MHC-II Expression on Oral Langerhans Cells Differentially Regulates Mucosal CD4 and CD8 T Cells

Abstract

In murine periodontitis, the Th17 response against Porphyromonas gingivalis in CLN is abrogated by diphtheria toxin-driven depletion of Langerhans cells (LCs). We determined the impact of MHC-II presentation in LCs on Th17 cells in the oral mucosa of mice. Using an established human-Langerin promoter-Cre mouse model we generated LC-specific deletion of the H2-Ab1 (MHC-II) gene. MHC-II expression was ablated in 81.2% of oral-resident LCs compared to >99% of skin-resident LCs. MHC-II (LC^ΔMHC-II) depletion did not reduce the number of CD4 T cells nor the frequency of Th17 cells when compared to wild-type mice. However, the frequencies of Th1 cells decreased and Helios⁺ Treg cells increased. In ligature-induced periodontitis, the number of CD4 T cells and Th17 cells were similar in LC^ΔMHC-II and wild-type mice. Normal numbers of Th17 cells can therefore be sustained by as little as 18.8% of MHC-II expressing LCs in oral mucosa. Unexpectedly, oral mucosa CD8 T cells increased >25-fold in LC^ΔMHC-II mice. Hence, these residual MHC-II expressing LCs appear unable to suppress the local expansion of CD8 T cells while sufficient to sustain a homeostatic CD4 T cell response. Reducing expression of MHC-II on specific LC subpopulations may ultimately boost CD8-mediated intraepithelial surveillance at mucosal surfaces.

Introduction

Langerhans cells (LCs) are a population of antigen-presenting cells residing in the stratified epithelia of the skin (Kaplan, 2017), oral mucosa (Capucha et al., 2015), cervical-vaginal mucosa (Iijima et al., 2008), and the ocular surface (Gillette et al., 1982). All LCs express the C-type lectin receptor langerin (CD207), which captures antigen to accumulate in Birbeck granules (Capucha et al., 2015, Romani et al., 2010). LCs also express major histocompatibility complex class II molecules (MHC-II) on their surface, presenting antigens to MHC-II-restricted T cells in the skin and the stratified-squamous mucosal barrier (Capucha et al., 2015). In the uninflamed skin or buccal epithelium, LCs are the only MHC-II expressing cells, whereas in the gingival epithelium LCs represent approximately 80% of the MHC-II⁺ population (Capucha et al., 2015, Kaplan, 2017). Unlike the oral mucosa, the skin becomes impermeable to bone marrow-derived LC precursors at embryonic day 16. In contrast, the oral mucosa is more permissive to continuous recruitment of LCs precursors throughout life (Capucha et al., 2018, Capucha et al., 2015).

Localized within the surface epithelial barrier, LCs are uniquely positioned to be among the first cells to encounter antigen. LCs, therefore, may provide critical defense against invasive organisms attempting to breach the stratified epithelium. Invasive oral pathogens include the opportunistic oral fungus Candida albicans, which is responsible for oral thrush (Conti et al., 2014, Naglik et al., 2008), and Porphyromonas gingivalis, a keystone pathogen of periodontitis progression (Costalonga and Herzberg, 2014).

To study their function in vivo, LCs can be ablated using the human-Langerin promoter-driven diphtheria toxin receptor (huLangerin-DTR) or diphtheria toxin-A (huLang-DTA) murine models (Bittner-Eddy et al., 2016). Loss of LCs cause the Th17 response to be abrogated in response to Candida albicans in the skin (Igyarto et al., 2011) or to P. ginigvalis in the oral mucosa (Bittner-Eddy et al., 2016). While LCs presenting MHC-II are required for the Th17 response to cutaneous C. albicans and oral P. gingivalis infection, a Th1 response to these pathogens can occur independently (Bittner-Eddy et al., 2016, Igyarto et al., 2011). Hence, LCs induce differentiation of IL-17A producing CD4 T cells (Th17) (Bittner-Eddy et al., 2016, Bobr et al., 2010, Igyarto et al., 2011, Kaplan et al., 2005).

To mount a Th17 response to C. albicans in the skin, LCs also require myeloid differentiation primary response 88 (MyD88)-dependent signaling (Haley et al., 2012). The C. albicans microbial-associated molecular patterns (MAMPs) are engaged by surface Toll-like receptors (TLR) and the signal is transduced by the intracellular adaptor molecule MyD88 (Adachi et al., 1998). Absence of MyD88 in LCs resulted in reduced antigen-specific CD4 T cell clonal expansion and reduced Th17 response during engagement of C. albicans in the skin (Haley et al., 2012). Therefore, MyD88 and MHC-II molecules on skin LCs appear to be important for the differentiation and effector function of Th17 cells.

While cutaneous LC are well-studied (Bobr et al., 2010, Igyarto et al., 2011, Kaplan, 2010, 2017, Kaplan et al., 2005, Kaplan et al., 2007) data from cutaneous epithelium may not extrapolate to mucosal epithelium. Although the cytology and phenotypes of LCs are similar, the differentiation of LCs is clearly influenced by their tissue of residency, resulting in unique transcriptomic signatures and niche-specific functions (Capucha et al., 2018, Capucha et al., 2015, Chorro et al., 2009, Horev et al., 2020, Merad et al., 2002). Skin and oral LCs appear to support the differentiation of Th cells differently (Capucha et al., 2015, Horev et al., 2020, Kaplan, 2017). Therefore, we tested whether the differentiation of Th17 cells in response to microorganisms colonizing the oral cavity can be explained by antigen recognition involving MyD88 transduction or MHC-II antigen presentation in oral LCs. To answer these questions, the human-Langerin promoter-Cre mouse model was utilized. Floxed-gene deletions were induced uniquely in LCs to determine the effect of excising the MyD88 or H2-Ab1 (MHC-II) genes on the expansion and differentiation of oral mucosal Th17 cells.

Results

Human-Langerin promoter-Cre conditionally ablates MyD88 or MHC-II expression in Langerhans cells

Utilizing the huLangerin-DTA mouse model, LC ablation at ontogeny prevents Th17 cell differentiation without affecting the Th1 response to oral infection with the periodontal pathogen, Porphyromonas gingivalis (Bittner-Eddy et al., 2016). To determine whether TLR-dependent antigen recognition or antigen presentation on LCs sustains a Th17 response, huLang-Cre mice were bred with either floxed MyD88 or floxed MHC-II mice (Figure 1a, 1b).

FIGURE 1. Conditional excision of genes encoding MyD88 and MHC-II in Langerhans cells.

