Staphylococcus aureus Extracellular Vesicles Elicit an Immunostimulatory Response in vivo on the Murine Mammary Gland

Staphylococcus aureus is a major pathogen responsible for bovine mastitis, the most common and costly disease affecting dairy cattle. S. aureus naturally releases extracellular vesicles (EVs) during its growth. EVs play an important role in the bacteria-bacteria and bacteria-host interactions and are notably considered as nanocarriers that deliver virulence factors to the host tissues. Whether EVs play a role in a mastitis context is still unknown. In this work, we showed that S. aureus Newbould 305 (N305), a bovine mastitis isolate, has the ability to generate EVs in vitro with a designated protein content. Purified S. aureus N305-secreted EVs were not cytotoxic when tested in vitro on MAC-T and PS, two bovine mammary epithelial cell lines. However, they induced the gene expression of inflammatory cytokines at levels similar to those induced by live S. aureus N305. The in vivo immune response to purified S. aureus N305-secreted EVs was tested in a mouse model for bovine mastitis and their immunogenic effect was compared to that of live S. aureus N305, heat-killed S. aureus N305 and to S. aureus lipoteichoic acid (LTA). Clinical and histopathological signs were evaluated and pro-inflammatory and chemotactic cytokine levels were measured in the mammary gland 24 h post-inoculation. Live S. aureus induced a significantly stronger inflammatory response than that of any other condition tested. Nevertheless, S. aureus N305-secreted EVs induced a dose-dependent neutrophil recruitment and the production of a selected set of pro-inflammatory mediators as well as chemokines. This immune response elicited by intramammary S. aureus N305-secreted EVs was comparable to that of heat-killed S. aureus N305 and, partly, by LTA. These results demonstrated that S. aureus N305-secreted EVs induce a mild inflammatory response distinct from the live pathogen after intramammary injection. Overall, our combined in vitro and in vivo data suggest that EVs are worth to be investigated to better understand the S. aureus pathogenesis and are relevant tools to develop strategies against bovine S. aureus mastitis.


INTRODUCTION
Mastitis is an inflammatory response of the mammary gland that often results from a bacterial infection and that induces local to systemic symptoms in small and large ruminants (Bradley, 2002;Le Maréchal et al., 2011b). In dairy farms, mastitis severely impacts both the animal health and the quality of milk, causing important economic losses in the dairy industry (Le Maréchal et al., 2011b;Peton and Le Loir, 2014). The Grampositive pathogen Staphylococcus aureus is one of the most important etiological agent of mastitis worldwide (Bradley, 2002;Le Maréchal et al., 2011b). Signs of S. aureus mastitis range from subclinical to gangrenous infection in ruminants and rely on strain-specific features such as the production and secretion of specific virulence factors that increase invasiveness or enable the mammary epithelial colonization of S. aureus (Le Maréchal et al., 2011a,c;Peton and Le Loir, 2014). Clearance of S. aureus from the infected udder is impaired as the pathogen is able to adhere, internalize, survive and multiply into the mammary epithelium (Alekseeva et al., 2013;Bouchard et al., 2013;. This ability of S. aureus to induce chronic infections negatively affects animals and notably, S. aureus mastitis are reportedly difficult to cure and show a high recurrence rate (Conlon, 2014;Peton and Le Loir, 2014;. The bovine strain S. aureus Newbould 305 (Prasad and Newbould, 1968;Bouchard et al., 2012), hereafter referred to as S. aureus N305 , has been used as a model strain for S. aureus mastitis in numerous studies including several from our group (Bouchard et al., 2013;Breyne et al., 2014Breyne et al., , 2017aPeton et al., 2016). Despite many efforts dedicated to understand the pathogenesis of S. aureus mastitis, the infectious process is still poorly understood and a better knowledge on host-pathogen interactions is required to allow the development of effective preventive or curative strategies. S. aureus secretes many virulence factors as well as exports both envelope-associated proteins through classical Secdependent pathways and cytoplasmic proteins through nonclassical secretion mechanisms (Bendtsen et al., 2005;Hecker et al., 2010). One of these latter mechanisms is the release of extracellular vesicles (EVs) which has been extensively described in eukaryotes (van der Pol et al., 2012;Roy et al., 2018). These EVs are spherical nano-sized particles with a lipid bilayer secreted naturally by pathogenic and non-pathogenic bacteria from budding of the cellular membranes (Prados-Rosales et al., 2011;Deatherage and Cookson, 2012;Al-Nedawi et al., 2015). Donor cells use EVs to transport various proteins which can be delivered to local or distant cellular targets to interact with and modify them. The first evidence of this novel secretion process was obtained in Gram-negative bacteria already in the 1960s Work et al., 1966;Chatterjee and Das, 1967). EVs are now recognized as important vehicles of intra-and inter-species cellular communication across all three kingdoms of life (Deatherage and Cookson, 2012;Celluzzi and Masotti, 2016). The protein content of bacterial vesicles includes factors involved in virulence, biofilm formation, modulation of the host immune response, resistance to antibiotics, bacterial survival and intra-and interspecies communication and cooperation (MacDonald and Kuehn, 2012;Brown et al., 2015;Kim et al., 2015). Most studies have been conducted on Gram-negative bacteria (Horstman and Kuehn, 2000;Ellis and Kuehn, 2010;José Fábrega et al., 2016). Consequently, our knowledge regarding Gram-positive EVs still remains limited (Brown et al., 2015). Since 2009, a few works reported the production and secretion of EVs by S. aureus, with particular emphasis on their protein content characterization and their impact on host cells (Lee et al., 2009(Lee et al., , 2013bGurung et al., 2011;Hong et al., 2011;Kim et al., 2012;Thay et al., 2013;Choi et al., 2015;Jeon et al., 2016;Bae et al., 2017;He et al., 2017;Im et al., 2017;Jun et al., 2017;Askarian et al., 2018). In analogy with Gram-negative bacteria S. aureus EVs harbor, inter alia, numerous virulence factors, can have cytotoxic effects on host cells in vitro (Gurung et al., 2011;Jeon et al., 2016) and trigger a pro-inflammatory response both in vitro and in vivo (Hong et al., 2011;Kim et al., 2012).
However, the potential contribution of S. aureus EVs to bacterial pathogenesis has only been explored for human isolates. Therefore, in the present report we aimed to at first characterize EVs produced by the bovine mastitis strain S. aureus N305 to investigate their role in the context of mastitis. To obtain this aim, we evaluated whether these purified EVs are capable to induce a stimulation of the host immune response comparable to either live or heat-killed S. aureus N305 and LTA. Our data demonstrate that S. aureus N305-secreted EVs induce an innate immune response on bovine mammary epithelial cells in vitro and a mild pro-inflammatory response in mammary tissues in vivo.

