A Flow Cytometry‐Based Assay for High‐Throughput Detection and Quantification of Neutrophil Extracellular Traps in Mixed Cell Populations

Abstract

Neutrophil extracellular traps (NETs) are web‐like structures composed of decondensed chromatin and antimicrobial proteins that are released into the extracellular space during microbial infections. This active cell death program is known as NETosis. To date, florescence microscopy is the widely accepted method for visualization and quantification of NETs. However, this method is subjective, time consuming and yields low numbers of analyzed polymorphonuclear cells (PMNs) per sample. Increasing interest has emerged on the identification of NETs using flow cytometry techniques. However, flow cytometry analysis of NETs requires particular precautions for sample preparation to obtain reproducible data. Herein, we describe a flow cytometry‐based assay for high‐throughput detection and quantification of NETosis in mixed cell populations. We used fluorescent‐labeled antibodies against cell markers on PMNs together with a combination of nucleic acid stains to measure NETosis in whole blood (WB) and purified PMNs. Using plasma membrane‐impermeable DNA‐binding dye, SYTOX Orange (SO), we found that cell‐appendant DNA of NETting PMNs were positive for SO and DAPI. The combination of optimally diluted antibody and nucleic acid dyes required no washing and yielded low background fluorescence. Significant correlations were found for NETosis from WB and purified PMNs. We then validated the assay by comparing with time‐lapse live cell fluorescence microscopy and determined very good intraassay and interassay variances. The assay was then applied to a disease associated with NETosis, systemic lupus erythematosus (SLE). We examined PMA‐induced NETosis in peripheral PMNs from SLE patients and controls and in bone marrow PMNs from multiple murine models. In summary, this assay is observer‐independent and allows for rapid assessment of a large number of PMNs per sample. Use of this assay does not require sophisticated microscopic equipment like imaging flow cytometers and may be a starting point to analyze extracellular trap formation from immune cells other than PMNs. © 2018 The Authors. Cytometry Part A published by Wiley Periodicals, Inc. on behalf of International Society for Advancement of Cytometry.

Short abstract

Introduction

Neutrophils, also known as polymorphonuclear cells (PMNs), are the first and most numerous innate immune cells recruited to sites of infection 1. Elimination of pathogenic material is achieved through a number of processes that include phagocytosis and the generation of reactive oxygen species via respiratory burst 2. This is assisted by the release of microbicidal peptides from cytoplasmic granules. The life cycle of the neutrophil includes maturation in the bone marrow before release into the circulation 3. Clearance of senescent neutrophils is completed by the reticuloendothelial system 3. In the bone marrow granulocyte‐macrophage progenitor cells give rise to pre‐neutrophils which are a proliferative neutrophil precursor population. These then develop into immature neutrophils before differentiating into mature neutrophils 4. Neutrophils have a low level of gene expression but are equipped with proteins required to kill microbes by the time they reach the circulation 5. Neutrophil elastase (NE) and myeloperoxidase (MPO) are stored in the azurophilic granules of naïve neutrophils 6. NE is serine protease that is neutrophil‐specific. It primarily degrades virulence factors and destroys bacteria 6. MPO catalyzes the oxidation of halide ions such as chloride in the presence of hydrogen peroxide, to generate hypochlorous acid and other reactive products that mediate antimicrobial action 7.

Over a decade ago, a new form of active cell death program in neutrophils was described and named NETosis 8, 9. NETosis is morphologically different from other forms of cell death such as apoptosis or necrosis 10. Neutrophil activation by microbes and other stimuli promotes reactive oxygen species (ROS) generation and the action of NE and MPO after their translocation from the azurophilic granules to the nucleus 6. NE cleaves histones and stimulates chromatin decondensation in the nucleus 6. In the late stages of this process, MPO binds to chromatin to drive further decondensation and the release of mesh‐like structures with a chromatin backbone with attached globular domains, termed neutrophil extracellular traps (NETs) 6, 9. The locally elevated concentration of antimicrobial proteins (e.g., citrullinated histones, MPO and NE) trap and promote the clearance of bacteria, fungi, parasites and viruses by NETs (Fig. 1A) 2, 9. This dependency on neutrophil‐specific factors explains in part why extracellular trap (ET) formation is observed mainly in PMNs, which are the most efficient producers of ROS, primarily through NADPH oxidase‐dependent pathways 6, 11, 12. While NETs have antimicrobial roles, they also contribute to autoimmune diseases such as systemic lupus erythematosus (SLE) 13. Defective NET degradation has been associated with increased risk of lupus nephritis, with NETs observed in the kidney biopsy of a SLE patient with impaired NET degradation 14.

Figure 1.

Visualizing NETosis on immunofluorescence confocal microscopy. (A) Schematic diagram representing a resting PMN undergoing activation and NET formation after encounter with pathogens. (B) Neutrophils were stained with anti‐CD66‐FITC and Hoechst. Surface expression of CD66 was upregulated following 30 nM PMA stimulation. (C) The long extended NET‐DNA structures are clearly visible with SO and DAPI following 30 nM PMA treatment. Images were acquired using 40× oil objective on a Zeiss LSM800 confocal microscope.

As its original description has become apparent that in addition to nuclear DNA, mitochondrial DNA (mtDNA) could also be used by lymphocytes, monocytes, and PMNs to form ETs without induction of cell death 15, 16. In addition, none of the proteins associated with NETs are present in B cell ETs 16. Hence, although ETs classically contain citrullinated histones, MPO and NE; their specific composition continues to be defined 17.

