Neutrophils Encompass a Regulatory Subset Suppressing T Cells in Apparently Healthy Cattle and Mice

Neutrophils that reside in the bone marrow are swiftly recruited from circulating blood to fight infections. For a long time, these first line defenders were considered as microbe killers. However their role is far more complex as cross talk with T cells or dendritic cells have been described for human or mouse neutrophils. In cattle, these new roles are not documented yet. We identified a new subset of regulatory neutrophils that is present in the mouse bone marrow or circulate in cattle blood under steady state conditions. These regulatory neutrophils that display MHC-II on the surface are morphologically indistinguishable from classical MHC-IIneg neutrophils. However MHC-IIpos and MHC-IIneg neutrophils display distinct transcriptomic profiles. While MHC-IIneg and MHC-IIpos neutrophils display similar bacterial phagocytosis or killing activity, MHC-IIpos only are able to suppress T cell proliferation under contact-dependent mechanisms. Regulatory neutrophils are highly enriched in lymphoid organs as compared to their MHC-IIneg counterparts and in the mouse they express PDL-1, an immune checkpoint involved in T-cell blockade. Our results emphasize neutrophils as true partners of the adaptive immune response, including in domestic species. They open the way for discovery of new biomarkers and therapeutic interventions to better control cattle diseases.


INTRODUCTION
Neutrophils are major partners of the innate immune system and are considered the first line of defense against microbes (1). They originate from the bone marrow (BM) and circulate in large numbers in the blood from where they are swiftly recruited to sites of inflammation or infection to fight danger. They are functionally equipped to rapidly phagocytose invading microbes via a variety of receptors (2) and rapidly deliver large amounts of reactive oxygen species (ROS) into phagocytic vacuoles to kill microbes (3). If not tightly controlled these dangerous weapons lead to tissue damage and neutrophils are often a signature of uncontrolled inflammation or infection.
Neutrophils are the dominant cell population circulating in blood in a wide range of animal species and humans (4). However, in cattle or mouse blood they are less numerous than lymphocytes (5,6). Compared to human or mouse neutrophils, bovine neutrophils present peculiarities such as unique receptors or the lack of N-formylated chemotactic peptide receptor (5). Neutrophil counts are an important signature of cattle condition. In lactating cows, they represent the most part of an abnormal somatic cell count in milk from mastitic cows. Indeed, their swift recruitment into the mammary gland is of critical importance in the fight against invasive pathogens (7). For a long time, neutrophils were thought to only travel from their birth place, the BM, to blood and tissues in response to inflammatory signals or infections. However, this has recently been revisited by a number of authors who have identified neutrophils residing in spleen (8,9), lymph nodes (10), or lung vasculature (11). These new locations and features of resident neutrophils have been discovered in mice and humans but are not well documented in other animal species, including cattle, under normal conditions. Long considered as "suicide bombers," neutrophils were recently upgraded as global players of the immune response along with the discovery of new functions (12). They fully participate in shaping adaptive immunity by suppressing T cell proliferation and activity (13,14) or by promoting IgM and IgG production by splenic B cells (9). This broad range of phenotypes and functions was recently linked to discrete populations of neutrophils characterized by unbiased single cell analyses (15). Like macrophages and dendritic cells, neutrophils are now recognized as plastic cells able to respond and adapt to their environment. Again, although neutrophil diversity and plasticity are now well documented for mouse and human neutrophils, they remain unknown for domestic species including cattle.
Here we conducted a parallel thorough characterization of neutrophils present in mouse BM or circulating in cattle blood at steady state. We define this term as animals without any apparent sign of disease or infection. In other words, animals where the immune system is at homeostasis. We discovered a new population of regulatory neutrophils that displayed suppressive activity on T-cells, at steady state. Animal protocols were approved by the "Val de Loire" Ethics Committee for Animal Experimentation and were registered to the French National Committee for Animal Experimentation under N°2016091610026164.V3 (mice) or N°2016040410531486 (cattle). Cattle remained in their environment (UE-PAO) and no animal was sacrificed for this work. Six-to eight-week-old C57BL/6 mice and OT-II transgenic mice for the OVA MHC class II complexspecific TCR (16) on a C57BL/6 genetic background were bred at the resident PFIE animal facility before use.

Animal Protocols and Sampling
Bovine blood was collected from Holstein Friesian cows at the jugular vein into vacutainer K2 EDTA tubes. Broncho-tracheal lymph nodes, spleen biopsy, and sternum BM were collected post-mortem at a commercial abattoir from Blonde d'Aquitaine, Limousine, Charolaise and Highlands cows. Mice were euthanized by CO 2 inhalation. Blood (from heart), spleen, inguinal lymph nodes, and femurs were collected.

