Human gingiva, including sulcus/pocket epithelium and underlying connective tissue, were obtained from either healthy controls (n = 20) or patients with periodontitis (n = 20) with alveolar bone loss confirmed by radiography. Patients and healthy controls clinical characteristics and demographics are summarized in Supplementary Table S1. Written informed consent was obtained from each patient. Human gingival tissues were used to analyze SLIT2 protein expression by ELISA. The Medical Ethics Committee of the Affiliated Stomatology Hospital of Guangzhou Medical University approved all protocols dealing with patients (approval number: KY2019032).

Total of 47 wild-type C57BL/6 and 47 Slit2 transgenic (Slit2-Tg) C57BL/6 mice were used in this study. Wild-type mice were purchased from Guangdong Medical Animal Experiment Center (Guangdong, China). Slit2-Tg mice were obtained from Prof. Lijing Wang’s lab, Vascular Biology Research Institute, School of Life Sciences and Biopharmaceutics, Guangdong Pharmaceutical University, Guangzhou, China. Slit2-Tg mice were generated according to the previously reported protocol (Yang et al., 2010). All the mice used were healthy and immune-normal, euthanized after their experimental periods. All studies were performed in 8-week-old male mice unless otherwise indicated. The Experimental Animal Ethics Committee of Guangzhou Medical University approved all animal care and study protocols (GY2020-004).

Bone marrow-derived macrophages (BMMs) were harvested from the bone marrow of six C57BL/6 mice as described previously (Spiller et al., 2016). BMMs were expanded in T75 culture flasks supplemented with Gibco RPMI 1640 medium (Life Technologies, Carlsbad, CA, United States) and 10% FBS. For the in vitro studies cultures were supplemented with 30 ng/ml recombinant mouse macrophage colony-stimulating factor (M-CSF, CB34, Novoprotein, Shanghai, China). Cultures were used for further mRNA and protein expression analysis.

A ligature-induced periodontitis (LIP) model was developed in Slit2-Tg and wild-type mice, as described previously (Abe and Hajishengallis, 2013).

Micro-CT scanned the maxilla to evaluate the morphological changes in the alveolar bone. The distance from the mesial buccal cemento-enamel junction to the alveolar bone crest (CEJ-ABC) of the second molar was measured as a reference for bone loss. The level of bone resorption was calculated as described previously (Abe and Hajishengallis, 2013).

Periodontitis-affected periodontal tissue (PAPT; 0.5 cm× 0.5 cm × 1.0 cm) surrounding the teeth cut from mice maxilla, including gingiva, periodontal ligament, and a part of alveolar bone, was stored at −80°C for use. Frozen periodontal tissue samples were solubilized in lysis buffer containing 10 mM of Tris–HCl, pH 7.4, 150 mM of NaCl, 1 mM of EDTA, and 1% Triton X-100 at 4°C for 20 min. The tissue lysates were subjected to centrifugation at 15000 × g at 4°C for 20 min. The supernatant was collected, and their protein contents were determined using Coomassie Plus Protein Assay Reagent (Pierce, Rockford, IL, United States). Cell lysate protein (50 μg) was separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and transferred to polyvinylidene difluoride (PVDF) membranes. The membranes were then blocked with 5% non-fat dry milk for 1 h and incubated overnight at 4°C with primary antibodies: anti-TRAF6 (1:1000 dilution, ABclonal, Wuhan, China), anti-P38 (1:1000 dilution, ABclonal), anti-P-P38 (1:1000 dilution, Cell Signaling Technology, MA, United States), or anti-TRAP (1:1000 dilution, ABclonal). The membranes were then washed and incubated with HRP-conjugated secondary antibodies (1:5000 dilution, ABclonal) at room temperature. An enhanced chemiluminescence detection system (Thermo Scientific, MA, United States) detected the immunoreactive protein bands. Densitometry data for band intensities results were generated by analyzing the images using ImageJ software. β-actin was used as a reference housekeeping protein.

After scanning micro-CT images, the collected maxillae were fixed in 4% formaldehyde overnight, followed by decalcification in 10% EDTA, which was exchanged every week for 1 month. The specimens were cut in 3 mm thickness along the long axis of the molars. Next, the samples were dehydrated and embedded in paraffin blocks. Serial sections of 5 μm thickness were cut and mounted on poly-l-lysine-coated slides. Hematoxylin and eosin (H&E) staining was performed separately on consecutive tissue sections. Images were captured using a microscope. The CEJ-ABC distance between the maxillary first and second molars was measured as a net bone loss on tissue sections, using Image J software (Wang et al., 2015).

