Oral Neutrophils Characterized: Chemotactic, Phagocytic, and Neutrophil Extracellular Trap (NET) Formation Properties

Maintenance of oral health is in part managed by the immune-surveillance and antimicrobial functions of polymorphonuclear leukocytes (PMNs), which migrate from the circulatory system through the oral mucosal tissues as oral PMNs (oPMNs). In any microorganism-rich ecosystem, such as the oral cavity, PMNs migrate toward various exogenous chemoattractants, phagocytose bacteria, and produce neutrophil extracellular traps (NETs) to immobilize and eliminate pathogens. PMNs obtained from the circulation through venipuncture (hereafter called cPMNs) have been widely studied using various functional assays. We aimed to study the potential of oPMNs in maintaining oral health and therefore compared their chemotactic and antimicrobial functions with cPMNs. To establish chemotactic, phagocytic, and NET forming capacities, oPMNs and cPMNs were isolated from healthy subjects without obvious oral inflammation. Directional chemotaxis toward the chemoattractant fMLP was analyzed using an Insall chamber and video microscopy. fMLP expression was assessed by flow cytometry. Phagocytosis was analyzed by flow cytometry, following PMN incubation with heat-inactivated FITC-labeled micro-organisms. Furthermore, agar plate-based killing assays were performed with Escherichia coli (Ec). NET formation by oPMNs and cPMNs was quantified fluorimetrically using SYTOX™ Green, following stimulation with either PMA or RPMI medium (unstimulated control). In contrast to cPMNs, the chemotactic responses of oPMNs to fMLP did not differ from controls (mean velocity ± SEM of cPMNs: 0.79 ± 0.24; of oPMNs; 0.10 ± 0.07 micrometer/min). The impaired directional movement toward fMLP by oPMNs was explained by significantly lower fMLP receptor expression. Increased adhesion and internalization of various micro-organisms by oPMNs was observed. oPMNs formed 13 times more NETs than stimulated cPMNs, in both unstimulated and stimulated conditions. Compared to cPMNs, oPMNs showed a limited ability for intracellular killing of Ec. In conclusion, oPMNs showed exhausted capacity for efficient chemotaxis toward fMLP which may be the result of migration through the oral tissues into the oral cavity, being a highly “hostile” ecosystem. Overall, oPMNs' behavior is consistent with hyperactivity and frustrated killing. Nevertheless, oPMNs most likely contribute to maintaining a balanced oral ecosystem, as their ability to internalize microbes in conjunction with their abundant NET production remains after entering the oral cavity.

Maintenance of oral health is in part managed by the immune-surveillance and antimicrobial functions of polymorphonuclear leukocytes (PMNs), which migrate from the circulatory system through the oral mucosal tissues as oral PMNs (oPMNs). In any microorganism-rich ecosystem, such as the oral cavity, PMNs migrate toward various exogenous chemoattractants, phagocytose bacteria, and produce neutrophil extracellular traps (NETs) to immobilize and eliminate pathogens. PMNs obtained from the circulation through venipuncture (hereafter called cPMNs) have been widely studied using various functional assays. We aimed to study the potential of oPMNs in maintaining oral health and therefore compared their chemotactic and antimicrobial functions with cPMNs. To establish chemotactic, phagocytic, and NET forming capacities, oPMNs and cPMNs were isolated from healthy subjects without obvious oral inflammation. Directional chemotaxis toward the chemoattractant fMLP was analyzed using an Insall chamber and video microscopy. fMLP expression was assessed by flow cytometry. Phagocytosis was analyzed by flow cytometry, following PMN incubation with heat-inactivated FITC-labeled micro-organisms. Furthermore, agar plate-based killing assays were performed with Escherichia coli (Ec). NET formation by oPMNs and cPMNs was quantified fluorimetrically using SYTOX TM Green, following stimulation with either PMA or RPMI medium (unstimulated control). In contrast to cPMNs, the chemotactic responses of oPMNs to fMLP did not differ from controls (mean velocity ± SEM of cPMNs: 0.79 ± 0.24; of oPMNs; 0.10 ± 0.07 micrometer/min). The impaired directional movement toward fMLP by oPMNs was explained by significantly lower fMLP receptor expression. Increased adhesion and internalization of various micro-organisms by oPMNs was observed. oPMNs formed 13 times more NETs than stimulated cPMNs, in both unstimulated and stimulated conditions. Compared to cPMNs, oPMNs showed a limited ability for intracellular killing of Ec. In conclusion, oPMNs showed exhausted capacity for efficient chemotaxis toward fMLP which may be the result of migration through the oral tissues into the oral cavity, being a highly "hostile" ecosystem. Overall, oPMNs' behavior is consistent with hyperactivity and frustrated killing. Nevertheless, oPMNs most likely contribute to maintaining a balanced oral ecosystem, as their ability to internalize microbes in conjunction with their abundant NET production remains after entering the oral cavity.

