Suscettibilità Genetica alla Parodontite Aggressiva: Analisi di Polimorfismi in Geni Candidati

BACKGROUND & RATIONALE Aggressive Periodontitis (AgP) is a complex multifactorial periodontal disease where genetic and environmental factors are thought to play a paramount pathogenic role. Patients with AgP are characterized by a rapid and severe periodontal destruction in mainly younger individuals. While putative perio-pathogenic bacteria constitute the primary extrinsic agent in the etiology of AgP, the risk of developing periodontal disease is not the same for all individuals, suggesting that host factors, potentially related to the genetic make-up, are involved in determining the disease susceptibility. IL-1a and IL-1b, two functionally similar molecules, have been recognized as central proinflammatory cytokines. IL-1ra is an anti-inflammatory non-signaling molecule that competes for receptor binding with IL-1a and IL-1b. The overall contribution of IL-1 to the pro-inflammatory response depends on the balance between these three molecules. Polymorphisms of the Interleukin-1 (IL-1) gene cluster family have been described in association with inter-individual differences of IL-1 levels upon bacterial challenge and they have been proposed as potential genetic markers for periodontal diseases. AIM OF THE STUDY In our previous studies, we have investigated the role of IL-1 genes in explaining the individual susceptibility to both plaque-induced gingivitis and AgP. With respect to AgP, we examined 6 polymorphisms derived from genes of the IL-1 cluster comparing 40 AgP patients with 96 periodontally healthy controls. We tested the association through both the analysis of allelic/genotypic distributions as well as by constructing a map in LD units of the candidate region 2q13–14. Our analyses highlighted the presence of recombination hot spots between the IL-1B+3954 and IL-1B-511 variants and around the IL-1RN (VNTR) marker. The multilocus modelling of association with disease gave a location for the association peak at the IL-1B+3954 marker. Moreover, haplotype analysis identified the IL-1B+3954-IL-1B-511 haplotype as the one having the lowest p-value in the region. In view of this preliminary results, we decided to further investigate the role of IL-1 cluster genes in the pathogenesis of generalized AgP. In particular, we were aware that the main limitations of our previous studies were the small number of the genetic variants investigated and the limited size of the study population. Thus, for the present study, these two constrains have been settled by implementing the study cohort as well as by covering the chromosomal region of interest with an adequate number of single nucleotide polymorphisms (SNPs). MATERIALS AND METHODS Study Population A total of 95 generalized AgP patients (29 males and 66 females, mean age 43.4 ± 7.7) were recruited among those seeking care for periodontal treatment at two different University centers: the Research Center for the Study of Periodontal Diseases, University of Ferrara (56 subjects) and the Department of Odontostomatological, Orthodontic and Surgical Disciplines, Second University of Naples, (39 subjects). Patients were enrolled for genetic genotyping only if they fulfilled the criteria for a clinical diagnosis of generalized AgP at the time of the first (initial) examination according to AAP World Workshop definition. 121 systemically healthy individuals, (60 males and 61 females, mean age 30.1 ± 4,7), were recruited as controls. They were selected if they showed no interproximal attachment loss greater than 2 mm at any of the fully erupted teeth. All subjects enrolled for the study were unrelated Caucasian Italian and chosen among current and permanent residents of Ferrara and Naples areas, respectively. Written informed consent was provided by all participants before study inclusion. The study protocol was approved by the respective local ethical committees. Genotyping Genomic DNA was extracted from peripheral blood leukocyte using the QIAamp blood Kit (QIAGEN Inc., Germany) or a standard proteinase K digestion and phenol/chloroform extraction method. A minimum of 1g of genomic DNA (diluted in 0,2x TE buffer at 10 ng per microliter) from patients and controls was arrayed on 96-well plates and DNA concentrations were determined by fluorescence measurement with molecular probes (Invitrogen, Molecular Probes, USA). For each of the 216 samples of DNA extracted, the genotyping of the 70 polymorphisms selected was realized on Sequenom MassARRAY® platform (Sequenom, San Diego, CA), using protocol iPLEXTM. Statistical analysis A c2 test (p>0.05) was used to assess that the genotype distributions of the polymorphisms fulfilled the criteria of Hardy–Weinberg equilibrium (HWE). The analysis of the potential association between the candidate genes/regions and AgP, in the two samples considered, was verified through the application of the following specific statistical programs: The PLINK software [version 1.00] (http://pngu.mgh.harvard.edu/~purcell/plink/) was used to compare the allele counts in cases and controls by Fisher’s exact test and