Porphyromonas gingivalis Regulates TREM-1 in Human Polymorphonuclear Neutrophils via Its Gingipains

The Triggering Receptor Expressed on Myeloid cells 1 (TREM-1) is a cell surface receptor of the immunoglobulin superfamily, with the capacity to amplify pro-inflammatory cytokine production and regulate apoptosis. Polymorphonuclear neutrophils (PMNs) are the first line of defence against infection, and a major source of TREM-1. Porphyromonas gingivalis is a Gram-negative anaerobe highly implicated in the inflammatory processes governing periodontal disease, which is characterized by the destruction of the tooth-supporting tissues. It expresses a number of virulence factors, including the cysteine proteinases (or gingipains). The aim of this in vitro study was to investigate the effect of P. gingivalis on TREM-1 expression and production by primary human PMNs, and to evaluate the role of its gingipains in this process. After 4 h of challenge, P. gingivalis enhanced TREM-1 expression as identified by quantitative real-time PCR. This was followed by an increase in soluble (s)TREM-1 secretion over a period of 18 h, as determined by ELISA. At this time-point, the P. gingivalis-challenged PMNs exhibited diminished TREM-1 cell-membrane staining, as identified by flow cytometry and confocal laser scanning microscopy. Furthermore engagement of TREM-1, by means of anti-TREM-1 antibodies, enhanced the capacity of P. gingivalis to stimulate interleukin (IL)-8 production. Conversely, antagonism of TREM-1 using a synthetic peptide resulted in reduction of IL-8 secretion. Using isogenic P. gingivalis mutant strains, we identified the Arg-gingipain to be responsible for shedding of sTREM-1 from the PMN surface, whereas the Lys-gingipain had the capacity to degrade TREM-1. In conclusion, the differential regulation of TREM-1 by the P. gingivalis gingipains may present a novel mechanism by which P. gingivalis manipulates the host innate immune response helping to establish chronic periodontal inflammation.


Introduction
Periodontal diseases are the most common inflammatory infections in humans, caused by complex polymicrobial biofilms attaching on the tooth surface and causing inflammation by the tooth-supporting (periodontal) tissues [1]. The balanced relationship between the biofilm microbiota and the host response of periodontal tissues is commensurate with health. In contrast, a dysbiotic relationship can lead to periodontal disease [2], which is characterized by the destruction of the periodontal tissues (periodontitis), and eventually tooth loss. Polymorphonuclear neutrophils (PMNs) are the first cells to be recruited to the site of inflammation, in response to the developing subgingival biofilmassociated infections in the periodontal tissues [3,4]. Beyond the protective role against bacterial infections PMN may also cause collateral damage to the periodontal tissues. Although PMNs have been extensively studied for their involvement in the local inflammatory responses to periodontal disease, not much is known on their potential role in the amplification of inflammation.
The Triggering Receptor Expressed on Myeloid cells 1 (TREM-1) is a cell surface receptor of the immunoglobulin superfamily, constitutively expressed by monocytes and PMN [5]. It is activated upon bacterial recognition by host cells, triggering a number of intracellular signalling events that result in enhanced proinflammatory cytokine production [5,6]. Bacterial or fungal infections can cause up-regulation of membrane bound TREM-1, as well as release in its soluble (s)TREM-1 form [7,8] rendering it a useful early inflammatory biomarker for systemic infection [9]. Recent evidence in periodontitis patients demonstrated elevated sTREM-1 salivary and serum levels [10], or gingival crevicular fluid levels [11], positively associated with the presence of putative periodontal pathogens in subgingival biofilms [12].
The Gram-negative anaerobe bacterium Porphyromonas gingivalis, is one of the major pathogens implicated in the chronic inflammatory responses governing periodontal disease by either impeding or modulating innate immune defence mechanisms in the periodontium [13][14][15]. Since P. gingivalis shifts the commensal biofilm composition towards a dysbiotic flora resulting in pathological host response and subsequently in periodontitis it is also known as ''keystone-pathogen'' [16,17]. Important P. gingivalis virulence factors that deregulate innate immune responses are the potent arginine-specific and lysine-specific cysteine proteinases aka ''gingipains'' [18][19][20].
It was recently shown that P. gingivalis induces TREM-1 expression in monocytes, concomitantly with an increased release of sTREM-1 [21,22]. Also, mice inoculated with P. gingivalis exhibited higher TREM-1 gene expression, compared to their corresponding uninfected controls [16]. In PMN TREM-1 activation was shown to propagate degranulation, respiratory burst, phagocytosis, and cytokine release in response to bacterial infections with Staphylococcus aureus and Pseudomonas aeruginosa [23,24]. Accordingly, in the periodontal pocket the local excessive inflammatory reaction of the PMNs in response to the ''keystone pathogen'' P. gingivalis could be regulated by TREM-1. The aim of this study was to investigate the effect of P. gingivalis on TREM-1 regulation in PMNs and to evaluate the involvement of its gingipains in this process.