(a and b) Cartoons of the Myd88^tm1Defr (LC^ΔMyD88) and H2-Ab1^tm1Koni (LC^ΔMHC-II) alleles before and after predicted Cre recombinase action at loxP sites. The structure of the huLangerin promoter Cre recombinase transgene is not shown. Boxes represent exons, solid black triangles show the relative position and direction of loxP sites. Dashed lines indicate the size and position of the expected PCR amplicons. Note that Cre recombinase action on the H2-Ab1^tm1Koni allele will delete both amplicon priming sites. (c and d) Single-cell suspensions obtained from the cervical lymph nodes of two F₂ LC^ΔMHC-II or LC^ΔMyD88 mice were magnetically enriched for CD11c⁺ cells, stained with rat anti-mouse mAbs and sorted on a FACSAria II cytometer into Langerhans cells (LC) or langerin positive dendritic cell (LN⁺ DC) populations. LC were identified as Zombie Aqua^lo, CD45⁺, CD11c^hi, CD3⁻, B220⁻, CD207⁺ and CD103^−ve, while LN⁺ DC were instead CD103^+ve. Flow cytometry plots for LC^ΔMyD88 and LC^ΔMHC-II mouse strains are shown. (e and f) DNA was isolated from sorted LC populations and PCR performed using diagnostic primer pairs. PCR amplicons were separated on 2% agarose gels. All four DNA samples were additionally amplified using primers located within the huLang promoter (285 bp amplicon). MW = DNA molecular weight marker (New England Biolabs, 100 bp ladder).

The floxed deletions were confirmed to be specific to LCs without affecting langerin (LN)⁺ DC (Andrusaite and Milling, 2020, Bittner-Eddy et al., 2019) by analyzing cells sorted from the cervical lymph nodes (CLN) of LC^ΔMyD88 (Figure 1c) and LC^ΔMHC-II mice (Figure 1d). Targeted deletion of exon 3 of the MyD88 gene (Figure 1e) and exon 1 of the H2-Ab1 gene (Figure 1f) occurred exclusively within LCs as confirmed using PCR amplification. Importantly deletions within the MyD88 or MHC-II gene exons were not detected in LN⁺ DC (Figure 1e, 1f), even though the huLangerin promoter-Cre (huLang-Cre) construct could be amplified in both cell types (Figure 1e, 1f). Since LCs and LN⁺ DC were distinguished in CLN preparations via CD103 staining, we also conclude that LC migration to CLN was not impeded by the absence of either MyD88 or MHC-II exons.

Oral mucosa Th17 cell frequency is unaffected by ablation of MHC-II or MyD88 in Langerhans cells

Upon visual inspection, both LC^ΔMHC-II and LC^ΔMyD88 mouse strains caged in specific pathogen free (SPF) conditions appeared normal and did not display obvious spontaneous mucosal or skin inflammatory phenotypes. To determine the contribution of MyD88 signaling and MHC-II antigen presentation by oral LCs to T cells in SPF mice, we quantified the number of resident interstitial T cells in the oral mucosa of wild-type, LC^ΔMHC-II and LC^ΔMyD88 mice. These T cells are assumed to have differentiated and/or expanded in response to microorganisms that have naturally colonized the oral cavity. Interstitial oral mucosa T cells from LC^ΔMHC-II, LC^ΔMyD88 mice and wild-type siblings were separated from non-immune cells and circulating blood-resident lymphocytes (Figure 2a, left). Interstitial oral mucosa T cells were identified as CD45⁺, CD3⁺, TCRβ⁺ cells (Figure 2a, middle) expressing either CD4 T or CD8 T co-receptors (Figure 2a, right). CD8 T cells were similar in wild-type and LC^ΔMyD88 mice (Figure 2b). CD8 T cells in LC^ΔMHC-II mice, however, were 25 to 35-fold higher (Figure 2b). Similarly, CD8 T cells were also elevated in the murine skin (manuscript in preparation). Normalized interstitial oral mucosa CD4 T cell numbers were similar in LC^ΔMHC-II, LC^ΔMyD88 mice and wild-type siblings (Figure 2b). Within the CD4 T cell compartment, the frequency of Th17 cells were similar (Figure 3a, 3b) across all three strains. The frequency of Th1 cells, however, was reduced in LC^ΔMHC-II as compared to WT or LC^ΔMyD88 mice. CD4 Treg cells were also enumerated. Although contentious (Thornton and Shevach, 2019), we used Helios as a marker to differentiate thymus-derived Treg from peripheral/induced Treg cells. The frequency of CD4 T cells expressing FoxP3 was significantly increased in LC^ΔMHC-II mice as compared to WT siblings (Figure 3c, 3d). This increase in frequency can be attributed to the Helios⁺ Treg population.

FIGURE 2. Cell number in the oral mucosa are not affected by ablation of H2-Ab1 or MyD88 in Langerhans cells.

Mice housed in SPF conditions were given 1.25 μg of rat anti-mouse CD45 conjugated to FITC three minutes prior to sacrifice to identify interstitial cells and exclude blood-resident immune cells from downstream analysis. Oral mucosa was harvested, processed and single cell suspensions stained with anti-mouse mAbs to identify interstitial CD4 T cells and CD8 T cells by flow cytometry. (a) Representative flow cytometry plots of live mucosal cells obtained from LC^ΔMHC-II, LC^ΔMyD88 or huLang-Cre^−ve littermate control mice (“WT”) showing gating strategy to identify CD4 T cells and CD8 T cells. Cell numbers were normalized across experiments to 100,000 live non-immune cells. Live non-immune, immune interstitial and blood-resident cells are distinguished in the first column of panels. Live interstitial CD4 T cells were identified as Zombie Aqua^lo, CD45:FITC⁻, CD45:PE⁺, CD3⁺, β TCR⁺ and CD4⁺, CD8α⁻ while interstitial CD8⁺ T cells were CD4⁻, CD8α⁺. (b) Summary data of total numbers of interstitial CD4 T cells and CD8 T cells found in the oral mucosa of LC^ΔMHC-II, LC^ΔMyD88 or huLang-Cre^−ve littermate control mice. Tests of normality and outlier exclusion were performed prior to one-way ANOVA with appropriate post-hoc analysis to compare CD4 T cells and CD8 T cells across the three groups of mice. Each bar represents an n of 10 to 13 mice from 3 separate experimental replicates. **** = p<0.0001; n.s. = not significant. Means plus or minus SEM are plotted.

FIGURE 3. CD4 T cell phenotype was altered after ablation of H2-Ab1 in Langerhans cells.

