2.1 Research subjects

Chronic PD patients were enrolled from outpatient department of stomatology at the Huzhou Central Hospital, Affiliated Cent Hosp Huzhou University between August 2012 and August 2016. PD was diagnosed according to the Armitage standard (1999) [26]. The inclusion criteria were as follows: 1, patients with at least 14 teeth for periodontal evaluation (including at least 4 molar teeth); 2, patients without kin relationship; 3, patients without systemic diseases; 4, patients without history of orthodontic treatment; 5, patients with no use of antibiotics 3 months before the experiment. The exclusion criteria were as follows: 1. the patients received systematic PD treatment; 2. the patients received teeth cleanse treatments 1 year before the experiment; 3. women were in pregnancy and breast-feeding period. The patients in the control group were healthy individuals who were admitted to the outpatient department of stomatology at the Huzhou Central Hospital, Affiliated Cent Hosp Huzhou University during the same period of time. The individuals in the control group had a full dentition periodontal probing depth of ≤ 3 mm, no attachment loss, gingival recession of ≤ 1 mm, and no teeth bleeding during the dental probing.

Informed consent: Informed consent has been obtained from all individuals included in this study.

Ethical approval: The research related to human use has been complied with all the relevant national regulations, institutional policies and in accordance the tenets of the Helsinki Declaration, and has been approved and supervised by the ethics committee at Huzhou Central Hospital.

2.2 Denaturing high performance liquid chromatography (DHPLC) [27, 28]

A total of 5 mL peripheral blood was acquired from each patient with empty stomach in the morning. The blood samples were anti-coagulated by ethylenediamine tetraacetic acid (EDTA) according to the instructions of a DNA Kit (Qiagen Company, Hilden, Germany). The DNA concentration in the samples was measured. The absorbance at 260 nm and 280 nm was measured, and their ratio was always between 1.8 and 2. The reverse transcription quantitative polymerase chain reaction (RT-qPCR) amplification primers for the testing sites were designed and verified using the Primer Premier 5.0 software, and primers were synthesized by Shanghai Sangon Biotechnology Co., Ltd. (Shanghai, China). The primer sequences were: CCATCGGGGAATCAGTGACT (forward sequence of rs5498) and ATGACTGCGGCTGCTACCA (reverse sequence of rs5498), GGAGATGTTTCACGAAGTTTGT (forward sequence of rs3181092) and AAGGGTCACATTTTGAGTATCC (reverse sequence of rs3181092). The RT-qPCR reaction conditions were as follows: pre-denaturation at 94oC for 2 min, 35 cycles of denaturation at 94oC for 1 min, annealing at 52oC for 1 min and extension at 72oC for 1 min, and a final elongation step at 72oC for 5 min. The RT-qPCR products were stored at 4oC and all of the genotypes were detected by DHPLC under the conditions of partial denaturation. During DHPLC, the column temperature was 59.3oC, and the flow rate of mobile phase was 0.9 mL/min. The genotyping of fragment length gene at each site was accomplished in two steps: in the first step, samples with bimodal DHPLC curves were classified as heterozygous genotype; in the second step, the RT-qPCR samples producing unimodal curves were equally mixed with samples with homozygous genotypes confirmed by sequencing, and then measured by DHPLC again. The samples with unimodal curves were classified as wild-type homozygote, while the samples with bimodal curves were classified as homozygous mutates, and the mutation was further confirmed by gene sequencing.

2.3 Determination of periodontal index and the severity of illness in PD patients

Plaque index (PI) was measured using the bacteria plaque index and scoring method of Silness and Loe [29]: 0 = no bacterial plague in gingival zone; 1 = thin bacterial plaques on the surface of gingival zone, which cannot be observed directly by visual examination and can be scraped out with a dental probe; 2 = moderate amount of bacterial plaques visible in gingival zone or on adjacent surfaces; 3 = large amount of soft fouling in gingival sulcus, gingival zone and adjacent areas. Bleeding index (BI) was determined using the bleeding indexing method proposed by Mazza et al. (1981) [30], and was recorded in 6 levels. Probing depth was used to measure the distance from the bottom of gingival pockets to the upper edge of the gingival zone, and 6 loci were measured for each tooth in mm units. Clinical attachment level (CAL) was defined as the value of probing depth minus the distance from the cementoenamel junction to the gingival zone. In case of gingival recession, CAL was the value of probing depth plus the distance of gingival recession.

