This survey was carried out in accordance to the Justus Liebig University Animal Care Committee guidelines. Protocols were approved by the Ethic Commission for Experimental Animal Studies of the Federal State of Hesse (Regierungspräsidium Giessen) (A9/2012; JLU-No. 521_AZ), in accordance to the prevalent European Animal Welfare Legislation: ART13TFEU and the current applicable German Animal Protection Laws.

All NET-related experiments were performed with tachyzoite stages of the apicomplexan parasite N. caninum [strain Nc1 (42)], which was cultivated in vitro as described elsewhere (7, 11). In brief, N. caninum tachyzoites were maintained by serial passages either in primary bovine umbilical vein endothelial cells (BUVEC) or permanent African green monkey kidney epithelial cells (MARC-145). Viable N. caninum-tachyzoites were collected from infected host cell layer supernatants, pelleted (400 × g, 12 min), washed thrice in sterile PBS, counted in a Neubauer hemocytometer (Marienfeld-Superior, Germany) and re-suspended in sterile RPMI 1640 medium (Gibco) until further experimental use.

MARC-145 cell layers were maintained in cell culture medium DMEM (Sigma-Aldrich) supplemented with 1% penicillin (500 U/ml; Sigma-Aldrich), streptomycin (500 mg/ml; Sigma-Aldrich), and 10% FCS (Gibco) and cultivated at 37°C and 5% CO₂ atmosphere until confluency. Confluent MARC-145 layers were infected with viable N. caninum tachyzoites (20 × 10⁶ parasites/25 cm²).

Isolation of primary BUVEC was performed according to the method reported by Taubert et al. (7). In brief, the umbilical cords retrieved from newborn calves were enriched with 1% penicillin–streptomycin (Sigma-Aldrich, St. Louis, MO, USA) and refrigerated in 0.9% HBSS–HEPES buffer (pH 7.4; Gibco, USA). Endothelial cells were isolated using 0.025% collagenase type II (Worthington Biochemical Corporation, USA), filling the lumen of the ligated umbilical vein and incubating for 20 min at 37°C in 5% CO₂ atmosphere. Then, the umbilical vein was mildly massaged; the collagenase-cell suspension was retrieved and 1 ml FCS (Gibco, USA) was aggregated to inactivate the collagenase type II. After two centrifugations (400 × g, 10 min, 4°C), the isolated BUVEC were kept in complete ECGM (endothelial cell growth medium; PromoCell, Heidelberg, Germany), plated in 25 and 75 cm² plastic culture flasks (Nunc, Roskilde, Denmark), and incubated at 37°C in 5% CO₂ atmosphere until confluency.

Healthy adult dairy cows (n = 3) were bled by puncture of the jugular vein and 30 ml blood was collected in 50 ml sterile plastic tubes (Greiner), containing 0.1 ml heparin (Sigma-Aldrich) as anticoagulant. Approximately 20 ml heparinized blood was re-suspended in 20 ml PBS with 0.02% EDTA (Sigma-Aldrich), slowly layered on the top of 12 ml Biocoll Separating Solution^® (Biochrom AG), and centrifuged (800 × g, 45 min). After the extraction of plasma and mononuclear cells, the pellet was washed in 25 ml distilled water and gently shaken during 40 s to lyse erythrocytes. Osmolarity was rapidly normalized using an appropriate volume of Hanks balanced salt solution (4 ml, HBSS 10×, Biochrom AG). To complete the erythrocyte lyses, this step was repeated twice and the PMN were later re-suspended in RPMI medium (Gibco). Calculation and viability of the cells were performed in a Neubauer hemocytometer as described elsewhere (12). Finally, bovine PMN were cultured at 37°C and 5% CO₂ atmosphere for 30 min until further use. As neutrophils have a short lifespan, PMN isolation was performed not exceeding 3 h after blood collection.

