A systematic review of the protein composition of whole saliva in subjects with healthy periodontium compared with chronic periodontitis

Context Periodontitis is a chronic multifactorial inflammatory disease linked to oral microbiota dysbiosis. This disease progresses to infection that stimulates a host immune/inflammatory response, with progressive destruction of the tooth-supporting structures. Objective This systematic review aims to present a robust critical evaluation of the evidence of salivary protein profiles for identifying oral diseases using proteomic approaches and summarize the use of these approaches to diagnose chronic periodontitis. Data sources A systematic literature search was conducted from January 1st, 2010, to December 1st, 2022, based on PICO criteria following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines and by searching the three databases Science Direct, Scopus, and Springer Link. Study selection According to the inclusion criteria, eight studies were identified to analyze the proteins identified by proteomics. Results The protein family S100 was identified as the most abundant in patients with chronic periodontitis. In this family, an increased abundance of S100A8 and S100A9 from individuals with the active disease was observed, which strongly relates to the inflammatory response. Moreover, the ratio S100A8/S100A9 and the metalloproteinase-8 in saliva could differentiate distinct periodontitis groups. The changes in protein profile after non-surgical periodontal therapy improved the health of the buccal area. The results of this systematic review identified a set of proteins that could be used as a complementary tool for periodontitis diagnosis using salivary proteins. Conclusion Biomarkers in saliva can be used to monitor an early stage of periodontitis and the progression of the disease following therapy.


Introduction
According to the 216 Global Health Burden, out of 328 diseases, permanent caries and periodontal disease ranked in positions 1 and 11 among the most prevalent human diseases [1]. Periodontitis is a chronic multifactorial inflammatory disease linked to oral microbiota dysbiosis. A recent global meta-analysis showed that the prevalence of apical periodontitis in at least one tooth was 52% of pooled samples worldwide [2]. Although biofilm formation is the primary etiological factor, other host-related factors, such as genetics, immunologic and environmental factors, are associated with periodontitis, making it challenging to establish its diagnosis in the early stages [3]. This disease progresses to infection that stimulates a host immune/inflammatory response, with progressive destruction of the tooth-supporting structures. Moreover, periodontitis has been strongly associated with systemic diseases such as diabetes, cardiovascular disease, rheumatoid arthritis, and the development of Alzheimer's disease [4,5].
Periodontal injury is strongly related to the myeloid cells, which can infiltrate the gingival tissue, destroying alveolar bone due to the production of metalloproteinase (MMP)-12 [6]. More recently, a study reported that in addition to MMP-12, calgranulins are also implicated in innate immune responses [7]. Furthermore, these proteins are members of the S100 family involved in various physiological functions, including calcium regulation, metabolism, cell multiplication, and inflammation [8].
The clinical diagnosis of periodontitis requires the use of a standardized periodontal probe to measure pocket depth (PD) and clinical attachment loss (CAL) by circumferential evaluation of erupted teeth concerning the gingival margin and cement-enamel junction (CEJ) [9]. However, these indices could be subjective because they depend on the pressure applied to insert the periodontal probe in the crevice, the anatomic details of teeth, limitations in the mouth opening, and some discomforts [10]. Therefore, there has been an increasing interest in proposing a diagnostic test using biomarkers that could detect periodontitis at its early stages, the active phases of the disease and monitoring the progression of periodontal therapy. For instance, different types of samples have been used to detect the evolution of periodontal diseases, such as periodontal samples and gingival tissue, where the transforming growth factor-β downregulation had a relationship with tissue destruction [11]. In addition, non-invasive sample like saliva has become a potential source of biomarkers compared to a crevicular fluid due to its easy and rapid collection and higher quantity [12].
Saliva contains organic compounds, exfoliated oral epithelial cells, and microorganisms and may include blood, respiratory secretions, gastric acid from reflux, and food debris [13]. In healthy individuals, saliva is produced from 0.5 to 1.5 L per day [14]. The major salivary glands: parotid, submandibular and sublingual, secrete about 90% of human saliva, and the rest by minor salivary glands located throughout the oral mucosa within a pH from 6.0 to 7.0 [15].
This systematic review aimed to present a robust critical evaluation of the evidence of salivary protein profiles for identifying oral diseases using proteomic approaches and summarize the use of these approaches to diagnose chronic periodontitis.
To address the objective, Participants, Interventions, Comparisons, Outcomes, and Study design (PICOS) criteria were set. Three PICOS questions were formulated according to the Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) statement guidelines [16].
The questions underpinning this review were as follows (Table 1): PICOS question 1. In healthy individuals over 18 years old, what is the composition and concentration of proteins in the whole saliva compared to chronic periodontitis subjects? PICOS question 2. In patients with periodontitis over 18 years old, which proteins are present in saliva after non-surgical periodontal therapy regarding composition and concentration? PICOS question 3. In adults over 18 years with healthy periodontium or periodontitis, what methods are applied for saliva collection, and what characteristics of the techniques for chemical analysis?

