Immunofluorescence Imaging of Neutrophil Extracellular Traps in Human and Mouse Tissues

Summary

Neutrophil extracellular traps (NETs) are associated with various diseases, and immunofluorescence is often used for their visualization. However, there are various staining protocols, and, in many cases, only one type of tissue is examined. Here, we establish a generally applicable protocol for staining NETs in mouse and human tissue.

Abstract

Neutrophil extracellular traps (NETs) are released by neutrophils as a response to bacterial infection or traumatic tissue damage but also play a role in autoimmune diseases and sterile inflammation. They are web-like structures composed of double-stranded DNA filaments, histones, and antimicrobial proteins. Once released, NETs can trap and kill extracellular pathogens in blood and tissue. Furthermore, NETs participate in homeostatic regulation by stimulating platelet adhesion and coagulation. However, the dysregulated production of NETs has also been associated with various diseases, including sepsis or autoimmune disorders, which makes them a promising target for therapeutic intervention. Apart from electron microscopy, visualizing NETs using immunofluorescence imaging is currently one of the only known methods to demonstrate NET interactions in tissue. Therefore, various staining methods to visualize NETs have been utilized. In the literature, different staining protocols are described, and we identified four key components showing high variability between protocols: (1) the types of antibodies used, (2) the usage of autofluorescence-reducing agents, (3) antigen retrieval methods, and (4) permeabilization. Therefore, in vitro immunofluorescence staining protocols were systemically adapted and improved in this work to make them applicable for different species (mouse, human) and tissues (skin, intestine, lung, liver, heart, spinal disc). After fixation and paraffin-embedding, 3 µm thick sections were mounted onto slides. These samples were stained with primary antibodies for myeloperoxidase (MPO), citrullinated histone H3 (H3cit), and neutrophil elastase (NE) according to a modified staining protocol. The slides were stained with secondary antibodies and examined using a widefield fluorescence microscope. The results were analyzed according to an evaluation sheet, and differences were recorded semi-quantitatively.

Here, we present an optimized NET staining protocol suitable for different tissues. We used a novel primary antibody to stain for H3cit and reduced non-specific staining with an autofluorescence-reducing agent. Furthermore, we demonstrated that NET staining requires a constant high temperature and careful handling of samples.

Introduction

Neutrophil extracellular traps (NETs) were first visualized by Brinkmann et al. as a pathway of cellular death different from apoptosis and necrosis in 2004¹. In this pathway, neutrophils release their decondensed chromatin into the extracellular space to form large web-like structures covered in antimicrobial proteins that were formerly stored in the granules or cytosol. These antimicrobial proteins include neutrophil elastase (NE), myeloperoxidase (MPO), and citrullinated histone H3 (H3cit), which are commonly used for indirect immunofluorescence detection of NETs². This method not only identifies the quantitative presence of these proteins; indeed, it has the advantage of specifically detecting NET-like structures. In the NETs, the mentioned proteins co-localize with extracellular DNA, which can be detected by an overlap of the fluorescence signals of each stained protein and the extracellular DNA. In contrast to the overlapping signals due to extracellular DNA and protein co-localization in NETs, intact neutrophils show no co-localization. Here, the NET components are usually stored separately in the granules, nuclei, and cytosol³.

Since their first discovery, it has been shown that NETs play a central role in numerous diseases, particularly those involving inflammation. NETs show antimicrobial functions during infection through trapping and killing extracellular pathogens in blood and tissue^4,5. However, NETs have also been connected to autoimmune diseases and hyperinflammatory responses, like systemic lupus erythematosus, rheumatic arthritis, and allergic asthma^6,7,8. NETs promote vaso-occlusion and inflammation in atherosclerosis, platelet adhesion, and are speculated to play a role in metastatic cancer^9,10,11. Nevertheless, they are thought to have anti-inflammatory properties by reducing proinflammatory cytokine levels¹². While NETs are gaining more interest in a broader field of research, a robust NET detection method is fundamental for future research.

Even though the visualization of NETs in different tissue using immunofluorescence imaging is complex and requires customization, apart from electron microscopy, it is currently one of the most renowned methods for visualizing the interactions between NETs and cells and is predominantly used in formalin-fixed paraffin-embedded tissues (FFPE)^13,14. However, comparing NET imaging is difficult, as different laboratories use their own customized protocols. These protocols differ in their use of antibodies, antigen retrieval, or permeabilization method and are often optimized for a specific type of tissue^{3,13,15,16,17,18,19,20,21,22,23,24,25,26,27}.

After Brinkmann et al. published the first methodic study using immunofluorescent visualization of NETs in FFPE tissue, we wanted to optimize this protocol for a wider variety of tissues and species¹⁵. Additionally, to establish a broadly applicable immunofluorescence protocol, we tested different modified protocols from studies that used immunofluorescence methods in FFPE tissue to detect NETs^{3,13,16,17,18,19,20,21,22,23,24,25,26,27}. Furthermore, we tried a new H3cit antibody for more specific extracellular staining²⁸. We hypothesize that by systematically adapting current staining protocols to different species and tissue, in vitro imaging can be improved, resulting in a better representation of the interaction between neutrophils and NETs both locally and systemically.

