Sex-based differences in human neutrophil chemorepulsion

Abstract

A considerable amount is known about how eukaryotic cells move towards an attractant, and the mechanisms are conserved from Dictyostelium discoideum to human neutrophils. Relatively little is known about chemorepulsion, where cells move away from a repellent signal. We previously identified pathways mediating chemorepulsion in Dictyostelium, and here we show that these pathways, including Ras, Rac, PKC, PTEN, and ERK1 and 2 are required for human neutrophil chemorepulsion, and as with Dictyostelium chemorepulsion, PI3K and PLC are not necessary, suggesting that eukaryotic chemorepulsion mechanisms are conserved. Surprisingly, there were differences between male and female neutrophils. Inhibition of Rho-associated kinases (ROCKs) or Cdc42 caused male neutrophils to be more repelled by a chemorepellent and female neutrophils to be attracted to the chemorepellent. In the presence of a chemorepellent, compared to male neutrophils, female neutrophils showed a reduced percentage of repelled neutrophils, greater persistence of movement, more adhesion, less accumulation of PI(3,4,5)P₃, and less polymerization of actin. Five proteins associated with chemorepulsion pathways are differentially abundant with three of the five showing sex dimorphism in protein localization in unstimulated male and female neutrophils. Together, this indicates a fundamental difference in a motility mechanism in the innate immune system in men and women.

Introduction

In a process called chemoattraction, many eukaryotic cells will move toward the source of an attractant, and the mechanism is highly conserved from the social amoebae Dictyostelium discoideum to human neutrophils (1, 2). Relatively little is known about chemorepulsion, where cells move away from a repellant. Dictyostelium cells secrete a chemorepellent protein called AprA that causes cells at the edge of a colony of cells to move away from the colony, possibly to find new sources of food (3, 4). Neutrophils can also move away from a repellent, but it is unclear if the chemorepulsion mechanism is conserved.

In Dictyostelium, AprA-induced chemorepulsion is mediated by the G-protein coupled receptor GrlH, G proteins Gα8 and Gβγ, Rho GTPases RasC, RasG, and RacC, components of the target of rapamycin complex 2 (TORC2), protein kinase C (PKC), and extracellular signal-regulated kinases (ERKs) of the MAPK pathway (4–7). However, unlike cAMP-induced chemoattraction in Dictyostelium (8), AprA-induced chemorepulsion does not require PI3 kinases 1 and 2 and phospholipase C (PLC) (4, 7).

The predicted structure of AprA is similar to the structure of the human extracellular protease dipeptidyl peptidase IV (DPPIV) (9), and AprA has DPPIV-like protease activity (10). At physiological concentrations, DPPIV repels human and murine neutrophils (9). DPPIV induces neutrophil chemorepulsion through the protease activated receptor 2 (PAR2) (11). Like DPPIV, PAR2 agonists AC55541, 2f-LIGRL-NH₂, and SLIGKV-NH₂ induce neutrophil chemorepulsion (12). Male mouse neutrophils were more repelled than female neutrophils in DPPIV or AC55541 gradients (12). However, the female neutrophils that were repelled had a more direct pathlength than male neutrophils (12). Neutrophils from Par2^−/− mice were not repelled by DPPIV or the PAR2 agonists SLIGKV-NH₂ and AC55541, but showed sex differences in speed and directness (12). Neutrophil repellents were not sufficient to repel neutrophils when competing against an fMLF gradient (12, 13).

There were no observed sex differences during chemorepulsion from the PAR2 activators trypsin and tryptase (12). However, DPPIV and tryptase themselves have sex-based differences. There is increased DPPIV mRNA expression and increased percentages of CD4+DPPIV+ and CD8+DPPIV+ T cells in women compared to men, which may contribute to sex differences observed in chronic HIV infection (14). One study observed increased serum DPPIV activity in men compared to women despite no difference in levels of serum DPPIV (15), although other studies found no sex-based differences in serum DPPIV activity in healthy patients (16, 17). Men have higher mean basal serum tryptase (18), and in patients with renal failure, men also have higher mast cell tryptase (19). We could find no evidence of sex-based differences for trypsin in humans.

In this report, we examined the effect of pharmacological inhibitors of potential pathway components on the ability of SLIGKV-NH₂ to induce human neutrophil chemorepulsion. We found similarities in the Dictyostelium and neutrophil chemorepulsion pathways, but with differences between male and female neutrophils.

Materials and Methods

Neutrophil isolation and culture

Human venous blood was collected with the approval from the Texas A&M University Institutional Review Board from healthy volunteers who gave written consent. Neutrophils were isolated as previously described (13) with the exception that the polymorphprep (Accurate Chemical, Westbury, MA) gradient centrifugation was performed for 45 minutes. Cells were resuspended in RPMI-1640 (Lonza, Walkersville, MD) with 2% BSA (Rockland Inc, Limerick, PA) (RPMI-BSA). Cell spots and staining with eosin and methylene blue were done following (13). Isolated neutrophil preparations were 95.5% ± 0.4 neutrophils (mean ± SEM, n = 8). We never used the same donor twice for a given experiment. The age ranges for the donors were 18–56 years for males and 18–35 years for females.

Insall chamber assays

Ras inhibitory peptide (RIP) (Cayman Chemical, Ann Arbor, MI), Rac inhibitor NSC23766 (Cayman Chemical, Ann Arbor, MI), Cdc42 inhibitor ML 141 (Cayman Chemical, Ann Arbor, MI), Rho-associated kinase inhibitor (Rho-assoc Kin inhibitor or Rho Kinase inhibitor) (Cayman Chemical, Ann Arbor, MI), Protein Kinase C (PKC) inhibitor Gӧ6976 (Cayman Chemical, Ann Arbor, MI), and phosphatidylinositol 3 kinase (PI3K) inhibitor LY294002 (BioVision, Milpitas, CA) were reconstituted to 10 mM in DMSO (VWR Life Science, Radnor, PA). The DMSO stocks were then diluted in RPMI-BSA to 10 μM with 2.5 × 10⁶ neutrophils in 0.5 ml for 30 minutes in a humidified 5% CO₂ incubator. Phosphatase and tensin homolog (PTEN) inhibitor SF1670 (Cayman Chemical, Ann Arbor, MI) was reconstituted to 5 mM in DMSO and then diluted in RPMI-BSA to 500 nM for 30 minutes as described above. Phospholipase C (PLC) inhibitor U73122 (Alfa Aesar, Haverhill, MA) was reconstituted to 5 mM in DMSO, diluted to 1 μM in RPMI-BSA, and used as above for 1 hour. ERK1 and ERK2 inhibitor PD98059 (Cayman Chemical, Ann Arbor, MI) was reconstituted to 30 mM in DMSO, diluted to 30 μM in RPMI-BSA, and used as described above for 1 hour. Cells were then placed on acid-etched glass coverslips coated with 20 μg/ml human fibronectin (Corning, Bedford, MA), as previously described (9, 12, 13), and allowed to adhere for 20 minutes. Insall chambers were used to generate gradients as previously described (9, 12, 13). The neutrophil chemoattractant N-formylmethionyl-leucyl-phenylalanine (fMLF; Cayman Chemical, Ann Arbor, MI) was diluted in RPMI-BSA and used at 1 nM and the chemorepellent SLIGKV-NH₂ (Tocris, Avonmouth, Bristol, UK or Sigma, St. Louis, MO; henceforth designated SLIGKV) was diluted in RPMI-BSA and used at 50 ng/ml. Neutrophils were tracked by videomicroscopy as previously described (9, 12, 13). All cells in the starting frame (≥ 10 neutrophils per experiment) were tracked for 40 minutes. For each donor, neutrophils were tracked in a no-gradient/no stimulus control. Neutrophils were never used past 5 hours from the end of the isolation step. We used two Insall chambers/microscope/camera set-ups in parallel to allow 6 to 8 experimental conditions to be measured per set of donor neutrophils.