Preparation of Cells
Tubes containing blood were centrifuged at 1000xg for 10 min at 20°C before removal of the plasma layer and buffy coat. In indicated experiments, buffy coat was used in proliferation assay (see below). Red blood lysis buffer (Sigma 94 R7757) (4 vol/1 vol of blood) was added for 5 min at room temperature to lyse red blood cells. Cells were washed twice in D-PBS with 2mM EDTA. Bovine blood neutrophils were labeled with primary anti-G1 and anti-MHC-II Ab, followed by secondary fluorescent-conjugated Ab before purification by FACS (see next paragraph for labeling protocol). Mouse neutrophils were generally purified from the BM by positive magnetic selection with anti-Ly-6G PEconjugated Ab (1A8; BD Biosciences) and anti-PE microbeads (Miltenyi Biotec) as described previously (17) except for cell sorting, where mouse neutrophils were enriched from BM by positive magnetic selection with anti-CD11b microbeads (Miltenyi Biotec) before labeling with antibodies. Bovine BM cells were collected post-mortem at local commercial abattoirs, by scratching an open rib (directly on the carcass) with a spatula, cells were diluted in PBS. Spleen biopsies, the broncho-tracheal lympho nodes (LN) and BM cells were transported at 4°C to our laboratory. Bovine and murine LN and spleen biopsies were disrupted mechanically and all cells were filtered through 100mm nylon cell strainer (BD Falcon), and washed twice. Cells were suspended in RPMI-1640 supplemented with 2mM L-glutamine, 10mM HEPES, and 1mg/ml of BSA with extremely low endotoxin level (≤1.0 EU/mg) (hereafter referred to as RPMI complete medium). Cell counts were determined after staining with Türk solution and numerated with a Malassez's chamber.

Flow Cytometry
Bovine cells were suspended in PBS with 10% of horse serum (Gibco), 2mM EDTA, and labeled for 30 min with primary antibodies (see Supplementary Table S1). Mouse cells were suspended in PBS 2mM EDTA. After saturation with anti-CD32/ CD16, murine cells were incubated 30min with fluorescent mAb. After washes in D-PBS (300xg, 10 min, 4°C), both murine and bovine cells were labeled 30min with the corresponding fluorescentconjugated secondary antibodies. Cells were washed and fixed with BD cell fix diluted four times in PBS. Data were acquired with a LSR Fortessa ™ X-20 Flow cytometer (Becton Dickinson) and results analyzed with Kaluza software (Beckman Coulter).
For neutrophil subsets purification, cell concentrations were adjusted to 10 7 cells/ml and sorted with a MoFlo Astrios EQ high speed cell sorter (Beckman Coulter) according to our previously published protocol (18). Sorted cells were spread on microscope slides (Superfrost, Thermo) by cytocentrifugation (3 min, 700xg) and stained with May-Grünwald and Giemsa with the RAL 555 kit (RAL Diagnostics).

Neutrophils Functional Assays
Sorted neutrophils viability was evaluated after 1 h stimulation with 100ng/ml of LPS (Escherichia coli 0111:B4, Sigma). Neutrophils were then incubated 15min at room temperature with anti-annexin V antibodies in binding buffer (BD Bioscience). Cells were washed and incubated 15min at room temperature with streptavidin-APC-Cy7 (BD Bioscience). Cells were washed and incubated 5 min with 20µg/ml of propidium iodide (BD Bioscience) and then directly analyzed with LSR Fortessa X-20 flow cytometer. Phagocytosis was measured using pHrodo ™ Red E. coli BioParticles ® Conjugated (MolecularProbes ® ) following the manufacturer's instructions. Briefly, purified neutrophils were first incubated in a 96 wells microplate for 30 min at 37°C in RPMI complete medium with or without 2µg/well of cytochalasin D (Sigma), and then for 1 h with 20µg/wells of pHrodo E. coli BioParticles. Fluorescence was directly measured with the LSR Fortessa ™ X-20 Flow cytometer. ROS produced by neutrophils were quantified using the CellROX ® Orange Flow Cytometry Assay Kits (MolecularProbes ® , C10493) following the manufacturer's instructions. Briefly, purified neutrophils were first incubated for 1 h at 37°C in RPMI complete medium with or without 400µM of TBHP in a 96-wells black microplate and then for 30 min with 100nM CellROX ® . Fluorescence was measured with the LSR Fortessa ™ X-20 Flow Cytometer. To test the bacterial killing capacity of neutrophils, we used the Escherichia coli P4 strain, isolated from a clinical case of bovine mastitis which causes severe infections in mice (19). Escherichia coli P4 bacteria were grown in 10 ml BHI medium overnight at 37°C without agitation. Bacteria were then diluted in BHI medium (1 vol/100 vol) and incubated for 6 h at 37°C without agitation. Bacterial concentration was determined by the optical density at 600 nm and adjusted at 4.10 5 CFU/ml in RPMI complete medium. Purified neutrophils were infected at a MOI of 0.2 in RPMI complete medium in 1.5ml Eppendorf tubes for 90 min at 37°C under agitation on a Rotator SB3 (Stuart equipment). Bacteria in RPMI complete medium without neutrophils were used as the reference to asses killing. Bacterial dilutions (in PBS 0.25% SDS) were plated on TSA agar plates and incubated for 16 h at 37°C before numeration of CFUs.