Protocol

This study included mouse tissues derived from experiments approved by the Hamburg State Administration for Animal Research, Behörde für Justiz und Verbraucherschutz, Hamburg, Germany (73/17, 100/17, 94/16, 109/2018, 63/16). The tissues used were mouse lung and colon from a septic model and burned skin. We used 8 week old male and female mice. The European Directive 2010/63/EU on the protection of animals used for scientific purposes was followed for all the experiments. The anonymized human samples included tissues from neonatal enterocolitis, burned skin, biliary atresia, spondylodiscitis, and myocardium. According to the Medical Research Ethics Committee of Hamburg, the samples did not need informed consent, but the study was approved by the committee (WF-026/21).

1. Sample fixation

1. Use the following protocol derived from Abu Abed and Brinkmann for sample fixation, dehydration, paraffin embedding, sectioning, and mounting^3,15.
   1. Prepare 4% formaldehyde solution by dissolving 40 g of paraformaldehyde (PFA) in 800 mL of Tris-buffered saline, pH 7.4 (TBS).
   2. Stir the mixture at 60 °C under a fume hood until the PFA has dissolved. Bring the solution to room temperature (RT), and adjust the volume with TBS to 1,000 mL.
   3. Adjust the pH to 7.4. Store at 4 °C for 2-3 weeks or at −20 °C for up to 1 year.
2. For sample fixation, place fresh tissue in TBS buffer, and dissect into pieces smaller than 20 mm x 30 mm x 3 mm. Immerse the tissue sections in 4% paraformaldehyde solution for 12-24 h.
   ​NOTE: The original protocol used a 2% PFA solution and a fixation time of 8-20 h. With this method, some tissue sections were not completely fixated after 20 h, so the PFA concentration was increased to 4% PFA.
3. Transfer the samples into labeled tissue processing cassettes. Use solvent-resistant markers or pencils for labeling.
4. Start the sample dehydration by immersing then cassettes in 70% ethanol for 1 h. Then, immerse in 80% ethanol, 90% ethanol, 96% ethanol, and twice in 100% ethanol, with each step lasting 1 h.
5. Immerse twice in 100% xylene (dimethylbenzene) and then twice in 60 °C paraffin for 1 h each time.
6. Use embedding molds for mounting, and let the paraffin solidify before removing the mould.
7. Use a microtome for cutting 3 µm thick tissue sections.
8. Use tweezers to lay the cut sections in a 37 °C water bath, and let them float on the surface to stretch the tissue cuts.
9. Place the sections onto adhesive glass slides (Table of Materials), and let them dry overnight in a 40 °C heat chamber.

2. Sample rehydration

1. For deparaffinization, stack the slides in a slide rack, and submerge twice for 5 min each in the xylene replacement medium limonene (Table of Materials) and once for 5 min in a 1:1 limonene/ethanol mixture.
   CAUTION: Limonene is flammable at 50 °C and can cause allergic reactions. Use under a fume hood with eye protection and gloves. Keep away from open fire.
2. Rehydrate the samples in a descending ethanol series by submerging the slide rack twice in 100% ethanol and once in 96% ethanol, 90% ethanol, 80% ethanol, and 70% ethanol for 5 min each.
3. Submerge the slide rack for 5 min in deionized water (DI water) to clean off any remaining ethanol.

3. Autofluorescence blocking and antigen retrieval

1. Prepare pH 6 citrate target retrieval solution (TRS) (Table of Materials) according to the datasheet. This TRS is concentrated 10 times. Therefore, dilute 1:10 with DI water. Preheat in a plastic staining jar to 96 °C.
   NOTE: TRS breaks the methylene bridges formed during sample fixation to expose the antigen epitopes. This allows the primary antibodies to bind their target antigen. Store concentrated TRS at 2-8 °C for 8 months. Always prepare fresh TRS.
2. Take the slides out of the slide rack, and place them in a wet chamber. Pipette one to two drops of autofluorescence-reducing agent (Table of Materials) onto each sample. Incubate for 5 min.
   NOTE: The autofluorescence-reducing agent blocks the inherent tissue autofluorescence caused by endogenous fluorophores and fixatives. Store the autofluorescence-reducing agent at RT for up to 1 year.
   NOTE: Work only with small sections of tissues to avoid drying out of the samples or exceeding incubation time.
3. Stack the slides back into the slide rack, and rinse for 1 min in 60% ethanol using an upward and downward motion until all the unused autofluorescence-reducing reagent has been washed off.
4. Submerge the slide rack for 5 min in DI water.
5. For antigen retrieval, transfer the slides into a staining jar with the preheated TRS. Incubate for 10 min at 96 °C in a water bath.
   NOTE: The 96 °C water bath can be substituted with a microwave or steam cooker if the TRS is preheated to 96 °C and 10 min of incubation time at 96 °C is ensured.
6. Take the staining jar out, and let it cool slowly to RT for approximately 60-90 min.
   NOTE: The protocol can be paused here for up to 4 h, leaving the slides in the TRS.
7. Remove the TRS, but keep the slides in the jar. Rinse twice with Tris-buffered saline with 0.05% Tween (TBST, pH 7.4) for 3 min each to wash off any leftover TRS residues.
8. For permeabilization, prepare 0.2% Triton in phosphate-buffered saline (PBS), pH 7.4, and fill it into the staining jar to permeabilize the samples for 10 min.
   NOTE: Permeabilization enhances the binding of the primary antibodies with the intracellular target proteins in the fixed samples.
9. Rinse the slides again twice in TBST for 3 min each time.
10. Prepare the wet chamber, and lay down slides. Wipe excess water from slides with paper tissues. Circle every sample with a hydrophobic barrier pen to keep the antibody solution from running off.
    NOTE: Do not let the samples dry out. It is best to work with small sections.

