The investigations were approved by the Local Independent Ethics Committee of the University of Ulm (number 459/18; 94/14). After obtaining informed written consent from healthy human volunteers, blood was drawn into syringes containing 3.2% trisodium citrate (Sarstedt, Nürnbrecht, Germany). Neutrophils were isolated by Ficoll-Paque (GE Healthcare, Uppsala, Sweden) density gradient centrifugation and subsequent dextran sedimentation followed by hypotonic lysis of remaining erythrocytes, as described previously (14–16). Polymorphonuclear granulocytes (mainly consisting of neutrophils) were adjusted to a concentration of 2 × 10⁶ cells/ml using Hank's balanced salt solution with calcium and magnesium (HBSS, Thermo Fisher, Darmstadt, Germany), the pH of which was adjusted to 7.3, when not indicated otherwise.

Isolated neutrophils were incubated in a light-protected water bath at 37°C with the fluorescent dyes bis(1,3-dibutylbarbituric acid) trimethine oxonol (DiBAC₄(3)), 50 nM, or 5-(and-6)-carboxy-SNARF-1 (SNARF), 1 μM (Thermo Fisher) for 20 min or with dihydrorhodamine 123 (D123) (Santa Cruz Biotechnology, Heidelberg, Germany), 1.8 μM, for 30 min to determine MP (16), pH_i, and ROS generation [what is measured by D123 is the byproduct H₂O₂ which is itself derived from O2•-, a product generated by NOX activity (37)], respectively. To determine porcine neutrophil pH_i, the cells were instead incubated with 2′,7′-bis-(2-carboxyethyl)-5-(and-6)-carboxyfluorescein, acetoxymethyl ester (BCECF, Abcam, Cambridge, United Kingdom), 25 nM, for 30 min, because exposure of porcine neutrophils to SNARF resulted in changes in forward scatter area (FSC-A), which was interpreted as possible cellular activation (data not shown). Following incubation, neutrophils were centrifuged for 5 min (340 g) and resuspended in Roswell Park Memorial Institute medium (RPMI) with magnesium and calcium with a pH adjusted to 7.3. DiBAC₄(3) was again added. After a resting period of 10 min in a light-protected water bath at 37°C, neutrophils were stimulated with 1 μM PAF or 100 ng/ml complement factor 5a (C5a, Complement Technology, Tyler, Texas, USA), when not indicated otherwise. Cells were analyzed using a Canto II flow cytometer (BD Biosciences, Heidelberg, Germany). Neutrophils were identified by FSC-A and sideward scatter area. Non-fluorescent reference microspheres (10, 15, and 20 μm diameter; Polybead® Polystyrene Microsphere, Polysciences, Hirschberg, Germany) were used to quantify the shape changes assessed by the FSC-A (14). Differences in absolute FSC-A values between experiments are explained by recalibration of the flow cytometer between different experimental series. The fluorescence of DiBAC₄(3) were converted in changes of the MP as described previously (16) with a small modification: RPMI with varying extracellular concentrations of sodium and potassium were used to calculate the amount of change in DiBAC₄(3) fluorescence per mV.

Near-real time kinetics were measured by acquiring neutrophils for 10 min continuously embedded in a heating unit (TC-1234A Temperature Controller, Warner Instruments LLC, Holliston, Massachusetts, USA), ensuring a steady temperature level of 37°C. Analysis was performed by a specifically designed algorithm using the statistic language R (R Core Team, Vienna, Austria).

Isolated neutrophils were incubated in a light-protected water bath at 37°C. Following PAF stimulation as indicated above (normally 1 μM) for 10 min, neutrophils were stained with 0.1 μg/ml APC anti-mouse/human CD11b antibody (#101212, Biolegend, San Diego, USA) and 0.5 μg/ml PE anti-human CD62L antibody (#304806, Biolegend) for 5 min at room temperature. Cells were fixed using FACS Lysing Solution (BD, Heidelberg, Germany). Proper isotype controls were used (CD11b: APC Rat IgG2b, κ Isotype Ctrl; CD62L: PE Mouse IgG1, κ Isotype Ctrl; Biolegend).

