Whole blood was obtained from healthy donors and anticoagulated using 200 ATU/mL Heparin (Rafludin, Celegene, UK). Platelet poor plasma (PPP) was isolated as previously described (Tilley et al., 2013). Approval for the collection of whole blood was obtained from the Ethics Committee in RCSI (REC 1121). Informed consent was provided in accordance with the Declaration of Helsinki. In the absence of healthy donors, pooled human PDP (Cambridge Biosciences, London, UK) was used for experimentation.

Staphylococcus aureus Newman Wildtype NCTC 8178 (Duthie and Lorenz, 1952) was cultured anaerobically overnight at 37°C in Brain Heart Infusion (BHI) broth as outlined previously (Kerrigan et al., 2008). Cultures were centrifuged at 3800 x g for 7 minutes, washed in PBS and adjusted to an OD_600nm = 1.0A.

Human aortic endothelial cells (Promocell, Germany) were sheared at 10 dyn/cm² for 24 hours in Endothelial Cell Growth Media MV (Cruinn, Wexford, Ireland). Spent media was aspirated and replaced with fresh EMV media containing 10ng/mL of TNFα for 4 hours, followed by a 1-hour incubation with platelet poor plasma (PPP). Wells were subsequently blocked with 1% Bovine Serum Albumin (BSA) in EMV media for 1 hour. Plates were infected with Staphylococcus aureus Newman Wildtype for 24 hours. Negative control plates received fresh media after BSA blocking and both groups were incubated for 24 hours at 37°C, 5% CO₂. A multiplicity of infection (MOI) of 10 was used to infect HAoECs for 24 hours under static conditions.

Exosomes were isolated from healthy and infected sheared human endothelial cell supernatants. Briefly, supernatants were centrifuged at 300 x g for 5 minutes to remove cells, 2500 x g for 10 minutes to remove dead cells and debris, and 3800 x g for 7 minutes to remove bacteria. Supernatants were transferred to an ultracentrifuge tube and centrifuged at 20,000 x g for 90 minutes to remove microparticles and larger vesicles. Supernatants were passed through a 0.2µm filter to ensure filtrate was free of Staphylococcus aureus. The filtrate was transferred to a new ultracentrifuge tube and centrifuged at 120,000 x g for 120 minutes.

For characterisation by Western Blot, exosomes were resuspended in RIPA buffer and passed through a 0.45 µM filter. Samples were sonicated at 15 A for 3 seconds and BCA assays were performed to determine protein concentration. Protein expression was determined using 1:500 rabbit anti-Alix IgG primary antibody (Clone: A302-938A, Bethyl laboratories), 1:200 rabbit anti-CD63 (Clone: MX-49.129.5, Santa Cruz Biotech), 1:1000 rabbit anti-Flotillin1 (Clone: AB_941621, Abcam), and 1:500 mouse anti-Tsg101 (Clone: 4A10, Genetex).

Exosomes were studied by flow cytometry using CD63+ (Clone TEA3/18) superparamagnetic capture beads (Abbexa, Cambridge, UK). Briefly, superparamagnetic capture beads were added to exosomes in 1X assay buffer and incubated in the dark overnight at room temperature. Next, a biotin primary antibody was added to each tube and incubated for 60 minutes at 4°C. The beads and exosome mixture were washed and placed on a magnetic rack for 5 minutes. Secondary antibody (Streptavidin-Phycoerythrin) was added for 30 minutes at 4°C, and samples were analysed using the Attune NxT Flow Cytometer.

To visualise exosomes, transmission electron microscopy (TEM) was performed. Superparamagnetic capture beads were incubated with exosomes in the dark overnight at room temperature. Following this, the exosome-bead mixture was placed on PELCO^® 200 Mesh Special Metal Grids (Ted Pella Inc, California, USA) and allowed to dry. Samples were visualised using a Hitachi Transmission Electron Microscope at 150,000X.