4. Blocking nonspecific antibody binding

1. Take ready-to-use blocking solution with donkey serum (Table of Materials), and pipette one drop onto every sample. Incubate for 30 min at RT.
   NOTE: This blocking step minimizes the binding of the secondary antibody to nonspecific binding sites on the sample. After blocking these nonspecific epitopes and applying the primary antibody, the secondary antibody will bind to the primary antibody and not interact with the surface of the tissue. This ready-to-use blocking reagent is stored at 2-8 °C for 6 months.
2. Remove excess blocking solution from slides by tapping the edge of the slides against a hard surface. Do not rinse them.

5. Primary antibody

1. Dilute primary antibodies in antibody dilution buffer (Table of Materials). Use approximately 100 µL of the final solution for every sample. For the NE and H3cit (R8) antibody and isocontrol, dilute to a concentration of 5 µg/mL. For MPO and corresponding isocontrol, use 10 µg/mL. Prepare one tube for every primary antibody and one for every isocontrol antibody.
   NOTE: H3cit (R8)/MPO or NE/MPO and their isocontrols can be combined for NETs staining. For the combination of H3cit (R8)/MPO, dilute to a final concentration of 5 µg/mL for H3cit (R8) and 10 µg/mL for MPO to a final volume of 100 µL per sample. For NE/MPO, dilute to a final concentration of 5 µg/mL for NE and 10 µg/mL for MPO to a final volume of 100 µL per sample. For isocontrol, use the same concentration as for each corresponding antibody. Always use a new pipette tip for each antibody used. Primary antibodies can be stored at 4 °C for 1-2 weeks and at −20 °C or −80 °C for up to 1 year.
2. With two samples on one slide, use one for the isocontrol antibodies and one for the MPO/H3cit or MPO/NE antibody dilution. Use approximately 100 µL for every sample, and spread the antibody solution evenly.
   NOTE: Do not let samples dry out. Be careful to avoid mixing up the isocontrol and primary antibodies.
3. Store in a wet chamber overnight at 4 °C.
4. The next day, tap off the excess primary antibody solution from the slides, and stack the slides in a cuvette. Rinse three times for 5 min each time with TBST to wash off the remaining primary antibody solution.

6. Secondary antibody

1. Prepare the secondary antibody solution. Use two different fluorescent secondary antibodies: donkey-anti-goat for MPO and donkey-anti-rabbit for H3cit staining. Dilute each one to a concentration of 7.5 µg/mL with antibody dilution buffer (Table of Materials).
   NOTE: For double-staining, use two fluorescent antibodies with different excitation areas and minimal spectral overlap. Combine both secondary antibodies, and dilute to a final concentration of 7.5 µg/mL for each antibody for a final volume of 100 µL per sample. Protect the antibodies from light. The secondary antibodies can be stored at 2-8 °C for 6-8 weeks or at −80 °C for up to 1 year.
2. Keep the slides in the wet chamber, and add 100 µL of the secondary antibody solution onto every sample. Let them incubate for 30 min at RT and protected from light.
3. Tap off the excess secondary antibody solution, and stack the slides in the slide rack. Submerge three times for 5 min each in a staining jar filled with PBS to wash off any remaining unbound antibodies.
4. Prepare DAPI (4',6-diamidino-2-phenylindol) solution (Table of Materials), and dilute to a concentration of 1 µg/mL with DI water. Submerge the slide rack in a staining jar with DAPI solution, and incubate for 5 min at RT in the dark.
   NOTE: DAPI is used as a fluorescent stain for double-stranded DNA. The prepared DAPI solution can be used multiple times. Store the prepared DAPI solution at RT. The DAPI concentrate can be stored at −20 °C for up to 1 year.
5. Submerge the slide rack for 5 min in a staining jar with PBS to wash off the excess DAPI solution.