PAF-induced effects on neutrophils were analyzed in the presence of the subsequently listed pharmacological modulators (their proposed targets are given in brackets): Amiloride [200 μM, 10 min, Na⁺-channels (15)], BIX [5 μM, 30 min, sodium-proton exchanger 1 (NHE1) (38); Tocris, Wiesbaden, Germany], 5-Nitro-2-(3-phenylpropylamino)benzoic acid [NPPB, 100 μM, 10 min, Cl⁻ channels (39)], and ebselen [10 nM, 30 min, glutathione peroxidase and peroxiredoxin enzyme mimetic and NOX2 inhibitor (40, 41); Tocris]. Following pre-incubation, samples were divided to obtain paired results and treated as control or stimulated with PAF.

Following the described staining and centrifugation, neutrophils were resuspended in RPMI adjusted to a pH of 6.6, 7.0, 7.4, or 7.8 to simulate extracellular alkalosis or acidification. Neutrophils were incubated for 10 min at 37°C and subsequently stimulated with PAF.

Isolated neutrophils were stimulated for 10 min, diluted 1:500 with RPMI adjusted to a pH 7.3 and measured by a cell counter working through electronic current exclusion (Cell Counter CASY, OLS OMNI Life Science, Bremen, Germany).

Neutrophil chemotactic activity was assessed using a Neuro Probe A96 chemotaxis chamber (Neuro Probe, Gaithersburg, Maryland, USA). Isolated neutrophils at 5 × 10⁶ cells/ml in HBSS + 0.1% bovine serum albumin (BSA) were stained with the fluorescent dye BCECF (1.6 μg/ml) for 30 min at 37°C, subsequently centrifuged for 5 min (340 g), and resuspended in HBSS + 0.1% BSA. A total of 33 μl chemoattractant PAF (final concentration 1 μM) was added into the wells of the lower plate. Thereafter, a silicone gasket and a framed filter with 3 μm pores were placed upon the lower wells. On top of it, the upper plate was attached, and the dyed neutrophils were pipetted into the corresponding wells. During incubation for 30 min at 37°C, neutrophils migrated from the upper wells toward the lower wells containing PAF, but became adherent to the filter, resulting in increased fluorescence. The fluorescence of the cells in the filter was measured at a wavelength of 485/538 nm using a Fluoroskan Ascent (Thermo Scientific, Rockford, Illinois, USA) with the Ascent Software Version 2.6.

Platelet–neutrophil complex (PNC) formation, defined as a neutrophil with at least one platelet in direct proximity, was assessed as described previously by light microscopy and flow cytometry (36, 42). For blood smears, whole blood was diluted 1:1 with phosphate-buffered saline containing calcium and magnesium (PBS) and incubated for 15 min at 37°C with or without 1 μM PAF on a spinning wheel (Snijders Labs, Tilburg, Netherlands) at 3 rpm. Blood smears were created and stained with Hemacolor® Rapid staining (Merck). In each sample, 50 neutrophils were counted by two independent and blinded individuals. For flow cytometry analysis, whole blood was diluted 1:5 with PBS. Following 15 min of incubation with or without 1 μM PAF, the sample was stained with anti-CD41 APC (125 ng/ml, #303710, Biolegend) and anti-CD61 PerCP (250 ng/ml, # 336412, Biolegend; both monoclonal mouse anti-human antibodies) for 15 min at room temperature followed by 30 min incubation with 1 ml of 1X BD FACS Lysing solution (BD Biosciences, San Jose, California, USA). Samples were centrifuged for 5 min at 340 g and resuspended in 100 μl PBS + 0.1% BSA and stored at 4°C in the dark until further analysis (normally within 1 h).