SYTOX Orange (SO) is an intercalating dye that has negligible intrinsic fluorescence but large fluorescence enhancement (~450‐fold) upon binding to double‐stranded DNA 18, 19, 20. Therefore, it stands to reason that cell‐appendant NET DNA could be detected using flow cytometry with SO 21. In addition, SO does not cross the plasma membrane of live cells 20. There has been increasing interest for the identification of NETs utilizing flow cytometry techniques 22, 23, 24, 25. However, flow cytometry analysis of NETs requires particular precautions for sample preparation to obtain reproducible data. In particular, this method should differentiate NETosis from other forms of cell death such as necrosis or apoptosis 21. In this report, we describe an optimized protocol utilizing a combination of SO and 4′,6‐diamidino‐2‐phenylindole (DAPI) for two‐color visualization of NETs on flow cytometry 20. SO⁺DAPI⁺ PMNs activated with phorbol 12‐myristate 13‐acetate (PMA) was dose‐ and time‐dependent, as previously described 8, 25. To verify that SO⁺DAPI⁺ PMNs were not due to other forms of cell death, we determined the following findings: (1) minimal SO⁺DAPI⁺ PMNs with epoptoside‐induced apoptosis; (2) correlation in the percentage of SO⁺DAPI⁺ PMNs with NETosis rate obtained using time‐lapse live cell microscopy; and (3) colocalization of extracellular DNA and NE on flow cytometry. Last, this flow cytometric‐based assay can be used for both human and murine samples, including purified PMNs or mixed cell populations extracted from whole blood (WB) or bone marrow (BM). In addition, this is the first flow cytometry assay, which does not require purification as isolation procedures may activate the PMNs 22, 24, 25, 26.

Materials and Methods

Subjects

Patients with SLE attending the outpatient clinics of National University Hospital who fulfilled at least four of the revised American College of Rheumatology revised classification criteria or healthy donors (HD) were invited to take part in the study 27. The study was approved by National Healthcare Group Domain Specific Review Board E (reference code: 2014/00600) and National University of Singapore Institutional Review Board (reference code: 09–256). The study was carried out in accordance with the principles of the Declaration of Helsinki. All subjects gave written informed consent prior to recruitment. Blood was collected in ethylenediaminetetraacetic acid (EDTA) tubes for flow cytometry and microscopy. Purified PMNs and WB were used for these experiments.

Animals

All mice were bred in the Biological Resource Centre, A*STAR in Singapore. Mice were 8–10 week old females. Breeding pairs for C57BL/6 J (B6) mice were originally obtained from the Jackson Laboratory, ME, US. The derivation of the B6.Sle1 (Sle1) and B6.Sle1Sle2Sle3Sle5 (Sle123) congenic mice bearing NZM2410‐lupus susceptibility intervals and the generation of the conditional BAC Tg7 have been previously described 28, 29, 30. Breeding pairs for all lupus‐prone mice were all a kind gift from Prof. Ward Wakeland, UT Southwestern Medical Center, Dallas, TX, US. All mice procedures conformed to National Institutes of Health guidelines and were conducted according to an Institutional Animal Care and Use Committee approved protocol (reference code: 161176).

Blood Sample Collection and Processing

Venous blood from SLE patients and HDs was collected in EDTA tubes. PMNs were purified from WB using Histopaque‐1077 (Sigma‐Aldrich, MO, US)/Histopaque‐1119 (Sigma‐Aldrich, MO, US)‐based density gradient centrifugation. Briefly, gradients were centrifuged at 400g for 30 min at room temperature (RT) without brake and PMNs at the interface between Histopaque‐1077 and Histopaque‐1119 were collected. Red blood cells were lysed using ACK lysing buffer (Lonza, Basel‐Stadt, Switzerland). The PMNs were washed in phosphate buffered saline before being resuspended in RPMI (Gibco, MA, US) supplemented with 10% fetal bovine serum (FBS) and 1% penicillin/streptomycin. PMN purity after density gradient centrifugation was determined using flow cytometry (Supporting Information Table S1). The final cell suspension was typically >95% PMNs.

Murine Bone Marrow Cell Preparation

BM cells were harvested from mouse femurs and tibias. Briefly, after separating the bones, the BM cells were flushed using RPMI onto a 50 ml conical tube through a 70 μm cell strainer. They were then centrifuged at 450 g for 5 min at RT and resuspended in RPMI supplemented with 10% FBS and 1% penicillin/streptomycin.

Murine PMN and Platelet Preparation

BM cells were harvested from mouse femurs and tibias as described above. Murine PMNs were purified from BM cells using Histopaque‐1077 (Sigma‐Aldrich, MO, US)/Histopaque‐1119 (Sigma‐Aldrich, MO, US)‐based density gradient centrifugation. They were then centrifuged at 400g for 30 min at RT and resuspended in RPMI supplemented with 10% FBS and 1% penicillin/streptomycin. For platelet (PLT) isolation, peripheral blood was collected via cardiac puncture. Platelet rich plasma was collected after centrifugation (150g for 10 min at RT without brake) and washed in Tyrode's buffer containing prostaglandin E1 (1 μM final concentration) (Cayman Chemical, MI, US). The suspension was centrifuged for 150g, 10 min at RT without brake to pellet contaminating erythrocytes and leukocytes. The platelet pellet was obtained after centrifugation for 800g, 10 min at RT without brake and resuspended in RPMI supplemented with 10% FBS and 1% penicillin/streptomycin.

