TNFα, IL-6, miR-103a-3p, miR-423-5p, miR-23a-3p, miR-15a-5p and miR-223-3p in the crevicular fluid of periodontopathic patients correlate with each other and at different stages of the disease

Periodontitis is one of the main frequent intraoral diseases. Pathogenesis triggers are the immune responses with pro-inflammatory cytokines production and non-coding RNAs expression. The purpose of the present study was to evaluate the involvement of selected miRNAs in various stages of periodontitis and their relationship with the levels of inflammatory mediators in gingival crevicular fluid (GCF). For this study, 36 subjects (21 with periodontal disease, 15 healthy controls) were selected with an age mean of 59.1 ± 3.7 years. Clinical parameters included plaque index, gingival index, sulcus bleeding index, pocket depth, and clinical attachment level. The GCF samples were taken using capillary paper. The levels of miRNAs in GCF were estimated using a Real-Time PCR and TNFα and IL-6 levels were assessed by enzyme-linked immunosorbent assay (ELISA). The results indicated that the miRNA-103a-3p, miRNA-23a-3p, miRNA-15a-5p, and miRNA-223-3p were significantly upregulated with respect to healthy controls. Significant differences were observed for miRNA-23a-3p, miRNA-103a-3p and miRNA-423-5p levels in accord with the disease stages. Inflammatory mediators evaluated in GCF correlate well with the clinical parameters and the severity of the periodontal disease. miRNAs can represent biomarkers of disease stage and can be investigated as a possible therapeutic target, as well as levels of TNFα and IL-6 may drive the disease progression by acting as prognostic markers.

www.nature.com/scientificreports/ collected with a minimally invasive procedure, bestowing great potential for the diagnostic and prognostic value of periodontal disease. Recent investigations on inflammatory response and bone tissue homeostasis have reported a significant increase in salivary expression levels of miR-146a, miR-155 and miR-223, inhibiting osteoclastogenesis and driving the nuclear factor kappa-light-chain-enhancer of activated B cells (NF-kB) pathway activation, in patients affected by periodontitis [16][17][18] . Moreover, in chronic periodontitis, it has been confirmed a relationship between salivary miRNAs expression and periodontitis' pathological processes, with the highlighting of the contribution of miRNAs (miR-142-3p, miR-146a, miR-155, miR-203, and miR-223) involved in the regulation of bacterial infections, inflammation, and immune response 16,19,20 .
MiR-223, in particular, has been reported as key regulator of the innate immune responses mostly in association with the ability of the myeloid lineage differentiation 21 , inflammatory response modulation and infection development 22 .
Similarities were reported between miR-223 and miRNAs 15 and 23 based on inflammatory target pathways regulation 23,24 .
Among the miRNAs overexpressed during inflammation, the miR-103 is reported to target tumor necrosis factor (TNF)α, interleukin (IL)-17, IL-1β and the pathway NF-kB 25 . These miRNAs are also related to the osteoclast development process and bone metabolism, such as the miRNA 423 26 .
Furthermore, literature evidence suggests an association of miRNAs with inflammatory cytokines such as IL-1α, IL-1β, TNFα, IL-6 and IL-8, which are known to play an active role in periodontal tissue disease and whose unregulated production appears to be involved in chronic leukocyte recruitment and promoting tissue destruction 5 .

Results
Demographic and periodontal parameters. The demographic characteristics of the enrolled subjects are summarized in Table 3. The mean age of participants was not significantly different in the periodontal disease group in comparison with the healthy controls (HC). Also, gender distribution, Body Mass Index (BMI), and lifestyle habits were not statistically different within and between the groups.
As the more accurate indicators of the healthy periodontal support structure around a tooth, clinical attachment level (CAL) and maximum probing depth (MPD) have been used to define the disease stage. The clinical attachment level was measured with probe from cemento-enamel junction (CEJ) to the bottom of periodontal pocket. The maximum probing depth was measured from the gingival margin to the bottom of the gingival sulcus/ pocket. As shown in Table 4, patients were grouped into 3 disease stages: mild which correspond to grade II (8 patients), moderate as grade III (6 patients) and severe as grade IV of new periodontitis classification (7 patients).
TNFα and IL-6 quantification in periodontitis and HC groups. To define the inflammatory condition in patients with periodontitis, TNFα and IL-6 levels were quantified by ELISA assay in patients' GCF. As reported in Table 5, a significant and progressive increase was observed in relation to the disease stage. Significant higher levels of both cytokines were observed in periodontitis patients at mild, moderate and severe stages with respect to HC (p < 0.001), underlying the involvement of these inflammatory cytokines in the pathogenesis and progression of periodontitis. Table 3. Demographic characteristics of periodontitis and HC groups. Data are expressed as the mean ± standard deviation (SD) or number (n) and percentage (%). BMI, body mass index. www.nature.com/scientificreports/ miRNAs expression levels in GCF in periodontitis and HC groups. The expression levels of miRNAs 103a-3p, miR423-5p, miR23a-3p, miR15a-5p, and miR223-3p in periodontitis patients and HC were analyzed by Real-time PCR. Figure 1 shows that miR103a-3p, miR23a-3p, miR15a-5p, and miR223-3p were significantly upregulated in GCF collected in the periodontitis group in comparison with the HC, while miR423-5p showed no difference in fold change.

