Slight Pro-Inflammatory Immunomodulation Properties of Dendritic Cells by Gardnerella vaginalis: The “Invisible Man” of Bacterial Vaginosis?

Bacterial vaginosis (BV), the most common genital infection in reproductive-aged women, is associated with increased risk of sexually transmitted infections. Its etiology remains unclear, especially the role of Gardnerella (G.) vaginalis, an anaerobic bacterium characteristic of the BV-alteration of the vaginal ecosystem. In the genital mucosa, dendritic cells (DCs) sense bacteria of the microenvironment via receptors and then orchestrate the immune response by induction of different T cell subtypes. We investigated the interactions between G. vaginalis and human monocyte-derived DCs using a wide range of bacterial concentrations (multiplicity of infection from 0.01 to 100), and the effects of this pathogen on PHA-induced lymphocyte proliferation. As observed by electron microscopy and cytometry, G. vaginalis reduced the internalization ability of DCs by forming extracellular clusters and induced neither DC maturation, nor DC secretion of cytokines, except at the highest dose with a very early DC maturation state. The same profile was observed on lymphocytes with significant increases of proliferation and cytokine secretion only at the highest bacterial concentration. Our findings indicate that G. vaginalis possesses slight immune-stimulating activities against DCs and T cells, reflecting thus a defective inflammatory response and giving rise to the atypical, non- or low-grade, inflammatory clinical disease profile.


Introduction
Bacterial vaginosis (BV) is the most common low genital infection among reproductive-aged women, with a prevalence of 29% among 14-to 49-year-old US women and almost 40% in individuals at high risk for sexually transmitted infections (STIs) [1]. BV is associated with serious medical complications, including adverse pregnancy outcomes, endometritis, and pelvic inflammatory disease such as endometriosis [2,3]. BV also increases women's risk of acquiring STIs, particularly HIV infections [4,5].
Clinically, one-half of BV-positive women are asymptomatic while the others suffer only from mild symptoms, such as homogeneous white vaginal discharge and amine (fishy) odor [6]. These signs are associated with a vaginal pH > 4.5 and the presence of characteristic "clue cells" on microscopic examination. These four manifestations constitute Amsel's clinical criteria [6]. The microbiological diagnosis Only one study reported effects of BV on the percentage of Treg cells in peripheral blood mononuclear cells (PBMCs). However, this study did not test the possibility of a specific impact of G. vaginalis and it did not objectify differences in the distribution of Treg in BV+ versus BV− HIV-negative women, decreased Treg being solely observed in BV+/HIV+ women compared to BV−/HIV+ women [25]. Many studies have attempted to measure cervicovaginal production of cytokines in BV, but disparate results were obtained. Most articles reported an elevation in IL-1 , and less consistently in IL-6 and IL-8, in BV-affected women [26][27][28][29]. Additionally, BV mucosal fluid was found to increase proliferation of T cells in allogeneic mixed-leukocyte reaction (MLR) [23]. Finally, the specific effects of G. vaginalis on DC and T cells have never been evaluated yet.
Unlike conventional vaginitis that is characterized by burning, dysuria, dyspareunia, and frequent pruritus, BV causes scant inflammatory signs without primary pain or pruritus in affected women [14]. Likewise, a relative paucity of inflammatory cells and a near normal number of vaginal neutrophils are characteristic of BV status. In view of the literature data, we hypothesized that BV corresponds to a unique local immunological environment, with a lowgrade inflammation, potentially mediated by so far unknown immunomodulatory mechanisms of action of G. vaginalis on the vaginal immune system, particularly on the DCs and T cells. In the present study, we investigated this hypothesis in in vitro models by monitoring (i) the internalization, maturation, and cytokine secretion of moDCs; (ii) lymphocyte proliferation and subset cytokine production, after cell exposure to G. vaginalis or to commensal or pathogenic microorganisms potentially found in the vaginal mucosa.

