A dysbiotic microbiome triggers TH17 cells to mediate oral mucosal immunopathology in mice and humans

See allHide authors and affiliations

Science Translational Medicine  17 Oct 2018:
Vol. 10, Issue 463, eaat0797
DOI: 10.1126/scitranslmed.aat0797

Probing periodontitis pathology

Periodontitis is a common inflammatory disease that can lead to bone loss. Dutzan et al. used a mouse model of ligature-induced periodontitis and samples from patients to demonstrate that TH17 cells drive disease. People naturally deficient in TH17 cells were less likely to develop periodontitis. These pathogenic cells were dependent on the microbiome, and targeting TH17 cells with a small-molecule inhibitor dampened inflammation and bone loss in mice. Their results reveal mechanisms behind the immunopathology and could lead to effective treatments for this disease.


Periodontitis is one of the most common human inflammatory diseases, yet the mechanisms that drive immunopathology and could be therapeutically targeted are not well defined. Here, we demonstrate an expansion of resident memory T helper 17 (TH17) cells in human periodontitis. Phenocopying humans, TH17 cells expanded in murine experimental periodontitis through local proliferation. Unlike homeostatic oral TH17 cells, which accumulate in a commensal-independent and interleukin-6 (IL-6)–dependent manner, periodontitis-associated expansion of TH17 cells was dependent on the local dysbiotic microbiome and required both IL-6 and IL-23. TH17 cells and associated neutrophil accumulation were necessary for inflammatory tissue destruction in experimental periodontitis. Genetic or pharmacological inhibition of TH17 cell differentiation conferred protection from immunopathology. Studies in a unique patient population with a genetic defect in TH17 cell differentiation established human relevance for our murine experimental studies. In the oral cavity, human TH17 cell defects were associated with diminished periodontal inflammation and bone loss, despite increased prevalence of recurrent oral fungal infections. Our study highlights distinct functions of TH17 cells in oral immunity and inflammation and paves the way to a new targeted therapeutic approach for the treatment of periodontitis.


Periodontitis is one of the most prevalent human inflammatory diseases and poses a serious economic and public health burden (1). In this condition, exaggerated inflammatory responses in the oral mucosal tissues surrounding the dentition (gingiva) lead to immunopathology and destruction of supporting bone (2). To date, the pathogenic drivers of exaggerated destructive inflammation are incompletely understood, and treatment largely aims at reduction of microbial stimulation rather than at targeting of specific immune pathways through host-modulation therapies (2). In the lesions of human chronic periodontitis, expression of the cytokine interleukin-17 (IL-17) (36) and an abundance of T helper 17 (TH17) cells have been reported but not causally linked to periodontal disease pathogenesis (3, 57). Here, we test the hypothesis that TH17 cells are drivers of pathogenic mucosal inflammation in periodontitis.

In various mucosal surfaces, TH17 cells are critical regulators of barrier immunity. Yet, amplification and dysregulation of IL-17–secreting cells in the setting of disease has been linked to immunopathology (810). TH17 cells have been implicated in the development of psoriasis in the skin and colitis at the lower gastrointestinal (GI) tract (810). The physiologic immune protective role of TH17 cells is particularly evident at the oral mucosal barrier. Defects in TH17 cells and in IL-17 cytokine signaling underlie susceptibility to oral fungal infection (candidiasis) in both human and murine models (11). Specifically, patients with primary immunodeficiencies affecting either TH17 cell differentiation or function (12, 13) all have susceptibility to oral candidiasis. Although the homeostatic role of TH17 cells at the oral mucosa is well delineated, the role of TH17 cells in the development of oral immunopathology, particularly in humans, has not been conclusively defined.

Periodontitis provides an attractive setting to study TH17-associated chronic inflammation in humans (14). This disease is particularly common, and affected tissues are readily accessible for obtaining mucosal biopsy samples and corresponding microbiome samples directly from specific mucosal microenvironments, allowing for direct evaluation of the host-microbiome interplay at a specific site. Furthermore, TH17 cells in the oral mucosa not only have a well-established role in oral antifungal immunity (11) but are also associated with local inflammatory lesions in periodontitis. Therefore, oral tissue is a unique setting in which to dissect distinct regulation and function of TH17 cells in tissue immunity and inflammation.

Previous work in our laboratory has highlighted unique requirements for homeostatic development of TH17 cells at the oral mucosa (15). Unlike other barrier sites, such as the skin and lower GI tract, TH17 cells may arise at the oral mucosa independently of commensal colonization. Ongoing damage through mastication is a unique tissue-specific cue for the development of homeostatic TH17 cells in oral tissues in health (15). However, the mechanisms implicated in the amplification and dysregulation of TH17 cells in chronic periodontitis are poorly understood. We aimed to characterize the phenotypic and functional characteristics of human periodontitis-associated TH17 cells and investigated TH17 cell induction and functionality in periodontitis through complementary studies in experimental models and human systems.

Here, we document that expansion of TH17 cells in periodontitis is mediated by the disease-associated microbiome and is dependent on both IL-6 and IL-23, revealing divergent regulation of oral TH17 cells in health versus disease (15). Through complementary human and animal experimentation, we demonstrate that TH17 cells are required for periodontal tissue destruction and can be therapeutically targeted. Collectively, our use of concurrent and consistent evidence from both experimental and clinical models reveal microbially driven TH17 cells as pathogenic drivers in periodontitis.


