Neutrophil-derived extracellular vesicles modulate the phenotype of naïve human neutrophils

Abstract

Neutrophils (PMN) regulate inflammation in many ways, including communication with other immune cells via extracellular vesicles (EVs). EVs released by human neutrophils activated with N-formylmethionyl-leucyl-phenylalanine (fMLF) (PMN-fMLF EVs) had an outside-out orientation and contained functionally important neutrophil plasma membrane proteins, including flavocytochrome b558, and enzymatically active granule proteins, elastase and myeloperoxidase. Treatment of naïve PMN with PMN-fMLF EVs primed fMLF-stimulated NADPH oxidase activity, increased surface expression of the complement receptors CD11b/CD18 and CD35, the specific granule membrane protein CD66, and flavocytochrome b₅₅₈, and promoted phagocytosis of serum-opsonized Staphylococcus aureus. The primed oxidase activity reflected increased surface expression of flavocytochrome b558 and phosphorylation of SER345 in p47^phox, two recognized mechanisms for oxidase priming. Taken together, these data demonstrate that stimulated PMN released EVs that altered the phenotype of naïve phagocytes by priming of the NADPH oxidase activity and augmenting phagocytosis, two responses that are integral to optimal PMN host defense.

Graphical Abstract

Summary sentence:

Extracellular vesicles from fMLF-stimulated human neutrophils prime the NADPH oxidase and promote surface expression of functionally important membrane proteins and phagocytosis of Staphylococcus aureus.

Introduction

Neutrophils (polymorphonuclear phagocytes or PMN) are the first cellular responders to microbial invasion and drive an exuberant inflammatory response after migrating to a site of infection [1]. In addition to their well-established ability to phagocytose and directly kill microbes, PMN communicate with other immune cells, including other phagocytes, to participate in regulation of the immune response [2, 3]. By signaling to human macrophages to clear spent PMN from sites of infection and by secreting cytokines, PMN actively contribute to the orchestration of the acute inflammatory response in a variety of clinical settings, including infection [4]. In contrast to the extensive study of PMN-macrophage communication, data on signaling between PMN are limited.

Agonists such as N-formylmethionyl-leucyl-phenylalanine (fMLF) induce PMN to release extracellular vesicles (EVs) [reviewed in [5]], which are particles enveloped in a lipid bilayer [6] and have the capacity to transfer mediators such as proteins, lipids, and nucleic acids [7-9] and alter gene expression in target cells [10, 11]. Their biogenesis, structural features, and effects on target cells greatly depends on several variables, including the agonist prompting their generation, the functional state of the PMN during stimulation, and the environmental conditions [12, 13]. EVs from PMN act upon multiple cell types with varied responses [9, 14, 15] but, based upon their effects on phagocytes, are often characterized as anti-inflammatory. For example, EVs released from fMLF-stimulated human PMN do not increase IL-8 or TNFα secretion by human monocyte-derived macrophages (HMDM) but instead prompt release of transforming growth factor β1. Furthermore, PMN-derived EVs block HMDM release of proinflammatory cytokines in response to endotoxin or zymosan [16, 17].

To explore how PMN-derived EVs might change the functional properties of naïve PMN, we recovered EVs from PMN stimulated with fMLF (PMN-fMLF EVs). PMN-fMLF EVs had an outside-side out orientation and possessed functionally important plasma membrane proteins and enzymatically active human neutrophil elastase (HNE) and myeloperoxidase (MPO). Naïve PMN exposed to PMN-fMLF EVs exhibited priming of the NADPH oxidase, increased surface expression of several functionally important membrane proteins, and increased capacity to ingest opsonized Staphylococcus aureus. These data demonstrate that PMN-derived EVs can alter the phenotype of naïve PMN and thereby locally modify inflammation.