Human endothelial cells were cultured as previously described and seeded onto glass cover slips adhered to 6-well plates. Confluent endothelial cells were incubated with 5µM of a Near Infrared BF2-azadipyrromethene (NIR-AZA) fluorescent probe for 2 hours. Endothelial cells were fixed with 4% paraformaldehyde (Sigma) for 10 minutes before being washed three times with ice cold PBS (Sigma). Coverslips were mounted with 4′,6-diamidino-2-phenylindole [(DAPI) Sigma] and analysed using the Zeiss AxioObserver Z1 Microscope.

Exosomal-RNA (exo-RNA) from healthy and S. aureus infected human endothelial cells was isolated using Total Exosomes & Protein Isolation Kit as per manufacturer’s specifications. Small RNA composition (20–200 nucleotides) of exo-RNA was assessed using an AATI fragment analyser. RNA was reverse transcribed to complementary DNA (cDNA) and the expression of 384 miRNA were assessed using TaqMan^® Human Microarray A cards & RT-PCR for healthy and S. aureus infected samples. Individual qPCR assays were performed to confirm the expression of select miRNA from the TaqMan^® Human Microarray A cards. For individual assays, data was normalised to a semi-synthetic oligonucleotide cel-miR-39-3p (UCACCGGGUGUAAAUCAGCUUG) and both TaqMan^® Human Microarray A cards and individual miRNA expression data was calculated using the comparative cycle threshold (2^-ΔΔCt) method.

Total RNA was isolated from THP-1 monocytes using the mirVana™ isolation kit as per manufacturers specifications. RNA was quantified using a NanoDrop spectrometer. Semi-synthetic miR-99a (rArArCrCrCrGrUrArGrArUrCrCrGrArUrCrUrUrGrUrG) and miR-99b (rArArCrCrCrGrUrArGrArUrCrCrGrArUrCrUrUrGrUrG) oligonucleotides (Integrated DNA Technologies, Leuven, Belgium) were used to generate standard curves (2 – 8200 fM) for absolute quantification of miR-99a and miR-99b levels. RNA was reverse transcribed to cDNA and RT-qPCR was performed as described above with data normalised to RNU48 and total miRNA concentrations determined from standard curves.

Monocytes (10⁶ cells/well) were cultured with exosomes isolated from healthy and S. aureus infected endothelial cells for 24hrs. The suspension was centrifuged, washed, and resuspended in human FcR binding inhibitor (eBiosciences, UK). Monocytes were incubated with CD11b Monoclonal Antibody (M1/70), APC, and MHC Class II (I-A/I-E) Monoclonal Antibody (M5/114.15.2), PE, or isotype control. Expression of these activation markers was analysed by flow cytometry. For delivery of exosomes to monocytes, pellets were resuspended in RPMI media 1640+ GlutaMAX and incubated with monocytes for 24 and 120 hours

Human IL-6 (240 µg/mL) and IL-10 (180 µg/mL) capture antibodies were adhered to 96 well plates overnight and subsequently blocked with a 1% BSA solution for 1 hour. Monocytes were co-cultured with healthy and S. aureus infected endothelial cells, and the supernatant was added to the plates for 2 hours at room temperature. IL-6 (120 ng/mL) and IL-10 (90 ng/mL) detection antibodies were added to the plates for 2 hours at room temperature followed by 0.25 mg/mL of Streptavidin-HRP for 20 minutes. Plates were washed between each step with 0.05% tween-20 in PBS. Plates were treated with an equal mixture of H₂O₂/tetramethylbenzidine, and the reaction was stopped after 20 minutes with 2N H₂SO₄. The absorbance at 450/570nm was determined in a 1420 Victor V3, Perkin Elmer plate reader, and standard curves were generated to determine unknown IL-6 and IL-10 sample values.