Single cell suspensions prepared from the oral mucosa were cultured and polyclonally activated with PMA/ionomycin in the presence of brefeldin A. Cells were surface stained to identify CD4 T cells as described in Figure 2. Representative flow cytometry plots showing % of CD4 T cells expressing IL-17A and IFN-γ (a). Th1 and Th17 cells were defined based on expression of IFN-γ or IL-17A, respectively. (b) Summary data of Th1 and Th17 frequency. (c) Representative flow cytometry plots showing intranuclear staining of transcription factors FoxP3 and Helios in CD4 T cells from the oral mucosa. Treg cells were defined as FoxP3⁺ (d) Summary data of frequency of total Treg cells and Helios⁻ and Helios⁺ subpopulations. Data from a total of 10 to 13 mice per group resulting from 3 independent experiments are plotted with means ± SEM. Two-way ANOVA plus appropriate post-hoc test was utilized to compare means in “b” while two-tailed Student’s t-test compared means in “d” after exclusion of outliers and normality test. **** = p<0.0001; n.s. = not significant.

Having established that there is no difference in Th17 cells found in the oral mucosae of LC^ΔMHC-II mice and wild-type sibling under SPF conditions, we asked whether there might be a difference during an active oral infection. In order to test this hypothesis we took advantage of the well-established ligature-induced periodontitis (LIP) model that disrupts the endogenous microflora and causes bacterial dysbiosis and inflammation in the oral mucosa (Abe and Hajishengallis, 2013). We reasoned that the effects of MHC-II ablation in oral LCs maybe two-fold. First, the inability to present antigen to resident CD4 T cells restricts their clonal expansion locally and second, prevents the differentiation of new CD4 T cells in response to the emergent bacterial species found in the dysbiotic environment. We therefore examined the total CD4 T cell response. LIP increased the number of interstitial CD4 T cells in LC^ΔMHC-II and LC^WT mice compared to non-ligated SPF counterparts. Significantly, LIP resulted in a similar number of CD4 T cells in LC^ΔMHC-II and LC^WT mice (Figure 4a). Next, we asked if there might be a difference in the phenotype of these CD4 T cells. However, similar to our previous observation in non-ligated SPF mice we saw no significant difference in the number of Th17 cells present (Figure 4b). Therefore, we concluded that in uninfected mice or during ligature-induced dysbiosis the contribution of MHC class II presentation on LCs does not affect the numbers or the differentiation of Th17 cells homing to the oral mucosa.

FIGURE 4. Molar ligatures drive a significant, but comparable, increase in CD4 T cell and Th17 cell numbers in the oral mucosa of LC^ΔMHC-II and Wild Type mice.

Circumferential ligatures were placed around both maxillary 2^nd molars to induce dysbiosis of the microflora and stimulate an inflammatory response. At day 10, mice were sacrificed following iv administration of 1.25 μg of rat anti-mouse CD45 mAb conjugated to FITC. Oral mucosae were harvested, processed and single cell suspensions cultured and polyclonally activated with PMA/ionomycin in the presence of brefeldin A for 6 hrs. Cells were surface stained with anti-mouse mAbs to identify interstitial CD4 T cells and then intracellularly with rat anti-mouse IL-17A and IFN-γ mAbs. Live interstitial CD4 T cells were identified by flow cytometry as Zombie Aqua^lo, CD45:FITC⁻, CD45:PE⁺, CD90.2⁺, CD3⁺, β TCR⁺, CD4⁺ and CD8α⁻. (a) Summary data of total numbers of interstitial CD4 T cells found in the oral mucosa of LC^ΔMHC-II and wild type littermate control mice with and without ligature placement. Cell numbers were normalized across experiments to 100,000 live non-immune cells. Data are from 8 to 10 mice in two independent experimental replicates and are plotted with means ± SEM. Two-tailed Student’s t-test was used to compare ligated with non-ligated mice, and ligated LC^ΔMHC-II and wild type siblings. ***** = p<0.00001; **** = p<0.0001; n.s. = not significant. (b) Summary data of total numbers of interstitial Th17 cells found in the oral mucosa of LC^ΔMHC-II and wild type siblings following ligation of maxillary 2^nd molars. Cell numbers were normalized across experiments to 100,000 live non-immune cells. Data are from 8 mice in two independent experimental replicates and are plotted with means ± SEM. Two-tailed Student’s t-test was used to compare ligated LC^ΔMHC-II and wild type siblings. n.s. = not significant.

MHC-I expression is unaffected by MHC-II depletion

The expansion of CD8 T cells in the oral mucosa of LC^ΔMHC-II mice might be linked to increased MHC-I expression on LCs. MHC-I and MHC-II surface expression has been reported to be associated, in that deletion of MHC-II ubiquitination molecule MARCH1 alters MHC-II trafficking and subsequently decreases MHC-I levels on dendritic cells (Wilson et al., 2018). If increasing the amount of MHC-II expression on a cell can decrease MHC-I levels, perhaps decreasing MHC-II levels would do the opposite, resulting in more MHC-I available to prime CD8 T cells. We therefore examined the expression of MHC-I on oral mucosa LCs isolated from epithelial sheets using flow cytometry (Figure 5a). The level of MHC-I on oral LCs was similar in LC^ΔMHC-II mice and wild type siblings in oral mucosa epithelium (OM-E) and in CLN (Figure 5b). Since the increase in oral CD8 T cells occurs without an increase in MHC-I expression on oral LCs, CD8 T cell numbers in the oral mucosa must be indirectly regulated by LCs via an MHC-II-dependent mechanism.

FIGURE 5. MHC-I levels on LC are not affected in LC^ΔMHC-II mice.

Single-cell suspensions were obtained from epithelial sheets of oral mucosa (OM-E) from LC^ΔMHC-II mice and huLang-Cre^−ve littermate control mice (“WT”) after Dispase II digestion and from CLN. Cells were stained with rat anti-mouse mAbs and analyzed with an LSR II cytometer. (a) Mean fluorescence intensity (MFI) histogram of PE-conjugated anti-MHC-I mAb on LC in epithelial sheets identified as Zombie Aqua^lo, CD45⁺, CD11c^hi, EpCAM^hi, CD207⁺ population. (b) Summary data of MFI of PE- or PacB-conjugated anti-MHC-I mAb on LC in OM-E or CLN, respectively. Shown are MFI means ± SEM of wild type (black) and LC^ΔMHC-II siblings (white). N = 8 mice in each group. LC from CLN were identified as described in Figure 1. Means were compared by two-tailed Student’s t-test.

