Local levels of inflammatory mediators in uncontrolled type 2 diabetic subjects with chronic periodontitis Abstract Aim: The aim

of this study was to evaluate the levels of a wide panel of cyto/ chemokines in the gingival crevicular fluid (GCF) of uncontrolled type 2 diabeti c subjects as compared with non-diabetic subjects with periodontitis. Methods: Twenty-six uncontrolled type 2 diabetic subjects (glycated haemoglobin levels >7.5%) and 20 non-diabetic subjects with chronic periodontitis were enrolled in this study. The levels of 14 cyto/chemokines were measured in the GCF of healthy and diseased sites of the diabetic and non-diabetic subjects usin g multiplex bead immunoassays. Results: The concentrations of eotaxin, macrophage inflammatory protein-1a, granulocyte-macrophage colony-stimulating factor, interleukin (IL)-6, tumour necrosis factor-a and IL-12 were higher in healthy and diseased sites of diabeti c than non-diabetic subjects, after adjustment for multiple comparisons (p < 0.0035). Conclusion: Uncontrolled type 2 diabetes mellitus modulated the local levels of several cyto/chemokines at both healthy and diseased periodontal sites in favour of a proinflammatory profile, which may partially explain the greater susceptibi lity of diabetic subjects to periodontal breakdown Chronic periodontitis is an infectiousinflammatory disease with continuous release of several immunoinflammatory factors that are essential for pathogen clearance, but may damage periodontal tissues. The host responses against pathogens in periodontitis involve the activation of inflammation and the innate and adaptive immunities, particularly due to the upregulation of proinflammatory proteins and the downregulation of anti-inflammatory mediators (Madianos et al. 2005, Tatakis & Kumar 2005). Type 2 diabetes mellitus (DM) is a chronic polygenic disorder characterized by defects in insulin action and/or deficiencies in pancreatic insulin secretion, which eventually lead to loss of pancreatic insulinsecreting cells (Israili 2011). Several studies have proposed a bidirectional relationship between DM and periodontitis (Lakschevitz et al. 2011, Chapple & Genco 2013). DM plays an important role in the pathogenesis of periodontitis, as there is an increased prevalence, severity and progression of periodontitis in diabetic patients, especially those with uncontrolled glycemia (Lakschevitz et al. 2011). In addition, periodontitis may have a negative impact on

DM control while the periodontal therapy may lead to improvements in the glycemia (Lakschevitz et al. 2011, Chapple & Genco 2013). Although a well-known and strong relationship has been established between DM and periodontitis, to date, there are insufficient data regarding the actual cellular and molecular mechanisms related to diabetesassociated periodontitis. Periodontitis and DM appear to share common pathogenic processes, including increased immunoinflammatory responses showing similar biological mediators (Thunell et al. 2010). Some studies have demonstrated that DM exacerbates the levels of some immunoinflammatory mediators at diseased periodontal sites (Bulut et al. 2001, Engebretson et al. 2006, Duarte et al. 2007a,b, Javed et al. 2012). A recent comprehensive Literature Review (Javed et al. 2012) disclosed that the few available studies on this topic have investigated only a limited number of cyto/chemokines in the gingival crevicular fluid (GCF) of diabetic subjects with periodontitis, mainly using enzyme-linked immunosorbent assays (ELISAs). As GCF contains connective tissue degradation products, host and bacterial enzymes, cyto/chemokines and other inflammatory mediators, efforts have been expended on the identification of GCF components able to predict a site or a patient at risk for disease progression. However, the volume of GCF available from each site is very small and insufficient for the quantification of a large number of mediators using ELISA. This limitation hampers a deep comprehension of the immunoinflammatory process in a given GFC sample. Currently, commercially available multiplex technologies, able to detect large numbers of biomarkers in a reduced volume of sample (e.g. GCF), provide the possibility to better understand the multifaceted immunoinflammatory process associated with the periodontal breakdown (Zhou et al. 2010). Therefore, this study evaluated the levels of a wide panel of cyto/chemokines in the GCF of uncontrolled type 2 diabetic subjects with periodontitis, using

