Saliva as a Source of Biomarkers for Periodontitis and Periimplantitis

Saliva has the potential to be used as a diagnostic and monitoring tool for various diseases if biomarkers of an adequate sensitivity and specificity could be identified. Several reviews and even meta-analyses have been performed in recent years, which have found some candidate biomarkers for periodontitis, like macrophage inflammatory protein-1 alpha, interleukin-1ß, interleukin-6, matrix metalloproteinase-8, or hemoglobin. However, none of those are currently in use to replace conventional periodontal diagnostics with a periodontal probe. For periimplantitis, to date, heterogeneity of different study protocols and implant types did not permit to discover clear biomarkers, which were able to distinguish between healthy and diseased implants. Few proinflammatory cytokines, similar to periodontitis, have been characterized as adjunct tools to clinical diagnosis. The additional determination of antimicrobial peptides, bone turnover markers, and bacteria could help to enhance sensitivity and specificity in a combined model for periodontitis and periimplantitis. Furthermore, proteomic approaches might be preferred over single biomarker determinations. A global consensus is also needed to harmonize salivary sampling methods as well as procedures of biomarker analysis to ensure future comparability.


INTRODUCTION
Periodontitis is one of the most prevalent noncommunicable diseases worldwide, affecting around 796 million people in its severe form (1). The prevalence of periimplantitis has been reported inconsistently, but according to recent data, it is about between 18 and 10% per subject and per implant, respectively, and its occurrence is linked to periodontitis (2,3). The diagnosis of periodontitis and periimplantitis is still based on clinical evaluation using a periodontal probe and radiography and is currently defined after the combined AAP/EFP workshop on a new classification system 2017 (4). Those classical methods can reflect the current state of inflammation and attachment loss, but are limited in the detection of early tissue degradation with low predictive potential.
In the last decades, saliva came into the focus as a noninvasive diagnostic fluid for oral and systemic diseases (5,6). Its collection is easy, at a low cost, and does not need trained medical staff (7). Since it contains, e.g., hormones, growth factors, enzymes, antibodies as well as microbes and their products, it might be useful in the early detection of systemic diseases, like cancer, autoimmune disorders, cardiovascular disease, diabetes, or virus-related diseases (7,8). The diagnostic potential for saliva in ambulatory care and under self-collection conditions has recently been summarized in a meta-analysis for SarS-CoV-2 RT-PCR testing (9).
For periodontitis, one of the first studies about marker analysis in saliva was published by Wilton et al. (10). The conclusion of this very first study still mirrors the state of knowledge in 2021, affirming that saliva could be seen as a source of indicators for disease activity or response to treatment more than discriminating diagnostic potential. Furthermore, it was concluded in this article that the determination of markers in saliva gives no more information than one could get by direct clinical examination. In the United States, enthusiastic prospects were described in the 2010s, seeing even serum analysis for systemic diseases replaced by noninvasive saliva analysis and designating salivary diagnostic as a "game changer" for patient evaluation (5,(11)(12)(13).
In the last few years, an increasing number of studies about saliva and periodontitis with around 70 publications per year were performed principally at universities in the United States, Europe, South America, and Asia (Figure 1). To give an overview about the current state of salivary biomarkers in connection with periodontitis and periimplantitis, a systematic research was performed in PubMed to screen articles published until December 2020, using the keywords saliva AND biomarkers AND periodontitis. Six hundred and twenty-three articles were found, including systematic reviews and meta-analyses. Of those articles, 120 were not considered, since they were animal studies, articles in foreign languages without a detailed description in the English abstract, listed twice or did not include saliva analysis, periodontitis, or periimplantitis cases (Figure 2A and Supplementary Material). The research revealed that the very first studies were conducted in Russia, Austria, Japan, Germany, and the United States. Later on, more study groups around the world discovered potential salivary biomarkers associated with periodontitis and since 2010 also with periimplantitis. Evidence is growing fast with an increasing number of studies since the 1980s (Figure 2B), which allow systematic reviews and meta-analyses for certain parameters. Periimplantitis, however, has not yet been extensively studied, and the research in PubMed for articles published until December 2020 using the keywords saliva AND biomarkers AND periimplantitis revealed only 15 articles. There exist already several systematic reviews that are summarizing the most promising diagnostic markers in different categories: bacteria-derived salivary markers, hostderived salivary biomarkers associated with inflammation, and biomarkers linked to soft or hard tissue destruction (14)(15)(16). However, a lot of studies about biomarker candidates have in part not been included in those reviews due to missing sensitivity and specificity calculations, detailed description of material and methods, or other factors that could thwart the reliability of results. This selection is sometimes distorting, since biomarkers such as MIP-1 alpha are included in a meta-analysis (16), although it was investigated by 10 studies only. However, there are already longitudinal studies including MIP-1 alpha, giving it a strong candidacy as a diagnostic marker (17).
This article should give an overview about the current state of knowledge not claiming to be seen as systematic reviews that are already existing in this field (16,18,19), but to give a reflection on how this dynamic field of research evolved and yield a tremendous potential in the diagnosis and monitoring of periodontitis and periimplantitis.