Bacterial Strain and Growth Conditions
S. aureus N305 (ATCC 29740) was grown in Brain Heart Infusion (BHI) (Difco, pH 7.4) broth at 37 • C under vigorous shaking (150 rpm/min). The phases of bacterial growth were determined by measurement of optical density at 600 nm (OD 600 ) and routinely the colony forming units (CFU) were counted on BHI agar using the micromethod (Baron et al., 2006).

Purification of S. aureus N305-Secreted EVs From Culture Supernatants
EVs were purified from S. aureus N305 culture supernatants using a method adapted from (Gurung et al., 2011). Sub-cultured cells at the end of exponential phase were diluted 1:1,000 in 1 L of fresh BHI medium and were grown until the stationary phase. After the cells were pelleted at 6,000 g for 15 min, the supernatant fraction was filtered through a 0.22 µm vacuum filter (PES) and the filtrate was concentrated around 100-fold using Amicon ultrafiltration system (Millipore) with 100 kDa filter. The resulting filtrate was subjected to ultracentrifugation at 150,000 g for 120 min at 4 • C and were applied to a discontinuous sucrose density gradient (8-68%). After centrifugation at 100,000 g for 150 min at 4 • C, each fraction of the gradient was collected. The fractions with density around 1.08-1.13 g/cm 3 were then recovered by sedimentation at 150,000 g for 120 min and suspended in Tris-Buffered Saline (TBS) (150 mM NaCl; 50 mM Tris-Cl, pH 7.5). Purified EVs were checked for absence of bacterial contamination and stored at −20 • C before use. The EVs amount were measured based on protein concentration using the Bradford reagent (Bio-Rad) and visualized by SDS-PAGE. Hereafter, the S. aureus-secreted vesicle dose correspond to the quantity of S. aureus-secreted vesicle proteins.

Negative Staining Electron Microscopy (EM)
Negative staining electron microscopy was performed at the Microscopy Rennes Imaging Center platform (MRic TEM) (University of Rennes 1, Rennes, France). Purified EVs were applied to copper grids and were negatively stained with 2% uranyl acetate as previously described (Gurung et al., 2011). The samples were visualized on a transmission electron microscope Jeol 1400 TEM (Jeol, Tokyo, Japan) operating at 120 kv accelerating voltage.

Cryo-Electron Tomography (Cryo-ET)
Vitrification of purified EVs was performed using an automatic plunge freezer (EM GP, Leica) under controlled humidity and temperature (Dubochet and McDowall, 1981). Mix-capped gold nanoparticles of 10 nm in diameter (Duchesne et al., 2008) were added to the sample at a final concentration of 80 nM to be used as fiducial markers. The samples were deposited to glowdischarged electron microscope grids followed by blotting and vitrification by rapid freezing into liquid ethane. Grids were transferred to a single-axis cryo-holder (model 626, Gatan) and were observed using a 200 kV electron microscope (Tecnai G 2 T20 Sphera, FEI) equipped with a 4k × 4k CCD camera (model USC 4000, Gatan). Single-axis tilt series, typically in the angular range ±60 • , were acquired under low electron doses (∼0.3 e − /Å 2 ) using the camera in binning mode 2 and at a nominal magnifications of 29,000x. Tomograms were reconstructed using the graphical user interface eTomo from the IMOD software package (Mastronarde, 1997). Slices through the tomograms were extracted using the graphical user interface 3dmod of the IMOD package. Measurements were performed using the measuring tools available in the slicer panel of 3dmod.

