Noncanonical dendritic cell differentiation and survival driven by a bacteremic pathogen

Oral bacteremic pathogen, P. gingivalis, drives a non-canonical route of dendritic cell differentiation, and promotes survival of its protective dendritic cell niche.

Abstract

Maintenance of blood DC homeostasis is essential to preventing autoimmunity while controlling chronic infection. However, the ability of bacteremic pathogens to directly regulate blood DC homeostasis has not been defined. One such bacteremic pathogen, Porphyromonas gingivalis, is shown by our group to survive within mDCs under aerobic conditions and therein, metastasize from its oral mucosal niche. This is accompanied by expansion of the blood mDC pool in vivo, independently of canonical DC poietins. We presently know little of how this bacteremic pathogen causes blood DC expansion and the pathophysiological significance. This work shows that optimum differentiation of MoDCs from primary human monocytes, with or without GM-CSF/IL-4, is dependent on infection with P. gingivalis strains expressing the DC-SIGN ligand mfa-1. DC differentiation is lost when DC-SIGN is blocked with its ligand HIV gp120 or knocked out by siRNA gene silencing. Thus, we have identified a novel, noncanonical pathway of DC differentiation. We term these PDDCs and show that PDDCs are bona fide DCs, based on phenotype and phagocytic activity when immature and the ability to up-regulate accessory molecules and stimulate allo-CD4⁺ T cell proliferation when matured. The latter is dependent on the P. gingivalis strain used to initially “educate” PDDCs. Moreover, we show that P. gingivalis-infected, conventional MoDCs become resistant to apoptosis and inflammatory pyroptosis, as determined by levels of Annexin V and caspase-8, -3/7, and -1. Taken together, we provide new insights into how a relatively asymptomatic bacteremia may influence immune homeostasis and promote chronic inflammation.

Introduction

DCs are the major sentinels of the immune system [1] and the key cells bridging innate and adaptive immune responses [2]. The ability of DCs to recognize danger signals and activate the adaptive immune response, as well as provide tolerance to self-antigens [3], is a major facet of immune homeostasis [4]. A balance of environmental cues regulates these processes in the steady state and inflammatory states [5]. Disruption of this balance impairs the immune response and promotes chronic inflammatory diseases [6].

Central to many chronic inflammatory diseases are fluctuations of the blood DC pool and the dysfunction of resultant DCs. These include autoimmune diseases [7], rheumatoid arthritis [8], coronary artery disease [9], and CP [10]. We previously showed in CP subjects an elevation of blood mDCs without a commensurate rise in serum levels of canonical DC poietins and DC growth factors. Microbes, including CMV [11], pneumococcal infections [12], and the “keystone” [13] mucosal pathogen P. gingivalis [10, 14–18], are reported to enter the bloodstream and invade DCs. We hypothesized that bacteremia with P. gingivalis directly stimulates differentiation of mDCs from progenitors and results in blood DC expansion.

Of particular relevance to this question are the PRRs on DCs, including C-type lectins, TLRs, and nucleotide-binding oligomerization-like receptors [19]. Various pathogens involved in chronic infections target the C-type lectin DC-SIGN [20] for survival and immunosuppression [21, 22]. P. gingivalis, in particular, disrupts host homeostasis by various mechanisms, such as inhibition of chemotaxis and processing of ROS [23]. In regards to the present work, this obligate anaerobe binds to DC-SIGN on inflamed tissue mDCs in situ [10] and MoDCs in vitro [24] by its mfa-1. DC-SIGN routes P. gingivalis into passive intracellular compartments, where it survives under aerobic conditions [10] and inhibits Th1 effector responses [25]. Cross-linking DC-SIGN has been described to promote DC differentiation from progenitors [26]. Canonical pathways to DC differentiation, involving cytokines and DC-poietin signaling pathways and direct bone marrow stromal interaction, have been well-studied [27], but the direct role of infection is presently unclear.