THP-1 monocytes (ATCC^® TIB-202™) were cultured in RPMI media 1640+ GlutaMAX supplemented with 10% Foetal bovine serum (FBS) and penicillin/streptomycin. Monocytes were seeded onto 6-well plates at a concentration of 1x10⁶ cells/well and a transfection mixture containing OptiMEMTM (Biosciences), Lipofectamine 3000, and miRNA mimetics and antagomirs (miR-99a, miR-99b, and negative control scrambled RNA) was added for 48 hours. Monocytes were subsequently infected with S. aureus for 24 hours at an MOI of 1. Following infection, cell culture supernatants containing monocytes were harvested by centrifugation at 1500 x g for 5 minutes and washed with ice cold PBS. Cleared lysates were separated on a 10% SDS-PAGE gel. Proteins were electroblotted onto PVDF membranes (Roche) for 1 h. Membranes were probed using a 1:1000 monoclonal rabbit anti-mTOR antibody (Clone: 7C10, Cell Signalling Technologies, Dublin, Ireland). Cytokine production was monitored from supernatants for IL-6 and IL-10 as previously outlined.

Data are presented as mean ± standard error of the mean. Experiments were carried out in triplicate with a minimum of three independent experiments. Statistical differences between groups were assessed by ANOVA with Tukey post hoc tests or t-tests, as indicated. P-value < 0.05 was considered to be significant.

Exosomes are actively released from cells to facilitate intercellular communication in both healthy and disease states, and profiling exosomes can be particularly difficult due to their small size and low refractive index (Gardiner et al., 2014). However, exosomes commonly express various membrane and biogenesis related proteins which can distinguish subsets of extracellular vesicles. We confirmed by Western blot that our enriched samples did indeed contain exosomes since they expressed exosomal-specific markers including tetraspanin CD63, biogenesis related Alix and Tsg101, and membrane-associated Flotillin-1 ( Figures 1A, B ). Tsg101 was only detected in exosome samples and not microparticle fractions demonstrating that exosomal preparations were not contaminated with other vesicles or cellular components ( Supplementary Figures 1, 2 ). In addition, we used an immunostaining approach wherein the enriched samples were analysed for their expression of CD63, a signature exosome surface marker ( Figure 1C ). Transmission electron microscopy was performed with 300 nm magnetic beads coated with the CD63 ligand. Images were acquired before and after incubation with exosomes purified from the endothelial cells, which demonstrated a collection of <100 nm sized vesicles that distributed around the beads (arrows). We next investigated if there was a difference in exosomal release from human endothelial cells following infection. Superparamagnetic CD63+ capture beads were used to detect exosomes from our enriched supernatants which were then analysed using flow cytometry. We found that there was no significant difference in exosomal release from human endothelial cells in healthy or infected conditions ( Figure 1D , Supplementary Figures 3, 4 ).

Since similar levels of exosomes were found in both the healthy and infected endothelial cells, we next sought to characterise the exosomal contents in each condition to determine their miRNA expression profiles. Exosomes were isolated from healthy and S. aureus infected endothelial cells, then exosomal small RNA (smRNA) was quantified using a fragment analyser ( Figure 2A ). A 17-fold increase in smRNA was observed in exosomes isolated from the infected endothelial cells, in contrast to the exosomes from the healthy endothelial cells. Spectrometer RNA profiles indicated that exosomes isolated from healthy and infected cells contain smRNA between 4 and 40 nucleotides in length, consistent with miRNA. We performed quantitative miRNA expression analysis using TaqMan™ Array Human MicroRNA A Cards containing 384 miRNAs, to determine the miRNA profile of the exosomes derived from the healthy and infected endothelial cells ( Figure 2B ). Briefly, we detected the presence of 49 differentially regulated miRNAs altered in response to S. aureus infection, of which 17 were upregulated and 14 downregulated. In particular, the microRNA-99 family members (miR99a/b) were highly upregulated, both of which are putative downstream targets of mTOR – a protein associated with the anti-inflammatory response to microbial infection (Weichhart et al., 2011; Cao et al., 2017; Tsai et al., 2018). To validate exosomal miRNA expression, individual TaqMan arrays were performed which revealed that miR-99a and miR-99b were upregulated in contrast to exosomes derived from healthy endothelial cells ( Figure 2C ). Collectively, these results suggest that there are significant differences in the miRNA profile in exosomes isolated from healthy and S. aureus-infected human endothelial cells. Specifically, exosomal miRNA-99a/b may be associated with infection through the mTOR regulation pathway.