Some oral mucosa LCs retain expression of MHC-II in LC^ΔMHC-II mice

Ablation of LCs driven by the expression of huLang-DTA abrogates the Th17 in skin and oral mucosa model (Bittner-Eddy et al., 2016, Bobr et al., 2010, Kaplan et al., 2005). Therefore, a normal Th17 response in the oral mucosa of LC^ΔMHC-II mice may be due to either failure to delete the floxed H2-Ab1 gene or by acquiring MHC-II molecules from neighboring antigen presenting cells via trogocytosis. To test the first possibility, we isolated epithelial sheets from oral mucosa and skin (Figure 6).

FIGURE 6. MHC-II persists on some oral LC but not on skin LC.

Single-cell suspensions were obtained from epithelial sheets of oral mucosa and skin of LC^ΔMHC-II mice and huLang-Cre^−ve littermate control mice (“WT”) after Dispase II digestion. Cells were stained with rat anti-mouse mAbs and analyzed on an LSR II cytometer. (a) Gating strategy to identify LC from the Zombie Aqua^lo, CD45⁺ population. Boolean ‘AND’ gates were applied in FlowJo software to define epithelial resident CD11c^hi EpCAM^hi CD207⁺ LC that did or did not express MHC-II molecules on their surface. (b) In LC^ΔMHC-II mice percent of LC expressing MHC-II are displayed in purple while MHC-II negative LC are shown in red. Data are from 8 mice in two experimental replicates, means ± SEM are shown.

Within live CD45⁺ cells, a population of CD11c⁺ CD207⁺, cells were identified in OM-E. To exclude any LN⁺ DCs that may have remained within the epithelial peel, only EpCAM^hi cells (Nagao et al., 2009) were considered to be true LCs (Figure 6a, S1). Furthermore, direct analysis of the epithelial peel showed only minimal contamination (~2%) of EpCAM^int/lo LN⁺ DCs (Figure S1). In contrast to the epithelium and lamina propria, in CLN we established that LCs and LN⁺ DC could not be distinguished solely by EpCAM expression (Figure S2).

Using Boolean gating, the presence of MHC-II was then determined on OM-E resident LCs (Figure 6a). Surprisingly, 18.8 ± 5.7% of the LCs of LC^ΔMHC-II mice retained MHC-II expression (Figure 6b left). In contrast, less than 1% of skin LCs remained MHC-II⁺ (Figure 6b right). EpCAM^int/lo CD11c⁺ CD207⁺ cells, which were likely contaminating LN⁺ DCs, did not show evidence of MHC-II depletion in the OM-E (Figure S1).

We next examined whether the presence of MHC-II on some LCs was due to trogocytosis. Using cell sorting followed by PCR, we established that oral MHC-II^−ve LCs lacked exon 1 of the H2-Ab1 gene as expected (Figure 7a). Surprisingly, the MHC-II^+ve population retained exon 1, despite the concurrent presence of the huLang-Cre transgene (Figure 7a). Moreover, the mean fluorescence intensity (MFI) of MHC-II on MHC-II⁺ LCs in LC^ΔMHC-II and wild type mice was similar (Figure 7b). Taken together, the residual population of MHC-II⁺ LCs is best explained by variable huLang-Cre activity in oral LCs and not trogocytosis.

FIGURE 7. Oral LC that display MHC-II in LC^ΔMHC-II mice retain a complete H2-Ab1 gene.

(a) Pooled single-cell suspensions were obtained from the epithelial sheets of oral mucosa harvested from 5 LC^ΔMHC-II mice. Cells were stained to identify LC and MHC-II expression using rat anti-mouse mAbs and sorted into LC MHC-II^−ve and MHC-II^+ve fractions using a FACSAria II cytometer. LC were identified as Aqua Zombie^lo, CD45⁺, CD11c⁺, CD90⁻, EpCAM^hi, and CD207⁺. Flow cytometry plots display the percentage of LC sorted into the MHC-II^−ve and MHC-II^+ve fractions. DNA was extracted from the sorted populations and PCR used to determine the presence or absence of the huLang promoter and the H2-Ab1 gene as described in Figure 1. PCR products were separated on a 2% agarose gel (285 bp and 449 bp amplicons). (b) Summary data of MFI of BV786-conjugated anti-MHC-II mAb on LC identified from epithelial sheets of oral mucosa (OM-E). Data is displayed as mean ± SEM from 8 WT mice and from the residual MHC-II^+ve LC population identified from 8 LC^ΔMHC-II mice. Means were compared by two-tailed Student’s t-test.

CD103⁻ CD11b^hi cells make up the majority of MHC-II^+ve LCs in LC^ΔMHC-II mice.

In LC^ΔMHC-II mice, a fraction of oral LCs retains MHC-II expression despite the presence of the huLang-Cre construct, yet in skin LCs MHC-II loss is complete. Given that CD103⁺ LCs do not reside in the skin but are found in the oral mucosa, we tested whether the subpopulation of CD103⁺ oral LCs is the population that retained MHC-II expression. To analyze MHC-II expression, three populations of oral LCs were identified based on the presence of CD103 and CD11b (Figure 8a). CD103⁻ CD11b^hi LCs were the most abundant subpopulation in the oral mucosa, representing 61.8 ± 4.8% of total oral LCs. CD103⁺ CD11b^hi LCs were 33.7 ± 3.9 % of the LCs population. CD103⁺ CD11b^lo LCs made up the remaining 4.6 ± 3.2% of the LC population.

FIGURE 8. Proportion of LCs retaining MHC-II in LC^ΔMHC-II mice varies by subpopulation.

Single-cell suspensions were obtained from an epithelial sheet of oral mucosa of LC^ΔMHC-II mice after Dispase II digestion, stained with rat anti-mouse mAbs and analyzed on an LSR II cytometer. LC were identified from the Zombie Aqua^lo, CD45⁺ CD11c⁺, EpCAM^hi, and CD207⁺. (a) LC population was further subdivided by CD103 and CD11b levels. (b) Percentage of MHC-II⁺ LC within each LC subpopulation. (c) Normalized total number of residual MHC-II⁺ LC out of LC within each subset. Data are from 8 mice in three experimental replicates, means ± SEM are shown. ANOVA with appropriate post-hoc analysis was utilized to compare residual expression of MHC-II across subgroups of LCs. *** = p<0.001.