Bovine PMN (n = 3) were re-suspended in medium RPMI 1640 lacking phenol red and without serum and then confronted in duplicates with vital N. caninum tachyzoites (37°C, 4:1 ratio: 1 × 10⁶ N. caninum tachyzoites versus 2.5 × 10⁵ bovine PMN/200 μl). For NET blockage, the following inhibitors were used: the NOX-inhibitor DPI [10 µM, Sigma-Aldrich, according to Farley et al. (43)], the leukocyte elastase-inhibitor Suc-Ala-Ala-Pro-Val chloromethyl ketone [CMK; 1 mM, Sigma-Aldrich, according to Scapinello et al. (44)], the MPO-inhibitor 4-aminobenzoic acid hydrazide [ABAH; 100 µM, Merck, according Parker et al. (45)], the SOCE-inhibitor aminoethoxydiphenyl borate [2-APB; 100 µM, Sigma-Aldrich, according to Conejeros et al. (46)], UO126 as inhibitor of ERK1/2 [50 µM; Sigma-Aldrich, according to Muñoz-Caro et al. (12)], SB 202190 as specific inhibitor of p38 MAPK [10 µM; Sigma-Aldrich, according to Muñoz-Caro et al. (12)], the G-protein-coupled receptor (GPCR) antagonist NF-449 for P2Y2 blockage (GPCR-NF-449, 10 µM; Santa Cruz Biotechnology), and N-α-benzoyl-N5-(2-chloro-1-iminoethyl)-l-Orn amide for PAD4 inhibition (Cl-amidine, 200 µM, Merck). For blocking experiments, PMN were pre-exposed with the corresponding inhibitor in serum-free medium RPMI 1640 without phenol red (RT, GPCR-NF-449, and Cl-amidine: 120 min, all other inhibitors: 30 min) prior to exposure to N. caninum tachyzoites. To disrupt NETs and facilitate their DNA quantification, 50 µl of micrococcal nuclease buffer (New England Biolabs) including 0.1 U/μl micrococcal nuclease (New England Biolabs) was supplied to each well and incubated (15 min, 37°C). Next, all the samples were centrifuged (300 × g, 5 min). The supernatant of each sample was deposited in duplicate into a 96-well flat-bottom plate (100 µl per well). DNA from NETs was assessed using Pico Green^® (Invitrogen), an extracellular DNA-linking fluorescent stain. Fifty microliters of Pico Green^® (diluted 1:2,000 in 10 nM Tris buffer with 1 mM EDTA) was added to each well. NET production was quantified according to the fluorescence intensities obtained in the spectrofluorometric analysis (484 nm excitation wavelength and 520 nm emission wavelength) performed by an automated plate monochrome reader (Varioskan Flash^®, Thermo Scientific). For negative controls, PMN in normal serum-free medium RPMI 1640 without phenol red were employed. Zymosan (1 mg/ml; Sigma-Aldrich) stimulated PMN served as positive controls according to Muñoz-Caro et al. (12). Diverse PMN-tachyzoites ratios were applied (1:1, 1:2, 1:3, 1:4) for dose-dependency evaluation. For analyses on the role of CD11b in parasite-triggered NETosis, the CD11b receptor was blocked via preincubation of bovine PMN in monoclonal mouse anti-bovine CD11b antibodies (MCA1425, diluted 1:5 in PBS; AbDSerotec, 30 min, RT). As antibody control, an irrelevant monoclonal antibody (mouse anti-bovine CD4, AbDSerotec) was used as described elsewhere (12). To resolve NET formation, DNase I (90 U/well, Roche Diagnostics) was supplemented 15 min before the end of incubation period.

Following PMN: N. caninum tachyzoite co-cultivation (ratio 1:4, 120 min, on 15 mm round glass coverslips pretreated with poly-l-lysine), fixation of the samples (4% paraformaldehyde, Merck, 15 min, 37°C), and three washings in PBS, the samples were blocked with BSA (2%, Sigma-Aldrich), incubated in antibody solutions (1 h, RT), and mounted on Prolong Gold^® with 4′-6-diamidino-2-phenylindole (DAPI) staining (Invitrogen, 1:1,000, 5 min, RT, in the dark). For the identification of antimicrobial peptides within extracellular DNA structures, the following antibodies were applied: anti-histone (H3) monoclonal [DyLight, ab139848, Abcam (1:1,000)], anti-MPO (Alexa Fluor 488, ABIN906866, https://Antibodies-online.com, 1:1,000), anti-NE (AB68672, Abcam, 1:1,000), and anti-pentraxin (SAB2104614-50UG, Sigma-Aldrich, 1:1,000) antibodies. The immunofluorescence images were taken by a digital camera from an inverted Olympus IX81 fluorescence microscope.