Bacterial Growth Conditions
Porphyromonas gingivalis wild-type W50 strain and gingipain knock-out mutant K1A and E8 strains were used in this study [18,19,25]. E8 strain is deficient in both Arg-gingipain A and Arggingipain B (rgpArgpB), whereas K1A strain is deficient in Lysgingipain (kgp). All three strains were grown anaerobically on Columbia Blood Agar (CBA) plates for 3 to 4 days at 37uC, followed by further anaerobic sub-culturing for 2 to 3 days at 37uC in Brain Heart Infusion broth, containing 0.5% hemin and 0.2% menadione.

Isolation of PMN and Bacterial Challenge
Blood was taken from healthy donors (BD Vacutainer, Allschwil, Switzerland) and PMNs were isolated by density gradient centrifugation using the PolymorphprepTM system according to the manufacturers protocol (Axis-Shield) and in strict accordance to the protocol 2010-0126/0, approved by the Institutional Review Board of the University of Zürich. All participants signed the informed consent. After isolation, PMNs were re-suspended in antibiotic-free cell-culture medium (RPMI-Glutamax, Invitrogen, Life Technologies, Basel, Switzerland) supplemented with 10% heat inactivated human serum at a final density of 2610 6 cells/ml. For each experiment, PMNs from a different healthy blood donor were isolated (24-40 years of age, both males and females).
For the experiments P. gingivalis was washed once with 1 ml of PBS and opsonized with fresh 30% human serum for 30 min at 37uC, where indicated followed by additional washes with PBS before addition to the PMNs, at multiplicity of infection (MOI = bacteria:PMN ratio) 1, 10 and 100, for 4 h or 18 h. Experiments were carried out in triplicate cultures in V-bottom 96 wells plates and were centrifuged at 3806g for 5 min to synchronize adherence of bacteria to PMNs. At least three independent experiments were performed.

Cytotoxicity Assay
The potential cytotoxicity of PMNs by P. gingivalis W50 or its derivative E8 and K1A strains was determined by measuring the extracellularly released lactate dehydrogenase (LDH) activity, over a period of 18 h using the CytoTox 96 Non-radioactive Cytotoxicity Assay (Promega, Mannheim, Germany), according to the manufacturer's instructions. The absorbance was read at 490 nm by a spectophotometric plate reader (Epoch, BioTek, Luzern, Switzerland). The LDH enzyme activity released from damaged cells into the supernatant was expressed as a percentage of total (intracellular+extracellular) LDH activity.

RNA Extraction and cDNA Synthesis
Upon completion of the experiments, the PMNs were collected by centrifugation, after removal of the culture supernatants (which were stored at 280uC, until further analysis). The PMN pellets  Reverse Transcriptase, Oligo(dT) 15 Primers, and PCR Nucleotide Mix according to the manufacturer's protocol (Promega, Mannheim, Germany), at 40uC for 60 min, and 70uC for 15 min.

Quantitative Real-time Polymerase Chain Reaction (qPCR)
TaqMan qPCR was employed for gene expression analysis, in a StepOne Real Time PCR System and software (Applied Biosystems, Life Technologies, Basel, Switzerland). Beta-2 microglobulin (B2M) was used as endogenous control in the samples (house-keeping gene). For the amplification reactions, the Applied Biosystems TaqMan Gene Expression Master Mix and Gene Expression Assay kits were used (assay IDs TREM-1: Hs00218624-m1, B2M: Hs00984230-m1). The standard PCR conditions were 10 min at 95uC, followed 40 cycles at 95uC for 15 sec and 60uC for 1 min. The expression levels of the target transcripts in each sample were calculated by the comparative Ct method (2 2DCt formula) after normalization to the house-keeping gene.