The present study investigated the ability of the bacteremic pathogen P. gingivalis and its isogenic mutant strains DPG/minor mfa-1 fimbria to infect DC progenitors and induce differentiation into bona fide DCs. We show that indeed P. gingivalis induces differentiation of monocytes into what we term PDDCs, and furthermore, these PDDCs are highly phagocytic when immature and immunostimulatory when matured, although the latter depends on the P. gingivalis strain, to which DC progenitors were “educated”. Also important to the regulation of chronic inflammation is the routine apoptosis of DCs [28]. As it is established that DC-SIGN ligation disrupts DC function and promotes differentiation, we investigated whether this interaction also inhibited apoptosis of DCs and could promote survival of DCs that harbor P. gingivalis. Hence, we further show that this low-grade infection of DCs extends their lifespan and renders them less inflammatory, as determined by reduced apoptosis and pyroptosis, respectively. We conclude that in the setting of low-grade bacteremia, DCs that are commandeered by microbial pathogens [29] in this manner and become more resistant to apoptosis may become detrimental to the host.

MATERIALS AND METHODS

Bacterial culture growth conditions, bacterial labeling, and MoDC infection

WT Pg381, which expresses minor and major fimbriae (Pg min⁺/maj⁺), isogenic major fimbria-deficient mutant DPG-3, which expresses only the minor fimbriae (Pg min⁺/maj⁻), and the double-fimbriae mutant MFB (Pg min⁻/maj⁻) were maintained anaerobically (10% hydrogen, 10% carbon dioxide, and 80% nitrogen) in a Coy Laboratory vinyl anaerobic system glove box at 37°C [30] in Acumedia Wilkins-Chalgren anaerobe broth. Erythromycin (5 μg/ml) and tetracycline (2 μg/ml) were added according to the selection requirements of the strains [31]. Bacteria were washed once in PBS and stained with CFSE (eBioscience, San Diego, CA, USA), as described for flow cytometry [32]. Briefly, bacteria in PBS were stained with CFSE at a final concentration of 5 μM for 30 min at 37°C and protected from light. The bacterial suspension was washed five times in PBS, and Pg was resuspended to an OD at 660 nm of 0.11, which was determined previously to be equal to 5 × 10⁷ CFU [33]. MoDCs were pulsed with Pg strains at a MOI [34] of 0.1, 1, and 10 for 24 h (Supplemental Fig. 1). Low MOI values were used to mimic a natural blood mDC infection observed in CP patients, as well as to avoid overwhelming the host response. In select experiments to prevent bacterial uptake, cells were treated for 1 h with cytochalasin D (Enzo BML-T109) at a final concentration of 2 μM before infection.

MoDC culture and phenotype

The Committee on Research Involving Human Subjects Research at Stony Brook University approved all protocols involving human subjects. Informed consent was obtained from all healthy volunteers before commencement of the study. Conventional MoDCs were generated, as we have described previously [35, 36]. Briefly, monocytes were isolated from mononuclear fractions of the peripheral blood of healthy donors and seeded in the presence of GM-CSF (100 ng/ml) and IL-4 (25 ng/ml) at a concentration of 1–2 × 10⁵ cells/ml in RPMI 1640 (cellgro; Mediatech, Manassas, VA, USA) containing 10% heat-inactivated FBS (Lonza, Walkersville, MD, USA) and antibiotic/antimycotic (Thermo Scientific HyClone, Logan, UT, USA) for 5–7 days, after which, flow cytometry was performed to confirm the immature DC phenotype (CD14^loCD83⁻CD1c⁺DC-SIGN⁺). Cell-surface markers of DCs were evaluated by four-color immunofluorescence staining with the following mAb: CD83 (11-0839-42; eBioscience), CD14 (17-0149-42; eBioscience), CD209 (45-2099042; eBioscience), and CD1c (130-090-507 MACS; Miltenyi Biotec, Auburn, CA, USA).

Pathogen-differentiated MoDCs

Monocytes were isolated from total PBMCs using RoboSep automated positive selection of CD14⁺ cells (18058; Stemcell Technologies, Vancouver, BC, Canada). Monocytes were then split evenly in six-well plates and infected immediately with MOI = 1 of DPG-3, MFI, or MFB strains of P. gingivalis, in the presence or absence of GM-CSF and IL-4. Initial infection was considered as Day 0, and differentiation was determined at each time-point after by collecting and analyzing scattergraph characteristics and expression of CD14 and DC-SIGN by flow cytometry. For siRNA knockdown of DC-SIGN, 10 μM DC-SIGN-targeting siRNA (s26944 Ambion Silencer Select, Life Technologies, Grand Island, NY, uSA) or an off-target, scrambled sequence was transfected into monocytes using Lipofectamine RNAiMAX (Invitrogen, Life Technologies), diluted in Opti-MEM I reduced serum medium (31985-062; Invitrogen, Life Technologies) at the time of infections.

