Bioparticle Microarrays for Chemotactic and Molecular Analysis of Human Neutrophil Swarming in vitro

Neutrophils, the most abundant white blood cell in the bloodstream¹, are gaining attention as potential diagnostic and therapeutic targets^2,3 because they may be involved in a variety of medical conditions including gout⁴, sepsis³, trauma^5,6, cancer^1,7,8, and various autoimmune diseases^5,9. Neutrophil swarming is a multistage, tightly regulated process with a complexity that makes it a particularly interesting focus of study^5,10,11. During swarming, neutrophils isolate a site of inflammation from the surrounding healthy tissue^5,10,11. Proper regulation of neutrophil swarming is essential to promote wound healing and ultimately inflammation resolution^5,12. Neutrophil swarming has primarily been studied in vivo in rodent^12,13,14,15 and zebrafish^10,11,12,15 models. However, the nature of these in vivo animal models gives rise to limitations⁵. For example, the mediators released by neutrophils during swarming are not easily accessible for analysis⁵. Additionally, there are many potential sources for a given mediator in vivo, so an in vivo experiment must introduce a genetic deficiency to inhibit cellular production and/or interaction in order to investigate the role of that mediator in a given process¹³. An in vitro experiment circumvents this complication by enabling neutrophil observation without the context of additional cells. Additionally, research describing human neutrophil coordinated migration is limited¹⁶. On an in vitro swarming platform, human neutrophils can be directly analyzed. An in vitro swarming platform could expand upon the knowledge gained from in vivo studies by providing opportunities to fill the gaps left by the limitations of in vivo studies.

To address the need for an in vitro platform that mimics in vivo neutrophil swarming, we developed a microstamping platform that enables us to pattern bioparticle microarrays that stimulate neutrophil swarming in a spatially controlled manner. We generate bioparticle microarrays on glass slides in a two-step process. First, we use microstamping to generate a microarray of cationic polyelectrolyte (CP) spots. Second, we add a solution of bioparticles that adhere to the CP spots via electrostatic interaction. By first patterning the CP layer, we can selectively pattern negatively charged bioparticles to generate the desired neutrophil swarming pattern. The positively charged layer holds the negatively charged bioparticles through the vigorous washing step that removes the bioparticles from the areas on the glass slide that do not have the CP. Additionally, the CP used here, a copolymer of acrylamide and quaternized cationic monomer, is biocompatible, so it does not induce a response from the neutrophils. It has a very high surface charge that immobilizes the micron-sized bioparticles to the glass slide, thus inhibiting neutrophils from removing the particles from the patterned position on the glass slide. This results in bioparticle clusters arranged in a microarray. When we added neutrophils to the microarray, they formed stable swarms around the bioparticle clusters. Through tracking neutrophil migration, we found that swarming neutrophils actively migrate toward the bioparticle clusters. Furthermore, we used this platform to analyze certain mediators that neutrophils release during swarming. We found 16 mediators that are differentially expressed during swarming. Their concentrations follow three general trends over time: increase, decrease, or spike. Our in vitro neutrophil swarming platform facilitates the analysis of spatially controlled human neutrophil swarming, as well as the collection and analysis of mediators released by neutrophil swarming. In a previous publication, we demonstrated that patients with certain medical conditions (trauma, autoimmune disease, and sepsis), had neutrophils that functioned differently than those from healthy donors⁵. In future research studies, our platform could be used to analyze neutrophil function among a variety of patient populations. This platform can quantitatively analyze the complex coordination involved in neutrophil swarming. Additional studies can be done to provide insight on the neutrophil function of a specific patient population or neutrophil response to a pathogen of interest.

The authors acknowledge the healthy volunteers who kindly donated their blood. Blood specimens were obtained after informed volunteer consent according to institutional review board (IRB) protocol #2018H0268 reviewed by the Biomedical Sciences Committee at The Ohio State University.