RNA Extraction and Gene Expression Analysis
Total RNAs were extracted from cell-sorter purified cells using NucleoSpin RNA kit with a DNase treatment (Macherey Nagel) and reverse transcribed with iScript ™ Reverse Transcriptase mix (Bio-Rad) according to the manufacturer's instructions. Primers (Eurogentec) are listed in Supplementary Table S2. Primers validation was performed on a serial diluted pool of cDNA (a mix of cDNA from spleen, lung, LN, blood, and BM cells for both species) with a LightCycler ® 480 Real-Time PCR System (Roche). Gene expression was then assessed with the Biomark HD (Fluidigm) in a 48x48 wells plate, according to the manufacturer's instructions. The annealing temperature was 60 and 62°C for bovine and mouse samples respectively. Data were analyzed with Fluidigm Real-Time PCR software to determine the cycle threshold (Ct) values. Messenger RNA (mRNA) expression was normalized to the mean expression of three housekeeping genes for each animal species to obtain the DCt value. Principal component analysis (PCA) and hierarchical clustering were performed with DCt values in R studio (Version 1.3.959, © 2009-2020 RStudio, PBC), using respectively the FactoMineR and pheatmap packages. Ward's minimum variance method was applied for clustering, with dissimilarities squared before clustering (ward.D2). DCt values were centered to the median for clustering.

Measure of T-Cell Suppressive Activity of Neutrophils
For the mouse, splenocytes from OT-II mice were collected, homogenized to single-cell suspensions through nylon screens and resuspended in RPMI medium (Gibco) supplemented with 10% decomplemented fetal bovine serum (Gibco), 2 mM L-glutamine (Gibco), 100 U penicillin, and 100 mg/ml streptomycin (Gibco). 10 5 cells/well were distributed in a 96-wells round bottom plate (BD Falcon). OT-II splenocytes proliferation was induced by addition of 2 mg/ml of the OVA peptide 323-339 (PolyPeptide Group). As indicated, purified neutrophils were added to the culture at a ratio of 1 neutrophil:10 splenocytes in a final volume of 200µl. Wells without neutrophils were used as reference for maximal proliferation.
As indicated, neutrophils were separated from splenocytes by placing them in a HTS Transwell-96 permeable device with 0.4 mm pore and polycarbonate membrane and adapted receiver plate (Corning, reference CLS3381). To test the role of MHC-II and CD11b molecules in the suppression mechanism, neutrophils were also treated 1 h before incubation with splenocytes with 15µg/ml anti-CD11b mAb (clone M1/70) or rat IgG2bk as isotype control; anti MHC-II mAb (clone 2G9) or rat IgG2ak as isotype control. Plates were incubated at 37°C with 5%CO 2 . Cell proliferation was quantified after 3 days of culture using CyQUANT Cell Proliferation Assay tests (Thermo Fisher) according to the manufacturers' instructions.
For the bovine, a mixed lymphocyte reaction assay was set up by mixing blood cells from two genetically distant cows with cow N°1 as the responding animal and cow N°2 as the stimulating animal. Briefly, blood was centrifuged at 1000xg, 15min, 20°C, and buffy coats were collected and diluted four times in PBS. PBMCs were collected at the interface of 1.077 density Percoll gradient (GE Healthcare) after centrifugation at 400g, 15min, 20°C, without brake. PBMCs from cow n°1 (responding) remained untreated while PBMCs from cow n°2 (stimulating) were incubated for 30 min at 38.5°C in 5% CO 2 with 50µg/ml of mitomycin C from Streptomyces caespitosus (Sigma, M4287), to block their proliferation. After three washes, 10 5 stimulating cells were mixed with 10 5 responding cells in 96 wells plates (ratio 1:1) in a total volume of 150µl in RPMI complete medium supplemented with 100U/ml penicillin and 100mg/ml streptomycin. The negative control was responding cells alone and the reference maximal proliferation was responding and simulating cells together at ratio 1:1. Proliferation that started at day 6 was stopped at day 9. As indicated, at day 4, 10 5 of purified syngenic neutrophils from the responding animal were added to the well in a final volume of 200 µl. In one experiment, neutrophils were placed in a HTS Transwell-96 permeable device as for the mouse system. To quantify proliferation, plates were centrifuged at day 6 and day 9 for 5 min at 300xg, supernatants were discarded and cells were frozen at −80°C before addition of reagents from the CyQUANT Cell Proliferation Assay Kit (Thermo Fisher) as indicated by the manufacturer.

Statistical Analysis
Individual data and the median were presented in the figures. Statistical analyses were performed with Prism 6.0 software (GraphPad). Analyzes were performed on data from two to six independent experiments, Mann Whitney non-parametric tests, or two-way ANOVA test were used. Represented p-values were: * p < 0.05; * * p < 0.01, and * * * p < 0.001.

SUPPLEMENTARY INFORMATION
Supplemental information can be found with this article online. Should the reader need additional details, please email a request to corresponding authors.

