Bovine mammosphere-derived epithelial cells (MDECs) were isolated from the udder tissue of Holstein cows after euthanasia for reasons unrelated to this study, and cultured according to our protocols for generating mammary epithelial cultures enriched for stem and progenitor cells (20). CM containing all factors secreted by MDECs was collected as described previously for viability, chemotaxis, phagocytosis and reactive oxygen species (ROS) assays, as well as for treating neutrophils used for RNA deep sequencing (RNAseq) (23), using RPMI medium (Corning Incorporated, Corning, NY) + 2% filtered fetal bovine serum (FBS) (R&D Systems, Minneapolis, MN) as the base medium. CM for assays designed to evaluate features associated with neutrophil extracellular traps (NETs), including neutrophil elastase (NE) activity assays, immunofluorescent (IF) antibody binding assays, and extracellular versus intracellular DNA visualization, was generated using RPMI + 1% bovine serum albumin (BSA) (Sigma Aldrich, St. Louis, MO), as FBS has been documented to affect NET stability (36) and serum free culture has been shown to induce NET formation (37).

Blood collection for neutrophil isolation and serum collection was approved by the Cornell Institutional Animal Care and Use Committee (IACUC # 2014–0038). We obtained 40 ml peripheral blood by puncture of the coccygeal vessels of lactating Holstein cows (n = 27) into evacuated tubes containing ethylenediaminetetraacetic acid (EDTA) and another 10 ml blood from the same animal into serum tubes. For serum collection, tubes were incubated at room temperature (RT) for 30 min, then centrifuged at 2000 x g for 10 min at 4°C. Serum was removed from clot, and frozen in aliquots at -20°C. To separate neutrophils, blood was removed from tubes containing EDTA, diluted 1:1 with phosphate-buffered saline (PBS), layered over Ficollpaque Plus (GE Healthcare, Chicago, IL), and centrifuged at 1000 x g for 30 min at RT with no brake. Plasma, interphases, and Ficollpaque Plus, were removed and discarded. Remaining cell pellets consisting primarily of red blood cells (RBCs) and neutrophils were manually dislodged and RBCs were lysed with a 0.2% NaCl solution. An equal volume of a 1.6% NaCl solution was added, and cells were mixed gently by inversion. Cells were washed by centrifugation at 250 x g for 10 min at 4°C. Pellets were dislodged and RBC lysis was repeated. Neutrophils were resuspended in RPMI medium with 1% HEPES (Corning Incorporated), mixed 1:1 with trypan blue (Corning Incorporated) to visually assess viability, and counted.

To determine if MDEC CM is toxic to neutrophils in the planned functional experiments, neutrophils were incubated with RPMI + 2% FBS (control), MDEC CM, or 100 µM busulfan (positive control) at 37°C with 5% CO₂ for 1 hour (h). Cells were then washed with RPMI + 2% FBS by centrifugation at 300 x g for 5 min at RT and maintained for 5 h in RPMI + 2% FBS. To distinguish live from dead cells, 0.5 µg/ml propidium iodide (PI) was added just prior to analysis, at approximately 6 h post isolation. Caspase activity assays were performed using TF2-VAD-FMK as a fluorescent indicator for activated caspase-1, -3, -4, -5, -6, -7, -8 and -9 in apoptotic cells, according to manufacturer’s instructions (Abcam, Waltham, MA). PI was added to cells immediately prior to analysis, again at about 6 h post isolation. Fluorescence for both assays was analyzed on a BD LSRFortessa X-20 flow cytometer with FACSDiva software (BD Biosciences, San Jose, CA). FlowJo software (FlowJo LLC, Ashland, OR) was used to calculate the mean fluorescent intensity (MFI) of labeled neutrophils. The gating scheme for these, and other flow cytometry assays is shown in Supplementary Figure 1A , and MFI data used to calculate “MFI (% Control)” for this and all other flow cytometry-based assays are presented in Supplementary Table 2 .

For chemotaxis assays, 24-well tissue culture plate wells were fitted with 12 mm, circular glass coverslips and 6.5 mm transwell inserts with a 3.0 µm pore size (Corning Incorporated). RPMI + 2% FBS (control), MDEC CM, 50 ng/ml bovine interleukin-8 (IL8) (Kingfisher Biotech, Saint Paul, MN) or 10 ng/ml bovine C-X-C motif chemokine ligand 6 (CXCL6) (Novus International, St. Louis, MO) diluted in RPMI + 2% FBS were added to the bottoms of the wells, and 50,000 neutrophils were added to the tops of the inserts. Plates were incubated and 37°C with 5% CO₂ for 1 h, inserts were removed, and plates were centrifuged at 250 x g for 5 min at RT to collect neutrophils that had migrated on the surfaces of the coverslips. Treatment media were gently removed, and neutrophils that had migrated were fixed to coverslips with cold 70% EtOH and stained with hematoxylin and eosin (Sigma Aldrich). Coverslips were removed from wells and mounted on glass slides for imaging. Brightfield images of ten random fields per coverslip were captured using an Olympus CKX41 inverted microscope and Infinity 2 digital camera (Olympus Corporation, Center Valley, PA), and neutrophils were manually counted, in a blinded manner, from images.