SALIVA COLLECTION METHODS
Numerous saliva collection methods have been used, ranging from unstimulated whole saliva collection (20), stimulated whole saliva collection to different saliva collection systems, which could have a significant influence on biomarkers, such as proteins, in saliva (21). Proceeding of collected saliva samples, centrifugation, and storage are not overall defined and vary according to the following diagnostic method. The storage and further processing might have an important impact on the results; therefore, Henson et al. defined already in 2010 protocols for a standardized molecular analysis of salivary diagnostic constituents (22). Possible interactions of analytes with investigated biomarkers cannot be excluded, and the individual microbiome and proteome of the study population could influence the significance of the results (23). In most of the cross-sectional studies, unstimulated saliva was collected and concentrations were determined and compared between groups afterward. These procedures mostly do not take into account individual parameters, which could have a significant influence on biomarker concentrations. Individual salivary flow, gender, but mostly specific responses to inflammatory stimulus were often not included in the assessment of biomarker levels in the saliva samples. One should also be aware of the limited significance when only a single saliva sampling took place at a certain time point. Studies with stress-related biomarkers have shown, for example, that certain individuals had an immediate release of a substance, while in others, this release could be found only later, when a second sampling was performed (24).

DIAGNOSTIC METHODS
The majority of the studies were using enzyme-linked immune assays for biomarker detection but more innovative methods are upcoming, such as omics technologies, which are seen as a novel and holistic approach in the management, diagnosis, prognosis, and monitoring of oral diseases (25). Recently, lateralflow immunoassays were also shown to be suitable for the detection, prediction, or treatment outcome of periodontitis and periimplantitis (26). Furthermore, novel biodetection systems via protein fingerprinting with data processing were proposed as a convenient system for the examination of periodontal disease (27). Methods, which reduce hands-on time and easy sample preparation, like magnet-beating were shown to be suitable for preanalytic processing of saliva for automated point of care (PoC) protein analysis (28). Those PoC devices can be based on various techniques for the detection of periodontopathogens, proteins, metabolites, and small molecules (29,30).
Another promising, rapid, and label-free diagnostic biometric tool in saliva can be provided by vibrational spectroscopy (31). Diagnostic models can eventually be constructed by combining protein and microbial profiles and computing diagnostic powers via areas under the receiver-operating characteristic (ROC) curve (32). For evaluating disease progression or stability, an individual approach was seen to be most suitable using unique patient profiles for salivary expression profiles of IL-1ß, IL-6, MMP-8, and MIP1-alpha (33,34).

SALIVARY BIOMARKERS DISCRIMINATING PERIODONTAL HEALTH AND DISEASE
Proinflammatory cytokines and proteinases have been extensively investigated in mostly cross-sectional studies. Ebersole et al. described that IL-1ß, IL-6, MMP-8, and MIP-1 alpha could be seen as suitable markers to discriminate health from gingivitis and periodontitis (19). Arias-Bujanda summarized accordingly in a recent meta-analysis that the highest values of sensitivity for periodontitis were obtained for IL-1ß, MMP-8, IL-6, and hemoglobin (15). MMP-8 is by far the best investigated biomarker for periodontitis and periimplantitis and a strong biomarker candidate for detecting alveolar bone destruction (35) (Figure 2C and Table 1). MIP-1 alpha has also great potential as a periodontitis biomarker since it showed high sensitivity and specificity and a good correlation with probing depths and the onset of bone loss (41). This biomarker is particularly interesting since it is the only one that has been used in a longitudinal study of children at risk for periodontal disease (41). Most research focused on different salivary markers, and analysis of microbes or their metabolites was scarce (42). The combination of salivary biomarkers and bacteria seems promising since periodontopathic bacteria were detectable comparably to subgingival plaque sampling (43), and a cumulative use of bacterial and host-derived biomarkers showed encouraging results (44). Furthermore, biomarkers like alanine aminotransferase levels and P. gingivalis ratio could be potential indicators for the progression of periodontitis (45). Longitudinal studies are missing to evaluate the ability of salivary copy counts of major periodontopathic bacteria predicting further periodontal breakdown (46). The combination of salivary biomarkers, MMPs, and bacterial biofilm generated ROC curves with a strong diagnostic ability (47). Commercially available tests have already been developed but there are still challenges regarding the introduction of new technologies to clinical practice and adoption by dental practitioners (11).