Adhesion assays

Neutrophil adhesion assays were performed as described (12, 13). 2.0 × 10⁶ cells/ml neutrophils were pre-incubated with 10 μM Ras inhibitor, 10 μM Rac inhibitor, 10 μM Cdc42 inhibitor, 10 μM Rho-associated kinase inhibitor, 10 μM PKC inhibitor, 500 nM PTEN inhibitor, 10 μM PI3K inhibitor, 1 μM PLC inhibitor, or 30 μM ERK 1 and ERK2 inhibitor for either 1 hour (PLC inhibitor and ERK 1 and 2 inhibitor) or 30 minutes (all others) at 37°C. 1.0 × 10⁵ neutrophils in the presence or absence of inhibitors were added to the wells such that the final concentration of stimuli were 10 nM fMLF, 500 ng/ml SLIGKV, 10 ng/ml TNFα, or an equivalent volume of RPMI-BSA or RPMI-BSA with 1:1,000 dilution of DMSO (No Stimulus/DMSO). Neutrophils were allowed to adhere for 45 minutes at 37°C. The wells were then washed 3 times with PBS, air dried, and stained with methylene blue. For each individual donor, the conditions of each well were run in duplicate. Adherent neutrophils were counted in four different 900 μm diameter fields of view as described (20).

PI(4,5)P₂ and PI(3,4,5)P₃ assays

Phosphatidylinositol 4,5-bisphosphate (PI(4,5)P₂) and phosphatidylinositol 3,4,5-trisphosphate (PI(3,4,5)P₃) assays were performed as described (13) with the following modifications. 1 × 10⁷ neutrophils were incubated at 37°C in 0.5 ml of RPMI-BSA in the presence or absence of 10 nM fMLF, 500 ng/ml SLIGKV, or an equivalent volume of RPMI-BSA (No Stimulus) for 5 or 10 minutes. 0.5 ml of 1 M trichloroacetic acid was used to stop the reaction before incubating the samples on ice for 5 minutes. PI(4,5)P₂ and PI(3,4,5)P₃ were assayed using PI(4,5)P₂ K-4500 and PI(3,4,5)P₃ K-2500 ELISA assay kits (Echelon Biosciences Inc., Salt Lake City, UT) following the manufacturer’s directions.

Isolation of cytoskeletal proteins and associated western blots

Preparation of cytoskeletons, gel electrophoresis, and western blot analysis to visualize F-actin and polymerized myosin IIA was done as previously described (13) with the exception that the cytoskeleton pellet was resuspended in 1× SDS sample buffer with β-mercaptoethanol and boiled at 98°C for 15 minutes. Samples were then electrophoresed on 4–20% polyacrylamide gel (Bio-Rad Laboratories, Hercules, CA) electrophoresis. Voltage was set at 45V for 20 minutes and then 95V for 75 minutes. Polyacrylamide gels were transferred onto PVDF membrane (GE Healthcare Life Science, Marlborough, MA, USA).

Western blots were blocked with TBS-0.1% Tween 20 with 5% non-fat milk for 1 hour at room temperature, then incubated overnight at 4°C with a 1:1,000 dilution of anti-β-actin (13E5) antibody or a 1:1,000 dilution of anti-Myosin IIA Rabbit antibody (both from Cell Signaling Technology, Danvers, MA) following the manufacturer’s directions to detect F-actin and polymerized myosin IIA in the cytoskeleton. Bound antibody was detected with ECL western blotting kits (Thermo Scientific, Rockford, IL). Western blot band intensities were quantified using Image Lab software (Bio-Rad Laboratories, Hercules, CA). In both F-actin and polymerized myosin IIA, each timepoint for each individual experiment was normalized to the F-actin or myosin IIA band intensity of time = 0 seconds value for that stimulus. As an example, the band intensity for F-actin with 90 seconds of fMLF stimulus was normalized to the band intensity for F-actin with 0 seconds of fMLF stimulus.

Live and fixed cell microscopy

For live neutrophil imaging, 2.5 × 10⁵ neutrophils in 0.2 ml of RPMI-BSA were allowed to adhere to fibronectin coated chambers in an 8-well cover glass slide (#94.6190.802, Starstedt Inc., Nümbrecht, Germany) for 30 minutes at 37°C on a well-by-well basis to prevent the cells from cooling before imaging. 1 μl of prewarmed 1 μM fMLF in RPMI-BSA, 2.5 μl of 20 μg/ml SLIGKV in RPMI-BSA, or 2.5 μl of RPMI-BSA (No Stimulus) was gently added to the corner of the well to prevent disturbing the cells and media. After 5 minutes, images were taken from the same region of the well for each sample.

For fixed neutrophil images of cells in a gradient over time, 5 × 10⁵ cells were allowed to adhere to fibronectin coated 8 well chamber cover-glass bottom slides and material was added to the corner of the well as above. After 10 seconds or 5, 10, and 20 minutes, 0.15 ml of media was removed from the corner of the well and the cells were fixed by adding 0.2 ml of 2.5% glutaraldehyde in PBS for 10 minutes. Wells were gently washed twice with PBS before cells were permeabilized with 2% Triton X-100 in PBS for 5 minutes. Wells were washed twice again with PBS before blocking with 2% BSA/ 300 μM glycine in PBS (pH 7.4) for 1 hour at room temperature. Wells were washed twice again with PBS before incubating the cells overnight at 4°C in a humid chamber with 1:250 anti-P-Myosin Light Chain 2 (T1B/S19) Rabbit antibody (#3674s, Cell Signaling Technology, Danvers, MA) in PBS/ 0.1% Tween 20. Cells were washed twice with PBS before incubating with 1:1,000 Alexa Fluor 488-conjugated AffiniPure F(ab’)2 Fragment Donkey anti-Rabbit IgG (H+L) (#711–546-152, Jackson ImmunoResearch, West Grove, PA) and 1:2,000 Phalloidin-iFluor 555 Reagent (ab176756, Abcam, Cambridge, United Kingdom) for 1 hour at room temperature protected from light. Wells were washed twice more with PBS, and then cells were incubated with 1:1,000 of DAPI (#422801, Biolegend, San Diego, CA) in PBS/ 0.1% Tween 20 for 10 minutes at room temperature. After one wash with PBS, 0.2 ml of PBS was added to the wells. Immunofluorescence images were captured with a 40× objective using a Ti2-Eclipse (Nikon, Tokyo, Japan) inverted fluorescence microscope. Images were deconvoluted to visualize protein localization and abundance. Cells were scored by a blinded observer as having phalloidin or phospho-myosin light staining in one primary region located either towards or away from the corner of the well where the stimulus was added, multiple staining (having 2 or more significant regions of staining), or diffuse staining (where staining was weak, dispersed, or circumferential). Images were analyzed using Fiji ImageJ for average circularity and average roundness of all neutrophils in a field of view (>8 cells per field of view with an average of 4 of more fields of view per stimulus per donor) following (21) method 3.2.2. “Performing image quantitation without an external drawing tablet” where thresholding was used to include the entire area of each neutrophil while excluding background and red blood cells.