multiplex bead immunoassays. Such cyto/chemokines are key markers of inflammation, innate and/or adaptive immunities, and are supposed to be critical for the pathogenesis of both periodontitis and DM. Some of these biomarkers [eotaxin, macrophage inflammatory protein [MIP]-1a, granulocyte-macrophage colony-stimulating factor (GM-CSF), granulocyte colony-stimulating factor (G-CSF) and certain interleukins (IL), namely IL-12, IL-2 and IL-7] have not been previously evaluated in the GCF or periodontal tissues of diabetic subjects. Material and Methods Subject population Twenty-six diabetic subjects and 20 non-diabetic subjects, all with chronic periodontitis (age range: 42 60 years), were selected from the population referred to the Periodontal Clinic of Guarulhos University (Guarulhos, SP, Brazil). Detailed medical and dental records were obtained. Subjects who fulfilled the following inclusion/exclusion criteria were invited to participate in the study. All eligible subjects were informed of the nature, potential risks and benefits of their participation in the study and signed an informed consent. The Guarulhos University's Ethics Committee in Clinical Research previously approved this study protocol. Inclusion and exclusion criteria All subjects should be >35 years old and present with at least 15 teeth (excluding third molars). They were all diagnosed with generalized chronic periodontitis, based on the criteria proposed by the American Academy of Periodontology (Armitage 1999). The subjects should present with more than 30% of the sites with concomitant probing depth (PD) and clinical attachment level (CAL) 4 mm and, at least six noncontiguous sites with PD and CAL >5 mm, distributed in different quadrants. Diabetic subjects should have type 2 DM, diagnosed by a physician, for at least the past 5 years. The glycated haemoglobin [HbA1c (%), high-performance liquid chromatography method] and the fasting plasma glucose [FPG (mg/dl), glucose oxidase method] levels were

assessed by the Guarulhos University Clinical Analysis Laboratory. To be included in this study, diabetic subjects demonstrated HbA1c > 7.5% and FPG > 99 mg/ dl at the beginning of the study. In addition, subjects were asked to bring at least two previous exams reporting HbA1c > 7.0% to attest history of uncontrolled glycemia within the last year. Diabetic subjects could be on a diet regime, oral hypoglycemic agents (metformin or glibenclamide) and/or insulin supplementation. HbA1c and FPG levels were also assessed for the nondiabetic subjects to confirm their non-diabetic status (HbA1c < 6.5% and FPG = 7099 mg/dl). Exclusion criteria were pregnancy, lactation, current cigarette smoking and cigarette smoking within the past 5 years, periodontal and antibiotic therapies in the previous 6 months, continuous use of mouthrinses containing antimicrobials in the preceding 2 months, any systemic condition, other than DM, that could affect the progression of periodontal disease (e.g. immunological disorders, osteoporosis). The continuing use of anti-inflammatory and immunosuppressive medications, endodontic lesions, orthodontic appliances and obesity (i.e. body mass index 30 and <40 kg/m2 and waist-hip ratio 0.85 and 0.90, for women and men, respectively) were also exclusion criteria. Clinical monitoring One examiner (T.S.M.), calibrated as described previously (Luckheeram et al. 2012), performed all clinical examinations. The intra-examiner variability was 0.19 mm for PD and 0.21 mm for CAL. The clinical parameters registered dichotomously [e.g. bleeding on probing (BoP) and suppuration (SUP)] were calculated by the Kappa-Light test, and the intra-examiner agreement was >0.85. The following parameters were assessed at six sites of all teeth, excluding third molars, using a manual periodontal probe (UNC15; Hu-Friedy, Chicago, IL, USA): visible plaque accumulation (PI) (Ainamo & Bay 1975), BoP, SUP, marginal bleeding (MB) (Ainamo & respectively. In healthy sites, eotaxin, MIP-1a, GM-CSF, IL-6, TNF-a and