Methods

Reagents and antibodies:

PMN isolation used heparin from APP Pharmaceuticals LLC (Schaumburg, IL), dextran T500 from Pharmacosmos (Holbaek, Denmark), Ficoll-Hypague PLUS from GE Healthcare (Piscataway, NJ), and endotoxin-free water and 0.9% sodium chloride from Baxter (Deerfield, IL). Hanks’ Balanced Salt Solution (HBSS; HBSS^+/+ and HBSS^−/− signify with or without Ca⁺⁺ and Mg⁺⁺) and phosphate buffered saline (PBS; PBS^+/+ and PBS^−/− with above meanings) were purchased from Mediatech (Manassas, VA); PBS was adjusted to pH 7.4 prior to use. Roswell Park Memorial Institute (RPMI) 1640 medium without phenol red (with L-glutamine; 25 mM 4-(2-Hydroxyethyl)-1-piperazineethanesulfonic acid [HEPES] added prior to use), HEPES (Gibco), and aldehyde/sulfate latex beads (4 μm diameter) were from Thermo Fisher Scientific (Waltham, MA), as were fetal bovine serum and goat serum. Bovine serum albumin (BSA) was from ICN Biomedicals Inc. (Aurora, OH). Human serum albumin (HSA) (25%) was purchased from Talecris Biotherapeutics (Raleigh, NC). Triton X-100 (Tx-100) was purchased from Fisher Biotech (Fair Lawn, NJ). Tryptic Soy Broth and tryptic soy agar were from BD Biosciences (San Jose, CA). fMLF, phorbol 12-myristate 13-acetate (PMA), cytochrome C, Cell Vue Claret, oxalic acid, methyl cellulose, and superoxide dismutase (SOD) were from Sigma-Aldrich (St. Louis, MO). Micro bicinchoninic acid (BCA) Assay kit was from Bio-Rad Laboratories (Hercules, CA). Sodium phosphate buffer, paraformaldehyde, uranyl acetate, formvar-carbon coated EM grids were from Polysciences (Warrington, PA, USA), and TEM-grade glutaraldehyde was from Ted Pella (Redding, CA).

Antibodies used:

Goat anti-mouse fluorescein (FITC), anti-human CD63-phycoerythrin (PE), and anti-phospho-p47^phox-SER345 were from Thermo Fisher Scientific. Anti-human CD63-PE and anti-human CD35-FITC used for EV flow cytometry were from BioLegend (San Diego, CA). Anti-human CD11b-PE, anti-human CD35-FITC, and anti-human CD66b-FITC used for PMN flow cytometry were from BD Biosciences (San Jose, CA). HNE antibody, goat anti-mouse horseradish peroxidase (HRP) and goat anti-rabbit HRP were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). Goat anti-mouse HRP was from Bio-Rad. Monospecific rabbit polyclonal antibody to human MPO was previously described [18]. Monoclonal antibodies 54.1 and 44.1, specific for gp91^phox and p22^phox [19] respectively, were provided by A. J. Jesaitis (Montana State University, Bozeman, MT). The murine monoclonal antibody 7D5 [20, 21] was purified in our lab by affinity chromatography (gamma bind sepharose) of spent culture media.

PMN isolation:

Human PMN were isolated from venous blood using dextran sedimentation and Hypaque-Ficoll density gradient separation [22] and used in experiments within 1 hour after isolation. Preparations were 92 to 95% PMN. Written consent was obtained from each volunteer blood product donor in accordance with a protocol approved by the Institutional Review Board for Human Subjects at the University of Iowa.

Generation of PMN-fMLF EVs:

PMN were incubated in RPMI with 10⁻⁶M fMLF, for 20 min at 37°C. Intact PMN were removed by centrifugation at 4,000 × g and filtered (0.45 μm) supernatant was ultracentrifuged at 200,000 × g (SW-41 Ti for 55 min followed by SW-55 Ti for 30 min) to sediment and concentrate the EVs. EV samples were resuspended in a small volume (100-200 μL) of PBS^−/−, aliquoted, flash frozen, and stored at −80°C until use. Before use in experiments, EVs were thawed on ice. EV protein concentration was measured by bicinchoninic acid (BCA) assay, and isolated EVs were characterized by transmission electron microscopy (TEM), flow cytometry, and 0.05% Triton X-100 lysis as described in [23].