Phagocytic capacity of pathogen-differentiated MoDCs

MoDCs were generated from CD14⁺ monocytes using all of our P. gingivalis strains as in the initial differentiation experiments above. After 24 h of infection with each respective strain, these DCs, along with uninfected controls in the presence or absence of GM-CSF and IL-4, were infected with a subsequent dose of CFSE-labeled (eBioscience) P. gingivalis. The amount of phagocytosis was measured as the percentage of CFSE-positive MoDCs at 6 and 24 h by flow cytometry. Additionally, these cells were measured for Annexin V-PE (eBioscience) staining at 24 h postsecondary infection. All infections were at a MOI = 1. To determine phagocytic uptake capacity after exposure to a pathogen different than P. gingivalis, fluorescent Escherichia coli particles (pHrodo phagocytosis kit; Invitrogen, Life Technologies) were incubated with PDDC and analyzed for presence of both CFSE–P. gingivalis and red fluorescent E. coli by flow cytometry.

CD4⁺ T cell proliferation

PDDCs, generated as above, were collected at 24 h postinfection or after a subsequent 24 h-induced maturation with E. coli LPS and TNF-α. CD4⁺ T cells were isolated using RoboSep automated negative selection (#19052; Stemcell Technologies) and stained with CFSE, according to the manufacturer's protocol (eBioscience). MoDCs (stimulators) were incubated with T cells (responders) at a ratio of 1:3 and 1:9, respectively, for 5 days. Cells were collected, and proliferation was measured by reduction of CFSE by flow cytometry.

Annexin V and caspase staining of MoDCs

Immature MoDCs were infected with strains of P. gingivalis, as above, for 24 h. Cells were analyzed using an Acuri C6 cytometer for Annexin V-FITC and PI (BMS500FI/300; eBioscience), as well as Annexin V-PE (eBioscience) and CFSE-labeled P. gingivalis. Intracellular polycaspase, initiator caspase-3/7, and effector caspase-1 and -8 were determined using Vybrant FAM FLICA kits (V3511-7, -8, -9; Molecular Probes, Life Technologies) and FLICA 660 Caspase-1 kit (9122; ImmunoChemistry Technologies, Bloomington, MN, USA), according to the manufacturer's instructions.

Flow cytometry gating and statistical analysis

Gates were chosen using suspension cell fsc and ssc characteristics. Nonadherent MoDCs typically fall within fsc-h 300–600 and ssc-h 100–200 (×1000). Undifferentiated monocytes, lymphocyte carryover, and debris were excluded from analysis. Monocytes displayed 95–99% differentiation into MoDCs by Day 5 of culture. For statistical analyses of expression within gates, experimental samples were compared with untreated controls or between other experimental samples using unpaired Student's t-tests of the means from at least three independent flow experiments. Statistical significance was calculated with the Holm-Sidak method with an α of 0.05 using GraphPad Prism 6 and is denoted by asterisks within figures as *P ≤ 0.05; ¶P ≤ 0.01; and TP ≤ 0.001.