Measurement of sTREM-1 and Interleukin (IL)-8 by ELISA
The levels of sTREM-1 and IL-8 in the cell-free culture supernatants were measured by commercially available specific enzyme-linked immunosorbent assay (ELISA) kits (DuoSet, R&D Systems, Abingdon, UK). The absorbance at 450 nm was measured using a microplate reader (Epoch, BioTek, Luzern, Switzerland), with a wavelength correction set at 570 nm to subtract background. A standard curve was generated using a four-parameter logistic curve fit for each set of samples assayed.

TREM-1 Engagement and Antagonism Experiments
The involvement of TREM-1 in the instigation of proinflammatory responses was further evaluated, by measuring the secretion of interleukin (IL)-8, as a representative cytokine. For investigating the engagement of TREM-1 in IL-8 production, Figure 4. Engagement and antagonism of TREM-1 in IL-8 stimulating-responses by P. gingivalis. PMNs were cultured in 96well plates pre-coated with anti-human TREM-1, or matching IgG 1 isotype control, or in the presence of LP17 (100 ng/ml) and exposed to P. gingivalis MOI 100 for 18 h. Upon completion of the experiments, the cell-free culture supernatants were collected and the concentrations of IL-8 were measured by ELISA. Bars represent mean values 6 standard errors of mean (SEM) from three independent experiments. The asterisk represents statistically significant difference between the control and the corresponding P. gingivalis-challenged groups. doi:10.1371/journal.pone.0075784.g004 Figure 5. Involvement of gingipains in TREM-1 gene expression and secretion. PMNs were exposed to P. gingivalis W50, the Arg-gingipain mutant E8 or the Lys-gingipain mutant K1A strain (MOI 100). After 4 h, the gene expression levels of TREM-1 were measured by qPCR analysis, normalized against the expression levels of the house-keeping gene. The results are expressed as the 2 2DCT formula (A). After 18 h, secretion of sTREM-1 by PMNs into the culture supernatants was measured by ELISA (B). Bars represent mean values 6 standard errors of mean (SEM) from three independent experiments. The asterisk represents statistically significant difference between the P. gingivalis-challenged and control groups. doi:10.1371/journal.pone.0075784.g005 TREM-1 Regulation in Neutrophils by P. gingivalis PLOS ONE | www.plosone.org PMNs were cultured in 96-well flat-bottom plates, which were precoated (4 h at 37uC) with either anti-human TREM-1 antibody (10 mg/ml) or an isotype control (10 mg/ml) (R&D Systems, Abingdon, UK, mouse IgG 1 , clone #193015). The PMNs were plated at a density of 2610 5 cells per well and challenged with P. gingivalis MOI 100, for 18 h. Conversely, the relative involvement of TREM-1 in IL-8 secretion was investigated by the use of its antagonist LP17 peptide (Pepscan Presto B.V., Lelystad, The Netherlands). By this approach, LP17 was added to the cells at concentration 100 ng/ml, simultaneously to P. gingivalis challenge (MOI 100).

Flow Cytometric Analysis for TREM-1 Detection on the PMN Surface
The PMNs were challenged with ascending P. gingivalis MOIs for 4 h and 18 h, as described above. At the end of the experimentation, the cells were washed twice, followed by subsequent staining with monoclonal anti-TREM-1 Alexa Fluor 488 (R&D Systems, Abington, UK) on ice for 30 min. The stained cells were washed, re-suspended and the analysis was performed using a FACSCalibur (Becton Dickinson, Oxford, U.K) flow cytometer and FlowJo Software (Tree Star, Ashland, OR, USA).

