Activation of bone marrow-derived dendritic cells and CD4+ T cell differentiation by outer membrane vesicles of periodontal pathogens

ABSTRACT Outer membrane vesicles (OMVs) released from gram-negative bacteria harbor diverse molecules to communicate with host cells. In this study, we evaluated the OMVs of periodontal pathogens for their effects on the activation of dendritic cells and CD4+ T cell differentiation. OMVs of Porphyromonas gingivalis ATCC 33277, Treponema denticola ATCC 33521, and Tannerella forsythia ATCC 43037 (‘red complex’ pathogens) were isolated by density gradient ultracentrifugation. Mouse bone marrow-derived dendritic cells (BMDCs) were treated with OMVs, and OMV-primed BMDCs were cocultured with naïve CD4+ T cells to analyze the polarization of effector helper T cells. The OMVs upregulated maturation markers, including MHC class II, CD80, CD86, and CD40, on BMDCs. OMVs of P. gingivalis and T. forsythia induced the expression of the proinflammatory cytokines IL-1β, IL-6, IL-23, and IL-12p70 in BMDCs. In T. denticola OMV-primed BMDCs, proinflammatory cytokines were poorly detected, which may be attributed to posttranslational degradation due to the highly proteolytic nature of OMVs. In cocultures of naïve CD4+ T cells with OMV-primed BMDCs, OMVs of P. gingivalis and T. denticola induced the differentiation of Th17 cells, whereas T. forsythia OMVs induced Th1 cell differentiation. These results demonstrate that OMVs derived from the ‘red complex’ periodontal pathogens induce maturation of BMDCs and differentiation of naïve CD4+ T cells to Th1 or Th17 cells.


Introduction
Periodontitis is a chronic inflammatory disease characterized by periodontal tissue destruction, often leading to tooth loss [1]. The proinflammatory response of periodontal tissue is induced by oral microbiota dysbiosis in the subgingival biofilm, with an increased proportion of the 'red complex' bacteria Porphyromonas gingivalis, Treponema denticola, and Tannerella forsythia [2]. The 'red complex' is highly associated with periodontal lesions and disrupts host immune systems by producing proteases, such as gingipains and dentilisin [2]. In addition, periodontal pathogen-derived molecules and bacteria themselves are related to the pathogenesis of systemic diseases, such as Alzheimer's diseases, cardiovascular disease, diabetes, and rheumatoid arthritis [3,4].
Extracellular vesicles (EVs) are nanosized with a diameter of 30-2,000 nm and enclosed by a lipid bilayer membrane [5]. EVs are secreted by all living cells and contain various biomolecules, such as proteins, lipids, nucleic acids, and metabolites, that can mediate cell-cell communication and change the physiological status of recipient cells [5]. EVs named outer membrane vesicles (OMVs) are released by gram-negative bacteria and those named membrane vesicles (MVs) are released by gram-positive bacteria [6]. Since bacterial EVs contain various microbe-associated molecular patterns (MAMPs) and virulence factors, bacterial EVs can stimulate host cells to induce inflammatory responses [7]. Innate immune responses to EVs derived from periodontal pathogens have been reported. OMVs of P. gingivalis, T. denticola, and T. forsythia induce Toll-like receptor (TLR) signaling pathways and inflammasome activation [8,9]. In addition, bacterial EVs can spread to host tissues through the bloodstream and induce local and systemic inflammatory responses [10]. Recently, it has been reported that EVs derived from periodontal pathogens are responsible for periodontitisassociated systemic diseases [11][12][13].
Dendritic cells (DCs) are professional antigenpresenting cells (APCs) that take up and process antigens and then present them to T cells to initiate proper adaptive immunity [14]. APCs express various pattern-recognition receptors (PRRs), which recognize MAMPs to trigger innate immune responses against invading pathogens. DCs are matured by PRR signaling and then migrate to draining lymph nodes to present antigens to naïve T cells. According to the cytokine milieu, naïve CD4 + T cells differentiate into a distinct lineage of helper T cell subsets to modulate immune responses in the appropriate direction [15]. CD4 + T cells in gingival tissue play a central role in orchestrating oral mucosal immunity [16]. In a healthy state, mechanical barrier damage caused by mastication induces Th17 cells for protective immunity of the oral mucosa [17]. However, when the burden of dysbiotic oral bacteria is increased, dysregulation of Th17 cell responses is induced, which leads to chronic inflammation, eventually developing into periodontal disease and bone resorption [18,19].
P. gingivalis OMVs induce neutrophil infiltration [20] and foam cell formation in the presence of lowdensity lipoproteins [21]. EVs of Filifactor alocis, a gram-positive periodontal pathogen, induce systemic bone loss through TLR2-mediated osteoclastogenesis [13]. OMVs of Aggregatibacter actinomycetemcomitans, a gram-negative periodontal pathogen, can cause neuroinflammation through activation of brain monocytes and microglial cells [22]. However, the OMVs of periodontal pathogens have not been studied for their effects on DC activation and CD4 + T cell differentiation. In this study, we analyzed the maturation of DCs and the mode of helper T cell differentiation induced by OMVs derived from the 'red complex' bacteria P. gingivalis ATCC 33277, T. denticola ATCC 33521, and T. forsythia ATCC 43037 using mouse bone marrowderived dendritic cells (BMDCs) and splenic naïve CD4 + T cells.

