Blood samples were drawn from healthy Holstein Friesian cows of a commercial dairy farm located in the Basque Country. The herd is enrolled in the national bTB eradication program. Animals on the farm tested negative for bTb by skin-test over the last five years. All animals used in this study (n=4) tested negative for Map shedding in feces, as confirmed by PCR, and were negative for Map-specific antibodies to Map, as determined by PPA-3 ELISA.

Animals used in this study were submitted only to procedures that according to European (Directive 2010/63/EU of the European Parliament and of the Council of 22 September 2010 on the protection of animals used for scientific purposes. Chapter 1, Article 1, Section 5, paragraphs b and f) and Spanish (Real Decreto 53/2013, de 1 de febrero, por el que se establecen las normas básicas aplicables para la protección de los animales utilizados en experimentación y otros fines científicos, incluyendo la docencia, Article 2, Section 5, Paragraphs b and f) legislation on experimental animals are exempt from its application. The animals, belonging to a registered commercial farm supervised by the local livestock authority (Servicio de Ganadería de la Diputación Foral de Bizkaia) were submitted only to the introduction of a needle in accordance with good veterinary practice and were not killed in relationship with this study.

Blood was taken from the jugular vein using a 16G x 11/2 hypodermic needle into a blood-collection bag containing 63 mL of citrate phosphate dextrose adenine (CPDA; TerumoBCT Teruflex) for a total capacity of 450 mL, which was used for monocyte isolation. One month later, 40 mL of blood from the same cows were collected with a 18G hypodermic needle into a 50 mL tube containing 7 mL of CPDA for PMN isolation. Blood samples were processed within two hours of the extraction.

Peripheral blood mononuclear cells (PBMCs) were separated using Histopaque 1077^® and monocytes were selected using mouse anti-human CD14-coupled microbeads and MS columns (Miltenyi) following the manufacturer’s recommendations. Greater than 90% of purity of monocytes was achieved as determined by flow cytometry using a mouse anti-bovine CD14 FITC-labelled antibody (BioRad).

For the production of monocyte-derived macrophages (MDMs), monocytes were plated at 1x10⁵ cells/well for co-cultures and at 2x10⁵ cells/well for MDM cultures in wells of 96-well plates for fluorimetric assays to quantify ET release. For immunofluorescence microscopy, monocytes were plated at 1x10⁵ cells/well for co-cultures and at 2x10⁵ cells/well for MDM cultures on 16-well chamber slides (178599PK Thermo Scientific). For contact independent cultures, monocytes were plated at 5x10⁵ cells/well for co-cultures and at 1x10⁶ cells/well in wells of 24-well plates. During maturation to MΦs, monocytes were cultured in RPMI 1640 media without phenol red (Gibco 11835030), supplemented with 2mM L-glutamine, 10% fetal bovine serum (FBS), 1% penicillin-streptomycin and 10 ng/mL recombinant bovine M-CSF (Kingfisher Biotech, Inc. RP1353B) at 37°C and 5% CO2. The media was replaced every three days and prior to the beginning of the mycobacterial infection assays. After 7 days of culture, MΦ morphology was confirmed by light microscopy (increased size, increased adherence, cytoplasmic granularity and presence of pseudopods).

A week after PBMC isolation, 25 mL of blood were drawn from the same cows by venepuncture and mixed with 25 mL of PBS supplemented with 2% FBS. 25 mL of the mixture were layered on top of 15 mL of Histopaque^® 1077 (Sigma) and centrifuged at 1200 x g for 45 min at RT with no brakes. PMNs were located at the lower end of the tube and were subjected to two flash hypotonic lysis using 25 mL of sterile distilled water during 20-30 sec, followed by 2.5 mL of a 10% NaCl solution and topped up with PBS-2% FBS, in order to eliminate remaining red blood cells. PMNs were then resuspended in RPMI 1640 with glutamine and without phenol red and counted. Over 90% purity was achieved as determined by FSC and SSC gating and visualization of characteristic nuclei after DAPI staining under a fluorescence microscope.

