The protocol was approved by the Institutional Review and Ethics Board of Shanghai Ninth People’s Hospital, Shanghai Jiao Tong University School of Medicine. All medical data and samples were collected according to standard clinical procedures. Periodontitis were initially diagnosed by dentist (L Bai) and the patients we selected also met the following criteria (Kinane et al., 2017; Qian et al., 2021): (1) gum bleeding within 15 s after probing; (2) at least one site probing depth > 6 mm; (3) at least one site with attachment loss > 5 mm; (4) at least one site with alveolar bone absorption more than 1/2 of the length of the root. Patients who had taken any antibiotic or probiotic, smoked, undergone periodontal therapy in the previous 6 months, or had severe systemic diseases were not included. At the same time, healthy people were also recruited. All participants fasted for 2 h before the examination. Saliva samples of healthy people and periodontitis patients were collected and then mixed, respectively, and stored at −80°C until further processing.

All animal experiments were approved by the Institutional Review and Ethics Board of Shanghai Ninth People’s Hospital, Shanghai Jiao Tong University School of Medicine. C57BL/6 mice were purchased from Charles River Laboratory Co., Ltd. (Beijing, China) and acclimatized the mice to the animal facility for 1 week before being randomly assigned to the experimental group. Il17a^–/– mice on the C57BL/6 background were kindly provided by You-Cun Qian (Chinese Academy of Sciences, Shanghai). Kaede transgenic mice were a generous gift from M. Tomura (Kyoto University). All mice were males (age = 6–8 weeks). Animals were kept under standard specific pathogen free (SPF) conditions at an appropriate temperature (22 ± 2°C) and humidity of 60% under a 12 h light/dark cycle. Food and water were available ad libitum. Genetically modified lines were bred in our facility. To minimize the impact of the maternal body on the composition of the microbiota, mice from different litters were mixed and then randomly assigned to the experimental group.

Ampicillin (Aladdin, Shanghai, China), metronidazole (Sigma-Aldrich, MO, United States), neomycin sulfate (Sigma-Aldrich), and vancomycin (Sigma-Aldrich), abbreviated AMNV. AMNV was administrated by drinking for 2 weeks, and 200 μl of the AMNV was administered by oral gavage every day during the second week. Concentration of drinking: ampicillin (0.3 g/L), metronidazole (0.3 g/L), neomycin sulfate (0.3 g/L), and vancomycin (0.15 g/L). Concentration of gavage: ampicillin (1 g/L), metronidazole (1 g/L), neomycin sulfate (1 g/L), and vancomycin (0.5 g/L).

The model was induced as previously described (Belayev et al., 1996). Briefly, the mice were under intraperitoneal anesthesia with 2% avertin (Sigma-Aldrich). The temperature was maintained using a homoeothermic blanket. The neck skin was cut along the midline. After the left common carotid artery, the internal carotid artery, and the external carotid artery separated carefully, a silicone-coated suture (Yushun, Henan, China) was inserted into the left external carotid artery, advanced into the internal carotid artery and wedged into the cerebral arterial circle to obstruct the origin of the middle cerebral artery. After 1 h of occlusion, the suture was withdrawn. Mice in the SHAM group underwent the same surgical procedures, but the suture was withdrawn immediately after the suture reached the origin of the middle cerebral artery.

Mice were euthanized 1 day after MCAO. Brains were immediately removed and frozen at −20°C for 30 min. Then mouse brains were coronally cut into 2 mm-thick sections. The sections were incubated with 2% 2, 3, 5-triphenyl-2H-tetrazolium chloride (TTC) (Sigma-Aldrich) at 37°C for 15 min and then photographed. Infarct size and volume were calculated by ImageJ software (National Institutes of Health, United States). To exclude the effects of brain edema, the relative percentage of infarct volume in hemisphere was calculated as follows: (volume of the contralateral hemisphere minus the volume of the non-lesioned ipsilateral hemisphere)/(volume of the contralateral hemisphere) (Lu et al., 2013).

The Modified Garcia Score was used to measure motor function: body proprioception, forelimb walking, climbing, lateral turning, and limb symmetry as previously described (Wang et al., 2018). Each test was scored from 0 to 3 (maximal score 15).