Size Distribution of S. aureus N305-Secreted EVs
The size distribution of EVs was estimated by three different methods: nanoparticle tracking analysis (NTA), tunable resistive pulse sensing (TRPS) and cryo-EM. NTA analysis was carried out using a NanoSight NS300 (Malvern Instruments, United Kingdom). EVs were thawed and diluted in TBS at 1:10,000 until an optimum visualization of a maximum number of vesicles. Data was analyzed by NTA 3.0 software (Malvern Instruments). All measurements were performed at 22 • C. TRPS analysis was carried out using the IZON qNano system (Izon Science). EVs were diluted 1:100 and applied to qNano instrument (Izon Science) at 22 • C using a nanopore NP140 after the calibration of the system with 70 nm standard carboxylated polystyrene particles (CPC70). Finally, the diameter of vesicles was measured using the images obtained by Cryo-EM from 90 round vesicles using the measuring tools available in the slicer panel of 3dmod (IMOD package).

In-solution Digestion and Identification of Proteins in S. aureus N305-Secreted EVs
Three independent biological replicates of EVs, purified as described above, were digested for NanoLC-ESI-MS/MS analysis. Purified EVs (approximately 50 µg) were pelleted at 150,000 g for 2 h at 4 • C and suspended with the solution of 6 M Guanidine-HCl (Sigma-Aldrich), 50 mM Tris-HCl (pH 8.0) (VWR C) and 2 mM DTT (Sigma-Aldrich). EVs were heated at 95 • C for 20 min and cooled in 50 mM NH 4 HCO 3 (pH 7.8) (Sigma-Aldrich). Then, samples were digested in solution using sequencing grademodified trypsin (Promega) with the ratio 1:50 of enzyme:protein for 15 h at 37 • C, as previously described by Lee et al. (2009). After digestion, the peptides were stored at −20 • C until further analysis. Nano-LC experiments were performed as previously reported (Le Maréchal et al., 2011a), with minor modifications. Briefly, the peptide mixture was loaded using a Dionex U3000-RSLC nanoLC system fitted to a Q-Exactive mass spectrometer (Thermo Scientific, USA) equipped with a nano-electrospray ion source (ESI) (Proxeon Biosystems A/S). Samples were first concentrated on a PepMap 100 reverse-phase column (C18, 5 µm, 300 µm inner diameter (i.d.) by 5 mm length) (Dionex). Peptides were then separated on a reverse phase PepMap column (C18, 3 µm, 75 µm i.d. by 250 mm length) (Dionex) using solvent A [2% (v/v) acetonitrile, 0.08% (v/v) formic acid, and 0.01% (v/v) TFA in deionized water] and solvent B [95% (v/v) acetonitrile, 0.08% (v/v) formic acid, and 0.01% (v/v) TFA in deionized water]. A linear gradient from 5 to 85% of solvent B was applied for the elution at a flow rate of 0.3 µL/min. MS data was acquired in positive mode and the spectra were collected in the selected mass range 250 to 2,000 m/z at a resolution of 70,000 for MS and at a resolution of 17,500 for MS/MS spectra. The peptides were identified from the MS/MS spectra using the X! Tandem pipeline software (Langella et al., 2017), matched against the genome sequence of the S. aureus N305 and S. aureus RF122, a bovine strain associated with severe symptoms in the host (Herron-Olson et al., 2007). A minimum of two peptides per protein was imposed with a false discovery rate (FDR) of <0.1% at the peptide level.

Protein Extraction and SDS-PAGE Electrophoresis
Proteins samples for total bacterial lysates and supernatants were extracted as previously described (Le Maréchal et al., 2009). For the extraction of proteins from the supernatant (SP), bacterial cultures were centrifuged at 7,000 g for 20 min at 4 • C and the supernatants were filtered through a 0.45 µm filter. Then, the proteins were precipitated with 10% TCA at 4 • C for 16 h and were centrifuged at 7,000 g for 90 min at 4 • C. Protein pellets were washed with ethanol 96% and the samples were stored at −20 • C. For total protein extracts (WC), cells were lysed with 200 µg/mL lysostaphin (Invitrogen) for 1 h at 37 • C in Tris-EDTA buffer (Sigma-Aldrich). 10 µg of each extract (WC, SP and intact EVs) were treated for 10 min at 100 • C in Laemmli buffer and separated by 12% SDS-PAGE (Laemmli, 1970) and the gel was subsequently stained with Bio-Safe Coomassie (Bio-rad).