For images of fixed unstimulated neutrophils, 5 × 10⁵ cells were allowed to adhere to fibronectin coated 96 well chamber cover-glass bottom slides (#89626, Ibidi, Fitchburg, WI). 0.15 ml of media was removed from the corner of the well and the cells were fixed by adding 0.2 ml of 2.5% glutaraldehyde in PBS for 10 minutes. Wells were gently washed twice with PBS before cells were permeabilized with 2% Triton X-100 in PBS for 5 minutes. Wells were gently washed twice again with PBS before blocking with 2% BSA/ 300 μM glycine in PBS (pH 7.4) for 1 hour at room temperature. Wells were gently washed twice again with PBS before incubating the cells overnight at 4°C in a humid chamber with 1:250 anti-RhoA MAb IgM clone:54D6.1.16 (ARH05, Cytoskeleton Inc, Denver, CO), 1:250 anti-RhoB Rabbit (NBP2–94104, Novus, Littleton, CO), 1:250 anti-Cdc42 MAb clone:4B3 (ACD03, Cytoskeleton Inc, Denver, CO), 1:500 anti-PIP4K2C Polyclonal (PA5–66180, Thermo Fisher Scientific, Waltham, MA), or 1:250 anti-CD58/LFA-3 (TS2/9) Mouse MAb (NBP2–22542, Novus, Littleton, CO) in PBS/ 0.1% Tween 20. Cells were washed twice with PBS before incubating with either 1:1,000 Alexa Fluor 488-conjugated AffiniPure F(ab’)2 Fragment Donkey anti-Rabbit IgG (H+L) (#711–546-152, Jackson ImmunoResearch, West Grove, PA) or 1:1,000 Alexa Fluor 488-conjugated AffiniPure F(ab’)2 Fragment Donkey anti-Mouse IgG (H+L) (#715–546-150, Jackson ImmunoResearch, West Grove, PA) and 1:2,000 Phalloidin-iFluor 555 Reagent (ab176756, Abcam, Cambridge, United Kingdom) for 1 hour at room temperature protected from light. For the RhoA antibody, 1:500 dilution of Biotin-SP-conjugated AffiniPure F(ab’)2 Fragment Donkey Anti-Mouse IgM (H+L) (#715–065-140, Jackson ImmunoResearch, West Grove, PA) was used. After 30 minutes, cells were washed twice with PBS and then treated with 1:1,000 dilution of Streptavidin, Alexa Fluor 488, conjugate (#43773A, Invitrogen, Waltham, MA) and 1:2,000 Phalloidin-iFluor 555 Reagent (ab176756, Abcam, Cambridge, United Kingdom) for 30 minutes at room temperature protected from light. After secondary antibody treatments, wells were washed twice more with PBS, and then cells were incubated with 1:1,000 of DAPI (Biolegend) in PBS/ 0.1% Tween 20 for 10 minutes at room temperature. After one wash with PBS, 0.2 ml of PBS was added to the wells. Immunofluorescence images were captured with 40× dry or 100× oil immersion objectives using a Ti2-Eclipse inverted fluorescence microscope. Mean fluorescent intensity of all neutrophils in a field of view (>10 cells per field of view with an average of 5 of more fields of view per antibody per donor) was quantified following (21), method 3.2.2. “Performing image quantitation without an external drawing tablet”, utilizing Fiji ImageJ, where thresholding was used to include the entire area of each neutrophil while excluding background fluorescence. Staining morphology was scored by a blinded observer.

Proteomics

2 × 10⁷ neutrophils in 1 ml of RPMI-BSA were collected by centrifugation at 500 × g for 3 minutes at 4°C. The supernatant was discarded and cells were washed twice with 1 ml ice cold PBS, at 500 × g for 3 minutes at 4°C between each wash. The cell pellet was resuspended in 0.3 ml of ice cold Ripa Buffer (#89900, ThermoScientific, Rockford, IL) with 10× Protease and Phosphatase Inhibitor Cocktail (#1861281, ThermoScientific, Rockford, IL). A pipette was used to vigorously lyse the cells, and the lysate was then snap frozen in liquid nitrogen and stored at −80°C.

In-gel protein preparation was performed as described on the University of Texas – Southwestern Proteomics Core webpage (https://proteomics.swmed.edu/wordpress/?page_id=553). In short, 75 μg of protein per test sample was combined with 1× SDS sample buffer with β-mercaptoethanol and boiled at 95°C for 10 minutes. 50 μg of protein was loaded onto a 4–20% SDS-PAGE gel and run at 70V for ~20 minutes (until ~5–10mm through the gel). The gel was stained with freshly prepared Coomassie blue dye for 1.5 hours with gentle shaking, destained for 1.5 hours, and further destained overnight in sterile water. In a cleaned glass petri dish, the protein bands were cut into 1–2mm cubes, and transferred to the University of Texas Southwestern Proteomics Core (https://proteomics.swmed.edu/wordpress/) for Thermo Fusion Lumos standard gradient mass spectrometry. The proteins were analyzed using Proteome Discoverer 2.4 and searched using the human protein database from UniProt. Raw and processed proteomic data has been uploaded to MassIVE at the University of California – San Diego Center for Computational Mass Spectrometry (https://massive.ucsd.edu/ProteoSAFe/dataset.jsp?task=002e367a56ef471da06a302861229930). Accession number: MSV000088857. For each donor, all protein counts were summed and then divided by the total counts per donor. Male and female values were compared to determine sex-based differential protein abundance.

Whole cell lysates and associated western blots

2 × 10⁶ neutrophils in 0.1 ml of RPMI-BSA were collected and washed as above. The cell pellet was resuspended in 0.1 ml of 1× SDS sample buffer with β-mercaptoethanol and boiled at 98°C for 15 minutes. Western blots were then stained as above overnight at 4°C with 1:500 anti-RhoA MAb IgM clone:54D6.1.16 (ARH05, Cytoskeleton Inc, Denver, CO), 1:1,000 anti-RhoB Rabbit (NBP2–94104, Novus, Littleton, CO), 1:1,000 anti-Cdc42 MAb clone:4B3 (ACD03, Cytoskeleton Inc, Denver, CO), 1:2,000 anti-PIP4K2C Polyclonal (PA5–66180, Thermo Fisher Scientific, Waltham, MA), 1:500 anti-CD58/LFA-3 (TS2/9) Mouse MAb (NBP2–22542, Novus, Littleton, CO), or 1:5,000 anti-GAPDH Mouse McAb (60004–1-Ig, Protein Tech, Rosemont, IL) following the manufacturer’s directions. Bound antibody was detected with ECL western blotting kits (Thermo Scientific, Rockford, IL). On each experiment day, neutrophils from one male and one female were analyzed. Western blot band intensities were quantified using Image Lab software (Bio-Rad Laboratories, Hercules, CA) and normalized to each test sample’s GAPDH as a loading control before normalizing to the date matched male donor.

Statistics

Prism (GraphPad, San Diego, CA) and Excel (Microsoft, Redmond, WA) were used for data analysis. Data are shown as mean ± SEM except where otherwise stated. Statistical significance was defined as p ˃ 0.05.

