Acceleration of HDL-Mediated Cholesterol Efflux Alleviates Periodontitis

Abstract

Periodontitis (PD) is a common inflammatory disease known to be closely associated with metabolic disorders, particularly hyperlipidemia. In the current study, we demonstrated that hypercholesterolemia is a predisposing factor in the development of PD. Logistic regression analysis revealed a strong positive correlation between PD and dyslipidemia. Data from in vivo (PD mouse model subjected to a high cholesterol diet) and in vitro (cholesterol treatment of gingival fibroblasts [GFs]) experiments showed that excess cholesterol influx into GFs potentially contributes to periodontal inflammation and, subsequently, alveolar bone erosion. Additionally, we compared the protective efficacies of cholesterol-lowering drugs with their different modes of action against PD pathogenesis in mice. Among the cholesterol-lowering drugs we tested, fenofibrate exerted the most protective effect against PD pathogenesis due to an increased level of high-density lipoprotein cholesterol, a lipoprotein involved in cholesterol efflux from cells and reverse cholesterol transport. Indeed, cholesterol efflux was suppressed during PD progression by downregulation of the apoA-I binding protein (APOA1BP) expression in inflamed GFs. We also demonstrated that the overexpression of APOA1BP efficiently regulated periodontal inflammation and the subsequent alveolar bone loss by inducing cholesterol efflux. Our collective findings highlight the potential utility of currently available cholesterol-lowering medications for the mitigation of PD pathogenesis. By targeting the acceleration of high-density lipoprotein-mediated cellular cholesterol efflux, a new therapeutic approach for PD may become possible.

Keywords: HDL; apolipoprotein A-I; fibroblasts; hypercholesterolemia; hyperlipidemia; lipoprotein.

Figure 1.

An abnormal cholesterol uptake in the human gingival fibroblasts (GFs) aggravates periodontitis (PD) phenotypes. (A) Lipid profiling using serum from healthy individuals (noninflamed, n = 9) and patients with chronic PD (inflamed, n = 10). (B) Messenger RNA (mRNA) levels of the indicated genes involved in PD pathogenesis from noninflamed (healthy) or inflamed (PD) human gingival tissues (n = 6). (C, D) Representative images of gingival tissues from human patients with PD (C) and maxilla region of ligature-induced PD mice (D) following micro–computed tomography (µCT), IL6 immunostaining with percentage of area of positive staining, and filipin staining with relative fluorescence intensity. Scale bar: 25 μm. (E) The cellular levels of total cholesterol, free cholesterol, and cholesteryl ester in human GF and periodontal ligament (PDL) cells that were treated with IL1β (2 ng/mL) or TNFα (50 ng/mL) for 24 h (n = 3). (F) mRNA levels of the indicated genes involved in PD pathogenesis in human GF and PDL cells, treated with cholesterol (200 μM) for 36 h (n ≥ 3). (G) Representative μCT images and the bone mineral density (BMD) and cementum–enamel junction and alveolar bone crest (CEJ-ABC) distance analysis in the maxilla region of ligature-induced PD mice fed a regular diet (RD) or a high-cholesterol diet (HCD) for 13 wk (n = 5, mice per group, total 10 mice). n indicates the number of biologically independent samples or human specimens or mouse number per group. Values are presented as mean ± standard error of the mean (SEM) based on the 2-tailed t test (A, B, E, G) and 1-way analysis of variance with Tukey’s test (F). *P < 0.05, **P < 0.01, ***P < 0.001.

Figure 2.

Administration of cholesterol-lowering drugs attenuates high-cholesterol diet (HCD)–induced metabolic periodontitis (PD). (A) Total cholesterol (TC), high-density lipoprotein cholesterol (HDL-C), and low-density lipoprotein cholesterol (LDL-C) levels in serum samples collected from mice that were administered with cholesterol-lowering drugs for 9 wk and HCD fed for 13 wk (n ≥ 23). (B–E) Representative micro–computed tomography (µCT) and hematoxylin and eosin (H&E) staining images (B, D) with bone mineral density (BMD) and cementum–enamel junction and alveolar bone crest (CEJ-ABC) distance analysis (C, E) in the maxilla region of male (B, C) and female (D, E) mice administered vehicle (DMSO) or cholesterol-lowering drugs for 9 wk and HCD feeding for 13 wk (n = 8, male mice per group, total 48 mice; HCD + vehicle, atorvastatin: n = 11, regular diet [RD], HCD + fenofibrate, niacin, ezetimibe: n = 10, female mice per group, total 62 mice). Scale bar: 100 μm. n indicates the number of mice per group. Values are presented as mean ± standard error of the mean (SEM) based on the 2-tailed t test (A, C, E). *P < 0.05, **P < 0.01, ***P < 0.001.