Isolation of outer membrane vesicles (OMVs)
OMVs were isolated as previously described with some modification [13]. Briefly, 400 ml of each bacterial culture supernatant was collected and centrifuged at 10,000 × g and 4°C for 10 min. Then, cell-free culture supernatants were filtered using a 0.2 μm polyethersulfone membrane filter system (Corning, New York, NY, USA). The filtered culture supernatants were subjected to ultracentrifugation at 120,000 × g and 4°C for 3 h. The supernatants were discarded, and the pellets were resuspended in sterile phosphate-buffered saline (PBS; Welgene, Daegu, Korea) followed by ultracentrifugation at 120,000 × g and 4°C for 2 h. The pellets were resuspended in PBS and mixed with 60% iodixanol solution (OptiPrep TM ; Sigma-Aldrich) to obtain a 40% solution. The iodixanoldiluted pellets were laid on the bottom of an ultracentrifuge tube, and 35%, 30%, 25%, 20%, and 5% iodixanol solution diluted with PBS was overlaid. The discontinuous density gradient layers were ultracentrifuged at 100,000 × g and 4°C for 18 h. The same volume (1.3 ml) of each fraction was harvested from top to bottom. The nanoparticle concentrations in each fraction were analyzed via nanoparticle tracking analysis (NTA) as described below. The nanoparticle-enriched fractions were mixed with PBS (10 ml) and ultracentrifuged at 120,000 × g and 4°C for 2 h. After removal of the supernatant, the OMV pellets were resuspended in PBS. The protein concentration of OMVs was determined by bicinchoninic acid (BCA) assays (Thermo Fisher Scientific Inc., Waltham, MA, USA) according to the manufacturer's instructions. The particle number in 10 μg proteins of each OMV was as follow: P. gingivalis OMVs (1.51 ± 0.54 × 10 10 ), T. forsythia OMVs (1.39 ± 0.40 × 10 10 ), and T denticola OMVs (3.83 ± 0.33 × 10 10 ). The isolated OMVs were aliquoted and stored at −80°C until use.

Transmission electron microscopy (TEM)
Five microliters of OMVs were loaded onto a glowdischarged formvar/carbon-coated copper grid (Electron Microscopy Sciences, Hatfield, PA, USA) for 1 min, washed twice with distilled water, and then stained with 2% uranyl acetate (Electron Microscopy Sciences) for 1 min. The negatively stained OMVs were imaged via TEM (LIBRA 120; Carl Zeiss, Jena, Germany) at 120 kV.

Nanoparticle Tracking Analysis (NTA)
The size and concentration of OMVs were measured with a NanoSight LM10 instrument (Malvern Instruments Ltd., Worcestershire, UK). Each OMV was diluted in nanoparticle-free PBS to adjust the proper concentration range (5 × 10 8 -2 × 10 9 particles/ml). NTA software (Ver. 2.5, Malvern Instruments Ltd.) was used for data analysis, and the acquisition settings used in this experiment were as follows: screen gain, 12; camera level, 15; and detection threshold, 3.

Mice
Eight-week-old C57BL/6N mice were purchased from Orient Bio (Seongnam, Korea). All mouse experiments were approved by the Institutional Animal Care and Use Committee of Seoul National University (SNU-210602-1).

