Simultaneous and Positively Correlated NET Formation and Autophagy in Besnoitia besnoiti Tachyzoite-Exposed Bovine Polymorphonuclear Neutrophils

Given that B. besnoiti tachyzoites infect host endothelial cells of vessels in vivo, they become potential targets for professional phagocytes [e.g., polymorphonuclear neutrophils (PMN)] when in search for adequate host cells or in case of host cell lysis. It was recently reported that B. besnoiti-tachyzoites can efficiently be trapped by neutrophil extracellular traps (NETs) released by bovine PMN. So far, the potential role of autophagy in parasite-triggered NET formation is unclear. Thus, we here analyzed autophagosome formation and activation of AMP-activated protein kinase α (AMPKα) in potentially NET-forming innate leukocytes being exposed to B. besnoiti tachyzoites. Blood was collected from healthy adult dairy cows, and bovine PMN were isolated via density gradient centrifugation. Scanning electron microscopy confirmed PMN to undergo NET formation upon contact with B. besnoiti tachyzoites. Nuclear area expansion (NAE) analysis and cell-free and anchored NETs quantification were performed in B. besnoiti-induced NET formation. Interestingly, tachyzoites of B. besnoiti additionally induced LC3B-related autophagosome formation in parallel to NET formation in bovine PMN. Notably, both rapamycin- and wortmannin-treatments failed to influence B. besnoiti-triggered NET formation and autophagosome formation. Also, isolated NETs fail to induce autophagy suggesting independence between both cellular processes. Finally, enhanced phosphorylation of AMP activated kinase α (AMPKα), a key regulator molecule of autophagy, was observed within the first minutes of interaction in tachyzoite-exposed PMN thereby emphasizing that B. besnoiti-triggered NET formation indeed occurs in parallel to autophagy.


INTRODUCTION
Besnoitia besnoiti is a cyst-forming apicomplexan protozoan parasite that causes bovine besnoitiosis which is traditionally endemic in Africa and Asia. Recent continuous reports on bovine besnoitiosis outbreaks in several European countries (1-9) indicated a re-emergence and spread of this disease in Europe (10) and led to the classification as emerging disease by the European Food Safety Authority (EFSA) in 2010. Overall, bovine besnoitiosis has a detrimental impact on both, individual animal welfare (e.g., pain, oedemas, fever, abortion, orchitis, male infertility, severe skin lesions) and cattle industry (losses).
So far, very little data is available on early host innate immune reactions during primary acute B. besnoiti infections in vivo (11) and in vitro (12,13) despite the fact that early host innate defense reactions should be critical for the outcome of the disease. In this sense, PMN play a pivotal role since they are the most abundant leukocyte population in blood and the first ones to be recruited to sites of infection. As reported for other mammalian species, bovine PMN own several efficient effector mechanisms to combat apicomplexan stages, such as phagocytosis (14), production of reactive oxygen species (ROS) (15) and in vitro excretion of antimicrobial peptides. Additionally, the release of neutrophil extracellular traps (NETs) in response to coccidian protozoa was reported (13,(16)(17)(18). NETs are commonly released via a PMN-derived cell death process known as NET formation (19). Suicidal NET formation was described as a NADPH oxidase (NOX)-dependent cellular mechanism which induces the extrusion of DNA and nuclear and cytoplasmic granule enzymes leading to the formation of DNArich networks being decorated with histones and various potent antimicrobial granular effector molecules, such as neutrophil elastase (NE), myeloperoxidase (MPO), lactoferrin, pentraxin, peptidoglycan recognition proteins, or calprotectin (19)(20)(21)(22). A variety of invasive pathogens such as bacteria, virus, fungi, protozoan, and metazoan parasites, might either be immobilized within released sticky NET structures or be killed via local high concentration of antimicrobial histones, peptides, and proteases (16,20,(23)(24)(25).
Autophagy is an essential intracellular degradation system, that recycles cell components as proteins and organelles and it is essential in the cellular response to stress (42). In neutrophils, autophagy has been described in PMN derived from mouse and human (43,44). Interestingly, first evidences suggest that autophagy is necessary and can prime PMN to undergo NET formation (42,45,46). Besides other molecules, autophagy is regulated by the metabolic sensor molecule AMP activated kinase α (AMPKα) and by the mechanistic target of rapamycin (mTOR) (47). The processes of autophagy and NET formation appear to be linked in PMA-activated PMN and in sterile inflammation (44) by a mechanism which seems dependent on mTOR activation (48).
So far, B. besnoiti-mediated NET formation seems to be NOX-, NE-MPO-dependent and capable to efficiently hamper tachyzoites from active host cell invasion (13). On this regard, AMPK is been described as critical molecule of the autophagic process and governing critical functions in PMN as ROS production, chemotaxis and phagocytosis (49,50). Despite this, nothing is known on the role of autophagy or autophagyrelated molecules such as AMPKα in B. besnoiti-triggered NET formation.
Aim of the current study was to analyze the presence of autophagy during B. besnoiti-triggered suicidal NET formation. Therefore, we first confirmed NET formation induction by B. besnoiti tachyzoites and then showed that both, NET formation and autophagy are performed independent of rapamycin (stimulator of autophagy via mTOR binding), wortmannin (inhibitor of PI3K-mediated autophagy), treatments. In addition, we studied the release of extracellular DNA induced by B. besnoitia tachyzoites in presence of the autophagy-related molecules: LY294002 (inhibitor of PI3K-mediated autophagy) parthenolide (NF-κB inhibitor) or WP1130 (ubiquitinase inhibitor). Interestingly, NET formation and autophagosome formation occur simultaneously in tachyzoite-exposed PMN and is accompanied by a rapid phosphorylation of AMPKα.