pHrodo green Escherichia (E.) coli bioparticles (Thermo Fisher Scientific, Waltham, MA) were diluted with 2 ml RPMI + 1% HEPES to a concentration of 1 mg/ml, according to manufacturer’s instructions. Bioparticles were opsonized with 1% cow serum (20 µl) that had been collected on site, by incubation at 37°C for 1 h, while gently rocking. Neutrophils were distributed into 4 ml tubes, with 1 x 10⁶ cells per tube. Treatments consisting of either RPMI + 2% FBS (control), MDEC CM, 20 ng/ml bovine granulocyte-macrophage colony-stimulating factor (GM CSF) or 50 ng/ml bovine insulin-like growth factor 2 (IGF2) (both from Kingfisher Biotech, and diluted in RPMI + 2% FBS), were added, and cells were incubated for 1 h at 37°C. Cells were washed with RPMI + 2% FBS by centrifugation, 100 µl bioparticles were added, and cells were incubated for 1 h at 37°C. Cells were washed with RPMI + 2% FBS by centrifugation, and resuspended in PBS. A small aliquot of cells with particles (20 µl) was removed and held for microscopic visualization, and 0.5 µg/ml PI was added to the remainder. Neutrophils, neutrophils with PI or bioparticles, neutrophils with PI and bioparticles, and bioparticles alone, were analyzed on the BD LSRFortessa X-20 flow cytometer with FACSDiva software. FlowJo software was used to calculate MFI according to the gating scheme depicted in Supplementary Figure 1A . The pHrodo-labeled bioparticles are pH-sensitive, showing little fluorescent signal at the neutral pH of culture medium or physiologic buffers but bright fluorescence in the acidic environment of the neutrophil phagosome. This feature reduces the risk of detecting free particles or particles adhered to the outside of cells in the fluorescent channel. Cells removed for visualization were counterstained with the nuclear indicator 4’6-diamidino-2-phenylindole (DAPI, Sigma Aldrich) at 10 µg/ml, transferred to slides, and imaged using a 100x objective (UPlanApo, 100x/1.50 oil HR, ∞/0.13–0.19/OFN22) on an Olympus FV3000 confocal microscope with cellSens imaging software (Olympus Corporation).

Neutrophils were distributed into 4 ml tubes, with 1 x 10⁶ cells per tube. Treatments consisting of either RPMI + 2% FBS (control), MDEC CM, 500 U/ml superoxide dismutase (SOD), 20 µg/ml peroxiredoxin 2 (PII), or 250 U/ml catalase (CAT) (all from Sigma Aldrich, and diluted in RPMI + 2% FBS) were added and cells were incubated for 1 h at 37°C. Cells were washed with RPMI + 2% FBS by centrifugation and 7.5 µM 2’,7’-dichlorodihydrofluorescein diacetate (H2DCFDA) (Sigma Aldrich), diluted in RPMI + 1% HEPES, was added. After a 15 min incubation at 37°C, cells were washed again before adding the neutrophil stimulant, phorbol myristate acetate (PMA) (Sigma Aldrich) at 25 ng/ml in RPMI + 1% HEPES. Cells were again incubated for 15 min at 37°C, put on ice to stop reactions. Cells were washed with RPMI + 1% HEPES by centrifugation and resuspended in PBS. A small aliquot of cells (20 µl) was removed and held for visualization, and 0.5 µg/ml PI was added to the remainder. Neutrophils, neutrophils with PI or H2DCFDA, and neutrophils with PI and H2DCFDA were analyzed on the BD LSRFortessa X-20 flow cytometer with FACSDiva software. FlowJo software was used to calculate MFI of samples and control cells according to the gating scheme shown in Supplementary Figure 1A . Cells removed for visualization were counterstained with 10 µg/ml DAPI, transferred to slides, and imaged using the 100x objective of the Olympus FV3000 confocal microscope with cellSens imaging software.

NE activity assays were carried out, according to manufacturer’s instructions (Abcam). Briefly, 1 x 10⁵ neutrophils in 100 µl NET Assay Buffer (RPMI + 1% BSA + 1 mM CaCl₂) were added to 24-well plate wells. 800 µl RPMI + 1% BSA or MDEC CM was added to each well and incubated for 1 h at 37°C. Treatment media were removed, cells were gently rinsed with RPMI + 1% BSA and neutrophils were treated with 25 ng/ml PMA and incubated for 4 h at 37°C. Neutrophils were rinsed twice with NET Assay Buffer before the addition of S7 nuclease, to digest NET DNA, for 45 min at 37°C. Supernatants from wells were transferred to microfuge tubes and ethylenediaminetetraacetic acid (EDTA, Sigma Aldrich) was added to a final concentration of 10 mM to inactivate the nuclease. Tubes were centrifuged to remove debris and supernatants were transferred to 96-well plate wells. NE substrate was added to each well and plates were incubated for 2 h at 37°C. Absorbance at 405 nm was measured using a Tecan Infinite 200 Pro plate reader (Tecan U.S., Morrisville, NC). NE standards of known concentrations were included on each plate. Absorbances versus concentrations of standards were plotted using GraphPad Prism (GraphPad Software, San Diego, CA), and sample concentrations were interpolated based on best fit values.