7. Mounting and storing the samples, microscopic analysis

1. Mount the samples with coverslips and mounting medium (Table of Materials).
   NOTE: Use only small amounts of mounting medium, and wipe off the excess medium with wet tissues. Wear gloves, and avoid accidentally wiping off any medium from the coverslips or slides. Avoid bubble formation, and use cotton swabs to press gently on the coverslips to remove any bubbles from the slides.
2. For imaging, use a widefield microscope or a confocal microscope.
   NOTE: Take an image of the isocontrol first. Lower the exposure time for the fluorescence filters (except the blue filter for DAPI) until almost no signal can be detected. Then, use this setting to examine the corresponding samples. Look for areas where a fluorescent signal can be detected in every individual sample using the fluorescence channel. After taking an image of the area, the program generates a compound image of all the channels and shows the different fluorescence signals as an overlap of different colors. The image can then be further analyzed for co-localization with imaging software such as ImageJ.
3. Store the slides at 4 °C for up to 6 months.

Results

Before starting our protocol optimization, we identified key steps for successful staining by searching PubMed for studies that used FFPE tissue for the immunostaining of NETs and compared their protocols. The most promising protocol differences were identified as the key steps for the protocol optimization, while steps that mostly corresponded to each other were not changed (Table 1).

Table 1: PubMed Research for FFPE immunostaining of NETs. This table shows the variables in the immunostaining protocol in the examined studies. The protocols used were divided into their essential steps and then compared with one another. The steps with the most promising differences were then taken as the key steps for optimization and adapted for our protocol. Steps that mostly corresponded to each other were not changed, like the incubation time for the primary antibody (overnight, 4 °C). Please click here to download this Table.

Based on our findings, we concluded that the antibodies chosen to target the epitopes was the first critical step. At least 10 different protocols used the triH3cit antibody (see Table 1; Primary antibody column). Nonetheless, Thålin et al. found that the H3cit (R8) clone displayed lesser off-target cross-reactivity to non-citrullinated histones and showed negligible inter-lot variability. Thus, we decided to compare triH3cit and H3cit (R8) staining results with each other²⁸.

Four studies used the human/mouse MPO antibody. Furthermore, two other protocols applied MPO (2D4) for mouse tissue and MPO (2C7) for human tissue (see Table 1; Primary antibody column). Therefore, we separately compared MPO (2C7) with human/mouse MPO antibody for human tissues and MPO (2D4) with human/mouse MPO antibody for mouse tissues. NE was detected using at least five different antibodies, but only three of them showed good staining results in the images provided. However, one antibody was no longer available on the market, so we compared the NE antibody from a rabbit host with one from a mouse host for our test series in human tissues. For mouse samples, there seemed to be no available and reliable alternative to the NE antibody raised in rabbit host at the start of this study. Since one NE and both H3cit antibodies are derived from the same rabbit host, they cannot be combined for double staining with this protocol. The secondary antibody is specific to the constant region of the primary antibody, which is determined by the host it was raised in. If two primary antibodies derived from the same host are used, the secondary antibody could bind to both primary antibodies, and the staining would be unspecific. However, double staining is preferred over single staining because more NET components can be detected and co-localized. Therefore, the staining result will be more specific. Consequently, we double-stained for H3cit and MPO to obtain a more robust detection protocol.

Despite the similarities in the antibodies used, the dilutions for the antibodies varied in nearly all the protocols; for instance, for triH3cit, the concentration range was from 0.5 µg/mL to 20 µg/mL^17,18. For every antibody used, we tried different dilutions and obtained satisfactory staining results in the entire range reported in the literature.

Furthermore, many similarities could be found in the incubation time for the primary antibody (overnight at 4 °C) and the usage and dilution of the secondary antibody (see Table 1; Incubation primary antibody column). Therefore, we did not change these steps in our test series and performed them according to the protocol described above.

The next critical step determined from the literature was the antigen retrieval. This step is essential because, due to formalin fixation, the epitopes for the antibodies are masked through methylene bridges, which can be reversed by heating the tissue section in a suitable buffer²⁹. Citrate TRS at pH 6 and EDTA TRS buffer at pH 9 were used equally often in literature and gave similar results (see Table 1; TRS buffer column). Thus, we decided on the pH 6 citrate TRS for our test series. For antigen retrieval, we tested two different heating methods: a microwave (first 1 min at 360 W and then following with 9 min at 90 W) and a water bath (60 °C for 90 min, 96 °C for 10 min).

The last step that showed some variability in the literature was the permeabilization with Triton X-100. This step required optimization because with detergent treatment, the cell membrane becomes permeable to antibodies, and intracellular epitopes can be reached³⁰. The previous protocols used different Triton X-100 concentrations ranging from 1% to 0.1% (see Table 1; Permeabilization column). Therefore, we tried two Triton X-100 concentrations (0.2% and 0.5%) and one series with no Triton permeabilization.

After identifying these key steps, we modified them and tried to optimize the protocol. Then the images were examined according to an evaluation sheet, and the differences were recorded semi-quantitatively and compared (see Table 2).

Table 2: Table of results for optimizing the protocol steps. This table shows the results for the adapted steps: autofluorescence-reducing agent, antigen retrieval, and permeabilization. Before beginning this test series, we tested for the best antibody combination and concentration. The slides were evaluated in 10 different areas, and then one representative area was scored from (-) for a negative result to (++) for a positive NET-containing result. The partially positive results included a higher diffuse background staining of non-neutrophil cells. Abbreviations: n/u = not used; - = negative result; +/-partially positive staining; + = moderate specific staining; + = good staining of NETs and neutrophils. Please click here to download this Table.