RESULTS

mfa-1⁺ P. gingivalis drives generation of PDDCs

In addition to the minor mfa-1 fimbriae, WT Pg381 also expresses fimA, which targets CXCR4/TLR2 [37]. We have shown previously that WT Pg381 positively influences DC differentiation [10], but the specific role of the fimbria was not addressed. Primary human monocytes were therefore infected with isogenic mutant strains of P. gingivalis with defined defects in mfa-1 and fimA (Figs. 1 and 2). Initial dose-response studies established that a MOI = 1 at the upper limit of the MOI range observed in blood mDCs of CP subjects in vivo [10] was optimum for DC differentiation, as determined by down-regulation of CD14 and up-regulation of DC-SIGN (Fig. 1). Our results show that the mfa-1⁺fimA⁻ isogenic mutant strain (i.e., DPG-3) is most potent in induction of DC differentiation (Fig. 1A and B). DPG-3 led to a significant increase of DC-SIGN (Fig. 1A) and a significant loss of CD14 (Fig. 1B) at 1 h and 24 h compared with control MoDCs. In contrast, the mfa-1⁻fimA⁻ strain (i.e., MFB) or the DC-SIGN ligand HIV-1 gp120 did not induce DC differentiation over controls (Fig. 1A and B). These results were consistent over a 72-h time-course of infection, and supplementation with GM-CSF/IL-4 conferred a synergistic effect on DC differentiation induced by mfa-1⁺ P. gingivalis (Supplemental Fig. 1). As mfa-1⁺ P. gingivalis caused DC expansion within 6 h, we examined whether fimA had a role in DC expansion by using the mfa1⁻fimA⁺ strain (i.e., MFI) but found that mfa1⁻fimA⁺ P. gingivalis failed to cause rapid expansion of DCs (Fig. 2A and B). Pretreatment of monocytes with HIV-1 gp120 to block uptake of mfa-1⁺ P. gingivalis strains by MoDCs blocked DC expansion in response to mfa-1⁺ P. gingivalis infection (Fig. 2A and B). The addition of gp120 did not prevent antibody detection of DC-SIGN on MoDCs (Supplemental Fig. 2).

Figure 1. P. gingivalis minor fimbriae-mediated invasion drives optimal DC differentiation.

CD14⁺ monocytes were treated with P. gingivalis strains DPG-3 (mfa-1⁺) and MFB (mfa1⁻fimA⁻) or the DC-SIGN ligand HIV gp120. Cells were analyzed by flow cytometry to determine differentiation by loss of CD14 and gain of DC-SIGN. (A) The DPG-3-treated cells show an up-regulation of DC-SIGN significantly higher than conventional MoDCs at 1 h and 24 h, where the other treatments do not. Monocytes pretreated with gp120 show a significant decrease in DC-SIGN at 24 h. (B) DPG-3-treated monocytes significantly down-regulate the monocyte marker CD14 compared with conventional MoDCs at 1 h and 24 h, whereas other treatments cause a down-regulation of CD14 at 1 h but no significant decreases at 24 h. TP ≤ 0.001; ¶P ≤ 0.01; *P ≤ 0.05.

Figure 2. P. gingivalis mfa-1 fimbriae, not fimA, drive rapid DC differentiation.

(A and B) CD14⁺ monocytes were treated with DPG-3 (mfa-1⁺fimA⁻) and MFI (mfa-1⁻fimA⁺) or the DC-SIGN ligand gp120 prior to DPG-3 infection. Cells were analyzed by flow cytometry to determine differentiation by loss of CD14 and gain of DC-SIGN. (A) Only DPG-3-treated monocytes display a significant increase of DC-SIGN. MFI does not drive a rapid increase of DC-SIGN expression, and blocking DC-SIGN prevents DPG-3 from causing a significant up-regulation of DC-SIGN. (B) MFI causes less CD14 down-regulation than DPG-3 at 1 h, and MFI and gp120 treatments have significantly higher CD14 expression at 6 h. (C and D) Monocytes are treated with DC-SIGN-targeting siRNA or scrambled siRNA fragments simultaneously with bacterial infections. (C) Disruption of DPG-3 uptake by siRNA at 6 h prevents up-regulation of DC-SIGN driven by DPG-3. At 24 h, only DPG-3-infected or DPG-3 and off-target (OT) controls show a significant increase in DC-SIGN expression. (D) Only DPG-3-treated monocytes have significantly decreased CD14 at 6 h, and all DPG-3-infected samples show a significant decrease of CD14 by 24 h. TP ≤ 0.001; *P ≤ 0.05.

To further confirm the role of DC-SIGN-targeting by P. gingivalis in DC differentiation, monocytes were treated with DC-SIGN-targeting siRNA simultaneously with bacterial infection (Supplemental Fig. 3A). The presence of DC-SIGN-specific siRNA lowered DPG-3-infection rates, dampened CD14 down-regulation (Fig. 2C), and significantly prevented DC-SIGN expression (Fig. 2D). Scrambled siRNA had no effect on differentiation. We analyzed expression levels of another DC phenotypic marker, CD1a, during siRNA transfections. Surface CD1a levels were similar to untreated monocytes when DPG-3 uptake was inhibited with DC-SIGN-targeting siRNA, but the presence of GM-CSF/IL-4 resulted in immature CD1a⁺ MoDCs regardless of the presence of DC-SIGN siRNA (Supplemental Fig. 3B).