INTRODUCTION
Polymorphonuclear leukocytes (PMNs) are terminally differentiated innate immune cells that descend from hematopoietic stem cells in the bone marrow and transit through the peripheral blood circulation as circulatory PMNs (cPMNs) (1,2). Traditionally, cPMNs exit the vasculature and migrate to sites of tissue damage, inflammation, and infection to perform various protective and antimicrobial functions contributing to the neutralization and elimination of pathogens and damaged cells (3).
The oral cavity typically harbors over 700 different species of colonizing bacteria evidently priming and activating the PMNs, which originate from a nearly sterile blood circulation (4). In a healthy state, approximately 30,000 oral PMNs (oPMNs) per minute arrive through the gingival crevicular fluid (GCF) which flows into the oral cavity from the periodontal sulcus (5). In the gingival crevice, oPMNs form a wall between the epithelium and the dental biofilm to protect the periodontal tissue and to maintain periodontal tissue homeostasis (6)(7)(8). Moreover, oPMNs migrate from all oral mucosal tissues (9). It has been suggested that in the oral cavity, PMNs carry out a unique immune surveillance function and symbiotically interact with the commensal oral microflora in order to maintain homeostasis and oral health (10). In cases of chronic inflammation, such as in periodontitis, a multifactorial chronic inflammatory disease of the periodontium leading to alveolar bone loss (11), an increased influx of oPMNs with a hyperactive phenotype extravasate into the oral cavity (12).
PMNs recognize, immobilize, internalize, and kill extracellular pathogens both intra-and extracellularly (13,14). PMN recruitment and migration is mediated by various endogenous and exogenous chemoattractants such as interleukin-8 (IL-8), bacteria-derived lipopolysaccharides (LPS), and N-formylmethionyl-leucyl-phenylalanine (fMLP). These chemoattractants initiate signal transduction events, leading to a multitude of cellular processes including diapedesis, chemotaxis, and migration of PMNs (15,16). Upon contact, PMNs adhere to and engulf bacteria, a process known as phagocytosis (17). Impaired phagocytic capacities of PMNs lead to the accumulation of bacteria, delay of bacterial clearance and disturbance of oral microbial homeostasis (18). Pathogen destruction is for a large part accomplished through reactive oxygen species (ROS) generation by activated PMNs. Another antimicrobial strategy of PMNs is the formation of neutrophil extracellular traps (NETs) released by highly activated neutrophils (19)(20)(21). NETs are formed after nuclear membrane and granule disintegration. Subsequently, nuclear material and contents of the granules mix, and NETs are released after cell membrane disruption, followed by cell death. This has been termed suicidal NETosis (21,22). It has been reported that PMNs can also release NETs without cell membrane rupture. Since these PMNs remain viable, it is called vital NETosis (23). NETs consist of a core DNA element, decondensed nuclear chromatin combined with various DNAbound antimicrobial proteins and peptides (24). The putative role of NETs is to entrap pathogens and they are induced by diverse stimuli such as microbes, host factors, microbial products, and immune complexes which bind to the receptors on the surface of PMNs (25).
Hitherto, blood-derived PMNs (cPMNs) have been extensively studied in relation to the pathogenesis of oral inflammation, and to periodontitis in particular (26,27). Several studies show that altered PMN activity, including impaired chemotaxis (16), phagocytosis (28), and NET formation (25,29,30), is associated with the disruption of tissue homeostasis and disease. However, few studies are available about the functional antimicrobial characteristics of oPMNs. Glogauer et al. found elevated numbers of oPMNs with a hyperactive phenotype in periodontitis patients reflecting the severity of periodontal disease and treatment response (31)(32)(33). It was confirmed by our group that oPMNs are hyperactive cells in the oral ecosystem by demonstrating a significant increase in ROS production (7). However, oPMNs' characteristics such as the chemotaxis, phagocytosis, and NET formation have yet to be fully elucidated. In this study, we aimed to compare oPMNs and cPMNs, in terms of their chemotactic, phagocytic, and NET forming behavior.

Human Subjects and Experimental Setup
Donors (n = 9) for the chemotaxis, NET formation (n = 9), and killing (n = 4) experiments were recruited at the Birmingham Dental School and Hospital (United Kingdom). For the phagocytosis and fMLP expression experiments, donors (n = 3) were recruited at the Academic Center for Dentistry Amsterdam (ACTA, The Netherlands). The study was approved by the Birmingham ethical committee (14/SW/1148)

oPMN Isolation
Oral rinses were collected and isolated according to previously described protocols (7,9). Briefly, 4 oral rinses of 30 s with 10 mL of sterile 0.9% sodium chloride solution (Versylene R , Fresenius Kabi, Sèvres, France) with intermission periods of 4 min were collected. Oral rinses were centrifuged at 500 RCF, after which the pellet was suspended in 10 mL PBS.

