Enhanced neutrophil extracellular trap generation in rheumatoid arthritis: analysis of underlying signal transduction pathways and potential diagnostic utility

Human subjects

All patients fulfilled the American College of Rheumatology classification criteria for RA, or for systemic lupus erythematosus, respectively. Healthy volunteers, matched for gender and age, were recruited at the hospitals or at the Blood Bank of the Swiss Red Cross, Basel. Inclusion criteria for healthy controls were fair general condition, age ≥28 and ≤70 years, and for blood donors fulfilling national criteria for blood donation. Exclusion criteria were current or previous systemic autoimmune disease, asthma and reconvalescence after major illness, surgery, current medication with corticosteroids, immunosuppressive agents, and malignant neoplasia or chemotherapy within 5 years before recruitment for the study. RA cases had a DAS ≥2.6, were from age ≥27 to ≤70 years, and had no other systemic autoimmune disease, including ankylosing spondylitis and psoriatic arthritis.

Exclusion criteria were corticosteroids ≥40 mg equivalent of prednisone daily, and those mentioned earlier for healthy controls. Informed, written consent was obtained from all subjects in the study, which was approved by the Cantonal Ethical Review Boards of Aargau-Solothurn and Basel/Basel-Land, Switzerland.

Preparation of plasma and serum

Plasma and serum were collected and processed as described previously [15]. Samples were studied immediately or stored at -80°C until analysis.

Cell isolation

PMNs were isolated with Dextran-Ficoll density centrifugation [8]. Cell viability was 96% to 98%, with a purity of >95% PMNs. Neutrophils seeded in 24-well plates were allowed to settle for 1 hour at 37°C under 5% CO₂ before further experimentation.

Cell-free DNA isolation and quantification

Cell-free DNA extracted from 850 μl plasma or serum by using the QIAamp Circulating Nucleic Acid Kit (Qiagen, Valencia, CA, USA) was quantified by TaqMan real-time PCR (StepOne Plus Real-Time PCR System, Applied Biosystems, Foster City, CA, USA) specific for the glyceraldehyde-3-phosphate dehydrogenase (GAPDH) gene [15].

Detection of neutrophil elastase (NE), myeloperoxidase (MPO), and cell-free nucleosomes

The concentrations of neutrophil elastase (NE) and myeloperoxidase (MPO) were measured with sandwich ELISA (Elastase/a1-PI Complex ELISA Kit, Calbiochem/EMD, Gibbstown, NJ, USA) and the human (MPO ELISA Kit, Hycult Biotech, Plymouth Meeting, PA, USA) respectively. Nucleosomes were measured by using the Human Cell Death Detection ELISA^PLUS (Roche Diagnostics, Basel, Switzerland). Cell-culture supernatants were incubated with DNaseI (10 U for 5 minutes) (Roche Diagnostics) before analysis [20].

MPO/DNA complex detection

MPO is present on extruded NETs. To detect such structures, NET-associated MPO/DNA complexes were quantified by using a modified capture ELISA [21]. In brief, NET-associated MPO in serum or culture supernatant was captured by using the coated 96-well plate of the human MPO ELISA Kit, (Hycult Biotech), after which the NET-associated DNA backbone was detected by using the detection antibody of the Human Cell Death Detection ELISA^PLUS (Roche Diagnostics).

PAD4/DNA-complex detection

To detect the presence of PAD4 on extruded NETs in culture supernatants after spontaneous NETosis, cell-free PAD4/DNA complexes were quantified by using a modified capture ELISA, akin to that described for MPO earlier. In brief, cell-free PAD4 was captured by using the coated 96-well plate of a commercial human PAD4 ELISA (USCN Life Science, Inc., Wuhan, China), and associated DNA was detected by using Human Cell Death Detection ELISA^PLUS kit (Roche Diagnostics).

ROS generation analysis

ROS was measured by using a 2′,7′-dichloro dihydrofluorescein diacetate (DCFH-DA) assay [22]. The 5 × 10⁵ cells in a final volume of 500 μl were incubated for 30 minutes with 25 μM DCFH-DA (Sigma-Aldrich, St. Louis, MO, USA). Fluorescence was measured with flow cytometry (FACSCalibur; BD Biosciences, San Jose, CA, USA).