Measurement of NETosis by Flow Cytometry

For all the following flow cytometry procedures, we maintained a standard operating procedure (SOP) (Supporting Information Table S1). Cells were seeded at a density of 1 × 10⁵/100 μl in U‐bottom 96‐well polypropylene plates and allowed to adhere for 30 min. Totally 100 μl of PMA (Sigma‐Aldrich, MO, US) was then added at the specified concentrations to induce NETosis, or they were left unstimulated. For some experiments, PMNs were incubated with PLTs at various PMN:PLT ratios to induce NETosis. Cells were incubated for up to 4 h at 37°C in 5% CO₂ and fixed by adding 100 μl of 4% paraformaldehyde (PFA; EMS, PA, US) for 15 min at RT. The final concentration of PFA was 1.3%. PFA at concentrations of 1%–4% fix but do not permeate the plasma membrane 21. Following fixation, samples were centrifuged at 100g for 10 min at RT. The supernatant was carefully aspirated and 100 μl of a staining master mix was added for 15 min at RT in the dark before analysis by flow cytometry. The master mix consisted of CD66‐FITC (BD Pharmingen, NJ, US) or Ly6G‐PerCP‐Cy5.5 (BD Pharmingen, NJ, US), 0.1 μM SO (Thermo Fisher Scientific, MA, US), and 0.3 nM DAPI (Thermo Fisher Scientific, MA, US) in phosphate buffered saline (PBS; Gibco, MA, US) (Supporting Information Fig. S1A). Final concentrations for CD66‐FITC and Ly6G‐PerCP‐Cy5.5 were 1 μg/ml and 2 μg/ml, respectively. Antibodies and DNA dyes were carefully titrated to negate the need for washing steps since they can disrupt the fragile NET structures 8.

Multiplex Adaptations to the NETosis Protocol

In some cases, additional fluorescent antibodies and often a more complex flow cytometry panel are essential for data analysis. In these cases, we observed increased background fluorescence and therefore removal of unbound antibodies by washing was necessary. Hence, the above protocol was adapted to include washing steps. Following fixation, samples were washed in PBS and centrifuged at 100g for 10 min at 4°C without the centrifuge break on. The antibody master mix consisted of CD66b‐PerCP‐Cy5.5 (BioLegend, CA, US) and NE‐Alexa Fluor 488 (Novus Biologicals, CO, US). 100 μl of antibody staining cocktail was added for 30 min at 4°C in the dark. Final concentrations for CD66b‐PerCP‐Cy5.5 and NE‐Alexa Fluor 488 were 1 μg/ml and 13 μg/ml, respectively. After washing in PBS twice, 100 μl of DNA dyes were added as the final step prior to flow cytometric analysis. As SO and DAPI emit fluorescence after binding to DNA, the washing step to remove unbound dye can be omitted 20, 31.

Flow Cytometry Acquisition and Analysis

All flow cytometry experiments were performed on a BD FACSCanto II (BD Biosciences, NJ, US) using BD FACSDiva, version 6.1.3 (BD Biosciences, NJ, US) software and analyzed using FlowJo, version 10.4.2 (BD, NJ, US). For the flow cytometry experiments performed on the BD FACSCanto II, the filters for DAPI, FITC and SO were 450/40, 530/30, and 585/42, respectively.

Measurement of NETosis by Immunofluorescence Confocal Microscopy

NETosis was verified using an established fluorescent microscopy protocol with the following modifications 32. Neutrophils were resuspended at 4 × 10⁶ cells/ml RPMI and 250 μl was added to each well of a 12‐well plate containing a 15 mm coverslip. An equal volume of PMA was added for 4 h, 37°C, 5% CO₂. Following stimulation, PMNs were fixed using an equal volume of 4% PFA, 15 min, RT. The coverslip was then removed using forceps and placed carefully onto a drop of CD66‐FITC for 30 min at 4°C in the dark. The cover slip was washed on PBS drops for 5 min twice before staining with DNA dyes for 15 min, RT. The cover slip was then prepared for microscopy by mounting onto a microscope slide with Fluorescence Mounting Medium (Dako, CA, US). Images were acquired using 10× air objective on Olympus FV1000 or 40× oil objective on Zeiss LSM800 confocal microscopes. For the microscopy experiments performed on the Zeiss LSM800 confocal microscope, the filters for DAPI/ Hoechst 33342, FITC, and SO were SP 470, SP 545, and SP 620, respectively.

Measurement of NETosis by Time‐Lapse Live‐Cell Immunofluorescence Microscopy

Purified PMNs were stimulated with 30 nM PMA and immediately stained with 1 μg/ml CD66‐FITC, 0.2 μg/ml Hoechst 33342 (Thermo Fisher Scientific, MA, US) and 0.1 μM SO. They were analyzed until 4 h poststimulation by confocal live‐cell imaging on the Zeiss LSM 710 confocal microscope using a 20× objective. Images were captured every 15 min and analyzed for quantification of NETosis using ImageJ (NIH, MD, US). Calculations were based on >700 PMNs in five different areas. NETosis rate was calculated according to previous descriptions as follows 9:

$$\text{NETosis rate}=100\times \mathit{\text{Objects counted}}\left(\mathit{SO}\mathit{\text{channel}}\right)/$$

        Objected counted (Hoechst channel).