to calculate odds ratios (OR) with 95% confidence intervals. PLINK was also used to test for population stratification and for allelic heterogeneity between the two subgroups of subjects (from Ferrara and Naples area). From population genotype data, haplotypes were reconstructed by using PLINK toolset which implements a Expectation Maximization (EM) algorithm and performs the sliding window haplotype association analysis by the Chi square test. The CHROMSCAN program (http://cedar.genetics.soton.ac.uk/pub/PROGRAMS), was used to construct a LD map in Linkage Disequilibrium Units (LDUs) which has additive distances and locations monotonic with physical and genetic maps. The level of significance was chosen as P<0.05. When necessary, P values were subject to Bonferroni’s correction for multiple comparisons. RESULTS Control for Hardy-Weinberg Equilibrium After quality control, among the 70 markers selected, eight polymorphisms were excluded from further analysis because two SNPs were found to be monomorphic and six markers had a call rate less than 20%. Among the remaining sixty-two SNPs successfully characterized, three markers in the sample of affected subjects and four SNPs in the group of controls showed significant deviation from HWE. These latter SNPs were excluded from the subsequent analyses. Test for population stratification Based on the whole genome SNPs data, the 216x216 matrix of the pairwise identity-by-state (IBS) genetic distances was constructed using PLINK and used for Multidimensional Scaling (MDS). Through MDS, it was possible to obtain a graph where each point represented an individual and the two axes generated an image of the data reduced in two dimensions. The uniform distribution of points on the graph clearly indicated the absence of heterogeneity between the two subpopulations. To further check for homogeneity among populations, the Cochran-Mantel-Haenszel (CMH) test for stratified samples was also performed. At the CMH test, the most heterogeneous marker observed was the rs895497 variant which resulted not significant different in the two populations [c2 [CMH] = 2.1, p-value=0.146, OR[CMH] = 1.4 (95%CI 0.90-2.08)]. From these tests it was possible to assume that the two subpopulations sampled, Ferrara and Naples, were nearly genetically homogeneous for the investigated chromosomal region and, therefore, in the subsequent analyses, could be considered as a single group divided into cases and controls. Single-point analysis From all allelic and genotypic tests, a significant level less than 5% was found only for genotypic distribution of rs6751201 polymorphism, which maps in the tenth intron of the SEC7 homolog gene (p-value=0,025). However, this association did not hold the statistical significance after Bonferroni correction. LD structure of 2q13-14 region From the CHROMSCAN program, an LD map of the region 2q13–14 under the Malécot model for multiple markers was constructed. The coverage choice for each LDU was of four SNPs, with a mean distance between SNPs of 15.75 ± 11 Kb. The block-step structure of the candidate region was built by plotting, on the kb distances, the LDU locations obtained from: a) HapMap public data from Caucasian population, b) the entire Italian dataset, c) unaffected Italian subjects and d) AgP patients. LD maps showed very similar profiles and a close correspondence was found between the four LD maps. The chromosomal structure of the 2q13- 14 region obtained from Italian Caucasian data, that shows a total length of 16 LDUs, confirmes the profile obtained from the HapMap database. Haplotypic reconstruction On the basis of the "block-step" pattern obtained both from the LD map and from pairwise analysis, the 58 polymorphisms used have been subdivided in four regions grouping 15 (blocks 1 to 4), 10 (Blocks 5 to 8), 13 (Blocks 9 to 13) and 16 (Blocks 14 to 17) SNPs respectively. Within each region, the haplotypes have been reconstructed using the Expectation-Maximization algorithm implemented in PLINK software by means of sliding windows of size three and six (shifting 1 SNP at a time). Haplotype distributions were compared in cases and controls to test for haplotypic association with AgP. None of the possible inferred haplotypes showed a significant association with AgP at the global test. Inferred haplotypes, which contained the gene variants mapping within IL1F6 and IL1F8 (region 1, block 4, SNPs: rs879711, rs895497, rs2305150, rs1562302), demonstrated a significant or borderline association with disease (pvalues ranging from 0.028 to 0.058) both using 3 and 6 windows. Combined haplotypes consisting of SNPs mapping within IL1B (SNPs: rs3917368, rs1143634, rs1143627, rs16944) demonstrated only a very weak association with disease (p-values: 0.046 and 0.053) only for windows of size six. No significant association was found in the region where IL1A and IL1RN map. CONCLUSIONS In conclusion, within the limitation of the present study, our findings failed to support the existence of a causative variant for generalized AgP within the candidate region, where IL-1A, IL-1B and IL1RN genes reside, in an Italian Caucasian population.