Monocytes are innate immune cells that play a central role in the host response to infection. They function to trigger an inflammatory response to eradicate the infectious pathogen, thereby are critical in modulating sepsis-induced inflammation and immunosuppression (Giamarellos-Bourboulis et al., 2006; Santos et al., 2016). In normal physiological conditions, endothelial cells are quiescent and do not interact with monocytes (Njock et al., 2015; Tang et al., 2016). In an inflammatory and infectious environment, vascular activation promotes the expression of surface adhesins which in turn binds monocytes (Tang et al., 2016). In addition, emerging evidence suggests under pathological conditions, exosomes may alter the contents of their cargo and deliver them to neighbouring immune cells to promote intercellular communication and microbial defence (de Jong et al., 2012; Wu et al., 2017; Barberis et al., 2021). Since our results revealed exosomal content released by endothelial cells is altered during S. aureus infection, we sought to determine whether endothelial-derived exosomes could be taken up by monocytes and whether this contributes to their immunomodulatory function in sepsis.

We first treated the endothelial cells with a Near Infrared BF2-azadipyrromethene (NIR-AZA) fluorescent probe which binds non-specifically to the plasma membrane and cytosol of cells (Gantier et al., 2012). Immunofluorescence analysis revealed the NIR-AZA probe was successfully internalised by endothelial cells ( Figure 3A ). Next, exosomes were isolated from NIR-AZA stained cells that were infected with S. aureus for 24 hours. To identify if monocytes were capable of taking up these extracellular vesicles, the stained endothelial exosomes were co-cultured with monocytes and real-time fluorescent microscopy was performed over 16 hours. At t=0 minutes, small fluorescent vesicles – the endothelial exosomes stained with the NIR-AZA probe – surround the unstained monocytes ( Figure 3B ). At 60 minutes, the monocytes become fluorescent as they begin to interact with the exosomes. The fluorescent intensity of the monocytes is greatest at 300 minutes, indicating that since the NIR-AZA dye has stained the plasma membrane and cytosol of the monocytes, they have successfully taken up the endothelial exosomes. To validate exosomal uptake by monocytes, we performed absolute quantification of miR-99a and miR-99b in monocytes that were co-cultured with the exosomes derived from healthy and infected endothelial cells. Real-time PCR revealed basal concentrations of miR-99a and miR-99b in monocytes were 400 aM and 205 fM, respectively. S. aureus infection alone had no impact on miR-99a/b expression in monocytes. Delivery of healthy endothelial exosomes to monocytes induced a significant increase from basal levels in miR-99a (P<0.05), but not for miR-99b. Alternatively, delivery of S. aureus-infected endothelial exosomes caused a significant increase in both miR-99a (P<0.01) and miR-99b (P<0.05) expression in monocytes ( Figure 3C ). The cystolic increase in miR-99 levels in monocytes following exosome delivery confirms their uptake, relative to the unstimulated monocytes.