In LC^ΔMHC-II mice, the majority (60.0 ± 15.8 %) of CD103⁺ CD11b^lo LCs retained expression MHC-II (Figure 8b, grey) and are more resistant to the H2-Ab1 gene deletion compared to the two other populations. The CD103⁺ CD11b^hi subpopulation was the most susceptible to H2-Ab1 gene deletion (Figure 8c, pink) while the CD103⁺ CD11b^lo subpopulation are the least affected (Figure 8b, grey) Although only 16.2 ± 8.0 % of CD103⁻ CD11b^hi cells expressed MHC-II, these cells make up 62.3 ± 6.5 % of all residual MHC-II⁺ oral LCs becoming the majority of the MHC-II presenters in the oral epithelium in LCΔMHC-II mice (Figure 8c).

In summary, the huLang-driven Cre recombinase does not excise the floxed H2-Ab1 gene in all subpopulations of oral LCs equally. In the LC^ΔMHC-II mouse model, only ~19% of oral LCs are available to present antigens on MHC-II to CD4 T cells in draining lymph nodes and within the oral epithelium. This level of MHC-II-dependent antigen presentation is sufficient to drive a normal Th17 phenotype. The residual MHC-II⁺ LC population, however, is unable to prevent the expansion of CD8 T cell numbers perhaps through insufficient local activation of CD4 Treg cells within the oral epithelium.

Discussion

LCs are essential to elicit a Th17 response to pathogenic microorganisms at stratified epithelial surfaces (Bittner-Eddy et al., 2016, Bobr et al., 2010, Igyarto et al., 2011, Kaplan et al., 2005). Surprisingly, the ablation of MyD88 or H2-Ab1 genes driven in LCs by a human Langerin (huLang) promoter did not change either the number or frequency of Th17 cells in the oral mucosa when compared to wild-type siblings. In LC^ΔMHC-II mice, the frequency of Th1 cells and Helios⁺ Treg cells was altered. Losing the ability to present MHC-II antigens on 81% of oral LCs has a starkly different effect on CD4 as compared to CD8 T cells. While the residual 19% of oral LCs expressing MHC-II is sufficient to sustain a normal Th17 response to the endogenous microflora, it is not sufficient to prevent an aberrant expansion of intraepithelial CD8 T cells in the oral mucosa. Hence, the number of LCs expressing MHC-II may have differential thresholds for the MHC-II-dependent indirect regulation of CD8 T cell numbers as compared to CD4 T cell subsets.

Interestingly, the residual MHC-II expression on LCs was unique to the oral mucosa and not to the skin of LC^ΔMHC-II mice. In LC^ΔMHC-II mice, less than 1% of skin LCs expressed MHC-II in agreement with previous reports (Igyarto et al., 2009, Yao and Kaplan, 2018) as compared to approximately 19% of oral LCs. This observation explains our seemingly contradictory PCR analysis that revealed absence of the H2-Ab1 gene in LCs isolated from CLN since skin LCs represents greater than 90 % of all LCs found in the CLN (Bittner-Eddy et al., 2016). In dermal analysis of LC^ΔMHC-II mice, complete skin abrogation of MHC-II expression on LCs greatly dampened the Th17 response to cutaneous pathogens (Igyarto et al., 2011, Igyarto et al., 2009, Yao and Kaplan, 2018). However, our study has demonstrated that in the oral mucosa of genetically identical LC^ΔMHC-II mice, where depletion of MHC-II expression on LCs is incomplete, a normal Th17 response to oral dysbiosis can still be observed. This suggests that the presence of even a minority population of MHC-II expressing LCs is sufficient to cross a necessary threshold for generation of a Th17 response during inflammation.

Why might skin LCs have uniform ablation of MHC-II driven by the huLang-cre construct whereas LCs in the oral mucosa do not? Skin and oral Langerhans cells have different ontogeny and cannot be viewed as a homogenous population, and follow unique developmental pathways. While the presence of Birbeck granules, expression of CD207 (langerin), CD11c, MyD88 and MHC-II are common to all LCs, numerous differences between skin and oral LCs have also been reported (Capucha et al., 2015, Horev et al., 2020, Merad et al., 2002). The skin becomes impermeable to bone marrow-derived LC precursors at embryonic day 16. After day 16, skin LCs develop from a self-renewing and radioresistant population of macrophages with some dendritic cell-like properties (Capucha et al., 2018, Doebel et al., 2017). In contrast, the oral mucosa continues to recruit bone marrow-derived LCs precursors through maturity. The turnover of LCs is, therefore, considerably faster in the oral epithelium than in the skin. Additionally, oral mucosal LCs are derived from both pre-dendritic and monocyte precursors and are neither radioresistant nor self-renewing, meaning that they do not have the tissue-resident macrophage properties of skin LCs (Capucha et al., 2015, Doebel et al., 2017). Lastly, oral LCs express slightly less EpCAM than skin LCs (Capucha et al., 2015, Kaplan, 2017). This may indicate skin LCs are more tightly moored to their surrounding epithelial tissue than oral LCs. Both populations of LCs still express significantly higher EpCAM than langerin expressing dermal dendritic cells (LN⁺ DCs) in the subepithelial tissues (Nagao et al., 2009).

Given the established differences between oral and skin LCs, we hypothesized that oral LCs may not uniformly express the huLang-Cre construct allowing for a residual population of MHC-II⁺ oral LCs to persist. In this scenario, MHC-II⁺ oral LCs retain exon 1 of the H2-Ab1 gene. An alternative hypothesis to account for residual surface MHC-II is that oral LCs might have pilfered MHC-II from neighboring antigen presenting cells through the process of trogocytosis (Yao and Kaplan, 2018). In LC^ΔMHC-II mice we demonstrated that only residual MHC-II⁺ oral LCs retain the DNA for exon 1 of the H2-Ab1 gene. Moreover, in the LC^ΔMHC-II mice and wild-type (huLang-Cre^−ve) siblings, the MFI of MHC-II on MHC-II⁺ oral LCs was equal. If trogocytosis had occurred, we would have expected the MFI of MHC-II to be significantly lower since only irregular patches of MHC-II⁺ membrane would be transferred (Yao and Kaplan, 2018). The evidence favors that a minority population of LCs exclusively in the oral epithelium retained functional expression of MHC-II. Our study, therefore, supports the conclusion that LCs found in the oral mucosa are distinct from the population found in the skin epithelium. Moreover, the results also highlight differences within oral LCs to modulate CD4 and CD8 T cell responses.