Inflammatory lesions in periodontitis are dominated by a TH17 cell signature

In the oral disease periodontitis, mucosal tissues surrounding the dentition (gingiva) become heavily infiltrated by inflammatory cells with high representation of T cells (Fig. 1, A to C) (6, 16). Comparison of inflamed gingival tissues from periodontitis patients with healthy gingiva from volunteers revealed that, among known TH-associated factors, IL-17 had the highest expression in periodontitis (Fig. 1D), suggesting a bias toward TH17 cell differentiation in disease lesions (16, 17). Ex vivo stimulation of cell preparations from healthy and diseased gingiva also revealed significantly increased numbers and proportion of CD45+IL-17+ cells in periodontitis samples compared to healthy tissues (P < 0.05; Fig. 1, E and F, and fig. S1A). Characterization of the subtypes of IL-17+ cells revealed that the overwhelming majority (~80%) are CD4+IL-17+ T cells (TH17 cells) with minimal CD8+ IL-17+ cells, TCRγδ+IL-17+ T cells, or innate lymphocytes (CD45+, CD3, CD19, CD20, CD1a, CD11c, CD14, FcɛR1α, CD16, and CD34IL-17+) (Fig. 1G). TH17 cells were significantly increased in periodontitis lesions compared to health (P < 0.05; Fig. 1, H and I, and fig. S1B) and, importantly, their proportion correlated with disease severity as reflected by tissue destruction-bone loss in millimeters (Fig. 1J). TH17 cells in periodontitis were almost exclusively memory cells (CD45RO+CD45RA; ~99%), predominantly tissue resident cells, primarily resident effector memory (rTEM; ~60%), and secondarily resident central memory T cells (rTCM; ~20%) (Fig. 1, K and M). After ex vivo stimulation, periodontitis-associated TH17 cells coproduced cytokines linked to pathogenicity (10) such as granulocyte-macrophage colony-stimulating-factor (GM-CSF; ~30%) and interferon-γ (IFN-γ; ~15%), whereas a subset of TH17 cells coproduced IL-22 (~15%) (Fig. 1, N and O).

Fig. 1 TH17 cells in human periodontitis.

(A) Hematoxylin and eosin staining of health and (B) diseased (periodontitis) gingiva. (C) CD3 immunohistochemical staining in periodontitis (original magnification, ×15). (D) mRNA expression for IFN-γ, IL4, IL17A, and FOXP3 in health (n = 3) and periodontitis (n = 6) (unpaired t test, mean ± SEM). (E to G) CD45+IL-17+ cells in health and periodontitis gingiva. Cell preparations from health and periodontitis were stimulated ex vivo (phorbol 12-myristate 13-acetate/ionomycin) before cytokine staining. (E) Representative FACS plot and (F) graph showing numbers of CD45+IL-17+ cells per standardized biopsy (n = 9 health, n = 6 periodontitis, Mann-Whitney test, mean ± SEM). (G) IL-17+ cellular sources in periodontitis [one-way analysis of variance (ANOVA), Holm-Sidak’s multiple comparisons test, mean ± SEM]. (H to J) CD4+ IL-17+ in health and periodontitis. (H) Representative FACS plot and (I) graph showing numbers of CD4+IL-17+ cells per standardized biopsy (n = 9 health, n = 7 periodontitis, unpaired t test, mean ± SEM). (J) Spearman correlation of CD4+IL-17+ cells with bone loss (in mm; n = 14 patients). (K to O) TH17 cells in periodontitis. (K) Representative FACS plot of CD4+IL-17+ and (L) CD4+IL-17+ cells expressing CD45RO, CD45RA, CCR7, and CD69. (M) Frequencies of CD4+IL-17+ rTEM, rTCM, TEM, and TCM memory T cells. (N) Representative FACS plot and (O) graph of CD4+IL-17+ cells coproducing IFN-γ (n = 9), GM-CSF (n = 4), or IL-22 (n = 4) after ex vivo stimulation. (M) ANOVA and Holm-Sidak’s or (O) Tukey’s multiple comparisons tests, mean ± SEM. All P values are indicated in graphs. ns, not significant.

TH17 cells selectively expand in experimental periodontitis

To mechanistically dissect the role of TH17 cells in periodontal inflammation, we used an established murine model of experimental periodontitis, ligature-induced periodontitis (LIP). In LIP, atraumatic placement of a silk suture (ligature) around the second molar tooth leads to local accumulation of bacteria and gingival inflammation followed by destruction of tooth-supporting bone (fig. S2, A to D) (18). Within the inflammatory lesions of experimental periodontitis, there was a profound increase in Il17a gene expression (30- to 80-fold increase over control) (Fig. 2A), whereas other TH-associated cytokines and inflammatory mediators were essentially unaltered (fig. S2E).

Fig. 2 Expansion of TH17 cells in experimental periodontitis.

(A) IL-17a mRNA expression in control gingival tissues (CTL, n = 10) and after ligature induced periodontitis (LIP, n = 11). Data combined from two experiments, Mann-Whitney test, mean ± SEM. (B and C) CD45+IL-17+ cells in CTL and LIP, in IL-17acreR26ReYFP mice. (B) Representative FACS plot and (C) graph indicating numbers of CD45eYFP+ cells per standardized tissue (n = 12 per group, data combined from three experiments, unpaired t test, mean ± SEM). (D to F) Proportions and numbers of IL-17+ cells in control and LIP. (D) Representative FACS plots and (E) graph showing percentage of eYFP+ cells. (F) Graph showing numbers of eYFP+ per standardized gingival tissue (n = 9, data combined from three experiments). (G) Ki67+ staining in IL-17+ cells in LIP [n = 6, data combined from two experiments, one-way ANOVA and Tukey’s multiple comparisons test, mean ± SEM for (E) to (G)]. (H) Numbers of CD4+eYFP+ cells per gingival tissue in LIP with or without FTY720 (n = 5, data combined from two experiments, Mann-Whitney test, mean ± SEM). All P values are indicated in graphs.

To investigate the cellular sources of IL-17, a reporter mouse designed to map the fate of cells that have activated IL-17 expression (IL-17acreR26ReYFP) was used. This IL-17 reporter mouse was subjected to experimental periodontitis, and cellular sources of IL-17 in control and after 5 days of LIP were evaluated. Consistent with an increase in IL-17a gene expression during LIP, and similar to the human findings, CD45+IL-17+ cell number (Fig. 2, B and C) and CD4+IL-17+ cell proportion and number (Fig. 2, D to F, and fig. S2, F to H) significantly increased in disease lesions (all, P < 0.05). These data confirm this as a highly relevant animal model to dissect TH17 cell induction and functionality in periodontal inflammation.