Chemotaxis Assay
Chemotaxis experiments were performed as described previously by Roberts et al. (16). Briefly, the chemotactic movements of oPMNs and cPMNs (n = 9) were investigated over time in Insall chambers (34

Time-Lapse Image Analysis
Time-lapse images, generated by video microscopy, were processed using Q pro-imaging software (Qimaging) and analyzed using Fiji (35) Figure 2B) was determined based on forward-(FSC) and sideward scatter (SSC), which, respectively represented the distribution of cells in the light scatter based on size and intracellular composition. cPMNs and oPMNs were tested for CD16 (cPMNs: 99.7%, oPMNs: 96.2%) and CD66b (cPMNs: 99.8%, oPMNs; 80.4%) as neutrophil markers (37), where CD16-negative and CD66b-negative gates were determined using unstained samples. Finally, in the CD16positive population, fMLP receptor expression was assessed on cPMNs and oPMNs. fMLP receptor negative gates were set in control conditions and stained with the corresponding isotype IgG1κ.
Bacterial and yeast suspensions were isolated from broth cultures by centrifugation, washed twice in sterile PBS and diluted to an optical density of 1 at 600 nm in 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES) buffer. Microbes were heat-inactivated at 60 • C for 1 h and stored in aliquots at −20 • C for further analyses.

Adhesion and Internalization Assays
Micro-organisms were extrinsically labeled for 1 h with 100 µg/mL fluorescein isothiocyanate (FITC, Sigma Aldrich) and washed in phosphate buffer (PB, 0.1 M, pH 7.2) to remove remaining unlabeled FITC molecules. Labeling efficiency (>70%) was defined by flow cytometry before tests.
Adhesion and internalization by cPMNs and oPMNs were flow cytometrically analyzed. Both adhesion and internalization assays were performed with un-opsonized heat-inactivated microbes. Immediately after isolation, PMNs were recovered in HEPES buffer (pH 7.4), supplemented with 1 mM calcium chloride, 0.5% BSA, and 1 mg/mL glucose (all Sigma-Aldrich). FITC-conjugated anti-human CD66b (clone G10F5, BD Biosciences) was used to check for neutrophil purity (37). PMNs were stained with APC-conjugated anti-human CD16 (Clone 3G8, BD Biosciences) for 30 min in the dark. Cells were incubated with heat-inactivated FITC labeled microbes at a previously established optimal PMN/microbe ratio of 1:3 for 30 min at 4 or 37 • C for adhesion and internalization, respectively. Internalization was distinguished from adhesion using the membrane-impermeable dye trypan blue (38). Accordingly, trypan blue (20 µg/mL, Sigma Aldrich) was added to quench the fluorescence of adherent FITC-labeled micro-organisms before the assessment of internalization. After 30 min of incubation, cells were fixed in 1% paraformaldehyde (PFA) in PBS. Flow cytometric acquisition and analysis was performed on a BD FACSverse TM flow cytometer (BD Biosciences) with a medium flow rate (63 µL/min) where at least 10,000 cells were analyzed per sample. Flow cytometry data were analyzed using associated FACSuite software (BD Biosciences). The gating strategy employed for adhesion and internalization assays is presented in Supplementary Figure 3. Briefly, the live population of cPMNs (Supplementary Figure 3A) and oPMNs (Supplementary Figure 3B) was determined based on forward-(FSC) and sideward scatter (SSC), which, respectively represented the distribution of cells in the light scatter based on size and intracellular composition. Subsequently, cPMNs and oPMNs were tested for CD16 (cPMNs: 99.8%, oPMNs: 96.8%) and CD66b (cPMNs: 99.8%, oPMNs; 80.4%) as neutrophil markers (37), where CD16-negative and CD66b-negative gates were determined using unstained samples. Finally, adhesion and internalization of FITC-labeled bacteria by PMNs was determined based on the percentage of FITC+CD16+ population. Accordingly, this population illustrated the percentage of FITC+ events (i.e., FITC-labeled microbes) detected in the CD16+ (i.e., PMN) population.

