Triggering mouth-resident antiviral CD8⁺ T cells potentiates experimental periodontitis

Abstract

Emerging evidence indicates that gingival-resident helper CD4⁺ T cells are major drivers of periodontal inflammation in response to commensal and pathogenic oral microorganisms. Whether tissue-resident memory CD8⁺ T cells (T_RM), which principally safeguard against viruses and cancer but also drive certain autoimmune and inflammatory conditions, impact periodontitis progression and severity remain unknown. We asked whether local reactivation of oral CD8⁺ T_RM of a defined antigen specificity could exacerbate ligature-induced periodontitis (LIP), a well-established model of periodontal disease in mice. Topical application of virus-mimicking peptides to the oral mucosa concurrent with LIP 1) intensified alveolar bone loss, 2) amplified gingival and cervical lymph node inflammation, and 3) stimulated gingival transcriptional changes in genes related to innate immune sensing and cell-mediated cytotoxicity. Therapeutic depletion of CD103-expressing oral CD8⁺ T_RM in advance of LIP prevented exacerbation of disease. These observations provide evidence that oral CD103⁺ CD8⁺ T_RM have the potential to participate in gingival inflammation, alveolar bone loss, and periodontitis.

Introduction

Shifts in the balance between the host immune response and oral microbiota facilitates the outgrowth of virulent oral pathogens and predisposes periodontitis¹, a condition affecting 20–50 % of the global adult population². Immune dysregulation and local pro-inflammatory cues within gingiva perpetuate periodontitis progression despite clinical interventions to reduce the microbial load^1,3,4.

The mouth harbors the second most diverse microbial community in the body, encompassing all three domains of life (Archaea, Bacteria, Eukarya)⁵. In the oral mucosa, the number of prokaryotic cells is estimated to be three-ten times higher than human cells⁶. The emergence of 16S sequencing technologies has enabled the intensive characterization of oral microbial communities in several human health and disease settings including periodontitis. To date, however, identification of a single ‘keystone’ periodontal pathogen remains elusive. The oral mucosa is also constitutively exposed to viruses, which may include persistent infections with herpesviruses⁷, Epstein-Barr virus⁸, human papilloma virus⁹, as well as seasonal exposures to coronaviruses including SARS-CoV-2¹⁰. Collectively, the oral ‘virome’ is estimated to outnumber human cells in the mouth approximately 30–100:1^11,12, yet interactions between the oral virome and local immune cells residing within the oral mucosa remain understudied compared to host/bacterial interactions in the mouth. Additionally, fluctuations in temperature, hydration, pH, and mechanical stress endured by the oral mucosa may permit antigen exposure (and antigen-reactive recall responses) to dietary- and auto-antigens, contributing to oral inflammatory conditions including periodontitis.

Delineation of the complex immunological mechanisms driving initiation and progression of periodontal disease has benefited from small-animal models of the disease, chiefly the ligature-induced periodontitis (LIP) model¹³. Typically performed in specific-pathogen-free (SPF) mice, work in the LIP model has linked ongoing bacterial stimulation of resident helper CD4⁺ T cells with disease onset and severity^14–16. Unlike humans, however, SPF mice are shielded from normal environmental microbial and viral exposures, and thus harbor underdeveloped adaptive immune systems¹⁷. As a result, SPF mice experience a dearth of mucosal-associated resident antiviral CD8⁺ T cells (CD8⁺ T_RM). Viral infection in SPF mice, however, results in the differentiation, redistribution, and permanent establishment of CD8⁺ T_RM in many mucosal tissues where they durably maintain regional antiviral vigilance^18–20. Yet, overexuberant CD8⁺ T_RM-driven immune responses have been implicated in several autoimmune and inflammatory conditions including vitiligo²¹, multiple sclerosis²², psoriasis²³, and colitis²⁴. To date, a mechanistic link between viral reactivation of oral CD8⁺ T_RM and periodontitis progression has not been drawn.

A rigorous characterization of the ontogeny, distribution, and functional implications of CD8⁺ T_RM in the oral mucosa has been mired by a lack of animal models facilitating the generation of abundant antigen-specific memory T cell populations in the mouth to manipulate and study. Furthermore, T_RM are difficult to isolate from many non-lymphoid tissues, especially those that (like the mouth) encompass a diverse collection of individual structures and tissue architectures (i.e., varying degrees of keratinization and epithelial stratification). SPF mice have little pathogen experience, and there remains a paucity of tools to for identifying antigen specific T_RM populations by microscopy or flow cytometric methods. We recently reported a novel ‘Viral-Prime, Epitope-Pull’ (VPEP) model, which manifests a readily identifiable memory CD8⁺ T cell population of known antigen specificity within the oral mucosa of SPF mice²⁵. Here, recombinant virus expressing the model antigen SIINFEKL (SIIN) is used to expand and establish trackable SIIN-specific CD8⁺ T_RM in the mouth. This reductionist approach, which enables the mimicking of viral reinfection events by swabbing the oral surfaces with SIIN peptide, revealed that oral CD8⁺ T_RM reactivation orchestrated rapid, robust, and multifaceted oral inflammatory responses. In the present study, we leveraged the VPEP model in the context of LIP to investigate whether viral antigen triggering of oral CD8⁺ T_RM exacerbates experimental periodontitis. We report that local reactivation of oral CD103⁺ CD8⁺ T_RM synergizes with LIP in a CD4⁺ T cell-independent manner to intensify gingival inflammation and alveolar bone loss. These data provide the first mechanistic evidence linking periodontitis advancement with the triggering of oral antiviral CD8⁺ T_RM responses.

Materials and methods

Animals, adoptive transfers, and viral-prime, epitope-pull (VPEP) model

C57BL/6J (B6) mice were maintained under SPF conditions. CD45.1⁺ OT-I mice²⁶ were fully backcrossed to C57BL/6J mice and maintained at the University of Minnesota or the Cleveland Clinic. CD103^{− /−} mice were generated in-house from P14⁺ CD103^{− /−} mice²⁷. Adoptive transfers and VPEP were performed as previously described²⁵. All mice were used in accordance with the Institutional Animal Care and Use Committees guidelines at the University of Minnesota or the Cleveland Clinic.