Map strains K10, Map K10-GFP (kindly provided by Dr. Jeroen de Buck, University of Calgary (32) M.bovis BCG Danish strain 1331 and M. bovis BCG Danish strain 1331-GFP were used in the assays. All strains were grown to exponential phase at 37°C under aerobic conditions. Map-GFP and BCG-GFP were cultured for three weeks on Middlebrook 7H9 OADC with Kanamycin (25μg/mL) to select for GFP-plasmid carriers and supplementation with Mycobactin J was added for Map. Map strain K10 and BCG Danish strain 1331 were used for ET isolation and quantification and were cultured in the same medium without kanamycin. Cultures were adjusted after reaching an OD of 0.7 (3x10⁸ bacteria/mL) and colony forming units (CFU) were confirmed on agar after using 10-fold serial dilutions of bacterial suspensions. In experiments comparing the response to live and killed bacteria, heat-inactivation was performed by heating bacteria for 30 min at 85°C. Inactivation of bacteria was confirmed by plating on agar plates.

PMNs and MDMs isolated from the same cow were cultured separately or co-cultured together in the same well (CC; see Figures 1A and 3A for details). In order to study cellular changes following paracrine signaling in the absence of cell-to-cell contact between PMNs and MDMs as well as to determine ET formation, a transwell co-culture system was also employed. PMNs were seeded in the transwell insert (pore diameter 0.4 μm; Thermo Fisher Scientific 140620) to generate insert PMNs (TW-PMN), whereas MDMs were seeded into the lower well (TW-MDM). To assess the need for direct cell contact, conditioned MDM media (TW-CM) was produced by incubating MDMs with Map, BCG and Zymosan for 4h at 37°C and 5% CO₂, which was subsequently added to the lower well underneath the insert containing PMNs (TW-PMN-CM). ET release quantification and mycobacteria killing were performed on 24 and 96 well plates. ET visualization was performed on 16-well chamber slides (Nunc Labtek).

Once cell cultures were set, these were stimulated with live or heat-inactivated Map K10-GFP, M. bovis BCG -GFP, Map K10, M. bovis BCG at either a MOI of 1 (ET release) or a MOI of 5 (survival assay), zymosan (1mg/mL; Sigma Z4250), or left untreated (negative control). The original bacterial strains with no GFP were used for the isolation and quantification of NETs in the fluorometric assay to avoid GFP fluorescence interference.

Cultures were incubated at 37°C and 5% CO₂ for 4h for ET release evaluation and for 24h for bacteria killing and cytokine level quantification.

ET release was assayed as described by Köckritz-Blickwede et al. (33) with some modifications. Briefly, ETs generated by cells were digested with 500 mU/mL of micrococcal nuclease (10107921001 Roche) for 10 mins at 37°C 5% CO2. The nuclease activity was stopped by the addition of 5 mM EDTA; thereafter, culture supernatants were collected and stored at 4°C overnight. Total DNA was extracted from naive cells with DNeasy^® blood & tissue kit (Qiagen) following the manufacturer’s instructions. Extracted DNA was solubilized in TE buffer. Both ETs and genomic DNA was quantified using Quant-iT™ PicoGreen™ assay (Thermofisher) according to the manufacturer´s instructions. Plates were read in a fluorescence microplate reader (Synery HTX, Biotek) with filter settings at 488 nm excitation and 520 nm emission. The percentage of DNA released as ET-DNA was calculated by dividing the amount of quantified ET-DNA by the total amount of genomic DNA.

To assess histone staining as a marker for ET formation, cells were seeded into 16-well chamber slides for subsequent antibody staining and microscopy according to Conejeros et al. (34). After 4 h of incubation with either Map-GFP, BCG-GFP or zymosan, cells were fixed for 15 min with 4% formaldehyde. Cells were washed twice with PBS and kept in PBS at 4°C until staining was performed. Subsequently, cells were permeabilized for 15 min with PBS 0.1% Triton X-100, blocked with 1% goat serum containing 0.05% Tween 20 and 3% BSA-PBS. The immunofluorescence staining was performed overnight at 4°C using a mouse anti-pan-histone antibody (Merk MAB3422; diluted 1:200). Thereafter, cells were washed three-times with PBS and incubated with anti-mouse Alexa Fluor 594 labelled antibody (Invitrogen A-1105, diluted 1:500) for 30 min at RT. Following further three washes with PBS, wells were detached from the slides and mounting medium containing DAPI was added before coverslips were added. Preparations were let dry and observed at 400x magnification on a DMi8 fluorescence inverted microscope (LEICA^®). Pictures were taken using a DFC3000 G camera coupled to the microscope.