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

Parasites
All NET formation -related experiments were performed with tachyzoite stage of the apicomplexan parasite B. besnoiti (strain Bb Evora04) which was initially isolated from the field in Portugal as previously reported (13).
For experiments under physiological flow conditions, primary bovine umbilical vein endothelial cells (BUVEC) were isolated according to the method described by Taubert et al. (51). Briefly, umbilical cords retrieved from newborn calves via Sectio caesarea were enriched with 0.9% Hanks balanced salt solution (HBSS)-HEPES buffer (pH 7.4; Gibco) supplemented with 1% penicillin (500 U/ml; Sigma-Aldrich) and streptomycin (500 mg/ml; Sigma-Aldrich) and kept at 4 • C until use. For isolation of host endothelial cells, the lumen of umbilical veins were infused with 0.025% collagenase type II solution (Worthington Biochemical Corporation). Veins were ligated and incubated for 20 min at 37 • C and 5% CO 2 atmosphere. Then, veins were gently massaged and collagenase-cell suspensions were harvested in 50-ml plastic tubes (Nunc) containing 1 ml FCS (Gibco) to inactivate collagenase type II. After two centrifugations (400 × g, 10 min, 4 • C), endothelial cells were resuspended in complete ECGM (endothelial cell growth medium; PromoCell), plated in 25 or 75 cm 2 plastic culture flasks (Greiner) and cultured at 37 • C and 5% CO 2 atmosphere until confluency. For flow assays, BUVEC were grown on Thermanox R coverslips (Nunc) until confluency.

Isolation of Bovine PMN
Healthy adult dairy cows (n = 9) served as blood donors. Animals were bled by puncture of jugular vein and 30 ml blood was collected in 12 ml heparinized sterile plastic tubes (Kabe Labortechnik). Approximately 20 ml of heparinized blood were diluted in 20 ml sterile PBS with 0.02% EDTA (Sigma-Aldrich), layered on top of 12 ml Biocoll R separating solution (density = 1.077 g/l; Biochrom AG) and centrifuged (800 × g, 45 min). After removal of plasma and mononuclear cells, the cell pellet was suspended in 25 ml bi-distilled water and gently mixed during 40 s to lyse erythrocytes. Osmolarity was rapidly restored by adding 4 ml of 10 × Hanks balanced salt solution (Biochrom AG). For complete erythrocyte lysis, this step was repeated twice and PMN were later suspended in sterile RPMI 1640 medium (Sigma-Aldrich). PMN counts were analyzed in a Neubauer haemocytometer. Finally, freshly isolated bovine PMN were allowed to rest at 37 • C and 5% CO 2 atmosphere for 30 min until further use.