Having demonstrated successful uptake of endothelial exosomes by monocytes in vitro through real-time fluorescent microscopy and absolute quantification, we investigated possible pro-inflammatory effects induced by the exosomes. CD11b and MHCII are pro-inflammatory activation markers expressed on the surface of monocytes; CD11b is involved in monocyte adhesion, migration, and stimulation of pro-inflammatory pathways during infection, whereas MHCII is associated with antigen presentation and the adaptive immune response. Both proteins are routinely upregulated in response to a pro-inflammatory stimuli (Cella et al., 1997; Duan et al., 2016). Flow cytometry was performed to determine expression of CD11b and MHCII in monocytes following delivery of healthy and infected endothelial-derived exosomes after 24 hours ( Figure 4A ). A significant increase of both CD11b (P<0.05) and MHCII (P<0.05) expression was observed in the monocytes cultured with infected endothelial exosomes, compared to the unstimulated control group. Exosomes from healthy endothelial cells had no effect on CD11b but significantly increased MHCII expression. This suggests that S. aureus infected endothelial exosomes significantly alters CD11b and MHCII expression of monocytes which could directly influence their pro-inflammatory response to infection.

To further investigate the pro-inflammatory effect of exosomal delivery had on monocytes, we analysed the production of two major inflammatory cytokines frequently associated with sepsis mortality: IL-6 and IL-10. Basal levels of IL-6 and IL-10 were assessed in unstimulated monocytes. Healthy endothelial exosomes had no effect on cytokine expression whereas exosomes from S. aureus infected endothelial cells resulted in an increase in IL-6 (P<0.05) and a decrease in IL-10 (P<0.05) after 24 hours ( Figure 4B ). These results suggest that exosomes from S. aureus infected endothelial cells are driving the pro-inflammatory cytokine response and diminishing the anti-inflammatory response.

We identified that exosomes derived from endothelial cells infected with S. aureus contain heightened levels of miR-99, which are confirmed to target mTOR (Jin et al., 2013; Cao et al., 2017; Tsai et al., 2018; Yin et al., 2018). The mTOR pathway is a critical network which serves as master regulators of several cellular functions including metabolism, growth, and proliferation. In addition, mTOR limits the production of pro-inflammatory cytokines and promotes anti-inflammatory cytokines during infection (Weichhart et al., 2008; Weichhart et al., 2011). We performed western blots to assess mTOR protein expression in monocytes that were either unstimulated or co-cultured with exosomes derived from healthy and infected endothelial cells. Unstimulated and healthy endothelial exosomes produced undetectable mTOR levels, consistent with findings that mTOR activity is low in the absence of an appropriate stimuli (Zhang et al., 2011; Waickman and Powell, 2012; Jones and Pearce, 2017). In addition, monocytes directly infected with S. aureus caused heightened mTOR expression. However, monocytes exposed to infected endothelial exosomes had markedly lower mTOR expression than those infected with the bacteria directly ( Figure 4C ). This suggests that perhaps exosomal miR-99 expression could directly influence mTOR activity in monocytes during infection. To investigate this, transfection assays were performed using mimetics and antagomirs against miR-99a/b ( Figure 5A, D ). Mimetics targeting miR-99a, miR-99b, and a negative control composed of scrambled RNA were transfected into monocytes and subsequently infected with S. aureus for 24 hours. Both miR-99a and miR-99b mimetics delivered to monocytes resulted in reduced mTOR production ( Figure 5C ). We also analysed the effects of suppressing miR-99a/b using antagomirs ( Figure 5A ). Transfection of both miR-99a and miR-99b antagomirs to infected monocytes induced an increase in mTOR expression therefore restoring mTOR activity in vitro ( Figure 5F ). This data demonstrates that both members of the miR-99 family target and suppress mTOR activity during infection.

Since mTOR was successfully downregulated in the presence of miR-99a and miR-99b, we sought to investigate whether the production of pro- and anti-inflammatory cytokines could be influenced directly by these miRNAs during microbial infection. Basal levels of IL-6 and IL-10 were measured in negative controls with lipofectamine 3000 and scrambled RNA – both of which were infected with S. aureus. These were compared to S. aureus infected monocytes transfected with miR-99a and miR99b mimetics and antagomirs. Addition of miR-99a/b mimetics resulted in significantly increased levels of IL-6 and reduced IL-10 ( Figure 5B ); conversely, knockdown of miR-99a/b using antagomirs caused reduced IL-6 and increased IL-10 production ( Figure 5E ).