Why subpopulations of oral LCs are uniquely resistant to MHC-II ablation driven by an engineered huLang promoter-Cre remains unknown. In the skin epithelium, LCs are universally CD103 negative, which distinguishes them from lamina propria resident CD103⁺ LN⁺ DCs (Kaplan, 2017). It has been previously reported in a murineLangerin-DTR model that two populations of LCs exist in the oral mucosa: those that are CD103⁺ CD11b^lo, and those that are CD103⁻ CD11b^hi (Capucha et al., 2015). Given that the CD103⁺ CD11b^lo population of LCs is unique to the oral mucosa, we hypothesized that the residual MHC-II expression on oral LCs would correlate with the expression of CD103. However, not all authors agree that CD103⁺ langerin⁺ cells represent true oral LCs (Arizon et al., 2012). Residency in the epithelium is critical for establishing genuine LC identity, whether driven by tissue specific signals or interactions with the microbiome (Capucha et al., 2018, Capucha et al., 2015, Horev et al., 2020). To be certain that residual MHC-II expression and its effects are only attributable to LCs and not to lamina propria LN⁺ DCs (Adachi et al., 1998), oral mucosal tissues were treated with Dispase^® II to separate the epithelium from the submucosa. Isolation of cells from epithelial peels coupled with a strict requirement for high expression of EpCAM on LCs excluded LN⁺ DCs from our analysis. In addition, the huLang promoter-Cre construct used to generate this conditional knockout mouse is reported to be active exclusively in LCs and not in LN⁺ DCs nor in other cells in secondary lymphoid organs that can express langerin (Bobr et al., 2010, Bursch et al., 2007, Kaplan et al., 2005, Kaplan et al., 2007). In CD207⁺ EpCAM^int/lo cells, which we defined as the LN⁺ DC subset, MHC-II was not ablated in the OM-E. CD207⁺ EpCAM^hi cells in the lamina propria likely represent transiting LCs in both WT and LC^ΔMHC-II siblings, irrespective of MHC-II expression.

Although the CD103⁺ CD11b^lo subset was clearly the most resistant to deletion of H2-Ab1 exon 1, the majority of residual MHC-II⁺ LCs were CD103⁻ CD11b^hi. We also observed a population of CD103⁺ CD11b^hi LCs, which were not reported in previous descriptions of CD103⁺ oral LCs (Capucha et al., 2018, Capucha et al., 2015). Interestingly, the majority of these cells had MHC-II ablated, which suggests they express the huLang promoter-Cre and are thus unequivocally LCs. However, resistance to H2-Ab1 gene deletion could not be correlated with expression of either CD103 or CD11b. Exactly why some oral LCs retain the gene for MHC-II while ablation is virtually complete in skin LCs therefore remains enigmatic. Since skin and oral LCs are differentially susceptible to MHC-II depletion, it is possible that LCs with different ontogeny have different penetrance of the huLang-Cre allele. CD103⁺ CD11b^lo oral LCs were reported to exclusively derive from pre-dendritic cells as opposed to CD11b^hi oral LCs, which originate from either pre-dendritic or monocyte precursors (Capucha et al., 2015).

It has been reported that when both LCs and LN⁺ DCs are ablated in a murine periodontitis model, Treg cell numbers were reduced and IFN-γ levels were elevated, but there was no change in IL-17A production by antigen-specific CD4 T cells (Aramaki et al., 2011). In contrast when LCs are exclusively ablated with the huLang-DTA model, the antigen-specific Th1 response to oral P. gingivalis is elevated and the Th17 response is abrogated (Bittner-Eddy et al., 2016). Interestingly, in the current study where MHC-II is absent on the majority of LCs, the Th17 response to the oral microflora found both in SPF conditions and dysbiosis induced by LIP remains unaltered. These results together suggest that a normal Th17 response is not dependent on other antigen presenting cells and that there is a threshold of MHC-II expression sufficient to differentiate such Th17 cells. We can potentially reconcile these contrasting findings because two different promoter systems were utilized to drive conditional transgene expression, one affecting exclusively LCs and the other model affects both LCs and LN⁺ DCs (Bobr et al., 2010, Bursch et al., 2007, Kaplan, 2017, Kaplan et al., 2007). Utilizing PCR of sorted cell populations, our current huLang-driven ablation model shows that the huLang promoter-Cre is not active in LN⁺ DCs to drive deletion of H2-Ab1 and MyD88 genes. Our flow cytometry data on LN⁺ DCs isolated from the oral mucosa confirms that these cells express MHC-II. Hence, LCs and not LN⁺ DCs are primarily responsible for controlling the Th17 phenotype at mucosal surfaces (Aramaki et al., 2011, Arizon et al., 2012, Bobr et al., 2010)

Despite the stability of the Th17 response across all three strains of mice, the frequencies of Th1 cells decreased and Helios⁺ Treg cells increased in LC^ΔMHC-II mice. However, the number of CD4 T cells within the oral mucosa was similar between WT and LC^ΔMHC-II siblings. This similarity demonstrates that depletion of MHC-II in LCs alters the phenotype, but interestingly not the magnitude of the CD4 T cell response in the murine oral mucosa. One interpretation of our data is that there may be a hierarchy within the oral LC sub populations that differentially drive naïve CD4 T cells to a particular phenotype. Certainly, differential polarization capability of diverse dendritic cell types has been described (Haley et al., 2012, Yin et al., 2021). For example, CD103⁺ CD11b^hi oral LCs have proportionally the greatest MHC-II ablation of any of the 3 oral LC sub populations. Given the decrease in the frequency of Th1 cells in LC^ΔMHC-II mice, it is tempting to speculate that these cells may favor Th1 differentiation.

The finding in the CD4 T cell compartment is in stark contrast to the CD8 T cells, where their numbers in LC^ΔMHC-II are highly elevated compared to WT siblings. How oral LCs are regulating CD8 T cell numbers remains unclear. It seems almost paradoxical that deleting MHC-II on oral LCs would only impact the numbers of MHC-I-restricted CD8 T cells. Neither oral nor skin LCs are generally thought to present antigen to or stimulate CD8 T cells to proliferate in vitro, although there is some disagreement on the topic (Capucha et al., 2015, Kaplan, 2010, Stoitzner et al., 2006). Other antigen presenting cells, such as the dendritic cells, interact with CD8 T cells in the submucosa, lamina propria, and buccal region (Capucha et al., 2015, Elnekave et al., 2010, Li et al., 2020, Li et al., 2012). For example, a reduction in LN⁺ DCs reduces Treg populations in an ear lesion model (Kautz-Neu et al., 2011). It has been reported in one study that buccal LCs are critical in priming of CD8 T cells in response to vaccination with plasmid DNA (Nudel et al., 2011), but another study established that buccal DCs and not buccal LCs present antigen to stimulate CD8 T cells (Capucha et al., 2015). Furthermore, the heightened numbers of CD8 T cells did not result from an increased expression of MHC-I on LCs in LC^ΔMHC-II mice. LCs themselves do not appear to regulate CD8 T cell numbers directly. Treg cells may be a candidate MHC-II-restricted T cell population responsible for regulating intraepithelial CD8 T cells as LCs are reported to have a regulatory role in some disease models (de Goër de Herve et al., 2012, Kaplan et al., 2005, McNally et al., 2011). Despite the increased frequency of Helios⁺ Tregs in the OM of LC^ΔMHC-II mice, the reduced MHC-II on LCs may not provide sufficient local antigen presentation to negatively regulate CD8 T cell numbers.