Scanning Electron Microscopy (SEM)
Bovine PMN were co-cultured with vital B. besnoiti tachyzoites (ratio 1:4) for 60 min on coverslips (10 mm diameter; Thermo Fisher Scientific) pre-coated with 0.01% poly-L -lysine (Sigma-Aldrich) at 37 • C and 5% CO 2 . After incubation, cells were fixed in 2.5% glutaraldehyde (Merck), post-fixed in 1% osmium tetroxide (Merck), washed in distilled water, dehydrated, critical point dried by CO 2 -treatment and sputtered with gold. Finally, all samples were visualized via a Philips R XL30 scanning electron microscope at the Institute of Anatomy and Cell Biology, Justus Liebig University Giessen, Germany.
For NET visualizing, Sytox Orange (Life Technologies) was used to stain DNA and anti-histone (clone H11-4, 1:1,000; Merck Millipore #MAB3422), anti-NE (AB68672, 1:1,000, Abcam), or anti-MPO (orb11073, 1:1,000, Byorbit) antibodies were used to stain specific proteins on ETs structures. Therefore, fixed samples were washed three times with PBS, blocked with 1% bovine serum albumin (BSA, Sigma-Aldrich) for 30 min at RT and incubated with corresponding primary antibody solutions for 1 h at RT. Thereafter, samples were washed thrice with PBS and incubated in secondary antibody solutions (Alexa Fluor 488 goat anti-mouse IgG or Alexa Fluor 488 goat anti-rabbit IgG, both Life Technologies, 60 min, 1:1,000, RT). Finally, samples were washed thrice in PBS and mounted in anti-fading buffer (ProLong Gold Antifade Mountant; Thermo Fisher Scientific). Visualization was achieved using an inverted IX81 fluorescence microscope equipped with an XM 10 digital camera (Olympus).
NETs are divided into two distinct forms: one is released away from neutrophils named cell-free NETs, and the other is those that are anchored to neutrophils (namely anchored NETs). For "cell free"-and "anchored"-NETs determination according to Tanaka et al. (52), the plate was directly centrifuged at 300 × g for 5 min after incubation. The supernatants were transferred into a new 96-well plate to measure "cell-free"-NETs and the pellets were used for "anchored"-NETs estimation. For both sampling methods, a 1:200 dilution of Pico Green R (Invitrogen) in 10 mM Tris base buffered with 1 mM EDTA was added to each well (50 µl), and then extracellular DNA was detected and quantified by PicoGreen R -derived fluorescence intensities using an automated multiplate reader (Varioskan, Thermo Scientific) at 484 nm excitation/520 nm emission.
Estimation of "Anchored" NETs on B. besnoiti-Infected BUVEC Under Physiological Flow Conditions BUVEC (n = 3) were cultured on Thermanox R (Nunc) coverslips pre-coated with bovine fibronectin (10 µg/ml, 2 h RT; Sigma-Aldrich) until confluency and infected with 2.5 × 10 5 freshly isolated B. besnoiti tachyzoites. Twenty-four h. p. i., coverslips were washed to remove residual tachyzoites and placed into a parallel flow chamber (53). Bovine PMN (2.5 × 10 6 PMN in 500 µl medium) were perfused into the system at a constant wall shear stress of 1.0 dyn/cm 2 (syringe pump sp100i R ; World Precision Instruments). For "anchored"-NET formation visualization, coverslips were fixed, washed, and stained for DNA and histones as described above. Images were taken under an inverted fluorescence microscope (DM IRB; Leica) equipped with a digital camera (Olympus).

Nuclear Decondensation-Based Quantification of NETs
Nuclear expansion-based quantification of NETs was performed according to the method described by Papayannopoulos et al. (54). Briefly, bovine PMN (n = 3) were pretreated with rapamycin (50 nM), wortmannin (50 nM) or plain medium (RPMI 1640, Sigma-Aldrich) for 30 min and then exposed to B. besnoiti tachyzoites for 3 h at a 1:4 PMN:tachyzoites ratio. After incubation, PMN were fixed by 4% paraformaldehyde (Merck) and stained with 5 µM Sytox Orange R (Life Technologies) for 30 min at RT. Five images were captured randomly for each condition using an inverted fluorescence microscope (Olympus IX 81) and nuclear area size of single cells was analyzed using ImageJ R software as described by Gonzalez et al. (55). Cells that presented decondensed nucleus and exceeded the threshold of 50 µm 2 were considered as PMN undergoing NET formation. Overall, 1,200-1,700 PMN were analyzed for each experimental condition in samples from three different donors.