Distinct cellular sources of IL-17 were involved in periodontal homeostasis versus inflammation. In controls, T cell receptor γδ (TCRγδ+) T cells were the major IL-17 source and represented 60 to 70% of total IL-17+ cells. TCRαβ+CD4+IL-17+ (TH17) cells constituted about 10% of total CD45+IL-17+ cells followed by innate lymphoid cell (ILC) IL-17+ cells (TCRβ, TCRγδ, Ly6C, Ly6G B220, CD11b, CD11c, Thy-1.2, and CD90.2+) and a minimal proportion of CD8+IL-17+ cells at steady state (Fig. 2, D and E). In contrast, in the lesions of experimental periodontitis, the dominant IL-17+ cell population was TH17 cells (Fig. 2, D and E). Because proliferating (Ki67+) IL-17+ cells in the lesions of periodontitis were primarily TH17 cells, followed by γδT IL-17+ cells (Fig. 2G), we attributed the accumulation of TH17 cells to increased local proliferation. Moreover, consistent with local expansion of TH17 cells, numbers of TH17 cells in LIP were essentially unchanged when mice were treated with the lymph node egress inhibitor FTY720 (Fig. 2H). Last, to assess whether TCR signaling was active in cells expanding during experimental periodontitis, we subjected Nur77eGFP mice to LIP. CD4+ T cells expanding in LIP were predominantly Nur77eGFP+. These findings suggest that CD4+ T cells in LIP have active TCR engagement (fig. S3, A to C). Collectively, our data document preferential local expansion of TH17 cells in experimental periodontitis.

Expansion of TH17 cells in periodontitis is dependent on IL-6 and IL-23

We next examined cytokine requirements for accumulation of TH17 cells during LIP. IL-6 has been shown to be required for homeostatic TH17 cell accumulation in gingiva (15). Il6-deficient mice had significantly reduced (almost undetectable) TH17 cells in control and a significant reduction of TH17 cell accumulation during LIP (both P < 0.05; Fig. 3, A and B). IL-1 and IL-23 have also been linked to TH17 differentiation and/or expansion in other settings (9). IL-1 was dispensable for gingival TH17 cell accumulation in LIP (Fig. 3, C and D), whereas IL-23 was required (Fig. 3, E and F). Il1r1-deficient mice had comparable TH17 cell numbers with wild-type controls, whereas Il23a (p19)-deficient mice displayed a significant reduction in TH17 cell accumulation during LIP (P < 0.05). Together, these data highlight divergent regulation of TH17 cells in healthy and diseased gingiva (LIP), with IL-6 being the necessary stimulus for homeostatic accumulation of TH17 cells, whereas both IL-6 and IL-23 (but not IL-1) are required for TH17 cell expansion in disease.

Fig. 3 Cytokine requirements for TH17 cell accumulation in periodontitis.

IL-17 production by CD4+ cells in control and LIP. Representative FACS plots show numbers of TH17 cells after LIP and bar graphs show numbers of TH17 cells in control and LIP in (A and B) Il6−/− and Il6+/+ mice (n = 8, data combined from three experiments), (C and D) Il1r1−/− and Il1r1+/+ mice (n = 5, data combined from two experiments), and (E and F) Il23a−/− and Il23a+/+ mice (n = 6, data combined from three experiments). P values determined by Mann-Whitney test. Data were expressed as mean ± SEM. WT, wild-type.

Expansion of TH17 cells in periodontitis is triggered by a dysbiotic microbiome

We next investigated the local microbiome as a trigger for TH17 cells in periodontitis. We examined the local microbial communities present on the ligature after 2 hours of placement (control) and after 5 days of disease induction (LIP). In comparison to control, LIP featured local microbial communities with shifts in the relative abundance of commensal species and significant alterations in community composition and structure (Fig. 4, A to C; P < 0.05). Therefore, LIP is associated with alterations to local microbial communities consistent with dysbiosis, which we hypothesized may stimulate local gingival TH17 cell amplification in periodontitis. To assess the contribution of the microbiome to TH17 cell accumulation, mice were placed on a broad-spectrum systemic antibiotic (ATB) cocktail (doripenem-vancomycin-neomycin) for 2 weeks and subsequently were subjected to LIP experiments in the presence of continuous ATB treatment. In the presence of the ATB cocktail, the numbers of CD45+IL-17+ cells remained unchanged despite LIP (Fig. 4D). Although ATB treatment significantly affected TH17 cell proportions (P < 0.05; fig. S4A) and cell numbers, it did not affect other IL-17 cellular sources (Fig. 4E). TH17 cell numbers were significantly lower during LIP in the ATB group (P < 0.05), whereas the numbers of TCRγδ+IL-17+ and ILC IL-17+ cells were not affected (Fig. 4E), suggesting that the microbiota specifically triggers TH17 cell accumulation in the gingival mucosa. Reduction of TH17 cell accumulation in the presence of ATBs correlated with nearly absent proliferation of TH17 cells (Fig. 4F). To assess the spectrum of bacteria that mediate TH17 cell accumulation, animals were placed on different single-regimen ATBs during LIP. Treatment with the broad-spectrum ATB doripenem (which targets Gram-positive, Gram-negative, and anaerobic bacteria) was effective in inhibiting TH17 cell expansion (Fig. 4G). Vancomycin and ampicillin, which typically target Gram-positive bacteria, did not affect TH17 cells (Fig. 4G and fig. S4B). Neomycin and metronidazole both significantly inhibited TH17 cell accumulation when used as single agents (P < 0.05; Fig. 4G). Neomycin is considered effective against Gram-negative and some Gram-positive bacteria, and metronidazole targets mostly anaerobic bacteria. ATBs that inhibited TH17 cells led to a concurrent significant inhibition of periodontal bone loss (P < 0.05), whereas vancomycin, which did not inhibit TH17 cells, also failed to inhibit periodontal bone loss (Fig. 4H). These findings suggest a causative link among periodontitis-associated microbiota, TH17 cell accumulation, and periodontal bone loss.

Fig. 4 Disease-associated bacteria trigger TH17 cell expansion in periodontitis.