Fluorescence and scanning electron microscopy

The 5 × 10⁴ cells isolated PMNs seeded on poly-L-lysine-coated coverslips (BD Biosciences) were stimulated with phorbol-12-myristate-13-acetate (PMA, Sigma-Aldrich) for 90 minutes and dehydrated with a graded ethanol series (30%, 50%, 70%, 100%) [8], coated with 2-nm platinum, and analyzed with a Nova NanoSEM 230 scanning electron microscope (FEI Co., Hillsboro, OR, USA). PMNs were incubated for 10 minutes with 5 μM Sytox Green dye (Invitrogen Life Technologies, San Diego, CA, USA) for assessment of NETs with an Axiovert fluorescence microscope (Carl Zeiss) coupled to a Zeiss AxioCam color CCD camera (Carl Zeiss Microimaging, Oberkoch, Germany) [8, 23].

Immunohistochemical staining and quantification of NETs

The 5 × 10⁴ isolated PMNs were seeded on poly-L-lysine-coated glass coverslips (BD Biosciences, San Jose, CA, USA) in tissue-culture wells and allowed to settle before stimulation, as described earlier. Coverslips were rinsed with ice-cold HBSS and the cells fixed with 4% paraformaldehyde and blocked overnight (HBSS with 10% goat serum, 1% BSA, 0.1% Tween20, and 2 mM EDTA) at 4°C. NETs were detected with rabbit anti-NE (Abcam, Cambridge, MA, USA), rabbit anti-MPO (Dako, Glostrup, Denmark), two different rabbit anti-PAD 4 (Abcam), mouse anti-PAD4 (Abcam), mouse anti-histone H1 + core proteins (EMD Millipore, Billerica, MA, USA), and rabbit anti-citrullinated histone H3 (citH3, Abcam). Secondary antibodies were goat anti-rabbit IgG AF555, goat anti-rabbit IgG AF488 (Invitrogen Life Technologies, San Diego, CA, USA), and goat anti-mouse IgG AF647. DNA was stained with 4′,6-diamidino-2-phenylindole (DAPI, Sigma-Aldrich), and NETs were visualized by using a Zeiss Axioplan 2 Imaging fluorescence microscope in conjunction with a Zeiss AxioCam MRm monochromatic CCD camera and analyzed with Axiovision 4.8.2 software (Carl Zeiss). A minimum of 20 fields (at least 1,000 PMNs) per case was evaluated for MPO/NE and DNA co-staining; nuclear phenotypes and NETs were counted and expressed as percentage of the total number of cells in the fields.

RA serum depletion, IgG purification, and quantification of NETs

After three washes with PBS, 200 μl protein G agarose (Pierce Biotechnology Inc, Rockford, IL, USA) was incubated with 200 μl ACPA + and ACPA- RA or control serum diluted in an equal volume of phenol red-free RPMI 1640 medium overnight at 4°C. The serum/protein G agarose mixture was centrifuged at 2,500 g for 5 minutes, and the supernatant (IgG-depleted serum) was carefully transferred into a new Eppendorf microcentrifuge tube. The protein G agarose pellet was gently washed 3 times with 500 μl ddH₂O, and the bound antibody was released by the addition of 50 μl 0.1 M glycine pH 2–3, and immediately equilibrated with 10 μl of 1 M Tris pH 7.5-9. All protein concentrations were determined with the MN Protein Quantification Assay (Macherey-Nagel GmbH, Düren, Germany), and isolation of IgG was verified with Coomassie staining of SDS-PAGE.

Neutrophils from healthy donors (n = 3) were isolated and cultured for 2 hours in 96-well culture dishes (Thermo Fischer Scientific, Waltham, MA, USA), supplemented with: serum, depleted serum, and purified IgG from ACPA-positive RA patients (n = 3), ACPA-negative RA patients (n = 3), and healthy individuals (n = 3) to a final concentration of 100 μg/ml.

