Association of IL16 polymorphisms with periodontitis in Brazilians: A case- control study.

Periodontal disease is a complex and multifactorial disease characterized by the aggression of the periodontal ligament and alveolar bone [ 1 ]. This disease has a high prevalence that varies widely between countries and has an estimated average of 10.8% worldwide; it affects 743 million people and is considered the sixth most prevalent health condition in the world [ 2 ]. According to the International Workshop for a Classification of Periodontal Diseases and Conditions of 1999 [ 3 ], updated in 2017 [ 1 , 4 ], periodontitis (PD) involves a chronic inflammatory process resulting from the individual's immune response against bacterial components present in tooth plaque [ 5 ]. This process may lead to functional and morphological compromises of the periodontium, and in more advanced cases, significant alveolar bone resorption and tooth loss [ 6 ]. Some diseases, genetic and environmental factors are involved in the development of this disease [ 7 – 9 ]. Cytokines play an important role in the pathogenesis of PD. Some of them, such as Interleukins (IL)-1β (IL-1β), Interleukin-6 (IL-6), and Tumor Necrosis Factor Alpha (TNF- ɑ) are activators of bone resorption through osteoclasts. These cytokines are also involved in neutrophil chemotaxis, the inflammatory process and the production of other pro-inflammatory cytokines [ 10 – 12 ]. They can be produced by monocytes after stimulation by other cytokines such as Interleucin-16 (IL-16) [ 13 ]. Little is known about the role of IL-16 in PD. However, IL-16 levels were lower in gingival crevicular fluid (GCF) of the teeth sites with PD compared to healthy tooth sites and this cytokine was correlated with the severity of disease in alcohol-consuming individuals and smokers [ 14 ]. IL-16, a chemokine initially described as a chemotactic factor for leukocytes, is produced by some cells, such as T lymphocytes, fibroblasts and dendritic cells [ 15 ]. This cytokine is associated with pleiotropic effects, exhibiting patterns of inflammatory or anti-inflammatory responses in diseases [ 16 ]. Its chemotactic role, especially for CD4+ cells, leads to cell recruitment and subsequently immune response. The gene for this cytokine, called IL16 , is located on chromosome 15 (15q26.3) [ 17 ]. Polymorphisms in IL16 have been associated with chronic inflammatory diseases and cancer in studies conducted in different populations, such as Chinese and Iranians [ 18 – 24 ]. Among the most studied polymorphisms are the single nucleotide polymorphisms (SNPs) consisting of the exchange of only one nucleotide in the gene sequence. These polymorphisms may interfere with the amount of the transcribed product (as described for SNP rs4778889 of the IL16 , present in the promoter region) or in the biological effect of the produced protein, when located in the exon region of the gene (such as IL16 rs11556218 and rs4072111 ) [ 18 ]. IL16 polymorphisms were previously studied in PD [ 23 , 24 ]. Other SNPs in cytokine genes such as IL-1β, IL-2, IL-17 and IL-18 were associated with PD in our population from southern Brazil [ 25 – 27 ]. In an attempt to help elucidate the role of the genetic effect and cytokine production that affect the pathogenesis of PD, this study evaluated a possible association of rs11556218 T>G , rs4072111 C>T and rs4778889 T>C of IL16 and the serum levels of IL-16 with PD in a Brazilian population. In addition, the relationship between these polymorphisms and PD in relation to gender, smoking habits and age of these individuals was evaluated.