(A) Microbiome composition at the operational taxonomic unit (OTU) level. The 10 most abundant OTUs are classified at the species level, and less dominant OTUs are shown combined at phylum level. Principal coordinates analysis (PCoA) plot of (B) global microbial community composition and (C) community structure in control and LIP. P values were determined using analysis of molecular variance (AMOVA), and 95% confidence ellipses were depicted. (D) Numbers of CD45+IL-17+ cells in control (n = 6) and after LIP without (n = 7) or with (n = 6) broad-spectrum ATB cocktail (ATB; doripenem-vancomycin-neomycin) (data combined from two experiments, one-way ANOVA and Tukey’s multiple comparisons test, mean ± SEM). (E) Numbers of IL-17+(CD4, TCRγδ, and ILC) in control (n = 6) and after LIP without (n = 7) or with (n = 6) ATBs (data combined from two experiments). (F) Numbers of CD4+IL-17+Ki67+ cells in control and LIP with or without antibiotics [n = 7; data combined from two experiments, Kruskal-Wallis test and Dunn’s multiple comparisons test, mean ± SEM for (E) and (F)]. (G) CD4+IL-17+ cell numbers (n = 6), (H) bone loss (in mm; n = 6 to 9), and (I) total oral microbial biomass (n = 5) in control and after LIP with or without ATB treatment (DOR, doripenem; VAN, vancomycin; NEO, neomycin; MET, metronidazole). Data combined from two experiments, one-way ANOVA and Tukey’s multiple comparisons test, mean ± SEM. All P values are indicated in graphs.

Next, we inquired whether TH17 cell accumulation was dependent on an increase in bacterial biomass. We reasoned that if increased biomass was necessary for TH17 accumulation, then ATBs that inhibit TH17 cells would also lead to a decrease in bacterial biomass. Whereas combination ATBs and neomycin alone were effective in both reduction of microbial load and TH17 cells, doripenem and metronidazole were efficient in inhibiting TH17 cell accumulation without reduction of microbial biomass (Fig. 4I). These data indicate that, rather than global reduction in biomass, specific alterations in microbial communities are likely to be responsible for microbial TH17 triggering in periodontitis.

Last, the specific shifts in microbial communities that are linked to TH17 cell induction were characterized. 16S ribosomal RNA (rRNA) gene-based sequencing of LIP-associated communities in the presence or absence of different ATB regimens revealed effects of each ATB on the oral microbiome (fig. S4C). Combination of broad spectrum ATBs and neomycin targeted all bacteria associated with LIP through significant reduction in bacterial biomass (P < 0.05; Fig. 4I and fig. S4C). Unlike neomycin, doripenem and metronidazole (which also inhibit TH17 cell accumulation) did not target major constituents of LIP-associated communities (cluster 1: Enterococcus sp., Lactobacillus sp., and Pseudomonas sp.; fig. S4C). Doripenem and metronidazole did inhibit other LIP- associated bacteria (cluster 2: Porphyromonadaceae spp., Lachnospiraceae spp., Erysipelotrichaceae spp., among other taxa); however, these constituents were also targeted by vancomycin, which did not inhibit TH17 cell accumulation (fig. S4C). Different ATB regimens also resulted in an increase in relative abundance of select species. Therefore, our analysis did not reveal single bacterial candidates that may stimulate TH17 cell expansion but demonstrated unique shifts of bacterial depletion and overgrowth which occurred with each ATB regimen. These data suggest that changes in the balance of the community may underlie dysbiosis and pathogenicity.

Inhibition of TH17 cell differentiation confers protection from inflammatory bone loss

To definitively implicate TH17 cells as potential drivers of periodontal inflammatory bone loss, a murine model with a genetic defect in TH17 cell differentiation was used. In this model, the TH17 cell transcription factor Stat3 was deleted within the CD4 compartment (Cd4creStat3fl/fl mice). Cd4creStat3fl/fl mice did not have detectable TH17 cells within their gingival mucosa, neither at steady state nor after induction of experimental periodontitis (Fig. 5A and fig. S5A). In contrast, other cellular sources of IL-17 (TCRγδ cells and ILC) remained essentially intact at both steady state and after experimental periodontitis (Fig. 5, C to F), thus validating this model for targeting of TH17 cells. Lack of TH17 cells in Cd4creStat3fl/fl resulted in a significant decrease in total CD45+IL-17+ cells (P < 0.05) only upon experimental periodontitis (i.e., no difference in total IL-17+ cells was detectable at control; Fig. 5, G and H, and fig. S5B). Therefore, TH17 cells have a minimal contribution to the IL-17 pool at steady state but a major contribution to IL-17+ cell expansion during disease.

Fig. 5 Genetic ablation of TH17 cells prevents periodontal bone loss.

Gingival IL-17+ cells in Cd4CreStat3 fl/fl mice and littermates in control and LIP. Representative FACS plots from LIP and graphs showing the number of IL-17+ cells from control and LIP for (A and B) TCRβ+CD4+IL-17+, (C and D) TCRγδ+IL-17+, (E and F) ILC (LinCD90.2+) (lineage: TCRβ, TCRγδ, Ly6C, Ly6G, B220, CD11b, and CD11c), and (G and H) CD45+IL-17+ (n = 6 per group, data combined from three experiments). (I) Bone loss (in mm) after LIP in Cd4CreStat3 fl/fl mice (n = 10) and littermates (n = 9). Data combined from three separate experiments. All P values were determined by unpaired t test and graphs depict mean ± SEM. All P values are indicated in graphs.

Cd4creStat3fl/fll mice displayed significantly reduced bone loss compared to littermate controls (P < 0.05; Fig. 5I). Last, whereas inhibition of Stat3 in CD4 T cells has been linked to expansion of the Foxp3+CD4+ regulatory T cells (Tregs) in other disease models (19, 20), in LIP, we did not detect changes in the proportion or absolute numbers of Tregs in Cd4creStat3fl/fll mice (fig. S5, C to E). However, differentiation of IL17+ cells was inhibited as retinoic acid related orphan receptor γt (RORγt+) and RORγt+Foxp3+ cells were significantly reduced in Cd4creStat3fl/fll mice (P < 0.05; fig. S5, F to K). Therefore, the decreased disease severity observed in Cd4creStat3fl/fll mice may be attributed directly to the lack of TH17 cells rather than to a compensatory increase in Tregs (21).