Killing Assay
In order to assess the killing capacity of cPMNs and oPMNs, phagocytosis assays were performed with live Ec (strain W3110, ATCC 27325). As oPMN samples contained numerous oral bacteria, we pre-incubated isolated oPMNs with an antibiotic and antifungal cocktail (final concentration of 2.5 µg/mL Amoxicillin, 25 µg/mL Tetracycline, 25 µg/mL Metronidazole, 2.5 µg/mL Fungizone, all Sigma-Aldrich). After 15 min of incubation at RT, oPMNs were washed in 20 mL PBS. After centrifugation (10 min at 500 RCF), the supernatant was stored to test for the possible presence of antibiotics in the oPMN samples which would impact Ec survival.
Ec was grown on BHI agar. A single colony was inoculated into BHI broth and grown overnight at 37 • C on a shaker. The following day, 20 µL of the bacterial suspension was inoculated into 20 mL of fresh BHI broth and Ec was grown at 37 • C on a shaker until its late log phase (OD600 = 0.9-1.0). Next, Ec was added to 250,000 PMNs (PMN/bacteria ratio of 1:3) and incubated at 37 • C in a shaking water bath. After 30 min of incubation, the cell-bacteria suspension was diluted 1000 x in PBS and plated (100 µL per plate) onto BHI agar plates in triplicate and grown overnight at 37 • C, 5% CO 2 . Additionally, cell counts and the viability of cPMNs and oPMNs was routinely assessed with trypan blue light microscopy which resulted in viabilities of 93.35% (cPMNs) and 65.18% (oPMNs) after incubation with bacteria (mean percentages, n = 4). Finally, colony forming units (CFU) were counted.
Several controls were included for killing experiments. As a first control representing 0% killing, Ec was incubated and plated alone. As a second control condition, oPMNs were plated in order to determine the colonies originating from oral bacteria, which had survived incubation with antibiotics. This resulted in up to 5 CFU per plate, demonstrating that antibiotics were sufficient in killing oral bacteria in oPMN samples. Lastly, Ec was incubated with supernatants from the oPMN washing step after incubation with antibiotics, in order to investigate whether traces of antibiotics in the PMN pellets would affect bacterial killing of Ec. However, no differences were observed between Ec alone and Ec incubated with supernatants, proving that, even if present, any traces of antibiotics in the oPMN samples had negligible concentrations as they did not affect the survival of Ec.

Visualization of NET Formation by cPMNs and oPMNs
cPMNs and oPMNs (1 x 10 6 cells in culture medium) were added to a transparent 48-wells plate (Greiner Bio-One, Alphen a/d Rijn, The Netherlands), previously coated with filter sterilized 1% BSA. After 30 min of baseline incubation (37 • C, 5% CO 2 ), selected wells were stimulated with either 75 nM phorbol 12myristate-13-acetate (PMA, Sigma Aldrich, Merck) or culture medium as control and incubated for 3 h in a humidified atmosphere of 5% CO 2 in air at 37 • C. Post-incubation, 50 nM extracellular nucleic acid dye SYTOX TM green (Invitrogen by Thermo Fisher Scientific, California, USA) was added to each well for the visualization of NETs using a fluorescence microscope (Leica Microsystems). The morphology of PMNs forming NETs was characterized by the release of extracellular web-like DNA strands. Images were acquired using 20 x and 40 x objectives (Leica Microsystems). Leica software (Version 4.2) was used for image acquisition and processing.

Online Multimedia Supplements
The online data supplements (Supplementary Videos 1-4) contain movie files demonstrating migration assays of cPMNs and oPMNs performed as described in chemotaxis assays.

Statistics
Chemotactic index, velocity, and speed of chemotactically active PMNs are presented in Tukey box and whisker plots (designed with GraphPad Prism, version 6.07, La Jolla, CA, USA). Chemotaxis data distribution was assessed with a D'Agostino-Pearson omnibus test and found to be not normally distributed. Statistical analyses of these data were performed by Friedman tests using GraphPad Prism software. Phagocytosis (adhesion, internalization, and killing assays), fMLP receptor expression, and NET formation data were normally distributed according to D'Agostino-Pearson omnibus tests. These data were analyzed with paired t-tests using GraphPad Prism software and presented as means + standard error of means (SEM). P < 0.05 were considered significant.

RESULTS oPMNs Exhibit Impaired Directional Chemotactic Accuracy Toward fMLP
The chemotactic movements of cPMNs (Figures 1A,B) and oPMNs (Figures 1C,D) toward PBS and fMLP over the entire time period of 20 min were compared (See Supplementary Videos). For all datasets, an equal number of PMNs were tracked, analyzed and presented in spider, vector and rose plots (Figure 1). Control conditions (PBS) of cPMNs ( Figure 1A) and oPMNs (Figure 1C), as expected, show very little movement without any obvious directionality of movement. The strength of the movement, evidenced by the length of cell tracks in spider plots and the distribution of dots in rose plots (Figures 1A,C), is evenly distributed indicating non-directional, minimal random movement of cells in response to PBS. cPMNs show obvious chemotactic responses to fMLP compared to control treated cells, as previously described by Roberts et al. (16) (Figure 1B). cPMNs showed a strong response evidenced by longer cell tracks (spider diagrams), with the largest proportion of cells moving North (primarily 120 • , vector plot), showing a "Northern Hemisphere" distribution of cells (rose plot) and a median strength of 90 • (rose plot) toward fMLP. When comparing fMLP conditions, cPMNs ( Figure 1B) have longer cell tracks than oPMNs (Figure 1D), where only a few cells show random chemotactic movement. oPMNs show no collective directional movement toward fMLP ( Figure 1D). Overall, oPMN conditions showed similar movement patterns in control and treated conditions indicating that oPMNs exhibit impaired directional chemotactic movement toward the chemoattractant fMLP.  2 | Chemotactic index, velocity, and speed of cPMNs and oPMNs in response to fMLP or PBS. Quantitative analysis of migration is presented in Tukey box and whiskers plots. The box represents the 75 and 25th percentiles with the median presented as a midline. Tukey whiskers represent the minimum and maximum values and outliers were defined as values exceeding 1.5 times the interquartile range. The chemotactic index, which is the directional accuracy of chemotaxis, is presented in (A). The cell velocity, representing the average speed (µmeter/min) of the cells moving in its most prominent direction, is presented in (B). The speed, which is the average speed (µmeter/ min) moving in any direction is presented in (C). n = 9, **p < 0.01, ***p < 0.001.