1. Microfabrication of bioparticle microarray

1. Using standard photolithography procedures, generate the master silicon wafer.
   1. Generate a proof of the desired design using a computer-aided design (CAD) software, then send to a photomask manufacturer to produce a chrome photomask. The design used here is 4 mm x 4 mm rectangular arrays of 30 µm diameter filled in circles with a 150 µm center-to-center spacing. This design can be modified as desired for different applications.
   2. Spincoat a 40 µm thick layer of a negative photoresist onto a silicon wafer. Bake wafer at 65 °C for 5 min and 95 °C for 10 min.
   3. Expose wafer to UV light through a chrome photomask with 150–160 mJ/cm² (Figure 1A).
   4. Bake wafer at 65 °C for 5 min and 95 °C for 10 min. Submerge wafer in photoresist developer for 10 min and rinse with isopropyl alcohol. At this stage, the pattern should be visible on the wafer (Figure 1B).
2. Thoroughly mix a 10:1 ratio of polydimethylsiloxane prepolymer and its curing agent (i.e., 20 g of prepolymer and 2 g of curing agent) and pour the uncured polydimethylsiloxane (PDMS) mixture over the master wafer in a Petri dish (Figure 1C).
3. Vacuum treat the uncured PDMS mixture until no air bubbles are present over the master wafer. Cure at 65 °C overnight.
4. Use a scalpel to cut around the exterior of the patterned section of the wafer, and slowly remove the cured PDMS slab. Place the PDMS slab on a clean cutting board with the patterned side facing up (Figure 1D).
5. Punch out individual stamps from the PDMS slab with an 8 mm biopsy punch (Figure 1E). For one glass slide, eight stamps will be needed.
6. Place each stamp face down on the adhesive tape to remove any debris.
7. In advance, prepare a 1.6 mg/mL solution of CP in water.
   1. Add the proper amount of the CP powder to water (e.g., 0.8 g to 500 mL).
   2. Mix on a stir plate at room temperature overnight, or until all the solid is dissolved into the water. The CP solution can be stored at room temperature for 6 months.
   3. If desired, make the CP solution fluorescent by adding poly-L-lysine labelled with fluorescein isothiocyanate (PLL-FITC).
      1. Aliquot about 10 mL of the CP solution. Add a small amount of PLL-FITC (0.05 mg) to the aliquoted volume. The amount can be altered to adjust the brightness of fluorescence as desired.
      2. Vortex the CP solution labeled with FITC for 20 s, or until the solution is a uniform, pale yellow color. Protect from light and store at 4 °C for up to 1 month.
8. With the stamps face-up, prime each stamp with 100 µL of 1.6 mg/mL solution of CP, ensuring no air bubbles form between the CP solution and the stamp (Figure 1F).
9. Invert the stamps onto a layer of CP solution (Figure 1G).
10. Remove the stamps from the CP solution after 1 h.
11. Dab each wet stamp face down onto a clean glass slide 6–8x to remove excess liquid.
12. Vacuum treat the stamps for 1–2 min.
13. Adhere an eight well, 9 mm diameter imaging spacer on the top of a clean glass slide as a guide for stamp placement. With this spacer, each glass slide can have eight microarrays.
14. Place a stamp face down on the glass slide in the center of each well of the imaging spacer (i.e., use eight stamps total).
15. Place a 5.6 ± 0.1 g balanced weight on top of each stamp and allow 10 min for stamping (Figure 1H). A balanced weight is required to ensure the stamp is pressed evenly onto the glass slide and promote even transfer of the CP to the glass slide.
16. Remove the weights and stamps from the glass slide (Figure 1I). Allow the CP layer to dry at room temperature for 24 h before adding the bioparticles, as described in step 1.17. If the CP is tagged with FITC, the effectiveness of the stamping can be checked at this point with a fluorescent microscope at 488 nm before proceeding to step 1.17 (Figure 1M).
17. Cut a blank PDMS slab to the size of the imaging spacer and use the 8 mm biopsy punch to create wells in the PDMS that align with the wells of an imaging spacer. Adhere the PDMS slab to the glass slide with the imaging spacer (Figure 1J).
18. Thaw a solution of bioparticles (e.g., E. coli or zymosan) and dilute to 500 μg/mL in water for injection (WFI).
    NOTE: The bioparticles do not need to be opsonized. Neutrophil surface receptors directly recognize molecules on these bioparticles^19,20,21.
19. Add 100 μL of bioparticle solution to each PDMS well on the glass slide (Figure 1K).
20. Rock the glass slide for 30 min.
21. Rinse the wells thoroughly with water. The bioparticle microarray can be stored in a dust-free environment at 4 °C for up to 3 months. At this point, the pattern should be checked with a fluorescent microscope at 594 nm before proceeding to step 2.1 (Figure 1N).