Results

PAR2 agonist-induced neutrophil chemorepulsion uses some of the same signal transduction pathways necessary for Dictyostelium chemorepulsion

To determine if the mechanisms mediating chemorepulsion are potentially conserved between Dictyostelium and human neutrophils, we pre-incubated human neutrophils with a variety of drugs to inhibit or significantly affect the activity of specific pathway components. After pre-incubation, the neutrophils were placed in an Insall chamber (22) with gradients of 0–1 nM N-formylmethionyl-leucyl-phenylalanine (fMLF), a potent chemoattractant (23, 24), 0–50 ng/ml SLIGKV-NH₂ (SLIGKV), a neutrophil chemorepellent (12), or no gradient along the y axis. Cell migration was recorded and tracked. Movement of the cells towards or away from the chemo-stimulant source was measured as forward migration index (FMI), with FMI being defined as the distance the cell moved on the y axis divided by the cell’s migration path over the course of the experiment. A negative FMI indicated chemoattraction and a positive FMI indicated chemorepulsion. Cell tracks were also analyzed for migration speed and persistence of cell movement, where we defined persistence as the Euclidean distance (the straight-line distance) between the start point and end point of a cell’s movement divided by the distance the cell traveled. If the persistence score is closer to 1.0, this indicates that the cell traveled in a more direct route from the start point to the end point. Cells that move randomly have a persistence score closer to zero. No significant differences in chemoattraction were observed between male and female neutrophils (Figs. S1A, C, and E). Compared to male neutrophils, SLIGKV repelled a smaller percentage, and attracted a higher percentage, of female neutrophils (Fig. S1B). In addition, female neutrophils had a higher persistence than male neutrophils during chemorepulsion from SLIGKV (Fig. S1F). There were no significant sex-based differences in speed in SLIGKV gradients (Fig. S1D).

To determine if SLIGKV-induced chemorepulsion is mediated through Ras, human neutrophils were pre-incubated with 10 μM ras inhibitory peptide (RIP) and then tracked in gradients of either fMLF or SLIGKV. As previously observed for fMLF (25), RIP blocked the ability of fMLF to attract neutrophils (Figs. 1A and S1G). RIP also blocked SLIGKV repulsion of male and female neutrophils (Figs. 1A and S1H). For both male and female neutrophils, RIP decreased neutrophil migration speed (Figs. S1I and S1K) and persistence (Figs. S1M and S1O) in the chemoattractant gradient, but did not significantly affect speed (Figs. S1J and S1L) or persistence (Figs. S1N and S1P) in the chemorepellent gradient. 10 μM Rac inhibitor NSC23766 caused fMLF to repel male and female neutrophils and caused SLIGKV to attract male and female neutrophils (Figs. 1B, S1G, and S1H). However, NSC23766 did not significantly affect male or female neutrophil migration speed or persistence in either the chemoattractant or chemorepellent gradients (Figs. S1I–P). As previously observed (26), 10 μM mammalian target of rapamycin complex 2 (mTORC2) inhibitor Gӧ6976 abolished FMI, and reduced migration speed and persistence during fMLF-mediated chemotaxis (Figs. 1C, S1I, and S1M) and this effect was consistent for both male and female neutrophils (Figs. S1G, S1K, and S1O). For SLIGKV-mediated chemorepulsion of male and female neutrophils, Gӧ6976 inhibited FMI and reduced cell speed, but did not significantly affect persistence (Figs. 1C, S1H, S1J, S1L, and S1N). However, in a SLIGKV gradient, Gӧ6976 decreased male neutrophil persistence compared to female neutrophils (Fig. S1P). Pre-incubation of both male and female human neutrophils with 500 nM phosphatase and tensin homolog (PTEN) inhibitor SF1670 inhibited FMI, migration speed, and persistence in both the chemoattractant and chemorepellent gradients (Figs. 1D, S1G–P).

Figure 1. PAR2 agonist-induced neutrophil chemorepulsion uses some of the same signal transduction pathways necessary for chemorepulsion in Dictyostelium discoideum.

Human neutrophils were pre-incubated with inhibitors of A) Ras (Ras Inhibitory Peptide), B) Rac (NSC23766), C) Protein Kinase C (PKC) (Gӧ6976), D) Phosphatase and tensin homology (PTEN) (SF1670), E) PI3 kinase (LY294002), F) Phospholipase C (PLC) (U73122), G) ERK1 and 2 (PD98059), H) Cdc42 (ML141), or I) Rho-associated kinases (Y-27632) for either 30 minutes (A - E, H, and I) or 1 hour (F and G) before being placed in gradients of buffer, 0–1nM fMLF, or 0–50ng/ml SLIGKV. Cell positions were filmed and tracked to record individual cell migrations. A negative FMI indicates chemoattraction, and a positive FMI indicates chemorepulsion. For each individual donor, at least 10 cells per experiment group were tracked for 40 minutes. Values are mean ± SEM for neutrophils from at least 6 donors (3 males and 3 females). H and I) Values are separated into groups by sex. * indicates p ˂ 0.05, ** p ˂ 0.01, *** p ˂ 0.001 (Mann-Whitney t tests). Related to figure S1.

We also observed that 10 μM phosphoinositide 3-kinase (PI3K) inhibitor LY294002 inhibits male and female neutrophil chemoattraction to fMLF (Figs. 1E and S1G). LY294002 did not significantly affect FMI during SLIGKV-mediated chemorepulsion of male and female neutrophils (Figs. 1E and S1H). For both male and female neutrophils, inhibition of PI3K did not affect cell migration speed, however, persistence was decreased in the fMLF chemoattractant gradient (Figs. S1I–P). In agreement with Xu et al and Czepielewski et al (27, 28), 10 μM phospholipase C (PLC) inhibitor U73122 reduced the FMI, migration speed, and persistence for male and female neutrophils in the fMLF gradient (Figs. 1F, S1G, S1I, S1K, S1M, and S1O). For both male and female neutrophils in SLIGKV gradients, U73122 increased the FMI without significantly affecting speed or persistence (Figs. 1F, S1H, S1J, S1L, S1N, S1P).

In agreement with Zhang et al (29), we observed that 30 μM extracellular signal-regulated kinase (ERK) 1 and 2 inhibitor PD98059 inhibited male and female neutrophil chemotaxis towards fMLF (Figs. 1G and S1G). 30 μM PD98059 also inhibited chemorepulsion of male and female neutrophils from SLIGKV (Figs. 1G and S1H). Inhibition of ERK 1 and 2 had no significant effect on neutrophil migration speed or persistence in either gradient (Figs. S1I–P).

These data suggest that Ras, Rac1, PKC, PTEN, and ERK 1 and 2 are necessary for SLIGKV-mediated chemorepulsion in human neutrophils, PI3K is not necessary, and that PLC may inhibit SLIGKV-induced chemorepulsion.

In agreement with previous observations (30), 10 μM Cdc42 inhibitor ML141 did not significantly affect the FMI in an fMLF gradient (Fig. 1H). When ML141 treated neutrophils were exposed to an SLIGKV gradient, male neutrophils were repelled by SLIGKV, but female neutrophils were attracted to SLIGKV (Fig. 1H). This is somewhat similar to our previous observation that when Cdc42 was inhibited, both male and female neutrophils were attracted to the chemorepellent Slit2-S (13). For both male and female neutrophils, ML141 decreased migration speed and persistence in fMLF gradients but had no effect in SLIGKV gradients (Figs. S1I–P). These data indicate that Cdc42 is not required for SLIGKV-induced neutrophil chemorepulsion of male neutrophils, but is required for female neutrophil chemorepulsion. One possible explanation is that chemorepellents activate both attraction and repulsion pathways, and that inhibition of Cdc42 blocks the repulsion pathway, allowing the attraction pathway to dominate the cell movement.

