Periodontitis is a chronic dental disease with a high risk of recurrence, in which the periodontal tissue is destroyed by bacteria accumulating in the teeth (1). Several risk factors such as poor oral health, hormonal changes, bacterial infection have been reported to be associated with the initiation and progression of periodontal disease (2). Periodontitis is mostly caused by gram-negative bacteria such as Aggregatibacter actinomycetemcomitans, Porphyromonas gingivalis, Prevotella intermedia, and Bacteroides forsythus (3). Recent reports have suggested the association of periodontitis with systemic diseases including cardiovascular disease, rheumatoid arthritis, type 2 diabetes mellitus (T2D), hypertension and inflammatory bowel disease (4-6).

Diabetes is accompanied by complications such as hypertension, cardiovascular disease, skin lesions, and obesity (7, 8). A multidirectional relationship between diabetes and periodontitis has been reported, with diabetes increasing the risk of periodontitis and periodontal inflammation negatively affecting control of blood glucose levels (9-11). Non-alcoholic fatty liver disease (NAFLD) is frequently observed in patients with T2D and is known to be associated with increased insulin resistance triggered by fatty acid accumulation, inflammation and reactive oxygen species production (12-14). Moreover, it has been reported in many studies that periodontitis contributes to the progression of NAFLD (15, 16).

It has been shown that hepatic CD36, which acts as a transcriptional regulator of PPARγ, is significantly upregulated in obese rats with periodontitis (17). Depletion of CD36 from hepatocytes attenuates fatty liver disease and ameliorates insulin sensitivity in obese mice (18). As another important fatty acid regulator, the fatty acid-binding protein 4 (FABP4) can induce systemic disease progression (19). FABP4 expression is increased in macrophages upon infection with P. gingivalis or Fusobacterium nucleatum, affecting both T2D and periodontitis progression (20). In addition, adipocytes from patients with T2D have been reported to contain reduced levels of FOXO1, an important regulator of glucose and lipid homeostasis (21).

The link between periodontitis and T2D or fatty liver disease is associated with the production and expression of inflammatory factors including proinflammatory cytokines, chemokines, and metabolic regulators (22). The higher levels of tumor necrosis factor α (TNFα), interleukin (IL)-1β, and CRP induce insulin resistance by impairing insulin action (23, 24). A marked elevation of components of the NLR Family Pyrin Domain Containing 3 (NLRP3) inflammasome at both the gene and protein levels have been reported in patients with periodontitis and T2D (25). Moreover, several inflammatory responses mediated by proinflammatory cytokines or inflammasome activation have been suggested to play an important role in the induction of hepatic inflammation in NASH (26, 27).

Since obesity is a major cause for chronic diseases including T2D and NAFLD (28), we used HFD-induced obese mice model. To investigate the correlation of periodontitis and T2D or fatty liver disease, HFD-fed mice were inoculated with P. gingivalis to develop the periodontitis, and then checked for crucial pathophysiological factors related to the link between the diseases.

Obese mice with P. gingivalis infection exhibit alveolar bone loss and increased CRP level

To investigate the correlation between periodontitis and liver malfunction, periodontitis was induced by P. gingivalis infection on the mandibular molars of 8-week old mice or by direct administration into the stomach using a feeding catheter. HFD was supplied to the mice to generate a mouse model of obesity (Fig. 1A). Body weight was measured once every 2 weeks during the experiment. P. gingivalis treatment increased the body weight of mice. Although the body weight of HFD-fed groups significantly increased compared to the NC group, there was no difference among HFD-fed (HFD), HFD-fed P. gingivalis-infected (HFD-p.g.), and HFD-p.g. (P.O.) groups (Fig. 1B). To confirm that the experimental periodontitis was appropriately induced by P. gingivalis infection, the extent of alveolar bone loss was measured. The mice infected with P. gingivalis in the oral cavity displayed a significant increase in alveolar bone loss compared to the NC group, as indicated by a decrease in the relative amounts of bone around the mandibular molars of the infected mice (Fig. 1C). We next performed serum biochemical analysis by measuring glucose, metabolic disease-related factors (cholesterol, HDL, LDL), liver enzymes (AST, ALT), and the levels of the systemic inflammatory marker (CRP) in blood. The levels of glucose, cholesterol, HDL, LDL, AST and ALT were significantly increased in HFD-fed mice. However, P. gingivalis treatment did not further elevate the levels of glucose, cholesterol and lipoproteins, as well as liver enzymes. CRP levels were significantly increased in P. gingivalis-treated groups compared to those in the NC group (Fig. 1D). These results indicate that HFD feeding can induce systemic metabolic disease with liver damage and that the P. gingivalis infection does not alter the overall levels of serum biochemical factors except the CRP level.