An animal-free human whole blood model of endotoxemia was used to investigate the effect of LPS exposure on the PAF-induced response of neutrophils, which was described previously in detail (36). In brief, 0.5 IU/ml heparin (B. Braun Melsungen AG, Melsungen, Germany) and 100 ng/ml LPS (#L2630, Merck) or PBS (control) were added to 9 ml blood, which was transferred into a Cortiva BioActive Surface (Medtronic, Meerbusch, Germany) coated tubing system using a coated connector (Medtronic) to interlink both ends. Following 1 h of rotation (3 rpm) on a spinning wheel (Snijders Labs) at 37°C, the blood was transferred to citrate anti-coagulated monovettes (Sarstedt). Neutrophil activation markers were analyzed directly in whole blood as described previously (36) with or without exposure to PAF (1 μM, 15 min, 37°C). The remaining whole blood was processed to isolate and to analyze neutrophils as described above.

The Federal authorities for animal research (#1362, Tuebingen, Germany) as well as the Animal Care Committee of the University of Ulm approved the experiments. The well-described porcine sepsis model was performed in adherence with guidelines on the Use of Laboratory Animals of the National Institutes of Health with anesthesia and surgical instrumentation as described previously (43–45). In brief, nine Bretoncelles-Meishan-Willebrand pigs (5 male-castrated, 4 female, mean weight 65.6 kg ± 8.6) were subjected to polymicrobial sepsis induced by inoculation of autologous feces into the abdominal cavity, followed by intensive care therapy for 60 h after the sepsis initiation. Blood was drawn before inoculation and when the mean arterial pressure was reduced by >10%, indicating cardiocirculatory shock and triggering the beginning of resuscitation. To reduce animal numbers, the analyzed pigs are a subgroup of another currently unpublished trial comparing standard intensive care therapy with or without a pharmacological intervention targeting the calcitonin gene-related peptide receptor. Because intervention started after the fulfillment of the sepsis criteria listed above, animals were included irrespective of their group allocation.

Results are presented as mean ± standard deviation (SD), when not indicated otherwise. In all experiments, a minimum of 3,000 neutrophils were measured. All data were considered to be paired and non-parametric. Statistical significance is indicated by ^*, ^**, and ^***, with significance levels of p < 0.05, < 0.01, and < 0.001, respectively. Statistical analysis was performed using GraphPad Prism 9 (GraphPad Software Inc., San Diego, California, USA) and Microsoft Excel (Version 16.42, Microsoft Corporation, Redmond, Washington, USA).

Neutrophils responded to PAF stimulation with a rapid increase in cell size, peaking after 20 min with a relative increase of 73 ± 19% (Figures 1A,B). A near-real time measurement demonstrated that most of the size change occurred within in the first minutes (Figure 1A). The increase in the FSC-A induced by PAF was similar to the stimulation with C5a (Figure 1B). Using reference microspheres, the calculated diameters of unstimulated cells were 16.9 ± 1.1 μm, and 26.1 ± 0.6 μm for PAF-stimulated cells (after 10 min, p = 0.06, n = 5, representative measurement in Figure 1C). The change in neutrophil size was verified by Coulter counter measurement. Following 10 min of incubation with PAF, neutrophils exhibited an increase of 3.6 ± 1.7% in diameter and 10.5 ± 6.4% in volume in comparison with unstimulated neutrophils (Figure 1D).

Next, the response elicited by PAF on the neutrophil MP and pH_i was investigated. PAF rapidly induced a depolarization peaking after 1 min (Figure 2A) and an intracellular alkalization attaining a maximum of +0.45 ± 0.07 after 5 min (Figure 2B). In addition, PAF induced an increase in ROS production (+26 ± 11% after 10 min, p < 0.001 Wilcoxon signed-rank test, n = 17, data not shown). Depolarization and intracellular alkalization revealed different kinetics (Figures 2A–C). The PAF-induced response was comparable to C5a-stimulation for cell size and pH_i but was more intensive for the MP (1 min: PAF +12 ± 4 mV, C5a: 7 ± 3 mV, Figures 2A,B). All PAF-induced effects showed a clear concentration-response relationship (Supplement 1). The EC₅₀-values are presented in Figure 2D and ranged between 14 and 128 nM.

