Characteristics of cellular senescence along with periodontitis progression.

(A) Representative image of and semi-quantification of SA-β-gal staining in healthy (n=4 field) and periodontitis (n=6 field) patient gingiva, scale bar=40μm or 20μm. (B) Representative images of IHC staining and semi-quantification of p16, p21 and H3K9me3 in healthy and periodontitis patient gingiva (n=3 field), positive cells were indicated by black arrow, scale bar=40μm. (C) Analysis strategy of ligature-induced periodontitis (LIP) mouse model. (D) Representative image of IHC staining and semi-quantification of p16 in mouse gingiva of health and LIP post 3, 7 and 14 days (n=3 field), scale bar=40μm. (E) Western blot images and semi-quantification of p16 protein levels in control (CON) and LIP post 7 days (LIP 7D) mouse gingiva (n=4 independent experiments). (F) qrt-PCR analysis of p16, p21and Tp53 in control (CON) and LIP 7D mouse gingiva (n=3 independent experiments). Ep: Epithelium; LP: Lamina propria; Alv: Alveolar bone; Teeth. Data are expressed as mean ± SD. *p<0.05, **p<0.01, ***p<0.001. ****p<0.0001.

Cellular Senescence of gingival fibroblasts in periodontitis.

(A and B) UMAP diagram and single-cell annotation of cells clusters for the healthy and periodontitis patient gingiva from public dataset GSE164241. (C) Histogram of gingival tissue cell ratio in healthy and periodontitis patients. (D) The violin plot showing cellular senescence score of cell groups in healthy and periodontitis gingiva. (E) GSEA enrichment analysis of cellular senescence pathway in fibroblasts among periodontitis compared to those in healthy gingiva. (F) The violin plot showing cellular senescence score in subgroups in gingiva of healthy, mild and severe periodontitis patient from public dataset GSE152042. (G) Immunofluorescence staining and semi-quantification of p16 positive fibroblasts in healthy and periodontitis patient gingiva. p16 (red), Vimentin (green), and nuclei (blue), Ep: Epithelium; LP: Lamina propria. White arrow indicates double positive cells, scale bar=40μm, n=3.

CD81 is identified as the potential marker of senescent gingival fibroblast.

(A) UMAP diagram illustrated the cell subclusters of fibroblasts from public dataset GSE164241. (B) Histogram of fibroblasts subclusters ratio in healthy and periodontitis gingiva respectively. (C) The violin plot showing cellular senescence score of each fibroblast subcluster in healthy and periodontitis gingiva. (D) GO enrichment analysis of each fibroblast subcluster. Fibroblast subcluster 1 shows enrichment of aging process highlighted by red. (E) Cellular localization of the top 20 marker molecules in fibroblasts subcluster 1. CD81 protein, located at cell membrane, was highlighted by red. (F) Density map of CD81 expression among fibroblast subcluster and re-annotation of fibroblast subcluster according to GO analysis. (G) Immunofluorescence staining and semi-quantification of CD81 positive fibroblasts in healthy and periodontitis patient gingiva. VIM (green), CD81 (red) and nuclei (blue), Ep: Epithelium; LP: Lamina propria. White arrow indicates double positive cells, scale bar=40μm, n=3. Data are expressed as mean ± SD. * P <= 0.05, **P<= 0.01, ***P<= 0.001, ****P<= 0.0001.

CD81+ gingival fibroblasts are terminally differentiational cell with high SASP expression.

(A) Heatmap showing the relative expression for SASP genes in each fibroblast subclusters. (B) Trajectory reconstruction of each fibroblast subclusters. (C) Monocle pseudotime analysis revealing the progression of gingival fibroblast clusters. (D) Upper panel: Heatmap showing the scaled expression of differently expressed genes in trajectory as in (C), cataloged into four gene clusters (labels on left). Bottom panel: GO analysis of expressed genes whose expression increases as the differentiation trajectory progresses. (E) SASP-related genes with increased expression as the differentiation trajectory progresses.

CD81+fibroblasts possibly recruit neutrophils via C3/C3aR1 axis.

(A) The relative number of interactions between CD81+fibroblasts and other cell type in periodontitis gingiva. (B) Significant increased Ligand-Receptor interaction derived from CD81+fibroblasts. C3-C3AR1 signaling axis increased between CD81+ fibroblast and neutrophil especially, which was highlighted by red. (C) The heatmap showing the communication patterns of the Complement signaling pathway between fibroblasts and immune cell type in healthy and periodontitis gingiva. (D) The expression level of four representative genes in Complement signaling pathway. (E) Representative image of and semi-quantification of IHC staining regarding C3 in healthy and periodontitis gingiva. Scale bar=40μm, n=3. (F) Elisa analysis of human-C3 secretion between healthy human gingival fibroblasts (H-HGF, n= 16 samples) and periodontitis human gingival fibroblasts (P-HGF n= 23 samples). (G) Elisa analysis of human-C3 secretion in healthy human gingival fibroblasts with (Pg-LPS group) or without (CON group) 1 μg/mL Pg-LPS stimulated, n=6 samples. (H) H&E image and representative spatial mapping of CD81 and SOD2 in healthy and periodontitis gingiva from public dataset GSE152042. Co-localization in CD81 and SOD2, a neutrophil marker, was found in the periodontitis gingiva. Ep: Epithelium; LP: Lamina propria. Data are expressed as mean ± SD. *P <= 0.05, **P<= 0.01, ***P<= 0.001, ****P<= 0.0001.