2. Sample Preparation

1. Collect at least 2 mL of fresh blood in K2-EDTA tubes from the desired donor. The expected yield of neutrophils is 1–2 x 10⁶ cells/1 mL whole blood. The imaging assay requires approximately 1.5 x 10⁵ neutrophils, and the analysis of the supernatant requires 1 x 10⁶ neutrophils. Use the blood within 4 h.
2. Separate red blood cells (RBCs) by adding an erythrocyte aggregation agent in a 1:5 ratio to the whole blood. Wait 45 min for a translucent layer (buffy coat) to separate from the layer of RBCs.
3. Remove the buffy coat and wash with phosphate buffered saline (PBS) using 1 mL buffy coat: 9 mL PBS.
4. Centrifuge for 5 min at 190 x g and 20 °C.
5. Aspirate the supernatant and resuspend the pellet at 5 x 10⁷ cells/mL.
6. Use a negative immunomagnetic selection kit to isolate neutrophils.
   1. Add 50 μL of antibody cocktail/1 mL cell suspension. Wait 10 min.
   2. Add 100 μL of magnetic beads/1 mL cell suspension. Wait 10 min.
   3. Add cell suspension to a round-bottom tube and place in a cylindrical magnet. Wait 10 min.
7. Pour supernatant into a centrifuge tube. Add up to 10 mL of PBS. Centrifuge for 5 min at 190 x g and 20 °C.
8. Aspirate supernatant. Resuspend white pellet in IMDM with 0.4% human serum albumin.
9. Stain nuclei with 20 µg/mL Hoechst 33342 for 10 min at 37 °C.
10. Add 5 mL of IMDM with 0.4% human serum albumin to rinse. Centrifuge for 5 min at 190 x g and 20 °C.
11. Resuspend cells at 7.5 x 10⁵ cells/mL in IMDM with 0.4% human serum albumin.
12. Add 100 µL of the cell suspension to a PDMS well containing a bioparticle microarray. Ensure the cell suspension is convex over the top of the PDMS well and does not contain any bubbles.
13. Seal with a 12 mm diameter coverslip. Cover the opening of the PDMS well with a 12 mm diameter coverslip. Press down gently onto the coverslip with tweezers so the excess cell suspension escapes to the edge of the well. Use a tissue to remove the excess cell suspension.

3. Running the assay and image analysis

1. Load microparticle array with cells on the live cell imaging station with a microscope equipped with a cage incubator set to 37 °C, 5% CO₂, and 90% relative humidity.
2. Use time-lapse fluorescent and brightfield microscopy to record images at 10x magnification every 10 s at 405 nm, 594 nm, and brightfield. In a typical experiment, images are collected up to 2 h.
3. Use an automated cell tracking software to track the migration of individual neutrophils toward the bioparticle cluster.
   1. Use the autoregression mode of a spot detection cell tracking software. Set the spot radius to 5 µm (the approximate size of a neutrophil nucleus). Set the minimum track length to 120 s and a maximum gap size of one frame.
   2. From the data generated by the cell tracking software, extract the files that contain the neutrophil position and speed. These files can be used with a graphing software to generate neutrophil migration tracks (Figure 2C) and a heat map of speed vs. time (Figure 2D), respectively.
4. Use the 405 nm fluorescent images to track swarm size over time on an image analysis software of your choice.
   1. Define regions of interest (ROIs) around each bioparticle cluster where neutrophils will swarm. Keep the same size ROI to analyze each bioparticle cluster.
   2. Analyze the mean fluorescent intensity of the 405 nm images within each ROI over time.
   3. Generate a calibration curve of mean fluorescent intensity to swarm size by taking manual measurements at various swarm sizes from 0 µm² to the maximum swarm size. Use this calibration to calculate the swarm size over time.