The Centers for Disease Control and Prevention (CDC) of USA and the American academy of periodontology (AAP) provide grading standards for periodontitis based on the PD surveillance data from different populations [31]: mild periodontitis: < 4 mm periodontal pockets and 1 to 2 mm CAL, alveolar bone adsorption less than one third of root length as detected by X-ray; moderate periodontitis: ≥ 4 mm CAL at ≥ 2 interproximal sites from different teeth, or ≥ 5 mm periodontal pockets at ≥ 2 interproximal sites from different teeth; severe periodontitis: ≥ 6 mm CAL at ≥ 2 interproximal sites from different teeth, or ≥ 5 mm periodontal pockets at ≥ 1 interproximal sites from different teeth.

2.4 Double-antibody-sandwich ABC-enzyme-linked immunosorbent assay (ELISA)

The blood samples were centrifuged at 4oC for 10 min at 3000 rpm/min, and the plasma aliquots were separated and transferred into centrifuge tubes according to the instructions of the Kit (ICAM-1 antibody [Abcam Inc., Cambridge, MA, USA, ab174445] [32] and VCAM-1 antibody [Abcam Inc., Cambridge, MA, USA, ab46118] [33]). The serum samples to be tested were diluted, added into enzyme coated plates, and incubated at 37oC for 30 min. Each well was added with 50 μL of ELISA reactive solution, and incubated at 37oC for 30 min. Subsequently, each well was added with 50 μL of chromogenic agent A and 50 μL of chromogenic agent B, mixed by shaking for 30 s, and colored at 37oC in the dark for 15 min. The optical density (OD) values of each well were measured at a wavelength of 450 nm. Standard curves were then drawn with OD value as abscissa and concentration as vertical coordinate. The concentrations were 25, 50, 100, 200 and 400 ng/L.

2.5 Statistical methods

Statistical analysis was carried out using SPSS 21.0 software (IBM Corp. Armonk, NY, USA). Measurement data were presented as mean ± standard deviations (s.d.). Data obeying normal distribution and homogeneity of variance between two groups were compared using independent sample t-test. Data not conforming homogeneity of variance between two groups were compared using Welch’s approximate t-test. Mann-Whitney U (non-parametric) test was used for data with skewed distribution. Count data were presented using rate or percentage. Pearson’s chi-square test is used for comparing inter-group frequencies when N ≥ 40 and the expected value of all cells T ≥ 5, and chi-square test with correction for continuity was used when N ≥ 40 and 1 ≤ T ≤ 5. Multivariate analysis was done using a binary logistic regression analysis. Chi-square tests were used to verify whether the genotypes were consistent with the Hardy-Weinberg equilibrium model. The p meant bilateral probability, and a p value < 0.05 was indicative to be statistically significant.

3.1 PD correlates to increased ICAM-1 and VCAM-1 levels

There were 196 chronic PD cases, with 102 males and 94 females at a mean age of 45.30 ± 7.20 years included in our study. There were 192 cases in the control group, with 92 males and 100 females at a mean age of 44.89 ± 5.80 years. Following demographic characteristic analyses in the two groups (Table 1), no statistical difference was detected in the gender, age, smoking and drinking history between the two groups (all p > 0.05). The levels of hemoglobin A1c (HbA1c), serum total cholesterol (TC), triglyceride (TG) and low-density lipoprotein (LDL) in the PD group were much higher than those in the control group, while that of high-density lipoprotein (HDL) in the PD group was much lower than that in the control group (all p < 0.05). In addition, ELISA showed that the ICAM-1 level in the PD group was 723.60 ± 158.02 ng/mL, which was higher than that that observed in the control group (589.96 ± 88.95 ng/mL) (p = 0.001). VCAM-1 level in the PD group was 1226 ± 201 ng/mL, higher than that observed in the control group (957.98 ± 192 ng/mL) (p = 0.001).