To confirm these results using an independent approach, we used mice with a specific deletion of the master regulator of TH17 cells, Rorc, within the T cell compartment (LckcreRorcfl/fl mice). Similar to Cd4creStat3fl/fl, LckcreRorcfl/fl mice displayed significantly reduced numbers of TH17 cells at steady state and during periodontitis (P < 0.05) without alterations in the γδT or ILC IL-17+ populations and had reduced periodontal bone loss compared to their littermate controls (Fig. 6A and fig. S6, A to D). Collectively, these data demonstrate that genetic inhibition of TH17 cell differentiation blocks the overall expansion of CD45+IL-17+ cells during experimental periodontitis and confers protection from periodontal inflammatory bone loss. Therefore, TH17 cells are directly implicated as a major pathogenic cell subtype in this inflammatory disease, which supports focus on TH17 cell–targeted therapeutic approaches.

Fig. 6 RORγt inhibition in experimental periodontitis.

(A) Bone loss (in mm) after LIP in LckCreRorcfl/fl mice (n = 5) and littermates (n = 6). Data combined from two experiments. (B) Bone loss (in mm) after LIP in the presence of GSK805 (n = 7) or vehicle (n = 6). Data combined from two experiments. (C) Volcano plot of genes differentially expressed during LIP in the presence of GSK805 or vehicle. Genes in red are P < 0.05 and fold change > 1.5 (GSK805 versus vehicle treatment). (D) Heatmap depicts genes of interest from the top 30 down-regulated genes with GSK805 in LIP [false discovery rate (FDR) < 0.05]. Each column represents a single sample. (E) Bone loss (in mm) with LIP in mice treated with anti–IL-17A or isotype control (n = 8 per group). (F) Bone loss (in mm) with LIP in mice treated with anti-Ly6G or isotype control (n = 5 per group, Mann-Whitney test). All other P values determined by unpaired t test, and graphs depict mean ± SEM unless otherwise stated.

Pharmacologic targeting of RORγt reduces inflammatory bone loss and suggests an IL-17– and neutrophil-mediated immunopathology in periodontitis

Next, the effect of GSK805, a small-molecule inhibitor of RORγt-mediated transcription, TH17 cell development, and function (22), was evaluated in experimental periodontitis. To this end, mice were subjected to LIP in the presence or absence of oral GSK805. GSK805-treated mice displayed significantly reduced disease severity (measured as bone loss; P < 0.05) compared to control vehicle-treated mice (Fig. 6B). Consistent with previous reports in other systems (22), GSK805 treatment preferentially targeted the expansion of TH17 cells and did not affect steady-state sources of IL-17 (fig. S7, A to D), thus establishing GSK805 as a selective TH17 inhibitor also in the setting of periodontitis.

These experiments additionally provided a platform to evaluate mechanisms by which RORγt/TH17 cell inhibition prevents periodontal bone loss. For this, RNA sequencing (RNA-seq) of gingival tissues to assess transcription was used in the presence or absence of RORγt/TH17 inhibition. RORγt targeting inhibited IL-17A/F (Fig. 6, C and D, and fig. S8) and downstream targets, such as factors mediating granulopoiesis and neutrophil recruitment (Csf3, Cxcl1, Cxcl2, and Cxcl5; Fig. 6, C and D, first panel). Proinflammatory genes (Il1β, Il1α, Il6, and Ptgs2) were also inhibited (Fig. 6, C and D, second panel), in line with a role of IL-17 as an amplifier of inflammatory responses. Expression of genes associated with myeloid cell activation (Cd14, Trem-1, and Cxcl3) was also reduced (Fig. 6D, third panel). Last, expression of molecules linked to tissue destruction and bone loss (Mmp3, Mmp10, Pprss22, Osm, and Car9; Fig. 6D, fourth panel) was down-regulated. In summary, our transcriptome analysis after RORγt/TH17-targeted treatment provided insight into differentially regulated genes and pathways, which collectively suggest inhibition of IL-17 signaling primarily affecting the function of neutrophils and other myeloid cells.

Although neutrophils are suspected to mediate tissue destruction in periodontitis (23), no direct cause-and-effect relationship has been reported thus far. We next investigated whether IL-17 and related neutrophil immunopathology is implicated in periodontal bone loss by blocking IL-17. In line with IL-17 playing a primary role in TH17 cell–mediated immunopathology, antibody inhibition of IL-17 led to significant reduction of periodontal bone loss (P <0.05; Fig. 6E). Reduction of neutrophil numbers (fig. S8B) by means of anti-Ly6G treatment also led to significant inhibition of periodontal bone loss (P < 0.05; Fig. 6F). The involvement of neutrophils in periodontal tissue destruction is consistent with neutrophil accumulation in lesions of periodontitis in mice and humans (fig. S8, C to E). Collectively, our data suggest that TH17 cell–derived IL-17 and neutrophils mediate pathology in periodontitis.

Humans with defects in TH17 differentiation present with reduced periodontal inflammation and bone loss

If TH17 cells are important in human periodontitis as our preclinical data suggests, then one would expect protection against periodontal inflammation and bone loss when TH17 cells are reduced or inhibited. To evaluate the clinical consequences of TH17 cell inhibition in humans, we studied patients with the Mendelian disorder autosomal dominant hyper–immunoglobulin E syndrome (AD-HIES), a unique real-life setting to decipher the pathophysiologic role of TH17 cells in humans. AD-HIES patients bear an autosomal dominant, loss-of-function mutation in the STAT3 gene (12), which leads to reduced STAT3 signaling and consequent defective differentiation of TH17 cells (24). These patients do not have circulating TH17 cells, and their T cells display reduced in vitro differentiation into TH17 cells (24). Among the described TH17 cell human defects, AD-HIES is the most common, thereby allowing for evaluation of a fairly large Mendelian cohort rather than select patient cases.