Confocal Laser Scanning Microscopy (CLSM) Analysis of TREM-1 Localization
To investigate the localization of TREM-1, as well as P. gingivalis, on the PMNs CLSM was used. The cells were cultured on poly-d-lysine-coated dishes (MatTek Corp., Ashton, MA, USA) and challenged with P. gingivalis MOI 100, for 18 h. Upon completion of the experiments, the dishes were washed and stained using anti-TREM-1 Alexa Fluor 488 (R&D Systems, Abington, UK) followed by fixation in 4% paraformaldehyde (PFA) for 10 min at room temperature. Thereafter, the discs were processed by fluorescent in situ hybridization (FISH) for the detection of P. gingivalis, using simultaneously two oligonucleotide probes (Microsynth, Balgach, Switzerland), one specific for P. gingivalis (P-Pging1006 labelled at the 59-end with Cy3), and one universal bacterial probe (EUB338 labelled at the 59-end with Cy5). The sequences of these probes as well as the standard hybridization conditions are provided elsewhere [26]. For nuclear staining, the samples were further incubated with 1 mg/ml 49,6diamidino-2-phenylindole (DAPI; Sigma-Aldrich, Buchs, Switzerland), to counter-stain DNA. The visualization was performed as described previously [22]. In brief, stained samples were examined using a DM IRB/E inverted microscope (Leica Mikroskopie, Wetzlar, Germany), fitted with a UV laser and an Ar laser (both from Coherent Inc., Santa Clara, CA, USA), a He-Ne laser (Uniphase Vertriebs, Eching/Munich, Germany), and a TCS SP5 computer-operated confocal laser scanning system (Leica Lasertechnik, Heidelberg, Germany). Filters were set to 430-470 nm to detect DAPI, to 500-540 nm for Alexa 488, to 570-630 nm for Cy3, and to 660-710 nm for Cy5. Confocal images were obtained using663 (numeric aperture 1.30) glycerol immersion objective. Zseries were generated by vertical optical sectioning with the slice

Statistical Analysis
A one-way analysis of variance (ANOVA) was used to analyze the statistical significances of the results, using Bonferroni post hoc test for comparisons between individual groups. The data were considered significant at P,0.05.

Cytotoxic Effects of P. gingivalis Strains
Initially, the potential cytotoxic effects of P. gingivalis wild-type W50 strain and its gingipain knock-out mutant K1A and E8 strains on the PMN cells were tested, for up-to 18 h, and at MOI of 100. The percentage of extracellularly released LDH was then measured, indicative of cell lysis. It was found that there were no significant differences between the unchallenged PMNs (mean 6 SD, 13.060.9%) or challenged with P. gingivalis W50 (mean 6 SD, 11.761.9%), P. gingivalis E8 (mean 6 SD, 12.462.2%), or P. gingivalis K1A (mean 6 SD, 11.160.11%).

Effects of P. gingivalis on TREM-1 Expression and sTREM-1 Release by PMNs
The effect of P. gingivalis wild-type W50 strain on TREM-1 mRNA expression was investigated. After 4 h, the wild-type P. gingivalis W50 strain induced a concentration-dependent upregulation of TREM-1 expression, which proved to be significant with MOI 10 and MOI 100 ( Figure 1A). This was confirmed by the release of sTREM-1 from PMNs following P. gingivalis infection. P. gingivalis at a MOI 100 caused a significant 13-fold increase of sTREM-1 release, compared to the control, which was absent at lower MOIs ( Figure 1B).
To determine if P. gingivalis infection could elicit TREM-1 shedding from PMNs, its cell-surface levels were measured by flow cytometry. Although there were no differences between groups after 4 h, membrane-associated TREM-1 was down-regulated after 18 h in response to increasing MOI of P. gingivalis (Figure 2). TREM-1 expression on the PMNs cell surface in response to P. gingivalis was further investigated by CLSM. Unchallenged PMNs expressed TREM-1 on their surface as seen by the presence of a strong positive green staining ( Figure 3B). However, when the PMNs were infected with P. gingivalis MOI 100, the localization of TREM-1 on the cell surface was decreased, and the nuclei appeared more rounded with vivid DAPI staining ( Figure 3C). Further, co-staining by FISH with a species-specific, or a generic eubacterial probe, indicated that P. gingivalis was localized on the PMNs ( Figure 3C, arrow).