Primary antibodies
Before adapting the protocol, we tried to find the best antibody combination. Here, the triH3cit showed more intracellular histone staining than the H3cit (R8). For detecting NETs, we decided to use the H3cit (R8) antibody for our protocol optimization. This antibody only bound to the extracellular H3cit and showed no staining of intracellular H3cit at this concentration (see Figure 1A,B).

For MPO staining, we compared human/mouse MPO antibody with the MPO (2C7) for human tissue (see Figure 1C,D) and the MPO (2D4) for mouse tissue (see Figure 1E,F). The MPO (2D4) and the MPO (2C7) antibodies could not achieve consistent staining for multiple tissue types, while the human/mouse MPO resulted in reliable, good staining for MPO. Thus, we selected human/mouse MPO for our staining protocol.

For NE, we tried a NE antibody from a mouse host on human tissue, which showed NE staining only in one out of five samples compared to the reliable staining of the NE antibody from a rabbit host. Additionally, the NE antibody from a rabbit host is applicable to human and mouse tissue. (see Figure 1G,H).

Figure 1: Primary antibody comparison in different tissues. (A) Human neonatal enterocolitis (NEC) tissue stained with H3cit (R8) (red). Here, only an extracellular signal can be detected. (B) Same tissue stained with triH3cit (R2,8,17) (red). This antibody creates a broader signal with intense intracellular staining of citrullinated histones (yellow arrows). (C,E) Human NEC tissue (C) and mouse volvulus tissue (E) with good staining for H3cit (R8) (red) and mouse/human MPO (green). The H3cit, MPO, and DAPI (blue) signal co-localization indicates NET formation (white arrows). (D,F) In comparison, (D) using the MPO (2C7) for human NEC tissue and (F)MPO (2D4) (green) for mouse volvulus tissue, no MPO signal could be obtained. (G) Burned human skin sample with very strong staining for the NE antibody from a rabbit host (magenta) compared to the (H) negative staining result for the NE antibody from a mouse host. For the isotype control, see Supplementary Figure Isocontrol 1. Please click here to view a larger version of this figure.

Deparaffinization
Here, the xylene was replaced with limonene, which showed equivalent to better deparaffinization of the tissue samples compared to xylene, with less autofluorescence in the background (see Figure 2A,B).

Autofluorescence-reducing agent
The ready-to-use autofluorescence-reducing agent based on Sudan Black can be applied for 2-20 minutes. Here, we applied it for 0 min, 5 min, and 10 min. When no blocking was used, some mouse samples showed more non-specific staining, and partial positive staining could be reached (see Figure 2C). The 5 min blocking time showed good results in all the tissue types except for H3cit and MPO in mouse lung and skin tissue (see Table 2). At 10 min of blocking time in some samples, the staining started to be less bright, so the 5 min time frame is the best choice for blocking autofluorescence (see Figure 2D,E).

Figure 2: Deparaffinization methods and usage of an autofluorescence-reducing agent. (A) Human spondylodiscitis tissue deparaffinized with limonene and stained for H3cit (red) and MPO (green). The stranded formation of the signals and partial co-localization indicate the presence of NETs (white arrow). (B) Same tissue sample deparaffinized with xylene, resulting in similar staining results, indicating that the widely used xylene can be substituted with a replacement medium. (C-E) Mouse lung tissue after induced sepsis showing different NE staining patterns (magenta) when different incubation times for the autofluorescence-reducing agent are used. With no autofluorescence reducer used, image C still shows a slight background staining of erythrocytes (red arrow). In contrast, image D shows that after 5 min incubation with an autofluorescence-reducing agent, a clear signal is emitted. After 10 min, the staining quality in image E declines, and the signal becomes less bright. For the isotype control, see Supplementary Figure Isocontrol 2. Please click here to view a larger version of this figure.

Antigen retrieval methods
For NE staining, we heated the samples in pH 6 citrate buffer for 10 min in the microwave (first for 1 min at 360 W and then for 9 min at 90 W), for 10 min in a water bath at 96 °C, or for 90 min in a water bath at 60 °C. Here, the higher temperatures in the microwave and water bath showed consistently moderate to good antigen retrieval (see Figure 3A). No significant difference was found between the microwave and the 96 °C water bath (see Table 2). Moreover, it was shown that in a 60 °C water bath, there was only partially positive to no staining (see Figure 3B). Only the human ileum and human myocardium showed good specific results. As no adequate staining for NE could be achieved with the 60 °C water bath, the 60 °C water bath test series for H3cit and MPO was discarded.

For double-staining with MPO and H3cit, we heated the samples in pH 6 citrate buffer in the microwave for 10 min (first for 1 min at 360 W and then for 9 min at 90 W) or in a water bath at 96 °C for 10 min. Here, both methods showed good specific results, with slightly more favorable results for the 96 °C water bath (see Figure 3C).Only mouse lung and skin tissue could achieve a moderate overall ranking (see Table 2).