Eukaryotic Cell Viability Assay
The viability of eukaryotic cells was evaluated as previously described with slight modifications . Briefly, MAC-T and PS cell lines were seeded in 96-well plates at densities of 10 4 cells per well, cultured to 80% confluence and incubated for 24 h with DMEM alone (mock control) and DMEM containing triton X-100 (0.01%) (positive control) or various quantities of S. aureus N305 EVs (0.01, 0.1, 1, and 10 µg per well). The cell viability was evaluated using 0.5 mg/mL Thiazolyl Blue Tetrazolium Bromide (MTT) (Sigma-Aldrich) according to manufacturer's protocol. The absorbance was evaluated at 570 nm and viability was expressed using 100% viability as mock control condition.

qRT-PCR Gene Expression
Confluent monolayers of PS cells were seeded in a 12-well cell culture plate at densities of 2.0 × 10 5 cells per well. Briefly, cells were washed twice with Hank's Balanced Salt Solution (HBSS) (D. Dutscher) and incubated for 3 h with DMEM (mock control), and DMEM containing viable (N305) and heat-killed S. aureus N305 (N305 HK ) cells at a multiplicity of infection (MOI) of 100 bacteria per cell, 10 µg of purified staphylococcal lipoteichoic acid (LTA) (InvivoGen, USA), and 10 and 20 µg of N305 EVs. Note that 20 µg of S. aureus N305 EVs corresponded to the relation of 1 µg per 10 4 cells in the viability assays. Once heat-treated for 30 min at 80 • C, the samples were re-plating to ensure that all bacteria had been inactivated. After the incubation period, total RNA was extracted using RNeasy mini-kit (Qiagen) and treated with DNAse-free DNA Removal Kit (ThermoFisher Scientific) to remove residual genomic DNA according to manufacturer's instructions. RT-qPCR was carried out using first-strand cDNA synthetized from 500 ng of total RNA samples by qScript cDNA Synthesis kit (Quantabio). The PPIA (peptidyl-prolyl cis-trans isomerase A), RPL19 (ribosomal protein 19) and YWHA (14-3-3 phosphoserine/phospho-threonine binding protein) housekeeping genes were used as reference genes for normalization. Amplification was performed on a CFX96 real-time system (Bio-Rad, France) and the primers used in this study are listed in Table S1. The samples setups included biological triplicates and experimental triplicates. Genes considered significantly differentially expressed corresponded to those with a P-value < 0.05 (student's t-test) when compared to the mock control.

Intraductal Inoculation of S. aureus N305-Secreted EVs in the Mouse Mammary Gland
The in vivo experimental mastitis study was conducted with a comparable protocol as previously described (Brouillette et al., 2004;Le Maréchal et al., 2011a;Peton et al., 2016). It was performed in accordance with the International Guiding Principles for Biomedical Research Involving Animals under approval the Ethical Committee of the Faculty of Veterinary Medicine in the University of Ghent, Belgium (no. EC2015_127). Forty Hsd:ICR (CD-1) outbred lactating female mice (Envigo, The Netherlands) were mated with male mice and were used 12 days after birth of the offspring. The pups were separated 2 h before of the intraductal inoculation in the fourth mammary gland pair. A mixture of oxygen and isoflurane (2-3%) was used for inhalational anesthesia of the mice and a bolus of PBSdiluted Vetergesic (i.e., buprenorphine 10 µg/kg, Val d'Hony Verdifarm NV, Belgium) was administered intraperitoneally (i.p.) as analgesic prior to any surgical intervention. The mammary gland duct was exposed through a small cut at the teat tip and each sample was slowly intraductally injected at a volume of 100 µL with a 32 gauge blunted needle. Six groups of mice were simultaneously inoculated: two groups each received S. aureus N305-secreted EVs (at concentrations of 1 and 10 µg in phosphate-buffered saline or PBS, both n = 7) and compared to a negative control group (sham) receiving PBS only (n = 7). Three independent positive control groups were included in the study set-up for comparative purposes, a first one receiving 117 CFU of viable S. aureus N305 in PBS (N305, n = 7), a second one receiving 100 CFU of heat-killed S. aureus N305 in PBS (N305 HK , n = 6), and a third one receiving 10 µg of lipoteichoic acid (LTA) in PBS (InvivoGen, USA) (n = 6). Twenty-four hours post-infection (p.i.), mice were sedated by an intraperitoneal administered mixture of ketamine (100 mg/kg Anesketin, Eurovet Animal Health BV, Bladel, The Netherlands) and xylazine (10 mg/kg; Xylazini Hydrochloridum, Val d'Hony-Verdifarm, Belgium) and, subsequently, euthanized through cervical dislocation.

Bacterial Load, Cytokine Profiling, and Histology
Upon necropsy, all mammary glands were isolated and mechanically homogenized. A serial dilution derived from 20 µL of homogenate was plated on Tryptic Soy Agar to obtain a number of CFU per amount of tissue (g). To another 100 µL-aliquot of the homogenate 400 µL of lysis buffer supplied with protease inhibitors (200 mM NaCl, 10 mM Tris-HCl pH 7.4, 5 mM EDTA, 1% Nonidet P-40, 10% glycerol, 1 mM oxidized L-glutathion, 100 µM PMSF, 2.1 µM leupeptin, and 0.15 µM aprotinin) was added for later extraction of proteins. Mammary gland lysates were frozen overnight and centrifuged the following day at 12,250 g for 1 h. After recovering of the supernatant, the sample was quantified through Bio-Rad protein staining followed by spectrophotometry at 595 nm (Genesys 10S). All the samples were then adjusted to reach the same protein concentration (5 µg/µL). Selected cytokine profiling was done using a bead-based multiplex immunoassay (ProcartaPlex, Thermo Fisher Scientific) for the simultaneous quantification of IL-1α, IL-1β, IL-6, TNF-α, MCP-1, CXCL2 (MIP-2), RANTES and BAFF and specific simplex immunoassays (ProcartaPlex) for mouse CXCL1 (KC) and IL-17A. All assays were performed in accordance with the manufacturer's instructions after in house validation for mammary gland matrix. Isolated mammary glands (n = 2 per condition) were fixed in 3.5% buffered formaldehyde, embedded in paraffin and sections were deparaffinized and stained with hematoxylin and eosin (H&E). Mammary gland tissues were visualized at x200 and x400 magnification.