4. Supernatant collection and protein detection

1. Incubate the neutrophils in the wells containing the bioparticle microarray at 37 °C and 5% CO₂ for 3 h. Take samples at desired time points. Typically, samples will be taken at 0, 0.5, 1, and 3 h. To overcome the limit of detection of the protein array assay, the entire volume of supernatant of a single well (200 µL) was used for each time point. Each time point was analyzed in triplicate.
2. Using a brightfield microscope, verify that swarms are formed on the microarray.
3. Aspirate the supernatant with a 200 µL pipette and load in a 0.45 μm centrifuge filter tube.
4. Centrifuge the supernatant at 190 x g and 20 °C for 5 min and collect the filtrated volume.
5. Store samples at -80 °C until the processing time.
6. Use a microarray kit that detects a range of human proteins to process samples.
   1. Add 200 μL of each sample to a separate dialysis tube provided with the kit.
   2. Place the dialysis tubes in a beaker containing at least 500 mL of PBS (pH = 8.0). Stir gently on a stir plate for at least 3 h at 4 °C. Change the PBS in the beaker and repeat this step.
   3. Transfer each sample to a clean centrifuge tube and centrifuge at 9,000 x g for 5 min to remove any precipitates. Transfer each supernatant to a clean tube.
   4. Biotinylate each sample by adding 36 µL of 1x labeling reagent from the kit per 1 mg of total protein in the dialyzed sample to 180 µL of dialyzed sample. Incubate at 20 °C for 30 min. Mix gently every 5 min.
   5. Add 3 µL of stop solution provided with the kit into each sample tube. Transfer each sample to a fresh dialysis tube and repeat steps 4.6.2–4.6.3. At this stage, the sample can be stored at -20 °C or -80 °C until you are ready to proceed.
   6. The glass slide provided with the kit is stored at -20 °C. Allow it to come to room temperature. Place the assembled glass slide in a laminar flow hood for 1–2 h at room temperature.
   7. Add 400 μL of the blocking buffer provided with the kit into each well of the assembled glass slide. Incubate at room temperature for 30 min.
   8. Centrifuge the prepared samples for 5 min at 9,000 x g to remove precipitates or particulates. Dilute 5x with blocking buffer.
   9. Remove the blocking buffer from each well. Add 400 μL of the diluted samples into the appropriate wells. Incubate for 2 h at room temperature while rocking.
   10. Decant the samples from each well. Wash 3x with 800 μL of the 1x wash buffer I provided with the kit at room temperature for 5 min each while rocking.
   11. In a clean container, submerge the assembled glass slide in 1x wash buffer I. Wash 2x at room temperature for 5 min each while rocking.
   12. Add 400 μL of 1x Cy3-conjugated streptavidin to each sub-array. Cover with plastic adhesive strips. Protect from light for the remainder of the protocol.
   13. Incubate for 2 h at room temperature while rocking.
   14. Decant the solution and disassemble the glass slide from the sample chambers.
   15. In the 30 mL centrifuge tube provided with the kit, carefully add the glass slide and enough 1x wash buffer I to cover the glass slide. Wash 3x for 10 min each at room temperature while rocking.
   16. In the 30 mL centrifuge tube, wash 2x with 1x wash buffer II for 5 min each at room temperature while rocking.
   17. Wash the glass slide with 30 mL of ddH₂O for 5 min. Remove the glass slide from the centrifuge tube and allow to dry for 20 min in a laminar flow hood. The prepared glass slide may be stored at -20 °C until ready to scan.
   18. Scan the glass slide with a microarray scanner at a fluorescence emission of 555 nm.