Mice were euthanized 1 day after MCAO, the ipsilateral hemisphere samples were collected. Total RNA was isolated from the ipsilateral hemisphere using Trizol (Thermo Fisher Scientific, MA, United States). cDNA was synthesized from RNA by using reverse transcription kits (Takara, Tokyo, Japan) and then detected by SYBR Green Mix (Thermo Fisher Scientific, Carlsbad, CA, United States) on a LightCycler 480 II (Roche, Switzerland). Gene expression was normalized by GAPDH. Primer sequences are listed in Table 1.

Stool sample were collected and stored at −80°C. Frozen stools were weighed and extracted DNA through the Bacterial DNA Kit (Tiangen Biotech, Beijing, China). qPCR was used for quantification of 16S ribosomal (r16S) DNA copy numbers as previously described (Jimeno et al., 2018). The concentration of DNA samples was adjusted to equivalent. r16S DNA sequences were amplified from stool DNA using 0.2 μmol/L of the universal bacterial r16S gene primers Eubacteria-F primer (5′ ACTCCTACGGGAGGCAGCAGT 3′) and Eubacteria-R primer (5′ ATTACCGCGGCTGCTGGC 3′) in conjunction with the SYBR Green mix. The standard curve was prepared using a plasmid containing a V3-V4 DNA fragment.

The frozen saliva is dissolved and passed through a 0.22 μm filter to obtain filtrate. DNA of saliva and its filtrate was extracted through the Bacterial DNA Kit. The way to detected r16S DNA copy numbers was described above.

Specific periodontal bacteria were also detected by qPCR in mouse stools. DNA of stools was extracted through the Bacterial DNA Kit and the concentration of DNA was measured through NanoDrop Lite UV-Vis spectrophotometer (Thermo Scientific, United States). Then, DNA samples were adjusted to the same concentration. Sequences were amplified from 1 μl stool DNA sample using 0.2 μmol/L of the specific bacterial primers (Supplementary Table 1) in conjunction with the SYBR Green mix on a LightCycler 480 II.

The level of inflammatory cytokines (IL-1β and TNF-α) were detected by enzyme linked immunosorbent assay (ELISA) kits (JL10484 and JL18442, Jiang Lai Biotechnology, Shanghai, China) following the manufacturer’s instructions.

For surface marker analysis of the immune cells of ileum and the colon, cell suspensions were incubated with anti-CD16/CD32 antibody (101320, Biolegend, San Diego, CA, United States) for 10 min at 4°C to block the non-specific binding. Then cells were stained with the appropriate antibodies for 25 min at 4°C.

For surface marker analysis of the immune cells of brain, samples were subjected to Percoll (17089109, Civata, Uppsala, Sweden) gradient centrifugation before flow cytometric staining. Cell suspensions were incubated with anti-CD16/CD32 antibody for 10 min at 4°C to block the non-specific binding. Then cells were stained with the appropriate antibodies for 25 min at 4°C.

The following antibodies were used for staining: CD45 (553079, BD Biosciences, San Jose, CA, United States; 103115, Biolegend), CD4 (100406, 100407, Biolegend), CD8 (100734, Biolegend), TCR-γδ (118108, 128123, Biolegend), B220 (103212, Biolegend), CD11b (101228, Biolegend), CD64 (141712, Biolegend), MHCII (107607, Biolegend), CD11c (117306, Biolegend), Ly6G (127607, Biolegend), CD86 (560582, BD), CD206 (141712, Biolegend).

For intracellular staining of nuclear hormone receptor retinoid-related orphan receptor γt (RORγt) and forkhead box protein P3 (Foxp3) in the ileum and the colon, cells were first stained for surface markers as detailed above. Then, the FIX and PERM cell fixation kit (00-5123-43, Invitrogen, CA, United States) was incubated with the cells for 40 min at 4°C. Next, permeabilisation kit (00-833-56, Invitrogen) was used with the antibody RORγt (562607, BD) and the antibody Foxp3 (17-5773-80, eBioscience, San Diego, CA, United States) for 30 min at 4°C during the cells staining.

