S. aureus strain AH1263 (USA300 CA-MRSA ErmS) (3) and isolates from a pwCF with the numbers 17 (isolated in 2001) and 81 (isolated in 2012) with same spa-type t617 but different nuclease activities (21) were cultivated on Columbia 5% blood agar plates (Thermo scientific), in the following referred to as low₁₇, and high₈₁, respectively. Strains were grown in brain heart glucose bouillon (BHI, Roth) at 37°C at 160 rpm. To determine the number of colony forming units (CFUs) of S. aureus, serial dilutions were plated on Luria-Bertani Agar plates and incubated over night at 37°C. For direct use of S. aureus in co-culture with murine bone marrow derived neutrophils (BMDNs), bacteria were counted using a Neubauer improved counting chamber.

To assess the differential nuclease activity of the S. aureus- isolates, 2.5 µl of sterile filtered S. aureus supernatants were incubated with 7µg calf thymus DNA for 1h at 37°C in a total volume of 50 µl 300 mM Tris-HCl supplemented with CaCl₂ and MgCl₂ (each 3 mM) and subjected to agarose gel electrophoresis (15). Densitometry analysis was performed using ImageJ. To analyze the amount of SpA in the supernatant of S. aureus USA300, low₁₇ and high₈₁ were cultivated for 24 h, the supernatant collected, sterile filtered, and 5 µl were used for a Protein A ELISA (abcam) according to the manufacturer’s instructions.

We used 8–16 week old C57BL/6 mice, both male and female with a weight range between 18 g and 28 g. The mice were kept in a barrier facility under specific pathogen-free (SPF) conditions. All animal experiments were approved by the LANUV NRW.

Femur and tibia of C57Bl/6 mice were flushed with 6 ml PBS. This cell solution was carefully resuspended and layered on top of 6 ml 62% Percoll in PBS with 20 mM HEPES and centrifuged at 2500 rpm, for 30 min at room temperature without brake. The supernatant was removed and the neutrophil containing layer at the bottom of the tube was washed with PBS. Cells were counted and adjusted as described in the following method sections.

To analyze the capability of BMDNs to form NETs and inhibit growth of S. aureus in vitro, Percoll-purified BMDNs (2x10⁶/ml in RPMI1640, 1% FCS, 10 mM HEPES, 1 mM CaCl₂, 1 mM MgCl₂) were co-incubated with S. aureus low₁₇ and high₈₁ for 1 or 2 h with a multiplicity of infection (MOI) of 10 at 37°C with gentle shaking (50 rpm). At these time points, cultures were gently detached from the well surface, and an aliquot was used to determine the CFUs by plating of serial dilutions. Here, the cell suspension was thoroughly mixed by vortexing and pipetting to ensure optimal dispersion of bacteria. In some experiments, cells and bacteria were co-incubated for 90 min and half of the samples were treated with DNaseI (Sigma, 120 U/ml) and incubated for further 45 min at 37°C. For L-selectin shedding analysis, BMDNs were incubated with purified SpA (Sigma, P3838, 1 µg/1x10⁵ BMDNs) or H₂O as vehicle control (2.5% v/v) for 10 min or 30 min at RT. Following centrifugation (400 g, 5 min), BMDNs were used for flow cytometry or ImageStream^®X imaging flow cytometry analysis to assess or visualize NET formation (anti-Ly6G, clone 1A8, BioLegend; anti-MPO clone 2D4, abcam; Sytox red, ThermoFisher), or L-selectin surface expression (anti-CD62L, clone DREG56, BioLegend), and the supernatant was used for the NET-specific ELISA.

To analyze the abundance of NET-fragments in solutions, an ELISA was performed directed against the NET-specific citrullinated histone H3. For capturing, a 96-well plate (Immunolon 4HBX) was coated overnight at 4°C with 50 µl/well of H3cit-antibody (4 µg/ml, ab5103, abcam) in coating buffer (15 mM Na₂CO₃, 35 mM NaHCO₃, pH 9.6), washed 3x with PBS, and blocked with 5% BSA/PBS for 2 h. Wells were washed 3x (1% BSA/PBS, 0.02% Tween20), and 50 µl of the respective samples were applied to each well and incubated for 2 h. After an additional washing step, 50 µl of the detection antibody (anti-DNA-POD, Cell death Detection ELISA Plus, Roche, 1:100 dilution in 1% BSA/PBS) was applied to the wells and incubated for 2 h, shaking at 140 rpms at room temperature. After 5 washes, 100 µl/well TMB-substrate (ThermoFisher) were added, incubated for 20 min at RT in the dark. Absorbance at 652 nm was measured using the Synergy Mx microplate reader.

