10-12-week-old male C57BL/6J mice were purchased from Jackson Laboratories (Bar Harbor, Maine). All mice were maintained in a specific pathogen free (SPF) environment and housed in the same room. Mice were provided with water and a standard laboratory diet ad libitum. They were supplied with hardwood chips as bedding and housed in a temperature-controlled, air-conditioned room on a 12-hr light-dark cycle. For each independent experiment, 10 mice were assigned to the Quercetin treated group, while 5 mice were assigned to the vehicle treated group. All the mice were cohoused for two weeks to standardize microbial communities prior to the start of experiments (33). On day 1 of the study, mice were randomly assigned to experimental groups by husbandry staff with no involvement in the study design. Thereafter, mice were housed in separate cages based on experimental groups. Cages were changed daily to prevent Quercetin redosing by coprophagy and prevent feces accumulation and limit coprophagy (34, 35). Mice either received twice daily oral administration of Quercetin (40mg/kg/twice per day; Cayman Chemical: Item No.10005169) or an equal amount of phosphate buffered saline (PBS: Life Technologies) for 5 days. Body surface area was used to determine the concentration of Quercetin to be used in the mice. By using body surface area, a 40mg/kg dosage in mice would equate to 195mg of Quercetin for a 60kg human (36). Quercetin was solubilized in an ethanol-water mixture to enhance its absorption (37). A one-inch straight stainless-steel oral gavage needle with a round-ball stainless steel tip was placed at the end of a 1mL syringe and it was used for oral administration. Quercetin and vehicle were administered in small, 50µl, doses to make sure all the compound was swallowed before the introduction of the remaining medication. The total dosing volume was 300 µl of either Quercetin or vehicle two times daily. Periodontal inflammation and bone loss were induced by ligature placement on day 6 of the experimental timeline. Quercetin-treated and vehicle-treated were anesthetized intraperitoneally with a 200 µl mixture of ketamine (10mg/mL) and xylazine. Black braided 5-0 threads were placed at the left side of the maxilla interdentally between the first and second molars, while the right sides remained un-ligated as controls. Twice daily oral administration of Quercetin or vehicle treatment resumed on day 7 and mice were euthanized one-week post-ligation. Oral swabs of teeth and gingival tissue were taken before the induction of periodontitis on day 1 and day 6 and final swabs were taken at the completion of the study on day 13. The gingival tissues around all maxillary molars, without including palatal or buccal mucosa, were carefully excised using a new sterile #15 scalpel blade under 3.5x magnification by a trained periodontist using Dental Loupes (38–41). The mice were excluded from the study if the underlying bone was damaged during excision of the gingival tissues or the ligatures were lost. Maxillary jaws and gingival tissues were assessed for bone loss and inflammatory mediators and cellular infiltrate, respectively (42–44). The Institutional Animal Care and Use Committee of Virginia Commonwealth University approved all procedures and experiments were carried out following their guidelines.

The bacterial strains were Porphyromonas gingivalis (strain ATCC 33277) and Fusobacterium nucleatum (strain ATCC 25586) and grown anaerobically (5% CO₂, 10% H₂, 85% N₂) at 37°C. P. gingivalis was maintained in Brain Heart Infusion broth supplemented with 0.05% Yeast Extract, 5g/mL Hemin, 0.5g/mL Vitamin K and 0.1% cysteine. F. nucleatum was grown in Brain Heart Infusion broth containing 5 mg/L Hemin, 0.5 mg/L menadione, 1g/L cysteine, 5g/L Yeast Extract, 0.001% N-acetyl muramic acid, and 5% fetal bovine serum. The bacterial suspensions were killed by heating at 80°C for 10 minutes. Sterility of killed bacteria was confirmed by enumeration of CFU on an agar plate.

The human monocytic leukemia cell line (THP-1; ATCC TiB-202) were grown in in RPMI 1640 (Life Technologies, Cat# 11875093) media supplemented with 10mM HEPES (Sigma, Cat# 1101100250), 1mM sodium pyruvate (Fisher Scientific, Cat# 11360070), 4500 mg/L glucose (Fisher Scientific, Cat# D16-500), 0.05 mM mercaptoethanol (Sigma, Cat# M6250-100ML) 10% fetal bovine serum, and 1% penicillin/streptomycin (Invitrogen, Cat# 15140122) at 37°C in a humidified 5% CO₂ atmosphere. The cells were maintained in a logarithmic phase of growth (2 × 10⁵ to 8 × 10⁵) by passage every 3 to 4 days and were subsequently used for in vitro stimulation assays.