IL-12 concentrations were higher in the diabetic group than in the nondiabetic group after adjustment for multiple comparisons (p < 0.0035). In addition, there was a tendency towards higher concentrations of IL-8 (p = 0.04), G-CSF (p = 0.03) and IL-7 (p = 0.04) and lower concentrations of IL-10 (p = 0.047) and IL-2 (p = 0.04) in the healthy sites of diabetic subjects, compared with non-diabetic subjects (Table 2). In diseased sites, eotaxin, MIP1a, GM-CSF, IL-6, TNF-a and IL12 concentrations were higher in the diabetic group than in the nondiabetic group (p < 0.05) after adjustment for multiple comparisons (p < 0.0035). Furthermore, there was a trend towards a higher concentration of IL-8 (p = 0.048) and lower concentrations of IL-10 (p = 0.03) and IL-2 (p = 0.03) in the diseased sites of diabetic, compared with nondiabetic subjects (Table 3). The results of the MLR analyses are shown in Table 4. In the four models created for healthy sites, non-diabetic subjects had lower levels of eotaxin (p = 0.001), IL-12 (p = 0.002) and IL-7 (p = 0.006) than the dependent group (i.e. diabetic subjects; OR <1). With regard to the MLR analysis for diseased sites, non-diabetic subjects had lower concentrations of eotaxin (p = 0.003), IL-6 (p = 0.017), TNF-a (p = 0.015) and IL-7 (p = 0.006) than the diabetic subjects (OR < 1). Conversely, non-diabetic subjects had higher concentrations of IL-10 and IL-2 than diabetic subjects in healthy and diseased sites (OR > 1). Discussion To the best of our knowledge, this is the first study to concomitantly evaluate the concentrations of a broad panel of cyto/chemokines in the GCF of healthy and periodontitis sites in uncontrolled diabetic and non-diabetic subjects, matched for the severity of sampled sites. The results indicate that diseased sites in periodontitis subjects with uncontrolled DM presented significantly increased levels of eotaxin, MIP-1a, GM-CSF, IL-6, TNF-a and IL-12, after adjustment for multiple comparisons, compared with the nondiabetic subjects. The concentrations

of these same proinflammatory biomarkers were also significantly elevated in the healthy sites of the diabetic patients, when compared with the non-diabetic patients. In addition, there was a tendency towards reduced concentrations of IL-10 (p < 0.05), an important antiinflammatory cytokine, in the healthy and diseased sites of the diabetic subjects. Together, these findings may suggest the role of hyperglycemia in intensifying the responses by increasing the levels of relevant proinflammatory markers in the presence as well as absence of disease. It is hypothesized that the imbalance in the levels of pro- and anti-inflammatory mediators at healthy sites may be a risk for future periodontal breakdown, as soon as bacterial challenge begins (Garlet 2010). However, this study is limited to one specific point in time and, the actual clinical consequence of these immunological findings should be assessed over a longer follow-up of these subjects by a prospective cohort study. Another limitation of this study is that no formal sample size calculation could be performed, as there was no precise evidence regarding the clinical consequences of the differences in the GCF levels of the studied inflammatory mediators between diabetic and nondiabetic subjects. Consequently, it is possible that a larger sample size would be necessary to observe additional significant differences, if it exists, in some biomarkers between groups. Therefore, at this stage, this study is only able to propose some biomarkers to account for the association between hyperglycemia and periodontal destruction. Chemokines are proinflammatory cytokines that have chemotactic activity for specific leucocyte subsets at the infectious and inflammatory foci. IL-8 (CXCL8) is a CXCL class chemokine, originally described for its ability to attract neutrophils (Yoshimura et al. 1987, Romagnani et al. 2004). MCP-1 (CCL2) is a potent chemoattractant for monocytes/ macrophages, MIP-1a (CCL3) is a strong chemoattractant for lymphocytes, monocytes/macrophages and eosinophils, whereas eotaxin (CCL11) is mainly involved in eosinophil chemotaxis (Maurer & von