NETs were quantified after IHC staining with mouse anti-human MPO antibody (Abcam) and rabbit anti-human citH3 antibody (Abcam), or the respective isotype controls, followed by incubation with goat anti-mouse IgG AF555 and goat anti-rabbit IgG AF488 (Invitrogen Life Technologies). DNA was counterstained with 4′,6-diamidino-2-phenylindole (DAPI, Sigma-Aldrich). NETs were visualized by using an Olympus IX81 motorized epifluorescence microscope (Olympus America Inc., Center Valley, PA, USA) in conjunction with an Olympus XM10 monochromatic CCD camera (Olympus) and analyzed with the Olympus CellSens Dimension software (Olympus). A minimum of 20 fields at 10× magnification (at least 500 to 1,000 PMNs) per case was evaluated for MPO/citH3 and DNA co-staining through ImageJ analysis software (National Institutes of Health Image Processing, Bethesda, MD, USA); nuclear phenotypes and NETs were determined, counted, and expressed as percentage of the total area of cells in the fields [24].

Protein isolation and Western blot analysis

Total protein was isolated with NucleoSpin TriPrep kit (Macherey-Nagel) from 3 × 10⁶ PMNs. Proteins from the nuclear and cytoplasmic fractions were isolated by using the Nuclear and Cytoplasmic Protein Extraction Kit (Thermo Scientific). Western blotting was performed by using AnykD Mini-PROTEAN TGX Gels (Biorad, Hercules, CA, USA) and nylon/nitrocellulose membranes (Biorad). Primary and secondary antibodies used were: rabbit anti-PAD4 (Abcam), rabbit anti-MPO (Cell Signalling Technologies, Beverly, MA, USA), mouse anti-β-Actin (Sigma-Aldrich), goat anti-Mouse and/or anti-Rabbit, human anti-HRP (Southern Biotech). HRP activity was detected by using SuperSignal West Pico Chemiluminescent Substrate (Thermo Scientific). Equal loading was verified by using beta-actin or histone H3, when appropriate. Western blots of citrullinated H3 (citH3) protein were performed according to Shechter et al. [25]. Densitometric analysis and protein quantification of the Western blots was performed by using the ImageJ software.

RNA isolation and quantitative real-time PCR

Total RNA was isolated by using RNeasy Mini Kit (Qiagen). TaqMan real-time quantitative RT-PCR was performed by using the Applied Biosystems StepOne Plus cycler (Applied Biosystems) and TaqMan Gene Expression Assay primer/probe sets (Applied Biosystems) for ELANE (HS00236952_m1), MPO (HS00924296_m1), PADI4 (HS00202612_m1), and β₂-microglobulin B2M (HS99999907_m1). Data were normalized by using the housekeeping gene B2M, after a selection procedure involving six different endogenous reference genes, as suggested in the MIQE guidelines [26]. Relative values were calculated with 2^-DDCt analysis [27].

Statistical analysis

All data are presented as mean ± SEM. Descriptive statistics for continuous parameters consisted of median and range, and categoric variables were expressed as percentages. Comparisons between patients and healthy controls were by the Mann–Whitney U test with a Welch posttest correction. Statistical significance in multiple comparisons was by one-way analysis of variance (ANOVA) with a Dunn posttest correction. P < 0.05 was considered statistically significant.

Receiver operating characteristic (ROC) curves were calculated, and the area under the curve (AUC) with corresponding standard errors of means was calculated. Data were processed in GraphPad Prism version 5.0b for MacOSX (GraphPad Software Inc., San Diego, CA, USA). Analysis of covariance (ANCOVA) was conducted with SPSS version 21.0 statistical software (IBM SPSS Inc., Chicago, IL, USA). Additional professional statistical assistance was provided by A. Schoetzau, Basel.