P. gingivalis infection did not affect the glucose metabolism in HFD-fed mice

We next measured blood glucose using intraperitoneal glucose tolerance test (IPGTT) to investigate the bidirectional relationship between periodontitis and glucose homeostasis. After 6 hours of fasting, the blood glucose levels were measured. Fasting blood glucose was higher in all HFD-fed groups than in the NC group but was not altered by P. gingivalis administration (Fig. 2A). The IPGTTs also did not show any significant difference in glucose levels in mice infected with P. gingivalis via different routes. In the HFD-p.g. (P.O.) group, the glucose level in IPGTT at 9 and 12 weeks was the highest, but these changes were statistically insignificant compared to that in the HFD-fed groups (Fig. 2B). These results suggest that P. gingivalis treatment did not significantly aggravate HFD-induced glucose intolerance in our experimental settings.

P. gingivalis exacerbates NAFLD, particularly by up-regulating lobular inflammation

Given that fatty liver disease can be exacerbated in association with periodontitis, we then analyzed lipid accumulation, inflammation and hepatocyte degeneration in liver tissues. Upon histological examination by hematoxylin and eosin (H&E) staining (Fig. 3A), the evaluation of steatosis, inflammation, and hepatocyte ballooning were conducted to calculate NASH score (Fig. 3B-E). Our results showed that livers of NC-p.g. group mice had minimal steatosis, and no signs of lobular inflammation and hepatocyte ballooning (NASH score 0-1). The total score of HFD-fed mice (NASH score 4-7) was significantly elevated compared to that in the NC group (Fig. 3B). The HFD-p.g. group (NASH score 5-9) exhibited significantly increased total score compared with HFD control group (NASH score 4-7). Overall, increased steatosis, lobular inflammation, and hepatocyte ballooning were displayed in the HFD groups (Fig. 3C-E). In the HFD groups infected with P. gingivalis (p.g. and p.g.[P.O.]), we observed extensive lobular inflammation foci (Fig. 3A). Consistently, the inflammation score was significantly increased in both P. gingivalis inoculated groups (Fig. 3D). This finding indicates that HFD-fed mice have steatohepatitis, which is diagnosed as a total score of 5 or higher, and P. gingivalis infection impairs fatty liver disease to a greater extent through the induction of increased inflammatory responses.

P. gingivalis infection can exacerbate fatty liver disease through CD36 and PPARγ upregulation

We next assessed the expression level of several risk factors that play a crucial role in the progression of fatty liver disease. The mRNA expression levels of genes related to inflammation and lipid or glucose metabolism in liver were determined (Fig. 4A). We analyzed the expression of several metabolic regulators including Foxo1, PparγC1α, Cd36, and Fabp4 in the liver. Decreased Foxo1 expression was observed in the HFD and HFD-p.g. groups. Importantly, the expression level of PparγC1α, a key regu-lator of adipocyte differentiation and lipid metabolism, was significantly upregulated upon P. gingivalis infection in HFD-fed mice. Moreover, HFD-fed mice showed increased Cd36 mRNA expression, which increased further following P. gingivalis infection. The Fabp4 mRNA expression was high in HFD-fed mice, but did not show additional elevation upon P. gingivalis exposure. The expressions of Tnfα and Il-1β were upregulated in HFD-p.g., but this change was statistically insignificant. The expression levels of Il-4 and Il-17 were not altered. The mRNA levels of inflammasome components including Nlrp3 and Casp1 were not changed upon HFD feeding or P. gingivalis infection. We next investigated the phosphorylation of Akt, AMPK and mTOR, which represent insulin responsiveness and metabolic regulation, respectively. Akt phosphorylation was slightly decreased in the HFD-fed mice group and, both intraoral and peroral administration of P. gingivalis did not alter the level of p-Akt (Fig. 4B). Moreover, the expression levels of p-mTOR and p-AMPK did not show any consistent pattern of change (Fig. 4C). Taken together, our findings demonstrate that P. gingivalis infection can exacerbate fatty liver disease by disturbing hepatic lipid metabolism through the CD36-PPARγ pathway.

Periodontitis is an inflammatory disease that leads to alveolar bone loss and further exerts various adverse impacts on systemic health. It is widely accepted that immune responses to peri-odontogenic pathogen play key roles in the progression of metabolic diseases (29). In the present study, we reported that P. gingivalis, a major candidate correlating periodontitis and metabolic diseases (30), can aggravate fatty liver signatures in mouse models of obesity and periodontitis.

A previous study has experimentally demonstrated that P. gingivalis or Lipopolysaccharide (LPS) from the bacteria can impair HFD-induced insulin resistance and glucose tolerance (31). Although we used slightly modified protocols from this study, however, in our study, neither intraoral nor peroral administration of P. gingivalis did not significantly increase the level of fasting glucose or glucose intolerance. This discrepancy might be due to the differences in the environment of animal breeding or virulent potency in P. gingivalis strains. Other studies have shown consistent results with our findings. Li et al. explored the onset and severity of diabetes using mouse model of both T1D and T2D with P. gingivalis infection. Despite the development of alveolar bone loss, periodontitis did not alter the glucose metabolism, determined by fasting glucose, body weight change and glucose tolerance (32). In addition, another study by Wang et al. proposed the similar findings. In the study, P. gingivalis inoculation did not significantly regulate the levels of fasting glucose in db/db mice, Tallyho/JngJ mice and streptozotocin-treated mice (33). Kuraji et al. suggested that ligature-induced periodontitis could be more effective than the infusion of peri-odontogenic bacteria without ligation (34). Further cumulative studies might be required to establish more reproducible multiple animal models for periodontitis with other systemic diseases, using a ligature-induced model with the inoculation of peri-odontogenic bacteria.