This study therefore provides intriguing new evidence for a role of MHC-II on LCs in regulating intraepithelial CD8 and CD4 T cells. Insufficient MHC-II-mediated antigen presentation on LCs allows for aberrant expansion or recruitment of the CD8 T cells, presumably through intermediary CD4 Tregs. We hypothesize that an increased number of CD8 T cells in the oral mucosa may increase resistance to intracellular pathogens and heightened tumor surveillance. If so, inhibiting MHC-II presentation selectively on LCs in the oral mucosa may increase resistance to challenges at mucosal surfaces. Finally, there may be unappreciated additional subsets of oral LCs yet to be described.

Materials and Methods

Mice

All animal experiments were reviewed and approved by the Institutional Animal Care and Use Committee of the University of Minnesota (Protocol 2110–39519A) and performed on age-matched (8–12 week) littermates. Animals were provided with food and water ad libitum on a 12 hour on/off light cycle under specific pathogen free (SPF) conditions. HuLangerin-Cre YFPlox/lox mice (NCI strain #01X66) express yellow fluorescent protein (YFP) and Cre recombinase in response to activation of the Human-Langerin promoter, which is exclusive to LCs and not langerin⁺ dendritic cells (Kaplan et al., 2007). These animals were bred to B6.129X1-H2-Ab1tm1Koni/J (JAX strain #013181) mice which carry loxP sites flanking exon 1 of the H2-Ab1 gene encoding the β chain of the MHC-II molecule I-A^b (Hashimoto et al., 2002) and B6.129P2(SJL)-Myd88tm1Defr/J (JAX strain #008888) mice which carry loxP sites flanking exon 3 of the MyD88 gene (Hou et al., 2008) to generate LC^ΔMHC-II and LC^ΔMyD88 mice, respectively. We utilized a heterozygous breeding scheme based on huLang-Cre segregation so that we obtained test mice that were MHC-II^loxP/loxP huLang-Cre⁺ and sibling “wild type” mice that were only MHC-II^loxP/loxP. All animals utilized were F₂ generation or later to ensure homozygosity of the floxed gene. Mutants were cohoused with their respective wild type siblings to ensure sharing of the oral microbiota.

Isolation of LCs and Langerin⁺ DC from cervical lymph nodes

CLN were harvested from LC^ΔMHC-II and LC^ΔMyD88 mice. Single-cell suspensions were prepared and enriched for CD11c⁺ cells using magnetic purification (Miltenyi Biotech). Cells were stained with rat or hamster anti-mouse mAbs. All antibodies and the vitality dye, Zombie Aqua, were purchased from BioLegend, San Diego, CA unless otherwise noted. Using cell sorting on a FACSAria II (BD Biosciences), LCs were identified as Zombie Aqua^lo, CD45⁺ (30-F11), CD11c^hi (N418), CD3⁻ (145–2C11), CD90.2⁻ (30-H12), B220⁻ (RA3–6B2), CD207⁺ (4C7), CD103⁻ (2E7). Similarly, langerin expressing dermal dendritic cells were identified as Zombie Aqua^lo, CD45⁺, CD11c^hi, CD3⁻ (or CD90.2⁻), B220⁻, CD207⁺, and CD103⁺. To confirm specific targeted floxed deletions, DNA was isolated from sorted populations and PCR amplification of gene fragments of interest was performed using diagnostic primer pairs (Table 1). DNA samples were also amplified using primers located within the promoter of the human Langerin (huLang)-Cre construct.

Table 1.

Primers designation and sequences.

Primer	Sequence	Reference
MHCII del-F	5’ AGG AGC AGT CAC CTC AAA CG 3’	(Hashimoto et al., 2002)
MHCII del-R	5’ CAG GAT TTT GCC ACG CCC TC 3’	(Hashimoto et al., 2002)
MyD88 del-F	5’ AGG CTG AGT GCA AAC TTG GT 3’	(Hou et al., 2008)
MyD88 del-R	5’ AGC CTC TAC ACC CTT CTC TTC T 3’	(Hou et al., 2008)
huLang-Cre-F	5’ GAG GCA AAT GAT TGG CAT TCT AC 3’	Personal communication with D. Kaplan
huLang-Cre-R	5’ CTG GGA AAA TTC AAG AAG AGC CT 3’	Personal communication with D. Kaplan

Identification of interstitial CD3⁺ T cells from oral mucosa

Mice were injected intravenously with 1.25 μg of rat anti-mouse CD45 conjugated to FITC and sacrificed three minutes later, allowing blood-resident immune cells to be excluded from downstream analysis of cells residing in the interstitium (Bittner-Eddy et al., 2013, Bittner-Eddy et al., 2017). Oral mucosa was harvested, processed, and single cell suspensions were prepared as previously described (Bittner-Eddy et al., 2013, Bittner-Eddy et al., 2017). Oral mucosal cells were cultured and polyclonally activated with PMA/ionomycin in the presence of brefeldin A for 6 hours. CD4 T cells were identified by surface staining and then stained intracellularly with rat anti-mouse IL-17A (eBio17B7; Thermo Fisher) and IFN-γ (XMG1.2) using Cytofix/Cytoperm^™ Fixation and Permeabilization Solution (BD Biosciences). Live interstitial CD4 T cells were identified as Zombie Aqua^lo, CD45: FITC⁻, CD45: PE⁺ (30-F11), CD3⁺ (145–2C11), β TCR⁺ (H57–597), CD4⁺ (RM4–5), CD8α⁻ (53–6.7). Tregs were identified within the CD4 T cell population by intranuclear staining with rat anti-mouse Foxp3 (MF-14) and hamster anti-mouse/human Helios (22F6) after fixation and permeabilization with True-Nuclear Transcription Factor Buffer set according to manufacturer instructions (BioLegend). Live interstitial CD8 T cells were identified as Zombie Aqua^lo, CD45:FITC⁻, CD45:PE⁺, CD3⁺, β TCR⁺, CD4⁻, CD8α⁺. Cell numbers were normalized across experiments to 100,000 live non-immune cells to account for potential cell loss during processing and counting.