Overnight cultures of S. aureus were diluted to an OD of 0.1, incubated for further 3.5 h, and diluted to an OD of 1 in 10 ml BHI medium. Bacteria were pelleted, washed with 10 ml ice cold PBS and finally resuspended in 10 ml PBS, and stored at 4°C overnight. Serial dilutions were plated to assess the number of CFUs in these cultures. C57BL/6 mice were anesthetized by intraperitoneal injection of ketamine (125 µg/g body weight, Pfizer), and xylazine (12.5 µg/g body weight, Bayer). The trachea was exposed, and 50 µl of the S. aureus suspension (6x10⁸ CFUs/mouse in PBS) or vehicle control were administered via a 29-gauge needle. In some experiments, 6x10⁸ CFUs were resuspended in PBS containing 50 µg purified SpA (13) (Sigma, P-3838). After 4 h or 24 h respectively, mice were sacrificed, blood was collected, and the lungs lavaged 4 times with 0.7 ml saline. Following lung perfusion with 3 ml PBS, lungs were enzymatically digested with hyaluronidase type I-s, collagenase type XI, and DNase I (Sigma), and filtered through a nylon mesh (70 µm) to achieve single cell suspensions. The number of neutrophils in the bronchoalveolar lavage (BAL) fluid and the lung was determined by using a Sysmex hemacytometer, and further discriminated by flow cytometry using an anti-CD45, an anti-Gr-1, and an anti-Ly6B.2 antibody. CFUs within the BAL, the lung, and the blood were assessed by serial plating. Total protein content within the BAL was determined by using a BCA-assay (Pierce). TNFR- or L-selectin shedding in the BAL or in the plasma was assessed using a DuoSet ELISA (DY425, DY576 R&D) according to the manufacturer’s instructions.

The release of superoxide in response to S. aureus was measured by a cytochrome c reduction assay. Percoll-purified BMDNs were resuspended in HBSS supplemented with 20 mM HEPES at a density of 3.95 x10⁶/ml and pre-incubated at 37°C for 60 min with SpA (Sigma; 1 mg/ml in H₂O; 1 µg/1x10⁵ cells), H₂O (7.2% v/v) or DMSO (0.5% v/v) as vehicle controls, or the matrixmetalloproteinase inhibitor GM6001 (abcam, 50 µM final concentration). Afterwards, 270 µl of BMDNs were well mixed with 30 µl S. aureus with an MOI of 10 and seeded in fibrinogen-precoated (3 h, 0.15 mg/ml HBSS) Immunolon-4HBX 96-well plates (70 µl/well). Cofactors such as CaCl₂ (1 mM), MgCl₂ (1 mM) and cytochrome c (0.1 mM) were added with or without TNFα (0.5 µg/ml) or superoxide dismutase (SOD, Sigma, 45 units). Absorbance was analyzed in a microplate reader (Mx Synergy) at 37°C at 550 nm and 490 nm every 4 min for a total time of 90 min. For calculation of nmoles O₂^- each measurement was corrected by its SOD control measurement and subjected to the Beer-Lambert law.

Results are presented as means ± standard error mean (s.e.m.). Differences between groups were analyzed using Student´s t-test, one-way ANOVA, or one-way ANOVA with Dunnett´s multiple comparisons, where appropriate using GraphPad Prism Software v9.3. A p-value of ≤ 0.05 was considered statistically significant.

To confirm the increased nuclease activity of high₈₁ compared to low₁₇ as described before (21), calf thymus DNA was incubated with the supernatant of S. aureus cultures or BHI medium as a control, respectively. To estimate the activity, the digested DNA was subjected to gel electrophoresis and analyzed for the abundance of DNA fragments at different sizes. The representative image shown in Figure 1A and the densitometry analysis (Figure 1B) clearly demonstrate the significantly lower amount of small DNA fragments following incubation with the supernatant of the low₁₇ isolate, compared to a common USA300 strain or the high₈₁ isolate (Figure 1B).