However, exceeding the incubation time of 40 min at 96 °C resulted in less intense antibody staining, while substantially more background staining could be observed (see Figure 3D).

Figure 3: Antigen retrieval methods. (A) Mouse volvulus tissue stained for NE (magenta) using a 10 min incubation time at 96 °C for heat retrieval shows a significantly stronger signal than (B) when incubating the sample for 90 min in a 60 °C water bath. (C) Furthermore, the 10 min incubation in 96 °C antigen retrieval also results in a strong signal for H3cit (red) and MPO (green) with combined NET staining (white arrows). D: However, boiling the samples for more than 40 min results in less specific extracellular H3cit staining and no MPO staining. For the isotype control, see Supplementary Figure Isocontrol 3. Please click here to view a larger version of this figure.

Permeabilization
We tried permeabilizing the samples with Triton X-100 in two dilutions (0.2% and 0.5%) for 10 min and compared this to 10 min in deionized water. Here, 10 min with Triton 0.2% achieved good results across all tissue types, even though the differences were small compared to the 0.5% Triton and deionized water conditions (see Table 2).

Supplementary Figure Isocontrol 1: Isotype controls for Figure 1. All images show good DAPI (blue) staining but no signal for the fluorescent antibody. Thisconfirms that the binding of primary antibodies in Figure 1 is specific to the target antigen and not a result of non-specific binding or protein interactions. (A) Human neonatal enterocolitis (NEC) tissue stained with the isocontrol antibody for H3cit (R8) (red). (B) NEC tissue stained with the isocontrol antibody for H3cit (R2,8,17) (red). (C) NEC tissue (E) and mouse volvulus tissue stained with the isocontrol antibodies for H3cit (R8) (red) and mouse/human MPO (green). (D) NEC tissue stained with the isocontrol antibodies for H3cit (R8) (red) and MPO (2C7) (green). (F) Mouse volvulus tissue stained with the isocontrol antibodies for H3cit (R8) (red) and MPO 2D4 (green). G: Burned human skin sample stained with the isocontrol antibody for the NE antibody from a rabbit host (magenta). (H) Same tissue stained with the isocontrol antibody for the NE antibody from a mouse host (magenta). Please click here to download this File.

Supplementary Figure Isocontrol 2: Isotype controls for Figure 2. All images show good DAPI (blue) staining but no signal for the fluorescent antibody. This confirms the specific binding of the primary antibodies in Figure 2. Images C-E still show some background staining in magenta due to the long exposure times. However, the NE signal in Figure 2 is distinguishable from the background staining. (A) Human spondylodiscitis tissue deparaffinized with limonene and stained with the isocontrol antibodies for H3cit (R8) (red) and mouse/human MPO (green). (B) Same tissue sample deparaffinized with xylene and stained with the isocontrol antibodies for H3cit (red) and MPO (green). (C) Mouse lung tissue stained with the isocontrol antibody for NE (magenta), with no autofluorescence-reducing agent used. (D) Same tissue stained with the isocontrol antibody for NE (magenta) and 5 min of incubation with an autofluorescence-reducing agent. (E) Same tissue stained with the isocontrol antibody for NE (magenta) and 10 min of incubation with an autofluorescence-reducing agent. Please click here to download this File.

Supplementary Figure Isocontrol 3: Isotype controls for Figure 3. All images show good DAPI (blue) staining but no specific signal for the fluorescent antibody. This confirms the specific binding of the primary antibodies in Figure 3.(A) Mouse volvulus tissue stained with the isocontrol antibody for NE (magenta) using a 10 min incubation time at 96 °C for heat retrieval. (B) Same tissue stained with the isocontrol antibody for NE (magenta) and heat retrieval for 90 min in a 60 °C water bath. (C) Same tissue stained with the isocontrol antibody for H3cit (red) and MPO (green) and with identical heat retrieval conditions as in A. The green dot shows a staining artifact of aggregated secondary antibodies. (D) Same tissue stained with the isocontrol antibody for H3cit (red) and MPO (green) and using a 40 min incubation time at 96 °C for heat retrieval. Please click here to download this File.

Supplementary Figure Isocontrol 4: Isotype controls for Figure 4. All the following isocontrols are from the same slide as the corresponding "failed" experiment. The failed attempts were not made intentionally, so some isocontrols appear usable, while the sample was not. (A) NEC tissue stained with the isocontrol antibody for H3cit (red) and MPO (green). This sample was processed faster without drying out. Therefore, no excessive background staining is visible. The green and red structures are secondary antibody aggregates. (B) Spondylodiscitis tissue stained with the isocontrol antibody for NE (magenta). Here, the deparaffinization was successful, so no paraffin remnants can be seen. (C) NEC tissue stained with the isocontrol antibody for H3cit (red) and MPO (green). Even though the same improperly stored secondary antibodies were used, no staining would be expected for this isocontrol image. Therefore, this image cannot be used to evaluate the corresponding image's specific binding. (D) Burned human skin stained with the isocontrol antibody for H3cit (red) and MPO (green). Here, the mounting was done more carefully, and no light scattering through air bubbles can be seen. Please click here to download this File.