TREM-1 Engagement and Antagonism in P. gingivalisinduced IL-8 Production
In order to assess the consequences of TREM-1 engagement on P. gingivalis-induced IL-8 release by PMNs the natural ligand of the TREM-1 receptor anti-TREM-1 antibody was used. After 18 h, P. gingivalis induced IL-8 secretion by the PMNs which was further enhanced by 35% in the presence of anti-TREM-1 (Figure 4). The presence of anti-TREM-1 alone was adequate to cause a significant enhancement of IL-8 secretion by 30% in absence of P. gingivalis compared to the IgG isotype control.
Using a complementary approach, we investigated whether the induction of IL-8 secretion by P. gingivalis could be blocked by adding the synthetic peptide LP17 that mimics a highly conserved domain of sTREM-1. We found that the addition of the LP17 antagonist (100 ng/ml) of TREM-1 resulted in 25% reduction of IL-8 secretion by the P. gingivalis-challenged PMNs, (Figure 4). The presence of LP17 reduced IL-8 secretion by the cells by almost 30%, even in the absence of P. gingivalis.

Investigation of Gingipain Involvement on TREM-1 Regulation by P. gingivalis
Next, the potential involvement of P. gingivalis gingipains in the regulation of TREM-1 was investigated, by employing the Lysgingipain mutant strain K1A, and the Arg-gingipain mutant strain E8. It was found that both P. gingivalis gingipain-mutant strains induced TREM-1 expression in PMNs to a similar extent as their parental wild-type W50 strain, after 4 h of infection ( Figure 5A). On the contrary, differences between the different P. gingivalis strains were observed on the sTREM-1 secretion levels. Both the P. gingivalis wild-type W50 and K1A strains caused increased sTREM-1 release, whereas, in contrast, the E8 strain did not significantly affect sTREM-1 secretion, compared to the control ( Figure 5B).
The localization of PMN cell-bound TREM-1 in response to the P. gingivalis gingipain mutant strains was further confirmed by CLSM ( Figure 6). In the unchallenged control cultures, there was a strong positive green staining for TREM-1, associated with the surface of the PMNs ( Figure 6B). However, when the cells were infected with P. gingivalis wild-type W50 or the Lys-gingipain mutant K1A strain, TREM-1 was no longer detected on the cell surface ( Figure 6C-D). In contrast, in PMNs challenged with the P. gingivalis Arg-gingipain mutant strain E8, partial TREM-1 staining was detected on the cells ( Figure 6E). This is complementary to the finding by ELISA, showing decreased sTREM-1 release in response to the E8 strain. Collectively, these data may indicate that absence of the Arg-gingipain from P. gingivalis may prevent shedding of TREM-1 from the PMN cell surface into the culture supernatant.

Effect of P. gingivalis Gingipains on sTREM-1 Degradation
The capacity of P. gingivalis to degrade sTREM-1 was further tested. A defined concentration of recombinant sTREM-1 (2000 pg/ml) was spiked into cell culture media in the absence or presence of any of the three P. gingivalis strains, over a period of 18 h. Significant reduction of sTREM-1 concentrations occurred in the presence of P. gingivalis wild-type and Arg-gingipain mutant E8 strain, but not in the case of the Lys-gingipain mutant K1A strain (Figure 7). The resulting sTREM-1 concentration accounted for approximately 30% of control levels. These findings are suggestive of the capacity of Lys-gingipain, but not Arg-gingipain, to degrade sTREM-1.