Figure 3.

Expression of cholesterol efflux mediators and effect of high-density lipoprotein (HDL) in human gingival fibroblasts (GFs) treated with proinflammatory cytokines. (A) Cellular level of total cholesterol, free cholesterol, and cholesteryl ester in cholesterol (25 μM) preloaded human GFs treated with HDL in the presence of IL1β (2 ng/mL) or TNFα (50 ng/mL) for 12 h (n = 3). (B, C) Messenger RNA (mRNA) levels of the indicated genes involved in periodontitis (PD) in cholesterol (25 μM) preloaded human GFs treated with HDL in the presence of IL1β (2 ng/mL) (B) or TNFα (50 ng/mL) (C) for 12 h (n = 3). (D, E) mRNA levels of cholesterol efflux mediators (ABCA1, SCARB1, ABCA5, ABCG1) in human GFs treated with IL1β (D) or TNFα (E) for 24 h (n ≥ 3). (F) Cellular levels of total cholesterol in cholesterol (25 μM) preloaded human GFs treated with IL1β or TNFα in the presence of HDL (50 µg/mL) for 24 h (n = 3). n indicates the number of biologically independent samples. Values are presented as mean ± standard error of the mean (SEM) based on 1-way analysis of variance and Tukey’s test. *P < 0.05, **P < 0.01, ***P < 0.001.

Figure 4.

Upregulation of APOA1BP in human gingival fibroblasts (GFs) treated with proinflammatory cytokines and in periodontal tissues. (A, B) Messenger RNA (mRNA) levels of high-density lipoprotein (HDL)–related genes (HDL component: APOA1, APOA2, APOE, PON1; HDL interaction: LIPC, LIPG, PLTP, CETP, LCAT) in human GFs treated with IL1β (A) or TNFα (B) for 24 h (n = 4). (C) APOA1BP mRNA levels in human GFs treated with IL1β or TNFα for 24 h (n = 4). (D) Protein expression and relative band intensity of APOA1BP in human GFs treated with IL1β or TNFα for 24 h (E, F) Representative APOA1BP immunostaining images with percentage of area of positive staining in gingiva from healthy individuals and patients with periodontitis (PD) (E) and ligature-induced PD mice (F). Scale bar: 25 μm. n indicates the number of biologically independent samples. Values are presented as mean ± standard error of the mean (SEM) based on 1-way analysis of variance and Tukey’s test. *P < 0.05, **P < 0.01, ***P < 0.001.

Figure 5.

APOA1BP-mediated cholesterol efflux alleviates periodontitis (PD) progression. (A, B) APOA1BP, MMP1, PTGS2, IL6, and IL8 messenger RNA (mRNA) levels in human gingival fibroblasts (GFs) infected with Ad-APOA1BP (Ad-A) in the presence of IL1β (2 ng/mL) (A) or cholesterol (200 µM) (B) for 48 h (n = 3). (C) Total cholesterol levels in human GFs infected with Ad-C or Ad-APOA1BP (Ad-A) in the presence of IL1β (2 ng/mL) or cholesterol (200 µM) for 24 h (n = 4). (D–F) Representative APOA1BP immunostaining with percentage of area of positive staining (D), micro–computed tomography (µCT), and hematoxylin and eosin (H&E) staining images with bone mineral density (BMD) and cementum–enamel junction and alveolar bone crest (CEJ-ABC) distance analysis (E) and representative MMP1 and IL8 immunostaining with percentage of area of positive staining (F) in the maxilla region of ligature-induced PD mice injected intragingivally with Ad-C and Ad-Apoa1bp (1 × 10⁹ Plaque-Forming Unit (PFU) per 3 μL) (n = 7, mice per group, total 21 mice). Scale bar: 100 μm (immunostaining image, above). Scale bar: 25 μm (immunostaining image, below). Scale bar: 100 μm (H&E staining image). n indicates the number of biologically independent samples or the number of mice per group. Values are presented as mean ± standard error of the mean (SEM) based on the 2-tailed t test (E) and 1-way analysis of variance with Tukey’s test (A, B, C). *P < 0.05, **P < 0.01, ***P < 0.001.