In accordance with previous findings, neutrophils responded to PAF stimulation with an increased chemotactic activity (15.8 ± 9.7 fold increase after 30 min, p < 0.05 Wilcoxon signed-rank test, n = 6), CD11b upregulation (+63.3 ± 41.9% after 10 min, p < 0.01 Wilcoxon signed-rank test, n = 8), and CD62L shedding (−87 ± 10% surface expression after 10 min, p < 0.01 Wilcoxon signed-rank test, n = 8) (Supplements 1, 2). In addition, PAF induced formation of PNCs (Figure 3). For all these effects, there was no significant difference when comparing neutrophils from male and female donors (Supplement 2).

Ion fluxes in stimulated neutrophils are crucially involved in cellular activity. Unspecific inhibition of sodium or chloride flux inhibited most cellular changes induced by PAF (Figures 4A,C), however, in parallel also partially altering unstimulated neutrophils (Supplement 3).

A more detailed analysis revealed that PAF-induced intracellular alkalization was completely inhibited by targeting NHE1 activity with BIX (+0.30 ± 0.07 vs. −0.02 ± 0.04, p < 0.01). Cellular swelling and the ROS generation were also reduced, but to a lesser extent, by targeting NHE1 (Figure 4B). PAF-induced depolarization and CD62L shedding displayed no difference during selective NHE1 inhibition. PAF-induced CD11b expression increased during selective NHE1 inhibition, whereas it was fully inhibited by amiloride. However, during incubation with the NHE1 inhibitor BIX, neutrophils expressed less CD11b than cells without an inhibitor (Supplement 3).

Mimicking glutathione peroxidase activity and inhibiting NOX2 with ebselen significantly lowered PAF-induced generation of intracellular alkalization, ROS generation and CD62L shedding (Figure 4D).

The first step to investigate the above-mentioned PAF-elicited effects under pathologic conditions was to expose neutrophils to buffers of various extracellular pH (pH_e). The pH_e had no influence on the PAF-induced increase in the FSC-A (Figure 5A). Depolarization after PAF stimulation was significantly enhanced under acid extracellular conditions (pH_e 6.6 = +23.6 ± 7.5 mV vs. pH_e 7.4 = +10.7 ± 3.1 mV, p < 0.01 Kruskal-Wallis test and Dunn's multiple comparison test, n = 6, Figure 5B). Regarding the pH_i, neutrophils responded with an equal intracellular alkalization when stimulated with PAF under different extracellular conditions. In addition, the pH_i of neutrophils shifted with the pH_e, therefore with rising pH_e, the pH_i increased in parallel (Figure 5C).

To translate the findings to a more clinically relevant setting, an ex vivo model of human endotoxemia was used analyzing the modulation of the PAF-induced response in systemic inflammation. As previously reported (36), sole contact with the tubing system of the model did not promote neutrophil activation (data not shown). However, the exposure to LPS (100 ng/ml) resulted in an activated neutrophil phenotype as reflected by of the upregulation of the surface expression of CD11b and downregulation of CD62L. These alterations in baseline expression of these activation markers also modulated the response generated by additional in vitro stimulation with PAF (increase in CD62L downregulation, reduction in CD11b upregulation, Figures 6A,B).

LPS exposure resulted in increased cellular size, pH_i, and generation of ROS, but not depolarization in neutrophils without further PAF stimulation. The PAF-induced increase in cellular size was significantly reduced after exposure to LPS (CTRL-PAF +64 ± 10% vs. LPS-PAF +8 ± 3%, p < 0.05, Figure 6C). There was no significant difference in PAF-mediated depolarization after exposure to LPS (Figure 6D). Transient intracellular alkalization elicited by PAF was significantly reduced in neutrophils after endotoxemia (CTRL-PAF +0.33 ± 0.05 vs. LPS-PAF +0.02 ± 0.05, p < 0.05, Figure 6E). The PAF-induced elevation in ROS generation was reduced after LPS exposure (CTRL-PAF +23 ± 12% vs. LPS-PAF −9 ± 15%, Figure 6F).