Flow Cytometry Assessment of Apoptosis

Apoptosis was induced using etoposide (Sigma‐Aldrich, MO, US). PMNs were seeded at a density of 1 × 10⁵/100 μl in U‐bottom 96‐well polypropylene plates and allowed to adhere for 30 min before being exposed to etoposide (30 μM for 4 h). The PMNs were made to react with annexin V‐Alexa Fluor 488 and propidium iodide (Thermo Fisher Scientific, MA, US) according to the manufacturer's instruction.

Statistical Analysis

Data were analyzed using Prism 7.03 (GraphPad Software, CA, US). Normal distribution was assessed using the Shapiro–Wilk test. Gaussian data was analyzed using a Student's t‐test for two comparisons and one‐way ANOVA with post hoc for three or more comparisons and Pearson's correlation coefficient. For time and dose response analyses, a two‐way ANOVA was performed with Dunnett's multiple comparison test. Nonparametric data were assessed using Mann–Whitney test for two comparisons and Spearman's rank correlation coefficient. Statistical significance was defined as a two‐tailed P value of <0.05.

Results

Choice and Titration of Antibodies and Dyes for NETosis Assay Using Flow Cytometry

To optimize a flow cytometry‐based assay for NETosis, we began by assessing a series of fluorescently labeled antibodies to distinguish between untreated and PMA‐treated human PMNs. We discriminated human PMNs from other possible leukocytes using a combination of antibodies to several cell‐surface markers, including CD14, CD16, CD115, CD193, and Siglec‐10 (Supporting Information Fig. S1A and data not shown). CD66 antibody was subsequently selected, as it labels both inactivated and activated PMNs and can be used for both flow cytometry and microscopy 33 (Fig. 1B). CD66 monoclonal antibodies react only with granulocytes and the surface expression of CD66 is upregulated from intracellular stores following stimulation with a wide variety of agents 34. Ly6G was selected to define murine BM‐derived PMNs 4.

PMA is a consistent inducer of NETosis in vitro triggering the release of NET‐DNA into the extracellular space; a mechanism that requires PMN lysis 35, 36. Our strategy was therefore to choose plasma membrane‐impermeable DNA dye for the detection of NET‐DNA to avoid unspecific labelling of DNA in other forms of cell death. We tested antibodies and DNA dyes previously used in microscopy to label NETs 32, 37 (Supporting Information Fig. S1B,C). SO and DAPI were selected as SO does not cross the plasma membrane of live cells and DAPI is membrane‐semi‐permeable penetrating plasma membranes less efficiently in live cells, making them less susceptible to DAPI‐staining 20, 38, 39, 40. It is thought that PFA at concentrations of 1%–4% can fix but not permeate the plasma membrane to induce artificial “NET formation” 21. On the contrary, we found that the percentage of SO⁺DAPI⁺ PMNs, depending on the concentrations and exposure duration of PFA, varied on flow cytometry (Supporting Information Fig. S1D,E). For our assays we used a final concentration of 1.3% PFA for 15 min for the detection for extracellular NET‐DNA by SO and DAPI. This combination of optimally diluted antibody and DNA dyes identified NETting PMNs with minimal background in fluorescence by confocal microscopy (Fig. 1C).

Gating Strategy to Determine the Percentage of PMNs Undergoing NETosis

Purified PMNs were processed as described above and stimulated with or without PMA. They were then stained with anti‐CD66, DAPI and SO, and NETosis were analyzed with flow cytometry using the gating strategy in shown in Figure 2. We initially gated on FSC‐A/FSC‐H and SSC‐A/SSC‐H to exclude cell doublets as described in an earlier report 41, 42 and observed minimal doublet formation with flow cytometry (Fig. 2A). The cells were then gated on FSC/SSC [region 1 (R1)] to exclude any lymphocyte and monocyte contamination (Fig. 2B). PMNs were then identified based on their high granularity (SSC^hi) and surface expression of CD66 (Fig. 2C). Flow cytometry analysis of SSC^hiCD66⁺ PMNs revealed an increase percentage of SO⁺DAPI⁺ PMNs following PMA treatment (Fig. 2C, D). The background reading for percentage of SO⁺DAPI⁺ PMNs in untreated samples ranged from 7% to 20%. We observed a decrease in SSC profile and increase in surface expression of CD66 in PMNs following PMA treatment, consistent with the earlier reports 34, 42.

Figure 2.

Gating strategy for quantifying NETosis using flow cytometry. (A) Single cells were gated using FSC‐A/FSC‐H and SSC‐A/SSC‐H. (B) Contaminating lymphocytes and monocytes were excluded from analysis by applying R1. (C) The PMN population was identified by SSC^hiCD66⁺. (D) NET detection by SO and DAPI positivity in untreated and PMA‐treated (100 nM) samples.

