Periodontitis is an oral condition that is characterized by progressive destruction of the tooth supporting tissues, including both collagenous connective tissue of the gingiva and the tooth anchoring jawbone (1). In periodontally healthy states, indigenous polymicrobial communities at mucosal surfaces maintain an ecological balance via inter-microbial and host-microbial interactions. However, genetic and acquired factors, most notably smoking, but also obesity, immune deficiencies, immunoregulatory defects, diabetes mellitus and other systemic diseases may disrupt this homeostatic balance, leading to selective growth of species with the potential for destructive inflammation. This condition, known as dysbiosis, underlies the development of periodontitis in susceptible hosts. The pathogenic process is not linear, but involves a positive-feedback loop between a dysbiotic microbiota on the tooth surface below the gum line and the host inflammatory response (2). The molecular pathways underpinning tissue destruction have not been clarified, and there are currently no biomarkers available to help clinicians predict onset of periodontitis or identify individuals at risk of a more aggressive disease course.

Besides the local and detrimental effects on the periodontium that jeopardize tooth retention, the disease is strongly connected to a number of other non-communicable diseases, including rheumatoid arthritis (RA) (3). Different models for a causal and reciprocal relationship between periodontitis and RA have been described (4). In support of a causal link between periodontitis and RA, results from experimental studies show that periodontitis can exacerbate arthritis (5). Our findings of increased alveolar bone loss in individuals with high levels of anti-citrullinated protein antibodies (ACPA), before RA onset, imply that periodontitis predates the onset of ACPA-positive RA (6). However, it remains elusive whether there is a causative link.

The gram-negative anaerobic oral bacterium Porphyromonas gingivalis (P. gingivalis) is frequently detected in subgingival plaque of periodontitis patients (7). P. gingivalis expresses multiple virulence factors for its growth, survival, and immune evasion (8), including the arginine gingipains (Rgp), potent extracellular cysteine proteases which efficiently degrade host proteins by cleaving polypeptides C-terminal of arginine (9, 10). Protein degradation by gingipains can facilitate further enzymatic processing by Porphyromonas peptidylarginine deaminase, resulting in citrullinated neoepitopes (11). Gingipains have also been described to contribute directly to the inflammatory response by cleavage of protease activated receptor-2 on the neutrophil surface, triggering formation of neutrophil extracellular traps with the release of endogenously citrullinated proteins (12). Notably, citrulline-reactive B cells have been successfully isolated from gingival tissue of periodontitis patients, suggesting that break of tolerance and ACPA production in RA may take place in inflamed gums (13). Moreover, gingipains are highly immunogenic (14), and we have previously shown an increased anti-Rgp antibody response, not only in individuals with periodontitis compared to controls, but also in RA, especially ACPA-positive RA, compared to controls even before RA onset (15, 16). We have previously demonstrated associations between anti-Rgp IgG and alveolar bone loss, as well as presence of ACPA in PAROKRANK, a case-control study including patients hospitalized for a first myocardial infarction. Interestingly, we also found a particularly strong association between anti-Rgp IgG and severe periodontitis in a subfraction of PerioGene North study (17). Therefore, P. gingivalis antibodies, more specifically gingipain antibodies, are gaining increased attention as a potential biomarker for individuals with periodontitis and concomitant risk for systemic autoimmunity (13, 16–18).

In summary, accumulating evidence suggests that periodontitis is linked to RA-specific autoimmunity through a loop in which P. gingivalis promotes ACPA formation. In the present study, we address this link by analysing the association between periodontal clinical parameters, anti-Rgp IgG and ACPA in 987 individuals within the large and well-characterized PerioGene North case-control study (19).

In this study, we demonstrate that serum anti-Rgp IgG levels cannot clearly discriminate individuals with periodontitis from periodontally healthy but define a subgroup of periodontitis patients with high periodontal inflammation and advanced alveolar bone loss. Moreover, we find no strong support for a link between periodontitis or anti-Rgp IgG and RA-specific autoimmunity, e.g. ACPA.

Antibodies against periodontal pathogens have been investigated for many years (24). In particular, antibodies against P. gingivalis have been shown to associate with periodontitis (25), and among 29 different P. gingivalis antigens, Hirai and co-authors recently showed that the gingipains elicited the most sensitive and specific IgG responses in patients with periodontitis (14), supporting the use of Rgp in our study.

The data presented herein clearly show that individuals with periodontitis present with higher anti-Rgp IgG levels than periodontally healthy controls, which is in accordance with our previous findings (17). It should be noted that high anti-Rgp IgG levels were also detected among some of the controls. Studies have suggested that periodontally healthy individuals are able to produce highly functional antibodies, protecting against P. gingivalis colonization, whereas periodontitis patients produce less functional antibodies with lower avidity (26). Hence, we speculate that the presence of anti-Rgp antibodies in periodontally healthy controls in our study reflects a healthy immune response, while high levels of anti-Rgp IgG that we see in a subset of periodontitis patients indicate high P. gingivalis counts and high pathogenic gingipain load. In support of this scenario, several studies showed decreased anti-P. gingivalis antibody levels in response to successful periodontal treatment, following lower bacterial counts (27–29).