PDDCs have enhanced antigen-capture function

We next explored the functionality of PDDCs to determine if they were physiologically relevant mDCs. To determine if phagocytic uptake and antigen-capture activity were intact, PDDCs were generated using CFSE-labeled P. gingivalis strains and then were pulsed with red fluorescently labeled, inactivated E. coli. Monocyte and MoDC rates of E. coli phagocytosis are shown compared with the rates of E. coli phagocytosis of CFSE-positive PDDCs. The mfa-1⁻ (MFI)-generated PDDCs had a significantly higher level of antigen-capture activity than conventional MoDCs but displayed lower antigen-capture activity compared with PDDCs generated by mfa-1⁺ P. gingivalis (WT, DPG-3) or nonfimbriated MFB (Fig. 3). Whereas MFB-generated PDDCs display potent phagocytic activity, this remains a very small population of cells, as MFB is not an effective inducer of differentiation, nor is it highly taken up by cells (Supplemental Fig. 4).

Figure 3. PDDCs have intact antigen-capture function.

CD14⁺ monocytes were treated with CFSE-labeled mutant strains of P. gingivalis for 24 h to generate PDDC. Inactivated, whole E. coli were fluorescently labeled and used to measure phagocytosis of CFSE-positive PDDC by flow cytometry. All of the PDDCs were highly phagocytic compared with conventional MoDCs. MFI-generated PDDCs were less phagocytic than the other treatments but still significantly higher than conventional MoDCs. TP ≤ 0.001; ¶P ≤ 0.01.

PDDCs capable of maturation and immunostimulation but dependent on pathogen “education”

As PDDCs generated from mfa-1⁺ P. gingivalis were highly phagocytic, we hypothesized that they were phenotypically immature. We thus wanted to determine the accessory molecule expression and allo-CD4⁺ T cell-stimulatory capacity of PDDCs after forced maturation using TNF-α/E. coli LPS cocktails (Figs. 4 and 5). We observed that PDDCs were generally immature DCs with low accessory molecule expression, albeit mfa-1⁻ strains (MFI, MFB) yielded DCs with higher CD83 expression (Fig. 4A). The PDDC groups also had a significant increase in their surface CD40 expression compared with immature MoDCs (Fig. 4A). All PDDC groups were capable of up-regulation of accessory molecule expression upon maturation with TNF-α/E. coli LPS, although CD83 and CD86 expression remained significantly lower than in matured MoDCs (Fig. 4B). The PDDC groups, with the exception of MFI, again had significantly elevated CD40 surface expression from MoDC controls (Fig. 4B). PDDCs were then cocultured with allogeneic, CFSE-labeled CD4⁺ T cells. Proliferation of CD4⁺ T cells after 5 days was measured by loss of the CFSE signal and was reported as the percentage of input CD4⁺ T cells proliferating. mfa-1⁺fimA⁻ P. gingivalis-differentiated DCs (DPG-3) displayed a significant decrease in their ability to induce T cell proliferation when matured, whereas mfa-1⁺fimA⁺ P. gingivalis- and mfa1⁻fimA⁺ P. gingivalis-differentiated DCs (WT Pg381, MFI) did induce T cell proliferation when matured (Fig. 5).

Figure 4. PDDCs are immature, but accessory molecule expression can be induced.

PDDCs were generated at 24 h and measured by flow cytometry for the accessory molecules CD83, CD86, CD40, and HLA-DR. (A) mfa-1⁺ PDDCs (Pg381, DPG-3) displayed a low, natural expression of accessory molecules. mfa1⁻ PDDCs had significantly higher levels of CD83, whereas all PDDC groups had significantly elevated CD40 surface expression. (B) When PDDCs were subsequently matured, CD83 and CD86 surface expression was significantly lower in all of the infected groups compared with mature MoDCs. All PDDC groups had significantly elevated CD40 surface expression after maturation compared with MoDCs. ¶P ≤ 0.01; *P ≤ 0.05.

Figure 5. PDDCs have an immunostimulatory capacity dependent on education.