3.2 A alleles of rs5498 in ICAM-1 and rs3181092 in VCAM-1 increase the risk of PD

In the control group, the Hardy-Weinberg equilibrium was verified with a goodness of fit test. The results showed that the genotype distributions at rs5498 in ICAM-1 gene and rs3181092 in VCAM-1 gene were consistent with the Hardy-Weinberg equilibrium (p > 0.05). Therefore, the samples in this study were random samples and were representative of the general population.

The genotypes and allele frequency distributions at rs5498 in ICAM-1 gene and rs3181092 in VCAM-1 gene are shown in Table 2. The genotypes and allele distributions at rs5498 in ICAM-1 gene were different between the PD group and the control group (p < 0.05). The GG genotype may reduce PD risk (odds Ratio (OR) = 0.529, 95% confidence intervals (CI) = 0.292 to 0.959, p = 0.035), and the G allele may also reduce PD risk (OR = 0.721, 95% CI = 0.539 to 0.963, p = 0.026). At rs3181092 of VCAM-1 gene, the AG genotype (OR = 0.561, 95% CI = 0.348 to 0.905, p = 0.017), GG genotype (OR = 0.263, 95% CI = 0.143 to 0.481, p = 0.001) and G allele (OR = 0.530, 95% CI = 0.398 to 0.705, p = 0.001) may all decrease the risk of PD.

3.3 AA genotypes of rs5498 and rs3181092 of ICAM-1 and VCAM-1 genes correlate to the increase of BI index, CAL and the severity of PD

The correlation of rs5498 and rs3181092 of ICAM-1 and VCAM-1 genes with clinical features of PD is shown in Table 3. The genotype distribution at rs5498 of ICAM-1 gene and rs3181092 of VCAM-1 gene was related to the BI index of PD patients: compared with patients with AA genotype, patients with AG + GG genotypes exhibited decreased BI index (all p < 0.05), CAL: compared with patients with AA genotype, patients with AG + GG genotypes exhibited diminished CAL (all p < 0.05) and the severity of PD: severe patients with AA genotype at rs5498 of ICAM-1 accounted for 48.81%, while mild patients with AG + GG genotypes accounted for 56.25% (all p < 0.05); severe patients with AA genotype at rs3181092 of VCAM-1 accounted for 38.89%, while severe patients with AG + GG genotypes accounted for 21.77%, which was lower than severe patients with AA genotype (all p < 0.05).

3.4 AA genotype of rs5498 and rs3181092 increases levels of SICAM-1 and SVCAM-1

The ELISA results (Table 4) showed that compared with the patients with AA genotype of rs5498, patients with AG + GG genotypes presented with decreased SICAM-1 and SVCAM-1 levels (all p < 0.05). Compared with the patients with AA genotype of rs3181092, there was a downward trend in SICAM-1 and SVCAM-1 levels in patients with AG + GG genotypes (all p < 0.05).

3.5 HbA1c, LDL and AA genotype are risk factors for PD

The related risk factors of PD were analyzed using the logistic regression analysis. The results revealed that HbA1c (Exp (B) = 1.512, 95% CI = 1.185 to 1.930, p = 0.001) and LDL (Exp (B) = 1.783, 95% CI = 1.204 to 2.641, p = 0.004) may amplify PD risk, while HDL may reduce the risk of PD (Exp (B) = 0.418, 95% CI = 0.208 to 0.840, p = 0.014).

Using the AA genotype at rs3181092 in VCAM-1 gene as a reference, the genotype AG + GG may abate the risk of PD (Exp (B) = 0.344, 95% CI = 0.127 to 0.932, p = 0.036) (Table 5).