To validate the in vitro and ex vivo results, neutrophils isolated from a porcine polymicrobial sepsis model were investigated before and during sepsis. Clinical parameters and electrolytes are summarized in Table 1. The baseline FSC-A and pH_i of neutrophils without PAF stimulation remained stable before and during sepsis (data not shown). The PAF-induced change in cellular shape displayed a trend for an increase during sepsis (+0 ± 7% before vs. +12 ± 10% during sepsis, p = 0.08, Figure 7A). PAF-mediated depolarization was significantly diminished during sepsis (+17.1 ± 5.6 mV before vs. +2.6 ± 2.2 mV during sepsis, p < 0.05, Figure 7B). There was no difference between neutrophils before and during sepsis regarding the PAF-induced intracellular alkalization (Figure 7C).

The inflammatory mediator PAF induced a rapid sequence of cellular reactions in neutrophils in a concentration- and time-dependent manner as summarized in Figure 8. The initial depolarization within seconds can likely be explained by the activation of the NOX activity resulting in excessive electron extrusion (17). This is corroborated by increased ROS generation and the lack of cellular depolarization in patients with a NOX defect (46). Depolarization was followed by intracellular alkalization within minutes. The time course of the PAF-induced alkalization was in accordance with reports from bovine neutrophils (33). The change in pH_i elicited by PAF was less in the study by Hidalgo et al., however, this might result from the use of different dyes and/or species differences [(33) and the results presented in this work]. In this context, it is noteworthy that shifts in neutrophil pH_i are associated with other cellular functions, for example increasing chemotactic motility (21, 23, 47). Therefore, the impact of PAF-mediated transient alkalization requires further elucidation.

In parallel, neutrophils responded to PAF stimulation with an increase in cellular size as indicated by FSC-A. The rise in cell size was confirmed by analyzing the cell volume by applying the coulter counter principle, although with less pronounced changes. In this context, previous studies with C5a indicated that an increased FSC-A reflects a cellular elongation (14, 48) likely caused by actin polymerization (48). This possibly explains the differences between the PAF-induced changes in cell size reported by different methods. Overall, the PAF-mediated response was similar to that elicited by other chemoattractants, including the complement-derived anaphylatoxin C5a (14–16) and fMLF (18, 35, 46).

The cellular response of neutrophils induced by PAF was also investigated in the absence or presence of pharmacological inhibitors with a special interest in depolarization, intracellular pH, and swelling. In accordance with the results obtained on neutrophil stimulation with other chemokines (14, 15, 33), non-specific inhibition of sodium flux with amiloride and/or chloride flux with NPPB significantly reduced the intracellular alkalization and cellular swelling. Albeit amiloride also reduced depolarization, it should be noted that it also pre-depolarized unstimulated cells. Next, we investigated the role of NHE1 as an important representative ion transporter protein in stimulated granulocytes (15, 38, 49). The pharmacological NHE1 inhibitor BIX (38, 50) also inhibited intracellular alkalization and to a lesser extent, cellular swelling, but not early depolarization. This finding of the PAF-neutrophil interaction is in accordance with previous reports for fMLF and C5a (15, 23), as well as the involvement of NHE1 activity during cellular polarization and migration (49). Lastly, ebselen was used as an inhibitor of NOX2 and peroxide scavenger (40, 41, 51), which also dampened cellular swelling and alkalization but not depolarization. It is noteworthy, that exposure to either BIX or ebselen did not largely alter most cellular functions in unstimulated cells, and, therefore, might be potential targets for clinical evaluations regarding the modulation of the PAF-induced thromboinflammatory response of neutrophils.