CFSE-labeled CD4⁺ T cells were incubated for 5 days with PDDCs. The ratios of MoDCs to T cells used were 1:3 (left bar of each pair) and 1:9 (right bar of each pair). An allo-CD4⁺ T cell proliferation shows that PDDCs are efficient in stimulating CD4⁺ T cell proliferation after forced maturation with a cocktail of E. coli LPS/TNF-α, relative to matured, conventional MoDCs. The exception was mfa1⁺fimA (e.g., DPG-3)-differentiated DCs. ¶P ≤ 0.01.

mfa-1⁺ P. gingivalis rescues conventional MoDCs from apoptosis

As DC functions, including DC survival and apoptosis, are linked to immune homeostasis, we assessed the apoptosis rates of conventional, immature MoDCs in response to a MOI = 1 of mfa1⁺fimA⁺ WT Pg381 (Fig. 6). Whereas resting state MoDCs, in the presence of growth factor cocktail GM-CSF/IL-4, were viable out to 7 days, the absence of GM-CSF/IL-4 led MoDCs to display high rates of apoptosis by 24 h, as determined by Annexin V staining (Fig. 6A). The addition of inflammatory survival signals, including E. coli LPS/TNF-α, and of P. gingivalis strains promoted longevity of conventional MoDCs in the absence of GM-CSF/IL-4. In response to P. gingivalis, decreased Annexin V staining was mfa-1-dependent (Fig. 6A). mfa-1⁺ P. gingivalis strains (WT Pg381, DPG) provided significant decreases in Annexin V staining, whereas mfa-1⁻ P. gingivalis strains (MFB, MFI) did not (Fig. 6A). Prior blocking of DC-SIGN with HIV-1 gp120 restored DC apoptosis rates (Fig. 6A), emphasizing the role for DC-SIGN-targeting mfa-1 in rescue of DCs from apoptosis. The prevention of uptake of WT Pg381 by cytochalasin D treatments led to a restoration of natural apoptosis rates in a dose-dependent manner (Fig. 6B).

Figure 6. Infection of conventional MoDCs by mfa-1⁺ P. gingivalis reduces apoptosis.

MoDCs were washed and removed from the presence of growth factors after Day 5 of culture. MoDCs were treated as indicated for 24 h and analyzed by flow cytometry for Annexin V outer-membrane staining. (A) Untreated MoDCs lacking stimulation show high rates of apoptosis at 24 h, whereas mfa-1⁺-infected MoDCs (WT Pg381, DPG-3) or MoDCs treated with E. coli (Ec) LPS/TNF-α have significantly lower rates of Annexin V staining. mfa-1⁻ strains (MFB, MFI) or heat-killed (HK) WT Pg381 do not drive reduced rates of apoptosis upon infection. Further, pretreatment of MoDCs with gp120 to block DC-SIGN significantly prevents reduction of apoptosis. (B) MoDCs were pretreated with cytochalasin D (CytoD) to prevent bacterial uptake. The prevention of uptake led to significantly higher rates of MoDC apoptosis at the indicated MOI. All infections were significantly lower in Annexin V positivity compared with uninfected MoDCs. TP ≤ 0.001; ¶P ≤ 0.01; *P ≤ 0.05.

mfa-1⁺ P. gingivalis rescues conventional MoDCs from intrinsic and extrinsic apoptosis and to a greater extent, inflammatory pyroptosis

Annexin V staining does not distinguish intrinsic from extrinsic apoptosis nor from inflammatory pyroptosis [29]. Further investigation to distinguish these events through caspase activation was carried out. The results show that conventional MoDCs infected with mfa-1⁺ P. gingivalis (WT Pg381, DPG-3) had lower levels of caspase-3/7 and -1 activation (Fig. 7A). MoDCs infected with DPG-3 or heat-killed WT Pg 381 also had a significant decrease in capspase-8 activity (Fig. 7A). In contrast, MoDCs treated with LPS or mfa-1⁻ P. gingivalis strains MFI and MFB (Fig. 7A) had significantly higher levels of capsase-1 than conventional MoDCs. We further show that uptake of mfa-1⁺ DPG-3 effectively reduced caspase-1-mediated pyroptosis, whereas blocking of DPG-3 uptake with cytochalasin D led to a significant up-regulation of caspase-1 activity (Fig. 7B). Specifically, MoDCs pretreated with the DC-SIGN ligand HIV-1 gp120 displayed a blocked inhibition of caspase-1 activity by mfa-1⁺ P. gingivalis (Fig. 7C).