The upregulation of components of various inflammasomes in patients with periodontitis was reported by Bostanci et al. and Garcia-Hernández et al. (35). More importantly, the latter study demonstrated that NLRP3 inflammasome components were overexpressed in the gingival tissue of patients with periodontitis and uncontrolled T2D (25). Therefore, in the present study, we observed the expression of components of the inflammasome complex, including NLRP3, NLRC4, AIM2, CASP1 and IL-1β, in both periodontal tissues and liver tissues. However, in periodontal tissues, none of the inflammasome components was significantly regulated by HFD feeding or P. gingivalis infection (data not shown). One can envision that P. gingivalis-mediated inflammatory responses in periodontal tissues might be resolved at the time point when the samples were harvested. In liver tissues, the level of IL-1β expression was elevated in the HFD-fed group with intraoral P. gingivalis infection, which may indicate inflammasome activation could be involved in the regulation of live metabolism.

Despite the slight symptoms in alveolar bone loss and related inflammation, we here observed that the administration of P. gingivalis accelerated the progression of fatty liver disease. In particular, inflammation in liver tissue was increased in HFD-fed mice with both intraoral and peroral administration of P. gingivalis. Previous studies revealed that an increased level of hepatic CD36 could be a pivotal regulator in the impairment of liver metabolism (36). In addition, a recent study by Ipsen et al. reported that CD36 is abnormally increased in patients with NASH and hepatic steatosis (37). Consistent with these findings, our data showed that the level of CD36 in liver tissues was increased by HFD feeding and further increased by P. gingivalis administration to a greater extent. Moreover, the expression of PPARγ, a transcription factor in the downstream of CD36, was significantly upregulated in HFD-fed mice with P. gingivalis infection, compared to HFD-fed mice. Moreover, a large body of other previous studies have proven that PPARγ/CD36 pathway critically contributes to HFD-induced NAFLD through the increase of free fatty acid uptake and triglyceride synthesis in the liver. These studies collectively demonstrated that the increased expression of PPARγ results in elevated CD36 expression, which aggravates HFD-induced fatty liver and insulin resistance (18, 38, 39).

The limitation of our study is that the animal model we used here is not a perfect model for the study on the correlation between periodontitis and liver metabolism and that symptoms for both diseases are too mild. Therefore, further investigations using severe periodontitis model are required since damages in periodontal tissue and subsequent exacerbating effects of periodontitis on systemic disease are all cumulative. To this end, increased dosage of bacterial infection, shorter interval of bacterial inoculation, or combined treatment with molar ligation might be adopted for future study. In conclusion, our data revealed that the administration of P. gingivalis in HFD-fed mice aggravated NASH in mice through the upregulation of CD36 and PPARγ, pivotal regulators in lipid metabolism. These findings might propose the novel therapeutic targets in the treatment of periodontal metabolic liver diseases.

The detailed methods are described in the “Supplementary Materials and Methods”.

Induction of diabetes and periodontitis mice

8-week-old C57Bl/6 wild-type (WT) mice were housed in specific pathogen-free controlled conditions on 12 h day light/dark cycle. Mice were randomly divided into five groups: negative control group (NC), negative control with P. gingivalis treated group (NC-p.g.), high-fat diet group (HFD), high-fat diet with P. gingivalis treated group (HFD-p.g.) and high-fat diet with peroral administration P. gingivalis group (HFD-p.g.[P.O.]). For the P. gingivalis group, 10⁹/ml colony-forming unit (CFU) of P. gingivalis suspended in 100 μl PBS (Gibco, Waltham, Massachusetts, USA) with 2% carboxy-methylcellulose (Sigma, Saint Louis, Missouri, USA) was treated at the mandibular first molars, three times a week for 6 weeks. Each group was divided into two subgroups and fed with either normal chow or high-fat diet for 14 weeks. The induction of obesity and periodontitis model was confirmed by measuring body weight every 2 weeks and measuring fasting glucose level at 9, 12 and 14 weeks. This study protocol was confirmed by Ethics Committee of Pusan National University (ethics code: PNU-2019-2130). Mice were monitored for reaching criteria of humane endpoint. Mice exhibiting signs of distress, such as panting, breathlessness, abnormal bodily posture and shaggy fur, were sacrificed using CO₂ asphyxiation. At the time points of tissue harvesting, mice were sacrificed by CO₂ asphyxiation. Every effort was made to minimize the suffering of mice.

Histological analyses

To assess hepatic morphology and scoring, liver samples are fixed in 10% neutral buffered formalin and stained with hematoxylin and eosin. Histological scoring was performed by three-stage ranges system as follows: Degree of steatosis (0-3), lobular inflammation (0-3), hepatocyte ballooning (0-2) (40).

This work was supported by the National Research Foundation of Korea (NRF) grant funded by the Korea government (MSIT) (No. NRF-2018R1A5A2023879). We would like to thank Editage (www.editage.co.kr) for English language editing.

The authors have no conflicting interests.