Quantification was performed by taking snapshots of five fields containing visible ETs with histone staining. Analysis was performed using the Image J software package. Snapshots in 8-bit format were analyzed using the following pipeline. Threshold was adjusted depending on the micrograph, being 4 for low level and 255 for high level. The percentage of picture area occupied by ETs plus nuclei (total DNA) was measured by adjusting the settings for fluorescent particle size: 0-infinity and circularity: 0-1. The percentage of picture area occupied by nuclei were analyzed adjusting the settings for particle size: 200-5000 and circularity: 0.3-1. ET percentage was calculated as total DNA percentage minus nuclei percentage. In the case of cultures that did not contain ETs, nuclei were measured, and ET release was considered zero.

PMNs, MDMs and co-cultures inoculated with BCG -GFP and Map-GFP were prepared. At time 0 (4°C control) and 24 h, the supernatants were removed and centrifuged to pellet non-internalized mycobacteria. Supernatants were stored for cytokine analysis. Wells were washed twice with PBS to remove the remaining non-internalized bacteria. Bacterial pellets and washes of remaining bacteria were centrifuged together, and the final pellets were resuspended in 100 μL of filtered PBS. Adherent cells were lysed by vigorous pipetting with 0.5 mL of 0.1% Triton X-100 (Sigma-Aldrich) in sterile water for 10 min at RT.

Two serial 10-fold dilutions of each lysed sample and supernatant were performed in a total volume of 1 mL using filtered PBS. 200 μl of each dilution was inoculated in duplicate in Middlebrook 7H9 OADC-Kanamycin (25μg/mL) agar plates, supplemented with mycobactin J in the case of Map cultures. Seeded agar plates were allowed to dry at RT until humidity was no longer visible. Plates were sealed with tape to avoid desiccation and incubated at 37 ± 1°C for 6 weeks. CFU were counted and 0h (4°C control) CFU were considered total inoculated bacteria, 24h CFU supernatant bacteria were considered non-internalized and non-killed and 24h CFU cell bacteria were considered internalized and non-killed bacteria. Killed bacteria were estimated as total inoculated bacteria minus (24h CFU supernatant bacteria plus 24h CFU cell bacteria). CFU counts were multiplied by the inverse dilution factor and by the seeded volume. The mean of duplicate plates was calculated and the group mean for each culture type was calculated. Supernatants’ fraction and cell-lysis’ fractions were divided by the 100% survival control to calculate the survival percentage of each portion and the killing rate.

Commercially available direct ELISA kits were performed to determine bovine IL-1β (ESS0027 Invitrogen^®) and IL-8 (3114-1A-6 MABTECH) release in 24h culture supernatants using the provided manufacturer’s instructions. IL-1β ELISA is based on streptavidin-HRP and IL-8 on streptavidin-ALP. Absorbances were measured using an automated ELISA plate reader (Multiskan EX^®, Thermo Lab Systems, Finland) and in all cases standard curves were used to determine each cytokine amount in the supernatant samples.

Data were assessed for normal distribution and homoscedasticity using Shapiro-Wilk and Bartlett´s test, respectively. All data are presented as the mean +/- standard deviation (SD), with samples being analyzed in triplicates (ET release by fluorimetry) or duplicates (bacterial killing and cytokine release). For ET release, percentage differences between values obtained for the mean of each culture type were calculated using pairwise comparisons (t-tests) with pooled SD for all the stimuli except for values obtained for unstimulated and zymosan stimulated cultures. Here, values were not normally distributed and thus comparisons were calculated using non-pooled SD test.

The number of bacteria (expressed as CFU) in the inoculum (0 h) was considered as 100% viable bacteria. The number of bacteria being alive after 24 h incubation was calculated as the sum of internalized mycobacteria and mycobacteria being present in the supernatant, with the killing rate being the difference between the bacteria at 0h and the survival at 24 h, expressed as percentage. Mean values obtained for killing rates under different conditions were compared by multiple pairwise comparisons using t tests with non-pooled SD with Benjamini-Hochberg correction. To obtain killing rates for co-cultures, values obtained for the killing rate of individual cultures (PMN, MDM) were divided by two and obtained values were added to each other.

Differences in IL-1β levels were compared by Kruskal-Wallis test with the post-hoc Dunn’s multiple comparison test. IL-8 level differences were calculated through multiple pairwise comparisons using t tests with pooled SD.

To identify correlations between variables, Pearson (ρ_x,y) and Spearman (ρ) coefficients were calculated to detect linear and non-linear correlations and in each case the best fitting coefficient value was chosen.

