Bethesda, MD, USA) was used for protein semi-quantitative analysis.

Immunohistochemical

After deep anesthesia, the rats were perfused transcardially with pre-cold PBS and 4% paraformaldehyde solution (PFA) successively. Brains were removed and fixed in PFA at 4 °C. Paraffin sections were added with the primary antibody working solution of Cav-1 (Abcam, Cambridge, UK) overnight at 4 °C. Each section was added enzyme-labeled IgG polymer dropwise, incubated with DAB at room temperature.

Immunofluorescence microscopy and laser confocal microscopy

After deep anesthesia, the rats were perfused transcardially with pre-cold PBS and 4% PFA successively. Brains were removed and fixed in PFA at 4 °C. Tissue paraffin sections were blocked and then incubated with the anti-rat Albumin antibody (ABclonal, Wuhan, China), anti-Mfsd2a antibody (Novus, Colorado, USA) and anti-CD31 antibody (Santa, Texas, USA). Cell samples were treated with P. gingivalis or with PBS, fixed with 4% PFA, treated with TritonX-100. The primary antibody was anti-P. gingivalis (DSHB, Texas, USA) and rabbit anti-mouse Cav-1 (Abcam, Cambridge, UK). The sections were incubated with the goat anti-mouse secondary antibody (Proteintech, Wuhan, China) and the goat anti-rabbit secondary antibody (Proteintech, Wuhan, China) for 2 h. The sections were stained with DAPI. Images were ultimately acquired with the aid of a fluorescence microscope (ECLIPSE, Nikon, Japan) and a laser-scanning confocal microscope (C2, Nikon, Japan).

Establishment of the transwell BBB model in vitro and the permeability test

The mouse brain microvascular endothelial cell line bEnd.3 and the human astrocytoma cell line U87 were purchased from the Shanghai Cell Bank of the Chinese Academy of Sciences. Cells were cultured in DMEM medium supplemented with 10% fetal bovine serum (FBS) under the conditions of 37 °C with 5% CO2. The BBB is comprised of three kinds of specialized cells, including BMECs, astrocytes, and perivascular cells (pericytes). The co-culture transwell BBB model established by BMECs and astrocytes has been frequently reported in previous papers.44,45 bEnd.3 cells were incubated on the bottom of the upper transwell chamber (Corning 3402, Corning, USA), and U87 cells were incubated on the bottom of the lower transwell chamber. After 7 days, a certain amount of complete cell culture medium was added to the upper and lower chambers of the transwell unit to make the liquid level of the upper chamber higher than that of the lower chamber by 0.5 cm. After 4 h, if the original liquid level difference was maintained, the transwell BBB model should be considered to be successfully established.46 And then, P. ginigivalis ((MOI: 10, 100, 500) was added to the upper chamber for 24 h, and then the lower chamber liquid was aspirated for bacteria detection with the anaerobic culture method.

Transfection assays

For transfection, the bEnd.3 cells were plated on six-well flatbottom plates at a seeding density of 2 × 105 and grew to 80% confluence. The Mfsd2a plasmids were transfected into bEnd.3 cells for 24 h. The Mfsd2a siRNA was transfected into bEnd.3 cells for 24 h. The cells treated with empty vectors or scrambled siRNA were used as the negative control.

Flow cytometry

For flow cytometry, cells were divided into six groups and compared with each other in different experiments: control group, P. gingivalis, empty vector plasmid transfection+ P. gingivalis, Mfsd2a plasmid transfection + P. gingivalis, scrambled siRNA transfection + P. gingivalis, and Mfsd2a-siRNA transfection + P. gingivalis. After transfection, P. gingivalis was added to infect the BMECs for 24 h. Fluorescein isothiocyanate (FITC) dilution was then added to the wells for 10 min. The fluorescence intensity of FITC of the cells was detected by flow cytometry (FACS, Becton-Dickinson, Islandia, NY, USA) and analyzed using FlowJovX0.7 software.

Quantitative real-time polymerase chain reaction

In this study, we established a bacterial internalization model by the antibiotics protection assay. The bEnd.3 cells were infected with P. gingivalis with a specific multiplicity of infection (MOI) for 6 h. Then cells were washed three times with PBS and were further incubated in the culture medium containing 300 µg/mL of gentamicin and 200 µg·mL−1 of metronidazole (Sigma, St. Louis, MO, USA) for 2 h. Total RNA was extracted from cells using TRIzol reagent (Invitrogen Life Technologies, Gaithersburg, MD, USA) according to the manufacturer’s instructions. Real-time PCR analyses were conducted on an ABI Prism 7500 Sequence Detection System (Applied Biosystems, Foster City, CA, USA) in combination with a SYBR Premix Ex TaqTM II PCR Master Mix Reagents kit (Takara Bio, Inc., Dalian, China). P. ginigivalis 16s RNA primer sequence: Forward Primer: AGGCAGCTTGCCATACTGCG, Reverse Primer: ACTGTTAGCAACTACCGATG. Fold changes were calculated through relative quantification (2−△△CT) as previously reported. The quantity of P. gingivalis was displayed as the relative ratio to GAPDH expression level according to the methods reported in our previous studies.16,47 Each experiment was performed in triplicates.

Immunoprecipitation and mass spectrometry analysis

The immunoprecipitation was conducted as the kit instructions (Takara Bio, USA). Briefly, protein lysates were centrifuged, and the supernatant was removed and kept. Incubate the recommended amount of Cav-1 antibody overnight. The eluted antibody-protein complex with the neutralization buffer in the tube was collected. Subsequently, the sample was analyzed by mass spectrometry (Beijing Protein Innovation Co. Ltd, Beijing, China).

Protein–protein docking

The protein–protein docking was conducted by APExBIO Technology LLC (Shanghai, China). Cav-1 and P. gingivalis-arg-specific-gingipainA (RgpA) protein structures were queried from the UniProt database and the structure files were saved in pdb format. The pdb files were imported into the Maestro docking software. Cav-1 and RgpA were set as a ligand and the receptor, respectively, for the docking parameters. Thirty docking poses were output after docking and the Protein Interaction docking complex was used for interaction analysis. Protein docking data were analyzed and the graphs were drawn.

Statistical analysis

Normally distributed data were expressed as the means ± standard deviation (SD). Differences among the three group were analyzed by multiple comparisons using one-way analysis of variance (ANOVA). Differences among the two groups were analyzed by an independent two-tailed t test. SPSS 22.0 software package (SPSS Inc., Chicago, IL, USA) was used to perform the analysis. P < 0.05 was considered to be statistically significant.