For fluorescent antibody binding assays to visualize the effects of MDEC CM on proteins associated with neutrophil extracellular trap (NET) formation, neutrophils were plated at 1 x 10⁵ cells per well 24-well plate wells fitted with glass coverslips and incubated with either RPMI + 1% BSA (control) or MDEC CM, each +/- 25 ng/ml PMA for 1 h at 37°C. Treatment media were removed, cells were gently rinsed with PBS, and then fixed and permeabilized as previously described (38). For antibody binding, cells were blocked for 15 min with PBS + 1% BSA and primary antibodies against neutrophil elastase (NE), a serine protease secreted by neutrophils during inflammation and NET production (39), myeloperoxidase (MPO), an enzyme required for NET formation (40), and citrullinated histone H3 (H3), a histone modification associated with NETs (41), were diluted in PBS as described in Table 1 , were added. Neutrophils were incubated with primary antibodies or isotype control antibodies for 1 h at RT, rinsed 3 x 5 min with PBS, and then incubated with Alexa 488-conjugated secondary antibodies, diluted in PBS as described in Table 1 , for 1 h at RT. Neutrophils were rinsed 3 x 5 min with PBS and the DNA stain SYTOX^™ Orange (Thermo Fisher Scientific) was added at 1:30,000 in PBS. After a 5 min incubation at RT, cells were briefly rinsed in PBS and coverslips were transferred to slides. Cells were imaged using the 100x objective of an Olympus FluoView FV3000 confocal microscope (Olympus Corporation). The integrated density of pixels representing specific antibody binding in ten images per condition was determined using Fiji/ImageJ software (42) and normalized to account for number of cells per image.

To visualize neutrophil extracellular versus intracellular DNA, reagents from a NETosis imaging assay kit (Cayman Chemical, Ann Arbor, MI) were used, according to manufacturer’s instructions. Briefly, neutrophils were plated at a concentration of 80,000 cells per well were plated in 48-well plates and treated with RPMI + 1% BSA (control) or MDEC CM, with or without 25 ng/ml PMA for 1 h. 5 µM permeable nuclear red reagent was added and cells were incubated for 15 min at 37°C in the dark. Cells were rinsed with PBS by centrifugation and resuspended in NETosis Imaging Buffer with 5 µM extracellular green reagent, before imaging at 0, 3, 6 and 9 h using the 20x objective on a ZOE fluorescent cell imager (Bio Rad, Hercules, CA). The integrated density of pixels representing extracellular and intracellular DNA in each image taken at each timepoint was determined using Fiji/ImageJ software, and data were expressed as intracellular/extracellular DNA.

ELISAs were used to detect and quantify the concentrations of complement components C3 (C3) and C5a (C5a), bovine CXC motif chemokine ligand 6 (CXCL6), insulin like growth factor 2 (IGF2), superoxide dismutase 1 (SOD), peroxiredoxin 2 (PII) and catalase (CAT) in CM, as per manufacturer’s instructions (MyBioSource, San Diego, CA). Absorbance was measured on the Tecan Infinite 200 Pro plate reader (Tecan) at 450 nm. Absorbances versus concentrations of standards were plotted using GraphPad Prism (GraphPad Software) and sample concentrations were interpolated based on best fit values. For analysis of MDEC CM, CM from 3 MDEC lines was tested in duplicate in each assay and assays were repeated 3 times. RPMI + 2% FBS was included as control medium.

The multiplex assay was performed, as previously described (43). All antibodies used are monoclonal antibodies (mAbs) of IgG1 isotypes and were generated in mice. Equine antibodies used have been shown to be cross-reactive with bovine proteins (43). If not indicated otherwise, mAbs were produced in house (43–46). Briefly, fluorescent beads were coupled with bovine interleukin-10 (IL10) mAb 130, equine interferon gamma (IFNγ) mAb 3, and tumor necrosis factor alpha (TNFα) mAb 197–1. Coupled beads were individually sonicated and a bead mixture was prepared in PBS with 1% BSA and 0.05% sodium azide (PBN) with a final concentration of 5 x 10³ beads each. The cytokine standard mix containing recombinant fusion proteins for all 3 targets was prepared in 5-fold dilutions in PBN. Standards, samples, or PBN (blank values) were added to 96-well filter plates (MultiScreen HTS, Millipore, Burlington, MA) followed by the bead mixture. After 30 min incubation at RT in the dark, plates were washed 3 times with a plate washer (ELx50, Bio Tek, Winooski, VT), conditions that were also used for following incubation and wash steps. Subsequently, a mixture of biotinylated mAbs in PBN was added consisting of equine IL10 mAb 492–2, bovine IFNγ mAb CC-302 (Bio Rad), and bovine TNFα mAb 65–2. After another incubation and wash step, the reporter dye streptavidin R-PE (Thermo Fisher) was added, followed by a final incubation and wash. PBN was added to the plate and plates were read in an automated reader (Luminex 200, Bio Rad). Data were reported as median fluorescence intensity, and the standard curves were fitted with a 5-parameter logistic regression by the Bio-Plex software (Bio-Plex Manager 6.2, Bio Rad).