Bacterial Strains and Culture Conditions. G. vaginalis
ATCC14018 was grown in brain heart infusion (Biomérieux) supplemented with maltose (0.1%), glucose (0.1%), yeast extract (1%), and horse serum (10%) in 5% CO 2 at 37 ∘ C for 72 h. L. reuteri ATCC23272 was grown in De Man, Rogosa, Sharpe (MRS) medium (BD Difco™) at 37 ∘ C overnight. Candida albicans ATCC10231 was grown in Sabouraud broth at 37 ∘ C overnight. Microbial cells were harvested by centrifugation (11,000 ×g for 10 min), and the pellet was washed twice and then resuspended in RPMI 1640 (Cambrex Bio Science). Optical density (OD) measurements were performed at 620 nm to adjust the final concentration of the microbial suspension, and the exact number of colony forming units (CFU) was determined by plating serial dilutions of the inocula onto adapted agar plates (Columbia 5% Sheep Blood Agar, MRS, or Sabouraud). Before being added to the cell samples, the microbial cells were inactivated by exposure to UV for 1 h. The effectiveness of the inactivation was evaluated by plating 20 L of the irradiated inocula on adapted agar plates.

Ethics Statement.
The human cells used in this study were generated from the buffy-coat of healthy volunteers obtained Journal of Immunology Research 3 from the local French blood agency (Etablissement Français du Sang, EFS, Saint-Etienne). It is a statutory requirement that blood donors be given full necessary information (article R.1221-5 of the Public Health Code, 12/01/2009 and 11/06/2006 decrees). Written informed consent was obtained by the EFS from all volunteers involved in our study.

In Vitro Differentiation of Monocyte-Derived Dendritic
Cells. DCs were generated from PBMCs. Briefly, PBMCs were isolated from the buffy-coats of healthy volunteers by Ficoll-Histopaque (Sigma) density gradient centrifugation. PBMCs were washed twice in RPMI 1640 and resuspended at a final concentration of 5 × 10 7 cells per mL of phosphate buffered saline (PBS) supplemented with 2% fetal calf serum (FCS, Biowest-Abcys) and 1 mM EDTA. Monocytes were purified by negative selection using the EasySep Ý Human Monocyte Enrichment Kit, as recommended by the manufacturer (StemCell Technologies). They were then cultured for 5 days in RPMI 1640 supplemented with 1% L-glutamine (Sigma), 10% FCS, and 0.5% penicillin-streptomycin (Sigma), in the presence of 500 U/mL IL4 (R&D systems) and 800 U/mL granulocyte-macrophage colony-stimulating factor (GM-CSF, R&D systems). After 3 days of incubation, one-half volume of fresh culture medium containing 2x concentrations of IL4 and GM-CSF was added to each well.

Electron Microscope Observations.
DCs obtained as previously described were plated in a sterile 12-well plate at a concentration of 1 × 10 6 cells/mL. G. vaginalis was added to wells at a multiplicity of infection (MOI) of 10 for 1 h (Scanning Electron Microscopy, SEM) and at a MOI of 0.01, 1, or 100 for 3 h (Transmission Electron Microscopy, TEM). Cells were harvested, centrifuged (400 ×g for 10 min), rinsed with Natrium Cacodylate (0.2 M pH 7.4) for 10 min, and then fixed at 4 ∘ C overnight with glutaraldehyde 1.6% in Natrium Cacodylate buffer. The samples were then rinsed, postfixed with 1% Osmium Tetroxide (1 h, room temperature), rinsed again, and dehydrated with graded series of ethanol (70 to 100%) and eventually with 100% hexamethyldisilazane. Finally, after overnight drying, samples were placed on a Jeol SEM filter and metallised with carbon (40 s). For TEM, dried samples were embedded in a polymerized 2 mm thick Epon coating, and ultrathin sections were picked up with Formvar-coated copper grids (300 mesh). Sections were counterstained with 4% aqueous uranyl acetate. For negative staining, bacteria were grown overnight in M63B1-0.4% Glu medium and negatively stained with 2% phosphotungstic acid on Formvar-coated copper grids (300 mesh). Images were captured at the Centre d'Imagerie Cellulaire Santé (CICS) of the Université d' Auvergne with a Jeol JSM-6060LV (SEM) and a Hitachi H-7650 (TEM).