Ligature-induced periodontitis (LIP)

Experimental groups were determined based on the frequency of OT-I T cells in blood. For LIP, mice were anesthetized with ketamine/xylazine. Maxillary second molars were ligated circumferentially proximal to the gingiva using a 5–0 silk suture (Roboz Surgical, Gaithersburg, MD) as previously described^13,28 with the aid of a stereo microscope. Ligatures were left in place for three or seven days as indicated.

Microcomputed-tomography (MicroCT) analysis of alveolar bone loss

For bone volume experiments, only the right maxillary second molar was ligated. Seven days post-LIP, maxillae were dissected, fixed in 4 % paraformaldehyde for 48 h, and stored in 0.1 % sodium azide. MicroCT scans were obtained using an XTH225 scanner (Nikon Metrology, Brighton, MI) as described²⁸. The ligated and non-ligated hemispheres of the maxilla were registered three-dimensionally using Data Viewer (Bruker-Micro-CT, Kontich, Belgium). A volume of interest (VOI) was determined for both the ligated (right maxillary) and the non-ligated (left maxillary) molar sites and imported into the CT Analyzer (CTAN; Bruker), which determined the percentage of bone volume (BV) and total volume (TV) in the ligated and the non-ligated VOI. Bone loss was determined as the difference in BV at the non-ligated and ligated VOI.

Intravascular labeling, cell isolation, and flow cytometry

Intravascular labeling was performed as previously described²⁹. Spleens and cLNs were harvested, dissociated, and cells were passed through 70 μm mesh to obtain single-cell suspension. Gingival and tongue tissue were harvested using a dissecting microscope, an 18G needle, and straight fine forceps. The dissection carefully avoided nasal-associated lymphoid tissue (NALT)³⁰. Tissues were finely minced and enzymatically digested in collagenase IV (Sigma-Aldrich) containing 1 μg/ml DNase for 1 h at 37 ^◦C. Digests were further disrupted using a gentleMACS dissociator (Miltenyi Biotec). Tissue homogenates were passed through a 70 μm filter, and leukocytes were enriched on a 44–66 % Percoll gradient. Isolated mouse cells were stained with antibodies against TCRβ (H57–597), CD11b (M1/70), CD8α (53–6.7), CD8β (YTS156.7.7), CD4 (RM4–5), CD90.1 (OX-7), CD90.1 (HIS51), CD45.1 (A20), CD103 (M290), CD69 (H1.2F3), CD44 (IM7), CD62L (MEL-14), Ly6C (1A8), Integrin-β7 (FIB504), CD11c (N418), MHCII (M5–114), and CD86 (GL1). Ghost Dye Viability Dye (Tonbo Biosciences) was used to discriminate live vs. dead cells, and fluorescently conjugated streptavidin was used to detect intravascular biotinylated staining antibodies. Cell counts were determined by adding PKH beads (Sigma) directly to flow cytometry samples at a known concentration. Stained cell samples were acquired on an LSRII or Fortessa flow cytometer (BD) and analyzed with FlowJo software (TreeStar).

Immunofluorescence Microscopy (IF) and Clearing-Enhanced 3-Dimentional imaging (Ce3D)

Harvested tissues were embedded in optimum cutting temperature medium (OCT) and frozen on dry ice. For immunofluorescent microscopy, gingiva was embedded in OCT and 14 μm sections were prepared using a vibratome. Cut sections were mounted and fixed to SuperFrost Plus microscope slides (Fisher) in acetone and subsequently stained with antibodies against CD45.1 (A20), Thy1.1 (OX-7), CD45 (30-F11), CD45.2 (104), and E-Cadherin (DECMA-1). IF microscopy was performed using a Leica DM6000 B microscope. Counterstaining with DAPI or Sytox green was used to detect nuclei. For multiplexed fluorescent (Ce3D) imaging, oral tissues were fixed with BD Cytofix/Cytoperm, diluted 1:4 in PBS overnight at 4 ^◦C, followed by incubation for 8 h in a blocking buffer containing 1 % BSA, 1:100 mouse FC block, and 0.3 % Triton X-100 in PBS at 37 ^◦C. Tissues were then incubated with directly conjugated antibodies diluted 1:10–1:40 in blocking buffer for > 2 d at 37 ^◦C on a shaker. Stained samples were washed with PBS containing 0.2 % Triton X-100 and 1-thioglycerol (0.5 %) for 12–24 h at 24 ^◦C or 37 ^◦C. Tissue was cleared as previously described³¹. Tissue was mounted in fresh Ce3D medium and covered with a glass coverslip. Cleared tissue was imaged by tiling Z-stacks using a Leica Stellaris 8 microscope. A 16X glycerol-immersion coverslip-corrected objective was used to capture images with an XY voxel size of 971 nm, 4 μm step size, and 1-A.U. Pinhole. 3D reconstructions were performed using Imaris (Bitplane) software v9.2.1.

αCD4 monoclonal antibody and SAP-conjugated antibody treatment

αCD4 monoclonal antibody (clone GK1.5; eBioscience, San Diego, CA), was diluted in PBS and administered via intraperitoneal (i.p.) injection. For bone loss experiments, control mice received an equivalent dose of IgG2b antibody. Saporin-conjugated IgG2A and αCD103 (M290 − Advanced Targeting Systems) was diluted in PBS and administered as follows: 5 μg four days prior to LIP, 2 μg concurrent with LIP, 2 μg four days post-LIP.