Maxillae were dissected after euthanasia and fixed in 10% neutral-buffered formalin for at least 24 hours prior to imaging and scanned with a desktop micro-CT system (Brüker Skyscan 1173, Skyscan NV, Kontich, Belgium) at a resolution of 1,120 × 1,120 pixels (image pixel size of 15.82 µm) over 180^o,80 kV voltage, 80 µA current and 250 ms exposure time. Five x-ray projections were acquired every 0.2° and averaged. A standard Feldkamp reconstruction was done using NRecon software (Brüker) with a beam-hardening correction of 15% and a Gaussian-smoothing kernel of 1. Histomorphometric analysis was performed with the DataViewer MicroCT visualization Software (Brüker, Kontich, Belgium) with a 1,092 × 1,092 pixel size in all three spatial dimensions and setting the sagittal plane parallel to the X-ray beam axis. The amount of linear alveolar bone present was calculated and measured with DataViewer software using sagittal images measuring the distance from the cemental-enamel junction (CEJ) to the alveolar bone crest (ABC). Histogram settings were set to 255 to measure the CEJ-ABC junction distance and also to distinguish between hard tissues (i.e. alveolar bone) and enamel from the soft tissues. Linear measurements were taken (in millimeters) from the CEJ to the ABC in the interdental region between the first and second molars (M1-M2) or the second and third molars (M2-M3). All scans were reoriented with DataViewer to the same position for bone loss evaluation such that the CEJ and the root apex next to the measurement appeared in the micro-CT slice that was to be analyzed and standardize measurements according to previously established methods (44–51). Specifically, all images were oriented so that the CEJ of the first and second molar were parallel to the horizontal axis and the CEJ and root apex appeared in the same slice generating a standardized region of interest with set anatomical limits. A total of four CEJ-ABC measurements were made from the sagittal view in DataViewer. The average of the four linear measurements from each mouse was used for the subsequent analysis. Measurements were made by a blinded-investigator and analysis was performed three times for each mouse with similar results. Data are presented from 5 independent experiments from a total of 37 Quercetin-treated mice and 21 vehicle-treated control mice. One-Way ANOVA analysis using GraphPad Prism software from GraphPad Software Inc (La Jolla, CA, USA) was performed to compare the means of the 4 independent groups (Control+Vehicle, Ligature+Vehicle, Control+Quercetin, Ligature+Quercetin). 3D representative images were obtained using CTVox 3D visualization software (Brüker). Linear measurements were not made using this software as Brüker’s 3D visualization software solely offers 3D model creation, viewing and flexible control of Micro-CT scans (52–54).

Gingival tissues from Quercetin-treated and vehicle-treated control mice were removed using previously established protocols with slight modifications (44, 55). Using a #15 scalpel blade and 3.5x magnification Dental Loupes, a trained periodontitis excised the gingival tissues surrounding the three maxillary molars on the ligated or control side of the maxilla without including buccal and palatal tissues and damaging underlying tissues or bone (38–41). Tissues were fixed in 4% formaldehyde and sent to the Virginia Commonwealth University Cancer Mouse Models Core Laboratory where they were processed to paraffin. Briefly, fixed tissues were processed under a vacuum using the automated Tissue Tek Tissue Processor VIP (Sakura Finetek, Torrance, CA). The processor dehydrated the tissue using gradual increases in concentrations of ethanol from 70% to 100%, then was cleared with Citrisol (National Diagnostics) finishing with 4 changes of Paraplast Plus (VWR) paraffin wax at 60 degrees. Processed and paraffin embedded samples are then individually placed into a stainless steel mold containing molten Paraplast Plus with an embedding ring and allowed to harden on a cryo plate for 20 minutes before removing from mold. Using a standard rotary microtome, 5µm thick sections are floated on a 40-degree C water bath and then mounted onto positive charged slides and set on end to dry at room temperature. Progressive H&E staining of paraffin embedded sections is performed using the automated Agilent Dako CoverStrainer slide processing system. The instrument process involves baking, dewaxing, hydrating and staining through to the dehydrated, coverslipped and dried slide. Hematoxylin, Eosin and Bluing Buffer and mounting medium are purchased from Agilent Technologies as validated ready-to-use reagents for the Dako CoverStrainer. Histology images were acquired using the Q-Color 5 imaging system from Olympus Microscopy with a 10× magnification objective lens. Quantification of numbers of nucleated (hematoxylin-positive) in the gingiva was performed by a blinded examiner using CellSens software at 20× magnification. Counting was performed using the “Count and Measure” tool and the “Manual Threshold” option to choose an initial nucleated cell for reference. Once the cell was selected, subsequent cells were automatically selected until all nucleated cells in the connective tissue were highlighted. Data were reported as area stained (square micrometers) in each field of view. Nine regions/fields of view were analyzed per ligated side, and six regions/fields of view were analyzed per control side, per mouse, respectively. A minimum of 7 mice were examined per treatment group and data are representative of three independent analyses.