Flow Cytometry Analysis of DC Maturation and Viability.
On day 6, the immature DCs from each well were harvested, pooled, centrifuged, and reseeded at 1 × 10 5 cells/mL. UV-killed bacteria were then added at 10 L of suspension per well to reach a final concentration ranging from 10 3 to 10 7 CFU/mL, that is, a MOI between 0.01 and 100. Lipopolysaccharide (LPS) from Escherichia coli (Sigma) at a final concentration of 100 ng/mL was used as positive control. Immature DCs without addition of LPS or bacteria were used as negative control. After 48 h of maturation at 37 ∘ C in a 5% CO 2 atmosphere, DCs were collected, centrifuged, and resuspended in PBS with 1% bovine serum albumin (BSA, Sigma). Cell surfaces were stained with the appropriate fluorescence-labeled murine antibodies: APC-Cy7-conjugated anti-CD14 (LPS coreceptor specific to monocytes), PE-conjugated anti-CD86 (costimulatory molecule, activation marker), V450-conjugated anti-HLA-DR, PerCP-Cy5.5-conjugated anti-DC-SIGN (DC-specific ICAM-3 grabbing-nonintegrin or CD209, member of the CLR family, specific marker of immature DCs), Alexa Fluor Ý 488-conjugated anti-MR (Mannose Receptor or CD206, member of the CLR family, specific marker of immature DCs and macrophages), and streptavidin APC-conjugated anti-TLR4 (biotin antibody, an LPS receptor with activation functions). Antibodies were obtained from BD Biosciences, except anti-MR (Biolegend). Corresponding murine isotypematched and non-labeled antibodies (BD Biosciences or Biolegend) were used as controls. The cells were analyzed by a BD-LSRII flow cytometer with FACSDiva Software (BD Biosciences) at the CICS. Fluorescence compensation adjustments were performed. Gates were set on living DCs based on their forward/side scatter (FSC/SSC) properties. The analysis was halted at a count of 3,000 DCs. The level of staining was expressed as the mean fluorescence intensity (MFI). Culture supernatants were collected and stored at −20 ∘ C until cytokine analysis. To determine cell viability, dye LIVE/DEAD Ý beads (Fixable Blue Dead Cell Stain Kit, for UV excitation Life technologies) were added to cells and used as markers of dead cells. DC mortality was assessed by gating together dead cells and live DCs on SSC/FSC diagram. Two other gates were created from this gate to separate the two populations. From the dead cell gate, cells that expressed LIVE/DEAD marker were considered as effectively dead. This dead cell number was expressed as a ratio of the first created overall gate to obtain a dead DC percentage.

Lymphocyte Proliferation Assays.
The mitogenic response to plant lectins, as phytohemagglutinin A (PHA), is conventionally used to measure cell-mediated immunity in mammals in general and especially in humans [30]. These tests are named lymphocyte proliferation assays or lymphocyte transformation test (LTT). They were performed here to study the functional properties of G. vaginalis and especially its capabilities to modulate PHA-induced T cell proliferation. On day 1, PBMCs collected from a buffy-coat as previously described were adjusted to a concentration of 1 × 10 6 cells per mL of complete medium, that is, RPMI 1640 supplemented with 1% L-glutamine and 10% FCS. The cell suspension was deposited in a sterile 96-well plate (100 L per well). Each measurement was done in triplicate. Polyclonal proliferation of lymphocytes was induced with 2 g/mL of PHA (Sigma). Cells without PHA or bacteria were used as a negative control and cells with PHA without bacteria as a positive one. G. vaginalis was added to other PHA-treated wells to obtain concentrations ranging from 10 3 to 10 7 CFU/mL. After a 72 h incubation at 37 ∘ C under 5% CO 2 , 1 Ci of tritiated thymidine (Perkin Elmer) was added to each well and the cells were further incubated for 4 h. Labeling was stopped by cooling the plate to 4 ∘ C. The cells were then collected under vacuum onto a Whatman filter paper and incorporation of tritiated thymidine was measured using a counter (Tri-Carb 2300TR, Canberra-Packard). Proliferation results were expressed as mean cpm (counts per min) values of triplicate measurements. Identical incubations, however, without addition of tritiated thymidine, were carried out in parallel to collect supernatants for cytokine quantification.

Statistical
Analysis. All data were expressed as means + SD. After variance dispersion test, three different statistical tests were performed. Two-way ANOVA with post hoc Bonferroni test, Friedman's test with Nemenyi's group, and Kruskal-Wallis test with Dunn test were used to analyze the significant effect of bacteria on DCs or PBMCs with XLStat 7.5.2 software (Addinsoft, Paris, France). values lower than 0.05 were considered statistically significant.