Statistical Analysis
The data were presented as mean concentration ± standard error. The differences between the animal groups were assessed using one-way analysis of variance, followed by Tukey's range test. The statistical program Prism 6 (GraphPad) was used considering significant a P-value lower than 0.05.

RESULTS
The Bovine Mastitis-Associated S. aureus Strain N305 Produces EVs in vitro EVs secreted by S. aureus N305 were isolated from the cell-free supernatants of stationary phase cultures. For that purpose, we used centrifugation, filtration and density gradient ultracentrifugation, the standard method for the isolation and purification of membrane vesicles with higher purity (Yamada et al., 2012;Dauros Singorenko et al., 2017). Homogeneity and integrity of vesicles were evaluated by both negative staining electron microscopy and cryo-electron tomography (cryo-ET). Electron micrographs of purified EVs revealed nanosized vesicular structures with a typical cup-shape (Raposo and Stoorvogel, 2013; Figure 1A). Cryo-ET analysis showed homogeneously shaped spherical particles ( Figure 1B). The size distribution of EVs was estimated by three different methods: nanoparticle tracking analysis (NTA), tunable resistive pulse sensing (TRPS) and cryo-ET. Average EV sizes were 67 ± 13 nm (mean and standard deviation) for TRPS, 91 ± 23 nm for cryo-ET and 126 ± 2 nm for NTA ( Figure 1C). Although average size of EVs may vary according to the analytical method due to the limitations of each methodology (van der Pol et al., 2010(van der Pol et al., , 2014Maas et al., 2015;Sitar et al., 2015), these complementary approaches highlighted the monodisperse size distribution of S. aureus N305 EVs. In addition, the total particle count evaluated by TRPS and NTA was similar and close to 4 × 10 9 particles per mL of bacterial culture supernatant. These results demonstrate that the bovine mastitis-associated S. aureus strain N305 releases a high number of EVs homogenous in size and shape under laboratory culture conditions.

Bovine S. aureus N305-Secreted EVs Carry Virulence Factors
In addition to their structural characterization, cargo proteins of S. aureus N305-secreted EVs were determined through LC-MS/MS analysis of their proteomic profiles on three independent purified samples. The pattern of proteins associated with purified S. aureus N305-secreted EVs differed from that of the other bacterial cell fractions (Figure 2A). A total of 222 proteins were consistently identified (Table S2), with the majority (n = 160) predicted to be either cytoplasmic (n = 89) or putatively membrane-associated (n = 71) (Figure 2B), showing that shedding of EVs appears to be also a pathway for protein secretion in S. aureus N305. The latter were overrepresented in EVs when compared to the predicted whole membrane proteome (32 vs. 26%). More than half (34/58) of the number of predicted lipoproteins from the whole proteome (i.e., proteins with a signal peptidase II cleavage site) were identified, indicative for their relative enrichment in EVs ( Figure 2C).

S. aureus N305-Secreted EVs Are Not Cytotoxic Against bMEC in vitro
We first evaluated whether S. aureus N305 EVs could affect the viability of eukaryotic cells. For that purpose, two bovine mammary epithelial cell (bMEC) lines, MAC-T and PS were treated for 24 h with growing EVs doses: 0.01, 0.1, 1 and 10 µg per well. The analysis of viability by MTT analysis did not reveal any differences between the MAC-T and PS control cells and the cells exposed to EVs (Figure 3). These results show that S. aureus N305 EVs do not induce the cytotoxic effect in both MAC-T and PS cells in the tested conditions (Figure 3).

S. aureus N305-Secreted EVs Induce an Immunostimulatory Response in bMEC in vitro
To examine whether S. aureus N305-secreted EVs could induce the host's immunity in vitro, particularly the innate defense, the PS cell line was then treated with S. aureus N305-secreted EVs (10 and 20 µg per well). Live and heat-killed S. aureus N305 (MOI 100:1; 25 µg) and LTA, a pro-inflammatory component of the S. aureus envelope (von Aulock et al., 2003) were used as complementary positive controls. The expression levels of host genes coding for key pro-inflammatory cytokines (IL-1β, IL-8 and TNF-α) and for the antimicrobial peptides β defensing-1 (DEFβ1) were compared to those of untreated PS cells (Figure 4). A significant induction of all genes was observed after treatment with live S. aureus compared to untreated PS cells. In contrast, no differences in IL-1β, TNF-α, and DEFβ1 expression were observed following treatment with heat-killed bacteria, while the IL-8 expression was slightly (fold-change = 1.7) but significantly increased compared to untreated PS cells. Treatment with the positive control LTA increased the expression level of tested genes. Finally, we observed a significant and dose dependent increase of IL-8, IL-1β, TNF-α, and DEFβ1 expression level in presence of S. aureus N305 EVs (Figure 3, EV 10 and EV 20 ), either to a similar (IL-1β, DEFβ1) or slightly lower (IL-8, TNF-α) level than those following treatment of PS cells with live S. aureus N305. These results demonstrated the ability of S. aureus N305secreted EVs to stimulate bMEC in vitro in a way similar to that of live bacterial cells.