RA-derived PMN exhibit increased spontaneous NETosis

As in very recent observations [13], we observed that RA-derived PMNs underwent greater degrees of NETosis than did control PMNs in vitro (data not shown). To study this facet in more detail, we examined the kinetics of spontaneous NET extrusion, for which purpose, PMNs isolated from peripheral blood samples were allowed to settle for 1 hour and then cultured for a period of up to 3 hours in vitro (Figure  1A), NETs being detected by immunohistochemistry for neutrophil elastase (NE) and DAPI (4′,6-diamidino-2-phenylindole) (Figure  1B). In addition, we quantitatively assessed the degree of in vitro NETosis in these cultures by determining the concentration of cell-free nucleosomes in the respective supernatants (Figure  1C), specifically their association with myeloperoxidase (MPO), indicative of the NETotic origin of this material [2, 21] (Figure  1D). We also measured NET-associated MPO enzymatic activity by using tetramethylbenzidine (TMB) as a substrate (data not shown). The results clearly indicate that RA-derived PMNs generated NETs more rapidly, to a greater magnitude, and more extensively than did control healthy PMNs, which was particularly evident at the 3-hour stage of in vitro culture (Figure  1B to D). Accounting for variances in PMN counts, the difference between the healthy control and RA subjects remained highly significant in an analysis of covariance (ANCOVA) of the nucleosome assay (P < 0.001).

During NETosis, the morphology of the PMN nucleus changes from the familiar lobulated to a diffused and then to a spread phenotype (Figure  1E) [28]. By examining and enumerating these features, it was observed that at baseline (T1), the nuclei from healthy control PMNs were predominantly lobulated, whereas in this instance, the majority of RA-derived PMN nuclei exhibited a delobulated or diffused nuclear phenotype (Figure  1F). In RA-derived PMNs, this delobulated population decreased over time, giving rise to NETotic cells with a spread phenotype (Figure  1F). In contrast, in normal PMNs, we noted a steady progression in the proportion of delobulated cells (Figure  1F). The spontaneous progression of nuclei to the NETotic-spread phenotype was more pronounced in RA than in normal PMNs, a feature most evident after 3 hours (T3) (Figure  1F).

RA-derived PMNs demonstrate increased expression of NET-associated signaling elements, nuclear localization of PAD4, and augmented H3 citrullination

NETosis has been shown to depend on a number of biochemical signaling elements, among which are the generation of ROS by nicotinamide adenine dinucleotide phosphate (NADPH) oxidase, the action of NE in combination with MPO, and histone citrullination by PAD4 [2, 3, 5]. RA-derived PMNs exhibited increased basal intracellular ROS levels (Figure  2A), as well as increased levels of NE (Figure  2B) and MPO (Figure  2C and D), as determined with real-time PCR and/or Western blotting.

Surprisingly, neither PAD4 mRNA expression nor PAD4 levels in total cellular protein showed any discernible difference between RA and control PMNs (Figure  2E and F, respectively). Because PAD4 translocates to the nucleus on PMN activation [4, 29], where it citrullinates histone proteins, such as H3, we examined its presence in nuclear and cytoplasmic PMN fractions. Compared with control PMNs, PAD4 was preferentially located in the nucleus of RA-derived PMNs (Figure  2G). The nuclear presence of PAD4 was associated with increased citrullinated histone H3 (citH3) levels with Western blot analysis in PMNs from RA cases compared with controls (Figure  2H). Furthermore, citrullinated histone H3 could be readily detected on NET structures (Figure  2I).

Potential extracellular localization of PAD4 on NETs

Because we observed elevated nuclear translocation of PAD4 in RA PMNs, we examined whether this enzyme is extruded into the extracellular environment during NETosis. Unfortunately, the visualization of such an event by fluorescence immunohistochemistry proved to be difficult with a variety of commercially available antibodies, and we obtained only rudimentary evidence for the presence of PAD4 on NETs by this means (refer to Figure  2J).We were, however, able to detect PAD4/cell-free DNA complexes readily in culture supernatants from isolated PMNs, the levels of which were increased in cases with RA (Figure  2K). It is, therefore, quite probable that PAD4 is associated with NETs structures after aberrant NETosis in RA.

PMNs from RA patients present an enhanced NETotic response to PMA, and normal PMNs are strongly affected by RA serum and synovial fluid

In autoinflammatory or malignant conditions, such as SLE or cancer, an elevated NETotic response of PMNs to an external activation signal has been shown [9, 13]. In our experiments, we noted that when RA-derived PMNs were treated with PMA, they responded far more vigorously with regard to NETosis than did controls, as detected by SEM and fluorescence microscopy (Figure  3A and B, respectively).