All statistical tests were performed using R studio desktop (version 1.2.5033; (RStudio Team, 2019. RStudio: Integrated Development for R. RStudio, Inc., Boston, MA URL http://www.rstudio.com/), and a p-value of < 0.05 was considered statistically significant.

In a first set of experiments, we assessed the induction of ET in PMNs and MDMs, either alone or in co-cultures, in response to Map and BCG. In order to assess whether this mechanism would rely on live or heat-killed bacteria, all culture types were incubated with both live and heat-inactivated Map and BCG. Fluorometric analysis revealed that both cell types cultured individually extruded similar detectable DNA levels in response to zymosan ( Figure 1F ), whereas lower ET levels were detected when cells were cultured together ( Figure 1F ). Levels of ET formation in response to live mycobacteria were similar for PMNs and MDMs and lower for CCs although no statistical differences were found ( Figures 1B, D ). However, in response to killed mycobacteria, PMNs showed higher ET release ( Figures 1C, E ) compared to MDMs (K-BCG, p=0.03; K-Map, p=0.065) and to CC (K-BCG, p=0.04; K-Map, p=0.043).

Although fluorometric quantification analysis revealed significant differences in ET release among cell culture types, further analyses were performed to discriminate between DNA from ET release and DNA liberated by other means. Image analysis quantification of ETs showed that PMNs liberated significantly higher levels of ETs (6.52% - 7.23%) compared to MDMs (0%) and CCs (1.13% - 2.03%) and MDMs (0%) when stimulated with both mycobacteria, independent of whether the bacteria were alive or heat-killed ( Figures 2A–D ). As seen before, MDMs produced ET-like structures only when stimulated with zymosan ( Figure 2E ). MDMs and PMNs, cultured either alone or in co-cultures and stimulated with zymosan lost normal nuclei structure, ending in nuclear DNA extrusion ( Figure 2E ). However, co-cultures seemed to contain fewer ETs, which could be due to MDMs phagocytosing liberated ETs by PMNs or because direct contact of both cell types inhibits ET release.

The microscopic evaluations also revealed that ETs released by PMNs were the main source of histone-staining (shown in red), as this staining was mainly within the extruded ETs and not always associated with DAPI-stained nuclei ( Figures 2A, C ; PMNs and PMNs and CC). In agreement with data shown in Figure 1 , MDMs showed internalized mycobacteria but did not show ET release ( Figures 2A, C , MDMs), and anti-histone staining was always associated with the nuclei (as recognized by DAPI staining), independent of bacterial preparation used to stimulate MDM ( Figures 2A–D ; MDM). In addition, it seemed that MDMs showed less mycobacterial internalization when co-cultured with PMN, whereas the opposite was observed with PMNs ( Figures 2A, C ; CC). However, in Map stimulated CC, PMNs nuclei were not always labelled by the anti-histone antibody, and many PMNs showed condensed nuclei, potentially indicative of apoptosis ( Figure 2C ; CC). Furthermore, ETs were rarely detected and MDMs presented a red punctate pattern in the cytoplasm ( Figures 2A, C ; CCs) which could support the hypothesis that MDMs phagocytosed histones liberated by PMNs.

Having established that both mycobacterial species induce ET formation in PMNs, but less so in MDMs or co-culture of both cells, we next assessed if ET release was influenced by secreted mediators or required direct cell-cell contact ( Figure 3A ). To further address the possible paracrine effect of MDMs, PMNs were also cultured in the transwell system (TW-PMN-CM) containing macrophage-conditioned media (CM) ( Figure 3B ). Additional controls included unstimulated and stimulated PMN kept in the transwell insert on their own (TW-CTRL and TW-PMN, respectively) to exclude an impact on ET release by the transwell culture condition compared to PMNs kept in the lower well (PMN) (see Figure 3A ).

PMNs incubated either in the transwell insert (TW-CTRL) or directly in the well (PMN-CTRL) showed no differences in baseline ET release, indicating that the culture conditions did not impact on ET release.

PMNs released equivalent levels of ETs in response to both mycobacteria when cultured either alone in the lower well (PMN) in the presence of MDM conditioned media (TW-PMN-CM), or cultured in the insert of the transwell (TW-PMN) ( Figures 3C, D ). In contrast, MDMs do not release ETs in transwell co-culture with PMNs when stimulated with mycobacteria ( Figures 3C, D ), confirming results obtained previously.