Enzyme-linked immunosorbent assay (ELISA)
The levels of IL-1β, IL-4, IL-6, IL-23, and IL-12p70 in the culture supernatants of BMDCs stimulated with OMVs of periodontal pathogens were measured using ELISA kits (BioLegend and R&D Systems, Minneapolis, MN, USA) according to the manufacturer's instructions. The optical density of each well was measured with an Epoch2 microplate reader (BioTek Instruments Inc., Winooski, VT, USA) at wavelengths of 450 nm and 540 nm.

Degradation of proinflammatory cytokines by periodontal pathogen OMVs
In 96-well cell culture plates, 1 ng/ml of recombinant murine IL-1β (R&D Systems), IL-6 (Peprotech), IL-23 (R&D Systems), and IL-12p70 (Peprotech) were incubated with OMVs (1 and 10 μg/ml) in 200 μl/well RPMI 1640 complete medium for 24 h at 37°C in humidified aerobic conditions (5% CO 2 ). BMDCs were stimulated with 100 ng/ml Pam3CSK4 for 24 h at 37°C. Culture supernatants of Pam3CSK4-treated BMDCs were incubated with T. denticola OMVs in the presence or absence of 2 mM phenylmethylsulfonyl fluoride (PMSF, Thermo Fisher Scientific), a serine protease inhibitor, for 1 h. In addition, T. denticola OMVs were heated at 95°C for 10 min and incubated with culture supernatants of Pam3CSK4-treated BMDCs for 1 h. The levels of the remaining cytokines in the medium or culture supernatants were measured by ELISA.

Statistics
The mean value ± standard deviation (SD) was determined for each group. Student's t-test was used to determine the significance of differences between two groups. One-way analysis of variance (ANOVA) with Dunnett's post-test was used to determine the significance of differences among more than two groups. A p-value less than 0.05 was considered statistically significant. All statistical analyses were performed using GraphPad Prism software (GraphPad Software Inc., San Diego, CA, USA).

Characterization of OMVs from periodontal pathogens
OMVs of P. gingivalis ATCC 33277, T. denticola ATCC 33521, and T. forsythia ATCC 43037 were isolated from bacterial culture supernatants by ultracentrifugation combined with iodixanol density gradient ultracentrifugation (Figure 1a). The size and number of OMVs were measured by NTA, which can directly measure the concentration and size of nanoparticles (OMVs) in a sample by light scattering technologies [24]. Approximately 750 μg protein of P. gingivalis OMVs, 210 μg protein of T. denticola OMVs, and 1200 μg protein of T. forsythia OMVs were harvested from 400 ml of each bacterial culture supernatants. The OMVs showed a round and cupshaped morphology in the negatively stained TEM images, and their sizes ranged from 50 nm to 400 nm (Figure 1b).

Maturation of BMDCs induced by OMVs of periodontal pathogens
Activation of APCs by microbial components is essential for initiating naïve CD4 + T cell differentiation [25]. Activated DCs are mature and express high levels of major histocompatibility complex (MHC) class II and costimulatory molecules, including CD80, CD86, and CD40, on the cell surface to signal to naïve CD4 + T cells [25]. To analyze the maturation of BMDCs, they were stimulated with OMVs of periodontal pathogens for 24 h, and the expression levels of MHC class II and costimulatory molecules on the cell surface were analyzed via flow cytometry. All OMVs significantly induced the expression of MHC class II, CD80, CD86, and CD40 (Figure 2a,b). These results suggest that OMVs of periodontal pathogens induced the maturation of BMDCs.