In addition, morphometric assessment indicated that RA-derived PMNs exhibited a larger decrease in cells with a delobulated phenotype and a greater progression toward a NETotic-spread nuclear phenotype than control PMNs (Figure  3C), a feature accompanied by excessive release of cell-free nucleosomes in culture supernatants (Figure  3D). PMA appears to activate PAD4, as it enhanced translocation from the cytoplasm to the nucleus (Figure  3E). The stimulatory effect of PMA on the release of nucleosomes into the supernatant was abrogated by Cl-amidine, a chemical inhibitor of PAD4, indicating that PAD4 signaling is necessary for NETosis induced by PMA [4, 30] (Figure  3D). These data indicate that PMNs in RA are susceptible to increased NETosis after stimulation by a secondary signal, such as that mediated by PMA.

Because SLE sera and RA sera and synovial fluid (SF) have been shown to confer an increased NETotic response [9, 13], we examined whether RA-derived sera or SF exerted similar effects on normal PMNs. As noninflammatory controls, we used healthy serum or osteoarthritis SF. Both RA sera and SF induced a pronounced increase in NETosis, which was paralleled by an increase in the nucleosome content of the supernatant (Figure  4A,B), as well as in ROS production (Figure  4C) when compared with healthy serum or osteoarthritis SF, respectively.To assess whether antibodies participate in the effects of RA serum on normal PMNs, we depleted IgG from serum of ACPA-positive and -negative RA patients and healthy controls. Compared with nondepleted sera, IgG depletion of both ACPA-positive and -negative sera markedly reduced NET induction to levels of normal serum (Figure  4D). Whereas the reconstitution of ACPA-negative IgG to serum did not increase NET formation significantly compared with controls, it was practically reversed to the original value in the ACPA-positive cases. This indicates a prominent role for ACPA in the induction of NETs in ACPA-positive RA, while suggesting that an alternative mechanism is responsible for the increased NETosis in ACPA-negative RA patients.

These data are in accordance with recent findings that RA-derived serum and SF induce NETosis in normal PMNs, and that ACPA and also IgM RF are to a large part responsible for this effect [13, 14].

Serum from RA patients shows elevated levels of the principal components of NETs, indicating enhanced NET extrusion during clot formation, which has potential clinical utility

As we had previously observed increased levels of cell-free DNA (cfDNA) in RA sera [15], we determined whether this resulted from enhanced NETosis, and whether this could have diagnostic applications. cfDNA concentrations were indeed significantly higher in serum samples from RA cases compared with age-matched healthy control sera (Figure  5A). In parallel, the concentrations of cell-free nucleosomes, NE, and MPO were significantly elevated in RA serum compared with control sera (Figure  5B to D). The association of a significant fraction of MPO with markedly elevated cell-free nucleosomes in RA serum, which constitute a main component of NETs (Figure  5E), clearly suggests that NETosis is indeed the source of nucleosome material present in RA serum [21].Since there was no significant elevation of these parameters in simultaneously obtained plasma samples that were processed immediately, these data demonstrate a propensity for RA PMN to undergo increased NETosis during coagulation (Figure  5A to E).

To ascertain whether NET-associated serum components could be diagnostically useful, we conducted ROC analyses. For serum cell-free nucleosomes, this yielded the surprisingly high AUC value of 0.97 (see Additional file 2 and Figure  5F). Of interest is that no significant difference was found in this value regardless of whether the RA cases were ACPA positive or not (Figure  5G), although a slight trend for serum nucleosome levels was higher in ACPA-positive cases than in ACPA-negative cases (Figure  5H). The AUC for serum nucleosomes was significantly higher than for any of the other parameters examined (Figure  5I to K). With the cut-off set at 0.78, the ROC AUC translates into a sensitivity of 91%, with a specificity of 92% for differentiating between RA cases and healthy controls.

In contrast to RA cases, cell-free nucleosome serum values of 14 cases with SLE showed a slight, but statistically significant increase (Additional file 2). This translated into an ROC AUC of 0.7639 (see Additional file 3), which is below the clinically useful value for diagnostic purposes.