Stebut 2004, White et al. 2013). In this study, the concentrations of eotaxin and MIP-1a were higher in healthy and diseased sites of the diabetic subjects. To the authors' knowledge, no study has evaluated the levels of eotaxin and/or MIP-1a in diabetic subjects with periodontitis. In non-diabetic subjects, previous reports have suggested a role for eotaxin and MIP-1a in the pathogenesis of periodontal diseases (Kabashima et al. 2002, Hanes & Krishna 2010, Tymkiw et al. 2011). Although not statistically significant considering the adjustment for multiple comparisons, there was a tendency (p < 0.05) towards increased levels of IL-8 in the periodontal sites of subjects with DM, when compared with the non-diabetic subjects. Previous investigations have also demonstrated higher levels of IL-8 in the GCF or periodontal tissues of periodontitis subjects with type 2 DM, when compared with non-diabetic subjects (Engebretson et al. 2006, Venza et al. 2010, Amir et al. 2011). Furthermore, in accordance with our results, Venza et al. (2010) observed that the gingival expression of MCP1 did not differ between uncontrolled diabetic and non-diabetic subjects. Taken together, these results suggest that hyperglycemia is able to increase the production of specific members of the chemokine family that are probably involved in intensifying the periodontal breakdown in diabetic subjects. Hyperglycemia also upregulated the levels of GM-CSF, IL-6 and TNF-a in healthy and diseased periodontal sites. In addition, increased concentrations of IL-6 and TNF-a in diseased sites were associated with DM, according to MRL analysis and, there was a tendency towards higher levels of G-CSF (p = 0.03) in the healthy sites of the diabetic subjects. GM-CSF and G-CSF were firstly defined by their abilities to generate granulocytic macrophage colonies and granulocytic colonies, respectively, from bone marrow precursor cells. Currently, both mediators are proinflammatory cytokines that play key roles in the colonystimulating factor (CSF) network, inflammation and immunity of several

diseases (Hamilton 2008). Some studies have suggested a local role of GM-CSF in non-diabetic subjects with aggressive and chronic periodontitis (Shaddox et al. 2011, Tymkiw et al. 2011, de Lima Oliveira et al. 2012), whereas a previous investigation did not observe differences in the GCF levels of this cytokine between healthy and periodontitis sites (Hanes & Krishna 2010). As persistent excessive levels of CSFs contribute to chronic inflammation (Hamilton 2008), the present findings suggest that both GM-CSF and G-CSF may be critical mediators in the diabetes-associated periodontitis. Interestingly, there is a link between the biological functions of CSFs and other proinflammatory cytokines such as IL-1b, TNF-a and IL-6 (Hamilton 2008). IL-1b, TNF-a and IL-6 are classical cytokines involved in inflammation that are able to synergistically stimulate the degradation of the connective tissue and bone resorption directly and indirectly (Cavaillon 2001). In agreement with the present results, several previous studies have also demonstrated that the local levels of IL-6 and TNF-a were higher in type 2 diabetic subjects when compared with the non-diabetic subjects with periodontitis (Kurtis et al.1999, Duarte et al. 2007a, Cole et al. 2008, Venza et al. 2010, Amir et al. 2011, Kardesler et al. 2011). In this study, IL-1b concentrations in healthy and diseased sites did not differ between diabetic and non-diabetic subjects. Indeed, the actual role of IL-1b in the periodontal tissues of diabetic subjects is still controversial, probably due to methodological differences among studies (i.e. periodontal disease severity, DM duration, control of glycemia). Some studies have demonstrated that diabetic subjects had higher GCF levels of IL-1b, when compared with non-diabetic controls that were matched regarding periodontal disease severity (Salvi et al. 1997, Bulut et al. 2001). However, in agreement with this study, other investigations have observed that the periodontal concentrations of IL-1b did not differ between periodontitis subjects with and without DM (Cutler et al. 1999,