We observed that the FMI in male and female neutrophils pre-treated with 10 μM Rho-associated kinase (ROCK) inhibitor Y-27632 in a fMLF gradient was not adversely affected by inhibitor treatment, but cell migration speed and persistence were reduced (Figs. 1I, S1I, S1K, S1M, and S1O). When Y-27632 pre-treated neutrophils were exposed to an SLIGKV gradient, male neutrophils were more repelled by the chemorepellent and female neutrophils were attracted to the chemorepellent (Fig. 1I). Y-27632 did not significantly affect cell speed or persistence for male and female neutrophils in an SLIGKV gradient (Figs. S1J, S1L, S1N, and S1P). These data indicate that Cdc42 and rho-associated kinases have different roles in male and female neutrophils in response to a PAR2-agonist chemorepellent that is independent of cell migration speed and persistence.

ERK 1 and 2 inhibitor shows a sex-specific effect on neutrophil adhesion in the presence of SLIGKV

To determine if the inhibitors affect cell-substrate adhesion to possibly cause the observed effects on FMI, migration speed, or persistence, we pre-incubated cells in the presence or absence of inhibitors and then added a uniform concentration of SLIGKV, fMLF, or the neutrophil adhesion stimulator Tumor Necrosis Factor α (TNFα) (31). Cells were allowed to attach to fibronectin for 45 minutes and the non-adherent cells were washed off. The adhered cells were then counted. For both male and female neutrophils, in the presence of no stimulus or with SLIGKV, the inhibitors did not significantly affect adhesion to fibronectin (Fig. 2), although inhibition of ERK 1 and 2 decreased male neutrophil adhesion compared to female neutrophil adhesion in the presence of SLIGKV (Fig. 2D). As previously observed (32, 33), inhibition of Rho-associated kinases in male and female neutrophils increased adhesion in the presence of fMLF or TNFα (Figs. S2).

Figure 2. ERK1 and 2 inhibitor shows a sex-specific effect on neutrophil adhesion in the presence of SLIGKV.

Neutrophils were pre-incubated with the indicated inhibitors or DMSO for either 30 minutes (Ras, Rac, Cdc42, Rho-associated kinase, PKC, PTEN, or PI3 kinase inhibitor) or 1 hour (PLC inhibitor and ERK 1 and 2 inhibitor) before being allowed to adhere to a fibronectin coated well containing SLIGKV or buffer (control) for 45 minutes. The plates were gently washed and the adherent cells were air-dried, stained, and counted. Values are mean ± SEM, n = 8 (4 males and 4 females). * indicates p ˂ 0.05 (Wilcoxon rank t test). Related to figure S2.

SLIGKV transiently increases PI(3,4,5)P₃ levels, primarily from polarized male neutrophils

The phospholipid PI(4,5)P₂ is phosphorylated to PI(3,4,5)P₃ at the leading edge of a cell during chemoattraction to promote the formation of branching actin filaments to extend a pseudopod (34). The 5- and 10-minute timepoints were chosen to coincide with when neutrophils are starting to respond to, and then migrate from, the chemorepellent. Neutrophils increase PI(3,4,5)P₃ levels and consequently decrease PI(4,5)P₂ levels within the first 20 seconds after fMLF exposure and then slowly trend towards baseline levels over the next few minutes (35). Pre-incubation of neutrophils for 5 minutes with the chemoattractants fMLF and Slit2-N have no significant effect on PI(4,5)P₂ levels but increase PI(3,4,5)P₃ levels (13, 36, 37). To determine if SLIGKV affects PI(4,5)P₂ and PI(3,4,5)P₃ levels, neutrophils were incubated with a uniform concentration of fMLF, SLIGKV, or control buffer for 5 or 10 minutes before quantification of the extracted phosphoinositides. fMLF and SLIGKV did not significantly affect PI(4,5)P₂ levels at 5 and 10 minutes (Figs. 3A and 3B). fMLF increased levels of PI(3,4,5)P₃ in male but not female neutrophils at 5 minutes, and increased levels of PI(3,4,5)P₃ in the combined male and female neutrophils at 10 minutes (Figs. 3C and D). SLIGKV increased PI(3,4,5)P₃ levels at 5 minutes in male neutrophils (Figs. 3C and D). These data indicate that SLIGKV causes a transient increase in PI(3,4,5)P₃ levels even though PI3K is not necessary for SLIGKV-induced chemorepulsion (Fig. 1E), and that male neutrophils have a stronger PI(3,4,5)P₃ response than female neutrophils at 5 minutes.

Figure 3. SLIGKV increases PI(3,4,5)P₃ accumulation in male neutrophils.

Neutrophils were incubated with RPMI-2% BSA buffer (No Stimulus), fMLF, or SLIGKV for 5 or 10 minutes and then PI(4,5)P₂ and PI(3,4,5)P₃ levels in cells were measured (A - D). Values are means ± SEM for neutrophils from 4 male and 4 female donors. * indicates p < 0.05, ** p ˂ 0.01 (Mann-Whitney t-test (A and C), or Mixed Effects Analysis with Dunnett’s multiple comparisons test (B and D).

Male and female neutrophils show sex-based differences in cytoskeletal protein accumulation after chemorepellent stimulus resulting in a morphological difference at 5 minutes

Cell migration requires cytoskeletal rearrangement to maintain polarity, accumulate F-actin at the leading edge, and induce contractile forces at the rear of a cell for effective migration (38, 39). As previously observed (13, 40–42), a uniform concentration of fMLF induced a rapid increase in F-actin accumulation for both male and female neutrophils, with female neutrophils showing an increase starting at 60 seconds compared to 90 seconds in male neutrophils (Fig. 4A). We previously observed that AprA-induced chemorepulsion in Dictyostelium (7) and Slit2-S chemorepulsion of human neutrophils (13) did not affect F-actin accumulation after repellent stimulus from 0 seconds to 90 minutes or 0 seconds to 5 minutes, respectively. We hypothesized that SLIGKV would similarly have no effect on F-actin and myosin IIA accumulation. However, exposure to a uniform concentration of SLIGKV caused an increase in F-actin accumulation in both male and female neutrophils, peaking at 300 seconds (Fig. 4B). This F-actin accumulation was significantly higher in male neutrophils compared to female neutrophils.

Figure 4. Male and female neutrophils show sex-based differences in cytoskeletal protein accumulation after chemorepellent stimulus resulting in a morphological difference at 5 minutes.