When neutrophils are added to the bioparticle microarray, neutrophils that contact the bioparticle clusters become activated and initiate the swarming response. The bioparticle microarray was validated using time-lapse fluorescent microscopy to track neutrophil migration toward the bioparticle clusters (Video S1). The migration of individual neutrophil nuclei is tracked as they migrate toward the bioparticle cluster. When neutrophils reach the bioparticle cluster, their nuclei overlap with other nuclei in the cluster. Thus, it is not possible to accurately track a neutrophil within the cluster using this method. Zymosan and E. coli bioparticle clusters both result in the generation of neutrophil swarms. For our results, Figure 2B uses data from neutrophil swarms generated by E. coli particles. Figure 3 and the other panels of Figure 2 use neutrophil swarms generated by zymosan particles. The results obtained demonstrate that bioparticle clusters stimulates neutrophil activation when a neutrophil contacted the cluster, which ultimately led to the formation of stable neutrophil swarms around each cluster after 30–60 min (Figure 2A, top). In contrast, neutrophils did not show collective migration in the absence of bioparticle clusters (Figure 2A, bottom, and Video S2). Using the fluorescent intensity of stained neutrophil nuclei at 405 nm, the average neutrophil swarm size around 30 µm diameter E. coli bioparticle clusters was found to be 1,490 ± 680 µm² (mean ± SD, Figure 2B, top). The fluorescence intensity of a given region of interest where no bioparticle clusters are present was approximately constant over time, which confirmed the absence of collective migration in this setting (Figure 2B, bottom). Tracks of neutrophil migration show that neutrophils converged on a bioparticle cluster when one was present (Figure 2C, top). Conversely, no convergence was observed in the control system (Figure 2C, bottom). The speed (distance travelled/time) of swarming and nonactivated neutrophils was measured and a statistically significant difference in the speed distributions was found (ANOVA, p < 0.0001), as shown in Figure 2D. The average speed for swarming neutrophils was 20.6 ± 13.0 µm/min (mean ± SD), and the average speed for control neutrophils was 2.0 ± 2.2 µm/min.

Additionally, the concentration of 16 proteins that neutrophils released during swarming were analyzed (Figure 3). The normalized concentration of each protein at different time points in the swarming process (t = 0, 0.5, 1, and 3 h) were calculated, where the normalized concentration is (C – C_min) / (C_max – C_min). 10 proteins increased in concentration throughout swarming. These proteins were adipsin, galectin-3, GROa, IL-6R, MIP-1a, MMP-8, MMP-9, Nidogen-1, TIMP-1, and TLR2. Two proteins (pentraxin 3 and RANK) decreased in concentration throughout swarming. The remaining four proteins (clusterin, PF4, RANTES, and Trappin-2) increased during the first hour of swarming but decreased thereafter. In other words, the concentrations of those proteins "spiked" during swarming. Of the 16 proteins identified, 12 proteins were identified in our previous publication⁵. Adipsin, galectin-3, nidogen-1, pentraxin 3, TIMP-1, and TLR2 were shown to be swarming-specific, while clusterin, IL-6R, MMP-8, and MMP-9, RANK, and trappin-2 were not⁵.

