(6) Gingival Crevicular Fluid Metabolite Changes After Periodontal Therapy

Introduction: The purpose of this study is to identify metabolites in the gingival crevicular fluid (GCF) correlated with probing depth reduction after periodontal treatment.

Methods: Forty two subjects with Stage III/IV periodontitis and 24 healthy controls were examined and sampled every 6 months for 24 months. GCF was harvested from 6 Ramfjord teeth plus the 2 most diseased sites “D” in diseased subjects “D” (D/D) or healthy sites “H” in healthy subjects “H” (H/H) utilizing a porous-silver disk inserted in the gingival sulcus for 30 seconds. Samples were stored in BSS:PBS:D20:NaN3 buffer at -80°C and eluted on the day of proton-NMR analysis. A gradient-enhanced two-dimensional Total Correlation Spectroscopy (2D 1H-1H TOCSY) identified unique resonances matching to known metabolite standards. Metabolite signals were log-transformed to a normal distribution, normalized to compare relative abundances and scaled within one standard deviation. Each metabolite was compared between the three site/groups (H/H, D/D, H/D) using mixed-effects linear models, with a fixed effect for group and a random effect for subjects to account for dependence among sites in the same subject. Comparisons were adjusted for multiple comparisons (Benjamini 2015) to assess the false discovery rate (FDR). Partial Least Square-Discriminant Analysis was used to construct and examine the feasibility of a prediction model with “group” (H/H, D/D, H/D) as the outcome and “metabolite” as predictors. Analyses were conducted using R version 4.0.3 (R-Core Team2020) including the packages lme4 version 1.1-26 (Bates 2015) and mixOmics version 6.12.2. Only sites with PD change of at least ± 2 mm were included in the treatment response analysis.

Results: Thirty five metabolites associated with health or periodontitis were identified. Cross sectional comparison of D/D vs H/H identified an association of putrescine-lysine, cadaverine-lysine and tyramine with D/D, while uracil, tyrosine and alpha-ketoglutarate were linked to H/H. The comparison of H/H vs. H/D revealed different metabolic profiles, with alpha-ketoglutarate and methionine associated with H/H, while putrescine-, GABA- and alanine-lysine coherent resonances found even in H/D. Within diseased patients, D/D vs. H/D, the comparison associated uracil, tyrosine and uridine with H sites while taurine, cadaverine-lysine and propionate with D sites (Figure 1). The treated D sites revealed low likelihood to find propionate, taurine, glucose, putrescine-lysine, cadaverine-lysine coherent resonances as well as a high likelihood to find metabolites associated with health as uracil, creatine, methionine, isoleucine, aspartate, threonine, valine, arginine, trehalose with an FDR < 20% (Figure 2).

Conclusions: The metabolic profile of diseased sites differs significantly from the ones of healthy sites. Periodontal treatment reduces the metabolites associated with disease and increases the presence of metabolites associated with health.

Lead Author’s Contributions to the Research Project: The Lead Author was responsible patient sampling and re-call, sample lab processing and storage prior to NMR analysis, data collection after NMR analysis, data sorting/organization prior to statistical analysis, results interpretation and the composition of this abstract.