Neutrophils for RNA deep sequencing were incubated with RPMI + 2% FBS (control) or MDEC CM at 37°C for 1 h. Cells were then rinsed with PBS by centrifugation at 2000 x g for 10 min at RT and PBS was removed before cell pellets were snap frozen in liquid nitrogen and held at -80°C until RNA was extracted. Total RNA for sequencing was extracted with a QIAGEN RNeasy Plus Mini Kit as described previously (38), and RNA quantity was measured using a NanoDrop spectrophotometer (Thermo Fisher). The concentration of total RNA was confirmed using a Qubit RNA HS kit (Thermo Fisher), and a Fragment Analyzer (Agilent, Santa Clara, CA) was used to determine adequate RNA integrity. PolyA+ RNA was isolated with the NEBNext Poly(A) mRNA Magnetic Isolation Module (New England Biolabs, Ipswich, MA). UDI-barcoded RNAseq libraries were generated with the NEBNext Ultra II RNA Library Prep Kit (New England Biolabs). Each library was quantified with a Qubit dsDNA HS kit (Thermo Fisher) and the size distribution was determined with a Fragment Analyzer (Agilent) prior to pooling. Libraries were sequenced on an Illumina instrument. At least 20M reads were generated per library. Reads were trimmed for low quality and adaptor sequences with TrimGalore v0.6.0 (47), a wrapper for cutadapt (48, 49) and fastQC (50). [Parameters: -j 1 -e 0.1 –nextseq-trim=20 -O 1 -a AGATCGGAAGAGC –length 50 –fastqc] Reads were mapped to the reference genome/transcriptome Bos Taurus ARS-UCD1.2 (51) using STAR v2.7.0e (52) [Parameters:–outSAMstrandField intronMotif, –outFilterIntronMotifs RemoveNoncanonical, –outSAMtype BAM SortedByCoordinate, –quantMode GeneCounts] SARTools and DESeq2 v1.26.0 were used to generate normalized counts and statistical analysis of differential gene expression (53–55). [Parameters: fitType parametric, cooksCutoff TRUE, independentFiltering TRUE, alpha 0.05, pAdjustMethod BH, typeTrans VST, locfunc median].

Custom Silencer Select siRNA targeting bovine CXC motif chemokine ligand 6 (CXCL6), insulin like growth factor 2 (IGF2), superoxide dismutase 1 (SOD), peroxiredoxin 2 (PII) and catalase (CAT) were designed by Thermo Fisher. Silencer Select Negative Control #2 siRNA (Thermo Fisher) was confirmed to be non-complementary to all bovine protein coding genes using BLAST (56) and was included as a negative control. Cells were plated and transfected, as previously described (57), and base medium consisting of RPMI + 2% FBS was added 24 h post-transfection to generate CM, as described above.

SYBR green-based RT-PCR was performed to determine relative expression levels of transcripts of interest and data were analyzed, as previously described (22, 58). Glyceraldehyde 3-phosphate dehydrogenase (GAPDH), an accepted reference gene for bovine cells (59), was used to normalize cDNA input across samples. PCR primer sequences are listed in Table 2 .

WB was used to determine the efficacy of siRNA silencing of genes in bovine MDECs, as previously described (38, 57). Antibodies used for WB are shown in Table 1 . Images of blots were captured on a ChemiDoc Touch Imaging System and band intensities of proteins of interest relative to reference protein bands were determined using Image Lab software (both from Bio Rad). We and other groups have previously used ubiquitously expressed secreted and structural proteins as references for equal loading across CM samples in WB (57, 60, 61). To determine a suitable reference protein for the bovine MDEC CM samples in this study, CM from 3 bovine MDEC lines were run on a WB, each as undiluted, diluted 1:2, and diluted 1:4. Membranes were probed with either an anti-fibronectin or anti-β-actin antibody. Densities of bands representing each protein were consistent across undiluted samples and dilution of samples resulted in lower band densities ( Supplementary Figures 1B, C ). We chose fibronectin as the reference protein based on the following. Blots to determine the efficacy of siRNA silencing were first probed with antibodies against proteins of interest, stripped, and then probed with the reference protein antibody. Since fibronectin bands were visible at 150 kD, well above the bands representing the proteins of interest, thus greatly reducing the likelihood of misinterpreting the data if stripping was not entirely effective.

Initial viability, apoptosis, chemotaxis, phagocytosis, and ROS accumulation assays were run as 3 experiments, each on a different day (i.e., experiment 1, 2, and 3; Supplementary Figure 2A ). For each experiment, CM from MDECs generated from 3 different cows (i.e., MDEC lines 1, 2 & 3) was used and tested on neutrophils isolated from the blood of 3 different cows (e.g., neutrophil a, b & c for experiment 1; Supplementary Figure 2A ). Within an experiment, data from the 3 neutrophil isolations treated with CM from each MDEC cell line were averaged to create the data points shown on the cell line specific graphs (i.e., MDEC Cell Line 1, 2 & 3 Graph; Supplementary Figure 2A ). Viability and apoptosis data from all 3 MDEC cell lines were compiled on single graphs. To analyze cytokines secreted by neutrophils, CM from MDECs generated from 3 different cows (i.e., MDEC lines 1, 2 & 3) was used and tested on neutrophils isolated from the blood of 3 different cows. For assays consisting of three or more groups, ANOVAs were used to determine if the means between groups were significantly different. ANOVAs were followed by Dunnett’s multiple comparisons tests set up to compare the mean of each test group to the mean of the control group. For assays consisting of two groups, Student’s t-tests were used to compare the means of the groups.