Thick tissue sections (5 μm) were deparaffinized and rehydrated, followed by heat-induced epitope retrieval (Xiao et al., 2017). Methanol containing 3% H₂O₂ was used to block the endogenous peroxidase for 20 min. Tissue sections were then blocked with 10% bovine serum albumin (BSA) and incubated for overnight at 4°C with primary antibodies: monoclonal anti-CD34 [EP373Y], (ab81289, Abcam, Cambridge, United Kingdom), anti-CD45 (20103-1-AP, Proteintech, United States), monoclonal anti-F4/80 [BM8], (123110, BioLegend, CA, United States) or isotype controls. After washing with PBS three times, tissue sections were incubated with a corresponding goat anti-rabbit secondary antibody (SolelyBio, Beijing, China) for 30 min at 37°C. Slides were incubated with diaminobenzidine (Cell Signaling Technology, MA, United States) followed by hematoxylin counterstaining. Finally, all stained sections were dehydrated through a series of graded alcohol baths of increasing concentration, cleared in xylene, and mounted with coverslips. Stained tissue sections were visualized under a microscope (Leica, Germany), and images were captured. The visual fields between the first molar and the second molar were photographed. The number of immunostaining-positive cells was counted in the entire area of each image.

Tissue sections were stained using acid phosphatase (TRAP) kit (G1050, Servicebio, Wuhan, China) following the manufacturer’s instructions. TRAP stained histological tissue section were examined under microscope, and images of predefined areas were captured. TRAP-positive cells appeared bright red. TRAP-positive osteoclasts, in the alveolar bone tissue section between the first molar and the second molar, were counted.

Periodontal tissue removed from mice maxillae (periodontitis-affected and control group) was homogenized in a mortar on liquid nitrogen, and total RNA was isolated using PureLink RNA Mini Kit (12183018A, Thermo Fisher Scientific, United States). RNA samples with 1.8 to 2.0 of the OD260/280 ratio were used for analysis, and total RNA reverse were transcribed using Takara PrimeScriptTM RT Master Mix in T100 Thermal Cycler (Bio-Rad, United States). The transcribed cDNA was used for RT-qPCR using Takara TB-Green Premix Ex Taq in LightCycler RT-PCR system (LC-480 II, Roche, Switzerland). GADPH was used as a reference housekeeping gene. The 2^–ΔΔCt method was used to analyze the relative mRNA expression levels. The primers used for RT-qPCR are listed in Supplementary Table S2.

ELISA analyzed the protein level in human and mice periodontal tissue protein extract, mice serum, conditioned medium of mice BMMs culture, and mice BMMs cell lysates. The procedure of periodontal tissue protein extract has been described in the Western blot analysis section above. For mice serum separation, blood samples were collected from orbital venous plexus, and the serum was separated by 5 min centrifugation at 2,000 rpm. BMMs culture conditioned medium and cell lysates were prepared as follows. On the fifth day of BMMs culture, 100 ng/ml LPS was added. After 48 h of LPS treatment, conditioned medium (CM) was collected, and attached cells in the culture harvested for cell lysate preparation. SLIT2 level in human periodontal tissue lysate was analyzed by human SLIT2 ELISA kit (Cusabio, Wuhan, China) following the manufacturer’s instruction. The protein levels of IL-6, IL-1β, and TNF-a in periodontal tissue protein extract and CM were measured by mouse ELISA Standard Kit (RayBiotech, Atlanta, United States). The protein level of CD16/32 in mice periodontal tissue protein extract and the BMMs cell lysate were measured by Mouse ELISA Standard Kit (Cusabio, Wuhan, China). The protein levels of SLIT2 in mice periodontal tissue protein extract and mice serum were measured by mouse SLIT2 ELISA Kit (Cusabio, Wuhan, China) following the manufacturer’s instruction.