Figure 1: Production of bioparticle microarray. A silicon wafer coated with negative photoresist is exposed to UV light through a chrome photomask (A). After the silicon wafer is baked and developed, a photoresist pattern remains on the surface of the silicon wafer. This is the master wafer (B). In a Petri dish, a PDMS mixture is added to the top of the master wafer. The PDMS is cured overnight at 65 °C to form the PDMS mold (C). The PDMS mold is cut from the master wafer and a biopsy punch is used to punch out individual PDMS stamps (D). A PDMS stamp (E) is coated with CP solution (F). The stamp is inverted onto a thin layer of CP solution (G). After incubating in the CP solution for 1 h, the stamp is blotted onto a glass slide to remove excess CP. Eight stamps are then pressed onto a clean glass slide with a 5.6 ± 0.1 g weight, aligned with an imaging spacer (H). When the stamp is removed, a CP pattern remains (I). A PDMS slab with precut wells is adhered to the glass slide (J). Then, a bioparticle solution is added over the CP pattern (K). The negatively charged bioparticles bind to the positively charged polyelectrolyte via electrostatic interaction. The excess bioparticle solution is washed away, leaving bioparticles patterned on top of the CP pattern (L). (M) Fluorescent image of the CP layer labeled with FITC (Scale bars = 50 µm, large image; Scale bar = 25 µm, inset). (N) Fluorescent image of patterned zymosan bioparticles conjugated with Texas Red (Scale bar = 100 µm, large image; Scale bar = 50 µm, inset). Please click here to view a larger version of this figure.

Figure 2: Neutrophil swarm growth around bioparticle clusters. In the presence of bioparticle clusters, neutrophils undergo collective migration toward the bioparticle clusters (bottom, control neutrophils). When no bioparticles are present, the neutrophils do not perform collective migration. (A) In the presence of zymosan bioparticle clusters, neutrophils form swarms in 30 min. Without bioparticle clusters, no swarms form (Scale bars = 50 µm). (B) Neutrophil swarms grow to an average size of 1,490 ± 680 µm² (mean ± SD) around 30 µm diameter E. coli bioparticle clusters. Control neutrophils exhibit a constant density that corresponds to no swarm growth (ANOVA, p < 0.0001, n = 32 neutrophil swarms, N = 1 donor, error bars = standard deviation). (C) Tracks of swarming neutrophils converge on the zymosan bioparticle clusters, while control neutrophils show no converging point (Scale bars = 50 µm). (D) Neutrophils swarming toward a zymosan target have a mean speed of 20.6 ± 13.0 µm/min (mean ± SD), while control neutrophils have a mean speed of 2.0 ± 2.2 µm/min. Each count on the heat map represents a neutrophil with the given instantaneous speed at the given time point. These heat maps are representative of one experiment (n = 1 swarm; N = 1 donor; ANOVA, 6,114 = swarming neutrophils, 32,116 = control neutrophils, p < 0.0001). Please click here to view a larger version of this figure.

Figure 3: Free mediators released by swarming neutrophils. Neutrophil swarming affects the production of various proteins over time. Protein concentration was measured at 0, 0.5, 1, and 3 h. The data points were fitted with smoothing splines (λ = 0.05). These proteins tend to follow one of three characteristic trends: decreasing over time, increasing over time, or spiking around 1 h and then decreasing (Error bars = standard deviation; n = 3 replicates; N = 1 donor). Please click here to view a larger version of this figure.

Video S1: Neutrophil swarming toward zymosan bioparticle clusters. Neutrophil nuclei are shown in blue. Zymosan targets are marked with red circles (Scale bar = 50 µm; original acquisition time = 60 min). Please click here to download this video.

Video S2: Nonactivated neutrophil random migration. Neutrophil nuclei are shown in blue (Scale bar = 50 µm; original acquisition time = 60 min). Please click here to download this video.

We developed a microstamping platform to generate uniform arrays of bioparticles to stimulate in vitro neutrophil swarming. The in vitro nature of our platform allows us to circumvent the complications that arise with in vivo swarming experiments, namely the poor ability to analyze mediators released by swarming neutrophils⁵. Additionally, in vivo models are typically performed in rodents^{11,12,13,15,22,23,} or zebrafish^11,12,15,23. Our platform uses human neutrophils, which enables us to more directly interpret our results in the context of human disease, though certain similarities between mouse and human neutrophils have been observed^5,11. Additionally, we maintain a spatially-controlled swarming environment that distinguishes our platform from in vivo models by providing a high level of reproducibility that facilitates the analysis of human neutrophil collective migration as well as the collection and analysis of mediators released by swarming neutrophils.