oPMNs Are Live, Active Cells With Impaired Directional Migration Capacities
Quantitative analysis of migration by cPMNs and oPMNs in response to PBS and the chemotactic agent fMLP is shown in Figure 2. The chemotactic index (directional accuracy of chemotaxis) of cPMNs migrating toward fMLP is significantly higher than in all other conditions (Figure 2A). The same trend was observed for the average velocity of the cells in its most prominent direction over the time course (Figure 2B). In response to fMLP, a significantly lower velocity was observed in oPMNs when compared to cPMNs. No significant difference in chemotactic index and velocity was observed for oPMNs in response to PBS or fMLP (Figures 2A,B). The chemotactic index (Figure 2A), velocity (Figure 2B), and speed ( Figure 2C), did not differ significantly between cPMN control conditions (PBS) and oPMNs in response to PBS or fMLP. The speed of cPMNs exposed to fMLP was slightly higher (1.53 ± 0.44 µm/min) but not significantly different than under control conditions (0.93 ± 0.26 µm/min) or of that observed for oPMNs (fMLP: 0.91 ± 0.23, PBS: 0.63 ± 0.21 µm/min). Altogether, with the exception of speed, oPMNs were not significantly different in their response to fMLP in comparison to control-treated cPMNs and oPMNs (Figure 2, compare Figures 1A,D). This data demonstrates that oPMNs are active, live cells with impaired migration capacities toward the chemotactic agent fMLP.

Impaired Directional Migration Accuracy of oPMNs Toward fMLP Is Explained by Low fMLP Receptor Expression
Since oPMNs exhibited an impaired directional migration accuracy toward fMLP, we hypothesized that fMLP receptors on oPMNs were saturated and therefore incapable of binding further fMLP, rendering oPMNs incapable of migrating toward fMLP. As such, the expression of the fMLP receptor was analyzed on viable cPMNs and oPMNs using flow cytometry (Figure 3). Gating strategies employed for these experiments are shown in Supplementary Figure 2. Firstly, PMN surface markers CD66b and CD16 were tested on both cPMNs and oPMNs. Both cPMNs and oPMNs expressed CD16 (99.7 ± 0.2% and 96.2 ±1.2%, respectively, percentages ± SEM) and CD66b (99.8 ± 0.03% and 80.4 ± 6.5%, respectively, percentages ± SEM). Secondly, in these PMN populations, fMLP receptor expression was tested. Indeed, fewer oPMNs expressed the fMLP receptor (47.9 ± 3.0%), whereas nearly all cPMN were positive for the fMLP receptor (99.0 ± 0.6%, p < 0.0005).

A Trend for Greater Adhesion by oPMNs than by cPMNs Was Observed
In vivo, PMNs migrate and adhere to pathogens in order to affect phagocytosis and the subsequent destruction of pathogens. Adhesion was investigated at 4 • C in order to prevent active internalization. cPMN and oPMN adhesion were investigated with Aa, Pg, Fn, Tf, Sm, Ss, Ec, and Ca. Adhesion by cPMNs ( Figure 4A) and oPMNs (Figure 4B) of FITC labeled microbes are illustrated by white arrows in   and small debris remained in these oPMN suspensions as seen in Figure 4B. Nevertheless, the purity of PMNs in these samples, as established by flow cytometry with CD16 and CD66b markers, was on average 99.7 and 99.8% for cPMNs, and 96.2 and 80.4% for oPMNs, respectively (Supplementary Figure 3). Adhesion of FITC-labeled micro-organisms by cPMNs and oPMNs, as shown in Figures 4A,B, was quantified using flow cytometry. The employed gating strategy is shown in Supplementary Figure 3. The graphs in Figures 4C-J show a trend of greater adhesion by oPMNs vs. cPMNs for the majority of the tested micro-organisms. In general, more oPMNs adhered to Fn, Ca, Tf, and Ss, although this difference did not reach statistical significance. The adhesive capacity of oPMNs to Ec and Ss was 2-3 times greater than for other tested microbes. Specifically, the adhesion of Ec was significantly higher (p = 0.0018) by oPMNs in comparison to cPMNs.