Neutrophils were incubated with fMLF or SLIGKV for the indicated times and then lysed to enrich cytoskeletal proteins. Insoluble cytoskeletal proteins were analyzed by western blot to quantify F-actin (A and B) or polymerized myosin IIA (C and D). All values are mean ± SEM from 8 male donors and 7 female donors. * indicates p ˂ 0.05, ** p ˂ 0.01, *** p ˂ 0.001 compared to the 0 second time point for each sex (Mann-Whitney t test), # indicates p ˂ 0.05, # # p ˂ 0.01 compared to the opposite sex at the indicated time point (Multiple t test, corrected with the Holm-Šidák method). (E) Neutrophils were incubated with RPMI-2% BSA buffer (No Stimulus), fMLF, or SLIGKV for 5 minutes and imaged without fixation using a 100× oil objective. Bars are 5 μm. * indicates the direction toward the source of buffer, fMLF, or SLIGKV. Images are representative of 4 independent experiments. (F and G) Neutrophils were incubated as in (E), fixed, stained, and quantified for average circularity and roundness. Values are mean ± SEM from 4 male and 4 female donors. ** indicates p ˂ 0.01, **** p ˂ 0.0001 compared to same sex no stimulus control, # # indicates p ˂ 0.01, # # # # p ˂ 0.0001 comparing males and females, same stimulus (2-way ANOVA, with the Holm-Šidák multiple comparisons correction).

fMLF stimulation of human neutrophils (43) and neutrophil-like HL60 cells (44) increases myosin light chain phosphorylation (p-MLC), the regulatory subunit responsible for myosin IIA-mediated contraction of actin filaments in the uropod during migration (45, 46). fMLF stimulus of rabbit neutrophils does not increase polymerized myosin accumulation in the cytoskeleton (47). In human neutrophils, a uniform concentration of fMLF increased myosin IIA accumulation in the cytoskeleton at 20, 40, and 120 seconds in male neutrophils and decreased myosin IIA levels at 40 seconds and increased myosin IIA levels at 1800 seconds in female neutrophils (Fig. 4C). Myosin IIA accumulation in the cytoskeleton was significantly higher in male neutrophils than female neutrophils at 40 and 120 seconds (Fig. 4C). A study on the dynamic front-rear coupling during chemotaxis of neutrophil-like HL-60 cells also observed myosin II polymerization, then myosin depolymerization, and then additional rounds of myosin II polymerization and depolymerization in migrating cells (48). We hypothesize that the decrease in myosin IIA at 40 seconds is capturing a period of myosin depolymerization. During chemorepulsion of Dictyostelium cells, AprA did not affect the levels of myosin II in the cytoskeleton (49). A uniform concentration of SLIGKV increased cytoskeletal myosin IIA in female neutrophils at 20 seconds and decreased cytoskeletal myosin IIA in male neutrophils at 1200 seconds (Fig. 4D). These data indicate that SLIGKV increases F-actin accumulation for both male and female neutrophils, with male neutrophils showing significantly more accumulated F-actin than female neutrophils, and SLIGKV temporarily increased polymerized myosin IIA in female neutrophils and temporarily decreased myosin IIA accumulation in male neutrophils, indicating that sex affects cytoskeletal protein accumulation in response to a chemorepellent.

Live male and female neutrophils had similar round morphologies with no stimulus, and similar irregular shapes after 5 minutes (300 seconds) in a fMLF gradient (Fig. 4E). However, after 5 minutes in an SLIGKV gradient, most male neutrophils appeared to have started polarizing and migrating while the majority of female neutrophils maintained a rounded cell shape (Fig. 4E). Both male and female neutrophils had similar values for average circularity and roundness after 5 minutes exposure to either no stimulus and fMLF gradients, and fMLF decreased average circularity and roundness compared to no stimulus control, indicating a more amoeboid (less rounded) morphology overall (Fig 4F and G). Male neutrophils in an SLIGKV gradient had decreased average circularity and roundness compared to both the male no stimulus control and female neutrophils in an SLIGKV gradient (Fig 4F and G), indicating that male neutrophils are more amoeboid while female neutrophils maintain a rounded shape at 5 minutes in an SLIGKV gradient. These data indicate that male neutrophils respond faster than female neutrophils to a chemorepellent.

Sex-based differences in the percent of stimulated cells, F-actin localization, and phospho-myosin light chain localization are visible up to 10 minutes after chemotactic stimulus

To determine if the sex-based differences in F-actin accumulation and cell morphology from Figure 4 affect F-actin and phospho-myosin light chain localization, neutrophils were fixed after 10 seconds or 5, 10, and 20 minutes exposure to fMLF, SLIGKV, or control buffer gradients and stained for F-actin and for phosphorylated myosin light chain. Cells were scored as having F-actin or phospho-myosin light staining in one primary region located either towards or away from the corner of the well where the stimulus was added, multiple staining (having 2 or more significant regions of staining), or diffuse staining.

At 10 seconds, in the SLIGKV gradient, compared to females, males had more cells with diffuse F-actin staining (Figs. 5A, B, and G). At 5 minutes, the small perturbation to cells caused by adding medium to the corner of the well for the control cells caused an increase in cells with multiple F-actin foci (Fig. 5H), and this then disappeared at 10 minutes (Fig. 5I). As previously observed (50), a fMLF gradient increased the number of cells with F-actin at multiple regions and toward the source of fMLF (Figs. 5H, 5I, S3C, and S3E–L). At 5 minutes in a SLIGKV gradient, some male but few female cells showed F-actin located away from the source of SLIGKV (Figs. 5C, D, and H). After 10 minutes, compared to females, more male neutrophils showed F-actin toward the source of fMLF, and F-actin away from the source of SLIGKV (Figs. 5E, F, and I). By 20 minutes, the sex-based differences in F-actin accumulation and localization had dissipated (Fig. S3C). These data indicate that male neutrophils respond faster than female neutrophils to SLIGKV.

Figure 5. Sex-based differences in F-actin and phospho-myosin light chain localization.

(A - F) Neutrophils were incubated in a gradient of SLIGKV for 10 seconds, 5 minutes, or 10 minutes followed by fixation, permeabilization, and staining for F-actin (red) and phospho-myosin light chain (green). Blue is DAPI staining of DNA. * indicates the direction toward the source of SLIGKV. Bars are 20 μm. Images are representative of 3 independent experiments (G - L). Neutrophils were incubated in gradient of buffer (Control), fMLF, or SLIGKV for 10 seconds, 5 minutes, or 10 minutes, fixed, and stained as above. Images were quantified by a blinded observer for F-actin localization and phospho-myosin light chain localization at each time point, scoring for diffuse localization, localization primarily toward or away from the source of stimulus, or localization at multiple regions at the periphery of the cell. At least 10 cells were scored for each timepoint, for each condition, in each experiment, and the average percent of cells in each category was calculated. Values are mean ± SEM from 3 male donors and 3 female donors. * indicates p ˂ 0.05, ** p ˂ 0.01, *** p ˂ 0.001, **** p ˂ 0.0001 (2-way ANOVA, with the Holm-Šidák multiple comparisons correction). Related to figure S3.