Autophagosome Detection by Immunofluorescence Analysis
LC3 protein is a marker for autophagosomes (56) with LC3-I being cytosolic and LC3-II being membrane-bound and enriched in the autophagic vacuole. Analysis of autophagosome formation in PMN was performed according to Itakura and McCarty (48). In brief, bovine PMN (n = 3) were deposited on poly-L -lysine (0.01%) pre-treated coverslips (15 mm diameter, Thermo-Fisher scientific), pretreated with rapamycin (50 nM) or wortmannin (50 nM) for 30 min before being exposed to B. besnoiti tachyzoites at a 1:4 PMN:tachyzoite ratio for 3 h. After incubation, cells were fixed with 4% paraformaldehyde (10 min), permeabilized by ice cold methanol treatment (3 min at 4 • C) and blocked with blocking buffer (5% BSA, 0.1% Triton X-100 in sterile PBS; all Sigma-Aldrich) for 60 min at RT. Thereafter, cells were incubated overnight at 4 • C in anti-LC3B antibody solution (cat#2775 Cell Signaling Technology) diluted 1:200 in blocking buffer. After incubation, samples were washed thrice with PBS and incubated 30 min in the dark and RT in a 1:500 dilution of goat anti-rabbit IgG conjugated with Alexa Fluor 488 (Invitrogen). After three washes in PBS, samples were mounted in Prolong Anti-fading reagent with DAPI R (Invitrogen) on glass slides and images were taken applying confocal microscopy (Zeiss LSM 710). To estimate LC3B-positive cells, the background fluorescence signal was determined in control conditions for FITC (green) and DAPI (blue) channels. Image processing was carried out with Fiji ImageJ R using Z-project and merged channel plugins and restricted to overall adjustment of brightness and contrast.

Statistical Analysis
Statistical significance was defined by a p value <0.05. p value were determined by applying non-parametric analyses: Mann-Whitney test when two experimental conditions were compared and Kruskal-Wallis test followed by Dunn's post-hoc test for multiple comparisons. Correlation between LC3B-positive and NETotic PMN was determined by Spearman correlation test. All graphs (mean ± SD) and statistical analyses were generated by the use of Graph Pad software (v. 7.03).

Visualization of B. besnoiti-Triggered NETs in Bovine PMN
SEM analysis showed that bovine PMN (Figure 1) exposed to vital B. besnoiti tachyzoites released NET-like structures and many B. besnoiti tachyzoites were firmly trapped by these filaments (Figures 1A-C). To verify that bovine PMN were indeed undergoing NET formation, the main components of NETs [i.e., histones (H11-4) and NE] were visualized by immunostaining. Co-localization analyses of extracellular DNA being adorned with H11-4, NE (Figures 1D-I) in parasiteentrapping structures confirmed classical characteristics of NETs. Furthermore, tachyzoites were entangled in these NETs structures confirming the observations in SEM analysis (Figures 1D,I; Control condition is shown in Figures S6 and S7).
Given that B. besnoitia tachyzoites develop within endothelial host cells, we wondered whether infected endothelium would also contribute to NET formation. Therefore, we chose an experimental approach which mimicked the in vivo situation in a small vessel: controlled physiological flow condition of 1.0 dyn/cm 2 shear stress was applied on endothelial cell layers and a fixed number of PMN were floating over B. besnoiti-infected endothelial cells in parallel plate chamber assay (53). Under these flow conditions, bovine PMN also underwent "anchored"-NET formation on B. besnoiti-infected BUVEC which was also corroborated by co-localization of extracellular DNA decorated with histones (Figure 2).