Discussion
We found that P. gingivalis regulates TREM-1 in PMNs, with potential implications in down-stream inflammatory responses. P. gingivalis induced TREM-1 gene expression independent of its gingipains, followed by sTREM-1 release from the cell surface, in a gingipain-dependent manner. Porphyromonas gingivalis is well known for its capacity to modulate innate immunity in periodontal inflammation via PMNs, resulting in survival benefits [14]. In agreement with earlier reports in vivo [27] P. gingivalis did not cause any cytotoxic effects. There is also evidence that P. gingivalis can decrease [28], delay [29][30][31] and inhibit [32] PMN apoptosis, in addition to preventing the clearance of apoptotic cells [27,[33][34][35], and that activation of TREM-1 can exert a protective effect against apoptosis [36]. Although these findings support an antiapoptotic effect of P. gingivalis-induced TREM-1 activation, a proapoptotic effect cannot be excluded, due to the long period of PMN exposure, also applied in other studies [37,38], and the rounding of the PMN cell nuclei, particularly in presence of P. gingivalis, as viewed by CLSM. P. gingivalis induced TREM-1 gene expression independent of its gingipains, but also sTREM-1 release from the cell surface into the culture medium, in a gingipain-dependent manner. These findings are in agreement with previous reports demonstrating that PMNs are involved in sTREM-1 generation during other infections [5,39]. Moreover, recent findings demonstrate an increased sTREM-1 release by monocytic cells, in response to P. gingivalis [22,40]. The occurrence of sTREM-1 in biological fluids appears to be an important predictor and diagnostic marker for sepsis, bacterial and fungal pneumonia [41]. A positive association has been shown between periodontitis and the levels of sTREM-1 in serum and saliva, potentially revealing a link between systemic and oral inflammation [10]. Interestingly, a positive correlation is shown between gingival crevicular fluid site-specific levels of sTREM-1 and the levels of P. gingivalis in subgingival biofilms [42]. The findings of the present in vitro study on PMNs reveal a potential mechanistic pathway between P. gingivalis and the increased levels of sTREM-1.
It was further confirmed, by engagement and antagonism experiments, that TREM-1 propagated pro-inflammatory cytokine production, representatively demonstrated by IL-8 in the present study. Earlier studies have also demonstrated that, in response to bacterial challenge, TREM-1 engagement enhances the release of IL-8 from PMNs [5,23], or monocytic cells [8,22,40]. Since IL-8 is a potent chemokine, its enhanced secretion mediated by TREM-1 may consequently lead to further PMN recruitment at the affected site (i.e. gingival sulcus or periodontal pocket), in an attempt to tackle with the increasing bacterial challenge.
We identified the gingipains as the responsible P. gingivalis virulence factor for the shedding of sTREM-1 by PMNs. Gingipains are considered detrimental in the capacity of P. gingivalis to evade host innate immune responses, by contributing to its capacity to resist complement elimination, prevent blood clotting, and acquire essential nutrients [43] and activate matrix metalloproteinases [44]. It was found that the P.gingivalis Arggingipain mutant strain was unable to enhance sTREM-1 release by the PMNs, marked by a concomitant intense TREM-1 presence on the cells, in contrast to the effects of the wild-type or the Lys-gingipain mutant strain. These results indicate that sTREM-1 shedding from the cell surface is attributed to its Arggingipains. Nevertheless, as gingipains are also known to degrade several proteins, we evaluated the capacity of P. gingivalis to degrade known concentrations of sTREM-1. We found that sTREM-1 was resistant to degradation by the Lys-gingipain mutant strain, as opposed to the Arg-gingipain mutant, or the wild-type P. gingivalis strain. This indicates that the Lys-gingipain is responsible for sTREM-1 degradation, once this molecule is shed from the cell surface by the Arg-gingipain. These findings are collectively in line with earlier studies confirming the enzymatic properties of gingipains in cleaving and degrading complement factors, immunoglobulins. metalloproteinases, the fibronectinintegrin binding, toll-like receptors and other molecules [43,[45][46][47][48][49]. Both the Lys-gingipain and the Arg-gingipain are also shown to be able to cleave IL-8 and modulate its biological activity, as well as to degrade and inactivate tumor necrosis factor-a [50,51].
The capacity of P. gingivalis to differentially regulate molecules via its gingipains may favor the evasion of the innate immunity mechanisms that would normally result in its phagocytosis and killing. In the instance of TREM-1 as revealed in the present experimental system, P. gingivalis may employ its Arg-gingipain to shed-off sTREM-1 from the PMNs, thus amplifying the inflammatory response. On the other hand degradation of sTREM-1 by the Lys-gingipain may eliminate the ability of PMNs to propagate an inflammatory response or to efficiently accomplish their antibacterial actions, including phagocytosis. The net outcome of TREM-1 regulation in PMNs by P. gingivalis may depend on the stage of infection. In early stages, via its Arg-gingipain, P. gingivalis may cause a sTREM-1-supported innate immune response, creating an inflammation-rich environment for its survival. In later stages, where the excessive inflammatory response may compromise its survival, P. gingivalis may employ its Lys-gingipain to control this and remain stealth. Hence, dual regulation of sTREM-1 release and degradation by two different ginigpains may be a novel mechanism by which P. gingivalis evades the host defenses and establishes chronic periodontal inflammation.