The cell suspension was prepared from mice periodontal tissue, including gingiva, periodontal ligament, and a part of alveolar bone for flow cytometry analysis (Mizraji et al., 2013). Prior to the test, cells were counted and cell viability was evaluated by the Zombie NIR^TM Fixable Viability Kit (423105, BioLegend, Beijing, China). Cells were stained with 1 μg of anti-F4/80 (123110, BioLegend, CA, United States) solution, monoclonal anti-CD11b [M1/70], (101206, BioLegend, CA, United States) or anti-rabbit IgG per 1 × 10⁶ cells for 30 min on ice in the dark. After two times washing with PBS, the cells were resuspended in 300 μl PBS and transferred to flow tubes. Flow cytometric analysis was performed using the FACSAria III Cell Sorter (BD Biosciences, San Jose, CA, United States).

The wild-type and Slit2-Tg periodontitis mice periodontal tissue samples without enzymatic treatment were used for RNA sequencing. Total RNA was extracted from periodontal tissue using the TRIzol kit (Invitrogen, Carlsbad, CA, United States), according to the manufacturer’s protocol. The RNA concentration was determined using Qubit, and RNA amount and purity of each sample was assessed with a NanoDrop spectrophotometer (NanoDrop 2000, Wilmington, DE, United States). RNA was isolated in 40 μl of DEPC water and stored in −80°C. Briefly, RNA-seq libraries were prepared by using the Illumina TruseqTM RNA sample prep Kit and were sequenced using an Illumina HiSeq. Cutadapt was used to obtain paired-end reads (Martin, 2011). The quality of the RNA-samples used for RNA-seq is presented in Supplementary Table S3. The reads were aligned with Hisat2 (v 2.1.0) to GRCm38 with default parameters (Kim et al., 2015). Only the data matched to the reference genome were used for subsequent analysis. The mapped reads of each sample were assembled using StringTie (Pertea et al., 2015). Then, the transcriptomes from all samples were merged to reconstruct a comprehensive transcriptome using perl scripts. After the final transcriptome was generated, StringTie (Pertea et al., 2015) and Ballgown (Frazee et al., 2015) was used to estimate the expression levels of all transcripts. Traditional singular enrichment analysis (SEA) (Rodriguez et al., 2016) was used for the enrichment analysis of GO terms and pathways. The enrichment P-value calculation was performed with Fisher’s exact test. RNA-sequence data are available at Supplementary Data S1. All raw RNA-sequencing data can be accessed from NCBI SRA database (SRA accession: PRJNA639904)¹.

All data are presented as mean ± standard deviation (SD). Statistical analysis was performed using SPSS 19.0 statistical software (SPSS IBM, Armonk, NY, United States). For where appropriate (comparison of two groups only), two-tailed t-tests were done. In all cases, data from at least three independent experiments was used, and P value of less than 0.05 was considered to indicate a statistically significant difference.

First, we examined the SLIT2 expression in human gingival tissues of 20 periodontitis patients and 20 age and gender-matched healthy individuals. The SLIT2 protein level in periodontitis-affected human gingival tissue was 4.4-fold higher compared to healthy gingival tissue (Figure 1A). The SLIT2 protein level in the periodontal tissue of Slit2-Tg mice was 2.1-fold higher compared to wild-type mice (Figure 1B). The SLIT2 protein expression in wild-type mice was upregulated by 4.6-fold compared to healthy periodontal tissue (Figure 1B). Similarly, a 3.9-fold higher expression of SLIT2 protein was observed in Slit2-Tg mice PAPT compared to healthy periodontal tissue (Figure 1B). Slit2-Tg mice PAPT showed a 1.8-fold higher expression of SLIT2 protein in comparison to wild-type mice PAPT (Figure 1B). Additionally, the serum of wild-type periodontitis mice showed 5.0-fold higher expression of SLIT2 protein compared to wild-type healthy mice (Figure 1C). Our results indicate the upregulation of SLIT2 expression in gingival tissue during periodontitis in human and in PAPT of wild-type mice or Slit2-Tg mice. Moreover, periodontitis condition upregulates the level of serum SLIT2 in wild-type mice showing the systemic effect of periodontitis.