Electron Microscopy Observation of G. vaginalis-DC Interaction.
In vivo, the first step of the immune response in the vaginal mucosa corresponds to the interactions between bacteria and immature DCs, which lead to the internalization of bacteria. To mimic this initial phenomenon in vitro, we put in contact DCs with G. vaginalis during 1 to 3 h and took SEM and TEM pictures. SEM produced 3D images of cell and bacteria surfaces after 1 h of contact (Figures 1(a) and 1(b)). G. vaginalis bacteria were rarely found on isolated cell forms (in contrast with what we observed with Lactobacillus, data not shown) and were in general organized in clusters (Figure 1(a)). However, DC dendrites interacted with this cluster of bacteria and sometimes surrounded it ( Figure 1(b)). TEM realized at 3 h, to allow the cells to have time to internalize bacteria, confirmed interactions of DC dendrites with clusters of G. vaginalis (Figures 1(c) and 1(d)). Figure 1(c) shows additionally that the cluster is composed of bacteria enrobed in an extracellular matrix which may interact with DCs. TEM also shows internalized intracellular G. vaginalis bacteria, but only in their isolated form, without intracellular clusters (Figure 1(e)). At a MOI of 0.01 no DCs with internalized G. vaginalis were found, very few at a MOI of 1, and only 35% at a MOI of 100 (count on 100 DCs). For each DC with internalized G. vaginalis, the number of bacteria ranged from 1 to 9 bacteria per cell. This experiment was performed in parallel with L. reuteri at the same concentrations. Internalization of Lactobacillus, unlike G. vaginalis, could be observed on few cells at a MOI of 0.01, a majority of cells at a MOI of 1, and up to 82% of cells at a MOI of 100 (count on over 160 DCs). The number of Lactobacillus internalized in those cells was higher than with that of G. vaginalis since it ranged from 6 to 18 bacteria per cell (data not shown). To conclude on this part, we showed that DCs were able to interact with G. vaginalis and to internalize it, but less effectively than lactobacilli, probably due to G. vaginalis ability to form extracellular clusters.

Flow Cytometric Analysis of Microbial Effects on DC Phenotype.
During the immune response, bacterium internalization can then induce DC activation and maturation, resulting in modifications detectable by cytometry of numerous membrane markers and permitting to distinguish immature DCs from mature ones. In our study, DC activation and maturation were assessed by changes affecting an extensive phenotype of the cell membrane. As expected, immature DCs were characterized by low levels of CD86, HLA-DR, and CD14 expression (the latter was compared to its initial level in monocytes before IL-4 and GM-CSF treatment, data not shown), together with high levels of DC-SIGN and MR expression (Figures 2 and 3). Incubation with low or medium doses of G. vaginalis (10 3 -10 6 CFU/mL, i.e., MOI from 0.01 to 10) did not alter immature DC cell surface phenotype, as shown by the absence of significant change in either membrane marker (Figures 2 and 3). At the highest concentration (10 7 CFU/mL, i.e., MOI 100), G. vaginalis caused a very slight increase in HLA-DR expression (statistically non-significant, Figure 3(b)) combined with moderate decreases in MR and CD14 expressions (both non-significant, Figures 3(e) and 3(f)), which reflect early signs of maturation. However, the lack of increase in CD86 expression and decrease in DC-SIGN and TLR4 expressions indicated that this maturation process was incomplete.
Taken together, our overall cytometric results indicate that G. vaginalis induced no maturation of DCs or an incomplete DC maturation process at high doses.
Unlike G. vaginalis, L. reuteri and C. albicans induced a clear dose-dependent higher expression of CD86 and HLA-DR upon maturation of DCs (Figures 3(a) and 3(b)). Combined with decreased expressions of DC-SIGN, MR, CD14, and TLR4 (Figures 3(c), 3(d), 3(e), and 3(f)), these results show that the last two microorganisms induced fully mature DCs, similarly to LPS extract.    this assertion, a viability marker was added during flow cytometric experiments. Immature DCs alone had almost no mortality. In comparison, exposure of DCs to LPS extract induced a slight, but significant, increase in mortality ( Figure 4). Three concentrations of G. vaginalis were tested to analyze their impact on DC survival. Results showed that, from 10 3 to 10 7 CFU/mL (MOI 0.01 to 100), G. vaginalis induced no increase in dead DCs.