S. aureus N305-Secreted EVs Induce Inflammatory and a Local Host Innate Immune Response in vivo
To evaluate the in vivo modulation of S. aureus N305secreted EVs on mammary gland inflammation, a well-defined experimental model of bovine S. aureus-induced mouse mastitis was used (Peton et al., 2016). Six groups of mice were inoculated with either PBS (negative control), live S. aureus N305 (first positive control), heat-killed S. aureus N305 (S. aureus N305 HK , second positive control), LTA (third positive control), S. aureus N305-secreted EVs (1 µg, EV 1 or 10 µg, EV 10 ). At 24 h p.i., macroscopic signs of inflammation were observed in the glands that received live S. aureus N305, LTA and EV 10 and in a much lesser extent in the glands that received S. aureus N305 HK and EV 1 . The mammary gland inoculated with S. aureus N305 had an average bacterial load of 8.94 ± 0.25 × log10 (CFU/g) at 24 h p.i. and showed a profound edema and hemorrhage. This severe clinical response was attenuated in the 2 other positive control groups (S. aureus N305 HK and LTA), and also in the EV 1 and EV 10 (Figure 5). Upon microscopical evaluation, a comparable influx of immune cells was observed in the alveoli of all treated mammary glands except for the PBS-inoculation (Figure 5). Of note, mammary glands treated with EV 1 had less immune cells in their alveoli compared to EV 10 -inoculated mice again indicating a stronger inflammatory response for the higher dose.
The local levels of cytokines IL-1α, IL-1β, IL-6, MCP-1 (CCL2), IL-17A, RANTES (CCL5), BAAF, MIP-2, and KC (CXCL1) were significantly higher in the mammary glands inoculated with live S. aureus N305 compared to PBS (Figure 6). Inoculation with S. aureus N305 HK and LTA also induced some of these cytokines (i.e., BAFF and KC) but this increase was much more modest (LTA: BAFF and KC P < 0.0005; S. aureus N305 HK : BAFF P < 0.0005 and KC P < 0.05). S. aureus  N305-secreted EVs significantly induced several local cytokine levels i.e., MCP-1 (CCL2), RANTES (CCL5), KC, MIP-2 and BAFF compared to PBS. In addition, the increase of MCP-1, BAFF, MIP-2, and KC appeared to be dose-dependent. The local IL-1β level showed a modest and also a dose-dependent increase, albeit non-significant compared to the PBS control. In terms of chemokines, S. aureus N305-secreted EVs elicited even a stronger local response than both the S. aureus N305 HK and LTA positive controls: for EV 10 the average BAFF level was only slightly lower to that in live S. aureus N305-injected glands (91 ± 12 vs. 72 ± 25 pg/mL), while average KC levels (328 ± 110 vs. 205 ± 75 pg/mL) were even higher. These in vivo results demonstrated that S. aureus N305-secreted EVs induce a predominantly chemotactic local immunostimulatory response.