miRNAs expression levels in periodontitis patients with different disease stages. To better
understand the distribution of miRNAs expression in patients with periodontitis at different severity, we analyzed their expression levels based on disease stages. In detail, although an increase in expression levels was observed for miR-15a-5p and miR-223-3p in comparison with HC, there were no significant differences between disease stages (Fig. 2a, b). The expression levels of miR-23a-3pshowed a fold change of 11.32 in patients with mild periodontitis compared to HC, while, in moderate and severe periodontal disease, the expression levels were 1.9 and 1.4-fold respectively. Furthermore, in our patients, miR-23a-3p was significantly increased in mild periodontitis patients also in comparison with moderate and severe periodontitis patients (p < 0.001) (Fig. 2c). About miR-103a-3p, increasing expression levels were observed in relation to the different disease stages, with significant differences between mild and severe stages (p = 0.023), with a variation of 1.27-fold (Fig. 2d). Similarly, miR-423-5p showed an up-regulated expression in relation to disease severity, with a significant increase in severe periodontal disease patients, equal to 3.68-fold, in relation to patients with periodontitis in mild (0.39-fold) and moderate stage (0.45-fold) (Fig. 2e).
The distribution of differential miRNAs expression was confirmed by the radar plot analysis. As shown in Fig. 3, in patients with mild periodontitis, a higher distribution area was reported for miR-23a-3p, supporting the involvement in the early phases of periodontal disease. Moreover, with a lower join area, miR-23a-3p was present also in patients with moderate disease. Furthermore, a major distribution of miR-103a-3p was observed Table 4. Oral clinical parameters of the periodontitis group. Data summarized as mean ± SD. CAL, clinical attachment level; MPD, maximum probing depth. www.nature.com/scientificreports/ in patients with moderate and severe periodontitis, while in patients with severe disease, we observed a higher distribution of miR-423-5p.

Correlation analysis of miRNAs expression levels and inflammatory mediators. The analysis
of the correlation of miRNAs with inflammatory responses in patients with periodontitis pointed out that the cytokines levels were significantly correlated between them and with the disease stage, confirming the increased release in parallel with the disease worsening. Results shown in Table 4 indicate a positive correlation between GCF levels of miR-103a-3p and TNFα in patients with moderate (r = 0.418, p < 0.001) and severe (r = 0.555, p < 0.01) periodontal disease. Moreover, we observed a positive correlation between miR-423-5p and cytokines in patients at severe disease stage (TNFα: r = 0.584, p < 0.01; IL-6: r = 0.526, p < 0.05), in accordance with the up-regulated miRNAs expression and higher inflammatory cytokines levels. Moreover, a negative correlation between miR-23a-3p and TNFα (r = -0.891, p < 0.01) and IL-6 (r = − 0.503, p < 0.05) was observed in the early stage of periodontal disease (Table 6). Finally, genes were retrieved by GeneCards (https:// www. genec ards. org/) database 27 using "inflammation and bone metabolism" as keyword. As reported in Table 7 the intersection of TNFα, IL-6, miR-103a-3p, miR-423-5p, miR-23a-3p, miR-15a-5p and miR-223-3p were portrayed by the GeneCards Inferred Functionality Score and the relevance score.