In addition to inducing an (electro-)physiological response in neutrophils, PAF elicited chemotactic activity, CD11b upregulation and CD62L downregulation, and PNC formation which is in accordance with previous studies (24, 25, 27, 52). This “platelet satelitism” of neutrophils is also found in several pathologies, including sepsis, augmenting neutrophil activity, including extravasation, neutrophil extracellular trap formation, and bacterial killing (53–56). It is tempting to speculate that inhibition of PNC formation, possibly caused by PAF release during sepsis (57, 58), might be a suitable target to modulate the thromboinflammatory response. For example, the striking relevance of PNC formation has been demonstrated in a murine model of hydrochloric acid-induced acute lung injury, in which blocking PNC formation significantly ameliorated organ dysfunction (59).

During inflammation, neutrophils are required to migrate through various tissues and are exposed to varying levels of extracellular of extracellular proton concentrations. The pH of inflammatory body fluids, for example, abscesses, where neutrophils amass, is acidic (60–64). The experiment involving varying pH_e indicated that even under acidic pH_e, neutrophils remain able to respond with a relative intracellular alkalization and cellular swelling despite a shift in their pH_i in relationship to the pH_e. In addition, PAF-induced depolarization was increased in acidosis, potentially indicating elevated ROS generation, which has been demonstrated in a previous study (65).

Next, the effects of PAF on neutrophils were translationally investigated in two inflammatory environments. The previously described human whole blood model (36) of LPS-driven inflammation allows the investigation of blood physiology and immunity in an animal-free environment in accordance with the 3R principles (66). LPS exposure of whole blood altered the neutrophil phenotype as shown by CD11b and CD62L levels, confirming a sepsis-like phenotype that has been described in detail previously (67–71). LPS exposure did not alter the MP in comparison with unstimulated neutrophils and only augmented the PAF-induced response by a small amount. By contrast, a similar in vitro experiment comparing the depolarization of neutrophils with or without LPS exposure did report an increase in MP change, however, for fMLF (72). As observed in previous studies with neutrophils from septic humans and/or mice, exposure to LPS prompted an intracellular alkalization and swelling (14, 15, 73). This baseline drift in unstimulated cells was accompanied by a decrease in the PAF-induced response of granulocytes, possibly indicating ceiling effects. In general, the intracellular pH of neutrophils can rise beyond the LPS-induced shift in resting cells and/or the measurement method is capable of detecting larger shifts, arguing against a ceiling effect. Nevertheless, it is possible that neutrophils can only generate a certain alkaline shift, being limited by the pH_e that was not relevantly altered after LPS exposure in the whole blood model as reported earlier (36). LPS exposure in the whole blood model did not significantly change the sodium, potassium, or proton concentration, thus failing to explain the changes in alkalization or depolarization as reported previously (36).

Lastly, we analyzed the PAF-induced (electro-)physiological response in neutrophils before and during a well-characterized porcine model of polymicrobial sepsis (43–45). The clinical data confirmed a cardiocirculatory shock, indicating sepsis, also in corroboration with a significant hemoconcentration, suggesting a blood-organ barrier breakdown and subsequent hypovolemia. In the course of sepsis, the PAF-induced depolarization was significantly impaired, which confirmed findings in C5a-stimulated neutrophils after 3 h of porcine hemorrhagic shock (16). It is noteworthy, that there were no large shifts in extracellular ion concentrations and/or blood pH that could potentially explain this finding. Moreover, the PAF-mediated alkalization remained largely stable during sepsis, whereas cellular swelling was slightly increased. Also, the baseline of neutrophil pH_i and FSC-A did not shift (data not shown), which is in contrast to previous results in murine and/or human sepsis (14, 15, 73). These apparent differences in the findings within the two models might be explained by either species differences (human vs. pig), differences in the time points analyzed (1 h vs. 5.8 h ± 3.5), and/or regarding the pH_i with different fluorescent probes.