Identification of Langerhans cells in oral mucosa epithelial sheets

To obtain epithelial sheets, oral mucosa was harvested as previously described (Bittner-Eddy et al., 2013, Bittner-Eddy et al., 2017) and then incubated in 2.2 U/ml Dispase^® II (neutral protease, grade II) (Roche) at 37°C for 50 min (Horev et al., 2020). Epithelial sheets were manually peeled from the submucosa and lamina propria, treated with collagenase and DNase, and processed for single cell suspensions as previously described (Bittner-Eddy et al., 2013, Bittner-Eddy et al., 2017). Cutaneous epithelial sheets from ears were also obtained as described (Horev et al., 2020) to serve as comparison controls to oral mucosal epithelia. All references to skin herein contained refer to skin isolated from murine ears. Epithelial sheets contain LCs, which were characterized by staining with hamster or rat anti-mouse mAbs and analyzed using an LSR II cytometer. LCs were identified from the Zombie Aqua^lo, CD45⁺ (30-F11) population, wherein CD11c^hi (N418), EpCAM^hi (G8.8), CD207⁺ (4C7), and either MHC II⁺ (M5/114.15.2) or MHC-II⁻ populations were defined. Boolean “AND” gates were applied in FlowJo software (TreeStar) to CD11c^hi EpCAM^hi CD207⁺ cells to exclusively identify resident LCs in the oral mucosal epithelia. Subpopulations of LCs were further defined by expression of CD103 (2E7) and CD11b (M1/70; Thermo Fisher). Expression of MHC-I (AF6–88.5) was also evaluated.

Sorting of Langerhans cells from the oral mucosal epithelium to assess MHC-II and excision of exon 1 of the H2-Ab1 gene

Oral mucosa from 5 LC^ΔMHC-II mice were harvested and processed to obtain epithelial sheets and single cell suspensions as described above. Cells were combined and stained to identify LCs as described above and sorted using a FACSAria II into MHC II⁺ and MHC-II⁻ populations. DNA was extracted from these two populations and exon 1 of the H2-Ab1 gene was PCR amplified using diagnostic primer pairs (Table 1). As a positive PCR control, DNA samples were also amplified using primers located within the promoter of the huLang-Cre construct.

Placement of ligatures

Ligatures were placed around the circumference of the maxillary second molars to drive dysbiosis of the endogenous microflora and to induce an inflammatory response in the oral mucosa. LC^ΔMHC-II and wild-type littermate control mice were anesthetized using a ketamine/xylazine mix and laid in a supine position. Lower jaw and tongue were retracted, and black braided silk (5–0) ligatures (Corza Medical) placed around maxillary second molars with the aid of a dissecting microscope. Ligatures were secured by one square knot on the palatal side as previously described (Abe and Hajishengallis, 2013). At day 10, mice were euthanized, and ligatures carefully removed prior to oral mucosa harvesting. Tissues were processed for flow cytometry and phenotyping of CD4 T cells as described above.

Statistical Analyses

Shapiro-Wilk normality test was utilized to determine distribution of data points within each experiment. Identification and exclusion of outliers was done via ROUT method with Q = 1%, meaning that no more than 1% of identified outliers will be false positives. Parametric tests were utilized when the data was normally distributed. When only two comparisons were being made Student’s t-test was performed. Whenever greater than two groups were compared ANOVA with appropriate post hoc analysis was utilized. Standard error of the mean is displayed because we compared means between groups. SEM considers sample size, making it a more reliable measure of variability than standard deviation. Statistical analysis of the data was performed using GraphPad Prism.

Supplementary Material

1

Supplementary FIGURE 1. Expression of EpCAM differentiates LCs and LN⁺ DCs within epithelial peels of oral mucosa. Single cell suspensions from the oral mucosa of LCΔMHC-II and huLang-Cre^−ve and littermate control mice (“Wild Type”) were generated from individually processed oral mucosal epithelium (top) and lamina propria tissue (bottom). Cells from 3 mice were pooled and stained with vitality dye Zombie Aqua, followed by rat anti-mouse CD45, CD90, MHC-II, CD11c, CD207, and EpCAM mAbs, and then analyzed by flow cytometry. Representative flow plots are shown along with the frequency of cells within marked gates. Live antigen presenting cells were identified as Zombie Aqua^lo, CD45⁺, CD90⁻, CD11c⁺ cells. LC were identified within the antigen presenting cell population as CD207⁺, EpCAM^hi while Langerin expressing dendritic cells (LN⁺ DC) were CD207⁺ EpCAM^int/lo. Expression of MHC-II and EpCAM on LCs and LN⁺ DCs were evaluated within the CD11c⁺ CD207⁺ population.

2

Supplementary FIGURE 2. Surface expression of EpCAM does not allow differentiation of migratory LC from LN⁺ DC in CLN. Single cell suspensions from C57BL/6J mice cervical lymph nodes (CLN) were magnetically enriched for CD11c⁺ cells. Enriched cells were stained with vitality dye Zombie Aqua, followed by rat anti-mouse MHC-II, CD11c, CD207, CD103 and EpCAM mAbs, and then analyzed by flow cytometry. Representative flow plots are shown along with the frequency of cells within marked gates. Migratory antigen presenting cells were identified as Zombie Aqua^lo CD11c⁺ MHC-II^hi cells (left panel). Lymph nodal LC were identified as CD207⁺, CD103⁻ while Langerin expressing dendritic cells (LN⁺ DC) were CD207⁺, CD103⁺ (center panel). EpCAM expression in an overlayed cytometry plot is similar in LC (CD103⁻) and LN⁺ DC (CD103⁺) populations (right panel).

Abbreviations:

CLN

Cervical lymph nodes

LCs

Langerhans cells

LN⁺ DCs

Langerin⁺ dendritic cells

LN

Langerin

MHC

Major histocompatibility complex

MyD88

Myeloid differentiation primary response 88

TCR

T cell receptor

Th1

T helper 1 lymphocytes

Th17

T helper 17 lymphocytes

Treg

CD4 T regulatory lymphocytes

OM

Oral mucosa

OM-E

Oral mucosa epithelium

LIP

Ligature induced periodontitis

Data and materials availability

All data are available in the paper or the supplementary materials. Materials will be made available by contacting MC and after completion of a material transfer agreement.