DISCUSSION
The Gram-positive pathogen S. aureus infects a wide range of tissues and is one of the most important bacteria in bovine mastitis negatively affecting milk production worldwide FIGURE 4 | S. aureus N305-secreted EVs induce an immunostimulatory response in vitro on PS bovine mammary epithelial cells. Expression of IL-1β, IL-8, TNF-α, and DEFβ1 by bovine mammary epithelial PS cells shown as fold changes at mRNA level measured by RT-qPCR after 3 h post stimulation with either viable S. aureus N305 cells (N305, green), heat-killed S. aureus N305 cells (N305 HK , yellow), 10 µg of purified staphylococcal lipoteichoic acid (LTA, blue), 10 µg and 20 µg of N305 EVs (EV 10 , white; EV 20 , red). Values were calculated as the mean ± SD obtained from three independent experiments after normalization to mock control DMEM. Asterisks indicate statistical significance as evaluated by one-way analysis of variance (ANOVA). ****P < 0.0001; ***P < 0.0005; **P < 0.005. (Peton and Le Loir, 2014). The different degrees of clinical manifestations can be correlated to inter-strains variations in terms of specific virulence factors (Le Maréchal et al., 2011b). Furthermore, our knowledge of the host-pathogen interactions, as well as the molecular basis associated with persistence of , 1 µg of EVs (EV 1 ) (test group) and 10 µg of EVs (EV 10 ) (test group). Arrows emphasize different morphological and histopathological manifestations in the mammary gland post-treatment: 1, healthy lactating mammary gland; 2, severely inflamed mammary gland with bacterial exudates; 3, slightly inflamed mammary gland; 4, moderately inflamed mammary gland. Macroscopic differences resulting from the different treatments of the mammary glands are clearly visible (e.g., prominent redness and inflammation in the S. aureus N305, LTA and EV 10 groups). Middle and right panels: Representative H&E stained tissue sections of the mammary gland from each group acquired at two magnifications are shown; middle panel: 20x, scale bar = 50 µm; left panel: 40x, scale bar = 20 µm. Arrow labeled 5 highlights the milk secrete in the alveolar lumen; arrows labeled 6 marks red blood cells; arrows labeled 7 marks immune cells in the lumen of alveoli. At 24 h p.i. the PBS group did not show any immune cell influx in the alveolar space, the S. aureus N305 group alveolar lumen had a profound hemorrhage and a stronger immune cell influx compared to the S. aureus N305 HK and LTA groups. The EV 1 and EV 10 groups had a dose-dependent recruitment of immune cells with an influx for EV 10 similar to that observed in the LTA group. S. aureus infections remains to be fully elucidated. Although EVs have been associated with multiple S. aureus infectious processes this is not yet the case in veterinary medicine (Hong et al., 2011;Kim et al., 2012). Thus, our objective was to investigate if EVs are secreted by the bovine udder isolate S. aureus N305 and their role in the context of mastitis. S. aureus N305 secreted EVs that displayed the basic features of extracellular prokaryotic membrane vesicles, i.e., a nanometric size range and a cup-shaped morphology and spherical structure (Raposo and Stoorvogel, 2013). In terms of their protein cargo, they appeared to be enriched with lipoprotein and membrane protein classes as also shown for human S. aureus strains and other bacterial species (Deatherage and Cookson, 2012;Rath et al., 2013;Brown et al., 2014;Askarian et al., 2018). Furthermore, S. aureus N305-secreted EVs share several proteins in common with those of S. aureus strains isolated from human clinical sources (Lee et al., 2009;Jeon et al., 2016) (Supplementary data, Figure S1 and Table S3), supporting the hypothesis that conserved regulatory mechanisms for cargo sorting may exist. A remarkable feature of S. aureus N305secreted EVs was the predominance of virulence factors that accounted for approximately 10% of their vesicular proteome. Similar observations exist for other pathogenic Gram-positive bacteria, such as Mycobacterium tuberculosis , Bacillus anthracis (Rivera et al., 2010), Streptococcus pneumoniae (Olaya-Abril et al., 2014), Listeria monocytogenes (Lee et al., 2013a), and Clostridium perfringens (Jiang et al., 2014). This feature suggests that virulent protein delivery via EVs represents an important common mechanism in the development or progression of infections. Consistent with this, in the current study proteins involved in key steps of mastitis pathogenesis such as adherence to host tissues, development of lesions and tissue damage, and evasion from the host immune system were observed. S. aureus N305-secreted EVs also contained numerous proteins associated with metal ion acquisition, a mechanism essential for local bacterial proliferation and for circumventing nutritional immunity, as well as proteins involved in resistance to antimicrobial agents. Additionally, several moonlighting proteins with secondary roles closely related with pathogenesis (e.g., enolase, GAPDH, autolysin, Tuf) (Modun and Williams, 1999;Heilmann et al., 2005;Antikainen et al., 2007;Widjaja et al., 2017) and lipoproteins involved in S. aureus Toll-like receptor 2 (TLR-2) activation and pathogenicity (Shahmirzadi et al., 2016) were identified. Collectively, our proteomic data suggest a role of S. aureus N305-derived EVs in the host-pathogen interaction.
Bovine mammary epithelial cells (bMECs) play an important role as the first line of defense against intramammary infections through the recognition of pathogens and the secretion of chemokines, cytokines and antimicrobial peptides that lead to neutrophil recruitment (Gray et al., 2005;Rainard and Riollet, 2006). Our in vitro data showed that bMECs exposure to S. aureus N305-secreted EVs led to a significant and dose-dependent increased expression of two pro-inflammatory cytokines (IL-1β, TNF-α), one chemokine (IL-8) and one bactericidal peptide (DEFβ1). Both cytokines are key elements of the early innate immune response in the mastitic mammary gland, comprising also the chemokine which is responsible for neutrophil recruitment and activation (Lahouassa et al., 2007), and the bactericidal peptide which is involved at the level of the oxygen-independent antimicrobial processes (Gurao et al., 2017). The ability to modulate the epithelial immune response has been described for EVs originated from both Gram-negative