Discussion

In this work, we aimed to adapt and optimize the existing protocols for imaging NETs to more tissue types, beginning with the actual staining process. The first critical step for this method is the selection of the most suitable antibodies. For NE, we tried an NE antibody from a mouse host on human tissue, which showed no reliable staining compared to NE from a rabbit host. Furthermore, Thålin et al. proposed H3cit (R8) as a more specific antibody for extracellular staining. We compared this antibody with the widely used triH3cit (R2, R8, R17). Our study showed that the H3cit (R8) antibody stained only for an extracellular signal at the used concentrations, enabling easier recognition of NETs on the slides. This finding supports the claims from Thålin et al. that the H3cit (R8) clone can provide a good NET signal with decreased off-target cross-reactivity to non-citrullinated histones²⁸. We also compared MPO antibodies for human and mouse tissue for MPO staining. Our results show that the mouse/human MPO antibody showed more reliable staining and can be used simultaneously on human and mouse samples, simplifying its use in the laboratory. Consequently, we recommend using this antibody combination, H3cit (R8) and mouse/human MPO, for future research and implemented it in our protocol.

However, there is a limitation for multiple staining attempts with this antibody selection. This protocol was modified to use primary antibodies from different hosts. Otherwise, one secondary antibody could bind to both primary antibodies. The H3cit antibody comes from the same host as the NE antibody, so we could not stain NE and H3cit together. Even though no triple staining (NE, H3cit, MPO) was done in this study, this protocol could serve as a foundation for triple-staining experiments in the future. One possible option for triple-staining or NE and H3cit double-staining could be exchanging one of these antibodies with a reliable antibody from a different host. In this case, a possible combination partner could be the NE antibody (Santa Cruz; sc-55549) from sheep used by Knackstedt et al.²⁴. Another option by Duler et al. were published in 2021, where they used a consecutive double-staining method and combined NE and H3cit from a rabbit host²⁶. An additional promising application of a multiple-staining method is for the discrimination between intra- and extravascular NET formation. Instead of triple staining for NET markers, double NET staining could be combined with additional vascular staining. Furthermore, an increasing number of studies have examined the intra- and extravascular distribution of NETs to gain more knowledge on the pathomechanisms of specific diseases, such as Alzheimer's disease. Smyth et al.³¹ described a promising and in-depth protocol similar to ours to differentiate between the two possible localizations. They modified their existing NET staining protocol by adding a biotinylated lectin to label endothelial cells and used a fluorophore-conjugated streptavidin for the immunofluorescence staining of lectin. By examining the co-localization of lectin and NET markers in the same image, they were able to identify intravascular NETs³¹. In this context, further studies are needed to expand the application of our protocol.

After finding the most suitable antibody combination, the next critical step is introducing an autofluorescence-reducing agent. Exogenous factors such as sample fixation using formaldehyde fixatives and endogenous factors such as erythrocytes can result in high autofluorescence, which makes determining co-localization difficult³². This problem mainly occurs in the blue (excitation area: 430-480 nm) and green fluorescence spectra (excitation area: 500-550 nm) and can result in inconclusive staining results when a fluorescence antibody with the same excitation area is used³. The literature reports Sudan Black to be an effective agent for reducing inherent tissue autofluorescence³³. Sudan Black makes it possible to obtain clearer staining results using the blue or green fluorescence channels. This excitation area will be necessary for future multiple staining attempts because the blue and green channels are needed when using a fourth fluorescent dye. Our study shows that the optimum incubation time depends on the type of tissue and antibody used in the staining process. Nevertheless, in this work, 5 min of autofluorescence-reducing agent generally gave good results on all tissue types and had the benefit of staining the sample black, thus making it easier to see on the slides. This prevented missing parts of the tissues in the staining process, especially with small samples.

Another essential factor for the success of this protocol is a constant high temperature for the antigen retrieval step. Our research showed that 10 of the 15 previous studies used temperatures above 96 °C for antigen retrieval^{13,15,16,17,19,21,22,25,26,27}. This was consistent with our results, where incubating the samples in a 96 °C water bath or microwave gave good results with no significant differences. Nonetheless, we recommend using a water bath for even heat distribution. When using the microwave, we found a temperature gradient in the buffer of up to 5 °C between the top and the bottom of the staining jar. Additionally, the water bath is easier to use and has a lower risk of the buffer boiling over. For permeabilization, we found that the Triton X-100 had a lesser effect on the staining. So, skipping this step in the protocol is possible without adversely affecting the outcome. Generally, human samples gave better results than mouse samples. A reason for this could be that humans have a higher percentage of neutrophils than mice^34,35. Additionally, mouse lung samples showed no visible macroscopic inflammation and fewer neutrophils, while mouse intestine samples were macroscopically inflamed and showed good staining results. Thus, the lower scores in our study could have resulted from the low inflammation and not due to the protocol not working optimally.