Expression of CD4 + T cell polarization cytokines secreted from BMDCs induced by OMVs of periodontal pathogens
Since the polarization of naïve CD4 + T cells is regulated by cytokines [25], we measured the cytokines released from BMDCs stimulated with OMVs of periodontal pathogens using ELISA. IL-12p70 was measured for Th1 cell polarization, and IL-4 for Th2 cell polarization. IL-1β, IL-6, and IL-23 were measured for Th17 cell polarization. P. gingivalis OMVs and T. forsythia OMVs significantly induced the secretion of IL-12p70, IL-1β, IL-6, and IL-23 from BMDCs (Figure 3a). T. denticola OMVs increased IL-6 but did not induce IL-12p70, IL-1β, or IL-23 secretion. No OMVs induced IL-4 secretion. As T. denticola OMVs activated BMDCs as efficiently as other OMVs, as judged by the expression of costimulatory molecules (Figure 2), we speculated that T. denticola OMVs might have the ability to degrade cytokines. As expected, T. denticola OMVs significantly induced Il12a, Il1b, Il6, and Il23a, similar to or less than P. gingivlias OMVs and T. forsythia OMVs (Figure 3b). To examine the proteolytic activity of the OMVs, we incubated recombinant cytokines with the OMVs and measured the remaining cytokines via ELISA. As shown in Figure 3c, T. denticola OMVs degraded IL-1β, IL-6, and IL-23 but not IL-12p70 in a dosedependent manner (Figure 3c). To determine whether T. denticola OMVs were able to degrade cytokines secreted from BMDM stimulated with Pam3CSK4, a synthetic lipopeptide known to mimic bacterial lipoproteins, the culture supernatants of BMDCs were harvested and then incubated with T. denticola OMVs. T. denticola OMVs significantly degraded IL-1β, IL-6, and IL-23 in BMDM culture supernatants (Supplementary Figure S1). Heat-or PMSF-treated T. denticola OMVs lost their proteolytic activity against IL-1β, IL-6, and IL-23. These results suggest that T. denticola OMVs induced the production of IL-1β, IL-6, and IL-23, which were subsequently degraded. P. gingivalis OMVs also degraded IL-1β and IL-23 but to a lesser extent than T. denticola OMVs. T. forsythia OMVs degraded IL-1β. These results indicate that OMVs of periodontal pathogens are likely to induce Th1 or Th17 polarization and that T. denticola OMVs possess strong proteolytic activity compared to other OMVs.

Th1/Th17 polarization of CD4 + T cells by BMDCs stimulated with OMVs of periodontal pathogens
Naïve CD4 + T cells were cocultured with OMVprimed BMDCs to analyze the effect of OMVs on T cell polarization. To exclude the possibility that OMVs of periodontal pathogens can degrade cytokines from OMV-primed BMDCs, we vigorously washed OMV-primed BMDCs at 5 h posttreatment and then cocultured them with naïve CD4 + T cells. BMDCs stimulated with OMVs of P. gingivalis and T. denticola induced IL-17A + T cells, while T. forsythia OMV-primed BMDCs mainly induced IFN-γ + T cells along with some IL-17A + T cells (Figure 4a,b Supplementary Figure  S2). IL-4 + T cells were not detected in any coculture. These results indicate that OMVs of P. gingivalis and T. denticola induced Th17 polarization, whereas T. forsythia OMVs favored Th1 polarization rather than Th17.

Evaluation of the cytokines responsible for CD4 + T cell differentiation induced by OMVs of periodontal pathogens
To evaluate which cytokine contributes to the differentiation of Th1 or Th17 cells induced by OMVprimed BMDCs, we treated BMDCs with neutralizing antibodies against IL-12 and IL-6 during coculture and then analyzed CD4 + T cell differentiation. Neutralization of IL-6 significantly increased IFN-γ + cells but reduced IL-17A + cells in the coculture of each periodontal pathogen derived OMV-primed BMDC with naïve CD4 + T cells (Figure 5a-c). Neutralization of IL-12p40 decreased IFN-γ + cells but increased IL-17A + cells in the coculture of T. forsythia OMV-primed BMDCs with naïve CD4 + T cells (Figure 5f). However, IFN-γ + and IL-17A + cells were not affected in the coculture of P. gingivalis OMV-and T. denticola OMV-primed BMDCs with naïve CD4 + T cells (Figure 5d,e). These results suggest that IL-6 is major cytokine in Th17 differentiation by OMVs of P. gingivalis and T. denticola, whereas IL-12 is responsible for Th1 differentiation by T. forsythia OMVs.