For intracellular cytokine staining, cells were first placed in culture with cell activation cocktail (with Brefeldin A) (423303, biolegend) for 5 h. Cells were stained for surface markers as detailed above. Then, the FIX and PERM cell fixation kit (00-5123-43, Invitrogen, CA, United States) was incubated with the cells for 40 min at 4°C. Next, permeabilisation kit (Invitrogen) was used with the antibody IL-17A (506916, Biolegend) for 30 min at 4°C during the cells staining.

Zombie (423107, Biolegend) was labeled to identify dead and live cells. Appropriate isotype-matched controls from the same vendors were included in intracellular staining in order to ensure proper compensation and staining specificity. Analysis was performed with FlowJo software (version 10, Tree Star).

Mice were euthanized 1 day after MCAO. Brains were removed following perfusion with saline and 4% paraformaldehyde (Biosharp, Guangzhou, China) in phosphate buffered saline (PBS) and then soaked in 30% sucrose in PBS. The brains were embedded in OCT (Sakura Finetek, United States) solution after sinking to the bottom of the liquid, and cryosections were prepared. Then brains were cut on a freezing microtome into 25 μm-thick sections and subjected to immunofluorescence staining. After antigen retrieval treatment, sections were incubated with a blocking buffer containing 5% normal goat serum and 0.3% Triton X-100 (Thermo Fisher Scientific) at 37°C for 1 h. Then the sections were incubated with anti-Iba1 (ab178846, Abcam, United States) or GFAP (16825-1-AP, Proteintech, United States) at 4°C overnight, followed by Fluorochrome-conjugated secondary antibodies (Thermo Fisher Scientific) at room temperature for 2 h the next day. Finally, the sections were counterstained with DAPI (Thermo Fisher Scientific). Images were captured using a fluorescence microscope (Leica, Germany).

The mice were anesthetized with 2% isoflurane (vol/vol), delivered 30% O₂ and 70% N₂O at a rate of 2 L/min, and kept at 37°C during the entire process. Photoconversion was performed by using a defocused violet laser source (405 nm, ZhongShanZiGu, China). The mice were placed in the supine position with a 2 cm incision in the abdomen. After exposing the distal small intestine, the body and the rest of the intestines were covered with aluminum foil. Intestinal tissues were applied with saline to maintain moisture. The exposed intestine was irradiated for 10 min and then returned to the abdominal cavity, and the peritoneum and skin were sutured.

Extraction of DNA used the OMEGA Soil DNA Kit (M5635-02, Omega Bio-Tek, Norcross, GA, United States) following the manufacturer’s instructions. NanoDrop NC2000 spectrophotometer (Thermo Fisher Scientific, Waltham, MA, United States) and agarose gel electrophoresis were used to test the quantity and quality of extracted DNA, respectively.

PCR amplification of the near full-length bacterial 16S rRNA genes was performed using the forward primer 27F (5′-AGAGTTTGATCMTGGCTCAG-3′) and the reverse primer 1492R (5′-ACCTTGTTACGACTT-3′). The PCR products were quantified with PicoGreen dsDNA Assay Kit (Invitrogen, Carlsbad, United States) and sequenced on PacBio Sequel platform at Shanghai Personal Biotechnology Co., Ltd. (Shanghai, China). Microbiome bioinformatics were performed on QIIME2 platform with slight modification (Bolyen et al., 2019). Analysis of sequencing data was based on amplicon sequence variants (ASVs) (Bokulich et al., 2018). After chimera detection, high-quality sequences with 97% similarity were clustered into the same ASV. Classification of ASVs was performed based on the Greengenes Database.

Simpson’s diversity index was calculated using the ASV table in QIIME2, and visualized as box plots. Beta diversity was analyzed using Bray-Curtis metrics and visualized via non-metric multidimensional scaling. Heat map were clustered by UPGMA (default clustering algorithm) (Ramette, 2007) according to the Euclidean distance of the species composition data, and are arranged according to the clustering results by default; otherwise, they are arranged according to the default order of samples/groups.