We performed micro-CT scanning to analyze the effect of LIP on alveolar bone loss. Intact periodontal tissue and normal alveolar bone were observed in healthy wild-type and Slit2-Tg mice (Figure 1D). Loss of periodontal tissue and alveolar bone was observed in LIP wild-type and Slit2-Tg mice (Figure 1D). The higher CEJ-ABC distance indicates the higher effect of periodontitis on periodontal tissue damage and alveolar bone loss. CEJ-ABC distance in periodontitis wild-type mice was 1.9-fold higher compared to healthy wild-type mice (Figure 1E). Similarly, CEJ-ABC distance in periodontitis Slit2-Tg mice was 2.3-fold higher compared to healthy Slit2-Tg mice (Figure 1E). Slit2-Tg periodontitis mice showed a 1.3-fold higher CEJ-ABC distance compared to wild-type periodontitis mice (Figure 1E). This finding was corroborated by the histological images as well (Figure 1F). Slit2-Tg mice periodontitis exhibited more bone loss and bone destruction in the alveolar crest height than in wild-type mice periodontitis (Figures 1D,E). This effect was further corroborated by the histological images (Figure 1F). The histological images showed the infiltration of inflammatory immune cells in gingival epithelial and alveolar bone resorption (Figure 1F). Histomorphometric analysis showed 1.3-fold higher periodontitis-induced alveolar bone loss in Slit2-Tg mice compared to wild-type mice (Figure 1G). Our results indicate that the overexpressed SLIT2 in periodontitis aggravates the consequences of the disease, such as the destruction of periodontal tissue and alveolar bone loss.

Inflammatory cytokines promote osteoclastogenesis, and higher osteoclastogenesis is associated with higher bone loss. In this study, we analyzed the role of the higher SLIT2 expression in periodontitis on osteoclastogenesis. More numbers of TRAP-positive osteoclasts were visualized in the PAPT section of Slit2-Tg mice compared to wild type mice (Figure 2A). A higher degree of vasculature in gingival tissue aggravates periodontitis. In this study, we performed CD34 immunostaining in PAPT to detect microvasculatures. Higher CD34 expression was observed in PAPT of Slit2-Tg mice compared to wild-type mice (Figure 2B). Quantitative analysis showed a 1.4-fold higher number of osteoclasts in PAPT in Slit2-Tg mice compared to wild-type mice (Figure 2C). Quantitative analysis revealed 1.4-fold higher microvessel density (MVD) in PAPT of Slit2-Tg mice compared to wild-type mice (Figure 2D).

Similarly, we performed immunohistochemical staining for leukocyte specific antigen, CD45, in gingival tissue sections to analyze the effect of SLIT2 overexpression on leucocytes infiltration in PAPT. More numbers of CD45 positive cells were observed in periodontitis-affected gingival tissue of Slit2-Tg mice compared to wild-type mice (Figure 2B). Quantitative analysis results revealed a 1.2-fold higher number of CD45-positive cells in PAPT of Slit2-Tg mice compared to wild-type mice (Figure 2E). This indicates that the higher degree of leucocyte infiltration in PAPT of Slit2-Tg mice, possibly amplify periodontal inflammation.

RNA sequencing (RNA-seq) of mice PAPT was performed to explore the underlying mechanisms of the regulatory effects of SLIT2 on periodontitis. Among 24 common inflammatory factors, Cxcr2, Il18, and Il1β were differentially upregulated by more than 2-fold in PAPT of Slit2-Tg mice compared to wild-type mice (Figure 3A (i)). The heat map corroborated the differential upregulation of Cxcr2, Il18, and Il1β mRNA in PAPT of Slit2-Tg mice compared to wild-type mice (Figure 3A (ii)). Highly expressed pro-inflammatory cytokines in periodontal tissue abrogate periodontitis. RT-qPCR results showed 2. 3-, 1. 8-, and 1.6-fold higher expression of Il6 (Figure 3B), Il1β (Figure 3C), and Tnf (Figure 3D), respectively, in PAPT of Slit2-Tg mice compared to wild-type mice. Protein level expression of IL-6 (Figure 3E), IL-1β (Figure 3F), and TNF-α (Figure 3G) were 8-, 4- and 2.6-fold higher, respectively, in PAPT of Slit2-Tg mice compared to wild-type mice. This result indicates that the SLIT2-overexpression intensifies the inflammation in periodontitis.

We further analyzed the differential expression of 32 osteoclastogenesis related genes in PAPT by RAN-seq. The expression of Fosl2, Il1β, Mapk14, Ncf4, and Pik3cb, was differentially upregulated in PAPT of Slit2-Tg mice compared to wild-type mice (Figures 4A (i),(ii)). We further analyzed the mRNA expression of the key osteoclastogenesis regulators (Acp5, Ctsk, and Nfatc1) in PAPT by RT-qPCR. Notably, the expression of Acp5, Ctsk, and Nfatc1 in PAPT of Slit2-Tg mice was 18. 4-, 2. 2-, and 1.8-fold higher, respectively, compared to wild-type mice (Figures 4B–D). Additionally, the expressions of the majority of osteoblastogenesis associated genes were not altered in PAPT of Slit2-Tg mice compared to wild-type mice. (Figure 4E and Supplementary Figure S1). These results indicate that the SLIT2 overexpression mainly triggers inflammation-induced osteoclastogenesis but might not affect the osteoblast function during periodontitis.