Neutrophils Represent Discrete Populations in Mouse and Cattle
In cattle, neutrophils are often isolated by simple centrifugation of freshly collected blood to separate them from the buffy coat. They segregate at the bottom of the tube with red blood cells (20,21). In mice, because of very small volumes, blood sampling is less practical. Therefore, high numbers of neutrophils are directly extracted from the BM by magnetic selection following labeling with antibody against Ly-6G which is highly expressed by mouse neutrophils (22). These methods, that lead to a fair level of purity for neutrophils, are convenient for most assays because they are quick and preserve these fragile cells. We prepared cattle and mouse neutrophils using these rapid procedures and analyzed them by flow cytometry (Figure 1). Following mouse bonemarrow cell preparation and labeling with anti-Ly-6G and magnetic separation ( Figure 1A), banded-cells of apparent homogeneity were observed after May-Grümwald-Giemsa (MGG) staining ( Figure 1B). However, the side scatter (SSC) and forward scatter (FSC) profile by flow cytometry displayed two discrete populations of heterogeneous size. Moreover, double labeling of these "pure" neutrophils with anti-Ly-6G and anti-CD11b distinguished CD11b hi from CD11b med cell populations ( Figure 1C). In cattle, after elimination of the buffy coat containing PBMCs, cells were collected from lower 2/3 of the tube ( Figure 1D) and analyzed after centrifugation and MGG staining. This revealed heterogeneity of this cell fraction with the presence of eosinophils in variable proportions ( Figure  1E). After labeling with anti G1, a marker that is highly expressed on the surface of bovine neutrophils (23,24) and anti CD11b, cells were analyzed by flow cytometry. After SSC and FSC gating on the granulocytes populations ( Figure 1F), we distinguished CD11b pos G1 low eosinophils (representing between 2 and 8% of the SSC x FSC granulocyte gate depending on the animal) from G1 hi neutrophils. Among G1 hi neutrophils CD11b labeling segregated two populations: a main population of CD11b med neutrophils (around 87% of the granulocyte gate) and a minor population (around 1.5%) of CD11b hi neutrophils ( Figure 1F).
As we were intrigued by the presence of different neutrophil populations in mouse BM and cattle blood at steady state, we performed a more thorough characterization by flow cytometry using a panel of markers ( Figure 2). In mouse, CD11b was used to gate myeloid cells ( Figure 2A) and a combination of anti-Ly-6C and anti-Ly-6G allowed us to distinguish bona-fide neutrophils (Ly-6G hi ) from monocytes (Ly-6G neg ) (25). Among Ly-6C + Ly-6G hi cells, a minor subset (1.2%) of MHC-II pos Ly-6C + Ly-6G hi neutrophils was clearly distinguished from the main population (92.5%) of the MHC-II neg neutrophils. In cattle, among the neutrophils that highly expressed the G1 marker, MHC-II neg neutrophils represented the main population (93%). However, a minor population of G1 hi neutrophils (around 1.5%) expressed MHC-II antigens on the surface as observed in the mouse ( Figure 2A). The two MHC neg et MHC pos subsets were sorted by flow cytometry in both species. After centrifugation onto glass slides and MGG staining, they were indistinguishable under a microscope ( Figure 2B). Their banded or segmented nuclei morphology, for mouse bone marrow or bovine blood neutrophils respectively, probably reflected their different maturation stage. Sorted MHC-II n e g and MHC-II pos neutrophils (purity >99%) were then labeled with a panel of antibodies ( Figure 2C). In both bovine and mouse, in comparison with classical MCH-II neg neutrophils, the MHC-II pos subset overexpressed CD11b on the surface as well as Lselectin CD62L. The CD14 LPS coreceptor was not detected on MHC-II neg neutrophils. By contrast around 50% of MHC-II pos neutrophils displayed this receptor on their surface in both the mouse and the bovine. In mouse, we also observed that the MHC-II pos subset expressed high levels of CD44 as well as CD274 (PDL-1) the ligand for PD-1 involved in T cell exhaustion (26). Since neutrophils are fragile cells which could be exacerbated by flow cytometry sorting, we next asked if the two subsets displayed different ex-vivo survival times. The two subsets were incubated with LPS for 1 hour and stained with annexin V and propidium iodide ( Figure 2D). The large majority of neutrophils remained alive after this treatment and no significant difference was observed between classical MHC-II neg neutrophils (97% alive in mouse or bovine) and MHC-II pos neutrophils (95% alive in mouse or bovine).
Therefore, in mouse BM or in cattle blood, two subsets of neutrophils could be distinguished at steady state by MHC-II labeling. Both subsets displayed similar survival profiles after short-term ex-vivo LPS stimulation.

MHC-II pos Neutrophils Are Enriched in Lymphoid Organs
Although neutrophils are mostly present in the BM reservoir and in circulating blood, recent data in the mouse model and in humans highlighted their wider distribution in tissues (27) including lymphoid organs. We next examined the distribution of MHC-II neg versus MHC-II pos neutrophils in different compartments in both mouse and cattle. Cells were extracted After labeling with anti-G1 and anti CD11b, dot-plot analysis of the "granulocytes" gate (right panel) revealed three subsets: eosinophils that were negative for the G1 marker (around 8% of granulocytes) and two subsets of G1 pos neutrophils that were CD11b med (87%) or CD11b hi (1.5%). The most representative animal is shown (n=4 for mouse, n=6 for bovine), although in cattle the proportion of eosinophils and CD11b hi neutrophils circulating in blood varied between animals. and prepared from BM, lymph nodes (LN, inguinal in the mouse and tracheobronchial in bovine), and spleen. For analysis of circulating neutrophils, total blood leukocytes were analyzed by flow cytometry (see Supplementary Figure S1 for isotype controls and Supplementary Figure S2 for gating strategy) after lysis of red cells. In the BM, neutrophils represented around 30% of total leukocytes in both species ( Figure 3A). This was expected since BM is the principal compartment hosting neutrophils. In blood, neutrophils represented 30 or 20% of total cells in mouse and cattle, respectively. In both the mouse and the bovine LN, even though neutrophils represented less than 1% of cells at steady state, they were consistently detected. Similarly, between 2 and 3% of cells collected from mouse or bovine spleen after lysis of red blood cells were neutrophils ( Figure 3A). In all analyzed compartments classical neutrophils as well as MHC-II pos neutrophils were detected ( Figure 3B). Interestingly, whereas MHC-II pos neutrophils represented only 2.4 ± 0.6 or 4.6 ± 1.4% of total neutrophils in the mouse bone-marrow or cattle blood, respectively, they represented around 40 or 60% of total neutrophils in mouse or bovine lymph nodes, respectively. In spleen, around 20% of total mouse neutrophils and 30% of bovine were MHC-II pos ( Figure 3B). Therefore MHC-II pos neutrophils were highly enriched in lymphoid organs in both species, at steady state.