Gingival tissue RNA-sequencing and analysis

Gingival tissue was isolated and immediately preserved in RNAlater (Thermo Fisher Scientific) at three- and seven-days post-LIP, following the specified treatment conditions. Tissues were homogenized in RLT buffer using a Beadblaster (Benchmark) and Omni Bead Ruptor ceramic bead tubes (1.4 mm), and RNA was extracted using a RNeasy mini kit (QIAGEN). Sequencing was carried out by the University of Minnesota Genomics Center (UMGC), libraries were prepared using the TruSeq Stranded mRNA Library Prep Kit (Illumina), and sequencing was performed on an Illumina NovaSeq 6000 platform, generating 150 bp paired-end reads. The sequencing depth achieved an average of 30 million reads per sample. Raw sequencing reads underwent quality control using fastqc and Trimmomatic and reads were aligned to the mouse genome (mm38) using STAR aligner. For the differential expression analysis, the edgeR implementation within CLC Genomics Workbench (CLCGWB) was employed and downstream analysis of bulk RNA-seq data was conducted using R software (R version 4.1.2 (2021–11–01). Differentially expressed genes with a fold change > 2 (false discovery rate, FDR < 0.05) were selected for GO enrichment analysis³². GO terms were categorized based on the Biological Process (BP) ontology. Enriched GO terms were identified using the hypergeometric test, and multiple testing correction (Benjamini-Hochberg method) was applied to control the false discovery rate. Enriched GO terms were visualized using the ggplot2 package. We used the ‘compareCluster’ function in the ‘clusterProfiler’ package to compare the enriched GO terms between different treatment conditions. This allowed us to identify common and condition-specific enriched pathways.

In vitro peptide stimulation and intracellular staining

For in vitro peptide stimulations, cells were isolated from the oral mucosa or cLNs and cultured for 5 h at 37 ^◦C in RPMI 1640 supplemented with 10 % FBS containing 1X Golgi-plug (BD) ± SIIN peptide (MedChemExpress) at a final concentration of 10 μg/ml. Isolated lymphocytes were fixed and permeabilized using a Cytofix/Cytoperm fixation/permeabilization kit (BD) and cells were labeled with intracellular antibodies against Ki67 (16A8), IFNγ (XMG1.2), and TNFα (TN3–19.12).

Statistical analysis

Unpaired and paired Student’s t tests were performed between the relevant comparisons. P < 0.05 was considered significant in all experiments. Statistical analysis was done in Prism (GraphPad Software). Sample size was chosen based on previous experience. No sample exclusion criteria were applied, and investigators were not blinded.

Results

Oral CD8⁺ T_RM reactivation exacerbates experimental periodontitis

The oral mucosa is continually exposed to intracellular pathogens including viruses. Whether antiviral CD8⁺ T_RM responses in the mouth can appreciably impact periodontal inflammation remains untested. We asked whether oral CD8⁺ T_RM reactivation in SPF mice concurrent with the initiation of the LIP model would aggravate disease. SIIN-specific TCR-transgenic OT-I T cells were adoptively transferred and expanded through systemic infection with vesicular stomatitis virus expressing SIIN (VSV-ova). Five-six days later, OT-I T cells were ‘pulled’ into the oral mucosa by swabbing the oral surfaces of anesthetized mice with SIIN peptide dissolved in chemical and mechanical irritants (nonoxynol-9 and dental pumice, respectively). Such mice are referred to as ‘VPEP’ mice (Viral-Prime, Epitope-Pull)²⁵. At least 30 days after VPEP, mice were orally swabbed with either an irrelevant gp33–41 peptide (gp33; derived from lymphocytic choriomeningitis virus; Armstrong strain) or SIIN (OT-I T cell reactivating peptide) in irritants. Immediately following, ligatures were placed around the maxillary second right molar, leaving the contralateral second molar non-ligated to serve as an internal negative control. Thus, two groups were established: gp33/LIP mice (control) and SIIN/LIP mice (experimental). Four days later, mice were re-exposed to gp33 or SIIN peptide in irritants. Gingival T cell infiltrate and alveolar bone loss was assessed three days later (Fig. 1A). SIIN peptide exposure, but not gp33, stimulated robust accumulation of OT-I T cells within ligated gingiva, whereas CD4⁺ T cells and neutrophils (PMNs) were not enhanced in gingiva of SIIN/LIP mice compared to gp33/LIP mice (Fig. 1B). We performed micro-computed tomography (microCT) to quantify LIP-associated bone loss, calculated as the change in bone volume between the non-ligated (reference) and ligated (target) molar (Fig. 1C). Bone loss was substantially elevated in SIIN/LIP mice compared to gp33/LIP mice. Iso-surface reconstructions revealed only small fenestration defects on the cortical bone of gp33/LIP mice with no– or partial-exposure of the second molar distal root, whereas fenestration defects on SIIN/LIP mice were substantially larger, exposing both the proximal and distal second molar roots (Fig. 1D). When normalizing bone loss to control (gp33/LIP) mice from two independent experiments, SIIN swabbing intensified LIP-associated bone loss approximately 25 % (Fig. 1E). We next investigated whether oral CD8⁺ T_RM reactivation concurrent with LIP would impact antigen trafficking and presentation within downstream cervical LNs (cLNs). Three days post-swabbing and LIP, cLNs were isolated and processed for flow cytometric analysis (Fig. 1F). Proliferating OT-I T cells (Ki67⁺) were significantly enriched in cLNs of SIIN/LIP mice (Fig. 1G). Moreover, conventional DCs (MHCII⁺ CD11c⁺) and monocyte-derived DCs (MHCII^int CD11c^int) were more numerous in cLNs of SIIN/LIP mice compared to gp33/LIP mice. These differences were not a direct consequence of increased cLN hyperplasia, as total LN cellularity was identical between gp33/LIP and SIIN/LIP groups. We next used Clearing-Enhanced 3-Dimentional imaging (Ce3D; day three post-LIP) and conventional immunofluorescent microscopy on decalcified maxillae (IF; day seven post-LIP) to visualize cellular influx into LIP gingiva histologically. Ce3D revealed a perceptible increase in CD45⁺ cells (pan-leukocyte marker) in both buccal and palatal gingiva of SIIN/LIP mice compared to gp33/LIP controls (Fig. 1I). OT-I T cell infiltration was pronounced within the interstitial gingiva spanning the second and third molar (white arrows) and within adjacent molar dental pulp (yellow arrows) in SIIN/LIP but not gp33/LIP mice (Fig. 1J). Collectively, these data demonstrate that oral CD8⁺ T_RM reactivation intensifies alveolar bone loss, gingival inflammation, and antigen transport/presentation within oral mucosa draining LNs.