Discussion
Periodontitis, one of the top ten highly prevalent diseases worldwide 28 , presents an amplified inflammatory immune response as a hallmark. The prevalent inflammatory microenvironment, affecting the periodontium and compromising tooth-supporting apparatuses (gingiva, cementum, periodontal ligament, and alveolar bone),   www.nature.com/scientificreports/ negatively influences the general state of health, leading to the persistence of chronic low-grade inflammation for lengthened periods and increasing the risk for cardiovascular, cerebrovascular and respiratory diseases, with a higher probability for metabolic diseases development [29][30][31] . Much of the tissue damage that occurs during inflammation can be attributed to TNFα and IL-6. The balance of TNFα and IL-6 mirrors the degree of the inflammatory burden contributing to periodontal inflammation 32 . TNFα, a pro-inflammatory cytokine released by macrophages, is known for its extensive role in periodontitismediated bone loss. Increased concentration of TNFα observed in periodontitis correlates closely with tissue destruction and immune response 33 . Meanwhile, IL-6, a multifunctional cytokine, has several biological activities, including B-lymphocyte differentiation, T-lymphocyte proliferation, and the stimulation of immunoglobulin (Ig) secretion by B-lymphocytes 34 . Particularly, IL-6 induces bone resorption by itself and in conjunction with other bone-resorbing agents 35 .
Our results showed that TNFα and IL-6 production levels were significantly increased in patients with periodontal disease in comparison with HC. Additionally, significant differences in TNFα and IL-6 levels were found between patients in accord with the mild, moderate, and severe periodontitis stages. These data agree with previous findings showing significantly elevated salivary concentrations of IL-6 and TNFα in periodontitis patients [36][37][38][39][40][41] and suggest that IL-6 and TNFα GCF levels may have the potential to distinguish different phases of periodontitis.
The miRNAs play a role in the production of pro-inflammatory cytokines by both a "direct" and "indirect" modulation, via the targeting activators or repressors. MiRNAs in body fluids have been reported as high stable molecules with resistance to degradation, suggesting great potential as biomarkers, also because they represent one of the most abundant classes of molecules involved in the regulation of gene expression 16 .
Despite significant advances in miRNAs involvement in periodontal disease, the regulatory mechanisms are still to be investigated, also due to heterogeneity of the miRNAs' expression.
Taking in mind that the inflammation in periodontal disease often results in alveolar bone loss and disruption of connective tissue homeostasis 42-44 , we have evaluated in GCF the expression of miR-15a-5p, miR-23a-3p, miR-223-3p, miR-103a-3p, and miR-423-5p, involved in the regulation of innate immune response, cell metabolism, and inflammation, in patients with periodontal disease at three different disease stages (mild, moderate and severe).
The expression of selected miRNAs was up-regulated in patients with respect to HC, in agreement with previous observations 12, 45 .
In accord with its role as a positive regulator of the inflammatory microenvironment and negative regulator of osteogenic differentiation in various cell types 46,47 , we have observed a high expression for miR-23a-3p in mild periodontal disease, probably acting as a starter for the bone metabolism alteration. Furthermore, miR-103a-3p, related to inflammation and bone metabolism targeting IL-1, prostaglandin E2 (PGE2), and TNFα, was upregulated in periodontal inflamed tissue disease 48,49 . In our patients, miR-103a-3p was increased mostly in the moderate disease level. About miR-423-5p, participating in osteoclast metabolisms 26 , a down-regulation in mild Table 6. Analysis of the correlation of miRNAs expression levels with TNFα and IL-6. The statistical significance shown are *p < 0.05; **p < 0.01; ***p < 0.001. Data were analyzed using Stata version 15. www.nature.com/scientificreports/ and moderate disease stages was observed with respect to HC, although an over-expression in severe disease stage patients was detected. These data are in accordance with miR-423-5p involvement in osteoclastogenesis, which represents a critical step in periodontal disease physiopathology 50 , characterized by an active inflammatory state, with high levels of TNFα, needed for osteoclastic cellular maturation. Furthermore, it has been observed that miR-423-5p has a better positive relation with TNFα and IL-6, with the ability in NFkB pathway up-regulation. Referring to miR-223-3p, Bauernfeind et al. suggested its antimicrobial effect against periodontal pathogens 51 and regulatory role of inflammation through neutrophil recruitment in chronic inflammatory sites. miR-223-3p is also involved in the differentiation of several immune cells, particularly macrophages, by influencing their activation patterns 52 . Therefore, has been reported that miR-223-3p is up-regulated in periodontal disease, to regulate tissue homeostasis, to control osteoblast differentiation, and to avoid alveolar bone loss, which are emblems of periodontitis 53 . In our study groups, although a slight increase in miR-223-3p expression levels has been observed in periodontal disease patients in comparison with HC, no significant differences between the three disease stages were observed, demonstrating a non-decisive role in the regulation of disease progression. miR-15a-5p is reported to be closely related to several diseases, mediating the inflammatory process, but its role is not plenty investigated in periodontal disease. Luan et al. found the up-regulation of miR-15a-5p in both humans and mice affected by periodontitis 7 . In our patients, although not significantly, miR-15a-5p expression showed a reduced trend during disease progression in parallel with the increase of TNFα and IL-6 levels in GCF.
Our results are encouraging, suggesting a different involvement of selected miRNAs in relation to the disease stages. The strong correlation found between miRNAs expression and inflammatory cytokine levels in relation to disease severity supports the deepening of their role as disease markers. Also, it is very interesting that this can be explored in GCF, supporting its high diagnostic, prognostic and therapeutic potential. However, their clinical application still needs to be defined and investigated. Gingival crevicular fluid samples collection. The GCF sample was taken from a single-rooted tooth, due to its easy access and to avoid errors associated with GCF sample collection, both in patients with periodontitis and HC. Before placing the absorbing paper strip within the sulcus, the supragingival plaque was removed and relative isolation was performed using cotton rolls and aspiration to prevent saliva contamination. Subsequently, the filter paper strip Periopaper (Oraflow, New York, NY, USA) was placed in the gingival sulcus until resistance was felt and left in this position for 30 s following the routine procedure. The filter paper strip was introduced in a sterile tube and stored at − 80 °C, until the subsequent analysis. Meanwhile, the demographic and habit data of the patients enrolled in both groups have been collected on specific Case Report Forms.