More numbers of macrophages (F4/80 positive cells) were observed in the PAPT section of Slit2-Tg mice compared to wild-type mice (Figure 5A (i)). Quantitative analysis of infiltrated macrophages from immunohistochemistry images showed 1.3-fold higher numbers of macrophages in PAPT of Slit2-Tg mice compared to wild-type mice (Figure 5A (ii)). This finding was further verified by the data from flow cytometry analysis (Figures 5B (i),(ii)). Flow cytometry analysis revealed 1.4-fold higher numbers of CD11b+F4/80+ cells in the PAPT of Slit2-Tg mice compared to wild-type mice (Figure 5B (ii)). Prior to the test, cell viability was evaluated and above 95% viability of cells was ensured (data not shown). This finding indicates that the SLIT2 overexpression in periodontitis facilitates the infiltration of macrophages in periodontal tissue.

Proinflammatory cytokines trigger the M1 macrophage polarization. The relative expression of Il1β and Il8 was upregulated in Slit2-Tg PAPT (Figure 5C). However, the expression of M2 macrophage polarization related cytokines (such as Il10 and Il4) remained the same in Slit2-Tg PAPT and wild-type mice (Figure 5D). The protein level of CD16/32, a marker of M1 macrophage polarization, was upregulated (2.8-fold) in PAPT of Slit2-Tg mice compared to wild-type mice (Figure 5E). Similarly, under inflammatory condition (LPS treatment), cell lysate from in vitro cultured macrophages of Slit2-Tg mice expressed 4.7-fold higher CD16/32 compared to wild-type mice (Figure 5F). Similarly, inflammatory condition upregulated IL-6, IL-1β, and TNF-α expression in Slit2-Tg mice macrophages by 3. 6-, 2. 6-, and 1.9-fold, respectively, compared to wild-type mice macrophages (Figures 5G–I). This result indicates that SLIT2 overexpression amplifies macrophage infiltration and M1 macrophage polarization in PAPT.

ROBO1, ROBO2, ROBO3, and ROBO4 are 4 receptors of SLIT2 ligand. In RNA-seq heat map, Robo1 and Robo2 gene expression were differentially upregulated, and Robo3 and Robo4 remained unchanged in Slit2-Tg mice PAPT (Figures 6A–D). Quantification of RNA-seq data showed 1.5-fold, and RT-qPCR analysis showed a 2.0-fold higher expression of Robo1 in PAPT of Slit2-Tg mice compared to wild type mice (Figure 6A (ii)). However, the expression of other receptors remained unchanged. This result indicates the possible role of the SLIT2/ROBO1 axis on SLIT2-mediated effects in periodontitis.

RNA sequencing data detected the expression of total of 53801 genes in mice PAPT (Supplementary Data S1). Among those, 134 genes were differentially expressed in Slit2-Tg mice. Among 134 genes, 90 were downregulated, and 44 were upregulated (Figure 6E). GO pathway enrichment analysis showed the significant changes in expression pattern genes associated with the 22 key signaling pathways related to inflammation, vital cell functions, and others (Figure 6F). We further analyzed the expression pattern of various genes associated with the MAPK pathway. RNA-seq data analysis showed a higher expression of Dusp1, Dusp5, Flna, Hspa1a, Hspa1b, Il1β, MAPK14, Pla2g4b, Pla2g4d, and Pla2g4f in Slit2-Tg mice PAPT (Figure 7A). Heat map from RNA-seq data of 20 MAPK related genes are shown in Figure 7B. Similarly, RT-qPCR result showed 2.1- and 1.6-fold higher expression of Traf6 and p38 in PAPT of Slit2-Tg mice compared to wild-type mice (Figures 7B,C). Western blot analysis showed a 2-fold higher phosphorylation p38 in PAPT of Slit2-Tg mice compared to wild-type mice (Figures 7D (i–iii)). Our finding indicates the possible role of activated MAPK pathway in SLIT2-mediated effects on periodontitis.