Consistent with blunted TH17 cell differentiation, AD-HIES patients exhibited significantly reduced numbers of TH17 cells within gingival tissues as compared to age- and gender-matched healthy volunteers and periodontitis patients (P < 0.05; Fig. 7, A and B). In line with the functional consequences of blunted tissue–TH17 cell responses and previous reports (25), we documented a selective susceptibility to oral candidiasis in this population, with 85% reporting recurrent oral thrush (Fig. 7C). Therefore, these patients constitute an appropriate cohort for investigating the consequences of local oral mucosal TH17 cell deficiency in periodontal immunity and inflammation.

Fig. 7 Reduced periodontitis susceptibility in patients with genetic defects in TH17 cell differentiation.

(A and B) IL-17 production by CD4+ (CD45+CD3+TCRγδCD56CD8) human gingival cells in healthy volunteers, AD-HIES patients, and periodontitis patients (periodontitis). (A) Representative FACS plots and (B) graph indicating numbers of CD4+IL-17+ cells per standardized gingival biopsy (healthy volunteers, n = 9; AD-HIES patients, n = 4; and periodontitis patients, n = 7). (C) Susceptibility to recurrent oral candidiasis in AD-HIES. Bar graph shows the number of patients with history of recurrent oral candidiasis (thrush, n = 31) or no history of Candida infections (no thrush, n = 5). (D and E) Periodontitis susceptibility in healthy volunteers, AD-HIES patients, and periodontitis patients. (D) Bar graph shows percentage of bleeding sites in healthy volunteers (n = 29), AD-HIES patients (n = 25), and periodontitis (n = 27) patients. (E) Bar graph shows (in mm) clinical attachment loss in healthy volunteers (n = 29), AD-HIES patients (n = 25), and periodontitis (n = 27) patients. All P values in this figure were determined by one-way ANOVA and Holm-Sidak’s test. Data were expressed as mean ± SEM.

Our large cohort of adult AD-HIES patients (n = 35) with confirmed STAT3 mutations was clinically evaluated for presence and history of periodontal inflammation and tissue destruction. In stark contrast to their susceptibility to oral fungal infection, AD-HIES patients did not demonstrate susceptibility to periodontal disease. A detailed clinical evaluation revealed significantly reduced periodontal inflammation and attenuated bone loss (reduced clinical attachment loss; P < 0.05) in AD-HIES patients compared to age- or gender-matched healthy controls and periodontitis patients (Fig. 7, D and E). These findings suggest that blunted TH17 cell responses confer protection from periodontal inflammatory disease in humans. Therefore, our clinical data from AD-HIES patients provide the human relevance of our murine mechanistic experiments and reinforce the distinct roles of oral TH17 cells in immunoprotection and immunopathology.


At mucosal barrier sites, the local immune system is tasked with the delicate act of maintaining a dynamic balance between host and environmental stimuli (26). Although much work in recent years has focused on the regulation of barrier immunity at the lower GI tract and skin, little is understood regarding the critical elements of oral barrier immunity. Yet, the oral mucosa is constantly exposed to a myriad of environmental stimuli (15), including a rich and diverse commensal microbial community. It is the initial portal of encounter for microbes, food, and airborne antigens as they enter the GI tract, all in the context of mechanical stress (mastication) (27). Hence, it is of great significance to understand how immunity is regulated at this critical barrier.

In our current work, we focused on the TH17 cell subset, which is not only pivotal for oral antifungal immunity (11, 25) but also has been documented to expand in the oral inflammatory disease, periodontitis (6, 28, 29). Our studies, using complementary human and murine datasets, demonstrate divergent regulation of oral TH17 cells in health and disease and implicate TH17 cells as drivers of disease pathogenesis and plausible therapeutic targets.

We investigated mechanisms by which TH17 cells are triggered to expand in the setting of periodontitis and potentially participate in disease pathogenesis. In human lesions of periodontitis, we document an expansion of TH17 cells that secrete cytokines associated with pathogenic activity (10, 30) and are predominantly resident memory TH17 cells (31). rTEM cells are thought to be retained in tissues to become terminally differentiated and to be capable of mounting rapid responses to site-specific cues (31). These findings imply that TH17 memory cells expand locally in response to tissue-specific cues and potentially acquire pathogenic functions. In this regard, experimental data in murine models of periodontitis points to a disease-associated microbiome being a local stimulus for TH17 cell expansion and related bone loss in periodontitis.

These studies provide an opportunity to compare and contrast regulation of gingival TH17 cells in health and disease. Our previous work had investigated the physiologic regulation of oral TH17 cells in health. We found that TH17 cells naturally accumulate in the oral mucosa with age (15). This physiologic accumulation of TH17 cells was dependent on ongoing mechanical stimulation, which occurs during mastication, and induces TH17 cell accumulation in an IL-6– and commensal-independent manner (15). Our current work investigates mechanisms involved in further amplification of TH17 cells observed in periodontitis. Amplification of TH17 cells in periodontitis is dependent on microbial triggering. Furthermore, whereas homeostatic TH17 cells are dependent solely on IL-6, disease-associated TH17 cell expansion requires both IL-6 and IL-23. IL-23 is known to be required for pathogenic TH17 cell development in other experimental disease settings (11).

We further investigated microbial triggering of TH17 cells in periodontitis. Microbial induction of TH17 cells has been well documented at various other barrier sites, including the lower GI (32) and skin (33). Barrier TH17 cells have been documented to have specificity for local commensals in the GI tract (34). Antigen-specific TH17 cell responses to commensal bacteria could explain the preferential proliferation of TH17 cells rather than of other cellular sources of IL-17 in the setting of periodontitis and are consistent with Nur77eGFP+ cells expanding in LIP. Innate, antigen-independent TCRαβ+IL-17+ cells have also been shown to expand at the oral barrier (specifically within the tongue) in response to Candida infection (35) and could conceivably be a contributing source of TCRαβ+IL-17+ in periodontitis.