NE activity, IF assays, and assays to determine ratios of intracellular/extracellular DNA, in addition to the RT-PCR, WB, chemotaxis, phagocytosis, and ROS production assays designed to determine which bioactive factors in MDEC CM were responsible for functional effects on neutrophils, were also run as 3 experiments on 3 different days (i.e., experiment 1, 2, and 3; Supplementary Figure 2B ). Here, each experiment consisted of CM from MDECs from one cow (i.e., MDEC Line 1) tested against neutrophils isolated from the blood of 3 individual cows (e.g., neutrophil j, k & l for experiment 1; Supplementary Figure 2B ). For all assays, except those to determine ratios of intracellular/extracellular DNA, data within each experiment were averaged to generate data points on graphs and data points were analyzed by ANOVA. Each ANOVA was followed by a Dunnett’s multiple comparisons test. For assays used to determine intracellular/extracellular DNA at 9 h, Student’s t-tests were used to compare the means of the groups.

RNA deep sequencing was performed on neutrophils isolated from 3 cows, each treated with CM from MDEC Line 1 and control medium.

GraphPad Prism (GraphPad Software) was used to analyze all data processed by the authors. RNAseq data was analyzed by the Cornell University Biotechnology Resource Center, as described above. P < 0.05 was considered significant and P-values for each comparison are both indicated in the manuscript text and on the graphs.

Immediately after isolation, neutrophils were stained with trypan blue and live cells were counted using a light microscope (VWR, Radnor, PA). Based on the exclusion of trypan blue dye, nearly 100% neutrophils were viable (data not shown). To determine if MDEC CM is toxic to neutrophils in the planned functional experiments, neutrophils were incubated with RPMI + 2% FBS (control), MDEC CM from 3 individual cell lines, or busulfan (positive control) for 1 h, and maintained for 5 h in RPMI + 2% FBS before being analyzed for dead cells and activated caspases associated with apoptosis, mimicking the timeline used in the functional assays. No differences in viability between neutrophils incubated in RPMI + 2% FBS and those incubated in MDEC CM from cell lines 1 (p = 0.1845), 2 (p = 0.0547), and 3 (p = 0.0530), were observed and the average viability of neutrophils incubated in MDEC CM was > 90% ( Supplementary Figure 3A ). In contrast, busulfan treatment did significantly reduce neutrophil viability when compared to the RPMI + 2% FBS control (p < 0.0001) ( Supplementary Figure 3A ). Caspase activity, as a measure of apoptosis, in live neutrophils incubated in MDEC CM from the 3 individual cell lines did not increase when compared to the RPMI + 2% FBS control (p = 0.8923, p = 0.8921 and p = 0.9534, respectively), while caspase activity in live neutrophils treated with busulfan significantly increased (p = 0.0216) [ Supplementary Figure 3B (i)]. To assure evidence of caspase activity was not missed by excluding dead cells from the analysis, data were reanalyzed taking all neutrophils (live and dead) into consideration. Again, neutrophils incubated in MDEC CM from the 3 individual cell lines did not exhibit elevated active caspase activity when compared to the RPMI + 2% FBS control (p = 0.2904, p = 0.1331 and p = 0.1918, respectively) [ Supplementary Figure 3B (ii)], whereas neutrophils treated with busulfan showed significantly more active caspase activity compared to those incubated with RPMI + 2% FBS control medium (p < 0.0001) [ Supplementary Figure 3B (ii)]. These data show that MDEC CM does not reduce neutrophil viability or induces apoptosis after 6 h of culture.

Bovine MDEC CM collected from Cell Line 1 stimulated neutrophil migration in a chemotaxis assay (p = 0.0091). As expected, recombinant bovine interleukin-8 (IL8), a known chemotactic factor for bovine neutrophils (62), did as well (p = 0.0022). The effect of each was statistically significant as compared to the RPMI medium control (dotted line) [ Figure 1A (i)]. Representative images of hematoxylin and eosin-stained neutrophils that migrated out of the transwell inserts they were plated in are shown in Figure 1A (ii). Bovine MDEC CM collected from Cell Line 1 did not significantly increase neutrophil phagocytosis as compared to the RMPI medium control (dotted line) (p = 0.100). The positive control bovine granulocyte-macrophage colony-stimulating factor (GM CSF), a cytokine reported to increase the phagocytic activity of bovine neutrophils (63) did not either (p = 0.159 [ Figure 1B (i)]. Representative images of neutrophils showing phagocytosed bio-particles (green) and nuclei (blue) are shown in Figure 1B (ii).