SALIVARY BIOMARKERS AND PERIIMPLANTITIS
To date, few studies exist about salivary biomarkers in periimplantitis; nevertheless, some proteinases and cytokines have been identified to possibly serve as a diagnostic or monitoring instrument for this disease. MMP-8 levels were increased in the saliva or periimplant crevicular fluid (37,38), notably in patients who also suffered from periodontitis (48). This was also observed in patients who suffered from cardiovascular diseases, where MMP-8 was seen as a PoC biomarker (49). Increased levels of T. denticola, IL-4, IL-10 were detected in the saliva of patients with implants and type-2 diabetes (50). A list of major biomarkers for periimplantitis validated by reviews are given in Table 1.

BIOMARKERS EVALUATING PERIODONTAL THERAPY
Besides clinical parameters such as Bleeding on Probing and Clinical Attachment Level, salivary biomarkers could also be useful to monitor the response of local tissue inflammation following periodontal therapy. A decrease in enzyme levels after scaling has been described as well as increases in anti-inflammatory biomarkers, such as melatonin (51)(52)(53). Antimicrobial peptides of the innate immune system, such as calprotectin, could also be used in oral fluid diagnosis to monitor treatment outcome (54-56), a biomarker linked to the innate immune system and first described in 2000 by Kojima et al. (57) connection with periodontitis. Salivary sTLR-2 also could have the potential as a prognostic or maintenance marker for periodontitis as well as antioxidants such as total antioxidant capacity, albumins, uric acid, superoxide dismutase, and glutathione peroxidase (58,59). Arginase activity in saliva was also successfully reduced after nonsurgical periodontal therapy in accordance with a reduction in clinical and microbiological parameters (60). The determination of 8-hydroxy-deoxyguanosine levels also showed significant effects of periodontal therapy and could be regarded as a disease activity marker (61,62), which was first described by Takane et al. already in 2002 (63). In patients with metabolic syndrome, successful periodontitis therapy could be mirrored in the reduction of oxidative stress markers (64). IL-1ß, MMP-8, OPG, and MIP-1 alpha, which had in part the highest sensitivity for periodontitis, are also able to reflect the response to therapy with the potential to be seen as monitoring markers for the periodontal status as well (65).

INFLUENCING FACTORS ON SALIVARY BIOMARKERS
The heterogeneity of evidence makes it difficult to compare and review existing studies. Biomarker analysis is dependent on many factors, like the collection system, the individual salivary flow rate (66), stimulated or unstimulated sampling, time of sampling, centrifugation and processing, the storage of samples, and the detection method. Furthermore, aggravating factors for periodontitis such as smoking (67)(68)(69) or stress, but also gender have a locally or systemically influence on the secretion of biomarkers into saliva (66,70,71). It was also shown that blood contamination of saliva samples could have an impact on biomarker levels (72). Proteomic analysis revealed that total protein concentration varies according to flow rate, duration of a possible stimulus, and its nature as well as circadian rhythms (73). Several systemic diseases are interconnected to periodontitis such as diabetes or cardiovascular diseases. Some biomarkers in saliva have been investigated in patients with periodontitis and atherosclerosis or diabetes, suggesting that inflammatory cytokines and biomarkers identified after metabolic profiling could be used in diagnosis and monitoring (74,75). Rheumatoid arthritis could influence the levels of some salivary biomarkers of periodontal disease, and its therapy could significantly lower IL-1ß or TNF-alpha (76). In metabolic syndrome, dietary changes had a positive influence on inflammatory variables of periodontal disease in saliva. Nutritional intervention can therefore have a positive effect on oxidative variables as well as bacterial counts in the saliva of periodontitis patients (77).