oPMNs Exhibit an Increased Internalization Capacity of Aa, Pg, and Ec
Internalization by cPMNs ( Figure 5A) and oPMNs ( Figure 5B) of FITC labeled microbes is indicated with white arrows in Figure 5. Internalization of pathogens by cPMNs and oPMNs, as shown in Figures 5A,B, was quantified using flow cytometry. The gating strategy employed for flow cytometry experiments is shown in Supplementary Figure 3.
A clear trend for more internalization of all tested microbes by oPMNs than cPMNs is shown in graphs of Figures 5C-J. Similar to the adhesion assay results, Ec phagocytosis by oPMNs was significantly higher (p = 0.0042) than by cPMNs. Accordingly, the internalization capacity of oPMNs for Ec and Ss was approximately 2 times greater than for other tested microbes. The internalization of the periodontal bacteria Aa and Pg was 3-fold higher in oPMNs than in cPMNs. Fn, Tf, Sm, Ss, and Ca internalization by oPMNs did not differ significantly in comparison to cPMNs, however, there is a trend of more internalization by oPMNs. Interestingly, significantly more bacteria were internalized than adhered to by both cPMNs and oPMNs. In conclusion, oPMNs are viable, active cells with functional adhesion and internalization properties.

oPMNs Have a Minimal Capacity to Kill Ec
The microbial killing capacity of cPMNs and oPMNs was assessed with live Ec. cPMNs were originated from a nearly sterile environment while oPMNs originate from an environment containing numerous oral bacteria. Therefore, oPMN samples were pre-incubated with antibiotics and subsequently plated in order to test oral bacteria survival. Incubation with antibiotics was sufficient to eliminate oral bacterial contamination in the oPMN samples as no more than 5 CFU were counted after overnight incubation of oPMN samples on BHI agar plates. No CFU were present on plated cPMN samples.
Compared to the control condition, cPMNs after incubation with Ec for 30 min, showed reduced numbers of CFU after the overnight cultures, corresponding to a 29% increase in intracellular killing (p = 0.02, Figure 6). In contrast, oPMNs showed 4% killing under the same conditions. This was significantly lower than cPMNs (p = 0.02) and the killing by Quantitative analysis of killing capacity (percentages) of cPMNs (white bar) and oPMNs (gray bar). Percentages are based on the presence of CFU on overnight Ec agar plate assays. The control condition, representing 0% killing, was Ec incubated alone. On average, cPMNs killed 29 ± 4% while oPMNs killed 4 ± 7%. cPMNs killed significantly more Ec than both the control condition (p = 0.02) and oPMNs (p = 0.02). Data are presented as mean percentages (+SEM), n = 4, *p < 0.05. oPMNs did not differ significantly from the control condition. We conclude that oPMNs show a minimal ability for the intracellular killing of Ec.

Hyper(re)active NET Formation by Unstimulated oPMNs
Previous research from our group reported a significant increase in extracellular ROS production by oPMNs when compared to cPMNs from healthy donors (7). Here, we investigated another antimicrobial property of PMNs: the formation of NETs. NET formation by cPMNs has been widely studied and confirmed in vivo and ex vivo using various functional assays. However, the ability of oPMNs to produce NETs was unknown. NET release by cPMNs and oPMNs was visualized (Figures 7A-D) and quantified (Figures 7E,F) using the membrane impermeable extracellular DNA dye SYTOX TM green. Immunofluorescence microscopy analyses showed an absence of NETs in unstimulated (culture medium) cPMNs ( Figure 7A). After stimulation with a well-established NET-inducer PMA (39), both oPMNs and cPMNs were capable of producing NETs (Figures 7C,D). Fluorimetric quantification of NET formation demonstrated significantly more NETs in stimulated conditions (PMA) by cPMNs in comparison to unstimulated control conditions ( Figure 7E). Interestingly, oPMNs showed apparent NET release in both unstimulated (control) and stimulated (PMA) conditions (Figure 7B), where no significant difference was observed between unstimulated and stimulated conditions for cPMNs ( Figure 7F). SYTOX TM green is a non-membrane-permeable nucleic acid dye which also stains Gram positive and negative bacteria (40). Accordingly, Figure 7D shows that NETs produced by oPMNs are saturated with oral bacteria co-isolated from the oral cavity, which illustrates the major function of NETs: entrapping and immobilizing bacteria. Unstimulated oPMNs show 32-fold increased NET formation over unstimulated cPMNs and 13-fold increased NET formation when compared to stimulated cPMNs, demonstrating the hyperactive phenotype of oPMNs.