RNA extraction and quantification from gingival crevicular fluid. Periopapers were incubated in
phosphate-saline buffer (PBS) solution pH 7 for 30 min at room temperature. To recover all the solution contained in the strip, each sample was centrifuged at 500 rpm for 10 min. Cell-free total RNA (including miRNAs) was isolated from 200 µL of PBS solution containing the biological material from the Periopaper using a miRNeasy Serum/Plasma kit (Qiagen, Hilden, Germany). RNA was eluted with 25 µL of RNAse-free water. Total RNA concentrations were quantified using NanoDrop ND 2000 UV-spectrophotometer (Thermo Scientific, Delaware, USA). ELISA assay. The human TNFα and IL-6 concentrations were measured by 'sandwich' ELISA (Millipore, Merck KGaA, Germany) on the GCF sample, according to the manufacturer's instructions. The absorbance of each well was detected with a spectrophotometer, allowing for the generation of a standard curve and subsequent determination of protein concentration. The range of the standard curves was 1.37-12,000 pg/mL for IL-6 and 24.58-6000 pg/mL for TNFα. Statistical analysis. The mean was taken as the measurement of the main tendency, and the SD was taken as the dispersion measurement for the statistical analysis of the results. Qualitative variables were reported as frequency and percentage. The normal distribution of data was tested by the Kolmogorov-Smirnov test. Student's t-test was performed to compare baseline variables between study groups. Radar plot analysis for the distribution of differential expression patterns of miRNAs was evaluated. Pearson's correlation analysis was calculated to assess the relationship between cytokine levels and miRNAs gene expression in patients. Statistically significant was set at p-value < 0.05. Data analysis was performed with Stata version 15 (StataCorp LLC, College Station, TX, USA) and GraphPad Prism 6 Software, version 6.01, 2012.
Ethical approval and consent to participate. This study conformed to the provisions of the Declaration of Helsinki. The protocol was approved by the Ethics Review Committee of G.d' Annunzio University, Chieti-Pescara (Protocol code 254/2019). Informed consent was obtained from all participants.

Strengths and limitations of the study
Our data underline the importance of crevicular fluid as a possible diagnostic tool and means to monitor the status of patients affected by periodontal disease. On the other hand, the relatively small number of studied samples limits the possibility of generalizing the obtained considerations and of using these data to define the diagnostic sensitivity and specificity of the studied miRNAs. The evaluation of miRNAs could be accompanied by the analysis of the proteome, transcriptome and microbiome, in order to obtain a complete picture of their role in periodontal disease.

Data availability
The data supporting the results of this article are included within the article and can be required to the corresponding authors.