RNA Profiling Distinguishes Classical From MHC-II pos Neutrophils in Both the Mouse and the Bovine
To analyze the transcriptional profile of MHC-II pos and MHC-II neg neutrophils, we sorted the two populations from either BM (mouse) or blood (cattle) to more than 99% purity and performed a transcriptomic profiling using 48 validated primer pairs (Supplementary Table S1) designed to cover a large set of neutrophils functions such as synthesis of cytokines or chemokines, enzymes stored in granules, surface receptors, as well as transcription factors. Some weakly expressed genes were removed from the panel (highlighted in Supplementary Table   S1). We then performed unsupervised principal component analysis (PCA) and observed that the different subsets were significantly discriminated ( Figure 4A). The first axis of the PCA (explaining 26.4% of the variance) separated bovine and mouse samples, as expected. The second axis (explaining 19.1% of the variance) clearly separated the MHC-II pos from the MHC-II neg neutrophils in each species (Figure 4A), indicating that they belonged to different subsets. As expected, in the bovine, gene expression profiles were more dispersed than in the inbred mouse, due to inter-individual variability. In both species, the two neutrophil subsets were also segregated by hierarchical clustering based on DCT values ( Figure 4B). Most of the genes were significantly less expressed in MHC-II pos neutrophils as compared to MHC-II neg . However, expression of the following genes was significantly higher in the MHC-II pos subset: elane, proteinase3, PD-L1 in murine and MHC-II; AHR in bovine cells. A trend for higher expression was observed in MHC-II pos neutrophils for CD14 (both species) and proteinase3 (bovine). Therefore, in both species the transcriptomic signature significantly differed between the two neutrophil subsets which suggested different biological roles.

MHC-II pos Neutrophils Produce Higher Levels of ROS Than MHC-II neg but Similarly Phagocytose Bioparticles and Kill Bacteria
In order to gain further insight into the functions of the two neutrophil populations, we first compared their phagocytic ability using conjugated pHrodo ™ Red E. coli BioParticles ™ that only fluoresce once inside the phagosome or endosome of the cell (28). Neutrophils were then analyzed by flow cytometry and the percentage of phagocytosis was directly correlated with the mean fluorescence intensity (see Supplementary Figure S3). Phagocytosis by MHC-II neg or MHC-II pos neutrophils was compared in individual animals from each species. Both subsets actively phagocytosed bioparticles which was dramatically reduced by treatment with cytochalasin D ( Figure  5A). We did not detect a significant difference in phagocytosis between MHC-II neg and MHC-II pos neutrophils neither in the mouse nor the bovine ( Figure 5A). We next measured their potential for total ROS production after incubation with the nonspecific chemical inducer tert-butyl hydroperoxide (TBHP) using the fluorescent probe CellROX that fluoresces when oxidized (see Supplementary Figure S3). Mouse bone-marrow neutrophils incubated with medium produced very low levels of ROS whereas bovine blood neutrophils produced higher levels ( Figure 5B). As expected, TBHP treatment dramatically increased ROS production by both MHC-II neg and MHC-II pos neutrophils in both species. However, in both conditions and for both species, MHC-II pos neutrophils produced significantly more ROS than MHC-II neg cells ( Figure 5B). We next analyzed the killing activity, under non-opsonic conditions, of the two subsets against the Escherichia coli P4 strain isolated from a case of bovine clinical mastitis (19). Both types of neutrophils efficiently killed E. coli and no significant difference was observed between MHC-II neg or MHC-II pos neutrophils ( Figure 5C).