DISCUSSION
Due to the heterogeneity of the diagnostic approaches, an organization of an International Consortium for Biomarkers of Periodontitis has already been demanded in 2015 but has not yet been established (14). Salivary biomarkers for periodontitis can still be seen to only complement regular clinical examination (78). The harmonization of saliva sampling protocols as well as definitions of power and other calculations would tremendously help to compare studies and subsume the most promising biomarkers for periodontitis. Most of the existing studies are not able to reach the quality criteria for a meta-analysis, and therefore, their results are not taken into consideration for the worldwide search for a reliable PoC diagnostic tool for periodontitis. However, considerable progress has been made to develop as sensitive and specific salivary diagnostic devices as for blood or urine testing (79).
To identify candidate biomarkers, changes in the proteome associated with periodontitis could be analyzed in databases (80). Enhanced interactions between the host and bacteria in periodontitis might also be reflected by an altered metabolomic profile of saliva (81). Salivary concentrations of inflammatory, bone turnover, and microbiological markers alone or preferably in combination could help to replace invasive diagnostic procedures and lead to a more precise and personalized dentistry for the twenty-first century (19). Multiplex panels of combined biomarkers could serve as screening tools with continued advances in this field (82). The combination of biomarkers and salivary concentrations of periodontopathic bacteria could also be used for evaluating periodontitis risk and therefore easily be used in large population surveys (83). Higher salivary MMP-8, MMP-9, OPG, and red complex periodontopathic bacteria could be used for accurate predictions of periodontal disease category, whereas T. denticola could be used together with MMP-8 for predicting periodontal disease severity (84). A. actinomycetemcomitans in combination with MMP-8 and MPO has the potential to be a trustworthy biomarker in periodontitis patients with ischemic stroke (85).
For periimplantitis, very few articles are available. Fifty percent of the articles found in the literature research comprised reviews, which indicates that mostly studies about salivary markers for periodontitis were taken into account.
Considering the panoply of studies who tried to identify reliable markers for periodontitis, it might be disappointing that very few of the more than 100 different biomarkers withstand criteria to be finally included in meta-analyses. In 2014, a critical review described the existing literature as "infant, " which is focused on validating metrics and identifying biomarkers with diagnostic potential, and further concluded that the evidence of the literature is graded as level 3 (86). A more recent metaanalysis pointed out that many promising biomarkers could not be considered due to missing validating studies of those with substantial intergroup differences (15). It was concluded that future studies should rely on latest methodological protocols (87), standardized protocols for clinical research, and focus on clearly unbiased controls (14), which can be confirmed by the present minireview. Future studies should orientate on previous methods, preferring unstimulated saliva collection in large study populations, and include confounding factors. In periimplantitis, the current heterogeneity of studies prevents a definitive evaluation of the potential of salivary diagnostic markers, and more randomized clinical trials are needed (39). Therefore, a global effort to find salivary biomarkers with high sensitivity and specificity to discriminate periodontitis from periodontal health in terms of clearly defining the best protocols, adequate sample sizes, saliva sampling techniques, the consideration of salivary flow on total biomarker concentrations, or the consideration of influencing factors would help to improve the reliability and comparability of those biomarkers. The major challenges to current saliva-based diagnostics for oral diseases are the small number of potential and valid biomarkers, the lack of real-time assessments or existing tests which are based on microbial and inflammatory cytokines that are not exclusively specific neither to periodontitis nor to periimplantitis (88).

CONCLUSIONS
No single or combination of biomarkers can so far disclose tissue destruction of periodontitis and periimplantitis effectively, and since promising biomarkers still need to be objectively demonstrated, the clinical measurements are still seen as the most reliable method of choice (89). Novel techniques of salivaomics (25), proteomics (90), peptidomics (91), metabolomics (92), and interactomics (23) might help to break the chains of single biomarker analysis and lead to a panel of biomarkers like it was already defined for IL-1ß, IL-6, MMP-8, MIP-1 alpha, and hemoglobin (16,19) to complement or even replace invasive clinical examination.

AUTHOR CONTRIBUTIONS
All authors contributed to the conception and correction of the manuscript and approved the submitted version.