Cytokine Secretion by Microbial-Matured DCs. In vivo,
mature DCs migrate to secondary lymphoid organs and present the antigen to T cells. During this interaction, DCs deliver three signals, notably a cytokinic signal polarizing the differentiation of lymphocytes into several subpopulations. To decipher the preferential pathway of T cell polarization induced by G. vaginalis, four cytokines assumed to be of particular interest were selected for ELISA measurements. Compared to untreated immature DCs, DCs incubated with G. vaginalis did not significantly increase the production of TNF-or IL-10, except at the upper dose of 10 7 CFU/mL (Figures 5(a) and 5(b)). Additionally, the production of IL-12p70 and IFN-was not or barely detectable in G. vaginalistreated or immature DCs (Figures 5(c) and 5(d)). Conversely, a strong dose-dependent increase in the production of TNFand IL-10 was induced by L. reuteri and C. albicans and, to a lesser extent, of IL-12p70 by L. reuteri and of IFN-by C. albicans (Figures 5(a), 5(b), 5(c), and 5(d)). The fold changes determined by comparing the level of cytokines produced by DCs exposed to 10 7 CFU/mL G. vaginalis (i.e., a MOI of 100) to that produced by immature DCs were equal to 8.9 for IL-10, 6.5 for TNF-, and only 1.1 and 1.3 for IL-12p70 and IFN-, respectively. By comparison, the fold changes induced by the same dose of L. reuteri or C. albicans ranged from 70 to 307 for IL-10 and TNF-production. Whatever the concentration of G. vaginalis, the level of DC-produced cytokines attained only 10 to 28% of that generated by LPS-treated DCs. Overall, these results show that G. vaginalis barely induces cytokinic secretion, in accordance with our findings on DC maturation (Figure 3

Slight Increase in PHA-Stimulated Lymphocyte Proliferation by G. vaginalis.
To investigate lymphocyte activation, the next classical stage of the immune response, we carried out functional tests of bacterium-induced modulation of lymphocytic proliferation, using a model of PHA-induced T cell proliferation. Addition of the strain to the medium caused a slight dose-dependent increase in PHA-stimulated lymphocyte proliferation in comparison to lymphocyte control assays without bacteria ( Figure 6). The increase in proliferation was only significant at a dose of 10 7 CFU/mL ( < 0.001) and attained an upper average value of 20% compared to that of PHA-stimulated control cells. Lower concentrations of G. vaginalis did not cause any significant modulation of PHAstimulated lymphocyte proliferation. Similar experiments we performed in parallel showed that L. reuteri and C. albicans did not induce a similar pattern (data not shown).

G. vaginalis-Dependent Increase in Cytokine Secretion by PHA-Stimulated Leukocytes.
To depict the type of immune response involved during the G. vaginalis-dependent increase in PHA-stimulated lymphocyte proliferation, secretion of cytokines by the four main subpopulations of T cells (IFN-, IL-4, IL-17A, and IL-10 corresponding to Th1, Th2, Th17, or Treg, resp.) or by APCs (IL-12p70 and TNF-) was measured in extracellular media of lymphocyte proliferation assays. PHA (2 g/mL) alone induced a strong secretion of IFN-, IL-17A, IL-10, IL-12p70, and TNF-from PBMCs (Figure 7). A clear dose-dependent increase in IFN-and IL-17A production was observed in PHA-stimulated PBMCs exposed to varying concentrations of G. vaginalis (Figures  7(c) and 7(d)). As shown previously for PHA-stimulated proliferation, only the highest dose of G. vaginalis tested induced a significant augmentation in the production of these two cytokines, compared to the PHA control without bacteria. For TNF-, IL-12p70, and IL-10, the increases were non-significant, even at high doses (Figures 7(a), 7(b), and 7(e)). Compared to the control conditions without bacteria, exposure to 10 7 CFU/mL G. vaginalis induced fold increases of 4.9, 4.0, 3.0, 2.2, and 2.2 for IFN-, IL-10, TNF-, IL-17A, and IL-12p70 production, respectively. Contrastingly, the secretion level of IL-4 was barely measurable or undetectable at any dose of G. vaginalis (not shown). Furthermore, at their highest concentration, L. reuteri and C. albicans caused an increase in cytokine secretion that, depending on the cytokine, was 4-to 5-fold higher than that caused by G. vaginalis (data not shown). Overall, these data show that G. vaginalis can induce a slight dose-dependent secretion of cytokines on both the inflammatory (IFN-, TNF-, IL-17A, and IL-12p70) and anti-inflammatory (IL-10) sides.