Effects of Autophagy on B. besnoiti-Stimulated NET Formation in Bovine PMN
To investigate effects of autophagy on B. besnoiti-triggered bovine NET formation, initially we used the mTOR-mediated autophagy inducer rapamycin (48) and the PIK3-mediated autophagy inhibitor wortmannin (57).
In a first experimental series, NET formation was measured based on PicoGreen R -derived fluorescence intensities as previously described (18,19) thereby rather targeting late phase of NETosis. Overall, confrontation of bovine PMN with B. besnoiti tachyzoites resulted in a significant increase of NET formation when compared to control groups (p = 0.02-0.03; Figures 3A-C). Given that we always experience high individual variations in NET-related assays, the reactions induced in PMN derived from each animal (n = 9) are also depicted ( Figure 3A). However, parasite-mediated NET formation was neither affected by rapamycin (tested in a range from 10 to 200 nM, Figure 3B) nor by wortmannin treatments (tested in the same range of concentration, Figure 3C). This lack of effect was also observed when non-stimulated PMN were treated with these compounds (100 nM) for control purposes (Figure 3D).
In a second series of experiments, an alternative method of NET quantification was applied which allowed for "cell free"and "anchored"-NETs distinction by following the methodology described by Tanaka et al. (52). Overall, zymosan treatment which was used for positive control resulted in a highly significant increase of both types of NETs, i.e., "cell free"-and "anchored"-NETs in bovine PMN (p < 0.0001; Figures 3E-H). In addition, confrontation of PMN with B. besnoiti tachzoites in principle also triggered both kinds of NETs as seen for "anchored"-NETs in the rapamycin-related (p = 0.008; Figure 3E) and the wortmanninrelated data set (p = 0.007; Figure 3G) and for "cell free"-NETs in the wortmannin-related dataset (p = 0.02; Figure 3H). Interestingly, the induction of "anchored"-NETs by parasite stages was more evident than the induction of "cell free"-NETs. In agreement with the data mentioned above, neither rapamycin nor wortmannin treatments led to altered parasite-triggered "cell free"-nor "anchored" NET formation (Figures 3E-H).
Since autophagy is a very complex process, and various signaling pathways are involved in autophagosome formation, we used more pharmacological regulatory factors to check if autophagy affects NET formation via appropriate pathways. In our experimental setting, LY294002 (PI3K inhibitor) and pathenolide (NF-κB inhibitor) nor WP1130 (deubiquitinase inhibitor) treatments did not alter anchored and cell free NET formation (Figure 3I,J).
In a third series of experiments, we chose to analyze parasitetriggered NET formation based on nuclear area expansion (NAE). In general, PMN undergo several stages of NET formation including NE-and MPO-dependent chromatin decondensation (54). Decondensed PMN nuclei are considered as a marker for early "NETotic" processes (55). In order to determine if B. besnoiti induces NAE in bovine PMN, as well as to estimate if rapamycin or wortmannin influences this parameter of NET formation, 1,200-1,700 cells were individually analyzed per condition (Figures 4C-H) and data illustrated via frequency histograms ( Figure 4I) and percentage of cells undergoing early NET formation ( Figure 4J). In agreement to data on the later phase of NET formation (Figure 3), confrontation with B. besnoiti tachyzoites significantly induced nuclear area expansion in a higher percentage of bovine PMN thereby indicating early NET formation processes (p = 0.03); illustrated in Figures 4B,F, data sets in Figures 4I,J, control condition is shown in Figures 4A,C. Interestingly, we also observed an increase of PMN populations showing NAE in case of rapamycin treatments (inducer of autophagy) of tachyzoite-exposed PMN, however, these reactions showed no significance in relation to untreated controls (PMN + B. besnoiti) due to the high individual variation of the donors already mentioned above ( Figure 4J). Given that significant differences were indeed detected referring to rapamycin-treated tachyzoite-exposed groups and to parasitefree rapamycin controls (p = 0.04), an influence of rapamycin and therefore of autophagy on parasite-triggered NET formation may be stated in that sense that induction of autophagy leads to enhanced early NET formation. The fact that this effect could not be detected by the other methods of NET quantification used before (see Figure 3) may be due to the targeted early/late phase of NET formation.

Effects of B. besnoiti Tachyzoite Exposure on Autophagosome Formation in PMN
Given that data on NAE indicated a link between autophagy and parasite-triggered early NET formation, we additionally analyzed the effect of tachyzoite exposure on PMN-derived autophagosome formation. During autophagy, the cytosolic form of microtubule-associated protein 1A/1B-light chain 3 (LC3-I) is conjugated to phosphatidylethanolamine to form LC3-II which allows LC3 to become associated with autophagic vesicles (58). We therefore used an antibody directed against the splice variant LC3B as a marker to investigate the effect of tachyzoite exposure to PMN-derived autophagy and of autophagy on B. besnoiti-induced bovine NET formation. Overall, exposure to B. besnoiti tachyzoites led to significant autophagosome formation in exposed bovine PMN (p = 0.01); for illustration, see Figures 5A,B; triangles, for data see Figure 5D. Notably, cells undergoing autophagy also showed NET formation against B. besnoiti tachyzoites, which were firmly entrapped in DAPI-labeled chromatin structures (illustrated in Figure 5B, arrows, more images are shown in Figure S3). Thus, we analyzed LC3B expression in tachyzoites-exposed PMN and in rapamycin and wortmannin pretreated PMN ( Figure 5C). The percentage of LC3B positive cells was increased in the B. besnoitia exposed condition but was not affected by rapamycin or wortmannin ( Figure 5D). However, an increase of 8% was observed in PMA-treated PMN that were incubated previously with rapamycin Figure S1. Finally we analyzed the data by Spearman correlation test indicating a positive correlation of NET formation and autophagy in tachyzoiteexposed PMN (Figure 5D). In a different approach we evaluated autophagosome formation in PMN confronted with isolated NETs obtained as described (59). Isolated NETs failed to induce autophagosome formation. Figures and analysis are shown in Figure S2.