To determine the capability of murine neutrophils to produce NETs in response to S. aureus, we co-cultivated BMDNs together with the two different CF isolates, and analyzed them via flow cytometry for a positive staining of myeloperoxidase (MPO), a typical component of NETs, and Sytox, which detects extracellular DNA. ImageStream^®X analysis revealed that extracellular fibers were stained positive for both, indicating that S. aureus stimulation of BMDNs induced NET formation (representative image shown in Figure 1C). No significant differences regarding the morphology, staining intensity, or the area of positive signals were observed following incubation with low₁₇ or high₈₁ (data not shown). Following 1 h of co-incubation with S. aureus low₁₇, flow cytometry analysis revealed a significant increase of MPO⁺/Sytox⁺ neutrophils compared to unstimulated control cells, with a further rise after 2 h of incubation (Figure 1D). Following incubation with high₈₁, a significant increase compared to control cells was detected after 2 h, whereas this increment was significantly lower compared to neutrophils incubated with low₁₇ (Figure 1D). Interestingly, incubation of BMDNs with the USA300 strain resulted in significantly more MPO⁺/Sytox⁺ neutrophils compared to the high₈₁ isolate, although they exhibited similar nuclease activities (Supplementary Figure S1A). In contrast to the flow cytometry, which focusses on the analysis of MPO⁺/Sytox⁺ cells still retaining their morphology, the NET specific ELISA detects NET-fragments in the supernatant of the co-cultures. Here, we observed significantly more NET fragments following incubation with S. aureus high₈₁, compared to control cells as well as to S. aureus low₁₇ stimulated neutrophils (Figure 1E), indicating that higher nuclease activities shed the NETs from the neutrophils, leading to higher amounts of NET fragments in the supernatant. To investigate whether the different nuclease activities influence the neutrophil-mediated killing or growth inhibition of S. aureus, we compared the CFUs following co-incubation of BMDNs with these two S. aureus isolates. Here, we observed a significant increase of CFUs of the high₈₁ -isolate after 2 h, in contrast to low₁₇, with only a modest increase of CFUs compared to the amount of seeded CFUs in this co-culture (Figure 1F). Incubation with USA300 resulted in a similar relative growth as observed for the low₁₇ isolate (Supplementary Figure S1B), indicating that strain specific attributes influence growth in this setting. These results suggest that the higher nuclease activity of the high₈₁ isolate diminishes the ability of neutrophils to kill or inhibit growth of S. aureus. To further confirm that the increased amounts of CFUs following incubation with high₈₁ was not due to an increased nuclease-mediated bacterial dispersion, we added DNaseI to these samples and observed no significant differences compared to the CFUs of non-treated samples (Supplementary Figure S1C).

Since this increase of nuclease activity was suggested to be an adaptive response to the neutrophil-rich and inflammatory environment within pwCF, we wanted to analyze the influence of this increased nuclease activity in a model of acute lung infection in mice. The blood was collected, the lung was lavaged, harvested and analyzed for the number of CFUs 24 h after intratracheal instillation with each of these two S. aureus isolates (Figures 1G–I). Interestingly, we observed a significant increase in the number of CFUs for the high₈₁ strain in BAL (Figure 1G), lung (Figure 1H), and blood (Figure 1I). Analyzing the number of neutrophils in the BAL (Figure 1J) or the lung tissue (Figure 1K) revealed that the two isolates do not display any differences in their ability to induce neutrophil recruitment. Intratracheal instillation of PBS as vehicle control causes recruitment of only minor numbers of neutrophils (Supplementary Figure S1D). Furthermore, although we observed increased CFUs for the high₈₁ isolate-infected mice, the protein content in the alveolar space was not affected (Figure 1L). However, although high amounts of bacteria were injected, mice dealt surprisingly well with the induced lung infection and none of the mice died. These observations led us to the question, how these isolates differ from the common MRSA USA300 strain.

To assess the virulence of the CF-isolates low₁₇ and high₈₁, both strains were compared to the USA300 S. aureus strain in a model of acute lung infection. To investigate the early phase following infection, mice were analyzed after 4 h regarding neutrophil recruitment to the alveoli (Figure 2A) and the lung (Figure 2B). Here, we observed less recruited neutrophils in the BAL of USA300-infected mice compared to the CF-isolates, but no differences within the lung tissue (Figure 2B). However, the amount of CFUs in the BAL (Figure 2C) and the lung (Figure 2D) was similar in all mice at this time point. Extending the observation interval to 24 h revealed that infection with USA300 did not induce an altered recruitment of neutrophils to the alveoli (Figure 2E), or the lung (Figure 2F). Interestingly, evaluation of the CFU numbers in the BAL (Figure 2G) or the lung (Figure 2H) revealed a significant increase in USA300 infected mice. Assessing the protein content in the alveolar space after 4 h, no differences were observed (Figure 2I). In contrast, the protein content in the BAL collected from USA300 infected mice 24 h after infection, displayed a significant increase (Figure 2J). In line with these observations, we surveyed significantly more animals with a severe burden, which had to be euthanized for ethical reasons, in those groups infected with USA300 (Figure 2K). These data suggest a reduced virulence of the CF-isolates compared to the USA300 strain.