Sensitive and Specific Determination of NETs by Flow Cytometry‐Based Assay

To assess the sensitivity of our assay, we examined NETosis with lower concentrations of PMA for shorter periods of time. Purified PMNs from five different HDs were stimulated with the indicated concentrations of PMA and analyzed hourly for up to 4 h (Fig. 3A). Using our assay, we observed a significant level of NETosis compared to untreated samples as early as 1 h following stimulation with the maximum concentration of PMA (30 nM) (Fig. 3A). Lower concentrations (1‐30 nM) of PMA reached a significant level of NETosis at 2 h and all concentrations of PMA resulted in significant levels of NETosis at 4 h (Fig. 3A). A comparison with a previously published assay by Gavillet and colleagues revealed a greater sensitivity in the detection of NETosis with our protocol (Fig. 3B and Supporting Information Table S2) 22. To confirm that SO⁺DAPI⁺ PMNs were recognizing NETosis, we used etoposide to stimulate apoptosis. While annexin V/PI staining showed an increase in early apoptosis at 4 h with 30 μM etoposide (Supporting Information Fig. S1F), there was no increase in the frequency of SO⁺DAPI⁺ PMNs (Fig. 3C). These findings clearly indicate that SO and DAPI do not recognize early apoptotic PMNs. Previous work has shown that NE binds to DNA and specifically localizes to NETs in a DNA‐dependent manner 43. Our immunofluorescent staining confirmed colocalization of NE with DNA fibers extruded from NETting PMNs (Fig. 3E). Furthermore, flow cytometry demonstrated that SO⁺DAPI⁺ cells were SO⁺NE⁺ in the PMA‐treated PMNs (Fig. 3F). Last, FSC^loSSC^loLy6G⁻ dead cells were intermediate SO⁺, NETs were SO⁺DAPI⁺ but dead cells were SO⁺DAPI⁻ (Supporting Information Fig. S2A‐C). This may be due to the low concentration of DAPI optimized for our flow cytometry assay detecting only extracellular NETs and not dead cells. Counterstaining dead cells requires 2 μM DAPI for flow cytometry, compared to 0.3 nM DAPI used in this assay 44. Hence, two‐color determination using SO and DAPI DNA dyes to detect NET‐DNA is required.

Figure 3.

Flow cytometry provides a sensitive and specific assay for detecting NETosis. Purified PMNs were stimulated with different concentrations of PMA as indicated. (A) NETosis was detected as early as 1 h following the high concentration of PMA shown and after 3 h with all concentrations used. (*P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001; n = 5). (B) PMNs from 3 HDs were untreated or stimulated with PMA (100 nM) for 4 h. The current assay demonstrates improved sensitivity for the detection of NETs compared to a previously published flow cytometry‐based assay 22. (C) Stimulation with 30 μM etoposide did not result in an increase in SO⁺DAPI⁺ PMNs. (*ns = not significant and ****P < 0.0001; n = 5). (D) Isolated BM PMNs and PLTs from B6 mice for incubated at various PMN:PLT ratios (1:0, 1:1, and 1:10) for 3 h. Formation of SSC^hiLy6G⁺CD41⁺ PMN‐PLT complexes and NETs were quantified by flow cytometry (*ns = not significant, **P < 0.01, **P < 0.001, and ****P < 0.0001; n = 3). (E) Colocalization of NET‐DNA and NE in PMA‐treated PMNs (100 nM for 4 h). Images were acquired using 40× oil objective on a Zeiss LSM800 confocal microscope. (F) Multicolor staining with SO, DAPI, NE‐Alexa Fluor 488 and CD66b‐PerCP‐Cy5.5 was conducted in purified PMNs from HDs, with and without PMA treatment (30 nM for 4 h), before subjecting to flow cytometry (*ns = not significant; n = 3).

Similar results were reproduced by an independent scientist (H.Y. Lee) following the SOP described in Supporting Information Table S1 (Supporting Information Fig. S3A,B).

Comparison of Flow Cytometry‐Based Assay with Immunofluorescence Microscopy and Time‐Lapse Live‐Cell Immunofluorescence Microscopy

Having demonstrated the sensitivity and specificity of the assay for NETosis, we went on to directly compare NETosis measured by microscopy with flow cytometry in the same HDs. We stimulated purified PMN with 100 nM PMA and measured NETosis every hour using both methods (Fig. 4A‐B). Similar to the increase in the double SO⁺DAPI⁺ population observed in flow cytometry following PMA exposure (Fig. 4A), NET structures were clearly visualized through colocalization of SO and DAPI (i.e., purple) at 2, 3, and 4 h poststimulation (Fig. 4B).

Figure 4.

Validation of the flow cytometry assay using microscopy techniques. PMNs were stimulated with PMA and prepared for quantification of NETosis as described. (A) NET formation was measured using flow cytometry. Graphs show the percentage of SO⁺DAPI⁺ PMNs following stimulation with 100 nM PMA every hour from five HDs. (**P < 0.01 and ****P < 0.001; n = 5). (B) In parallel, immunofluorescence microscopy was performed to visualize NETs in the same HDs. Representative imaging of UT and PMA‐treated PMNs from one of five HDs, acquired every hour. NET‐DNA was labeled with SO and DAPI. Images were acquired using 10× objective on Olympus FV1000 confocal microscope. (C) PMNs from 1 of 3 HDs were stimulated with 30 nM PMA for 0, 1, 2, 3, and 4 h and monitored using time‐lapse live‐cell immunofluorescence microscopy. Images were acquired using 20× objective on a Zeiss LSM800 confocal microscope. (D) Time course of PMA‐induced (30 nM) NETosis of PMNs from one HD, measured in triplicate, using flow cytometry. (E) Correlative assessment of levels of NETosis between time‐lapse live‐cell immunofluorescence microscopy and flow cytometry following PMA (30 nM) exposure at the indicated times. NETosis was measured in PMNs from 3 HDs on the same day (r = 0.90, ****P < 0.001; n = 3).