Discussion
The responses of T lymphocytes to dysbiotic oral bacteria play an essential role in the immunopathogenesis of periodontal disease [26]. As OMVs harbor   BMDCs were stimulated with 10 μg/ml of the indicated OMVs for 5 h. After three times of washing with PBS, the BMDCs were cocultured with naïve CD4 + T cells for 4 days in the presence of soluble anti-CD3ε antibody and analyzed via flow cytometry. The experiments were performed four times independently, and representative data are shown as dot plots (a) and the proportions of intracellular IFN-γ, IL-17A, and IL-4 in CD4 + T cells (b). The data are presented as the mean ± SD of triplicate assays and were analyzed by one-way ANOVA. * p < 0.05, ** p < 0.01, and *** p < 0.001 compared to the control (NT). NT, non-treatment; PG, P. gingivalis ATCC 33277; TD, T. denticola ATCC 33521; TF, T. forsythia ATCC 43037. various immunostimulatory molecules [7], the present study determined the role of the OMVs of periodontal pathogens in helper T cell polarization through DC maturation. The OMVs of periodontal pathogens activated DCs to secrete effector cytokines of Th1 and Th17 cells but not Th2 cells. P. gingivalis 33277 OMV-and T. denticola 33521 OMV-primed BMDCs induced the differentiation of naïve CD4 + T cells into Th17 cells. In contrast, T. forsythia 43037 OMV-primed BMDCs favored the polarization of Th1 cells rather than Th17 cells.
The maturation of DCs is essential for linking innate immunity to adaptive immunity [25]. There is some evidence that DC maturation by the OMVs of periodontal pathogens might occur through TLR signaling pathways. First, the 'red complex' bacteria induces immune responses mainly through TLR2 and TLR4 [27][28][29]. Second, the OMVs of periodontal pathogens harbor various MAMPs, such as lipopolysaccharide/lipooligosaccharide (LPS/LOS), lipoproteins, peptidoglycan, DNA, and RNA, which are the ligands of TLRs [8]. Several studies have reported that other bacterial OMVs can induce DC maturation through TLR signaling pathways [30][31][32]. Furthermore, since OMVs are potent immune stimulators and deliver bacteria-specific antigens without the bacteria being present, OMVs are receiving attention as an ideal vaccine against infectious diseases [33]. Therefore, further studies are needed to identify the molecular mechanisms of TLR signaling that are responsible for DC maturation induced by MAMPs in the periodontal pathogen OMVs.
For normalization of OMV treatment, one must choose either a biomolecule or a number of OMVs. According to Cecil et al. (2016), the amounts of protein, LPS, lipoprotein, peptidoglycan, DNA, and RNA were too different in the same number of OMVs of P. gingivalis, T. denticola, and T. forsythia [8]. In the report, the degree of TLR2 and TLR4 activation by the same amount of OMV proteins was similar to each other, but the degree of TLR2 and TLR4 activation by the same number of OMVs was very different, especially in the case of T. denticola OMVs. Although the particle number in 10 μg proteins of P. gingivalis OMVs and T. forsythia OMVs was similar, their ability to induce Th1 or Th17 differentiation was different. The particle number in 10 μg proteins of T. denticola OMVs was higher than that in 10 μg proteins of P. gingivalis OMVs and T. forsythia OMVs, but showed similar BMDC-activating abilities to P. gingivalis OMVs and T. forsythia OMVs. Therefore, we believe that protein quantification of OMVs is a reliable criterion for assessing host response to OMVs and that different T cell polarization by each OMVs might be due to their different qualitative rather than quantitative nature.
The present study demonstrated that P. gingivalis OMVs induced DC-mediated Th17 polarization. It seems that the ability of P. gingivalis OMVs to induce Th17 polarization is similar to that of P. gingivalis were performed three times independently. The data are presented as the mean ± SD of triplicate assays and were analyzed by Student's t-test. ** p < 0.01 and *** p < 0.001 compared to the control (isotype). N.S., not significant. PG, P. gingivalis; TD, T. denticola; TF, T. forsythia. whole cells. P. gingivalis induces Th17 responses in vivo and in vitro by activating DCs and monocytes [34][35][36][37]. P. gingivalis LPS has been found to upregulate Th17 cell differentiation from activated human naïve CD4 + T cells in the presence of Th17-driven cytokines [38]. Th17 cells induced by P. gingivalis affect the pathogenesis of periodontitis and systemic diseases, such as rheumatoid arthritis [39]. P. gingivalis oral infection induces alveolar bone resorption and aggravates the severity of arthritis, which may be associated with increased systemic Th17 responses [40,41]. Since P. gingivalis OMVs can spread to host tissues through the bloodstream [11], P. gingivalis OMVs may induce Th17 differentiation, leading to periodontitis and other systemic diseases. Since only the P. gingivalis ATCC 33277 strain was used in this study, it is necessary to analyze various P. gingivalis strains, including W50 and gingipain mutant strains, for DC maturation and T cell differentiation.