Activity assays of EV-associated enzymes:

Alkaline phosphatase activity was measured spectrophotometrically in EVs as described in [24]. Briefly, EVs with and without 2% Tx-100 were incubated at room temperature for 15 min. Alkaline phosphatase substrate p-nitrophenylphosphate (pNPP) was added to each sample and incubated at room temperature for 30 min. Absorbance was read at 405nm using a SpectraMax Plus (Molecular Devices; San Jose, CA). HNE activity of EVs was measured as release of the fluorophore aminomethyl coumarin from a non-fluorescent substrate as described [25]. MPO activity was quantitated as chlorination of taurine in the presence of hydrogen peroxide as previously described [26], performed in triplicate for each replicate.

Immunoblotting of EVs:

PMN-fMLF EVs were separated by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) under reducing conditions and transferred to nitrocellulose membranes for immunoblotting. Murine monoclonal antibodies against gp91^phox, p22^phox, and HNE were used as primary antibodies. For detection of MPO, a high-titer monospecific rabbit primary antibody was used.

Superoxide anion measurement:

Activity of the NADPH oxidase was assessed by quantitating extracellular superoxide generation as superoxide dismutase-inhibitable reduction of ferricytochrome C. PMN in suspension were primed with vehicle (PBS^−/−), 100 ng/mL lipopolysaccharide (LPS) plus 50 ng/mL LPS-binding protein (LBP), or PMN-fMLF EVs (10 μg/mL), by tumbling for 20 min at 37°C. PMN from each condition were then tumbled (1 × 10⁶ cells/mL, in triplicate) for 20 min at 37°C in PBS^+/+ with cytochrome C (100 μM), without or with fMLF (100 nM), and without or with SOD (50 μg/mL). PMN were pelleted at 500 × g, and supernatants diluted 1:4 before measuring absorbance at 550 nm using a Lamda35 spectrophotometer. The amount of superoxide anion was defined as SOD-inhibitable reduction of ferricytochrome C and calculated using the extinction coefficient of 21.1 mM⁻¹cm⁻¹ [27].

Flow cytometric assessment of plasma membrane protein expression in PMN:

PMN were primed for 40 min as described above. For detection of flavocytochrome b₅₅₈, PMN (1 × 10⁶ cells/aliquot, all conditions in duplicate) were incubated for 10 min in 10% pooled human serum in flow cytometry buffer consisting of PBS^−/− with 2% goat serum and 0.5 mM EDTA for 10 min to block nonspecific binding of antibodies to PMN, and sequentially with 7D5 (50 μg/mL), a murine monoclonal antibody directed against an extracellular epitope of gp91^phox [21] for 2h at 4°C, and a FITC-conjugated anti-mouse secondary antibody for 30 min in the dark at RT, as done previously [28]. Membrane expression of flavocytochrome b₅₅₈ on PMN was analyzed by flow cytometry as described above.

For detection of other surface proteins, PMN (1 × 10⁶ cells/aliquot, all conditions in duplicate) were incubated in PBS^+/+ with 4% fetal bovine serum and 2% nonfat dry milk for 10 min to block nonspecific binding, followed by treatment for 1h at 4°C with fluorescent-conjugated antibodies (FITC-CD35, PE-CD11b, FITC-CD66, or PE-CD63) at the concentrations recommended by the manufacturer. Staining of PMN membrane proteins was assessed by flow cytometry.

Immunoblotting for SER345 p47^phox:

In parallel with measurements of priming of the NADPH oxidase, suspensions of PMN were solubilized, and proteins separated by SDS-PAGE under reducing conditions and transferred to nitrocellulose membranes for immunoblotting. Blots were probed with an antibody against phospho-p47^phox-SER345 and donkey anti-rabbit IgG-HRP followed by Pierce Femto Reagent. Images were acquired and quantitated using Odyssey (Li-Cor, Lincoln, NE). Images of p47^phox phosphorylated on SER345 were acquired and quantitated. Blots were subsequently stripped and probed with an antibody against p47^phox that was generated in our laboratory [29]. Images of total p47^phox were acquired and quantitated, and the amount of p47^phox phosphorylated on SER345 under the specific experimental conditions was normalized for the total amount of p47^phox immunoblotted. The fold changes of normalized p47^phox-SER345 under various conditions was calculated relative to that in PBS-treated PMN.