Quantifications of infarct volume, bacteria content, Iba-1 or GFAP-positive cells in immunofluorescence staining, mRNA expression, protein expression, and immune cells in flow cytometry were represented in graphics with the mean value ± SD. Statistical analysis was performed using Prism (GraphPad Software, La Jolla, CA, United States). The differences between means of two experimental groups were analyzed by unpaired Student’s t-test or non-parametric test. The differences between means of three or more experimental groups were analyzed by One-Way ANOVA (Analysis of variance). Values of p ≤ 0.05 were considered statistically significant.

We designed a protocol to explore the effects of periodontitis salivary microbiota in ischemic stroke (Figure 1A). We used a cocktail of antibiotics AMNV to deplete the endogenous microbiota of mice (Figure 1A). The mice were then treated with sterile distilled water (ASDW) or saliva of periodontitis patients (ASPP) by gavage before being subjected to MCAO (Figure 1B). Assessment by TTC staining 1 day after MCAO demonstrated that the ASPP group had significantly larger cerebral infarct volume than the MCAO and ASDW group (Figures 1B,C). Modified Garcia Score showed significantly lower neurological scores in the ASPP group compared to the MCAO and ASDW group (Figure 1D). These results indicated that the gavage of saliva of periodontitis patients aggravated ischemic stroke in mice. Body weights and the fecal bacterial density were both comparable between the ASDW and the ASPP group (Supplementary Figures 1A,B).

To find out whether microorganisms in the saliva of periodontitis patients caused worse ischemic stroke outcomes, we first got the filtrate of saliva through the bacterial filter and tested the density of bacterial content in the saliva and filtrate by qPCR, the result showed bacteria were not detectable in the filtrate (Figure 1E). Then, we established MCAO model after mice were treated with AMNV and gavage of filtrate of the saliva of periodontitis patients (AFSP). The AFSP group manifested markedly smaller infarct volume than the ASPP group, suggesting that salivary microbiota mediated the effects of saliva on ischemic stroke (Figures 1F,G). Furthermore, we found that gavage of salivary microbiota of healthy individuals did not affect ischemic stroke in mice (Supplementary Figures 1C,D). Similar to previously reported (Lundmark et al., 2019; Belstrøm, 2020), the periodontitis salivary microbiota presented a distinct microbiota profile from that of healthy individuals in our study, with enrichment of periodontal pathogenic bacteria such as Prevotella intermedia, Prevotella oris, Fusobacterium nucleatum, Porphyromonas gingivalis, Porphyromonas endodontalis, and Tannerella forsythia (Supplementary Figures 1E–G). To further explore the changes of periodontitis-associated microbiota in the gut after saliva gavage, we tested these bacteria in feces and found that Prevotella intermedia, Fusobacterium nucleatum, Porphyromonas gingivalis, and Porphyromonas endodontalis were increased in the feces (Supplementary Figure 1H).

The inflammatory response following acute ischemic stroke is a critical element of brain damage. First, microglia/macrophage cells and astrocyte cells were analyzed by immunofluorescence staining of Iba-1and GFAP, respectively. The ASPP group had a significantly increased number of Iba-1⁺ cells and GFAP⁺ cells in the ischemic region compared to the MCAO and ASDW groups (Figures 2A,B). This result suggested that the more microglia/macrophages and astrocytes were rapidly activated in the ischemic region. Activated microglia/macrophage cells and astrocytes play a unique role in the secretion of pro-inflammatory cytokines. Results of qPCR and ELISA showed that the ASPP group had significantly increased expression of pro-inflammatory cytokines such as Interleukin-1β (Il-1β) and Tumor necrosis factor-alpha (Tnf-α) compared to the MCAO and ASDW groups (Figures 2C,D).

Immune cells in the ischemic brain were further analyzed by using flow cytometry. Lymphocytes were identified as CD45^highCD11b^– cells, macrophages or activated microglia were identified as CD45^highCD11b⁺ cells, resting microglia were identified as CD45^intCD11b⁺ cells, and neutrophils were identified as CD11b⁺Ly6G⁺ cells (Supplementary Figure 2A; Ponomarev et al., 2005). The ASPP group had a profound increase of lymphocytes (Figures 2E,F), macrophages/activated microglia (Figures 2E,G), and neutrophils (Figures 2H,I) in the ischemic brain compared to the MCAO and ASDW groups. The ASPP group has a decreased tendency in the number of resting microglia compared to the ASDW group (Supplementary Figure 3A).