Senolytics therapy alleviates inflammation and bone resorption in LIP model.

(A) Strategy of LIP mouse model treated by a senolytic drug Navitoclax. (B) Representative H&E staining image of each group, Inflammatory cells were labelled by black arrows, scale bar=20μm. (C) IHC staining and (a) semi-quantification of p16 in each group, Positive cells were labelled by black arrows, n=3, scale bar=20μm. (D) IHC staining and (b) semi-quantification of H3K29me3 in each group, Positive cells were labelled by black arrows, n=3, scale bar=20μm. (E) Immunofluorescence staining and (c) semi-quantification of CD81 (red), Vimentin (green), and nuclei (blue) in control and LIP mouse gingiva, n=3 mice, scale bar=20μm. White arrow indicates double positive cells. (F) IHC staining and (d) semi-quantification of C3 in each group, Positive cells were labelled by black arrows, n=3, scale bar=20μm. (G) IHC staining and (e) semi-quantification of MPO in each group, Positive cells were labelled by black arrows, n=3 field per group, scale bar=20μm. (H) IHC staining and (f) semi-quantification of CTSK in each group, Positive cells were labelled by black arrows, n=3 field per group, scale bar=20μm. Ep: Epithelium; LP: Lamina propria; Alv: Alveolar bone; Teeth. Data are expressed as mean ± SD. *P <= 0.05, **P<= 0.01, ***P<= 0.001, ****P<= 0.0001.

Metformin alleviates inflammation and bone resorption in LIP model via inhibiting the interaction between CD81+ fibroblasts and neutrophils.

(A) Strategy of LIP mouse model treated by metformin. (B) 3-D visualization of the maxilla and quantified by the Bone volume/ tissue volume (BV/TV) ratio (a) the cement-to-enamel to alveolar bone crest, CEJ-ABC distance (b) indicated by red line, n= 6 mice per group. (C) IHC staining and semi-quantification (c) of p16 in each group, Positive cells were labelled by black arrows, n= 6 field per group, scale bar=40μm. (D) Immunofluorescence staining and (d) semi-quantification of CD81 positive fibroblasts in each group. Vimentin (green), CD81 (red) and nuclei (blue). White arrow indicates double positive cells, n=3, scale bar=20μm. (E) IHC staining and (e) semi-quantification of C3 in each group, n=6, scale bar=40μm. (F) IHC staining and (f) semi-quantification of MPO, a neutrophils marker, in each group, Positive cells were labelled by black arrows, n=6, scale bar=40μm. Ep: Epithelium; LP: Lamina propria; Alv: Alveolar bone; Teeth. Data are expressed as mean ± SD. *P <= 0.05, **P<= 0.01, ***P<= 0.001, ****P<= 0.0001.

Bulk RNA-seq analysis of ligature-induced periodontitis mice model.

(A) Heatmap and (B) Volcano plots of differentially expressed genes in mouse gingiva at LIP 7D compared to the CON (n=3 samples each group). Representative senescence-related genes are indicated as green. blue dots indicate differentially down-regulated genes; red dots indicate differentially up-regulated genes. Significantly different expression genes with | log2FC | > 1 and FDR < 0.05. (C) GSEA enrichment analysis of cellular senescence gene sets in mouse gingiva at LIP 7D compared to the CON. (D) GO enrichment analysis with upregulated (red) and downregulated (blue) genes shown in (A). Aging biological process was significantly enriched and highlighted by red.

Bulk RNA-seq analysis of ligature-induced periodontitis mice model in regard to cellular senescence.

(A) Volcano plots of differentially expressed genes in mouse gingiva at LIP 7D compared to the control (CON). Representative senescence-associated secretory phenotypes (SASP) genes are indicated as green. (B and C) GSEA enrichment analysis of Citrate Cycle and Oxidative Phosphorylation gene sets in mouse gingiva at LIP 7D compared to the CON, which indicated mitochondrial dysfunction in periodontitis. (D-F) GSEA enrichment analysis of Pi3k-Akt, Mapk and Nf-Kappa B signaling pathway gene sets in mouse gingiva at LIP 7D compared to the CON, which indicated senescence-associated signaling pathway were activated in periodontitis.

(A) UMAP diagram illustrated the cell clusters of GSE164241. (B) Marker genes of each cell were shown in the dot plot. (C) UMAP diagram and single-cell annotation of cells clusters from GSE152042. (D) Histogram of gingival tissue cell ratio in healthy, mild and severe periodontitis patients from GSE152042.

(A) SA-β-gal staining and (B) semi-quantification of human gingival fibroblasts stimulated by different concentrations of Pg-LPS (n=3). Black arrow indicates SA-β-gal positive cells. Data are expressed as mean ± SD. *P <= 0.05, **P<= 0.01, ***P<= 0.001, ****P<= 0.0001.