To confirm that the effects of the MDEC CM on neutrophils were not specific to CM collected from Cell Line 1, chemotaxis and phagocytosis assays were repeated with CM generated from MDEC lines isolated from two additional cows (i.e., MDEC Lines 2 & 3). The overall patterns of these CM on neutrophil chemotaxis and phagocytosis were like those with CM from Cell Line 1 ( Supplementary Figures 4A, B ).

In contrast to the stimulatory effect of MDEC CM on neutrophil chemotaxis, CM collected from MDEC Line 1 significantly inhibited neutrophil ROS accumulation in both unstimulated neutrophils (p < 0.001) and neutrophils stimulated with phorbol myristate acetate (PMA), a protein kinase C agonist that leads to the release of superoxide anions (64, 65) (p < 0.001), when compared to RPMI medium control (dotted line) [ Figure 2 (i)]. In addition to showing less ROS accumulation (green), fragmented neutrophil nuclei (blue) in response to PMA stimulation also appeared to be reduced by MDEC CM treatment [ Figure 2 (ii) inserts designated by red boxes], although this was not formally quantified. We confirmed that the inhibitory effect on neutrophil ROS is a general feature of MDEC CM by repeating the experiments with CM generated from MDEC Lines 2 & 3 ( Supplementary Figure 4C ).

PMA-stimulated ROS accumulation by bovine neutrophils is known to lead to the release of neutrophil extracellular traps (NETs) (66), web-like structures comprised of chromatin and chromatin-associated proteins that trap pathogens and limit infection, but also damage host tissue and lead to delayed tissue repair (67). In order to quantify the effects of MDEC CM on NET production, we used a commercially available enzyme activity assay to measure neutrophil elastase (NE), a secreted serine protease that localizes to NETs (68), in culture media. There was no difference in secreted NE detected in the medium of unstimulated neutrophils cultured in the presence of MDEC CM from MDEC Cell Line 1 when compared to neutrophils cultures in control medium (p = 0.2029). Significantly less NE was detected in culture medium collected from PMA-stimulated neutrophils cultured in the presence of MDEC CM from MDEC Cell Line 1 when compared to those cultured in control medium (p = 0.0011) [ Supplementary Figure 5A (i)]. In human neutrophils, PMA stimulates NE secretion, while in bovine neutrophils, PMA may stimulate expression, but not secretion, of NE (69). Our results support the latter, as NE concentrations in all bovine neutrophil culture media were at the low end of the standard curve used to quantify NE in test samples [ Supplementary Figure 5A (ii)].

To robustly examine the expression of NET-associated proteins in bovine neutrophils, we used antibodies against NE, myeloperoxidase (MPO), and citrullinated histone H3 (cit-H3), on unstimulated and PMA-stimulated neutrophils cultured in the presence of MDEC CM from MDEC Cell Line 1 and unstimulated and PMA-stimulated neutrophils cultured in control medium. Analysis of NE antibody binding revealed a lower expression of NE in unstimulated (p = 0.0012) and PMA-stimulated (p = 0.0005) bovine neutrophils cultured in the presence of MDEC CM from MDEC Cell Line 1 when compared to those cultured in control media. [ Figure 3A (i)]. The expression of MPO did not differ in unstimulated neutrophils cultured in MDEC CM as compared to control medium (p = 0.1827), but MDEC CM did reduce MPO expression in PMA-stimulated neutrophils (p = 0.0217) [ Figure 3B (i)]. MDEC CM did not affect cit-H3 antibody binding in unstimulated or PMA-stimulated neutrophils as compared to the control media (p = 0.2256 and p = 0.2027 respectively) [ Figure 3C (i)]. Representative images of NE, MPO and cit-H3 antibody binding are shown in Figure 3 (ii) panels with antibody binding sites labeled in green and DNA labeled in orange. Images of neutrophils probed with mouse and rabbit isotype controls revealed little to no non-specific antibody binding based on the lack of green fluorescence ( Supplementary Figure 5B ).

As another way of exploring the effects of MDEC CM on neutrophil NET production, we evaluated overall intracellular (red) versus extracellular (green) DNA in unstimulated and PMA-stimulated neutrophils at multiple time points, using location-specific DNA dyes. Both unstimulated (p < 0.0087) and PMA-stimulated (p = 0.0004) neutrophils cultured in MDEC CM exhibited a higher ratio of intra- to extracellular DNA when compared to neutrophils cultured in control medium after 9 h of treatment [ Figure 4 (i)]. Images taken at 0, 3, 6 and 9 h capture the progression and patterns of intracellular/extracellular DNA in cultures over time [ Figure 4 (ii)]. As extracellular DNA is not NET-specific but also may be a product of cell death, we cannot conclude from these data that all green staining represents NETs. However, since the viability of unstimulated neutrophils incubated in control medium or MDEC CM was high after 6 h in culture ( Supplementary Figure 3A ) we propose that most of the green staining does represent NETs.

After determining that MDEC CM exerts functional effects on neutrophils, we took a two-pronged approach to determine how CM interacts with these innate immune cells. First, to define the neutrophil molecular pathways triggered by MDEC CM, we used an ELISA and fluorescent bead-based multiplex assay to measure cytokines secreted by neutrophils treated with MDEC CM. When these assays did not provide us with clear information (described in next paragraph), we followed up with RNA deep sequencing (RNAseq) to uncover transcriptional changes in neutrophils treated with MDEC CM. To approach the interaction of MDEC CM with neutrophils from another angle, we determined which MDEC secreted factors are responsible for the functional effects on neutrophils by analyzing and manipulating the MDEC CM.