B. besnoiti Tachyzoite Exposure Induces LC3B and p-AMPKα Protein Expression in Bovine PMN
Given that exposure to B. besnoiti tachyzoites induces autophagy in bovine PMN, we here analyzed the protein expression of LC3B and of phosphorylated AMPKα as a key regulator molecule of autophagy in a kinetic experiment (protein extracts of tachyzoiteexposed PMN isolated after 5, 30, 60, and 180 min of co-culture). For control reasons, protein extracts from pure B. besnoiti tachyzoites (three different isolates, 60 min of incubation) were included in the assays (Figure 6). Immunoblotting analyses revealed a non-statistically significant increase in LC3BII expression in B. besnoiti tachyzoite-confronted PMN when compared non-exposed PMN (Figure 6). Atg 5, which plays a role in the formation and elongation of autophagosomes does not show a detectable difference in protein expression when PMN were confronted with B. besnoiti tachyzoites (Figure S4). In addition, immunoblotting experiments revealed a very distinct expression profile on p-AMPKα. Here we used an antibody that is specific for AMPK that showed phosphorylation in the alpha subunit and thereby reflected AMPK activation. Whilst neither pure tachyzoites nor PMN alone showed any signal of this molecule, AMPKα expression was clearly induced by tachyzoite exposure of PMN thereby showing a fast response pattern with enhanced expression only within the first 30 min of contact (Figure 7, lower panel). In all samples, two distinct protein bands at the level of the expected size (∼62 and 50 kDa) were observed in AMPKα-positive Immunoblots, which most probably represents parts of a cleaved form of AMPK, a common process occurring in leukocytes (60). For AMPK, a transient pattern was observed in both, B. besnoiti confronted and non-confronted PMN (Figure 7, middle panel) showing peaks of expression at 30 and 180 min. For comparison, p-AMPKα band densitometry was obtained and normalized by vinculin signal, reinforcing the clear effect over AMPK phosphorylation showed in immunoblots. Due to the lack of signal of AMPK in animal 3, the normally used ratio of pAMPK/AMPK was not possible to apply for all the donors; however the graph corresponding to this ratio is shown in the Figure S5.