We find that dysbiotic changes in microbial communities, rather than a mere increase in total microbial load, are necessary for induction of TH17 cells that cause disease. Antibiotics that inhibited TH17 cell accumulation did not necessarily reduce microbial load or target specific high-abundance bacterial species. Rather, the protective antibiotics appeared to mediate shifts in the balance of commensal bacteria within communities. However, treatment with protective antibiotics did not reveal specific bacteria that stimulate TH17 cell accumulation but suggested that dysbiotic changes involving anaerobes mediate pathogenic inflammation. The human periodontal disease-associated microbiome is characterized by microbial overgrowth and dysbiotic shifts in bacterial communities, with an overgrowth of Gram-negative anaerobes (36, 37). However, in human disease, it is not practical to dissect whether increased biomass, specific constituents, or global shifts in microbial communities are directly important for disease induction, because evaluation of microbial communities is performed after establishment of disease, not allowing for identification of disease-initiating factors. The present study has succeeded in uncoupling disease initiation from nonspecific biofilm accumulation and further supports dysbiotic changes as the trigger for TH17 cell accumulation and downstream bone loss. In this regard, it should be noted that periodontitis is currently thought to be driven by dysbiosis of synergistic polymicrobial communities characterized by an imbalanced growth of a subset of commensal species (also known as pathobionts) rather than by a single or select few pathogens (14, 3840). Standard-of-care treatment of periodontal disease currently aims at removal of the periodontitis-associated microbial commensal communities (37, 41, 42) by means of mechanical dental debridement with adjunct systemic antibiotics for severe disease (43). Metronidazole, which effectively inhibited TH17 cell accumulation, is an ATB of choice in human periodontitis, because it classically targets anaerobes.

Although the link of commensal microbes to periodontal destructive inflammation is well documented clinically (42), the mechanisms by which commensal bacteria induce pathogenic inflammation in periodontitis are poorly understood. Our work establishes TH17 cells as the missing mechanistic link between microbial stimulation and induction of inflammatory immunopathology in periodontitis. Specific genetic targeting of TH17 cells (by knocking out Stat3 or Rorc in αβ T cells) led to inhibition of TH17 cell differentiation in gingival mucosal tissues, without interruption or augmentation of other cellular sources of IL-17. Inhibition of TH17 differentiation led to significantly reduced bone loss (50 to 70%), suggesting that TH17 cells constitute a major driver of periodontal immunopathology. Our genetic targeting of TH17 cells without interrupting other sources of IL-17 could be important for maintenance of homeostatic immune protection as is the case in the lower GI tract, where total inhibition of IL-17 becomes detrimental for barrier integrity (44). Similarly, in our pharmacological inhibition of RORγt, we selected a small-molecule inhibitor (GSK805) that preferentially targets TH17 cell differentiation (45) while sparing other sources of IL-17 (22). Small-molecule inhibitors of RORγt have been advocated and are currently being tested in TH17 cell–dominated diseases (4648). These inhibitors may be ideally suited for localized diseases, such as periodontitis, where local treatment can avoid systemic effects on thymocyte and lymph node development (47, 49).

Therapeutic inhibition of RORγt/TH17 cells in experimental periodontitis also afforded us the opportunity to investigate mechanisms whereby TH17 cells mediate periodontal immunopathology. By performing RNA-seq transcriptome analysis of periodontitis lesions after 5 days of LIP in the presence of RORγt or vehicle, we evaluated genes inhibited during RORγt/TH17 treatment. Zeroing in on specific gene regulation, we detected a signature consistent with inhibition of IL-17 and IL-17–dependent pathways. Although a dominant inhibition of IL-17– related genes was observed, we cannot completely rule out off-target effects of RORγt inhibition. Top genes inhibited included IL-17–dependent neutrophil recruitment, granulopoiesis, macrophage activation, and overall activation of proinflammatory pathways (50). Identification of inhibited genes also revealed an overall reduced activation of matrix and bone destruction. These findings suggest that the main mediator of TH17 cell–induced immunopathology is IL-17. Inhibition of the cytokine IL-17 significantly inhibited periodontal bone loss. To investigate mechanisms by which IL-17 mediates periodontal immunopathology, we focused on neutrophils as candidate downstream drivers of immunopathology. Top genes inhibited during small-molecule TH17 cell inhibition were related to neutrophil granulopoesis and neutrophil recruitment. Neutrophil accumulation is also known as a hallmark of human periodontitis and has been long speculated as a driver of pathology (23, 51, 52). Our experimentation revealed that reduction of neutrophil accumulation protects from periodontal bone loss. Together, these data suggest a microbial TH17 cell neutrophil axis as a driver for periodontitis immunopathology. Although it is well established that neutrophils secrete tissue-destructive proteases (53), further work is needed to dissect mechanisms by which neutrophils mediate downstream osteoclast activation and bone loss. In this respect, a recent study has shown a requirement for inflammation-associated receptor activator of nuclear factor κβ ligand (RANKL) expression in osteoblasts for induction of periodontal bone loss (28).

IL-17 has previously been shown to be a primary mediator of periodontal immunopathology in settings of altered background immunity such as immune deficiency (54, 55), diabetes (56), and aging (57). We have identified IL-23/IL-17 dysregulation in periodontitis associated with the rare genetic disorder, leukocyte adhesion deficiency type 1 (LAD1), and have exploited that target by IL-23 blockade in humans (55). Despite the commonality of TH17 cell amplification in both endemic and genetic forms of periodontitis, TH17 cell regulation and functionality appears to be distinct in genetic versus common forms of disease. In LAD1, defective neutrophil transmigration leads to tissue neutropenia, which underlies amplification of TH17 cell responses. Furthermore, TH17 cell–mediated bone destruction occurs in the absence of tissue neutrophils in LAD1. In contrast, in common periodontitis, neutrophil accumulation is a hallmark of disease and a driver of periodontal bone loss. Collectively, our previous and current work suggest distinct regulation and functionality of TH17 cells in genetic and common forms of periodontitis, which nonetheless may respond to a common therapeutic target.