Fig. 1. Oral CD8⁺ T_RM reactivation aggravates experimental periodontitis.

(A) VPEP model and experimental design. (B) Enumeration of gingival OT-I cells, CD4⁺ T cells, and PMNs seven days post-LIP in mice orally swabbed with gp33 or SIIN peptide. Data representative of two independent experiments with four-six mice per group per experiment. (C) Quantification of total bone within predefined coordinates surrounding the non-ligated second molar (Ref.; Left) and ligated second molar (Target; Middle), and their difference (Change; Right). (D) Representative maxilla iso-surfaces from LIP mice exposed to gp33 or SIIN peptide. Shaded regions highlight superficial differences between groups. (E) Alveolar bone loss, normalized to gp33/LIP mice from each of two independent experiments with two-four mice per group per experiment. Normal. = Normalized. (F) Experimental strategy to access the impact of oral CD8⁺ T_RM reactivation on cLN cellularity and gingival inflammatory infiltrate. (G) Abundance of Ki67⁺ OT-I T cells within cLNs of day three LIP mice exposed to either gp33 or SIIN peptide. White circles represent % Ki67⁺ OT-I T cells within cLNs of non-ligated VPEP mice (No LIP). (H) Enumeration of total LN cellularity and individual DC subsets within cLNs of day three LIP mice exposed to gp33 (grey) or SIIN (blue) peptide. White bars represent DCs of the indicated subset within cLNs of non-ligated VPEP mice. Data in G and H representative of two independent experiments with four-five mice per group per experiment. (I) Ce3D imaging of day three LIP gingiva from mice of the indicated treatment group, stained with antibodies against E-Cadherin (teal) or pan-CD45 (red). ‘1′ indicates location of first molar. ‘B’ and ‘P’ denote buccal and palatal orientation, respectively. Insets magnify regions of interest. Data representative of three-four mice per group. Scale bar represents 1 mm. (J) Immunofluorescence (IF) microscopy of interstitial gingiva separating the second and third molar of non-ligated VPEP mice (Left), gp33/LIP mice (Middle), and SIIN/LIP mice (Right) seven days post ligation. White and yellow arrowheads indicate OT-I T cells in gingiva and adjacent molar dental pulp, respectively. Representative images are from at least three decalcified maxillae sections, per mouse, of three or more individual mice per group. I.P. = interproximal; 2nd = second molar; 3rd = third molar; NS = non-specific. Scale bar represents 50 μm. Error bars in all graphs represent mean ± SEM. Dots in C, E, and G represent individual mice. *, P < 0.05; **, P < 0.01; ****, P < 0.0001 as determined by an unpaired Student’s t test between the relevant comparisons.

Oral CD8⁺ T_RM reactivation drives CD4⁺ T cell-independent alveolar bone loss

CD4⁺ T cells are considered major drivers of periodontal inflammation. Exacting studies employing the LIP model showed that, in response to a LIP-associated dysbiotic microbiota, gingival-resident ‘T helper 17′ CD4⁺ T cells (T_H17) and PMNs are critical for alveolar bone loss¹⁵. However, these experiments were performed in SPF mice lacking oral viruses and a corresponding oral CD8⁺ T_RM compartment. Fig. 1 indicated that CD8⁺ T_RM also have the potential to exacerbate experimental periodontitis. We therefore tested whether oral CD8⁺ T_RM reactivation could intensify LIP-associated alveolar bone loss independently of CD4⁺ T cells. We first confirmed that a single 200 μg administration of GK1.5 antibody (a CD4⁺ T cell depleting clone³³) efficiently eliminated CD4⁺ T cells systemically and in gingiva two days later (Fig. 2A). Next, VPEP mice were administered GK1.5 or control IgG2b antibody in advance of oral peptide swabbing and LIP. Four-days later, mice received an additional round of antibody and oral swabbing with either gp33 or SIIN peptide. Cellular profiling and bone volume was performed three days later (Fig. 2B). CD4⁺ T cell depletion was confirmed in spleen (Fig. 2C) and oral mucosa (tongue analyzed as a surrogate for gingiva to avoid damaging alveolar bone during tissue isolation) (Fig. 2D). Tongues from CD4⁺ T cell depleted SIIN/LIP mice showed a significant increase in OT-I T cells compared to gp33/LIP mice demonstrating oral CD8⁺ T_RM reactivation had been achieved (Fig. 2E). We next quantified DCs within draining cLNs where we observed an enrichment in monocyte-derived DCs in SIIN-swabbed, CD4⁺ T cell-depleted mice. Conventional DCs were also subtly increased in SIIN-swabbed cLNs, albeit not significantly (Fig. 2F & G). Contrary to predictions, alveolar bone loss following LIP was agnostic to the presence or absence of CD4⁺ T cells. (Fig. 2H). Instead, when normalizing bone loss to control animals (IgG2b/gp33/LIP) from two independent experiments, SIIN swabbing and LIP in CD4⁺ T cell-depleted mice resulted in approximately 22 % greater bone loss compared to gp33-swabbed mice (Fig. 2H-J). In line with Fig. 1D, both second molar buccal roots were routinely exposed following SIIN swabbing and LIP in CD4⁺ T cell depleted mice, whereas control animals (± CD4⁺ T cells) displayed no– or only partial-exposure of the second molar buccal roots (Fig. 2I). These data are consistent with the interpretation that CD8⁺ T cells can perpetuate gingival inflammation and alveolar bone loss through CD4⁺ T cell-independent mechanisms if restimulating antigen is present.