Protocol registration and review reporting
The protocol was registered in PROSPERO with ID no. CRD42021220377 and this review was conducted and reported according to the PRISMA statement (S1 Checklist) [16].

Eligibility criteria
Studies were selected with healthy individuals �18 years old and divided based on periodontitis. Subjects excluded from this study were those diagnosed with systemic inflammatory disorders, autoimmune diseases, diabetes mellitus, cancer, intake of antibiotics in the last three months, periodontal therapy in the previous year, and undergoing pregnancy or nursing. Studies were included if they presented a complete report on original research. Ongoing publications, conferences, poster presentations, congresses, meetings, and proceedings were excluded. There were no restrictions on the types of study design eligible for inclusion.

Search methods
The search used three electronic databases: ScienceDirect, Scopus, and Springer Link from January 1 st , 2010, to December 31 st , 2022, including an English language restriction. The search strategy was through combinations of medical subject headings (MeSH) terms and keywords. The search terms were "protein composition of saliva AND periodontal diseases OR chronic periodontitis OR periodontitis chronic OR periodontitis patients OR periodontitis subjects OR periodontal diseases subjects," "protein composition of saliva AND non-surgical periodontal treatment OR non-surgical periodontal therapy OR subgingival instrumentation OR subgingival debridement OR subgingival scaling OR root planning," "protein composition of saliva AND healthy periodontium," "protein composition of saliva AND periodontal inflammation."

Study selection
Once the studies were identified by searching the databases, record duplicates were removed and screened by title and abstract by two authors (AGSM and FMG). Eligibility criteria were applied under the PRISMA flow diagram to select the included studies. Disagreement was resolved through discussion or by a third independent reviewer (RESG).

Data management
Two authors (AGSM and FMG) extracted the data into custom Excel tables, verified by a third author (RESG). Data extracted included demographic information, methodology, and outcomes. The risk of bias was independently assessed for each study using the Newcastle-Ottawa Scale for Risk of Bias criteria (http://www.ohri.ca/programs/clinical_epidemiology/oxford.asp) for case-control and cohort studies [17]. Discrepancies between the examiners were solved in a consensus, and any disagreement through discussion, if needed, was resolved by RESG.

Study selection and characteristics
From the search of MeSH terms and keywords combinations, the total number of references obtained was 4841 citations (Fig 1). After the screening, eight studies that met the inclusion were assessed [3,[18][19][20][21][22][23][24]. Five of the eight studies focused on the proteomic features of individuals with chronic periodontitis (Table 2). Based on study designs, these eight papers were classified as three studies as case-control [3,18,21], and two were cohort studies [19,20]. The remaining three studies that met the inclusion criteria were related to the anatomic and proteomic features before and after non-surgical periodontal therapy (Table 3) [21][22][23]. Again, variations in saliva collection methods, definitions of periodontal diseases, and the measurement of protein concentration were found.

Risk of bias
Risk of bias assessment was performed according to the Newcastle-Ottawa methodological quality scale using Review Manager version 5.4 (Fig 2). Among the eight included studies, three fulfilled all domains (37.5%) with a low risk of bias, whereas five studies (62.5%) presented high bias. In contrast, the quality was good for both assessment selection and exposure/ outcomes. These biases were related to the representativeness of the cases/exposed cohort and comparability category. Based on the design or analysis, these domains evaluated the selection and comparability of cases and controls/cohorts.

Reported proteins
The reported proteins were analyzed in the 8 eight selected studies. Seven studies (87.5%) reported the identification of proteins with relevance to the periodontal process. One of the studies described a high concentration of proteins in patients with chronic periodontitis (p<0.05) compared to healthy controls [24].
Since different protein access numbers were used to identify the proteins, we unified them with ExPASy (https://www.expasy.org/). In addition, the molecular and cellular functions of the major proteins in the saliva were categorized using the Ingenuity Pathway Analysis (IPA) and reviewed in UniProt (https://www.uniprot.org/) ( Table 4).

Discussion
The present systematic review mined literature of the last twelve years to describe the main proteins in the saliva of subjects with chronic periodontitis. Some variations were detected in clinical parameters used in periodontal diagnosis. For example, only 25% of the studies  reviewed, two articles, used the same criteria to identify the presence and absence of periodontal disease [19,21].