MHC-II pos but Not MHC-II neg Neutrophils Exert Contact-Dependent Suppression of T Cells at Steady State
We wondered if surface proteins such as MHC-II (mouse and cattle) or PDL-1 (mouse) or enrichment of MHC-II pos neutrophils in lymphoid organs could be linked to regulatory functions on T cells. To address this, we set up in vitro assays that were either antigen specific in the mouse ( Figure 6A) or polyclonal in cattle ( Figure 6C). Using the OT-II transgenic mice that bear the OVA peptide 323-339-MHC class II complex-specific TCR (16) we observed strong proliferation of splenocytes when stimulated with OVA peptide for 72 hours. Taking this condition as the maximum proliferation (100%) we compared the impact of adding MHC-II neg and MHC-II pos neutrophils to the proliferating cells (ratio 10 splenocytes:1 neutrophil). The classical MHC-II neg neutrophils had no measurable effect on OT-II splenocytes proliferation (Figure 6B). By contrast, addition of MHC-II pos neutrophils purified either from mouse BM or blood, decreased the capacity of CD4 OT-II cells to proliferate by 66 ± 3% ( Figure 6B). The suppressive activity of mouse MHC-II pos neutrophils depended on contact with the proliferating T-cells as no effect was observed when cells were separated by a Transwell device ( Figure 6B). We then used blocking antibodies in order to define if MHC-II or CD11b, that were used to distinguish the two subsets as both are highly expressed on the surface of MHC-II pos neutrophils, were involved in this suppressive activity. Both anti-MHC-II and anti-CD11b partially relieved suppression by MHC-II pos neutrophils that reached 70 ± 7 and 75 ± 1% of total proliferation respectively ( Figure 6B). This indicated that both molecules were involved in the suppression mechanism although they were obviously not the only ones.
In cattle, we set up a mixed-leukocyte-reaction (MLR) by mixing PBMCs from a responder animal with mitomycin-C treated PBMCs from a genetically unmatched animal as stimulating cells (ratio 1:1) for a total of 9 days. After initial decline of total DNA content during 6 days due to cells dying in the wells, proliferation of T-cells from the responder animal was measured by an increase of the DNA content at day 9 in the control wells, correlating with polyclonal activation of T-cells (Supplementary Figure S4). Level of proliferation obtained under these control conditions was set as 100% ( Figure 6D). Addition of classical MHC-II neg bovine neutrophils to the PBMCs (ratio 1:1) from the responder animal did not change the proliferative capacity. By contrast, addition of MHC-II pos neutrophils from the responder animal to PBMCs (ratio 1:1) strongly suppressed T-cells as the proliferation was completely inhibited for 3 animals, and only 16% of proliferation remained for 1 animal, as compared to control wells ( Figure 6D). Decreasing the MHC-II pos neutrophils concentration, alleviated the suppressive effect (Supplementary Figure S5) showing the The percentage of T cell proliferation after addition of MHC-II pos or MHC-II neg neutrophils was calculated based on OT-II splenocytes proliferation with ova peptide only. Neutrophils were prepared from the BM of syngenic mice except for the "blood" sample as indicated on the graph. Different conditions for the assay are also indicated, with cells separated by a transwell device, or neutrophils incubated with anti-MHC-II, anti-CD11b, or isotype controls before addition to the proliferating splenocytes. (C) For the bovine, a mixed leucocyte reaction (MLR) was set up to assess polyclonal T cell proliferation. PBMCs from the responder animal were isolated and left untreated, while PBMCs from the stimulating animal were incubated with mitomycin C. PBMCs from the two cows were incubated at ratio of 1:1.
To assess the impact of neutrophils on cattle T cell proliferation, 10 5 of sorted MHC-II pos or MHC-II neg neutrophils from the responder animal were added to the reaction at day 4. DNA was quantified at day 6 and 9. (D) The proliferation was calculated by subtracting DNA values at day 9 from day 6 values. Proliferation measured in the MLR without neutrophils was defined as the reference 100% proliferation. The proliferation observed in the presence of sorted MHC-II pos or MHC-II neg neutrophils was calculated according to the DNA content (day 9-day 6) and expressed as percentage of the reference value. In one experiment, MHC-II pos neutrophils were separated from proliferating cells by a Transwell device. (B, D) Individual data and the median in each group are represented (3 independent experiments, n≥3 pool of 10 mice, n=4 cows, n=1 cow for transwell experiment; each represented value is the mean of technical triplicates). *P<0.05; **P<0.01; ****P<0.0001 (Mann Whitney non-parametric test).
dose-response effect of MHC-II pos neutrophils on T-cells. In one experiment we could separate MHC-II pos neutrophils from the proliferating PBMCs in a transwell device and observed that suppression was abolished, indicating that the suppressive activity of MHC-II pos neutrophils was contact-dependent as in the mouse ( Figure 6D). Therefore, both in the mouse and the bovine, a subset of MHC-II pos neutrophils can be distinguished from classical neutrophils as displaying suppressive activity on T-cells at steady state.