The characteristics of the studied population are described in Table 1 . No significant differences were observed in the groups according to age and gender, suggesting adequate pairing between cases and controls. The covariate smoking habit showed significant difference when the PD and control groups were compared. All results from the comparisons between the groups were adjusted to remove the effects of covariates acting as confounding biases. 1 Age in years, given by mean (±Standard deviation). Student’s t-test. 2 Chi-square test. The distribution of genotype frequencies for all studied polymorphisms in the control groups is in accordance with the expected for Hardy-Weinberg equilibrium. The distribution of alleles and genotype frequencies for the three studied polymorphisms is shown in Table 2 . No differences were observed in the allele and genotype frequencies between PD and controls for all polymorphisms evaluated. For these SNPs, no statistically significant association was observed between allele variants and PD under analysis in Haploview. 1 No significant difference was observed in allele and genotype frequency distribution between periodontitis and controls in the chi-square test ( P >0.05). The data were analyzed in a stratified form, with OR adjustment, in order to remove the possible effects of covariates in the evaluation of associations. Therefore, the values were adjusted for age and gender. The groups of smokers and nonsmokers were analyzed separately and only the significant results are presented in Table 3 (no significant results are presented in S2 Table ). 1 OR: Odds Ratio ; CI = Confidence interval 95%. 2 Pc: Bonferroni Correction of P-value . 3 Ref: Reference (OR = 1.00). 4 AIC: Akaike Information Criteria. In nonsmokers, the variant of the IL16 rs11556218 was statistically associated with a protection for PD in codominant, dominant, recessive and log-additive inheritance models ( Table 3 ). The codominant inheritance model was selected because it presented the lowest AIC. Individuals with the G/G genotype were less susceptible to developing PD compared to individuals with the T/T genotype (OR = 0.10; P = 0.006). The statistical significance for this model was maintained after BC ( Pc = 0.019). In smokers, this same polymorphism ( rs11556218 ) was statistically associated with the risk for PD in the dominant and log-additive inheritance models ( Table 3 ). The log-additive was considered the best model (OR = 2.08; P = 0.028) according to the lowest AIC value; however, significance was lost after BC ( Pc = 0.084). SNPs rs4778889 and rs4072111 were not significantly associated with PD in this study when smokers and nonsmokers were analyzed separately. Differences were also not observed between case and controls when the age covariate was considered for the analysis of all polymorphisms evaluated. Six possible haplotypes were obtained for the SNPs rs11556218(T>G)_rs4072111(C>T)_rs4778889(T>C) and the haplotype frequencies observed for the control group were: TCT 59.2%, GCT 11.7%, GCC 11.7%, TTT 6.1%, TCC 7.0% and GTT 3.0%. In PD patients, the haplotype frequency of TTT (11.3%) remained significant after testing with 10,000 permutations with haplotypes and single markers ( Pc = 0.03). In PD patients, the haplotype frequency of TTT (11.3%) was higher than in controls (OR = 2.45; 95% CI = 1.18–5.08; P = 0.01) and remained significant after testing with 10,000 permutations with haplotypes and single markers ( Pc = 0.03). The distribution of the other haplotype frequencies was similar between PD patients and controls. Thirty-nine nonsmokers were selected to determine IL-16 serum levels (9 controls and 30 PD patients, 20 men and 19 women). PD patients were classified according to the extent of the disease (localized n = 15 and generalized n = 15) and the degree of severity (mild n = 6, moderate n = 10, and severe n = 14). Lower serum levels for IL-16 were found in individuals with PD (median: 70.7 pg/ml) when compared to controls (89.4 pg/ml, P = 0.027) ( Fig 1A ). No significant differences were observed in serum levels for IL-16 between the forms of the disease ( Fig 1B ) or between females and males ( Fig 1C ). There were no statistically significant differences in the IL-16 serum levels when compared to the genotypes of rs11556218 T>G ( Fig 1D ), rs4778889 T>C ( Fig 1E ) and rs4072111 C>T ( Fig 1F ). (A) Comparison of IL-16 serum level between PD and controls; (B) between periodontitis groups; (C) between men and women with periodontitis; (D) among patients with and without G allele variant of rs11556218 for IL16 ; (E) among patients with and without the C allele variant of rs4778889 for IL16 ; and (F) among patients with and without the variant allele T of rs4072111 for IL16 .

The IL16 rs11556218 polymorphism and the IL-16 serum level were associated with protection for periodontitis in a nonsmoking Brazilians sample, although they seem to be independent events. No association was observed between polymorphisms in IL16 rs4778889 and rs4072111 and periodontitis. Environmental factors that alter the immunopathogenesis of periodontitis, such as smoking, influenced the association of IL16 with PD.