Phagocytosis assay:

Phagocytosis of green fluorescent protein (GFP)-expressing S.aureus was performed as previously described [30, 31]. For experiments using serum-opsonized bacteria, mid-logarithmic phase USA300 (LAC) S.aureus suspended in HBSS^+/+ were tumbled for 20 min at 37°C at the desired concentration with 10% pooled human serum. Phagocytosis experiments were performed in the presence of 10 μM diphenyleneiodonium (DPI) in order to block activation of the PMN NADPH oxidase [32] and subsequent generation of HOCl, which would otherwise bleach the GFP [30]. DPI does not alter phagocytosis by PMN [30].

PMN in suspension were incubated together with PMN-fMLF EVs (10 μg/mL) at 37°C for 10 min, after which DPI and opsonized GFP-labelled USA300 LAC S.aureus were added at 0.5:1 or 2:1 multiplicity of infection (MOI). Suspensions were tumbled for 10 min at 37°C and then centrifuged at 500 × g to isolate PMN containing ingested S.aureus. All PMN flow cytometry was performed as done previously [31] and using an Accuri C6 Plus flow cytometer with 488 nm laser, gating PMN based on size and performing additional analysis using the FlowJo program (FlowJo LLC). The GFP signal was measured to determine both the percentage of PMN that had phagocytosed GFP-expressing S.aureus (% phagocytosis) and the geometric mean fluorescence of the GFP+ PMN (a reflection of the number of GFP-expressing S.aureus per PMN, and thus a proxy for the phagocytic index).

Statistical Analysis:

Statistical differences between two groups were calculated using a paired Student’s t test. Statistical differences between three or more groups were calculated using a one-way ANOVA and either a Tukey posttest for multiple comparisons or Dunnett posttest for comparisons with the control. Where applicable, the statistical tests and p-values of posttest analyses are included in the text or the figure legends.

Results and Discussion

PMN-fMLF EVs contain PMN plasma membrane proteins and PMN granule proteins

To establish that the material recovered from fMLF-stimulated PMN represented EVs, we used complementary analytical techniques to determine if it met criteria for EV established in current guidelines [6]. Following total protein quantification by BCA, material recovered by differential ultracentrifugation of supernatants from fMLF-stimulated PMN was visualized by TEM. Round and semi-round structures between 100-200 nm were visible (Figure 1A, upper panel) and were eliminated by treatment with 0.05% Triton X-100 (Figure 1A, lower panel), an appearance and behavior consistent with EVs [6]. Furthermore, the lipophilic dye CellVue Claret stained the structures conjugated to beads and Triton X-100 treatment significantly decreased staining (Figures 1B and C). In addition, we measured two proteins on CellVue⁺ beads: CD63, a tetraspanin protein found at the cell surface, in multi-vesicular bodies (MVBs), and in small EVs [33], and CD35, a protein found at the cell surface and in secretory vesicles. CD63 stained the majority of CellVue⁺ beads (90.0% ± 6.7%), whereas CD35 was only present on a subset of CellVue⁺ CD63⁺ beads and its staining was more variable (30.3% ± 13.5%, Figure 1D). Enrichment in tetraspanin protein CD63 and detergent-sensitivity of staining for membrane proteins, features characteristic of EVs, suggest that the vesicles represented fMLF-PMN EVs arising from both the plasma membrane and inward budding multi-vesicular bodies.

Figure 1. Characterization of PMN-fMLF EVs by TEM, flow cytometry, and immunoblot.

PMN-fMLF EVs (top) and Tx 100-treated PMN-fMLF EVs (bottom) were positively-negatively stained and imaged using TEM at 10,000x magnification. Shown is a representative micrograph from three separate experiments (scale bar = 500 nm, A). In parallel, EVs were conjugated to beads, and stained with CellVue, anti-CD35-FITC, and anti-CD63-PE, and analyzed by flow cytometry. Representative dot plots of PMN-fMLF EVs (top) and Tx-100-treated PMN-fMLF EVs (bottom) are shown in (B) and quantification is shown in (C, n = 3). Data are expressed as % of CellVue⁺beads (right, paired Student’s t test, *** p < 0.001) and as % of CellVue⁺beads that were either single-positive for CD63 or double positive for CD63 and CD35 (bars represent mean ± SEM, n = 3). PMN-fMLF-EVs (3 μg for D, 4 μg for E, and 1.6 μg for F) were separated by SDS-PAGE and immunoblotted for (D) gp91^phox and p22^phox, (E) HNE, and (F) MPO. Arrows indicate the expected size of each protein on representative blots from five (D), four (E), or two (F) experiments. The unlabeled band on panel (F) is presumed to represent MPO-derived artifact associated with sample heating. Images have been cropped to contain the relevant lane. Both HNE and MPO were enzymatically active (see text).