Our previous data (17), based on 41 cases and 39 controls from PerioGene North, indicated a very good capacity for anti-Rgp antibodies to discriminate cases from controls (AUC=0.79). Thus, we wanted to further validate our data in the entire cohort. However, in the present study, anti-Rgp IgG had poor discriminatory ability to distinguish cases from controls (AUC=0.63). The discrepant data between the two studies suggests that the sub-cohort is unrepresentative of the entire PerioGene North cohort. Moreover, the results in the present study are in agreement with our previous findings in PAROKRANK, a case-control study designed to investigate periodontal health in relation to myocardial infarction, where anti-Rgp IgG also could not separate periodontitis patients from controls efficiently, but was still significantly elevated in individuals with severe forms of periodontitis compared to controls (17).

In here, we present significant associations between anti-Rgp IgG and specific clinical periodontal parameters, periodontal inflammation and alveolar bone loss. Thus, collectively, these data indicate that anti-Rgp antibodies are not likely to perform well as predictive biomarkers for periodontitis per se, while they may still be valuable indicators of risk for severe periodontal damage.

Gingipains are known for their proteolytic capacity, including the degradation of components of the complement system and extracellular matrix structures (30). They can directly contribute to gingival bleeding and tissue breakdown by degrading fibrinogen (31). Interestingly, gingipains have also been shown to promote osteoclastogenesis in vitro, thus implicating a possible role in bone resorption (32). Noteworthy, both gingipain vaccines and gingipain-inhibitors have been used successfully to prevent P. gingivalis-induced periodontal damage in experimental models (33, 34). Moreover, we have previously shown elevated anti-Rgp antibody levels in RA versus controls (15), and given the results in the present study – demonstrating an association between high anti-Rgp antibody levels and advanced alveolar bone loss – anti-Rgp IgG levels should potentially also be examined in relation to articular bone loss in RA. Notably, a recent study links oral bacteremia to RA flares (35), and other studies have detected P. gingivalis DNA in RA synovial joints (36, 37). Thus, the mechanistic role of gingipains in inflammation-mediated bone resorption should be further explored.

Several studies have linked periodontitis and periodontal pathogens to ACPA-positive RA (15, 35, 38), including a recently published meta-analysis demonstrating that RA patients with periodontitis have significantly higher ACPA levels than RA patients without periodontitis (39). However, until now, it has not been clear if periodontitis alone is sufficient to elicit a systemic ACPA response, as previous studies are sparse and present conflicting results. The present study is, to the best our knowledge, the first to investigate ACPA status in a large and well-characterized periodontitis case-control study. Although we observed a higher frequency of ACPA-positivity among periodontitis patients compared to controls, in line with our previous report (17), this was not significant after adjustment for the confounding effects of smoking and age. Given that only 15 of 478 individuals with periodontitis were ACPA-positive, we were underpowered to analyze ACPA in relation to periodontal parameters.

In accordance with the data herein, a cross-sectional study by Lew and co-authors did not find differences in ACPA levels when comparing periodontitis patients without RA and healthy controls (40). In here, we did not find an association between ACPA status and anti-Rgp. Similarly, Svard et al. found no association between anti-Rgp antibodies in saliva and periodontitis or ACPA status (41). In summary, the data we present are based on a large and well-characterized periodontitis cohort and suggest that there is not convincing evidence for that a systemic ACPA response is connected to periodontitis, or elevated anti-Rgp IgG levels in periodontitis patients. Still, this does not rule out that locally produced ACPA in inflamed gingival tissue are involved in onset and/or progression of RA in susceptible individuals. Notably, ACPA have been detected in gingival crevicular fluid from non-RA individuals (42).

The main strength with the present study is the significant power of PerioGene North, owing to its large number of study participants and the high resolution clinical periodontal data provided by trained specialists. Within this material, we have previously described serum levels of hs-CRP that are in line with results presented in a meta-analysis (43), implying that PerioGene North is well suited for investigating, and detecting, serum markers associated with periodontitis. Unlike the periodontitis cases in PAROKRANK, where more than half were affected by myocardial infarction, the overall presence of general diseases in PerioGene North is low (19). Therefore, the elevated anti-Rgp IgG levels observed in PerioGene North are less likely to be affected by the presence of other systemic diseases.

A limitation with our study was that the case group, due to the original purpose of PerioGene North to study genetic polymorphisms, were not matched to controls on age. However, all controls were 34 years or older, which is a strength as many other studies have included young adults as periodontally healthy controls. Still, this resulted in an older case group and all analyses were subsequently adjusted for age as well as for smoking, both well-known risk factors for periodontitis (44, 45). In our previous report (19), post-hoc analysis within different age groups confirmed that the distinguishable protein profile was not a result of the case group being older.

In conclusion, our data show that periodontitis per se is not associated with the presence of a systemic ACPA response and that ACPA-positive individuals in this cohort do not have an increased antibody response to Pg virulence factor arginine gingipain. While anti-Rgp IgG showed poor ability to separate patients with periodontitis from periodontally healthy in an efficient manner, elevated levels associate significantly with the subset of patients that have active periodontal inflammation and advanced alveolar bone loss. Prospective studies could reveal whether high anti-Rgp IgG levels can serve as a biomarker to predict a more aggressive disease course. Future studies should also address the actions of gingipains in the context of bone resorption, to clarify molecular mechanisms, and further explore whether blocking gingipains could be a future treatment for periodontitis.