For this case-control study, 424 individuals were selected in the dental clinics of the State University of Maringá (DOD-UEM) and the Uningá University Center, between January 2012 and September 2018, both located in Maringá, northwest region of the state of Paraná, Brazil. This study was approved by the Ethics and Research Committee on Human Beings of the State University of Maringá (UEM-No. 719/2011, 2011 and 1,866,509, 2016). The subjects who met all the eligibility criteria for this study (case and controls) and agreed to participate were informed about its nature and signed the free and informed consent form. All subjects were submitted to a periodontal clinical examination by a specialized periodontist who evaluated clinical parameters such as probing pocket depth (PPD), with an unc-15 periodontal probe (Hu-Friedy manufacturing, Chicago, IL, USA), clinical attachment level (CAL) and bleeding on probing (BOP) in four sites of each tooth (mesial, vestibular, distal and lingual). Clinical parameters were evaluated in order to classify individuals into case/control groups and the selection criteria was done according to the International Workshop for a Classification of Periodontal Diseases and Conditions of 1999 [ 3 ]. After the periodontal examination, participants were assigned to the case groups (n = 215) if they had at least 5 sites in non-adjacent teeth with PPD ≥ 5 mm, CAL ≥ 3 mm, and more than 25% BOP. The control group (n = 209) was formed by individuals without altered clinical insertion level, with a probing depth of ≤ 4 mm and with less than 25% bleeding on probing and plaque index of < 20%. Although the criteria used for the classification of patients with periodontal disease was based on the 1999 classification [ 3 ], it contemplates the criteria for the definition of PD established in 2017 by the Workshop on the Classification of Periodontal and Peri-Implant Diseases and Conditions [ 1 ]. In this last classification, periodontitis was defined as the presence of at least 2 non-adjacent teeth with detectable interdental clinical attachment level (CAL), or detectable buccal or oral CAL ≥3 mm with pocketing ≥3 mm in ≥2 teeth, with observable CAL that cannot be attributed to causes not related to PD. In addition to that, in the recent classification, it is established that significant descriptions of PD should include the proportion of sites that bleed on the probing, the number of with PPD above certain limits (commonly ≥4 mm and ≥6 mm) and with CAL ≥3 mm and ≥5 mm: in our study we evaluated > 25% of BOP and at least 5 sites in non-adjacent teeth with PPD ≥ 5 mm and CAL ≥ 3 mm. Furthermore, for the classification of staging and grading [ 4 ], the patients in the present study may be included in stages II and III (based on severity, complexity and with extension and description of the distribution as localized or generalized) and grade B (moderate rate of progression). Cases and controls were selected from adults over 30 years of age, with at least 20 teeth in the mouth, nonsmokers or smokers (including those who have smoked in the last 10 years). All individuals were considered ethnically mixed, due to the great miscegenation found in the Brazilian population. This classification was in accordance with the study previously conducted for the population of Paraná, southern Brazil, by Probst et al. [ 28 ] and confirmed for our region [ 29 ], whose population consisted mainly of Caucasians with European genetic origins (80.6%), with a lower presence of Africans (12.5%) and Amerindians (7.0%). Nevertheless, the self-declarations of ethnicity were collected according to the norms from the Brazilian Institute of Geography and Statistics (IBGE) ( S1 Table ), and no differences were found between the cases and controls indicating that there is no ethnic disproportion between them. To avoid possible confounding variables, pregnant women, Asians, individuals with chronic or acute inflammatory diseases, diabetes, cancer or autoimmune diseases, and those who had periodontal treatment in the last 6 months or who took antibiotics during this period were not included in our study. Peripheral blood samples were collected by venous blood puncture in a tube containing EDTA. DNA extraction from the samples was made from whole blood or buffy coat using the Salting Out method [ 30 ]. The DNA used for genotyping was diluted to a concentration of 50 ng/μl and stored at -20°C. The quality