In order to determine the orientation of the EVs, we assessed the latency of the integral membrane protein alkaline phosphatase. Intact PMN express alkaline phosphatase in two distinct locations, the plasma membrane and in secretory vesicles but with opposite orientations; in plasma membrane (and vesicles derived from plasma membrane), the catalytic domain of the enzyme is on the exterior, accessible to substrate, whereas the active site is intraluminal in secretory vesicles [24]. Consequently, detection of the enzymatic activity of alkaline phosphatase in secretory vesicles requires solubilization of the vesicle membrane, a phenomenon known as enzyme latency [34]. Because the substrate for the alkaline phosphatase assay, pNPP, is membrane-impermeable and because the access of substrate to catalytic domain has been used previously to assess the orientation of vesicles from PMN [35], we reasoned that the relative access of active alkaline phosphatase to substrate in the absence or presence of detergent solubilization of EVs would indicate their orientation. The addition of Tx-100 to PMN-fMLF EVs did not change alkaline phosphatase activity (1.5 ± 0.5 without Tx-100 vs. 1.3 ± 0.5 with Tx-100, expressed as mean ± SEM in A₄₅₀ units; n = 4). In contrast, and consistent with previous published work [35], the alkaline phosphatase activity of a subcellular fraction of PMN recovered from a Percoll density gradient and containing a mixture of plasma membrane and secretory vesicles [36] increased 4.5 ± 0.03 fold (p = .0001, n = 4) after the addition of Tx-100. Taken together, these data indicate that the EVs had the same outside-out orientation as does plasma membrane.

PMN-fMLF EVs contained both PMN plasma membrane and granule proteins. PMN-fMLF EVs expressed the NADPH oxidase plasma membrane components gp91^phox and p22^phox and the azurophilic granule proteins HNE and MPO as detected by immunoblotting (Figure 1D - F). Furthermore, HNE and MPO associated with PMN-fMLF EVs retained enzymatic activity. Elastase activity of isolated EVs was 14,904 ± 1,480 RLU/μg, whereas activity of purified HNE activity was 28,831 ± 4,329 RLU/μg and that of isolated PMN granules 27,875 ± 2,448/5 × 10⁻⁴ cell equivalents (n = 3 for all determinations). Active MPO was quantified as the capacity of EVs to generate HOCl in the presence of exogenous hydrogen peroxide and chloride [37]. 5 μg of EV protein generated 2.31 ± 0.39 nmoles HOCl (n = 3). As a frame of reference, 100 fmoles of purified MPO generated 0.86 ± 0.01nmoles HOCl (n = 5).

PMN-fMLF EVs prime the NADPH oxidase in naïve PMN

Activation of the NADPH oxidase represents one of the cellular responses critical to PMN-mediated host response. To determine if PMN-fMLF EVs influence NADPH oxidase activation in naïve PMN, we compared the ability of PMN-fMLF EVs to prime superoxide production from PMN with that of the established priming agent LPS/LBP [28], using SOD-inhibitable ferricytochrome C reduction to quantitate fMLF-induced superoxide anion production. LPS/LBP treatment enhanced fMLF-stimulated superoxide anion production approximately 2.5-fold (Figure 2). Similarly, PMN-fMLF EVs (10 μg/mL) primed fMLF-induced NADPH oxidase activity in naïve PMN (Figure 2). PMN-fMLF EVs did not prime PMN at 5 μg/mL (1.2 ± 0.2 -fold increase, n = 3, NS) , and PMN-fMLF EVs did not directly elicit superoxide production by PMN (0.93 ± 0.23 nmoles/10⁶ PMN vs 1.76 ± 0.62 nmoles/10⁶ PMN with PBS alone, n = 5, NS). These data demonstrate that EVs from fMLF-activated PMN served as a priming agent for NADPH oxidase activation in naïve PMN.