Gingival tissues from Quercetin-treated and vehicle-treated control mice were isolated and harvested using previously established protocols with slight modifications and as described in the previous sections (44, 55). Total RNA was extracted from gingival tissues using the RNeasy kit (Qiagen, Cat# 74136) and genomic DNA (gDNA) eliminator spin columns. The RNA concentrations were determined with NanoDrop. For each gingival side of each mouse approximately 1800 ng- 7500 ng of total RNA was isolated. Samples were never pooled of mice receiving the same treatment. 800 ng of total RNA was used for cDNA synthesis with High Capacity cDNA Reverse Transcription Kit, following the users’ manual (Applied Biosystems, Cat# 4368814). Primers were obtained by Invitrogen. Sequences for primers used were as follows: mIL-6 forward: 5′-TCTATACCACTTCACAAGTC GGA-3′, mIL-6 reverse: 5′-GAAT TGCCATTGCACAACTCTTT-3′; mTNF forward: 5′-CTGAACTTCGGGGTGAT CGG -3′, mTNF reverse: 5′-GGCTTGTCACTCGAATTTTGAGA-3′, TNFAIP3/A20 (mouse) forward: 5′-AGGTCGGTGTGAACGGATTTG -3′, and reverse: 5′-GGACAGTTGGGTGTCTCACATT-3′; TNFAIP3/A20 (human) forward: 5’-TTGTCCTCAGTTTCGGGAGAT-3’ and reverse: 5’-ACTTCTCGACACCAGTTGACTT-3’. Results were normalized with respect to the values obtained for the housekeeping gene GAPDH. ΔCt was calculated by subtracting the Ct value of the housekeeping gene (GAPDH) from the target gene Ct value. ΔΔCt was calculated by subtracting the control group target gene ΔCt from the experimental group target gene ΔCt. Relative mRNA expression was calculated by 2^-ΔΔCt. Data are presented as averages of 3 independent experiments from 26 Quercetin-treated mice and 11 vehicle-treated control mice.

Bacteria were obtained by oral swabbing of the teeth and gingival surface with Ultra-Fine polystyrene swab (Puritan Medical Products) and placed in DNA/RNA Shield (Zymo Research, Irvine, CA). Samples were set to ZymoBIOMICS for Targeted Metagenomic Sequencing. DNA Extraction was performed using ZymoBIOMICS-96 MagBeadDNA Kit (Zymo Research, Irvine, CA). The DNA samples were prepared for targeted sequencing with the Quick-16S NGS Library Prep Kit (Zymo Research, Irvine, CA). The primer sets used were Quick-16S Primer Set V3-V4 (Zymo Research, Irvine, CA). These primers were custom-designed by Zymo Research to provide the best coverage of the 16S gene while maintaining high sensitivity. The sequencing library was prepared using an innovative library preparation process in which PCR reactions were performed in real-time PCR machines to control cycles and therefore limit PCR chimera formation. The final PCR products were quantified with qPCR fluorescence readings and pooled together based on equal molarity. The final pooled library was cleaned up with the Select-a-Size DNA Clean & Concentrator (Zymo Research, Irvine, CA) then quantified with TapeStation (Agilent Technologies, Santa Clara, CA) and Qubit (Thermo Fisher Scientific, Waltham, WA). The ZymoBiomics Microbial Community Standard (Zymo Research, Irvine, CA) was used as a positive control for each DNA extraction and each targeted library preparation. Negative controls (blank extraction control, blank library preparation control) were included to assess the level of bioburden carried by the wet-lab process.