DISCUSSION
The initial description of NET formation was released in 2004 and was followed by the discovery of a novel programmed cell death pathway nowadays known as ETosis (19,20,61). ETosis comprises a unique series of cellular events by which FIGURE 3 | Effects of rapamycin and wortmannin treatments on Besnoitia besnoitia-triggered NET formation as detected by PicoGreen-based quantification. Bovine PMN were pre-treated with increasing concentrations of rapamycin and wortmannin for 30 min and then exposed to B. besnoiti tachyzoites for 3 h. After incubation, the samples were analyzed as follows. (A-D) samples were treated with micrococcal nuclease for 15 min, thereafter centrifuged and extracellular DNA present in the supernatants were analyzed using PicoGreen-based quantification. (E-H) samples were centrifuged immediately and the supernatants were analyzed for "cell free" NETs. Moreover, remaining pellets located at the bottom were investigated for "anchored" NETs. For both approaches, extracellular DNA was detected and quantified via PicoGreen-derived fluorescence intensities. In all assays, zymosan treatments (1 mg/ml) were used for positive controls. Statistical analyses were performed by Kruskal-Wallis test followed by Dunn's test for multiple comparisons. p-values are indicated in the graphs (Mean ± SD).
Concerning stimuli, both chemicals and pathogens can trigger ETosis (74,75). Several reports have demonstrated that ETosis is efficiently induced by protozoan and metazoan parasites, such as T. gondii, E. bovis, Cryptosporidium parvum, N. caninum, Dirofilaria immitis, Haemonchus contortus, and Schistosoma japonicum (18,23,34,37,76,77). Recently, Muñoz-Caro et al. (13) confirmed B. besnoiti tachyzoites as potent NET inducers by quantification of extracellular DNA using Picogreen R . In the current study, we aimed to take a more detailed look on B. besnoiti-triggered NET formation and to dissect these immune reactions into different NET types ("anchored"/"cellfree" NETs) and time-dependent reactions (early phase of NET formation vs. finalized NETs) by using different methodical approaches. Nevertheless, as proof of principle, ETosis events were first confirmed by SEM analysis in B. besnoiti tachyzoitesexposed bovine PMN. Immunostaining analyses revealed that these extracellular structures were mainly composed of DNA being decorated with histones, MPO and NE, thereby confirming classical components of ETs.
As also performed in the current study and leading to the confirmation of tachyzoites as NET inducers, NET formation can be estimated by quantification of extracellular DNA using specific probes as Picogreen R (19) following DNA digestion in cell culture microplates. The advantage of this approach is obvious since several compounds can be tested in the same experiment, however it is advised by several authors that NET quantification should always be confirmed by microscopy (55,78). In the current study we broadened the methodical panel to dissect between early events during NET formation by analyzing nuclear decondensation of PMN and late events by estimating the formation of "cell free" and "anchored" NETs. Overall, exposure of PMN to B. besnoiti tachyzoites led to a significant induction of nuclear decondensation and of "anchored"-NETs. Interestingly, the data were more consistent in terms of magnitude of the response in the formation of "anchored" (1.7-fold increase) than "cell free" NETs (1.4-fold increase). Thus, it appears that tachyzoites mainly induce "anchored" NETs. Of note was the inter-donor variation concerning NET formation quantification using this technique. However, this observation is in line with high interdonor variations in terms of quantity and quality of NETs induced by soluble mediators such as PMA and A23187 (59,79). In a further experimental approach using B. besnoitiinfected primary endothelium in a parallel plate flow chamber, FIGURE 5 | Autophagy and NET formation occurs simultaneously in B. besnoiti-exposed PMN. Bovine PMN were exposed to B. besnoiti tachyzoites for 3 h. Samples were fixed and permeabilized for LC3B-based immunostaining to determine autophagosome formation by confocal microscopy. we could confirm the formation of "anchored"-NETs under physiological flow conditions by co-localization of extracellular DNA with histones. This confirms a potential cross-talk between activated endothelium and leukocytes to be recruited to site of infection. Whether these extruded "anchored"-NETs might impact on intracellular located B. besnoiti tachyzoites needs further investigation. However, this observation might have implications on the outcome of cattle besnoitiosis since the tachyzoite stages are indeed infecting endothelium of vessels in vivo (10,12).
Autophagy is a physiological process within the body which maintains homeostasis or normal function of cells by protein degradation and turnover of destroyed cell organelles for new cell formation after cellular stress (80). Furthermore, autophagy has been shown to play a pivotal role in regulating early innate leukocyte-associated effector mechanisms against pathogens, such as phagocytosis (81), cytokine secretion (82), and NET formation (46). In this regards, mTOR pathway plays a key role in NET formation via regulation of autophagy pathways (48). Thus, Park et al. (50) showed using a Sytox Green-related assay FIGURE 6 | LC3B expression in B. besnoiti tachyzoite-exposed PMN. Protein extracts of B. besnoiti tachyzoite-exposed PMN (n = 3) were performed at different time points after exposure and analyzed for LC3B-I and LC3B-II expression by immunoblotting. The expression of vinculin was controlled as an internal reference protein.
for NET quantification, that rapamycin pretreatments primed human PMN enhancing NET formation in response to PMA. This reaction was reversed by a panel of different autophagy inhibitors. In this set-up, rapamycin treatments alone did not influence NET formation. In our experimental set-up using B. besnoiti tachyzoites instead of PMA, rapamycin treatments did not influence the degree of parasite-triggered bovine NET formation when using PicoGreen-based analyses on total NETs and "anchored"/"cell-free" NETs. In addition, treatments with the autophagy-inhibitor wortmannin failed to affect parasitetriggered NET formation. This observation was complemented with the use of the PI3K inhibitor LY294002, observing a nonsignificant decrease of "anchored" NETs formation. Same result was observed by the use of NF-κB inhibitor. However, when estimating early NET formation via nuclear area expansion, (NAE) analysis, we found that rapamycin pretreatments indeed primed bovine PMN for enhanced NET formation in response to tachyzoites. The discrepancy between the different methods of NET detection may be explained by two factors: first, autophagy appears to precede NET formation and may therefore rather be linked to early NET formation that to late NET-related effects, and secondly, the NAE-based assay appeared more sensitive for the detection of tachyzoite-triggered NET formation and may therefore have produced an improved resolution of the data. We therefore assume that early tachyzoite-triggered NET formation is indeed linked to autophagy in bovine PMN. Furthermore, the fact that formation of LC3B-positive autophagosomes was observed in bovine PMN while casting NETs supported the potential role of autophagy in PMN-derived responses against tachyzoite stages.
Autophagosomes are double-membraned vesicles formed during autophagy, which represent characteristic markers of autophagy. LC3 is a small, soluble protein, which is distributed ubiquitously in mammalian tissues and in cultured cells. During autophagy, LC3-I (a cytosolic form of LC3) is conjugated to phosphatidylethanolamine to form LC3-II, which is then recruited to autophagosomal membranes (58). Therefore, LC3-II is widely used as a marker for the microscopic detection of intracellular autophagosomes. The LC3 gene family comprises three members, LC3A, LC3B and LC3C, and LC3B represents the most used endogenous autophagic marker (58). Certain studies have revealed that autophagy is required for NET formation FIGURE 7 | B. besnoiti induces the phosphorylation of AMPK in bovine PMN. Protein extracts of 5 × 10 6 PMN confronted with 20 × 10 6 B. besnoiti tachyzoites from three different donors were obtained at different time points and the kinetics were analyzed for LC3B-I and LC3B-II (upper panel) and AMPKα T172 phosphorylation (middle panel) through western blot. For AMPK, the activation of the enzyme was detected in the first 30 min of interaction (middle panel). Vinculin detection via western blot was performed for the three donors as a protein quantity loading controls. (46,83) and that autophagy induction triggers NET formation (45,46,83). To detect autophagy in B. besnoiti tachyzoiteexposed PMN as a matter of principle, autophagosome formation was visualized by LC3B-based immunostaining. Confocal microscopy clearly showed that confrontation of PMN with B. besnoiti tachyzoites indeed caused a significant increase of autophagosome formation. As a highly interesting finding, we additionally observed that autophagic PMN also performed NET formation. However, neither rapamycin nor wortmannin pre-treatments had any influence on PMNderived autophagosome formation, reinforcing the observation that these processes were mTOR-independent. Our results are in line with those obtained in human PMN, where rapamycin by its own is not able to induce autophagy but increases the autophagosome formation induced by PMA (48). In line, mTOR-independent induction of autophagy was also reported in a distinct population of PMN from sepsis patients which showed increased PMA-triggered NET formation activity (45).
AMPKα is a key metabolic master regulator in eukaryotes with high impact on several important cellular mechanisms. AMPKα activation is initiated by changes in the metabolic status which result from inhibition of ATP generation during hypoxia, glucose deprivation and increased ATP consumption (84). Previous observations in PMN showed that AMPK activation decreased PMA-induced ROS production in human PMN (49), but enhanced PMN chemotaxis, bacterial killing, and phagocytosis (50). Moreover, AMPK promotes autophagy by directly activating Ulk1 which is a mTOR downstream enzyme during autophagosome formation (85). On the other hand, inhibition of AMPK in mice model induced histone 3 secretion, suggesting that AMPK activation contributed to murine NET formation (86). Since autophagy is a complex process and could be initiated via various signaling pathways, we tried to check if AMPK pathway is involved in B. besnoiti tachyzoite-induced autophagy. Our current data show that confrontation of PMN with B. besnoiti tachyzoites clearly induced AMPKα activation in a time-dependent manner. Thus, AMPKα phosphorylation was immediately induced from the very beginning of parasite-PMN interactions until 30 min of co-culture. So far, it is unclear if enhanced AMPKα activation is linked to autophagy or NET formation or both in tachyzoite-exposed neutrophils, but this will be a matter for further research.