Figure 7. Infection of conventional MoDCs by mfa-1⁺ P. gingivalis reduces caspase activation.

MoDCs were washed and removed from the presence of growth factors after Day 5 of culture. MoDCs were then treated, as indicated, for 24 h; caspase activity was detected by flow cytometry. (A) Caspase-3/7 and -1 were reduced upon infection with mfa-1⁺ strains (WT Pg381, DPG-3), and caspase-8 was reduced upon infection with heat-killed WT Pg381 and DPG-3 compared with uninfected MoDCs, and MoDCs, treated with E. coli LPS/TNF-α, heat-killed WT Pg381, MFB, and MFI, all displayed a significant up-regulation of caspase-1 activation. (B) Treatment of MoDCs with cytochalasin D prior to DPG-3 infection, to prevent its uptake, did not cause significant changes to caspase-3/7 or -8 activity but did drive a significant increase in caspase-1 activation. (C) Blocking of DC-SIGN with gp120 before infection with DPG led to significant caspase-1 activation. TP ≤ 0.001; ¶P ≤ 0.01; *P ≤ 0.05.

DISCUSSION

We have described a noncanonical pathway for differentiation of DCs that does not require exogenous growth factors or cytokines. These PDDCs are highly phagocytic when immature and can be induced to mature by E. coli LPS and TNF-α. These PDDCs do not express high levels of accessory molecules, as normally observed in DCs that have encountered pathogen, and show a lack of CD83 and CD86 expression upon forced maturation. The PDDC groups do show an elevated expression of CD40 from conventional MoDCs, suggesting that these DCs may have an increased capability to become activated and induce immunostimulatory responses, such as IL-12 secretion, upon CD40 ligation [38]. The allostimulatory capacity of PDDCs, however, is dependent on the bacterial strain used to educate the progenitors of PDDCs. Most notably, mfa1⁺fimA⁻ (i.e., DPG-3)-generated PDDCs were unable to stimulate CD4⁺ T cell proliferation but increased the longevity of DCs by reducing rates of apoptosis and pyroptosis. The understanding of how this low-grade infection modulates DC homeostasis during disease is critical toward efforts to prevent or treat chronic infections.

In autoimmune diseases, such as systemic lupus erythematosus, matured DCs are increased and cause autoreactive lymphocyte proliferation [39]. The discovery of the correct balance, resulting in resolution of infection, while preventing hyperactive immune stimulation, is of the utmost importance to human health. We show that PDDCs can function as APCs when fully matured but that this depends on the pathogen education. The PRR being targeted, DC-SIGN, is used by blood DCs to actively clear bacteremias [10]. Candida albicans, another opportunistic pathogen, reportedly induces expansion of macrophages and MoDCs by TLR2- and dectin-dependent signaling pathways in mice [40]. mfa-1 fimbriae on viable P. gingivalis provides an optimum differentiation signal, but it should be reiterated that the fimA fimbriae of P. gingivalis, a ligand for β-lectin and CXCR4/TLR2 [41], is not required for optimum differentiation. Nonetheless, the LPS of P. gingivalis is generally regarded as a weak TLR2 agonist and a TLR4 antagonist [42]. Variability does exist, however, in TLR signaling by P. gingivalis LPS, based on the predominant lipid A species present [43]. Cross-talk between DC-SIGN and TLRs is crucial for pathogen-tailored immune responses [44] and as our results suggest, may be crucial in DC differentiation. Efforts are in progress in our laboratory to identify intracellular routing and signaling pathways of these strains that might account for this distinction.

We show that the differentiation of DCs and their function as APCs depends on the aforementioned receptor ligand interactions involved in pathogen contact. Another DC-SIGN-targeting pathogen, hepatitis C virus, inhibits DC maturation and function [45], whereas the non-DC-SIGN-targeting Mycoplasma infection results in normal DC maturation and function [46]. Extensive studies highlighting the mechanisms involved in pathogen expansion of DCs remain lacking. Our present studies demonstrate the direct influence of the pathogen and its virulent fimbriae on DC differentiation, independently of exogenous cytokines or growth factors, and how these infections may lead to immune dysfunction.