Fig. 2. CD8⁺ T_RM driven exacerbation of LIP is CD4⁺ T cell independent.

(A) CD4⁺ T cell frequencies (Left) and absolute numbers (Middle) within spleen and gingiva of naïve C57BL/6J mice treated two days earlier with 200 μg of GK1.5 antibody (green circles), and representative flow cytometry (Right). Red numbers indicate average fold decrease comparing No Tx (no treatment) and GK1.5 treated mice. N = Four mice per group. (B) Experimental design. Representative flow cytometry plots (Left) and enumeration of CD4⁺ T cells (Right) in (C) spleens and (D) tongues of gp33/LIP or SIIN/LIP mice ± GK1.5 antibody treatment. (E) Quantification of OT-I T cells within tongues of mice from the indicated treatment groups seven days post-LIP. (F) Representative flow cytometry plots showing DC subsets within cLNs from mice treated as indicated. (G) Quantification of DC subsets within day seven LIP cLNs from mice of the indicated treatment groups, as gated in E. Data in D-G represents three–five mice per group. (H) Quantification of total bone within predefined coordinates surrounding the non-ligated second molar (Ref.; Left) and ligated second molar (Target; Middle), and their difference (Change; Right). (I) Representative maxilla iso-surfaces from gp33/LIP (Left) and SIIN/LIP (Right) mice treated with GK1.5 antibody. Shaded regions highlight superficial differences between groups. (J) Alveolar bone loss, normalized to gp33/LIP mice treated with isotype-control antibody from two independent experiments with two-five mice per group per experiment. Scale bars in all graphs represent mean ± SEM. Dots in A, C, D, E, H, and J represent individual mice. *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001 as determined by an unpaired Student’s t test between the relevant comparisons.

Oral CD8⁺ T_RM reactivation invokes a hyperinflammatory transcriptional responses in periodontitis

We next took a bulk-RNAseq approach to identify gene expression changes in LIP-associated gingiva driven by oral CD8⁺ T_RM reactivation. VPEP mice were orally swabbed with either gp33 or SIIN peptide in irritants, followed immediately by placement of ligatures around both maxillary second molars (to increase the amount of tissue per mouse for RNA isolation). One group of mice were sacrificed three days later, while the second group of mice received an additional round of swabbing with gp33 or SIIN peptide four days post-ligation and were sacrificed three days later (day seven post-ligation). In both cases, mice were sacrificed three days following their last oral peptide swabbing. Gingiva from non-ligated VPEP mice (which received oral SIIN swabbing > 24 days earlier during the VPEP process) provided a transcriptional baseline (No Treatment; No Tx). This experimental design allowed us to evaluate transcriptional changes induced longitudinally both by 1) LIP alone (e. g., gp33/LIP), and 2) LIP in the context of oral CD8⁺ T_RM reactivation (e. g., SIIN/LIP) (Fig. 3A). We focused on differentially upregulated genes (DUGs). Compared to non-ligated gingiva (e.g., No Tx), ligature placement induced significant upregulation (Fold change ≥ 2, FDR < 0.05) of 1142 genes in gp33/LIP gingiva compared to 1344 genes in SIIN/LIP gingiva at day three. At day seven, 1848 genes were upregulated in gp33/LIP gingiva compared to 2160 genes in SIIN/LIP gingiva. Most DUGs were shared between gp33/LIP and SIIN/LIP mice (985 or 73 % at day three, 1692 or 78 % at day seven) indicating ligature placement alone was the main driver of transcriptional changes regardless of timepoint or treatment group (Fig. 3B). Nevertheless, amongst shared genes, oral CD8⁺ T_RM reactivation concurrent with LIP drove expression an additional 16 % on average (Fig. 3C). This number is congruent with the approximate 22–25 % increase in bone loss observed in SIIN/LIP mice (Fig. 1E & Fig. 2I) and suggests that the amplification of LIP-associated genes downstream of oral CD8⁺ T_RM reactivation may contribute to exacerbated disease. For example, the top ten DUGs in SIIN/LIP gingiva at both day three and day seven timepoints were trending upwards or significantly increased when compared to gp33/LIP gingiva at the same timepoint. Amongst these genes were Il6, a major periodontitis gene and the sixth and fourth most prevalent transcript at day three and seven, respectively (Fig. 3D). Differences in genes encoding canonical CD8⁺ T_RM effector molecules (Ifng), perforin (Prf1), and granzymes A and B (Gzma, Gzmb) were even more pronounced in SIIN/LIP mice, especially at day seven (Fig. 3E). T cell activation (Cd69) and CD8⁺ T cell-mediated programmed cell death genes (Fasl) were also induced (Fig. 3F; Left & Middle). Consistent with histological data in Fig. 1I, CD45 transcripts (Ptprc) were elevated in SIIN/LIP mice (Fig. 3F; Right). Several Cxcl and Ccl chemokine genes, chemokine receptor genes (c[x]cr) genes, matrix metalloproteinase (Mmp) genes, toll-like receptor (Tlr) genes, and Fcγ-receptor (Fcgr) genes behaved similarly (e.g., progressively upregulated in SIIN/LIP gingiva over time) (Fig. 3G). Surprisingly, genes implicating cells and/or pathways known to be involved in LIP progression including CD4 (Cd4), IL-17A and F (Il17a, Il17f), Ly6G (Ly6g), TRAIL (Tnfsf10), and RANKL (Tnfsf11) were identical or only modestly increased in SIIN/LIP mice compared to gp33/LIP mice; many (Il17a, Il17f, TNFSF10) only induced after day three of LIP (Fig. 3H). These data support a model whereby oral CD8⁺ T_RM reactivation intensifies experimental periodontitis through noncanonical mechanisms (e.g., T_H17/PMN mediated bone loss)¹⁵.

Fig. 3. Antigen restimulation of oral CD8⁺ T_RM amplifies gingival transcriptional responses in LIP.