We further analyzed the polarization of macrophages/microglia using CD86 as a marker for pro-inflammatory M1 and CD206 as a marker for anti-inflammatory M2 (Peng and Nixon, 2021). The ASPP group had a substantial increase in CD45^highCD11b⁺CD86⁺ cells in the ischemic brain compared to the MCAO and ASDW group (Figures 2J,K). There was no significant difference in the number of CD45^intCD11b⁺CD86⁺ cells between the ASPP group and the MCAO and ASDW group (Supplementary Figure 3B). CD206 was nearly not detectable in CD45^highCD11b⁺ cells or CD45^intCD11b⁺ cells after ischemic stroke (Figure 2J and Supplementary Figure 3B). These results indicated that gavage of periodontitis salivary microbiota induced more infiltration of immune cells in the ischemic brain, including lymphocytes, activated M1 population of microglia/macrophage, and neutrophils. In addition, we also tested these immune cells in ischemic brain of mice after gavage saliva of healthy people and did not detect significant increase of these immune cells (Supplementary Figures 3C,D).

We next investigated the reason for the increased infiltration and the exacerbated inflammation in the ischemic brain induced by the gavage of periodontitis salivary microbiota. IL-17 is known to play a critical role in periodontitis and is tightly associated with dysbiosis of microbiota in the oral cavity (Cheng et al., 2014; Dutzan et al., 2018). Previous studies have indicated that oral pathogens may affect the expression of IL-17 in the gut (du Teil Espina et al., 2019; Feng et al., 2020; Kitamoto et al., 2020) and that IL-17-producing cells in the gut are closely associated with inflammatory response after ischemic stroke (Benakis et al., 2016). Therefore, we first used flow cytometry to detect IL-17-producing cells in the brain. Several studies have suggested that Th17 cells and IL-17⁺ γδ T cells were two major sources of IL-17A after ischemic stroke (Waisman et al., 2015; Zhang et al., 2021). In our experiment, Th17 cells were identified as CD4⁺IL17A⁺ cells and IL-17⁺ γδ T cells were identified as TCR-γδ⁺IL17A⁺ cells (Supplementary Figure 2B). Flow cytometry analysis revealed a significant increase of Th17 cells and IL-17⁺ γδ T cells in the ischemic brain of the ASPP group compared to the MCAO and ASDW group (Figures 3A–C). And we found gavage saliva of healthy people did not increase Th17 cells and IL-17⁺ γδ T cells in the ischemic brain compared to the ASDW group (Supplementary Figures 3E,F).

IL-17 has been shown to promote inflammatory cascade by inducing inflammatory chemokines to recruit other immune cells after stroke (Gelderblom et al., 2012; Waisman et al., 2015; Wojkowska et al., 2017; Zhang et al., 2021). Compared to the MCAO and ASDW groups, results of qPCR demonstrated that the ASPP group had significantly increased mRNA expression of chemokines, which was related to IL-17A, including C-X-C motif ligand 1 (Cxcl1), C-X-C motif ligand 2 (Cxcl2), and C-C chemokine ligand 2 (Ccl2) (Figure 3D). These results together indicated that gavage of periodontitis salivary microbiota elevated IL17-producing cells, which further induced chemokine production and the infiltration of other pro-inflammatory cells after ischemic stroke.

We further hypothesized that the increased accumulation of IL17-producing cells in the ischemic brain originated from the gut after gavage of periodontitis salivary microbiota. Flow cytometry analysis showed the ASPP group had comparable CD4⁺ T cells and TCR-γδ⁺ T cells with the ASDW and MCAO groups in the small intestine and colon (Supplementary Figure 4). Meanwhile, flow cytometry analysis revealed a marked increase of Th17 cells and IL-17⁺ γδ T cells in the small intestine of ASPP group (Figures 4A,B), but not the colon (Supplementary Figures 5A,B). Next, we also found that gavage of periodontitis salivary microbiota also increased Th17 cells and IL-17⁺ γδ T cells in the small intestine of mice without ischemic stroke (Supplementary Figures 6A,B).