The final library was sequenced on Illumina MiSeq with a v3 reagent kit (600 cycles). The sequencing was performed with >10% PhiX spike-in. The quality control, feature table construction, diversity and taxonomic analysis were performed in QIIME2. The t-Distributed Stochastic Neighbor Embedding (t-SNE) plot was created by Rtsne package and the differential abundance was tested by ALDEx2 package in R. The raw data with quality scores lower than 25 was trimmed in a quality control step. Alpha diversity was performed by testing the observed OTU, evenness and Shannon index. The significance of alpha diversity was analyzed using Kruskal–Wallis test and corrected using the Benjamini–Hochberg procedure with a false discovery rate of 5%. Beta diversity metrics was computed using Bray-Curtis distance, following by PERMANOVA statistical analysis and corrected using the Benjamini–Hochberg procedure with a false discovery rate of 5%. PERMANOVA is a multivariate ANOVA with permutations. It is meant to test differences between groups with a lot of variables and with permutations to avoid possible biases. Statistical differences reveal that the distribution and abundances of experimental groups are different (56). The ordination plot was generated by performing dimensionality reduction using t-SNE. Silva-132-99-nb-classifier was used to assign taxonomy to the OTUs, following by differential abundance analysis using LEfSe (57). The abundance of taxa was determined by dividing the hits of a specific taxon by the total hits of all taxa in the sample. Our analyses identified a total of 163 species which were present in levels greater than 0.1% in more than 5% of total samples. These bacteria were included for differential abundance analysis. Sequencing was performed in each individual sample once. Additionally, the OTUs were also assigned to taxonomy using HOMD 16s rRNA database, following by the same steps for differential abundance analysis.

THP-1 cells were suspended at a density of 3 × 10⁶ in 6-well plates and treated with 25ng/mL of phorbol 12-myristate 13-acetate (PMA: Sigma, Cat# P1585) overnight to differentiate into macrophages (43, 58, 59) The cells were then treated with Quercetin (5µg/mL) dissolved in ethanol for 2 hours prior to stimulation with bacteria and toll-like receptor agonists. The cells were challenged with heat-killed P.gingivalis (multiplicity of infection [MOI] of 1:100), F.nucleatum (MOI 1:50), P. gingivalis lipopolysaccharide (LPS) (10µg/mL) (In vivogen, Cat# tlrl-ppglps), Pam3Cys-Ser-(Lys)4 (Pam3CSK4) (10ng/mL)(In vivogen, Cat# tlrl-pms), and CpG oligonucleotide (ODN) 2006 (100µg/mL)(In vivogen, Cat# tlrl-2006-5) for up to 24 hours. Inflammatory cytokine levels (TNF and IL-6) were determined in cell-free culture supernatants using ELISA (TNF ELISA kit: Thermo Fisher, Cat# 88-7346-22 and IL-6 ELISA kit: Thermo Fisher, Cat# 88-7066-22). For in vitro assays evaluating A20 mRNA levels, the cells were challenged with heat-killed F.nucleatum (multiplicity of infection [MOI] of 1:10), and Pam3Cys-Ser-(Lys)4 (Pam3CSK4) (10ng/mL)(In vivogen, Cat# tlrl-pms) for up to 6 hours. Total RNA was isolated using RNeasy plus Mini Kit by QiaCube (Qiagen) and 800ng of total RNA was used for cDNA synthesis and subsequent qRT-PCR. All in vitro studies were performed with at least three sets of independent experiments with a minimum of 3 replicates per time point/stimulation.

THP-1 cells (3x10⁴) were grown on 8-chamber polystyrene vessel tissue culture treated glass slide (BD Falcon, USA) and pretreated with 25ng/ml PMA to differentiate to macrophages overnight. The cells were then treated with Quercetin (5µg/mL) dissolved in ethanol for 2 hours prior to stimulation with bacteria and toll-like receptor agonists. The cells were stimulated with heat-killed P.gingivalis (multiplicity of infection [MOI] of 1:100), and F.nucleatum (MOI 1:50), for up to 60 minutes. Following treatment, cells were washed in cold PBS and then fixed in 4% paraformaldehyde. The cell membrane was then permeabilized in buffer containing 0.3% Triton-100 and 0.1% NaN3 in PBS for 30min at RT. Nonspecific binding sites were blocked by incubation with 5% BSA, 0.1% NaN3 and 0.3% triton-X in PBS in a humidity chamber for 1h at RT or overnight at 4°C. Cells were incubated with anti-NF-κBp65 antibody (ProteinTech, 1:100) overnight at 4°C. After incubation, cells were rinsed twice for 5 min in PBS with gentle shaking and incubated with Alexa 488 goat anti-rabbit polyclonal antibody (Invitrogen, 1:250) for 1h at RT. Cells were washed 6 times in PBS with gentle shaking. Coverslips were mounted with antifade mounting medium with DAPI (Vectashield Vibrance) and kept under dark conditions at RT. The images were collected by Nikon confocal laser microscope at a magnification of 630X and zoom 2 to assess nuclear localization of NF-κB. For each group a minimum of 4 images were collected. Three sets of independent experiments were conducted with a minimum of 120 cells analyzed.