Low-grade bacteremia and its inflammatory by-products are thought to be the key links between CP and risk for accelerating systemic diseases, such as atherosclerotic vascular disease [47, 48]. There is little known of the predominate functions of blood DCs in low-grade chronic infections [49]. It is has been hypothesized that constant low-grade bacteremia during disease makes it possible for pathogen dissemination in planktonic form to distant sites [14]. We believe, however, that increased numbers of immature blood DCs that harbor pathogens provide a compelling argument for DCs serving a pathophysiologic function in chronic inflammatory diseases. DCs, although highly migratory, do not display relatively robust bactericidal capabilities [50]. This has been shown with P. gingivalis, which invades and survives within mDCs [10]. Hence, P. gingivalis may exploit the protective environment of DCs, while in turn, providing signals for DCs to survive in circulation. As this pathogen requires blood hemin for growth, it may use DCs as a “Trojan horse” to survive and obtain its favored substrate.

Interestingly, DCs infected via DC-SIGN are immature yet display reduced rates of apoptosis. This was found to be dependent on live mfa-1⁺ P. gingivalis, as heat killing likely denatures the fimbrial adhesins and prevents specific DC-SIGN-mediated intracellular routing. DC apoptosis, upon activation, is an important part of immune regulation, and failure of this mechanism is implicated in autoimmunity [28]. Circulating, immature mDCs are typically short-lived [51], and when matured, functional DCs undergo apoptosis to prevent overstimulation of adaptive immunity. Activation of DCs through external ligands, such as LPS, can provide a delay of cell death that allows for regulated immune stimulation [52]. We also highlight the role of caspase-1 inhibition, as this is typically an inflammatory mode of cell death that is important for killing of intracellular pathogens. In regards to intracellular pathogens, apoptosis of DCs through caspase-1 (pyroptosis) may permit intracellular pathogens to encounter more potent killing mechanisms, such as macrophages and neutrophils [53]. Overt pathogens, e.g., Salmonella, cause rapid inflammatory pyroptosis of DCs through caspase-1 activation [54]. As we have previous evidence that P. gingivalis induces a shift in DC bias from promoting a Th1 to a Th2 effector response and that P. gingivalis survives within mDCs, we believe the inhibition of caspase-1 provides compelling evidence for faulty immune clearance and pathogen dissemination. The P. gingivalis strains that lack mfa-1 (i.e., MFI, MFB) were shown to cause a significant up-regulation of caspase-1 activity. This correlates positively with our previous data that mfa-1⁺ strains prevent inflammatory cytokine production and survive within DCs, whereas unpublished work in our lab shows evidence that mfa-1⁻ strains are killed efficiently by MoDCs. Thus, the differential fimbrial-mediated invasion of DCs may control the switch between inflammatory and suppressive-immune responses. We do, however, acknowledge that caspase-3/7 and -8, typically associated with apoptosis, were also decreased and cannot rule out their importance in this pathway. The observed caspase inhibition may well be interconnected; a caspase-1 is also required for activation of caspase-7 and is essential for restricting Legionella replication in macrophages [55]. We report that caspase-1 and -7 activity is inhibited in MoDCs infected with mfa1⁺ P. gingivalis, thus reinforcing the notion that this pathogen manipulates DCs to provide a protective niche.

Our findings provide unique insights into novel DC differentiation pathways and how a mucosal, DC-SIGN-targeting pathogen dysregulates the immune response and potentially facilitates chronic infections systemically. Dysregulation of DC homeostasis may be a broad strategy of pathogens to evade immunity and facilitate travel to distant sites within the host. Faulty clearance of infection may accelerate other inflammatory diseases, thereby contributing to autoimmune disease [56, 57]. Based on the high disease prevalence of CP [58], these findings may have a significant impact on human systemic health. Our work raises many new questions about the outcomes of pathogen interactions with DCs and emphasizes the importance of continued progress toward understanding blood DC function in the resolution of chronic infection.

AUTHORSHIP

B.M. and C.W.C. designed experiments and wrote the manuscript. B.M., J.C., E.S., G.J.S., C.A.G., and C.W.C. discussed experiments. B.M., J.C., and E.S. performed differentiation experiments. B.M. performed apoptosis and caspase experiments.

DISCLOSURES

The authors have no conflicting interests.