(A) Experimental design to assess transcriptional changes in LIP gingiva driven by oral CD8⁺ T_RM reactivation. (B) Venn-diagram summarizing the number of DUGs identified in gp33/LIP vs. SIIN/LIP mice three- and seven-days post-LIP, compared to non-ligated VPEP mice. Red arrows highlight shared genes. (C) Log2 mean fold change in genes shared between gp33/LIP and SIIN/LIP mice at days three and seven (Left). Yellow shade highlights significance cutoff. (Right) Average % increase in shared genes between gp33/LIP and SIIN/LIP mice at the indicated timepoint. Expression values for shared genes in gp33/LIP mice were set at 100 %. Percent increase in shared SIIN/LIP genes calculated as 100 % + [(SIIN/LIP value − gp33/LIP value) ÷ gp33/LIP value × 100]. (D) Average expression of the top 10 DUGs in SIIN/LIP mice (blue circles) at day three and day seven, and their values in gp33/LIP mice (grey circles). Red line highlights Il6 expression. (E) Fold change in the indicated T cell effector genes comparing gp33/LIP vs. SIIN/LIP mice three (open circles) and seven days (closed circles) post LIP, normalized to No Tx controls. (F) Fold change in Cd69, Fasl, and Ptprc comparing gp33/LIP vs. SIIN/LIP mice three (open circles) and seven days (closed circles) post LIP, normalized to No Tx controls. (G) Heatmaps showing Log2 fold expression changes for the indicated genes in gp33/LIP vs. SIIN/LIP mice at three- and seven-days post-LIP, normalized to No Tx controls. (H) Fold change in the indicated periodontitis-associated genes comparing gp33/LIP vs. SIIN/LIP mice three (open circles) and seven (closed circles) days post LIP, normalized to No Tx controls. Group numbers are as follows: No Tx; N = 5, gp33/LIP; N = 4 per timepoint, SIIN/LIP; N = 4 per timepoint. Error bars in all graphs represent mean ± SEM. Dots in E, F, and H represent individual mice. Hashed line in E, F, and H denotes a value of 1. *, P < 0.05; **, P < 0.01; ****, P < 0.0001 as determined by a paired (C, D) or unpaired (E, F, H) Student’s t test between the relevant comparisons.

We next focused on uniquely upregulated genes in gingival tissue from gp33/LIP vs. SIIN/LIP mice at days three and seven days post-ligation. Gene ontogeny (GO) enrichment analysis was performed on the 157 and 156 unique DUGs in gp33/LIP mice (day three and seven, respectively) and on the 359 and 468 unique DUGs in SIIN/LIP mice (day three and seven, respectively) (Fig. 4A). Only one defined pathway, ‘double-strand break repair’, was identified in gp33/LIP mice (day seven only). In contrast, 20 unique and non-overlapping pathways were identified in SIIN/LIP mice at each of the two timepoints analyzed. At day three, most pathways in SIIN/LIP mice involved cell activation, proliferation, and cell–cell adhesion (Fig. 4A), whereas cell killing dominated day seven (Fig. 4B). Thus, oral CD8⁺ T_RM reactivation triggers two distinct waves of oral inflammation following LIP; the first recruiting and activating cells within gingiva while the second executing cell-mediated cytotoxicity. We surmised that unique DUGs in SIIN/LIP gingiva consistently expressed at both day three and day seven post-LIP may be of particular interest. Removing non-annotated and pseudogenes, we identified 31 upregulated genes in SIIN/LIP mice at both day three and day seven timepoints. Within the top 15 of these 31 unique SIIN/LIP genes were terminal deoxynucleotidyl transferase (Dntt), Il5ra, H2-Q7, and β−2-microglobulin (B2m) − a component of the MHC-class I molecule critical for antigen presentation to CD8⁺ T cells (Fig. 4D). Expression of several relevant genes may have been even more robust at timepoints not analyzed in this study. For instance, Ifng was shown to be upregulated ~ 50 fold in buccal mucosa within 12 h of oral CD8⁺ T_RM reactivation²⁵, yet we found Ifng transcript only modestly increased at day three in SIIN/LIP compared to gp33/LIP mice. We speculate that many genes, including Ifng, may have reached peak expression within the first 24 h following SIIN swabbing, and that expression values observed at days three and seven post-LIP represent a return to baseline³⁴. Indeed, oral CD8⁺ T_RM produced abundant IFNγ (and TNFα) within a five-hour window following ex vivo peptide restimulation (Supp. Fig. 1A & B). Using Ifng as one example, several scenarios were hypothesized for gene expression kinetics (Supp. Fig. 1C). Collectively, our transcriptional data supports three main conclusions; 1) oral CD8⁺ T_RM reactivation magnifies LIP-driven gene expression changes in diseased gingiva, 2) CD8⁺ T cell effector genes, rather than canonical periodontitis-associated genes, are uniquely invoked upon oral CD8⁺ T_RM reactivation and LIP, and 3) over the course of the disease model, oral CD8⁺ T_RM reactivation induced at least 827 unique DUGs which binned into dozens of defined inflammatory pathways.

Fig. 4. Unique genes and transcriptional pathways driven by oral CD8⁺ T_RM reactivation and LIP.

(A) Venn-diagram highlighting the number of unique DUGs in gingiva of gp33/LIP and SIIN/LIP mice at days three and seven, compared to No Tx. Red arrows highlight unique genes. Gene ontogeny enrichment analysis of genes uniquely expressed in gp33/LIP or SIIN/LIP gingiva at day 3 (B) and day 7 (C) post-ligature placement. GO terms were categorized using the ‘Biological Process’ ontology. (D) Heatmap showing Log2 fold expression changes in the top 15/31 DUGs shared at both day three and seven in gingiva of SIIN/LIP mice. Group numbers are as follows: No Tx; N = 5, gp33/LIP; N = 4 per timepoint, SIIN/LIP; N = 4 per timepoint.