Altered proteins in chronic periodontitis
Of the eight studies analyzed in this study, four reported that members of the S100 proteins were the most representative groups of proteins, with an increase in patients with chronic periodontitis [19,20,23,25]. However, Shin et al. reported a statistically significant difference in the quantified S100A8 in participants without periodontitis (430 pg/mL) compared to patients with periodontitis (11163 pg/mL) [20]. Main identified proteins: CSF-1, S100A8/A9, S100A12, IL-1β, MMP-8 and HGF were significantly elevated in saliva. The S100 family members belong to a calcium-binding protein group with various intracellular and extracellular functions [8]. Most of the S100 proteins are related to biological processes such as leukocyte migration, inflammatory response, chemotaxis, and aggregation of neutrophils [26]. Additionally, the S1009 protein induces phagocytosis, increasing the bactericidal activity of human neutrophils, and is critical in controlling microbial infection [27]. On the other hand, the S1008 protein has also been reported to be a potential biomarker in saliva for oral squamous cell carcinoma diagnosis [28].

Methods of saliva collection
In the present study, two studies reported increased concentrations of S1008 and S1009 using the quantitative proteomics [23,25]. In addition, another study showed that circulating monocytes from periodontitis patients had an altered expression of S100A12, suggesting its involvement in the pathogenesis of periodontitis [29].
The results in Table 2, which compared healthy periodontium with chronic periodontitis, showed differences attributable to study designs. For example, PICOS question 3 differed because of the saliva collection method and the use of protease inhibitors.
The non-surgical periodontal treatment requires a reduction of the subgingival biofilm since the procedure could be associated with tissue breakdown and the possible alteration of the composition and concentration of proteins during wound healing. A previous study reported that MMP-12 and S100/calgranulin concentration in saliva was stable and related to periodontal inflammation and adequate periodontal treatment [30]. Moreover, Porphyromonas gingivalis, an important pathogen bacterium that produces biofilm and is a keystone in periodontitis, enhanced expression of MMP 9 and interleukin-8 when grown in the culture of cell carcinoma. This event suggests the possibility of a direct relationship between orodigestive cancers and the P. gingivalis [31]. Therefore, we conclude that those biomarkers should measure the early detection and monitoring of an early stage of periodontitis.

Altered proteins concentration before and after periodontal treatment
Three studies matched the eligibility criteria to compare the evolution of patients with periodontitis that received non-surgical treatment; however, only two studies showed the identification of the salivary proteins (Table 3) [23,25]. These studies differ in the results because of the techniques used in quantifying proteins (mass spectrometry vs. immunoassays), the number of patients, and the method of collection (stimulated vs. unstimulated). Interestingly, results showed an up-regulation of colony-stimulating factor-1 (CSF-1), S100A8/A9, S100A12, interleukin (IL)-1β, matrix metalloproteinase (MMP)-8, and hepatocyte growth factor (HGF), in patients with periodontitis with a statistical significance of p<0.001 [25]. When these patients followed a non-surgical treatment, a reduction of IL-1β and MMP-8 was measured. Moreover, the ratio S100A8/S100A9 in saliva can differentiate distinct groups of periodontitis [25].
On the other hand, the second study used a quantitative biochemical parameter (bicinchoninic acid protein assay) to determine protein concentrations, but the identities of the proteins needed to be reported [24].
The differences in the methodology between the studies imply that the evidence needed to be sufficient to compare changes in protein profile before and after non-surgical periodontal therapy. Therefore, minimally invasive treatment is required to evaluate changes in the altered protein concentration before and after periodontal treatment. For example, a recent study with ultrasonic instrumentation combined with air polishing showed relevant results for the non-surgical treatment of periodontal pockets without complications [32]. Another example of successful and minimally invasive treatment is ozone water irrigation, which presents an adequate clinical evolution and a reduction of inflammatory mediators in the saliva [33]. In this context, evaluating these proteins in saliva could help assess oral hygiene status [34].

Conclusions
Our systematic review results suggested that the salivary proteins profile is an early biomarker of periodontitis and early detection of other non-communicable diseases. Proteomics tools help elucidate and understand its changes under healthy conditions and the transition to periodontal disease. Furthermore, the family of S100 protein is a potential periodontitis biomarker and could help differentiate distinct subgroups of periodontitis. Thus, further high-quality studies are recommended; some of the study design's main remarkable points should be considered, such as the saliva collection and use of protease inhibitors to stabilize the sample. Although the evidence between the studies implicated was insufficient to compare changes in salivary protein profile, well-designed studies showed statistical significance in protein concentration with potential like early biomarkers. Nevertheless, future research is needed to diagnose the disease's early stages and monitor the effectiveness of treatment.