Discussion
DCs, the main sentinels of the immune system, are abundant in the human vaginal mucosa, both in the epithelium and in the lamina propria [21]. In this mucosal area, DCs can interact with luminal microorganisms, either indirectly via epithelial transport mechanisms or directly via dendrites extended across epithelial cells that take up bacteria from the vaginal lumen [31,32]. As BV is characterized by an imbalance of the normal H 2 O 2 -producing Lactobacillus flora toward a polymorphic anaerobic flora with predominant G. vaginalis, it is likely that the PRRs of vaginal DCs are subsequently affected by this alteration. To gain insight into the pathophysiological role of G. vaginalis in BV, we decided to characterize the interactions between this bacterium and DCs.
We first carried out an electron microscope study of bacteria-DC interactions, the first stage in the immune response, to determine whether G. vaginalis can internalize DCs. This was confirmed by using TEM, but we observed that the bacteria were very sparsely represented in free form in culture medium and rather formed clusters coated in an extracellular matrix. As compared to a Lactobacillus species which remained in free form, G. vaginalis was internalized by DCs much less efficiently. It can reasonably be assumed that this difference is related to the propensity of G. vaginalis to form clusters. In addition, G. vaginalis forms, in vivo, biofilms in the vagina, a process that is involved in BV pathogenesis [33]. These clusters might be the beginning of biofilm formation. Biofilms reduce the host-immune response by decreasing bacteria internalization owing to their large size and to the fact that their extracellular matrix can prevent antigen recognition by APCs [34]. This conformation might thus allow bacteria like G. vaginalis to be less internalized by DCs comparatively to strains like Lactobacillus that are unable to form biofilms.
We then studied the impact of internalization of G. vaginalis on DC maturation status. The maturation state of DCs is a critical determinant of the balance between their Journal of Immunology Research  Figure 5: Cytokine production by human DCs exposed to a range of G. vaginalis, L. reuteri, or C. albicans concentrations. (a, b) TNF-and IL-10 cytokine production. G. vaginalis induceda slight increase in IL-10 secretion at high doses but no significant production of TNF-. (c, d) IL-12p70 and IFN-cytokine production. G. vaginalis induced no significant production of the 2 cytokines, even at the highest bacteria concentrations. Data are means (+ SD) of measurements from 6 independent experiments, except for C. albicans (3 experiments). Imm-DCs: immature DCs; LPS-DCs: DCs matured by a 48 h exposure to 100 ng/mL E. coli LPS; G. vaginalis, L. reuteri, and C. albicans: DCs matured after a 48 h incubation at different concentrations (10 3 to 10 7 CFU/mL, i.e., MOI = 0.01 to 100) of G. vaginalis, L. reuteri, and C. albicans, respectively. Friedman's test with Nemenyi's comparison * < 0.05, * * < 0.01, and * * * < 0.001 as compared to the Imm-DCs.
tolerogenic and immunogenic abilities [35]. Immature or semi-mature DCs generally promote tolerogenic responses whereas mature DCs promote immunogenic responses [36]. Flow cytometry analysis showed that, whatever bacterial concentrations, G. vaginalis elicited minimal changes in the DC membrane phenotype, thus inducing a very incomplete maturation of human DCs. Concurrently, cytokine production remained very low compared to that from LPSinduced fully mature DCs or L. reuteri-or C. albicansmatured DCs, even at high doses of microorganisms. Taken together, our cytometric and cytokinic data characterize a non-inflammatory DC response at low G. vaginalis doses and a very slight pro-inflammatory DC response at the highest concentration. G. vaginalis concentrations interacting  in vivo with mucosal DCs in the human genital tract have not been widely assessed. However, in women with BV, the number of bacteria present can be equal or superior to 10 8 per mL of vaginal fluid, including about 10 7 CFU/mL of G. vaginalis [37,38]. Although unknown, the actual number of bacteria in direct contact with vaginal DCs within the mucosa is certainly much lower than this number. Thus, the concentrations of G. vaginalis interacting with vaginal DCs in the mucosa of women with BV probably correspond to the low or intermediate MOIs used in our model and causing no or very little DC maturation. The results obtained with G. vaginalis were compared to those obtained with two other microorganisms potentially present in the vaginal mucosa, a commensal bacterium (L. reuteri) and a pathogenic yeast responsible for mycotic vaginitis (C. albicans). Each induced a clear-cut maturation of the DCs, similar to that we previously observed with other pathogens and probiotic strains [39,40] but in strong contrast to the G. vaginalis DC response. In our model, G. vaginalis did not induce DC mortality, unlike the two other microorganisms (data not shown). Overall, our findings show that G. vaginalis is slightly pro-inflammatory, but less than the Lactobacillus strain we used as control.