Hoechst 33342 is a plasma membrane‐permeable DNA dye and has been used extensively to stain live cells and therefore may be used as an internal control to quantify total PMN‐DNA (DNA in intact PMNs and NETs) and create a ratio of NET‐DNA/total PMN‐DNA (i.e., NETosis rate) 35, 45. We therefore, used time‐lapse live‐cell immunofluorescence microscopy to determine the NETosis rate following PMA stimulation at the indicated times (Fig. 4C). Similarly, as early as 2 h following PMA stimulation, colocalization of SO with Hoechst 33342 was identified (purple; Fig. 4C and Supporting Information Video S1). These studies were done in parallel with flow cytometry using SO⁺DAPI⁺ to identify NETs (Fig. 4D). The NETosis rate obtained using time‐lapse live‐cell immunofluorescence microscopy exhibited a positive correlation with the percentage of NETting PMNs on flow cytometry (Fig. 4E; r = 0.90, ****P < 0.001).

Additionally, we examined the intraassay variance (PMNs from the same HD in triplicate run on the same day at the same time) and inter‐assay variance (PMNs from same HD on three different days) of the flow cytometry assay. The intraassay and interassay variances for the percentage of SO⁺DAPI⁺ PMNs 4 h poststimulation with 30 nM PMA were 3.98% and 4.66%, respectively.

Detection and Quantification of NETosis within Mixed Cell Populations in SLE Patients and HDs Using Flow Cytometry

All current methods currently used in assessing NETosis involve neutrophil isolation by density centrifugation methods, which may provoke cellular depletion and activation 46. Therefore, an obvious advantage of a new NETosis assay would be the ability to measure function in unseparated cells. We therefore went on to examine NETosis in lysed WB and isolated PMNs in the same donors. We observed a similar frequency of SO⁺DAPI⁺ NET‐appendant PMNs in untreated and PMA‐stimulated PMNs across between lysed WB and purified PMN preparations (Fig. 5A). Furthermore, statistical analyses confirmed a positive correlation in NETosis between the two PMN preparations (Fig. 5B; r = 0.79, ****P < 0.001). In summary, our flow cytometry assay permits the study of NETosis in mixed cell populations such as lysed WB, reducing sample manipulation.

Figure 5.

Detection of NETs in mixed cell populations. Flow cytometry assays were performed in each subject with or without isolation of PMNs. (A) Cumulative data for percentage of SO⁺DAPI⁺ PMNs in the total PMN population (either purified PMNs or WB PMNs) from HDs, untreated and stimulated with 30 nM PMA for 4 h (n = 19). (B) Correlative assessment of the percentage of NETting PMNs between purified PMNs and WB PMNs in SLE patients and HDs using flow cytometry.

Comparison of NETosis in PMNs from SLE Patients and Lupus‐Prone Mice with Controls

A previous study reported no differences in NET formation between normal‐density PMNs from SLE patients and HDs at baseline or following 2 h stimulation with PMA 47. A comparison of PMA‐induced NETosis by flow cytometry in purified PMNs and WB PMNs from SLE patients with HDs did not reveal any significant difference as a result of disease (Fig. 6A,B). However, most of the recruited SLE patients examined were on immunosuppressive therapy with quiescent disease activity, which may have consequences for PMN activation (Supporting Information Table S3). Therefore, we went on to examine NET formation in BM PMNs from young female lupus‐prone mice. The NZM2410‐derived Sle1 lupus susceptibility region confers a loss of tolerance to nuclear material, resulting in anti‐nuclear antibody titers, increased T and B‐cell activation and mild splenomegaly 28, 29. The introgression of the additional NZM2410‐derived Sle2 and Sle3 regions (Sle123), or a moderate 2‐fold increase in TLR7 (Tg7) results in severe lupus nephritis 48. Stimulation with PMA resulted in an increase in the frequency of SO⁺DAPI⁺, SSC^hiLy6G⁺ BM PMNs, which was augmented in Sle1 mice, compared to B6 controls (Fig. 6C,D). An analysis of BM PMNs across all lupus‐prone mice (Sle1, Sle1Tg7, and Sle123) showed an increase in NETosis in severe‐lupus prone Sle123 mice compared to the milder lupus‐prone Sle1 strain (Fig. 6E). Last, the formation of PMN and PLT complexes induced NET formation when BM PMNs and PLTs from B6 mice for incubated at various PMN:PLT ratios (1:0, 1:1, and 1:10) for 3 h (Fig. 3D) 49.

Figure 6.