DISCUSSION
We reported here that a population of MHC-II pos neutrophils was present in the BM reservoir, circulated in blood and was enriched in lymphoid organs, in the apparently healthy bovine and mouse. Both MHC-II pos and MHC-II neg neutrophils displayed the polylobed nucleus, and were undistinguishable by this gold standard of neutrophils characterization (14). Similar to classical MHC-II neg neutrophils, the MHC-II pos subset displayed important functions of neutrophils such as ROS production, phagocytosis, and bacterial killing. However, unlike classical MHC-II neg neutrophils, the MHC-II pos neutrophils were able to suppress T-cells, a function that is reported here for the first time for bovine neutrophils. Heterogeneity or plasticity of neutrophils has largely emerged in the literature in humans or mice and new models of neutrophil differential development are proposed (15). Here, mouse and bovine MHC-II pos suppressive neutrophils were detected in the BM reservoir. They were the only ones to exert suppressive activity on T cells. They displayed a clearly distinct transcriptomic profile as compared to MHC-II neg neutrophils. They were also highly enriched in lymphoid organs. Thus, they may represent a distinct subset, produced in the BM for rapid mobilization and regulation of T cells. On the other hand, CD11b upregulation that was observed on these neutrophils could also sign an activated state (29). Recently, CD11b hi primed neutrophils were also reported to circulate in blood in healthy mice and humans to quickly respond to danger (30). Neutrophils from healthy humans could be induced in vitro to exert ROS, CD11b and contact-dependent suppressive activity on T cells upon activation with specific stimuli (31). We detected MHC-II pos neutrophils in the BM and, at least in the mouse, they were able to suppress T-cell proliferation, suggesting that they were already present as regulatory cells in the reservoir. Moreover, the regulatory neutrophils from the two species did not undergo higher apoptosis or death as compared to their classical neutrophils counterparts upon short-term ex-vivo stimulation with LPS, indicating they were not "older" or hyperactivated (32). In addition, MHC-II expression by MHC-II neg neutrophils could not be induced in vitro by incubation with LPS (data not shown). Whether heterogenous neutrophils are released from the bone-marrow as distinct subsets under steady state conditions or correspond to activation or polarization states in situ remains an open debate (33) but what we observed here is that MHC-II pos neutrophils able to regulate T cells display a specific neutrophil "phenotype" (34).
Suppressive neutrophils that may be included within a broader category of cells termed "myeloid-derived suppressor cells (MDSCs)" (35) accumulate mostly under pathological conditions. Two main branches of MDSCs have been described: the monocytic MDSCs and the granulocytic MDSCs (G-MDSCs) (35). In cancer, G-MDSCs are associated with tumor progression and escape to the immune surveillance (36). In chronic infections such as tuberculosis (37,38) or AIDS (39), or acute inflammatory syndrome such as sepsis (40) MDSCs are targeted for new host-directed therapies. However, MDSCs, including regulatory neutrophils, are also beneficial to the host under some circumstances. Because they sustain the generation of regulatory T-cells (41) they may help avoiding graft rejection (42). They can positively regulate autoimmune disorders (43) or protect the lung from destructive inflammation during Pseudomonas aeruginosa or Klebsiella pneumoniae infections (44). They accumulate during pregnancy where they are important regulators of fetal-maternal tolerance (45) and in neonates where they could prevent overwhelming inflammation following microbial colonization (46). Recently, Aarts and colleagues demonstrated that neutrophils from healthy humans could become T-cell suppressive under specific stimulation (31). Along the same line, Fine et al. recently described that neutrophils circulate in healthy donors under several states included as "primed" cells that are first recruited to the injury site to combat infection (30). In both these studies, high levels of CD11b signed the phenotype of these neutrophils, alike the MHC-II pos regulatory neutrophil subset that we describe here. In cattle, there is no description to date of suppressive neutrophils or G-MDSCs. In a recent report, Li and colleagues (47) identified a subset of MHC-II pos neutrophils that accumulated in spleen during Ostertagia ostertagi parasitic infection in cattle and produced the suppressive cytokine IL-10. This subset was the master regulator of immune suppression in parasitized animals. We did not observe il10 transcription in any neutrophil subset nor IL-10 production in the supernatant of neutrophil-T-cells cocultures in our study (data not shown). Whether MHC-II pos regulatory neutrophils that we identified are related to the subset identified by Li et al. (47) remains an open question.
Regulatory neutrophils, or G-MDSCs, block T-cell proliferation through different mechanisms (14) including ROS production (48). We observed in both species that direct contact between MHC-II pos neutrophils and T-cells was necessary to induce suppression, as was also reported for G-MDSCs. Formation of a synapse is believed to greatly enhance suppression especially when the delivery of shortlived molecules such as H 2 O 2 is instrumental (35). Interestingly, we observed here a significantly higher production of total ROS upon chemical stimulation by MHC-II pos neutrophils as compared to their MHC-II neg counterparts which did not translate into higher killing of E. coli. Pliyev and colleagues demonstrated that NADPH dependent ROS were required for the expression of MHC-II in human neutrophils activated by GM-CSF and IFNg (49). Whether the higher ROS production recorded in our study is linked to the suppression mechanism remains to be investigated. In both the mouse and the bovine (data not shown), MHC-II was also involved in the mechanism, as blocking this molecule partially relieved the suppression of T-cell proliferation, emphasizing the need for synapse formation between the T-cell and the regulatory neutrophil. We established in the past the important role of CD11b expressed by neutrophils in phagocytic synapse and crosstalk with other cells, as well as specific signaling pathways (50,51). In the present study, we observed that CD11b was also involved in the suppression mechanism in mice, as reported for human neutrophils acquiring suppressive activity upon in vitro stimulation. Direct CD11bdependent interactions between neutrophils and T cells was required for suppression but additional molecules such as ICAM-3 on T-cells may also function as ligands for CD11b (31). Similarly, our results in the mouse model suggested that a synapse with Tcells, involving CD11b and MHC-II on the surface of neutrophils, was necessary to exert suppression. However, suppression by MHC-II pos neutrophils clearly involved other mechanisms that we will investigate in the near future. In that respect, surface expression and up-regulated gene transcription of PD-L1 by mouse regulatory neutrophils is an interesting avenue. This molecule is an important immune check-point and a favorite target for cancer treatment. MDSCs express high levels of CD274 (52) which is involved in T-cell suppression. Of note, CD14 pos monocytes expressing CD274 accumulate in blood of cattle infected by Mycoplasma bovis (53), a strongly immunosuppressive pathogen causing antibiotic resistant mastitis ant other diseases. CD274 is now a target for host-directed therapies of cattle infected by this pathogen (54). We could not investigate CD274 expression by cattle MHC-II pos regulatory neutrophils due to a lack of available Ab. We did not observe differences at the mRNA level between the two subsets (data not shown) but whether such differences exist at the protein level remains to be investigated.
To conclude, comparative analysis of mouse and bovine species allowed us to characterize a new subset of regulatory neutrophils that are able to suppress T cells. In the near future, we will investigate how these cells behave during clinical conditions in cattle such as mastitis, which remains one of the most important issue in dairy farming. We believe that such studies are of utmost importance to better understand the physiopathology of this disease, especially during chronic infections that remain difficult to treat. Our findings could lead to the discovery of new biomarkers and development of innovative host-directed therapies (54) targeting regulatory neutrophils for more effective clearance of pathogens and better control of mammary gland inflammation and damage.