Navarro-Sanchez et al. 2007, Cole et al. 2008, Correa et al. 2008). Hence, it is possible that the periodontal infection mainly controls the production of IL-1b in periodontal tissues, regardless of the diabetic status. CD4+T cells modulate the innate and adaptive immunities through their distinct phenotypes and cytokine profiles. Naive CD4+T cells may differentiate into the classical Th1 and Th2 subsets. IL-12 and IFN-c are critical Th1-cytokines with proinflammatory properties (Hamilton 2008). Pathogen-induced IL-12 production leads to Th1 polarization (O'Garra and Murphy 2009), whereas IFN-c is essential for macrophages activation (Mendonca et al. 2012). In this study, IL-12 concentrations were higher in healthy and diseased sites of diabetic subjects. Elevated levels of IL-12 have been observed in periodontitis (Orozco et al. 2006, Yucel et al. 2008). In addition, peripheral mononuclear cells from type 2 diabetic subjects stimulated by LPS produced more IL-12 than those of systemically healthy controls (Wu et al. 2010). In agreement with these results, a previous investigation from our research group did not find differences in the GCF levels of IFN-c between uncontrolled diabetic and non-diabetic subjects with periodontitis (Ribeiro et al. 2011). Similarly, Shin et al. (2010) did not observe differences in the periodontal levels of IFN-c between systemically healthy and diabetic subjects. All these findings reinforce our previous observation that, IFN-c may not be the major cytokine involved in uncontrolled DM-associated periodontitis (Ribeiro et al. 2011). IL-10 is a Th2-cytokine with anti-inflammatory functions that antagonizes IFN-c and CSFs. After pathogen clearance, IL-10 helps achieve homeostasis through the inhibition of Th1 and innate immune cells (Couper et al. 2008, Duell et al. 2012). Noteworthy, both healthy and diseased sites of diabetic subjects presented a tendency towards lower concentrations of IL-10 (p < 0.05). Furthermore, the MRL analysis demonstrated that non-diabetic had significantly higher levels of IL-10

than those of the diabetic subjects at both healthy and diseased sites. A similar tendency was previously observed for the gene expression of IL-10 in the periodontitis tissues of poorly controlled type 2 diabetic subjects (Duarte et al. 2012) as well as in type 1 diabetic subjects (Duarte et al. 2007b). Considering the antiinflammatory actions of IL-10, its low levels in the periodontal tissues of the diabetic subjects may play a critical role in the pathogenesis of the uncontrolled DM associatedperiodontal diseases, by reducing their protective immunity against pathogens. Studies have shown elevated levels of the T cell growth factors, IL-2 and IL-7, in periodontitis (Lee et al. 1995, Gorska et al. 2003, Tymkiw et al. 2011), suggesting their involvement in periodontal diseases. In this study, there was a tendency towards lower levels of IL-2 in the healthy (p = 0.04) and diseased sites (p = 0.03) of diabetic subjects, which was reinforced by the MLR results. IL-2 stimulates the generation of Th1 and Th2 populations and the development of long-lived memory cells. Moreover, IL-2 is also essential in the control of inflammation by promoting the development, functionality and maintenance of T regulatory (Treg) cells and inhibition of Th17 cell differentiation (O'Garra and Murphy 2009, Katzman et al. 2011). Interestingly, our previous studies demonstrated that IL-17, a proinflammatory cytokine produced by Th17 cells, is a key mediator in the periodontitis in uncontrolled type 2 diabetic subjects (Santos et al. 2010, Ribeiro et al. 2011). Therefore, we speculated that the deficiency of IL-2, driven by hyperglycemia, could decrease the number of Treg cells and increase the number of Th17 cells and, consequently, the levels of IL-17. The MLR results showed elevated GCF levels of IL-7, another relevant T cell growth and survival factor (Katzman et al. 2011), in the diabetic subjects. Although IL-7 is important to protect the host against pathogens by stimulating T-cell immunity, high levels of IL-7 can be damaging due to the dysregulated production of inflammatory cytokines

from T cells and monocytes, as observed in other diseases like rheumatoid arthritis (Kim et al. 2008). In conclusion, uncontrolled type 2 DM modulated the local levels of several cyto/chemokines at healthy and diseased sites in favour of a proinflammatory profile. These findings may partially explain the greater susceptibility of diabetic subjects to periodontal breakdown.