To define which molecular pathways in neutrophils are influenced by MDEC CM, we used an enzyme-linked immunosorbent assay (ELISA) to measure C-X-C motif chemokine 6 (CXCL6) and a fluorescent bead-based multiplex assay to measure interleukin-10 (IL10), tumor necrosis factor alpha (TNFα), and interferon gamma (IFNγ), in medium collected from neutrophils after MDEC CM treatment. RPMI + 2% FBS was included as a control medium to distinguish cytokines produced by neutrophils from those present in the CM used as treatment. IFNγ was not detected in any of the samples. Media collected from neutrophils treated with MDEC CM contained variable levels of CXCL6, IL10 and TNFα as compared to control media collected from neutrophils ( Supplementary Figure 6A , Supplementary Table 3 ). Concentrations of the anti-inflammatory cytokine IL10 were all higher in media from neutrophils treated with MDEC CM when compared to media from neutrophils treated with the control medium, with CM from 2 out of the 3 MDEC lines resulting in a significantly higher production (p = 0.0087, p = 0.0232, p = 0.1654). The concentrations of the inflammatory cytokines CXCL6 and TNFα, on the other hand, were highly variable across samples, so no general conclusions about the impacts of MDEC CM on neutrophil inflammatory cytokine production could be made ( Supplementary Figure 6A , Supplementary Table 3 ).

As we were not successful at predicting the specific molecular pathways responsible for the functional effects of MDEC CM on neutrophils by evaluating changes in neutrophil cytokine production, we performed RNA deep sequencing (RNAseq) on neutrophils isolated from 3 individual cows to identify global transcriptional changes induced by MDEC CM treatment. Each neutrophil culture was incubated with MDEC CM from Cell Line 1 or control medium prior to mRNA extraction and sequencing. Principle component (PC) analysis of the sequencing data showed that cow-of-origin (PC1, PC2) effect was stronger than treatment (PC3) effect [ Figure 5A (i, ii)], resulting in only 2 genes that were differentially expressed between treatment groups. Both genes were upregulated in neutrophils treated with MDEC CM when compared to control medium treatment ( Figure 5B ; Supplementary Table 4 ). The two upregulated genes were cytokine inducible SH2-containing protein (CISH) and CSK interacting membrane protein (SCIMP) (p-adj = 6.8x10^-08 and p-adj = 0.0458, respectively). CISH is a cytokine inducible protein expressed in hematopoietic cells that suppresses cytokine signaling, thereby negatively regulating granulocyte production and growth (70, 71), and SCIMP is a transmembrane adaptor that promotes Toll-like receptor 4 (TLR4)-modulated proinflammatory cytokine responses (72, 73). As these 2 genes were detected as differentially expressed between MDEC CM- and control medium-treated neutrophils despite the strong influence of neutrophil cow-of-origin effect, these data are robust. However, the overwhelming impact of cow-of-origin did mask any additional data we anticipated to acquire with this experiment and thus, we were unable to identify specific molecular pathway alterations in MDEC CM-treated neutrophils.

To identify bioactive factors in bovine MDEC CM that could be responsible for functional effects on neutrophil chemotaxis, phagocytosis, and ROS production, we searched a data set previously generated in our lab that globally characterized bovine MDEC CM using liquid chromatography-mass spectrometry (23). This analysis resulted in the identification of 347 proteins (23) and of these, we identified 11 proteins that have been reported to affect neutrophil functions (74–86) ( Table 3 ). Of those 11 proteins, we were able to use enzyme linked immunosorbent assays (ELISAs) to detect and quantify the concentrations of 7 proteins in MDEC CM; namely complement components C3 (C3) and C5a (C5a), bovine CXC motif chemokine ligand 6 (CXCL6), insulin like growth factor 2 (IGF)2, superoxide dismutase 1 (SOD), peroxiredoxin 2 (PII) and catalase (CAT) ( Supplementary Table 5 ). We selected 5 of these proteins (CXCL6, IGF2, SOD, PII and CAT) for functional follow up to determine if they indeed mediate the effects of MDEC CM on neutrophil chemotaxis, phagocytosis, and ROS production ( Table 3 ). Although we did detect C3 and C5a in MDEC CM, we decided to exclude them for functional follow-up, as substantial concentrations were also detected in control medium ( Supplementary Table 5 ), complicating the separation of complement component activity from MDEC CM from complement component activity from the base control culture medium.