DISCUSSION
The predominant immune cell that is constitutively recruited into the oral cavity is the PMN, being a protective and antimicrobial innate immune responder. In healthy conditions, about 30,000 PMNs transit per minute from the circulatory blood into the oral mucosal tissues and gingival crevices.
The crucial role of PMNs in (oral) health maintenance has been extensively studied. However, the majority of studies have focused on cPMNs, which may not necessarily reflect the roles and characteristics of PMNs within the oral cavity. Hitherto, little is known about the functional role of orally derived PMNs in relation to oral health and inflammation. Therefore, we aimed to compare oPMNs and cPMNs, in terms of their chemotactic, phagocytic, and NET forming capacities.
The main findings of the current study are that oPMNs obtained from oral rinses are viable, active cells with impaired directional chemotactic accuracy toward fMLP and a pattern of low fMLP receptor expression, rendering them terminally migrated cells. oPMNs adequately perform innate immune responses including adhesion and internalization of various microbes, NET formation, and, as previously investigated by our group, extracellular ROS production. Interestingly, adhesion and internalization by oPMNs of Ec was accompanied by a minimal capacity for intracellular killing. Collectively, the results demonstrate that oPMNs have functional potential. With increased adhesion, internalization and NET formation, oPMNs most likely contribute to maintaining a balanced oral ecosystem. PMNs are short-living cells with an estimated half-life of 6-8 h that remain in the circulation for a few hours before they extravasate into tissues (41). Obviously, oPMNs are more mature cells than cPMNs due to their transendothelial extravasation, oral transepithelial migration and exposure to the oral biofilm. Studies show that oPMNs, which have migrated from the circulatory blood into the oral cavity, have undergone functional changes characterized by the increased production of pro-inflammatory cytokines and enhanced apoptosis mechanisms (7,12). Once within the oral cavity, PMNs are exposed to a lower osmolality in saliva than in blood-rapid cell swelling and eventually lysis of PMNs in response to those osmolities found in saliva have been reported (42). In our setup, all experiments were completed within 3 h after isolation, and cells were tested for viability. oPMNs were washed and stored in a physiological isotonic buffer (same osmolality as blood), minimizing the inhibitory effect of hypotonic saliva. Furthermore, cPMNs incubated with saliva do not exhibit functional responses similar to oPMNs, indicating that the oPMNs' phenotype is altered due to local characteristics such as transmigration and interactions with the oral biofilm and not by the mere exposure to saliva (43).
In vivo, PMNs migrate from the vasculature into the oral cavity in order to affect phagocytosis and the subsequent destruction of pathogens by degranulation, ROS, and NET production. Effective PMN recruitment, invasion, and activation are crucial antimicrobial functions for periodontal health maintenance. In the presence of chronic gingival inflammation, as apparent in periodontitis, oPMNs migrate into the oral cavity at an increased rate (33,44). Hyperactive, hyperreactive, supernumerary, or dysregulated PMNs have been reported as key players in chronic inflammation as they can cause collateral tissue damage through the release of inflammatory and toxic substances or tissue-degrading enzymes (8,45). Therefore, the functional properties of oPMNs were investigated. Impaired directional migration toward fMLP by cPMNs from periodontitis patients in comparison to cPMNs from healthy subjects has been reported by Roberts et al. (16). Here, we investigated oPMN migration toward the chemotactic agent fMLP and showed the dysfunctional movement of viable oPMNs, which was explained by low fMLP receptor expression.
PMNs are recruited from the vasculature into the gingival crevice and adhere to the epithelium which is regulated by intercellular adhesion molecule-1 (ICAM-I), expressed by mucosal keratinocytes (46) and gingival fibroblasts (47). Adhesion and migration across epithelial barriers are facilitated by increased CD11b expression in PMNs (48). Rijkschroeff et al. reported significantly higher CD11b expression by oPMNs when compared to cPMNs from healthy individuals. The expression, and thus migration facilitation of oPMNs, is thus partly managed by CD11b expression and likely exhausted after reaching the oral cavity. Impaired chemotactic accuracy by oPMNs could be explained by their journey through the oral mucosal tissues and exposure to the abundant oral biofilms. PMNs constantly interact with symbiotic and dysbiotic oral biofilms to maintain oral homeostasis and health (49). Exposure to numerous oral bacteria, as apparent in oPMN samples, would explain the low expression of fMLP receptors on oPMNs. Accordingly, a number of the oPMNs' fMLP receptors were likely saturated and therefore incapable of binding further fMLP, rendering PMNs unable to migrate toward fMLP. Almost all viable cPMNs expressed fMLP, explaining the higher sensitivity to fMLP as evident in our migration studies. Thus, oPMNs may be desensitized to fMLP, as was used in our migration assay. In this study, fMLP was chosen as a chemoattractant since it was previously reported as the most effective agent in migration assays, as evidenced by a significant CI, velocity, and speed of cPMNs from healthy donors (16). Here, no significant difference was observed between the speed of oPMNs and cPMNs indicating that oPMNs are viable cells.
Apart from impaired directional migration, decreased cPMN phagocytosis has been demonstrated to play a major role in the etiopathogenesis of periodontitis (50), and thus represents a very relevant function to study in oPMNs. Since phagocytosis is considered to be a multiple-step process, we investigated the adhesion and internalization of various heat-inactivated microorganisms by cPMNs and oPMNs separately. Interestingly, oPMNs have increased phagocytic capacities when compared to cPMNs indicating that oPMNs are effective innate responders with a high potential for phagocytosis. Here, different species were used and showed distinct phagocytic responses by both cPMNs and oPMNs. Overall, the same trends were visible, however, not all microbes were adhered to or internalized similarly, demonstrating the importance of the use of different microbes in phagocytosis assays. Furthermore, significantly more bacteria were internalized than adhered to by both cPMNs and oPMNs. Although oPMNs showed an increased capacity to adhere to and internalize Ec, they were incapable of killing live Ec. It is possible that oPMNs remained capable of ingesting microbes while their actual digestion and destruction capacities were exhausted. We previously demonstrated elevated extracellular ROS production by oPMNs, which is a subsequent step to phagocytosis, further along the activation cascade of the PMN (44). Together with our new findings, showing hyperactive NET formation, we hypothesize that the oPMNs' hyperactive state represents their an ultimate attempt to limit bacterial dissemination. Possibly, as oPMNs are incapable of degrading ingested microbes due to inefficient phagolysosome formation and granule perishment, they proceed to enter their final state: NETosis (51). Nevertheless, as previously shown by others (52,53), some bacteria are capable of escaping NETs.
Lastly, the NET formation capacities of cPMNs and oPMNs were investigated. We demonstrated that oPMNs produce 13 times more NETs than stimulated cPMNs in both unstimulated and stimulated conditions. Thus, after arriving in the oral cavity, oPMNs are in a hyperreactive activation state as evidenced by increased NET and ROS formation activity (44). Unstimulated NET formation capacities could be explained by the previously reported hyperreactive activation state of oPMNs (12,54,55) and by the extracellular environment of oPMNs. In contrast to cPMNs, the extracellular environment of oPMNs is contaminated with saliva, oral bacteria, epithelial cells, and cell debris. Since NETs are reportedly induced by bacterial products (24,56), the hyperactive NET production of oPMNs is likely influenced by the contaminated extracellular environment. Accordingly, oPMNs are constantly stimulated by this extracellular environment containing oral biofilm components and salivary microorganisms, which may explain the significantly increased rates of NET formation, even in unstimulated conditions.
Undoubtedly, a non-vital PMN is unable to migrate or move in a minimal, non-directional fashion, and adhere to or internalize microbes, as shown in our experiments. Extracellular conditions and the activation state of a PMN can influence whether PMNs undergo NETosis or phagocytose a pathogen (51). Ineffective phagocytosis will possibly prompt the production of ROS and NETs. About 20-30% of cPMNs produce NETs in vitro (22,23). It is unlikely that after NETosis, a PMN could subsequently perform its live, antimicrobial functions and then die by apoptosis. Accordingly, evidence exists for distinct phenotypic subsets of neutrophils based on the expression of cell surface markers (12,(57)(58)(59), suggesting large phenotypic heterogeneity and functional versatility. In this study, different subsets were not investigated, since we were interested in the whole PMN population present in the oral cavity and in comparing this population to circulatory PMNs. However, it is conceivable that some oPMNs solely perform phagocytosis, while others are more committed to produce (suicidal) NETs.
In conclusion, the terminally migrated oPMN is a viable cell with a hyperactive phenotype, as evidenced by increased adhesion and internalization of microbes, and NET formation capacities. These findings contribute to a better understanding of the role PMNs play in maintaining oral health.