Statistical analysis of microbial samples is given in the Sequencing Data Analysis section above. Statistical analysis for all other experiments was performed using GraphPad Prism software from GraphPad Software Inc (La Jolla, CA, USA). All data were expressed as the mean ± the standard deviation. The difference between two groups was established by the unpaired t test with Mann-Whitney correction. Multiple group comparisons were performed by one-way ANOVA with Tukey’s post hoc test to identify differences between specific groups. A value of p<0.05 was considered to be statistically significant. The number of animals examined per group and number of times the experiments were carried out are given above and in the Figure Legends.

The efficacy of Quercetin to modulate the immune responses and periodontal disease phenotype was assessed using the murine ligature-induced periodontitis model. Quercetin was administered orally at a dosage (40mg/kg/twice daily) comparable to humans a week before ligations and continued throughout the experimental period as described ( Figure 1A ). As anticipated, alveolar bone loss was observed in both groups following ligations ( Figures 1B, C ). Supporting a protective effect, the linear bone loss in the distance from the cementoenamel junction and alveolar bone crest was reduced in mice which received oral Quercetin supplement compared to the control group ( Figure 1D ). Periodontal bone loss typically mirrors the degree of inflammation present, therefore we further assessed immune cell infiltration in the dissected gingival tissues using H&E staining. In corroboration with the bone loss data, ligated sites displayed increased inflammatory cell infiltration in both experimental and control groups whereas the immune cell infiltrate in the ligated tissues of Quercetin-supplemented mice was significantly reduced compared to that of the mice in the vehicle group ( Figures 1E, F ). We subsequently determined gingival tissue cytokine expression to further elucidate the effect of Quercetin on the inflammatory response in the oral mucosa. As expected, induction of periodontitis led to the increased TNF and IL-6 expression in the gingival tissues. Substantiating the efficacy of oral delivery of Quercetin and anti-inflammatory action, the gingival tissues dissected from Quercetin-supplemented mice displayed significantly diminished cytokine expression profiles compared with their vehicle-treated counterparts ( Figures 1G, H ).

To characterize Quercetin effects on the temporal dynamics of inflammation-induced microbial dysbiosis within the oral cavity, changes in the oral microbiota were assessed upon treatment with Quercetin or vehicle matched control before and after periodontitis onset as described in Figure 2A . Oral bacteria were collected at the baseline of the study (Day 1), time of ligature placement (Day 6), and after the disease onset (Day 13). Before the onset of periodontitis (Days 1 and 6), the oral microbiota of each treatment group displayed increased alpha diversity which is consistent with healthy microbiota compared to those collected following disease establishment (Day 13) ( Figures 2B, C ). After periodontitis onset, the Evenness index was decreased compared to healthy group indicating few taxonomic groups dominate during the disease ( Figure 2B ). Similar results were obtained with the Shannon index, reflecting less community richness and evenness in the oral microbiota of animals driven by experimental periodontitis ( Figure 2C ). Confirming the efficacy of oral Quercetin delivery, after periodontitis onset, Quercetin-supplemented mice exhibited significantly higher Evenness and Shannon indices relative to the vehicle-treated group ( Figures 2B, C ). These results suggest that Quercetin supplement aids to maintain the diversity of the oral microbiome. The taxonomic compositions of each treatment group were also distinguished by dissimilarity ( Figures 2D, E ). Through visualization of clusters using the t-SNE algorithm, two-dimensional ordination demonstrated similarity of the oral microbiome between the treatment groups prior to periodontitis induction ( Figure 2D ). However, after periodontitis onset, treatment groups were well-separated indicating heterogeneity between the oral microbial composition of Quercetin and vehicle treated mice ( Figure 2D ). Variation in community composition among treatment groups was also characterized by Bray-Curtis measures of taxonomic distance ( Figure 2E ). As expected, significance was only observed in the microbial composition between Quercetin and vehicle-treated mice after the induction of periodontitis ( Figure 2E ). Corroborating our alpha and beta diversity analyses, the mice from both treatment groups displayed similar taxonomic profiles before the development of periodontitis ( Figures 3A, B ). However, after periodontitis onset, there were several species-, genus-, and family- level differences between bacterial communities in Quercetin-treated and vehicle-treated mice ( Figure 3C ). Mice which received Quercetin had increased levels of Streptococcus, which is often associated with a symbiotic microbiome ( Figure 3D ). Additionally, mice which received vehicle treatment, displayed increased levels of several bacteria including Enterococcus, Neisseria and Pseudomonas, which are often associated with periodontitis or other inflammatory conditions ( Figure 3D ). Furthermore, when using HOMD reference database specific for the oral microbiome, Quercetin-treated mice displayed increases in Streptococcus sanguinis and Streptococcus parasanguinis, known commensal bacteria in the oral cavity ( Supplementary Figure 1 ).