Unlike P. gingivalis, the role of T. denticola in CD4 + T cell responses has not been well studied. The present study demonstrated for the first time that T. denticola OMVs induced DC-mediated Th17 polarization but not Th1. IL-6 released from both P. gingivalis OMVand T. denticola OMV-primed BMDCs was found to play a pivotal role in Th17 polarization. Among various cytokines from mature DCs, IL-6 is an essential cytokine for the commitment of Th17 cells [42,43]. IL-6 activates the STAT3 signaling pathway through gp130 on naïve CD4 + T cells [42], and STAT3 is a crucial transcription factor for Th17 cell differentiation [43]. In addition, IL-6 inhibits the differentiation of Th1 cells by inhibiting SOCS1 function [44].
In contrast to P. gingivalis OMV-and T. denticola OMV-primed BMDCs, T. forsythia OMV-primed BMDCs preferentially induced the differentiation of Th1 cells rather than Th17 cells. IL-12 was found to play a pivotal role in Th1 differentiation by T. forsythia OMV-primed BMDCs. IL-12 is known as an inducer of naïve CD4 + T cell differentiation into Th1 cells [45]. T. forsythia OMVs significantly induced IL-12p70 expression in BMDCs compared to OMVs of P. gingivalis and T. denticola. Indeed, the neutralization of IL-12 reduced Th1 cells but increased Th17 cells in the coculture of T. forsythia OMV-primed BMDCs with naïve CD4 + T cells. Furthermore, the S-layer glycan of T. forsythia was previously reported to restrain Th17 cell responses in a mouse model and human peripheral blood mononuclear cell study [46,47]. The S-layer glycan is a typical structure of T. forsythia and protects the bacterium from recognition by DCs [46]. Further studies are needed to identify the detailed molecular mechanisms underlying the ability of OMVs from different bacterial species to differentially regulate T cell polarization.
Among the three periodontal pathogen OMVs, T. denticola OMVs showed the highest proteolytic activity. Dentilisin is a chymotrypsin-like proteinase located in the outer membrane of T. denticola [48]. Veith et al. showed that dentilisin is one of the abundant proteins in T. denticola OMVs, and its topology on T. denticola OMVs faces the extracellular space [49]. It has been shown that dentilisin can degrade several host proteins, including proinflammatory cytokines, such as IL-1β, IL-6, IL-8, and TNFα [50]. Ginginpains are well-characterized proteolytic enzymes and virulence factors of P. gingivalis [51]. There are three types of gingipains. RgpA and RgpB are arginine-specific cysteine proteinases, and Kgp is a lysine-specific cysteine proteinase [52]. Gingipains are abundant on P. gingivalis OMVs [53] and contribute to evasion of the host immune surveillance system by degrading cytokines such as IL-1β, IL-6, and IFN-γ [54]. In this study, P. gingivalis OMVs degraded recombinant murine IL-1β and IL-23 but not IL-6 and IL-12p70. Therefore, OMV secretion is a useful tool of T. denticola and P. gingivalis for evading host immune systems.
In conclusion, this study demonstrated that BMDCs stimulated with P. gingivalis 33277 OMVs and T. denticola 33521 OMVs induced the differentiation of Th17 cells and that T. forsythia 43037 OMVs favored Th1 cells polarization rather than Th17 cells. IL-6 and IL-12 released from OMVprimed BMDCs played pivotal roles in Th17 and Th1 polarization, respectively. These results suggest that the OMVs of 'red complex' periodontal pathogens may play a central role in the pathogenesis of periodontitis by inducing Th1 and Th17 cells through activation of DCs. The limitations of this study were that 1) only one strain of each bacterium was used, 2) OMVs derived from different growth phases and various culture media were not analyzed, 3) only mouse cells, not human cells, were used, and 4) the mechanisms by which OMVs act on the BMDCs remain unsolved. In the future, these issues must be addressed to clarify the role of OMVs of 'red complex' periodontal pathogens in T cell differentiation.

Disclosure statement
No potential conflict of interest was reported by the author(s).