Phospho-myosin light chain localization at 10 seconds showed no significant effects of sex or gradient condition (Figs. 5A, 5B, 5J, S3E, and S3F). At 5, 10, and 20 minutes in the SLIGKV gradient, there was more diffuse staining in male cells and more female cells with phospho-myosin light chain localized in multiple areas of the cell, indicating that female neutrophils were condensing areas of phospho-myosin more rapidly than their male counterparts (Figs. 5C–F, K, and L). This trend of more male neutrophils with diffuse phospho-myosin light chain localization and more female neutrophils with phospho-myosin located in multiple regions of the cell was also observed in the control samples at 10 and 20 minutes and in the fMLF exposed cells at 20 minutes (Figs. 5K, 5L, S3D, S3K, and S3L). At 10 minutes in an fMLF gradient, there was an increased percentage of female cells with phospho-myosin localized to multiple regions, but there was no difference in the percentages of diffusely stained phospho-myosin in both male and female neutrophils (Figs. 5L, S3I, and S3J). Additionally, in the cells exposed to an fMLF gradient for 20 minutes, male neutrophils had an increased percentage of cells with phospho-myosin light chain located in the region of the cell away from the corner where the gradient initiated (i.e., the rear of a neutrophil migrating along the fMLF gradient) compared to female neutrophils (Fig. S3D, S3K, and S3L). These data indicate that male and female neutrophils localize phospho-myosin light chain in different regions of the cell, independent of the chemotactic stimulus used, and therefore, the processes that control phospho-myosin light chain localization may be differentially regulated in male and female neutrophils.

Some proteins related to the proposed chemorepulsion pathway are differentially abundant and localized in unstimulated male and female neutrophils

To determine if the sex-based differences in neutrophil chemorepulsion may be due to differences in protein abundance, unstimulated neutrophils were analyzed by proteomics. A list of protein components analyzed in Figure 1, their related family members, and G-protein subunits are shown in Supplemental Table 1. Of these proteins, RhoA, RhoB, Cdc42, and Phosphatidylinositol-5-phosphate 4-kinase type 2 gamma (PIP4K2C) were more abundant in male neutrophils (Supplemental Table 1 and indicated by blue stars in Fig. 6). This difference in protein abundance was similarly observed by western blot and immunofluorescence staining intensity of unstimulated neutrophils (Fig. 7). Female neutrophils had decreased RhoA levels as measured by western blots, however, this decrease was not observed when quantifying immunofluorescence intensity (Fig. 7A–C). This mismatch could be due to an obscured or shielded epitope that is opened after heat denaturation. RhoB, Cdc42, and PIP4K2C protein levels were decreased in female neutrophils by western blot and mean fluorescence intensity (Supplemental Table 1 and Figs. 7D–L). Proteomic analysis identified lymphocyte function-associated antigen-3 (LFA-3, also known as, and here referred to as, CD58), with an increased protein abundance in female neutrophils (Supplemental Table 1). CD58 is a cell adhesion molecule involved in leukocyte migration that initiates antigen-independent cell adhesion via the LFA-3/CD2 pathway and is integral to early immune response via expansion and induction of naïve and memory T helper cells (51). The CD58 gene (Ensembl ENSG00000116815) is located on chromosome 1, does not have an identified estrogen response element, but high throughput screening suggests that CD58 interacts with Estrogen Receptor 1 (52). Increased CD58 protein levels in female neutrophils were observed by western blot and immunofluorescence (Figs. 7M–O). These data corroborate the proteomic analysis, identifying, at minimum, RhoA, RhoB, Cdc42, PIP4K2C, and CD58 as having differential protein abundance in unstimulated male and female neutrophils.

Figure 6. Proposed pathway for PAR2-agonist induced neutrophil chemorepulsion.

The binding of SLIGKV to the extracellular surface of PAR2 causes a confirmational change of the intracellular C-terminus, releasing G-protein subunits alpha (α) and heterodimer beta-gamma (β/γ). Release of β/γ activates guanine nucleotide exchange factors (GEFs) and phosphoinositide 3-kinases (PI3K). GEFs activate Ras, Rac1, Cdc42, RhoA, and RhoB to their GTP-bound active state, thereby activating downstream pathways. Active RhoA assists in localization of Phosphatase and tensin homolog (PTEN) at the rear of the cell during migration. PTEN dephosphorylates PIP₃ to PI(4,5)P₂. Downstream of RhoA are Rho-associated kinases 1 and 2 (ROCK1/2). Cdc42 can activate Rac1 via p21-activated kinases (PAKs). PIP4K2C phosphorylates PI(5)P to PI(4,5)P₂ (path not shown). PI3K phosphorylates PIP₂ to PIP₃ and activates the mammalian target of rapamycin complex 2 (mTORC2) pathway. One of the primary effector proteins of mTORC2 is protein kinase C (PKC). Another activator of PKC is diacylglycerol (DAG) which is formed when Gα activates phospholipase C (PLC), converting PIP₂ to DAG and inositol trisphosphate (IP₃).

Protein components tested are indicated with a black outline, the inhibitor used (in parentheses), and color-coded for identification with graphs in Figure 1. Light colored components were not tested directly, but are included to clarify the connections between components. Components colored both pink and blue indicate an observed sex difference in response when this component was inhibited. A blue star indicates the protein is more abundant in male neutrophils. The proposed pathway was created using BioRender.com.

Figure 7. Some proteins in the chemorepulsion pathway are differentially abundant and have differences in localization in male and female neutrophils.

Unstimulated neutrophil lysates were analyzed by western blot to quantify levels of RhoA (A), RhoB (D), Cdc42 (G), PIP4K2C (J), or CD58 (M). All values are mean ± SEM from 4 males and 4 females. Unstimulated neutrophils were fixed, permeabilized, and stained for F-actin (red), and either RhoA (B), RhoB (E), Cdc42 (H), PIP4K2C (K), or CD58 (N) (green). Blue is DAPI staining of DNA. Bars are 10 μm. (C, F, I, L, O) Quantitation of the mean fluorescent intensity (green) in (B), (E), (H), (K), or (N) respectively, normalized to the average mean fluorescent intensity of each experiment’s male (one male and one female were used for each individual experiment) for each antibody. Images and quantitation are representative of 4 (RhoB, Cdc42, PIP4K2C) or 5 (RhoA and CD58) independent experiments. * indicates p ˂ 0.05, ** p ˂ 0.01 (Mann-Whitney t test).

To determine if proteins with differences in abundance also have differences in localization, unstimulated neutrophils were fixed and stained for RhoA, RhoB, Cdc42, PIP4K2C, and CD58, in addition to phalloidin and DAPI, to visualize filamentous actin and nuclei respectively. There were no obvious differences between male and female neutrophils in the localization of RhoA and Cdc42 (Figs. 7B and H). Cells were scored as having either diffuse (Fig. 8A), puncta (Fig. 8B), or clusters of puncta staining, here referred to as clusters (Fig. 8C). RhoB staining appeared weaker and more diffuse in female neutrophils whereas in male neutrophils RhoB tended to be more punctate and cortical (Figs. 7E, 8D, and 8E). PIP4K2C staining tended to show small, separated puncta in male neutrophils compared to clusters of puncta in female neutrophils (Figs. 7K, 8F, and 8G). CD58 staining in female neutrophils tended to show larger, denser clusters of puncta compared to male neutrophils (Figs. 7N, 8H, and 8I). Together, these results suggest that human male and female neutrophils tend to have differences in the abundance and localization of some of the pathway components mediating chemorepulsion.

Figure 8. Differences in localization of RhoB, PIP4K2C, and CD58 in male and female neutrophils.

Unstimulated neutrophils were fixed, permeabilized, and stained for F-actin (red), and either RhoB (A-D), PIP4K2C (F), or CD58 (H) (green). Blue is DAPI staining of DNA. Bars are 10 μm. Green staining in cells (E, G, and I) were scored as having either diffuse (A), puncta (B), or clusters of puncta (clusters, C) by a blinded observer and the percent of counted cells in each category were quantified in (E), (G), or (I) respectively. Images and quantitation are representative of 4 (RhoB and PIP4K2C) or 5 (CD58) independent experiments. * indicates p ˂ 0.05, ** p ˂ 0.01 (Mann-Whitney t test).