and -positive bacteria, including human S. aureus strains (Ismail et al., 2003;Bauman and Kuehn, 2006;Bomberger et al., 2009;Ellis and Kuehn, 2010;Kaparakis et al., 2010;Parker et al., 2010;Jun et al., 2017). Although the current study did not aim to unravel the molecular mechanism behind the response elicited by the S. aureus N305-secreted EVs, it observed no cytotoxicity on bMECs after 24 h of incubation. At first sight, this finding may seem unexpected given the abundance of virulence factors within these EVs. However, although they all harbor an arsenal of virulence factors, cytotoxic activity is not shared by all S. aureus-secreted EVs (Gurung et al., 2011;Thay et al., 2013;Jeon et al., 2016;Jun et al., 2017). The presence or absence of cytotoxicity may result from proteome differences between S. aureus-secreted EVs (Jeon et al., 2016). EVs produced by M. tuberculosis induce a TLR2-dependent proinflammatory response via their lipoprotein cargo in interacting directly with the plasma membrane receptor that stimulates intracellular signaling cascades (Prados-Rosales et al., 2011). A plausible hypothesis states that S. aureus N305-secreted EVs stimulate bMECs in a similar way, since they are also enriched in lipoproteins that are TLR2 ligands (Shahmirzadi et al., 2016). However, a variety of other mechanisms of action may exist. For example, Helicobacter pylori-secreted EVs exhibit NF-κB-dependent pro-inflammatory activities via inflammasome-dependent signaling through the cytosolic NOD1 receptor after their fusion with the epithelial plasma membrane and delivery of their cargo into the cytosol (Kaparakis et al., 2010). Another example is the pore-forming toxin cytolysin A delivered by Escherichia coli outer membrane vesicles which induces an epithelial pro-inflammatory response via alteration of the cellular Ca 2+ homeostasis (Uhlén et al., 2000;Söderblom et al., 2005). Interestingly, S. aureus N305-secreted EVs also harbors toxins (PSMs, leukocidin) that are able to trigger Ca 2+mediated host cell activation (Barrio et al., 2006;Forsman et al., 2012). Whether these EVs act extracellularly through ligandreceptor interactions, intracellularly after their internalization, or by inducing subtle perturbations such as on the cellular Ca 2+ homeostasis to modulate the epithelial immune and inflammatory response remains to be investigated. It will be of high interest to examine more closely the role of lipoproteins and toxins in this modulation.
Consistent with our in vitro results, intramammary inoculations with S. aureus N305-secreted EVs elicited a local response with a dose-dependent immune cell recruitment and the induction of a pro-inflammatory cytokine profile. LTA is a major immunostimulatory of Gram-positive bacteria and can induce secretion of cytokines in vivo (Fournier and Philpott, 2005;Rainard and Riollet, 2006). Of relevance, these EVs were able to induce a higher in vivo response than LTA and S. aureus N305 HK both at the histological and cytokine levels, which suggests their role in S. aureus N305 pathogenesis as immunostimulatory factors. The influx of inflammatory cells at inflammation sites is generally associated with elevated levels of CXC chemokines (Zlotnik and Yoshie, 2000). Accordingly, we detected an induction of the murine IL-8-like chemokines KC (CXCL1) and MIP-2 (CXCL2) reportedly involved in neutrophilic recruitment at inflammation sites (Rollins, 1997;Leemans et al., 2003;De Filippo et al., 2008). In addition, the levels of MCP-1, a monocyte chemoattractant (Rollins, 1997), RANTES, a monocytes, T cells, basophils and eosinophils chemoattractant (Arango Duque and Descoteaux, 2014) and BAFF, the B-cell-activating factor increased. The immune response induced by EVs was comparable to that observed with live S. aureus N305 although attenuated and not restricted to chemokine induction. Notably, the induction of IL-17, a critical cytokine for immune response and clearance of the pathogens at epithelial surfaces (Marks and Craft, 2009), was detected only with live S. aureus N305. These results showed that the immune response induced by S. aureus N305-secreted EVs might not be associated with IL-17-dependent T cell signaling. In addition live S. aureus N305 induced an increase production of several pro-inflammatory cytokines (i.e., IL-1α, IL-1β, IL-6), as previously reported (Breyne et al., 2014;Peton et al., 2016). These EVs appeared to mainly induce a chemotaxis related migratory response in vivo when compared to the responses induced by live S. aureus N305. This raises the question of the role of EVs and the biological significance of their pro-chemotactic effects during the S. aureus infectious process. In the infected udder, colonization and invasion of the mammary gland by bacteria is followed by a recruitment of polymorphonuclear neutrophilic granulocytes, which are responsible for clinical symptoms and determine the course of infection. S. aureus cells are able to survive within a variety of host cells including professional phagocytes such as neutrophils (Voyich et al., 2005) and monocyte-derived macrophages (Kubica et al., 2008) that may serve as a vehicle for persistence and dissemination of the infection (Garzoni and Kelley, 2009). One could view the chemotactic activity of S. aureus N305 EVs as a strategy to recruit phagocytic cells to allow the internalization of the bacterium and therefore its survival. This hypothetic strategy may also explain the persistence observed in S. aureus N305 infections. Additional studies are needed to better understand the role exerted by EVs in S. aureus pathogenesis, particularly with regard to strain-dependent clinical manifestations of mastitis and their involvement in chronic infection.
In summary, our study demonstrated at first that EVs are produced by the mastitis strain S. aureus N305 and that they induce an immunostimulatory response, both in bMEC in vitro and in a preclinical model of bovine mastitis. Furthermore, it provides evidence that S. aureus N305-secreted EVs principally modulate the chemotaxis of innate immune cells. These findings provide both novel insights in S. aureus mastitis pathogenesis and innovative avenues to control mastitis in which EVs are proposed as potential candidates for the development of vaccines as these are currently lacking in the treatment of Gram-positive udder infection.