The effects of G. vaginalis on DCs observed in our study were slighter than those reported in other studies, which showed activation and maturation of human moDCs when exposed to the mucosal fluid of women with BV [23,24]. In these previous studies, DCs were exposed to a mixture of numerous bacterial products secreted by the characteristic polymorphic BV flora and to molecules produced by the mucosal immune system of women with BV. By contrast, in our model, DCs were placed in the presence of G. vaginalis alone to find new evidence of its role as a putative BV etiological agent. Thus, the absence of or the very slight G. vaginalisinduced DC maturation by in vitro direct contact is not inconsistent with a DC maturation induced by substantially secreted products of the overall BV flora, like, for example, the LPS of Prevotella bivia. In the vaginal mucosa, G. vaginalis could have local specific modulatory effects on immune cells, independently of an overall effect on the maturation of BV flora.
We next performed functional tests of lymphocytic proliferation using a model of PHA-stimulated PBMCs to investigate the activation of lymphocytes, the following stage in the classical immune response. We observed no significant variation in lymphocyte proliferation at all doses of G. vaginalis that we used, except the highest one where a significant increase was measured. These findings are consistent with the variations observed in DC status and confirm that G. vaginalis induces very few immunologic effects at low doses and a slight pro-inflammatory response at the highest concentrations, which are unlikely to be encountered in vivo. This profile of the immunological response to G. vaginalis could explain the characteristic lack of external inflammatory signs during BV, in contrast with bacterial or mycotic vaginitis, despite increased pro-inflammatory TLR2 and TLR4 signaling and IL-1 secretion reported elsewhere [41][42][43]. In light of our results, we hypothesize that, depending on its actual amount in contact with the mucosal immune cells, G. vaginalis could either go unnoticed or induce low-grade inflammation in vivo.
To explore the mechanism of the slight G. vaginalisinduced increase in lymphocyte proliferation, a large panel of cytokines was measured in the cell supernatants including molecules secreted by APCs (IL-10, IL-12p70, and TNF-) and/or by T cells (IFN-, IL-4, IL-17A, and IL-10). We evidenced a similar G. vaginalis dose-dependent profile of secretion for all these cytokines, except for IL-4, which remained undetectable. The higher the dose of the pathogen was, the stronger the PHA-stimulation of cytokine secretion was. Thus, this cytokine secretion profile is clearly related to that of the proliferation of PHA-stimulated PBMCs, irrespective of the anti-or pro-inflammatory nature of the cytokines. With regard to the response to G. vaginalis of the four main subpopulations of T cells, cytokine secretion suggests a clear dose-dependent induction of Th1 (IFN-, IL-12p70) and Th17 (IL-17A) and Tregs (IL-10), but not of Th2 (IL-4). Thus, Th1, Th17, and Tregs could be involved in the immune response to high doses of G. vaginalis. This topic deserves further investigation in future studies of T cell polarization.
Taken together, our findings show that G. vaginalis, by forming clusters and reducing the internalizing ability of DC, induces a slight immunological host response including maturation of DCs, lymphocyte proliferation, and pro-Th1, pro-Th17, and pro-Tregs cytokine production. These immunomodulatory properties are consistent with the atypical clinical profile of BV, which is characterized by a lowgrade inflammatory process, as we observed in our in vitro model at the highest dose of the bacterium. Our results Compared to the PHA-stimulated control cells, a significant and dose-dependent increase in cytokine secretion was observed for IL-17a and IFN-with the G. vaginalis-treated cells. Other cytokines showed no significant increase in secretion. Data are means (+ SD). PBMC: PBMCs cultured for 72 h without any effector; PHA control: PBMCs only exposed to PHA (2 g/mL) for 72 h; G. vaginalis: PBMCs exposed for 72 h to PHA (2 g/mL) in the presence of 10 3 to 10 7 CFU/mL of G. vaginalis. Friedman's test with Nemenyi's comparison ( < 0.05). A, B, and C: means with different superscript letters are significantly different from each other.
show the potential immunological effects of the bacteria of the vaginal flora and suggest the existence of a mechanism whereby BV and its associated changes in flora composition could affect host vaginal immunity. Finally, these results lend weight to the putative role of G. vaginalis in the pathophysiology of BV and open up broader prospects, in particular for the understanding of the contribution of local immunological alterations to the increased risk of STIs in women with BV.