NETosis is increased in murine PMNs from lupus‐prone mice. Quantification of NETosis of (A) purified PMNs and (B) WB PMNs from SLE patients (n = 13) and HDs (n = 19), untreated or exposed to 30 nM PMA for 4 h. The percentage of SO⁺DAPI⁺ PMNs in the total CD66⁺ PMN population is shown. A single cell suspension of BM cells from 6 to 8 week old nonautoimmune prone B6 mice (n = 8) or lupus‐prone mice [Sle1 (n = 7), Sle1Tg7 (n = 4) and Sle123 (n = 4)] were stimulated with or without 30 nM PMA for 4 h. NETosis was assessed by flow cytometry. The percentage of SO⁺DAPI⁺ PMNs in the total in the total Ly6G⁺ PMN population is shown. (C) Flow cytometry gating strategy for measuring NETosis in BM PMNs following stimulation with 30 nM PMA for 4 h, with untreated 4 h as an overlay. BM PMNs are SSC^hiLy6G⁺. (D) A comparison of NETosis in B6 and the mild autoimmune prone strain, Sle1. (F) Sle123 mice which are susceptible to severe autoimmune pathology show increased NETosis compared to Sle1 mice which have mild autoimmunity.

Discussion

In this article, we detailed and validated a novel flow cytometry assay to detect NETs in mixed cell populations using a combination of fluorescently labeled monoclonal antibody against PMN cell‐surface markers with DNA dyes 50. NETs are very fragile even after fixation and have to be handled gently, otherwise the majority will get lost during preparation 32. In addition, a common problem with the generation of novel assays is reproducibility by another operator 51, 52. We generated a SOP and an independent scientist not involved in the original design of the assay reproduced the findings. Therefore, this flow cytometry assay is simple, reproducible and obviates losses of NETs encountered in the standard methods of analysis.

We validated the assay by showing: (1) minimal SO⁺DAPI⁺ PMNs with etoposide‐induced apoptosis; (2) a positive correlation in the percentage of SO⁺DAPI⁺ PMNs with NETosis rate obtained using time‐lapse live cell microscopy; and (3) colocalization of extracellular DNA and NE on flow cytometry. Additionally, we demonstrated that this flow cytometry‐based assay could be used for both human and murine samples, in purified PMNs or in mixed cell populations extracted from WB or BM. Moreover, this is the first assay, which does not require purification, which may activate the PMNs 22, 24, 25, 26.

Over the last few years, a number of flow cytometry‐based assays to determine NETs have been proposed 22, 24, 25. Gavillet et al. described an assay using antibodies against citrullinated H3 and MPO, in combination with DAPI 22. Flow cytometry assays are not expected to detect NETosis from every kind of stimulation since previous data has shown that PAD4‐mediated hypercitrullination of histones depends on the nature of the stimulation 21. A side by side comparison with the assay by Gavillet et al. revealed greater sensitivity in the detection of NETosis with our protocol (Fig. 3B and Supporting Information Table S2) 22. Furthermore, Nel et al. identified extracellular NETs as CD16⁻Vybrant DyeCycle (VDC) Ruby⁺MPO⁺ or CD16⁻VDC Ruby⁺citrullniated H4⁺ 24. VDC Ruby is a plasma membrane‐permeable DNA dye 53. However, the assay by Nel et al. can underestimate the earlier phases of NETosis when the NETs are still PMN‐appendant. Additionally, Masuda et al. utilized Sytox Green (SG), a plasma membrane‐impermeable DNA dye, to detect PMN‐appendant NET‐DNA 21, 25. Similar to our findings, percentage of SG⁺ cells correlated with NET area stained using DAPI and observed on fluorescent microscopy 25. However, SG⁺ cells were gated solely based on FSC and SSC properties, which limit the assay to purified populations and may include debris 54. To discriminate NETting PMNs from other contaminating leukocytes, a combination of antibodies to other cell‐surface lineage markers may be introduced to the flow cytometry panel using our multiplex adaptation described above. As we gated on SSC^hiCD66⁺ or SSC^hiLy6G⁺ for human and murine PMNs, respectively, and CD66 and Ly6G are well accepted markers for PMNs; this gate should not have much contamination from other leukocytes. However, we believe a limitation of our assay is underestimating the late phases of NETosis by identifying NETting PMNs as SSC^hiCD66⁺ or SSC^hiLy6G⁺, because completely lysed PMNs will fall outside this gate. Hence, we agree that gating based on FSC and SSC is inadequate but cells positive for MPO/NE‐DNA complex should be considered as PMNs that have formed NETs 21, 25, 54. Late apoptotic or necrotic cells induced by etoposide requires incubation for 24 h 55. However, our assay is optimized to avoid spontaneous cell death in neutrophils that may occur with lengthy incubation periods. As such, it is possible that this SO and DAPI will eventually stain late apoptotic and necrotic cells at later time points but that is not the intended use of this assay, where the incubation time was limited to 4 h. The background reading for percentage of SO⁺DAPI⁺ PMNs in untreated samples ranged from 7% to 20%. This could be due to some loss of integrity as a result of PFA fixation, given the findings we report in Supporting Information Figure S1D comparing PFA concentrations.