Our work further establishes human relevance by evaluating patient cohorts with known genetic defects leading to impaired TH17 cell differentiation (AD-HIES) (24). We document that AD-HIES patients have blunted TH17 cell mucosal responses and susceptibility to recurrent oral fungal infection but reduced periodontal inflammation (bleeding on probing) and bone loss compared to healthy volunteers. These findings are consistent with a key role for TH17 cells in oral inflammatory bone loss and provide a unique demonstration of the distinct roles of TH17 cells within the same tissue and in the human setting.

A limitation of our current study is that the bulk of data implicating TH17 cells in periodontitis pathogenicity comes from animal experimentation. Although patients with AD-HIES present with reduced periodontitis susceptibility, it is not possible to formally prove resistance to periodontitis in this setting. Future biologic treatment studies of TH17 inhibition are necessary to conclusively implicate TH17 cells in human periodontitis pathogenesis and to evaluate safety and efficacy of such TH17-based treatment in humans. In this regard, our current study not only reveals critical insights into the regulation of TH17 cells at the oral barrier but also provides mechanistic justification for TH17 targeting in human periodontitis.


Study design

This study evaluated mechanisms of TH17 cell accumulation in the disease periodontitis and in experimental periodontitis and assessed preclinical TH17 cell targeting. Oral clinical evaluation and biopsy sampling were performed in healthy volunteers, patients with periodontitis, and AD-HIES patients. Oral examinations included full mouth periodontal examination with a standardized instrument (periodontal probe Hu-Friedy UNC 15 probe). Gingival biopsies were obtained for histology and evaluation of TH17 cells by flow cytometry. All subjects were enrolled in an Institutional Review Board–approved protocol ( #NCT01568697 and #NCT00404560) at the National Institutes of Health (NIH) Clinical Center, and all provided written informed consent. For inclusion in this study, AD-HIES patients were diagnosed on the basis of clinical phenotype and a positive STAT3 mutation. Healthy volunteers were systemically healthy on the basis of medical history and select laboratory testing. Periodontitis patients presented with more than four sites of moderate-severe bone loss and visible signs of inflammation (see the “Inclusion and exclusion criteria” section in supplementary materials and methods).

For experimental periodontitis studies in mice, LIP was performed as previously described (18) in different strains of mice in the presence or absence of ATB treatments, RORγt inhibition, FTY720 treatment, or antibody treatments to block IL-17, Ly6G, or isotype controls. Bone loss was evaluated by morphometric analysis. Investigators were blinded to genotype and treatment during bone loss measurements. Group size for bone loss measurements (n = 5) was determined by power analysis calculations on the basis of previous pilot experiments for P < 0.05 and a power of 0.80. TH17 cells within tissues were evaluated by flow cytometry. Murine oral microbiome was evaluated by 16S rRNA gene sequencing, and transcriptional responses in gingiva in the presence or absence of RORγt inhibition were evaluated by RNA-seq and analysis. Experimental approaches are further detailed in the supplementary materials. Primary data are located in table S2.

Statistical analysis

Flow cytometry, real-time polymerase chain reaction, and bone loss measurement data were evaluated by Shapiro-Wilk normality test. Two-tailed unpaired t test or Mann-Whitney test was used when comparing two groups. When more than two groups were compared, ANOVA or Kruskal-Wallis with post hoc analysis was applied. P values < 0.05 were considered statistically significant. Data were analyzed using Prism 7 program (GraphPad Software). For RNA-seq analysis, gene expression was evaluated by three independent statistical methods (DESeq2, edgeR, and limma-voom). Differentially expressed genes (FDR < 0.05) were considered for further analyses on the basis of results from edgeR. For 16S rRNA gene-sequencing analysis, AMOVA was used to test for differences in community diversity and structure, as implemented in mothur (


Materials and Methods

Fig. S1. Frequencies of IL-17+ cells in human gingiva (related to Fig. 1).

Fig. S2. Ligature-induced periodontitis model (related to Figs. 2 to 6).

Fig. S3. TCR activation in gingiva (related to Fig. 3).

Fig. S4. Effects of ATB treatments (related to Fig. 4).

Fig. S5. Foxp3 and RORγt in Cd4CreStat3fl/fl mice (related to Fig. 5).

Fig. S6. IL-17–secreting cells in LckCreRorcfl/fl mice (related to Fig. 6).

Fig. S7. IL-17–secreting cells during RORγt treatment (related to Fig. 6).

Fig. S8. Neutrophils in LIP and periodontitis (related to Fig. 6).

Fig. S9. IL-17 cell sources in AD-HIES (related to Fig. 7).

Table S1. Primary data.

Reference (58)


Acknowledgments: We thank the NIDCR Veterinary Resource core; the NIDCR Combined Technical Research Core, Genomics core, and Microbiome and Genetics Core (CCR, NCI, NIH); and the NIH HPC Biowulf Cluster. We thank W. Yuan and V. Thovarai for assistance with 16S rRNA sequencing and data deposition and E. Boger for technical expertise. Funding: This work was supported by intramural programs of NIDCR, NIAID, and NCI and NIDCR grants DE024153 and DE026152 to G.H. and La Roche-Posay, CEDEF, Fondation Groupe Pasteur Mutualité, Société Française de Dermatologie, Philippe Foundation, and Fondation pour la Recherche Médicale to C.H. Author contributions: N.D., T.K., L.A., T.G.-W., C.E.Z., T.I., T.A., and C.H. performed experimental work. N.M.M., L.B., and A.F.F. performed clinical work. V.L., G.T., and Y.B. provided key reagents. D.M. and R.J.M. performed RNA-seq and analysis. N.D., P.I.D., S.M.H., Y.B., and N.M.M. were involved in experimental design and data analysis and interpretation and critically edited the manuscript. N.D. wrote the first draft of the manuscript. G.H. was involved in study conception, experimental design and supervision, and data analysis and interpretation and critically edited the manuscript. N.M.M. conceived and supervised the study and wrote the final version of the manuscript. Competing interests: The authors declare that they have no competing interests. Data and materials availability: All data associated with this study are present in the paper or the Supplementary Materials. RNA-seq read data have been deposited in the Gene Expression Omnibus site (GSE118166). The microbiome sequencing read data have been uploaded to SRA (accession number SRP156589) under bioProject code PRJNA484972.
View Abstract

Navigate This Article