To analyze the repercussions of parasite-induced NETs on tachyzoite infectivity, three different parallel experimental conditions were chosen: (1) N. caninum were cocultured with PMN (1:4 ratio, 2 h, 37°C) allowing for effective NET formation. (2) For infection control, an equal number of non-exposed tachyzoites was incubated in plain medium (2 h, 37°C). (3) The same amount of parasites were incubated with PMN (1:4 ratio, 2 h, 37°C) permitting a competent NET formation and furthermore treated with DNase I (90 U/well, addition of DNase I 15 min before the end of the incubation period) to resolve NET structures and to indirectly measure potential adverse effects of NETs on tachyzoite viability. In a next step, the tachyzoites of setups 1–3 were transferred to confluent BUVEC monolayers for infection (4 h, 37°C, 5% CO₂ atmosphere). Overall, three different BUVEC isolates were used in this host cell invasion experiment. After incubation, BUVEC layers were washed to remove PMN and free tachyzoites. The infection rates were estimated microscopically (24 h p. i.) in 10 randomly selected vision power fields (400× magnification).

Cattle PMNs were cocultured with viable tachyzoites of N. caninum (ratio: 4:1) for 10, 30, 60, and 120 min on 10 mm glass coverslips (Nunc) prepared with poly-l-lysine (Sigma-Aldrich). Then, cells were fixed in 2.5% glutaraldehyde (Merck), post-fixed in 1% osmium tetroxide (Merck), washed in distilled water, dehydrated, dried by CO₂ treatment, and sputtered with gold as described elsewhere (11, 12). SEM samples were analyzed using a Philips XL30 scanning electron microscope (Institute of Anatomy and Cell Biology, Justus Liebig University Giessen, Germany).

Statistical analyses were performed by using Graph Pad Prism^® 6 software. One- or two-factorial analyses of variance (ANOVA) with repeated measures were applied to compare co-culture/stimulation conditions using a normal distribution of data. Dunnett’s multiple comparison tests were performed in dose and kinetic assays as follow-up test to ANOVA. For comparing enzyme activities, Tukey’s multiple comparison tests were used. Differences were considered as significant at a level of *p ≤ 0.05; **p ≤ 0.01; ***p ≤ 0.001, and ****p ≤ 0.0001.

Scanning electron microscopy analyses revealed tachyzoite-triggered generation of a fine network of grosser and slimmer strands of fibers produced by bovine PMN and being solidly adhered to tachyzoites (see Figure 1). Kinetic studies reported several stages of NETosis: posterior to 30 min of exposure, smooth PMN-derived filament structures capturing tachyzoites were observed (Figure 1A). Here, PMN still presented undamaged cell morphology. Thereafter, parasites were entrapped in an extracellular network of long drawn-out fibers originating from disrupted PMN (Figures 1B,C) and conglomerates of N. caninum tachyzoites and rather thick and chunky meshworks of PMN-derived filaments (Figure 1D, 60 min) were observed.

DAPI-based fluorescence analyses further proved the presence of NET-like structures containing DNA (see Figure 2). Furthermore, N. caninum tachyzoites were located in close proximity to NETs and presumably were trapped in these extracellular chromatin-rich structures (Figure 2). Moreover, co-localization of extracellular chromatin with histones (H3), NE, MPO, and pentraxin in parasite-capturing structures validated the typical nature of NETs (Figure 2).

Neutrophil extracellular trap quantification experiments revealed N. caninum tachyzoites as strong triggers of NETosis, since these stages induced even stronger reactions than zymosan stimulation of PMN (= positive control, Figures 3–6). Kinetics on NETosis indicated a significant induction of NETs formation in both incubation time periods (1 and 2 h) compared with the negative control (p ≤ 0.01, Figure 3). As expected, DNase I treatments leading to NET disintegration reduced NETosis under the basal levels of the negative controls (Figure 3). Furthermore, increasing amounts of N. caninum tachyzoites led to enhanced levels of NET formation as significant differences were observed at a ratio of 1:3 and 1:4 (PMN: tachyzoites) in comparison with the negative controls (p ≤ 0.05 and p ≤ 0.01, respectively, Figure 4).