The interaction between host immune cells and the oral microbiome, and subsequent crosstalk between innate signaling pathways play a critical role in determining periodontal disease outcome. Macrophages are one of the key cells responding to microbial insult in periodontitis pathogenesis. We therefore sought to determine the effect of Quercetin on TLR signaling and cytokine responses in human macrophages challenged with periodontal bacteria and various TLR agonists. P. gingivalis and F. nucleatum were included as representative organisms as they possess unique virulence factors favoring microbial dysbiosis and periodontal tissue destruction and are frequently associated with numerous systemic complications (60–63). Briefly, macrophages were treated with Quercetin for 2 hours prior to challenge with bacteria or TLR agonist, LPS, and cytokine response was assessed post-infection. Corroborating its anti-inflammatory action, Quercetin treatment significantly reduced TNF and IL-6 production in human macrophages following infection with P. gingivalis ( Figure 4A and Supplementary Figure 2A ) and F. nucleatum ( Figure 4B and Supplementary Figure 2B ). Cytokine levels remained comparable in Quercetin-treated unstimulated control macrophages compared to vehicle-treated unstimulated control cells, suggesting that Quercetin functions to mitigate inflammation downstream of TLR signaling. To further characterize the effect of Quercetin on inflammatory response, macrophages were treated with Quercetin or vehicle and challenged with TLR agonists ([P. gingivalis LPS for TLR4], [Pam3CSK4 for TLR2], and [CpG oligonucleotide 2006 [ODN] for TLR9]) for up to 24 hours ( Figures 4C–E and Supplementary Figures 2C-E ). As expected, quercetin treated cells produced significantly less cytokine in response to P. gingivalis LPS ( Figure 4C and Supplementary Figure 2C ), Pam3CSK4 ( Figure 4D and Supplementary Figure 2D ) and ODN2006 ( Figure 4E and Supplementary Figure 2E ) compared to vehicle-treated group.

We then sought to determine the mechanisms that Quercetin can exert its effect on periodontal inflammation and define downstream signaling pathways. The TLR : NF-κB:A20 signaling axis is one of the critical pathways governing inflammation in the oral mucosa through modulating cytokine gene expression. We first assessed NF-κB nuclear translocation, a hallmark of NF-κB activation, in vehicle- and Quercetin-treated human macrophages using immunofluorescence and confocal imaging. Our results showed increased NF-κB translocation in vehicle-treated macrophages upon infection with F. nucleatum ( Figures 5A, B ) and P. gingivalis ( Figures 5C, D ) compared to Quercetin-treated cells indicating that Quercetin regulates oral bacteria-induced inflammatory cytokine response through its effect on NF-κB signaling pathway. The ubiquitin-editing enzyme, A20, is an inducible and broadly expressed cytoplasmic protein that inhibits TNF- and TLR-induced NF- κB activity and previous studies suggested its modulation by Quercetin treatment (9, 10, 64, 65). To determine the effect of Quercetin on A20 expression in the oral mucosa, we next examined A20 mRNA levels in response to Quercetin treatment in human macrophage-like cells challenged with F. nucleatum- and Pam3CSK4 and noted increased A20 expression in cells treated with Quercetin compared to those which received vehicle ( Figure 6A ). Further, we also assessed the effect of Quercetin on gingival tissue A20 expression in vivo. As expected, there was an upregulation of A20 mRNA in the ligated gingival tissues of vehicle treated mice, indicating enhanced inflammation in these tissues ( Figure 6B ). Consistent with the improved disease phenotype, the ligated tissues of Quercetin treated mice displayed similar A20 mRNA levels compared to the control tissues ( Figure 6B ). Overall, these results support in vivo data and confirm the anti-inflammatory effect of Quercetin as it applies to host-microbiome interactions in the oral cavity. It is also of note that consistent with the proven clinical safety profile of Quercetin, we observed no effect on cell viability or structure throughout the experiments which indicates that decreased cytokine production was not due to undesirable effects of Quercetin on macrophages.