Recent evidence suggests that aberrant NETosis and/or impaired NET clearance play a key role in the pathogenesis of SLE 14, 47, 56. Ribonucleoprotein immune complexes, prevalent in lupus, commit PMNs to NETosis through TLR7/8‐mediated shedding of FcγRIIA 57. Furthermore, inhibition of the enzyme, peptidylarginine deiminase 4 (PAD4) blocked NET formation and reduced proteinuria and immune complex deposition in lupus‐prone MRL/lpr mice 58. PAD4 catalyzes the conversion of arginine residues to citrulline in three of the four histones, H2A, H3 and H4 10, 59. PAD4 is the only PAD that has a nuclear localization signal and is most abundantly expressed in granulocytes, monocytes and mast cells 59. In PAD4‐null mice, hypercitrullination of H3 is not detectable and the strain fails to make NETs 3. In apparent contrast to these observations, MRL/lpr mice that were deficient in Nox2, and therefore lacked functional NADPH oxidase, developed an acceleration of the lupus phenotype 12. However, NADPH oxidase is not the only source of ROS. Mitochondrial ROS production may also generate NETs enriched in oxidized mtDNA, NE, and MPO 12, 60. More recently, it has been shown that in addition to PMNs, lymphocytes (B cells, T cells, and NK cells) and monocytes, can release ET‐mtDNA 15, 16. ET‐mtDNA released by B cells was stained with SG and visualized using fluorescence microscopy 16. None of the proteins associated with NETs could be found on B cell ET‐mtDNA using mass spectrometry 16. Hence, although ETs classically contain citrullinated histones, MPO and NE; their specific protein composition continues to be defined and varies depending on the source of ETs and disease states 17. As our assay has been optimized for PMN‐appendant NET‐DNA, it could potentially be adapted to detect and quantify cell‐appendant ET‐mtDNA from other immune cells.

In conclusion, the assay described here can be easily applied to the flow cytometry study of NETosis for both purified PMNs and mixed cell populations. Analysis of NETosis in the newly described subpopulations of PMNs in WB or BM may open new avenues for researchers in the field 4. Last, high throughput screening for pharmacological inhibitors of aberrant NETosis (e.g., PAD4) may be carried out using our assay as a therapeutic strategy in SLE.

Authorship Contributions

O.Z. and S.H.T. designed, performed experiments and analyzed the data; H.Y.L. performed independent experiments; T.S. and W.Y.O. performed the confocal and time‐lapse live‐cell immunofluorescence microscopy; A.L., P.A.M., L.H.K.L. and J.E.C. provided material and participated in discussions; S.H.T. and A.‐M.F. drafted the manuscript; and A.‐M.F. oversaw the project.

Disclosure of Conflicts of Interest

The authors declare no competing financial interests.

Supporting information

Supplemental Figure S1. Optimization of flow cytometry‐based assay.

(A) List of various antibodies and nucleic acid dyes used for detection of NETs. Markers selected for use in this manuscript are denoted with “+”. (B‐C) Detection of NETs production by untreated (left panel) or treated (right panel) PMNs using (B) anti‐Histone 1 (1:300 μl) and anti‐NE (1:400 μl) antibodies or (C) 2.5 μM Calcein AM and 0.3 nM DAPI following stimulation for 4 h with 100 nM PMA. The gating strategy were similar to the one described for SO and DAPI. (D‐E) PMNs were harvested from BM of Sle1 mice (n = 3). (D) Untreated PMNs were fixed with 2 different concentrations of PFA for 15 min and analyzed at 0 h. (**P < 0.01; n = 3). (E) Untreated PMNs were fixed with 1.3% PFA for various times indicated and analyzed. (F) PMNs were exposed to 30 μM etoposide for 4 h. Apoptosis was assessed using Alexa Fluor 488‐conjugated annexin V and propidium iodide (PI) (Thermo Fisher Scientific) according to the manufacturer's instruction and then followed by flow cytometry.

Supplemental Figure S2. DAPI as a useful stain for NET quantification.

(A) Flow cytometry gating strategy for measuring NETosis in BM cells following stimulation with 30 nM PMA for 3 h. NETting SSC^hiLy6G⁺ PMNs were SO⁺DAPI⁺ while FSC^loSSC^loLy6G⁻ dead cells and were mainly SO⁺. (B) Overlay of histogram from BM PMNs and dead cells revealed that NETs were SO^hi whereas dead cells were intermediate SO⁺. (C) BM PMNs were harvested from B6 mice (n = 3) and stimulated with 30 nM PMA for 3 h. NETs were SO⁺DAPI⁺ but dead cells were SO⁺DAPI⁻.

Supplemental Figure S3. Independent assessment of NETosis flow cytometry assay.

(A) Identification of NETs in untreated or PMA‐treated (30 nM for 4 h) using the gating strategy shown in Figure 2. SSC^hi and CD66 were used to identify PMNs and the percentage of SO⁺DAPI⁺ PMNs was then determined. (B) NET formation following PMN stimulation with different concentrations of PMA or 30 μM etoposide which stimulated early apoptosis.

Supplemental Video S1. Measurement of NETosis by time‐lapse live‐cell immunofluorescence microscopy.

Purified PMNs were stimulated with 30 nM PMA, monitored with time‐lapse live‐cell immunofluorescence microscopy and stained using CD66‐FITC, Hoechst 33342 and SO. All cells were viable initially, evident by the absence of SO staining. After about 1 h of stimulation, some CD66⁺ PMNs were positive for SO and Hoechst 33342 (i.e. purple) and the nucleic acid dyes stain the entire cell. In the next few minutes, the plasma membrane ruptures and the purple staining around the PMN indicates release of NETs into the extracellular space. CD66‐FITC signal increases following PMA treatment and remains throughout the duration of the experiment, indicating that NET‐DNA do not diffuse after plasma membrane rupture but form a stable structure of PMN‐appendant NET DNA.

Supplemental Table S1. SOP for NETs identification using flow cytometry.

Appendix S1: Cytometry Part A