For troubleshooting, minor errors in the protocol or not handling the samples carefully may have huge effects on the staining results. One major problem that we identified was sample dehydration. Here, we found that handling more than 10-15 slides at once was critical because every step is time-consuming and can result in dehydration of the samples. Dehydrated samples are no longer usable due to high background staining and no detectable specific fluorescence signal (see Figure 4A). Additionally, the samples should be completely deparaffinized before starting the staining protocol. Paraffin residue resulted in a strong background signal, and the cutting lines could be seen in the stained paraffin (see Figure 4B). Furthermore, after more than 20 freeze-thaw cycles, no signal for the secondary antibody to MPO could be detected (see Figure 4C). Another important step with a great impact on the image quality is the handling of the final mounting step. In this work, air bubbles under the coverslip resulted in the scattering of the fluorescent signal and, thus, a blurry image (see Figure 4D).

Figure 4: Troubleshooting tips. (A) This panel shows the consequence of a dried-out sample. The solid red background staining hinders any evaluation of the sample. (B) Here, the red arrow shows paraffin remnants, characterized by the corrugated appearance, after an insufficient deparaffinization. These remnants result in a high background signal and lower-quality images. (C) Inadequate antibody storage or more than 20 freeze-thawing cycles result in the aggregation of the secondary antibody (green arrows) and no binding to MPO. (D) Mounting that is not done carefully can result in air bubbles and, therefore, scattering of the fluorescent light (blue arrow). This can significantly impact the image quality, especially when the scattering blurs the target. For the isotype control, see Supplementary Figure Isocontrol 4. Please click here to view a larger version of this figure.

Even though NETs are gaining more popularity in scientific research, there are still not enough studies focusing on the methods for NET detection^3,13,17,26. To our knowledge, we present the first study comparing protocols from different studies for immunofluorescence imaging of NETs in FFPE tissue. The previously published customized protocols differed in their antibody or antigen retrieval methods and were often designed for only one tissue type, making comparing the NET imaging results more difficult. Therefore, in this study, we identified critical steps and established a protocol suitable for different mouse and human tissues. We achieved this by using a new primary antibody for H3cit staining and reducing the background staining with an autofluorescence-reducing agent. Furthermore, we showed that a constant high temperature is crucial, and careful handling of samples is fundamental for successful NET staining. Therefore, this study provides the essential steps for developing a generally applicable protocol for staining NETs.

Materials

List of materials used in this article

Name	Company	Catalog Number	Comments
			Dilution
Anti-Neutrophil Elastase antibody 100µg	abcam	Ab 68672	1:100
Anti-Histone H3 (citrulline R2 + R8 + R17) antibody  100µg	abcam	Ab 5103	1:50
Anti-Myeloperoxidase antibody [2C7] anti-human 100 µg	abcam	Ab 25989	1:50
Anti-Myeloperoxidase antibody [2D4] anti-mouse 50 µg	abcam	Ab 90810	1:50
Axiovision Microscopy Software	Zeiss	4.8.2.
Blocking solution with donkey serum (READY TO USE) 50ml	GeneTex	GTX30972
Coverslips	Marienfeld	0101202
Dako Target Retrieval Solution Citrate pH6 (x10)	Dako	S2369
DAPI 25 mg	Roth	6335.1	1:25000
DCS antibody dilution 500 mL	DCS diagnostics	DCS AL120R500
Donkey ant goat Cy3	JacksonImmunoResearch	705-165-147	1:200
Donkey anti rabbit AF647	JacksonImmunoResearch	711-605-152	1:200
Donkey anti rabbit Cy3	JacksonImmunoResearch	711-165-152	1:200
Fluoromount-G Mounting Medium	Invitrogen	00-4958-02
Glass slide rack	Roth	H552.1
Human/Mouse MPO Antibody	R&D Systems	AF 3667	1:20
Hydrophobic Pen	KISKER	MKP-1
Isokontrolle Rabbit IgG Polyclonal 5mg	abcam	Ab 37415	1:2000 and 1:250
MaxBlock Autofluorescence Reducing Reagent Kit (RUO) 100 ml	Maxvision	MB-L
Microscopy camera	Zeiss	AxioCamHR3
Microwave	Bosch	HMT84M421
Mouse IgG1 negative control	Dako	X0931 Aglient	1:50 and 1:5
Normal Goat IgG Control	R&D Systems	AB-108-C	1:100
PBS Phosphate buffered saline (10x)	Sigma-Aldrich	P-3813
PMP staining jar	Roth	2292.2
Recombinant Anti-Histone H3 (citrulline R8) antibody 100µg	abcam	Ab 219406	1:100
Recombinant Rabbit IgG, monoclonal [EPR25A] - Isotype Control 200µg	abcam	Ab 172730	1:300
ROTI Histol	Roth	6640
SuperFrost Plus slides	R. Langenbrinck	03-0060
TBS Tris buffered saline (x10)	Sigma-Aldrich	T1503
Triton X-100	Sigma-Aldrich	T8787
Tween 20	Sigma-Aldrich	P9416
Water bath	Memmert	830476
Water bath rice cooker	reishunger	RCP-30
Wet chamber	Weckert Labortechnik	600016
Zeiss Widefield microscope	Zeiss	Axiovert 200M