Figure 2. PMN-fMLF EVs enhance superoxide anion generation by PMN.

PMN were treated for 20 min at 37°C with PBS, LPS (100 ng/mL)/LBP (50 ng/mL), or PMN-fMLF EVs (10 μg/mL, labelled in Figures as fMLF-EVs). Each group of PMN was incubated for 20 min at 37°C with ferricytochrome C (100 μM) in the presence of fMLF (10⁻⁷ M) and the presence or absence of SOD (50 μg/mL). Absorption at 550 nm was measured spectrophotometrically and superoxide anion was defined as SOD-inhibitable ferricytochrome C reduction. Bars represent the average of 17 experiments ± SEM. p-values were determined using a repeated measures one-way ANOVA and Tukey posttest (** p < 0.01 and **** p <0.0001 vs. PBS).

PMN-fMLF EVs promote surface expression of flavocytochrome b558, complement receptors and phosphorylation of SER345 in p47^phox in naïve PMN

At least two cellular responses mediate priming of the phagocyte NADPH oxidase, [reviewed in [38]], namely granule exocytosis with concomitant redistribution of flavocytochrome b558 from the membranes of secretory vesicles and secondary granules to the plasma membrane [39] and targeted phosphorylation of SER345 in p47^phox [reviewed in [40]], as previously reported for priming with either TNFα [41] or LBP-LPS [42].

To determine if PMN-fMLF EVs prompted redistribution of membrane proteins, which would contribute to the observed priming of oxidase activity in naïve PMN by PMN-fMLF EVs, we assessed changes in expression of functionally important membrane proteins on the surface of PMN after exposure to PMN-fMLF EVs. Flow cytometry with the antibody 7D5, which specifically recognizes an extracellular epitope of flavocytochrome b₅₅₈ [21], demonstrated that treatment of PMN with PMN-fMLF EVs prompted redistribution of flavocytochrome b₅₅₈ from secretory vesicles and specific granules [28] to the cell surface (Figure 3A). Supporting this evidence that secretory vesicles fused with plasma membrane after PMN-fMLF EV treatment, PMN treated with LPS/LBP or with PMN-fMLF EVs (10 μg/mL) upregulated surface expression of CD11b (Figure 3B) and the secretory vesicle integral membrane protein CD35 (Figure 3C) [43]. Concomitantly, PMN-fMLF EVs also elicited limited degranulation of naïve PMN, as demonstrated by cell surface expression of CD66 and CD63, markers of secondary (specific) and primary (azurophilic) granules, respectively. PMN-fMLF EVs upregulated PMN surface expression of CD66 (Figure 3D) but did not prompt PMN surface expression of CD63 (Figure 3E). Stimulation of exocytosis with PMA (100 ng/mL) promoted increased expression of both CD66 and CD63, an expected response to such a potent agonist. These data show that PMN-fMLF EVs triggered plasma membrane fusion of secretory vesicles and specific granules, events that likely contributed to their capacity to prime the fMLF-stimulated NADPH oxidase activity.

Figure 3. Flow cytometry assessments of flavocytochrome b558 and cell surface markers on PMN-fMLF EV-treated PMN.

PMN were treated for 40 min at 37°C with PBS, LPS (100 ng/mL)/LBP (50 ng/mL), or PMN-fMLF EVs (10 μg/mL). (A) PMN were labelled for flow cytometry with the anti-flavocytochrome b₅₅₈ antibody 7D5 and a FITC-conjugated secondary antibody. FITC signal was quantified by flow cytometry. Data expressed as the geometric mean fluorescence of FITC⁺ cells and bars show the mean of four experiments ± SEM. p-values were determined using a paired Student’s t-test (* p < 0.05). (B) PMN were treated as in panel A or (in panels D and E) with PMA (100 ng/ml). Flow cytometry was performed using the following antibodies conjugated to fluorescent probes: (B) CD11b-PE, (C) CD35-FITC, (D) CD66-FITC, (E) CD63-PE. Data are expressed as the geometric mean fluorescence in FITC⁺ or PE⁺PMN, and bars represent mean ± SEM (n = at least three experiments). p-values were determined using a repeated measures one-way ANOVA and Dunnett posttest (* p < 0.05, ** p < 0.01 vs. PBS).