To evaluate whether CXCL6 in MDEC CM is responsible for the stimulating effect on neutrophil migration, we first repeated the chemotaxis assays with MDEC CM or recombinant bovine CXCL6 as chemoattractants. We found that MDEC CM and recombinant CXCL6 each stimulated neutrophil migration (p = 0.0035, p = 0.0086 respectively), and that these increases were statistically significant when compared to the RPMI medium control (dotted line) ( Figure 6A ). Next, we reduced CXCL6 in MDEC CM by silencing CXCL6 expression in bovine MDECs using RNA interference. We confirmed knockdown of CXCL6 by a CXCL6-specific, but not a negative, small-interfering RNA (siRNA) on both mRNA and protein levels [ Supplementary Figures 6A, B (i, ii)]. We then repeated the chemotaxis assays using MDEC CM, CM collected from MDECs transfected with CXCL6 siRNA, or CM collected from MDECs transfected with a non-specific (negative) siRNA and compared neutrophil migration in each type of CM to that in control medium. As expected, complete CM and CM from MDECs transfected with the negative siRNA stimulated neutrophil migration significantly when compared to control medium (dotted line) (p = 0.0016, p = 0.0060 respectively) ( Figure 6B ). In contrast, neutrophil migration in the presence of CM from MDECs transfected with the CXCL6 siRNA, thus with reduced CXCL6, was not significantly different from migration in the presence of control medium (p = 0.8860), indicating that CXCL6 in CM promotes neutrophil chemotaxis ( Figure 6B ).

We next used a similar approach, i.e., using a recombinant bovine protein and RNA interference, to determine whether IGF2 in MDEC CM is responsible for stimulating effect on neutrophil phagocytosis. MDEC CM did not lead to a significant increase in neutrophil phagocytosis (p = 0.2030) ( Figure 6C ). Use of recombinant bovine IGF2 did stimulate neutrophil phagocytosis, which was statistically significant when compared to control medium (p = 0.0099) ( Figure 6C ). Follow up experiments consisting of transfecting MDECs with an IGF2 specific siRNA, which decreased IGF2 gene expression in MDECs and IGF2 protein in MDEC CM [ Supplementary Figures 6C, D (i, ii)], showed that MDEC CM with decreased IGF2 did not lead to an increase in phagocytosis when compared to control medium (p = 0.0880) ( Figure 6D ). This suggests that IGF2 in MDEC CM is not responsible for any stimulatory effect MDEC CM has on neutrophil phagocytosis, however, since the level to which complete MDEC CM affects neutrophil phagocytosis varied depending on the MDEC Cell Line used for CM ( Figure 1B ; Supplementary Figure 4B ), no concrete conclusions can be drawn regarding the role IGF2 in MDEC CM on this process.

As described above, we first evaluated neutrophil ROS in the absence and presence of PMA in either MDEC CM or medium with the addition of recombinant SOD, PII, and CAT, alone or in combination (triple recombinant), using protein concentrations recommended in other studies (87, 88). As expected, unstimulated or PMA-stimulated neutrophils in MDEC CM showed a significant reduction of ROS when compared to those in the medium controls (p = 0.0107, p = 0.0044 respectively) [ Figure 7A (i, ii)]. The addition of SOD, PII, CAT, alone or the 3 proteins combined, to unstimulated neutrophils did not affect ROS accumulation (p = 0.5128, p = 0.6948, p = 0.0690 and p = 0.5967, respectively) [ Figure 7A (i)]. The addition of SOD or PII to PMA-stimulated neutrophils did not affect ROS (p = 0.0912 and p = 0.1758, respectively), while treatment with CAT or the combination of all 3 proteins led to a significant decrease (p = 0.0325 and p = 0.0042, respectively) [ Figure 7A (ii)]. To reduce the expression of SOD, PII and CAT, alone or in combination, in MDEC CM, siRNAs targeting SOD, PII and CAT were used. The negative siRNA, which does not correspond to any bovine mRNA sequence, was included as a control at the same concentration as the specific siRNAs ( Supplementary Figures 7A–E , 8A, B ).

ROS accumulation in unstimulated neutrophils was significantly decreased when cells were incubated with complete MDEC CM (p = 0.0431), or CM collected from MDEC transfected with a triple dose of negative siRNA (p = 0.0170) [ Figure 7B (i)]. CM from MDECs transfected with SOD, PII, CAT, all 3 combined or the negative siRNA did not affect ROS levels in unstimulated neutrophils (p = 0.1188, p = 0.0929, p = 0.0614, p = 0.5405 and p = 0.596, respectively) [ Figure 7B (i)]. These data suggest that the activity of each of these proteins in MDEC CM influences neutrophil ROS accumulation, but since the negative control did not perform consistently, concrete conclusions about the involvement of each single protein cannot be drawn. Still, the simultaneous reduction of all 3 proteins in MDEC CM did change the effect of CM on ROS accumulation [ Figure 7B (i)], demonstrating that the combination of these bioactive factors does reduce ROS in unstimulated neutrophils.

PMA-stimulated neutrophils treated with MDEC CM, CM with reduced SOD from MDECs transfected with either concentration of negative siRNA deceased when compared to the control (p = 0.006, p = 0.0295, p = 0.0183 and p = 0.0092, respectively) ( Figure 7B (ii)). In contrast, ROS in PMA-stimulated neutrophils did not change when cells were treated with CM with reduced PII or catalase alone, nor with all 3 proteins combined (p = 0.1075, p = 0.660 and p = 0.999, respectively) ( Figure 7B (ii)), supporting the independent roles of PII and CAT in MDEC CM in reducing ROS accumulation in neutrophils.