Discussion

In this report, we find that PAR2 agonist-mediated neutrophil chemorepulsion requires Ras, Rac, PKC, PTEN, and ERK 1 and 2, but PI3K and PLC are not required. This matches the Dictyostelium AprA chemorepulsion pathway (7), indicating that as with chemoattraction (1), there is a strong conservation of G protein-coupled receptor-mediated chemorepulsion signal transduction pathways between Dictyostelium and human neutrophils.

PAR2 is activated by N-terminal cleavage by serine proteases such as typsin, tryptase, and DPPIV (53). Cleavage of the PAR2 N-terminus reveals the tethered ligand, which folds down to touch the extracellular surface of PAR2 and activate the receptor and downstream pathways (53). DPPIV is expressed on the extracellular surface of many cells (54), including the surface of activated T cells (55), and is also present as an enzymatically active soluble form in plasma and other body fluids (56). At an inflammatory site, neutrophils lead the first wave of host defense, attracted by secreted chemokines and distress factors released by resident macrophages (57). Two to three days after the initial neutrophil invasion, active T cells are recruited to the inflammatory site (58), releasing a DPPIV gradient (59), activating PAR2 on the surface of persisting neutrophils, thereby potentially repelling neutrophils from the inflammatory site (9).

The neuronal chemorepellent Slit2-S acts through a receptor tyrosine kinase and also repels human neutrophils (13). Like SLIGKV, Slit2-S-induced chemorepulsion requires Ras and Rac1, relocalizes F-actin, does not require PI3K, does not affect neutrophil adhesion, or show sex differences in FMI or speed in the absence of inhibitors (13). Unlike SLIGKV, Slit2-S did not increase PI(3,4,5)P₃ or F-actin, visibly alter localization of phospho-myosin light chain, or show sex differences in persistence (13). Sex differences in the response of neutrophils to inhibitors were not measured in this report (13). Together, this indicates some differences in neutrophil chemorepulsion depending on the repellent and receptor.

Figure 6 shows canonical signal transduction pathways from a wide variety of studies (25, 43, 49, 53, 60–93). A heavy black outline indicates components inhibited in this report. Of those, only inhibition of PI3K failed to have an effect on neutrophil chemorepulsion. As observed for chemoattraction in neutrophils (65) and chemorepulsion in Dictyostelium (7), chemorepulsion in neutrophils appears to utilize several pathways acting in parallel. More information on inhibitor targets and their effect on fMLF-mediated chemoattraction and Slit2-S-mediated chemorepulsion in human neutrophils is described in Supplemental Table 2 (4, 13, 25–30, 43, 49, 64, 66, 67, 70–74, 77–79, 81, 88–116).

In the presence of a chemorepellent, compared to male neutrophils, female neutrophils showed a smaller percentage of repelled cells, a greater percentage of attracted cells, greater persistence of movement, more cell-substrate adhesion, less accumulation of PI(3,4,5)P₃, less accumulation of F-actin, differences in F-actin and phospho-myosin light chain localization, and a slower change in cell morphology. Cdc42 and ROCKs are required for female neutrophil chemorepulsion, but appear to inhibit or are not required for male neutrophil chemorepulsion. There was a sex difference in myosin II accumulation with the chemoattactant fMLF. These data indicate that there are sex differences in both chemoattraction and chemorepulsion, and these differences may be caused by inherent differences in specific protein localization and abundance of the components in the associated pathways (blue stars in Fig. 6). There are also sex-based differences in the morphology of neutrophil nuclei (117), type I interferons and immunometabolism that affect neutrophil maturation profiles and proinflammatory responses (118), and most recently, evidence of sex dimorphism in murine neutrophil transcriptomics, metabolomics, lipidomics, and chromatin accessibility (119). Together, this indicates that there are multiple sex differences between male and female neutrophils, and that there is a fundamental difference in motility mechanisms in the innate immune system in men and women.

Inhibition of PLC caused cells to be more repelled by the chemorepellent, indicating that this protein inhibits chemorepulsion in neutrophils. When PKC was inhibited, we observed the same sex difference in persistence seen with SLIGKV-induced chemorepulsion alone, indicating that the observed sex-based difference in persistence is mediated by one or more of the other pathways. Inhibition of ERK1 and 2 reduced male neutrophil-substrate adhesion in the presence of SLIGKV, suggesting that ERK1 and 2 promote male neutrophil adhesion during chemorepulsion. When either Cdc42 or ROCK1/2 were inhibited, males had increased chemorepulsion and female neutrophils were attracted to the chemorepellent, indicating that there are sex-based differences in how these components affect downstream signaling, such as cytoskeletal reorganization, which might be exacerbated by the observed decreased levels of RhoA and RhoB, and the increased levels of Cdc42, in female neutrophils, as well as the observed sex-based differences in RhoB localization (Figs. 7A–I, 8D, and 8E). Rho GTPases, such as RhoA, RhoB, and Cdc42 play an integral role in cell polarity, spatiotemporal regulation of the actin-myosin cytoskeleton, and cell migration/motility (81, 88–93, 120). Why there are these differences in male and female neutrophils is baffling.

Several neutrophil-associated diseases have a sex difference in disease incidence. For instance, acute febrile neutrophilic dermatosis (121) and rheumatoid arthritis (122) are more prevalent in women, while acute respiratory distress syndrome (123, 124), reactive arthritis (125), sexually acquired reactive arthritis (126), and gout (127) are more prevalent in men. An intriguing possibility is that the differences in male and female neutrophil chemorepulsion we observe in this report may be correlated with differences in neutrophil associated disease incidences in men and women.

Key Points.

1. Chemorepulsion mechanisms are conserved between Dictyostelium and human neutrophils.

2. Male and female neutrophils have differences in chemorepulsion mechanisms.

3. Male and female neutrophils have differences in protein levels and localization.

Glossary

AprA

Autocrine proliferation repressor protein A

Cdc42

Cell division cycle 42

CD58

Lymphocyte function-associated antigen-3 or LFA-3

CnrN

Cell number regulator N

DAG

Diacylglycerol

DPPIV

Dipeptidyl peptidase IV

FMI

Forward migration index

GEF

Guanine nucleotide exchange factor

GRP

Gastrin-releasing peptide

HL60

Human Caucasian promyelocytic leukemia cell line

IP₃

Inositol trisphosphate

LFA-3

Lymphocyte function-associated antigen-3 or CD58

LTB₄

Leukotriene B4

mTORC2

mammalian target of rapamycin complex 2

PAR2

Protease activated receptor 2

PIP4K2C

Phosphatidylinositol-5-phosphate 4-kinase type 2 gamma

PI(4,5)P₂

Phosphatidylinositol 4,5-bisphosphate

PI(3,4,5)P₃

Phosphatidylinositol 3,4,5-trisphosphate

PKC

Protein kinase C

PLC

Phospholipase C

p-MLC

phospho-Myosin light chain

PTEN

Phosphatase and tensin homolog

PVDF

Polyvinylidene difluoride

RIP

Ras inhibitory peptide

ROCK

Rho-associated kinase

SLIGKV

SLIGKV-NH2

Slit2-S

110kDa N-terminal Slit2 fragment

Sos1

Son of sevenless1