As noted, targeted phosphorylation of SER345 in p47^phox has been implicated in priming of fMLF-stimulated NADPH oxidase by TNFα or LPS [40-42]. To determine if PMN-fMLF EVs stimulate phosphorylation of SER345 in p47^phox, we measured priming of fMLF-stimulated oxidase activity and phosphorylation of p47^phox on SER345, in parallel on PMN treated with buffer, LBP-LPS, or PMN-fMLF EVs before subsequent stimulation with fMLF. Both LBP-LPS and PMN-fMLF EVs were equally effective in augmenting superoxide production in response to a suboptimal concentration of fMLF (LBP-LPS 3.5 ± 0.9-fold vs 4.2 ± 0.4-fold, n = 6) and both LBP-LPS (1.72 ± 0.34-fold, n = 3) and PMN-fMLF EVs (2.33 ± 0.49-fold, n = 3) increased phosphorylation of SER345 (Figure 4). Taken together, these data indicate that PMN-fMLF EVs increased increased surface expression of flavocytochrome b558 and phosphorylation of SER345 in p47^phox, two recognized mechanisms for oxidase priming, to prime activity of the NADPH oxidase in response to fMLF.

Figure 4. Analysis of p47^phox phosphorylation at SER345.

PMN lysates were separated by SDS-PAGE and immunoblotted for phosphorylated SER345 in p47^phox. The amount of p47^phox phosphorylated on SER345 detected was normalized to the total amount of p47^phox immunoblotted and compared to that for PBS-stimulated PMN. Both LBP-LPS (1.72 ± 0.34-fold, n = 3) and PMN-fMLF EVs (2.33 ± 0.49-fold, n = 3) increased phosphorylation of SER345. Shown is a representative of three experiments.

PMN-fMLF EVs enhance phagocytic function of naïve PMN

Because complement receptors (CRs) contribute to the phagocytosis of serum-opsonized particles and PMN-fMLF EVs stimulated PMN translocated CD11b/CD18 (the β2 integrin receptor CR3) and CD35 (CR1) to their cell surface, we examined the effects of PMN-fMLF EV exposure on phagocytosis by naïve PMN. We fed serum-opsonized GFP-expressing S.aureus (MOI 0.5:1) to PMN that had been pretreated for 10 min with PBS or PMN-fMLF EVs (10 μg/mL) and quantified GFP-positive PMN by flow cytometry. More PMN engaged in phagocytosis in EV-treated PMN (Figures 5A and 5B). PMN-fMLF EVs did not increase the fluorescent intensity among GFP⁺ PMN (Figure 5C), suggesting that PMN-fMLF EVs promoted phagocytosis in a subset of PMN but did not alter their phagocytic index under these conditions. Doubling the concentration of PMN-fMLF EVs did not increase the percent of PMN that ingested S.aureus or the fluorescence of each PMN containing S.aureus (data not shown). However, increasing the MOI to 2:1 not only resulted in a higher percentage of PMN ingesting S.aureus after PMN-fMLF EV treatment (Figure 5D) but also promoted more S.aureus ingestion per PMN (Figure 5E). PMN-fMLF EVs increased the percentage of naïve PMN that ingested opsonized S.aureus, likely due in part to the concomitant increased surface expression of complement receptors CD11b/CD18 and CD35 prompted by exposure to PMN-fMLF EVs. The heterogeneity in PMN response to PMN-fMLF EVs may reflect the phenotypic diversity of circulating PMN and the presence of PMN subpopulations [44] or simply the stochastic nature of PMN encountering S.aureus while in suspension, with both the percentage of phagocytosing PMN and the organisms ingested per PMN increased at the higher MOI.

Figure 5. PMN-fMLF EVs enhance PMN phagocytosis.