Therapeutic elimination of oral CD103⁺ CD8⁺ T_RM abrogates experimental periodontitis

CD103 is a defining marker of tissue-residency amongst CD8⁺ T cells in the oral mucosa²⁵. Recirculating memory CD8⁺ T cells lack CD103 expression. Reactivation of oral CD8⁺ T_RM aggravated alveolar bone loss following LIP approximately 22–25 % (Fig. 1E & Fig. 2J), however it remained unclear whether this phenomenon was bolstered by activation and/or recruitment of recirculating CD8⁺ T cells into LIP gingiva. Thus, we wished to disentangle the contribution of bona fide oral CD8⁺ T_RM from recirculating memory CD8⁺ T cells to LIP-associated bone loss. We strategized targeting CD103⁺ cells for antibody-mediated depletion. However, many immune cells (in addition to CD8⁺ T_RM) express CD103 including naïve T cells, regulatory T cells, NK cells, and DCs. To obviate off-target effects, OT-I T cells were transferred and expanded through VPEP in CD103-deficient (CD103^KO) host mice, making OT-I T cells the only cell population capable of expressing CD103 upon oral CD8⁺ T_RM establishment and thus susceptible to antibody-mediated depletion with Saporin-toxin (SAP) conjugated αCD103 antibodies (SAP-103). VPEP mice received prophylactic treatment with SAP-103 or SAP-conjugated isotype control antibodies (SAP-IgG). Ligatures were placed around the maxillary second molar concurrent with gp33 or SIIN oral peptide swabbing, leaving the contralateral second molar non-ligated to serve as an internal reference control. Four-days later, mice received an additional round of oral swabbing and antibody treatment. Gingival cellular profiling and bone volume analysis was performed three days later (day seven post-LIP). Three experimental groups were thus established: those receiving gp33 (control) swabbing and LIP in the context of an intact oral CD103⁺ CD8⁺ T_RM compartment (IgG/gp33), those receiving SIIN swabbing (OT-I T cell reactivating) and LIP in the context of an intact oral CD103⁺ CD8⁺ T_RM compartment (IgG/SIIN), and those receiving SIIN swabbing (OT-I T cell reactivating) and LIP in the context of a depleted oral CD103⁺ CD8⁺ T_RM compartment (103/SIIN) (Fig. 5A). SIIN treatment in SAP-IgG treated mice increased gingival CD103⁺ CD8⁺ T_RM approximately five-fold, whereas SAP-103 treatment effectively eliminated CD103⁺ CD8⁺ T_RM in SIIN-exposed LIP gingiva (Fig. 5B; Left). In agreement with Fig. 1B, CD4⁺ T cell and PMN abundance were not significantly impacted by SIIN swabbing (Fig. 5B; Right). Based on MicroCT analysis, SIIN swabbing and LIP in mice with an intact CD103⁺ CD8⁺ T_RM compartment (IgG/SIIN) exacerbated disease approximately 25 %, whereas this response was abolished in SIIN swabbed mice prophylactically depleted of CD103⁺ CD8⁺ T_RM (103/SIIN) (Fig. 5C-E). These data demonstrate that oral CD103⁺ CD8⁺ T_RM, rather than their CD103^neg recirculating counterparts, catalyze alveolar bone loss in experimental periodontitis when restimulating antigen is present.

Fig. 5. CD103⁺ CD8⁺ T_RM catalyze LIP-associated alveolar bone loss in the presence of restimulating antigen.

(A) Experimental design to assess the contribution of oral CD103⁺ CD8⁺ T_RM from CD103^neg recirculating CD8⁺ T cells on aggravated bone loss following LIP. Experiments were performed in CD103^KO host mice. (B) Fold change in the number of CD103⁺ OT-I T cells, CD4⁺ T cells, and PMNs isolated from LIP gingiva from mice of the indicated treatment group, normalized to their abundance in IgG/gp33 treated mice. Data representative of two independent experiments with three-four mice per group per experiment. (C) Quantification of total bone within predefined coordinates surrounding the non-ligated second molar (Ref.; Left) and ligated second molar (Target; Middle), and their differences (Change; Right). (D) Representative maxilla iso-surfaces from LIP mice exposed to SIIN peptide with an intact (Left; blue) or depleted (Right; green) oral CD103⁺ CD8⁺ T_RM compartment. Shaded regions highlight superficial differences between groups. (E) Alveolar bone loss, normalized to IgG/gp33 mice, from each of two independent experiments with three-five mice per group per experiment. Error bars in all graphs represent mean ± SEM. Dots in C and E represent individual mice. **, P < 0.01; ***, P < 0.001 as determined by an unpaired Student’s t test between the relevant comparisons.

Discussion

The impact of oral viruses on the progression and severity of periodontitis remains an open question. The aim of this study was to evaluate whether reactivation of oral antiviral CD8⁺ T_RM could affect experimental periodontitis. Periodontitis encompasses a collection of inflammatory disorders that compromise the integrity of dental supporting structures (gingiva, cementum, periodontal ligament, and alveolar bone), manifesting a broad range of symptoms including unprovoked bleeding of the gums, halitosis, extreme temperature sensitivity, mobile teeth, pain during mastication, and ultimately tooth loss³⁵. Eventually, periodontal disease can significantly impact quality of life and cause serious social and economic burden³⁶. Chronic inflammation associated with periodontitis can also aggravate autoimmune or inflammatory co-morbidities including diabetes mellitus, arthritis, atherosclerosis, and certain neurodegenerative diseases³⁷. Thus, a further elucidation of the mechanisms driving periodontitis onset and progression may have far-reaching implications for both local and systemic health.