DATA AVAILABILITY STATEMENT
The original contributions presented in the study are included in the article/Supplementary Material. Further inquiries can be directed to the corresponding authors.

ETHICS STATEMENT
The animal study was reviewed and approved by Val de Loire Ethics Committee for Animal Experimentation.

AUTHOR CONTRIBUTIONS
MR designed and did most of the experiments for cattle neutrophils and ED-D for mouse neutrophils. They analyzed data and prepared all manuscript figures. YV performed the flow cytometry analysis and sorting of mouse and cattle neutrophils. PC helped with the mixed leukocyte reaction in cattle. PG helped with transcriptomic analysis, critically analyzed the data, and revised the manuscript. PR brought valuable expertise on cattle neutrophil purification methods, critically analyzed the data, and revised the manuscript. NW and AR obtained grants, supervised all aspects of the work, critically analyzed the data, and wrote the manuscript. AR also contributed to the important experiments in cattle. All authors contributed to the article and approved the submitted version. Supplementary Table 2 | Sequences of primers used in this study. Primers were designed using Geneious software, in intron-spanning regions when possible. The annealing temperature was set at 60 and 62°C for bovine and mouse samples respectively. Housekeeping genes used as the reference to calculate DCT for each species are indicated in the yellow boxes and weakly expressed genes that were removed from the Principal Component Analysis presented in Figure 4 are indicated in the grey boxes.

FUNDING
Supplementary Figure 1 | Isotype controls for neutrophil diversity analysis in mouse bone marrow and cattle blood. Mouse and bovine neutrophils were labelled as described in Figure 2. Appropriate isotype controls for all antibodies in each experiment were used to correctly set the analysis and sorting gates. Dot plots from one representative animal are depicted (3 independent experiments, n=4 mice, n=6 for bovine).
Supplementary Figure 2 | Gating strategy for neutrophil diversity analysis in all organs. Mouse and bovine neutrophils were labelled as described in Figure 2. Similar procedures were set up with isotype controls for all antibodies in each experiment to correctly set the analysis gates. Dot plots from one representative animal are depicted (3 independent experiments, n=4 mice, n=6 for bovine).
Supplementary Figure 3 | Gating strategy for analysis of phagocytosis and ROS production by neutrophils. Neutrophils were labelled and sorted as described in Figures 2 and S1. (A) After purification by cell sorting from the BM (mouse) or blood (bovine) phagocytosis by MHC-II pos or MHC-II neg neutrophils was assessed using pHrodo E.coli bioparticles with or without previous treatment with cytochalasin D. Dot plots from one representative animal are depicted (3 independent experiments, n=3 pool of 10 mice, n=3 for bovine). (B) Oxidative stress was measured in MHC-II pos and MHC-II neg sorted neutrophils using the CellROX Orange probe that reacts with all ROS species. Cells were activated with TBHP or incubated with medium alone and levels of ROS were measured by flow cytometry among the live cells (unstained with eFluor780 viability dye). Dot plots from one representative animal are depicted (3 and 4 independent experiments for mice and cattle respectively, n=3 pool of 10 mice, n=5 for bovine).
Supplementary Figure 4 | Kinetics of bovine Mixed Leukocyte Reaction and analysis of neutrophil suppressive activity. (A) PBMCs from the responder animal were isolated and left untreated, while PBMCs from the stimulating animal were incubated with mitomycin C to block their proliferation. PBMCs from the two cows were incubated at ratio of 1:1. Sorted MHC-II pos or MHC-II neg neutrophils from the responder animal were added to the reaction at day 4. (B) DNA was quantified at different time points with CyQUANT Cell Proliferation Assay tests according to manufacturer's instruction and fluorescence was read at 530nm. DNA extracted from PBMCs cultivated separately decreased along the assay indicated the absence of proliferation (dotted lines). In the MLR reaction, while DNA content declined between day 1 and 6, PBMCs proliferation could be measured between day 6 and 9 (black). The effect of adding sorted MHC-II neg neutrophils (blue) or MHC-II pos neutrophils (orange) to the proliferating cells could then be measured. One representative experiment is shown and data represent the mean ± SEM of technical triplicates. Four independent experiments were conducted with different pairs of cows.
Supplementary Figure 5 | Bovine MHC-II pos neutrophil suppressive activity is dose-dependent. The suppressive assay was performed as described in Figures 6  and S4. At day 4, 1x10 5 , 5x10 4 or 1x10 4 purified MHC-II pos neutrophils from the responder animal were added to the MLR reaction. One experiment was performed (technical duplicates are depicted).