Direct cell contact induced the lowest ET release compared to the rest of the co-culture types independent whether cells were stimulated with either mycobacteria (p<0.05), and this effect was also present when cells were stimulated with zymosan ( Figure 3E ). Overall, the response to zymosan followed the response seen to either mycobacteria with the exemption of MDMs that showed a similar ET release compared to PMNs (60-70%) upon stimulation with zymosan.

After demonstrating that ET release can be triggered by mycobacteria after 4 hours, we next assessed whether ET formation at an early time point may be beneficial in terms of bacterial clearance. For this purpose, the killing capacity of PMNs, MDMs and co-cultures against both mycobacteria after 24 h was quantified. Total inoculated CFU (0 h) and CFU recovered at 24h for all cultures are shown in Figures 4A, C . MDMs alone and CCs were more effective at killing BCG compared to PMNs alone, whereas PMNs were more effective at killing Map compared to MDMs.

The resulting distribution of live and killed BCG and Map in culture cells and supernatant is shown in Figures 4B, D . Statistical analysis revealed significant differences between MDMs, PMNs and CCs showing higher killing of BCG in MDM cultures compared to PMNs (84.85% vs 34.15%, p=0.0016), and higher Map killing in PMN cultures than in MDMs (38.36% vs 9.14%, p=0.045). PMN culture killing rate against BCG was lower than in co-cultures (34.14% vs 88.19%; p=0.001), whereas MDM culture killing rate against Map was lower compared to co-cultures (9.14% vs 26.10%; p=0.065), although it was not statistically significant. Considering that CCs include half of the amount of each cell type, and that each cell type shows a certain killing capacity in individual culture, a theorical killing rate for CCs against each mycobacterium was calculated by dividing each individual culture (PMN, MDM) killing rate by two and adding both obtained values. Results indicate that co-cultures infected with BCG showed improved killing results (expected killing: 59.49% vs observed killing: 88.19%) than individual cultures. In contrast, this synergic effect was not evident in co-cultures infected with Map, showing only 2.52% of improvement (expected killing: 23.58% vs observed killing: 26.1%).

When bacterial survival inside adherent cells of each cell-culture type was compared ( Figures 4B, D ), significant differences were not observed for either BCG or Map. However, non-internalized BCG survival in PMN cultures (46.23%) was higher than in MDM cultures (3.44%, p=0.021) and in co-cultures (8.65%, p=0.021), whereas non-internalized Map survival in PMN cultures was lower than in MDM cultures (54.46% vs 32.17%, p=0.03). Map-GFP survival inside adherent cells was higher than BCG-GFP as expected.

Having assessed the killing capacity of both cell types to different mycobacterial strains, we next investigated whether their cytokine response follows a similar pattern. Considering that IL-8 is a major ET inducer in PMNs (23) and IL-1β is a key mediator of the inflammatory response that attracts MΦs to the infection site (22), we analysed the levels of both cytokines in the supernatant. IL-1β release after 24h showed significant differences between the cells involved as well as in response to both mycobacteria. MDM cultures released higher concentrations of IL-1β ( Figure 5A ) in response to BCG compared to PMN (p<0.001) and CC (p=0.009) and in response to Map compared to PMN (p=0.035). Furthermore, MDM released significantly more IL-1β in response to BCG compared to Map.

Similar as seen for IL-1β, MDMs secreted more IL-8 into the supernatant in response to both mycobacteria ( Figure 5B ) compared to co-cultures (BCG; p< 0.001, Map; p=0.013) or PMNs alone (BCG; p=0.001, Map; p= 0.005). As before, MDMs challenged with BCG produced significantly more IL-8 than challenge with Map (p=0.011).

Correlation analyses were performed between cytokine levels, and mycobacteria killing rates to evaluate the strength of relationship between both parameters ( Table 1 ). The correlation study was performed including the data from all the culture types together and then subdivided by culture type and stimuli. IL-1β and IL-8 levels were positively correlated at 24h including all culture types in the analysis indicating that both cytokines are secreted in the same culture conditions in our assay. This correlation improved when eliminating the CCs from the analysis suggesting that cytokine release follows a different pattern in CCs maybe due to the difference in cell number. BCG-stimulated cultures showed a positive correlation between IL-1β and killing rate, which increased when eliminating CCs from the analysis. On the other hand, Map-stimulated cultures showed a negative correlation between IL-1β and killing rates, and in this case, power was increased when analysing PMNs alone. IL-8 and killing capacity were positively correlated in MDM cultures. IL-8 and killing rates were negatively correlated in Map-stimulated cultures.