To further corroborate the molecular characteristics of N. caninum-mediated NETosis, functional blocking experiments with DPI, a potent blocker of NOX, were performed. The PMN treatment with DPI resulted in a clear reduction of tachyzoite-triggered NET formation (Figure 5A), despite that this diminution was statistically not significant. Additionally, treatments of PMN with MPO and NE inhibitors (ABAH and CMK, respectively) also generate a decrease of tachyzoite-triggered NET formation (Figure 5A) underlining the pivotal role of NE and MPO in this process.

We here investigated the actual role of NET-associated molecular signaling pathways, receptors, and Ca⁺⁺ influx in N. caninum-triggered NETosis. The use of inhibitors affecting ERK1/2-(UO126) and p-38 MAPK-(SB202190) signaling routes in functional NET-derived studies caused a diminishment of tachyzoite-induced NET production (Figure 5B), proving a key role of ERK1/2- and p38 MAPK in N. caninum-triggered activation of NETosis-related signaling pathways.

Given that NOX-dependent ROS synthesis in bovine neutrophils is being reported as a Ca⁺⁺-associated process (12, 47, 48), we here furthermore tested whether N. caninum-triggered NETosis was influenced by SOCE. Treatments of PMN with 2-ABP produced a decrement of parasite-mediated NETs liberation (Figure 5B) proving that intracellular Ca⁺⁺ mobilization is necessary for efficient parasite-induced NETosis.

Until now, no information is available on PMN receptors related to N. caninum-triggered NETosis. Therefore, we analyzed whether antibody-mediated blocking of bovine CD11b leads to the diminishment of tachyzoite-triggered NETosis. Indeed, pretreatment of PMN with anti-CD11b led to a decrease of NET formation, but, however, these reactions were barely no significant.

Since no data are available on the role of histone hypercitrullination in parasite-induced NETosis so far, we here also intended to analyze whether the PMN exposure to Cl-amidine (200 µM), a specific inhibitor of PAD4, might have an impact on PAD4-derived histone hypercitrullination and chromatin decondensation during parasite-triggered NETosis. Cl-amidine pretreatment of bovine PMN resulted in diminished NET production when compared with non-treated but N. caninum tachyzoites-exposed PMN (Figure 5C). The same experiment was performed with zymosan as positive control (1 mg/ml) and negative controls (PMN cultured in plain medium alone).

We further intended to determine whether N. caninum-triggered NETosis is an energy and ATP-dependent process as seen for other PMN effector mechanisms (49–51). Therefore, the blocker of the ATP-specific G-protein receptor P2Y2 (NF-449) was used here for functional inhibition experiments. In fact, PMN-pretreatment with NF-449 led to a reduction on parasite-triggered NETosis (Figure 5D) when compared to non-treated parasite-exposed PMN.

Host cell penetration is a vital requisite of the parasite N. caninum to survive and reproduce successfully within a host. Therefore, to analyze the effects of NET-mediated parasite entrapment on subsequent tachyzoite host cell infectivity, PMN-pre-exposed tachyzoites were transferred to BUVEC monolayers as suitable specific host cells and infection rates were later calculated. In the same way, an equal amount of tachyzoites, which had not been in contact with PMN before were used to infect BUVEC. The prior confrontation of parasites with PMN and subsequent NET development significantly (p ≤ 0.0001) prevented N. caninum tachyzoites from host cell invasion (Figure 6). As such, infection rates decreased from log 60% = 1.778 (resulting from non-exposed tachyzoites = infection controls) to log 20% = 1.3 induced by PMN-pre-exposed tachyzoites. To prove that this impairment was due to NETosis, parallel samples containing the same numbers of tachyzoites and PMN were treated with DNase I treatment (leading to NET disentangle) 165 min after PMN-tachyzoite-exposure (i.e., after a time period, which allowed efficient NET formation) and then used for BUVEC infections. As depicted in Figure 6, the infectivity of PMN-pre-exposed N. caninum tachyzoite was completely restored by DNase I treatment proving that, first, the ensnarement of tachyzoites within NETs hampered a large proportion of tachyzoites from active host cell invasion, and, second, that NETs had no lethal effects on tachyzoites of N. caninum within a period of 3 h.