S. aureus is known for its large variety of proteins contributing to its toxicity and virulence. One of these factors is SpA, which has been demonstrated to induce TNFR-1 shedding, at least partially through ADAM17 (13, 14). This metalloproteinase also influences L-selectin shedding, contributing to the host response during lung infection as demonstrated recently (26). To further elucidate the role of SpA in the immune response against invading S. aureus, we first analyzed the secretion of SpA of the USA300 strain in comparison to the two CF-isolates and observed significantly more SpA release by the USA300 strain compared to both isolates (Figure 3A). Accordingly, investigating the percentage of IgG2a-positive neutrophils, indicative of SpA binding or interaction, in the co-culture with these three S. aureus strains revealed a significant increase of IgG2a-positive cells for the USA300 strain compared to control cells or neutrophils incubated with the CF-isolates low₁₇ and high₈₁ after 2 h (Figure 3B). Incubation of BMDNs with recombinant SpA induced fast shedding of L-selectin from the neutrophil surface (Figure 3C), indicating ADAM17 activity. To confirm the physiological relevance in vivo, we analyzed the BAL and the plasma samples from mice infected with the three S. aureus strains for the abundance of soluble L-Selectin (sL-Sel) using an ELISA assay (Figures 3D–G). Indeed, mice infected with the USA300 strain displayed significantly more sL-Sel in the BAL (Figure 3D), but not in the plasma (Figure 3E) 4 h post infection. Similarly, analyzing the amount of sL-Sel 24 h post infection revealed significantly more sL-Sel in the BAL (Figure 3F), but not in the plasma (Figure 3G) compared to those mice which were infected with low₁₇ or high₈₁. No differences were observed between these two CF-isolates. To further corroborate these findings, we also analyzed the amount of soluble TNFR (sTNFR) in these samples. Interestingly, the abundance of sTNFR in USA300 infected mice in comparison to mice infected with low₁₇ and high₈₁ was significantly increased after 4 h in the BAL (Figure 3H) and plasma (Figure 3I). However, analyzing the infected mice after 24 h revealed a less pronounced, but still significant difference between USA300 and low or high-infected mice in the BAL (Figure 3J), whereas no significant difference was observable in the plasma (Figure 3K).

Co-stimulation of TNFα-primed BMDNs with fibrinogen and USA300 and H₂O (7.2% v/v) as vehicle control induced the release of ROS, whereas neutrophils challenged with the two CF-isolates had a significantly weaker response (Figure 4A). To assess whether the formation of ROS in this assay is influenced by the amount of SpA, the same assay was performed with neutrophils pre-incubated with recombinant SpA (Figure 4B). Here, we observed a significant increase of ROS formation compared to control cells without additional SpA (Figure 4A). The response of neutrophils to both isolates, low₁₇ and high₈₁, is similar compared to neutrophils incubated with USA300 without additional SpA, while the addition of SpA to BMDNs in the presence of USA300 further increased the production of ROS. In order to confirm that the difference between USA300 and low₁₇ or high₈₁ induced ROS formation is dependent on the activation of ADAM17, BMDNs were pre-incubated with the vehicle control DMSO, which did not further influence ROS production (Figure 4C), or the matrix metallopeptidase inhibitor GM6001, which inhibits also ADAM17 activity (Figure 4D). Here, we observed that inhibition of ADAM17 in USA300 stimulated neutrophils reduced ROS production to the levels of those neutrophils challenged with the CF-isolates. Interestingly, no strain-specific or SpA-/GM6001- mediated differences were visible in BMDNs without TNFα priming (Supplementary Figures S2A, B). To exclude that an altered uptake of bacteria might influence ROS production, we performed a phagocytic assay. BMDNs were incubated with the S. aureus isolates and analyzed for the number of intracellular bacteria following different times of incubation, revealing no significant differences (Supplementary Figure S2C).

We further investigated whether the addition of SpA during infection of mice might influence the disease, and worsen the outcome of those mice being infected with CF-isolates low₁₇ and high₈₁. WT mice were either infected with USA300 or one of the clinical CF-isolates low₁₇ and high₈₁ supplemented with 50 µg SpA. 24 h following infection mice were analyzed for neutrophil recruitment to the alveoli (Figure 4E) or the lung (Figure 4F) and their bacterial burden (Figures 4G, H). Similar to previous experiments, we did not observe any significant differences regarding neutrophil recruitment to the alveoli (Figure 4E) or the lung (Figure 4F), although there was a tendency towards an increased neutrophil recruitment in the BAL in response to high₈₁ following supplementation with SpA. However, analyzing the bacterial burden of the BAL (Figure 4G) and the lung (Figure 4H) revealed no significant differences between the USA300 infected mice and those mice which received one of the isolates supplemented with SpA, indicating that SpA abolishes the previously observed difference in virulence between USA300 compared to low₁₇ or high₈₁. Interestingly, the addition of SpA also eliminates the growth advantage of the high₈₁ compared to the low₁₇ isolate, which might also be attributable to the slightly higher number of recruited neutrophils.