(A) The heatmap representing metabolic pathways in each fibroblast subclusters, which indicated fatty acid biosynthesis, arachidonic acid metabolism, and steroid biosynthesis were significantly upregulated in CD81+ fibroblasts. (B) The flow chart representing the metabolism of arachidonic acid, which could be converted into prostaglandins (PGs) and Thromboxane As (TXAs) by COX-1 or COX-2. (C) The dot plot representing that PTGS1 gene (encoding COX1 protein) and PTGS2 gene (encoding COX2 protein) are significantly higher in CD81+ gingival fibroblasts compared to other fibroblasts subclusters.

(A) Immunofluorescence staining and semi-quantification of p16 (red), Vimentin (green), and nuclei (blue) in control and LIP mouse gingiva, n=3 mice, scale bar=50μm. (B) Immunofluorescence staining and semi-quantification of CD81 (red), Vimentin (green), and nuclei (blue) in control and LIP mouse gingiva , n=3 mice, scale bar=50μm. (C) Immunofluorescence staining of p16 (red), VIM (green), CD81 (cyan), and nuclei (blue) in LIP mouse gingiva, scale bar=20μm. white arrow indicates triple positive cells. Ep: Epithelium; LP: Lamina propria; Alv: Alveolar bone; Data are expressed as mean ± SD. *p<0.05, **p<0.01, ***p<0.001. ****p<0.0001.

(A) Representative image of and (B) semi-quantification of IHC staining regarding C3 in control and LIP mouse gingiva. Scale bar=40μm, n=3. (C) Representative image of and (D) semi-quantification of IHC staining regarding MPO in control and LIP mouse gingiva. Scale bar=40μm, n=3. Black arrow indicates neutrophil cells. (E) Immunofluorescence staining and (F) semi-quantification of CD81 (red), C3 (green), and nuclei (blue) in control and LIP mouse gingiva, n=3 mice, scale bar=50μm. White arrow indicates double positive cells. (G) Immunofluorescence staining of CD81 (red), MPO (green), and nuclei (blue) in control and LIP mouse gingiva. Scale bar=50μm. Ep: Epithelium; LP: Lamina propria; Alv: Alveolar bone; Data are expressed as mean ± SD. *p<0.05, **p<0.01, ***p<0.001. ****p<0.0001.

(A) The violin plot showing cellular senescence score in mouse gingiva of healthy (H), LIP (P) and LIP treated with metformin (PM) groups from public data GSE242714. (B) H&E staining images in each group. Inflammatory cells were labelled by black arrows, scale bar=40μm. (C) Representative image and (a) semi-quantification of Masson staining, in which collagen fibers were stained into blue, in each group. Collagen fiber was labelled by white arrows, n= 6, scale bar=40μm. (D) IHC staining and (b) semi-quantification of p21 in each group. Positive cells were labelled by black arrows, n=6, scale bar=40μm. (E) IHC staining and (c) semi-quantification of H3K9me3 in each group. Positive cells were labelled by black arrows, n=6, scale bar=40μm. Ep: Epithelium; LP: Lamina propria; Alv: Alveolar bone; Teeth. Data are expressed as mean ± SD. *p<0.05, **p<0.01, ***p<0.001. ****p<0.0001.

(A) In vitro experiment model of Pg-LPS induced senescence of human gingival fibroblasts (HGFs) treated with or without metformin (MET). (B) SA-β-Gal staining and (C) semi-quantification of Pg-LPS induced senescence of human gingival fibroblasts (HGFs) treated with or without metformin (MET). Black arrow indicates positive cells, n=3, scale bar=40μm. (D) Western blot image of human C3, CD81 and p16 protein levels of Pg-LPS induced senescence of human gingival fibroblasts (HGFs) treated with or without metformin (MET).

(A) Immunofluorescence staining and (B) semi-quantification of CD81+ p16+fibroblasts in Pg-LPS induced senescence treated with or without metformin (MET). CD81 (red), P16 (green) and nuclei (blue). White arrow indicates double positive cells, n=3, scale bar=20μm. (C) Immunofluorescence staining and (D) semi-quantification of CD81+ C3+fibroblasts in Pg-LPS induced senescence treated with or without metformin (MET). CD81 (red), C3 (green) and nuclei (blue). White arrow indicates double positive cells, n=3, scale bar=20μm. Data are expressed as mean ± SD.* P <= 0.05, **P<= 0.01, ***P<= 0.001, ****P<= 0.0001.

Schematic overview of the CD81+ senescent gingival fibroblast-neutrophil axis in periodontitis progression.

We propose that the initial periodontal inflammation is triggered by the CD81+ senescent gingival fibroblast induced by bacterial virulence like Pg-LPS. CD81+ senescent gingival fibroblast could exaggerate inflammation in the periodontal tissue via secreting SASPs and recruiting neutrophils by C3. In addition, Navitoclax and Metformin could alleviate the cellular senescence of the fibroblast and rescue the uncontrolled inflammation and bone resorption.