PMN were pretreated with PBS or PMN-fMLF EVs (10 μg/mL) for 10 min prior to being fed serum-opsonized GFP-expressing SA (MOI 0.5:1 for A-C and MOI 2:1 for D-E) for 10 min. After centrifugation to isolate PMN that had ingested S.aureus, association of GFP-expressing S.aureus with intact neutrophils was measured by flow cytometry. PMN were gated based off size discrimination, thereby eliminating any unassociated S.aureus. Panel (A) shows representative histograms of fluorescence of PMN alone without S aureus (left), and of PBS-treated (middle) and PMN-fMLF EV-treated (right) PMN after ingestion of S aureus at 0.5:1 MOI. Phagocytosis was defined by % of PMN that are GFP⁺ (B and D), and the geometric mean fluorescence of GFP⁺ PMN (C and E). Bars represent mean ± SEM (n = 4 for B-C and n = 3 for D-E). p-values were determined using a paired Student’s t test (** p < 0.01 and *** p < 0.001).

In summary, we demonstrate that PMN-fMLF EVs contained several PMN plasma membrane and enzymatically active granule proteins and displayed the same outside-out orientation as that of the plasma membrane. Furthermore, we provide evidence for the novel observations that PMN-fMLF EVs prime the NADPH oxidase and phagocytic activity of naïve PMN via stimulated fusion of secretory vesicles and specific granules along with targeted phosphorylation of SER345 on p47^phox.

Although our data implicate PMN-fMLF EVs as the mediators of the phenotypic changes in naïve PMN, it is possible that there are at least two other factors contribute. First, trace amounts of fMLF, the stimulus used to promote EV generation, may be present in the final PMN-fMLF EV preparation. Whereas concentrations on the order of that used to stimulate PMN would trigger homologous desensitization of naïve PMN to subsequent stimulation with fMLF, the presence of fMLF at < 100 nM would prime the NADPH oxidase of PMN [45]. Furthermore, fMLF in the 10 to 100 nM range can enhance phagocytosis of opsonized S. aureus [46]. Such low concentrations of fMLF are not reliably quantitated, thus making it difficult to exclude this possibility with confidence. Second, it is possible that the initial stimulation with fMLF prompted PMN or the small percentage of contaminating leukocyte subsets (5 to 8%) to release bioactive molecules that could alter responses of naïve PMN to subsequent stimulation with fMLF. Addressing this possibility experimentally would require a wide survey of cytokines, chemokines, growth factors, and other bioactive molecules that are secreted by stimulated leukocytes and might be present in the PMN-fMLF EVs, followed by detailed analyses to implicate each as a potential causative agent in the observed priming. Even if feasible, such studies are beyond the scope of the work described. Either trace amounts of fMLF or leukocyte-derived bioactive molecules could act together with PMN-fMLF-EVs to augment phagocytosis and fMLF-mediated NADPH oxidase activity of naïve PMN.

Nonetheless, these data provide insights into how EVs from PMN might mediate PMN to PMN signaling and thus participate in regulation of the innate immune system in infection and inflammation.

Abbreviation list

BCA

bicinchoninic acid

BSA

bovine serum albumin

CR

complement receptor

DPI

diphenyleneiodonium

EV

extracellular vesicle

FITC

fluorescein

fMLF

N-formylmethionyl-leucyl-phenylalanine

GFP

green fluorescent protein

HBSS

Hanks’ buffered salt solution

HEPES

4-(2-Hydroxyethyl)-1-piperazineethanesulfonic acid

HNE

human neutrophil elastase

HRP

horseradish peroxidase

HSA

human serum albumin

HMDM

human monocyte-derived macrophages

LBP

LPS-binding protein

LPS

lipopolysaccharide

MOI

multiplicity of infection

MPO

myeloperoxidase

PBS

phosphate buffered saline

PE

phycoerythrin

PMA

phorbol 12-myristate 13-acetate

PMN

polymorphonuclear leukocytes

PMN-fMLF EVs

EVs from fMLF-treated PMN

RLU

relative luminescence units

RPMI

Roswell Park Memorial Institute medium

SDS-PAGE

sodium dodecyl sulfate polyacrylamide gel electrophoresis

SOD

superoxide dismutase

TEM

transmission electron microscopy