ETHICS STATEMENT
This study was conducted in accordance to Justus Liebig University Giessen Animal Care Committee Guidelines.

ACKNOWLEDGMENTS
The authors would like to acknowledge Anika Seipp, Institute of Anatomy and Cell Biology, JLU Giessen, Germany, for her technical support in scanning electron microscopy analyses. To Hannah Salecker for the technical support in B. besnoitia cell culture and Western blots. We further thank all staff members of JLU Gießen teaching and research station Oberer Hardthof.

SUPPLEMENTARY MATERIAL
The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fimmu. 2019.01131/full#supplementary-material Figure S1 | Autophagy induction in bovine PMN. Bovine PMN were treated with PMA (20 nM) and rapamycin (100 nM) for 1 h, and bovine PMN treated with media was used as control. Samples were fixed and permeabilized with pure cold methanol for LC3B-based immunostaining to determine autophagosome formation by confocal microscopy. (A,B) showing merged images with the staining for LC3B (green), DAPI (blue) in the control group. (C,D) showing merged images with the staining for LC3B (green), DAPI (blue) in the treated group. The right graph shows the percentage of autophagosome-positive cells (E). Figure S2 | Isolated NETs fails to induce autophagosome formation in bovine PMN. Isolated NETs from B. besnoiti-confronted PMN were isolated as described by Barrientos (59). Bovine PMN were treated with isolated NETs for 1 h, and bovine PMN treated with media was used as control. Samples were fixed and permeabilized with pure cold methanol for LC3B-based immunostaining to determine autophagosome formation by confocal microscopy.