During the development of our microstamping protocol, several challenges arose that required careful troubleshooting. First, the CP used for the microstamping is highly hydrophilic, and the PDMS stamps are hydrophobic. Because the CP does not have a high affinity for PDMS, our procedure was carefully designed to avoid the formation of bubbles and promote wetting. By first priming the stamp with CP while face-up (step 1.8), we minimize the formation of bubbles between the CP and the stamp. The stamp is then inverted onto a layer of CP and incubated for 1 h. This long incubation time ensures that every section of the stamp is wetted. Second, the process of removing excess CP before stamping on a clean glass slide (step 1.11) can be inconsistent. While performing step 1.11, the stamp must be carefully examined. When the pattern begins to become visible, the stamp is ready for step 1.12. Additionally, the required vacuum time to dry the stamps (step 1.12) can vary. This is primarily dependent on the weather. On a warm, humid day, 2 min of vacuum time is required. On a cool, dry day, 1 min of vacuum time is sufficient.

We have shown that neutrophils from healthy donors form stable swarms around bioparticle clusters. With time-lapse fluorescent microscopy, we can quantify swarm size and track neutrophil migration, which enables us to analyze neutrophil chemotaxis quantitatively⁵. For example, we have previously shown that this platform can be used to calculate the chemotactic index (CI, the cosine of the angle between the neutrophil velocity vector and the position vector between the neutrophil and the nearest bioparticle cluster), speed (distance the neutrophil travels divided by time), radial velocity (the speed multiplied by CI), and the total distance traveled (the difference between initial and final neutrophil position) of individual migrating neutrophils⁵. Unlike most in vitro studies^24,25,26, our platform does not have an artificial chemotactic gradient, so neutrophil intercellular communication is the sole driving force of neutrophil migration. Additionally, the supernatant generated by swarming neutrophils is easily accessible. We can collect and analyze the supernatant for intercellular mediators released by neutrophils without interference from other cell types that are present in vivo. 16 proteins were observed whose expression during swarming could be described as one of three trends: increase (10 proteins), decrease (two proteins), and spike (four proteins). Six of these proteins confirmed previously reported trends during swarming over time⁵. Some of the identified proteins were previously shown to be swarming-specific, while others were expressed differentially by activated non-swarming neutrophils⁵. Proteins that increase in concentration over time are likely associated with the pro-inflammation response. Some of the proteins that increase in concentration over time are already known to be involved in the pro-inflammatory response (e.g., galectin-3 and MMP-9)^27,28. The relationship between other proteins and inflammation is less well understood. The proteins that spike or decrease during swarming may be involved in the regulation of inflammation. However, further research is necessary to understand the role in inflammation of many of the proteins that are differentially expressed during swarming. Analyzing released mediators along with neutrophil migration can help us better understand the complex picture of inflammation and how neutrophils impact the surrounding tissue during swarming.

In future studies, our platform can be used to thoroughly study the ability of unhealthy neutrophils to generate stable swarms around patterned bioparticle clusters. Various medical conditions have been related to neutrophils, including sepsis³, trauma⁶, and cancer^1,8,17,18, which suggests neutrophils in these patients may have altered function. This platform can be used to examine differences between the intercellular mediators released by healthy and unhealthy neutrophils. Additionally, this platform could be modified to pattern live microbes and used to analyze neutrophil response to live microbes in vitro.

In conclusion, we have developed a novel platform for analyzing neutrophil swarming in vitro. The highly controlled nature of our platform allows us to mitigate issues that arise during in vivo neutrophil swarming experiments. Neutrophil swarming on a bioparticle microarray is easily quantifiable via time-lapse fluorescent microscopy. Additionally, we can collect the mediators released by the neutrophils without the interference of other tissues that are present in vivo. This platform can be used in future research to quantify differences between the migration behaviors of neutrophils from healthy and unhealthy donors.