RORγt has been demonstrated to play an important role in intestinal homeostasis and promote the expression of IL-17 (Ivanov et al., 2006; Eberl, 2012; Kumar et al., 2021). Consistently, there was an increased expression of RORγt in CD4⁺ T cells and TCR-γδ⁺ T cells in the small intestine of ASPP group (Figures 4C,D), but not the colon (Supplementary Figures 5C,D). Treg cells express Foxp3 and are related to the alteration of IL-17 (Huber et al., 2011). However, CD4⁺Foxp3⁺ Treg cells were not significantly different in the small intestine among the groups (Supplementary Figure 6C,D). These results demonstrated that gavage of periodontitis salivary microbiota promoted IL-17-producing cells in the small intestine, which was similar to the changes in the ischemic brain, and the mechanism may be up-regulated expression of RORγt.

We then investigated the possibility of migration of immune cells from the gut to the brain in response to ischemic stroke. The MCAO group showed a significant decrease of CD4⁺ T cells in the small intestine 1 day after ischemic stroke compared to the SHAM group (Figures 5A,B). Conversely, the MCAO group manifested increases of CD11b⁺Ly6G⁺ neutrophils and CD45⁺B220⁺ B cells in the intestine compared to the SHAM group (Figures 5A,B). The other populations such as γδ T cells, CD8⁺ T cells, dendritic cells, and macrophage cells were comparable between MCAO and SHAM groups (Figures 5A,B and Supplementary Figures 7A,B). These results indicated that the reduced intestinal CD4⁺ T cells may migrate to the ischemic brain.

Kaede transgenic mice were utilized to further illustrate the migration of immune cells. Kaede transgenic mice express the Kaede fluorescent protein, which could achieve photoconversion from a green (KaedeG) to a red (KaedeR) fluorescence after being exposed to violet light (Tomura et al., 2008). Photoconversion was achieved by exposure of distal small intestines of Kaede transgenic mice to violet light (Figure 5C), and 72 h after photoconversion the mice were subjected to MCAO and sham operation. Immune cells of brain were analyzed by flow cytometry 24 h after stroke (Figure 5C). Flow cytometry analysis showed that MCAO significantly increased the numbers of KaedeR⁺CD4⁺ cells and KaedeR⁺TCR-γδ⁺ cells in the brain (Figures 5D,E).

These results suggested that intestinal immune cells rapidly responded to ischemic stroke and CD4⁺ T cells and γδ T cells could migrate to the brain after ischemic stroke. Meanwhile, we also tested the migration of macrophages. KaedeR⁺CD11b⁺CD64⁺ cells were not detectable in the brain of the SHAM and MCAO group, suggesting that macrophages did not migrate from the small intestine to the brain 24 h after ischemic stroke (Supplementary Figure 7C).

Finally, we tested whether IL-17A was required for periodontitis salivary microbiota to exacerbate ischemic stroke. Wild-type mice and Il17a^–/– mice were subjected to MCAO and ischemic brains were analyzed by flow cytometry 24 h after stroke. Flow cytometry analysis showed significantly reduced number of CD45^highCD11b^– cells, CD45^highCD11b⁺ cells, CD45^highCD11b⁺CD86⁺ cells, and CD11b⁺Ly6G⁺ cells in Il17a^–/– mice after MCAO (Figures 6A,B). These results indicated that IL-17A played a vital role in the mobilization of immune cells by MCAO to induce neuroinflammation.

Then we repeated the protocol on Il17a^–/– mice to explore the effects of periodontitis salivary microbiota in ischemic stroke (Figure 6C). TTC staining 1 day after MCAO demonstrated that the difference in cerebral infarct volume we previously observed in wildtype mice was disappeared in Il17a^–/– mice (Figures 6D,E). These results demonstrated IL-17A was required for the periodontitis salivary microbiota to exacerbate ischemic stroke.