of the extracted DNA was analyzed by optical density in a Thermo Scientific Nanodrop 2000® (Wilmington, USA). The genotyping of IL16 polymorphisms (rs11556218 , rs4778889 and rs4072111) was performed according to Luo et al. [ 18 ] with modifications. Primer sequences were based on those described above [ 18 ] and tested in Primer Blast (NCBI) ( https://blast.ncbi.nlm.nih.gov/Blast.cgi ). Polymerase chain reaction (PCR) was performed in 10 μl with final concentration of 7.5 ng of DNA, 2.0 ng of each primer, 1.25 mM of MgCl2, 0.2 mM of dNTP (Invitrogen®, Frederick, MD, USA), 1X PCR buffer (5X Green Goat® Buffer, Promega, USA) and 0.5U of Taq Polymerase (GoTaq® Polymer DNA Polymerase, Promega, USA). The cycles were performed on the Veriti thermocycler (Applied Biosystems): five-minute cycle at 95°C; 35 cycles of 30 seconds at 95°C, 45 seconds at 61°C (for rs11556218 ), 67°C (for rs4778889 ) or 63°C (for rs4072111 ) and one minute at 72°C; at the end, 10-minute at 72°C. PCR products were digested for 16 hours at 37°C with NdeI digestion enzymes (Invitrogen®, Frederick, MD, USA) for rs11556218 , BsmAI (Invitrogen®, Frederick, MD, USA) for rs4072111 and AhdI (Invitrogen®, Frederick, MD, USA) for rs4778889 . The results were observed by electrophoresis in 3.5% agarose gel with SYBR™ Safe (Invitrogen Life Technologies, Grand Island, NY, USA) with a run of 70V for 15min and then 90V for 35min, 300mA and 150W. The result was visualized under ultraviolet light. To ensure the quality of the results, we compared the frequency of the less frequent allele with other populations and analyzed whether the distribution of genotype frequencies for all analyzed SNPS was in accordance with Hardy-Weinberg equilibrium for the control group. Serum levels of IL-16 were determined using an Enzyme-Linked Immunosorbent Assay (ELISA) kit (Sigma-Aldrich Co LLC, N°RAB0261, St. Louis, Missouri, USA) according to the manufacturer's instructions. The absorbance of each well was read in ELISA reader (ASYS HITECH GMBH—Eugendorf, Austria) using absorbance at 450 nm. The IL-16 serum levels (in pg/ml) of samples tested were estimated using the calculation of the standard curve. Reactions were made in duplicates. To obtain the minimum number of samples appropriate for this study with adequate statistical power (≥80%), the quantitative calculation software QUANTO [ 31 ] was used in the study design. At the significance level of 0.05, two-tailed test, K P = 0.50, for Minor Allele Frequencies (MAFs) of rs11556218 , rs4072111 and rs4778889 , power values of 93.2%, 80.5% and 92.2% were obtained, respectively, to detect a risk effect of 2.0, under a dominant model. Associations of IL16 polymorphisms and descriptive analyses were performed using software SNPStats ( https://www.snpstats.net/start.htm ) [ 32 ] and the software OpenEpi version 3.01 ( https://www.openepi.com/Menu/OE_Menu.htm ). The association tests were based on linear or logistic regression according to the response variable. The analysis for SNPs was performed by choosing models of genetic inheritance (codominant, dominant, recessive, overdominant and log-additive) according to the Akaike information criterion (AIC) and adjusting for possible interactions with the studied covariates [ 32 ]. The Odds Ratio (OR) with confidence intervals (CI) of 95% was used. For the analysis of multiple SNPs, the Haploview and SNPStats software were used [ 32 , 33 ] were used. The haplotype frequencies were determined using the expectation-maximization (EM) algorithm and the possible association of these haplotypes with PD was evaluated by the chi-squared test; furthermore, the linkage disequilibrium (LD) analysis between the SNPs and the permutation test for statistically significant results were performed. For statistically significant results, Bonferroni correction (BC) method was applied by multiplying the P values by the number of SNPs evaluated, as there was no detectable LD between them. The frequency distributions of genotypes were evaluated to ensure Hardy-Weinberg equilibrium for all SNPs in the populations [ 32 ]. Statistical analyses for ELISA assays were performed using Wilcoxon's U test, on software R version 3.5.2. To avoid the effect of biases such as smoking, smokers and nonsmokers were also analyzed separately. It was considered significant for all tests P-value <0.05.

(DOCX) Click here for additional data file. (DOCX) Click here for additional data file.