Periodontitis appears to be sustained by persistent immune responses directed against subgingival dental plaque microorganisms, a diverse community of approximately 700 distinct species^1,38,39. A decades-long debate about which species contribute to virulence remains unresolved. Current thought is that no single ‘keystone’ pathogen is causative, but the etiology involves oral dysbiosis-driven inflammation^39–41. Community diversity is lost, 1) enabling the outgrowth of inflammation-tolerant species, 2) altering community behavior and metabolic output, and 3) increasing virulence. Accordingly, efforts to prevent or mitigate periodontitis have historically focused on reducing microbial load and virulence through improved oral hygiene, antibiotics, and chemical inhibitors of microbial growth. Yet, viruses routinely afflict the human mouth, and it has been suggested that host inflammatory responses to oral viruses may contribute to periodontal disease progression and/or ‘flare-ups’ occurring under conditions of immunosuppression or prolonged stress^42–45. Local immune responses against viruses predominantly involve CD8⁺ T_RM, and CD8⁺ T cells bearing a residency phenotype (CD103 and/or CD69⁺) are enriched in human periodontitis relative to healthy adjacent tissue⁴⁶. Patients with progressive periodontitis were shown to harbor oral cytomegalovirus (CMV)⁴⁷, Epstein Barr virus (EBV)⁸, and herpes simplex virus 1 (HSV-1)⁷ infections, supporting the concept that host immune responses to viruses in the oral mucosa may augment disease. Moreover, antiviral therapy (valacyclovir, acyclovir) was shown to mitigate gingival inflammation and other clinical parameters of periodontitis severity^48–50, providing further evidence in humans that oral viruses are likely consequential. In one particularly illustrative case report, a patient with genetic predisposition to mucosal herpes infections (due to a primary immunodeficiency) demonstrated reduced periodontal inflammation and stabilized periodontal disease in response to antiviral treatment⁵¹.

Seminal mechanistic periodontology studies leveraging in vivo SPF mouse models have unequivocally established the role of oral biofilm-driven CD4⁺ T cell responses in disease^14–16. Investigating the impact of CD8⁺ T cells on periodontitis has been stymied by a lack animal models and tools for stimulating antigen-specific CD8⁺ T cell recall responses in the oral mucosa of SPF mice. The recently reported VPEP model provided an experimental platform for addressing the combined pathogenic potential of both gingival CD4⁺ T cells and oral CD8⁺ T_RM in LIP²⁵. We hypothesized that synergy between dysbiotic biofilm-driven CD4⁺ T cell responses and synchronized antiviral CD8⁺ T_RM triggering would intensify gingival inflammation and alveolar bone loss. Not only was this the case, but we found that oral CD8⁺ T_RM could autonomously amplify disease (i.e., in the absence of CD4⁺ T cells) if virus-mimicking peptides were present. One limitation of this study is that the GK1.5 antibody indiscriminately eliminates both conventional CD4⁺ T cells and immunosuppressive regulatory T cell (T_reg). Previous work suggests that depletion of T_H17 cells, in a setting replete with T_reg cells, would likely have attenuated LIP¹⁵. Additionally, our model system is based on high-affinity TCR interactions which promote robust inflammatory responses in the oral mucosa during LIP. This system likely overrides the nuanced development of medium and low-affinity T cell responses occurring in human periodontitis, which may provide a more regulated microenvironment. Nevertheless, our data places new emphasis on the potential role of CD8⁺ T_RM in periodontitis by unshrouding their capacity for amplifying localized inflammatory responses in diseased gingiva.

Our experimental approach for the targeted elimination of CD103⁺ CD8⁺ oral T_RM revealed that this subset was paramount for exacerbation of LIP in the presence of T cell restimulating antigens. Oral CD8⁺ T_RM reactivation also increased the abundance of DC subsets within oral mucosa-draining cLNs and boosted transcriptional responses within LIP gingiva ~ 16 %. While the magnitude of transcriptional responses has not been previously correlated with bone loss, the bolstering of LIP-driven genes within diseased gingiva represents a potential mechanism of exacerbated periodontitis downstream of oral CD8⁺ T_RM reactivation. Also of note, CD4⁺ T cells and PMNs were not enhanced in gingiva of SIIN/LIP mice, nor were genes implicating T_H17/PMN crosstalk. Collectively, these data suggest that the triggering of oral antiviral T cells may pivot the inflammatory response in periodontitis rather than amplifying existing pathways. While not assessed in the present study, priming of de novo T cell responses against oral microbiota-derived antigens was likely enhanced downstream of oral CD8⁺ T_RM reactivation. This scenario may enable a feed-forward loop whereby oral CD8⁺ T_RM reactivation indirectly enhances sensitivity to oral microorganisms, enabling the outgrowth of virulent (and inflammation tolerant) subsets at the expense of commensals. Indeed, the transcriptional pathways ‘response to protozoan’, ‘defense response to protozoan’, ‘defense response to bacterium’, and ‘positive regulation of response to biotic stimulus’ were noted exclusively in gingiva of SIIN/LIP mice. Follow-up studies will define the impact of oral CD8⁺ T_RM reactivation on the gingival microbiome. Novel gene expression signatures in gingiva not typically associated with periodontitis were also invoked upon oral CD8⁺ T_RM reactivation and LIP, namely innate immune sensing and cytotoxicity genes. While we are actively pursuing the role of CD8⁺ T_RM-derived effector molecules on periodontitis phenotypes, these signatures may provide a reference point for identifying conditions in human periodontitis where viral triggering of resident CD8⁺ T cells are contributory. In such situations, adjunctive antiviral therapies may diminish localized oral antiviral CD8⁺ T_RM responses and improve patient outcomes. The anatomic characteristics of the oral cavity provide easy access for the delivery of local immunomodulatory treatments, including those which may target CD8⁺ T_RM. Future work will evaluate the prevalence, phenotype, specificity, and anatomical positioning of oral CD8⁺ T_RM within healthy and diseased human gingiva biopsies, which may inform novel therapeutic strategies for their regionalized manipulation.

Supplementary Material

Appendix A. Supplementary material

Supplementary data to this article can be found online at https://doi.org/10.1016/j.mucimm.2025.02.003.

Abbreviations:

cLNs

Cervical lymph nodes

SAP

Saporin toxin

VPEP

Viral-Prime, Epitope-Pull

LIP

Ligature-induced periodontitis

T_RM

Tissue resident memory

SIIN

SIINFEKL from chicken OVA

gp33

gp33 peptide