ETHICS STATEMENT
The study was approved by the Birmingham ethical committee (14/SW/1148) and the Medical Ethical Committee of the Amsterdam University Medical Center, The Netherlands (2012-210#B2012406). All donors were systemically and periodontally healthy. Informed and written consent was obtained from all individuals prior to inclusion. Venous blood samples and oral rinses from each donor were collected.

AUTHOR CONTRIBUTIONS
CM, BL, and EN contributed to the conception and design of this study. CM performed all experiments, analyses, data interpretations, and wrote the first draft of the manuscript. JH helped with chemotaxis, killing, and NET formation experiments. LC assisted with adhesion and internalization experiments. EN and BL supervised the study. CM, JH, LC, IC, BL, and EN contributed to manuscript revisions, read, and approved the submitted version.

ACKNOWLEDGMENTS
The authors are grateful to Sander Spiekstra for his assistance with flow cytometry. They also thank Kaj Emanuel for his MATLAB expertise.

SUPPLEMENTARY MATERIAL
The shown and was determined with the Muse TM count & viability kit, which makes use of a membrane-permeant DNA staining dye to distinguish between (nucleated) live and dead cells. Viability is shown on the x-axis for all plots. Cell size index is shown on the y-axes of (C,E), which distinguishes the cells (green area) from debris (gray area). Nucleated cells are displayed on the y-axes of (D,F), discriminating the live nucleated cells (green area) from the non-nucleated dead cells and debris (gray area).