Role of Bacterial and Host DNases on Host-Pathogen Interaction during Streptococcus suis Meningitis

Streptococcus suis is a zoonotic agent causing meningitis in pigs and humans. Neutrophils, as the first line of defense against S. suis infections, release neutrophil extracellular traps (NETs) to entrap pathogens. In this study, we investigated the role of the secreted nuclease A of S. suis (SsnA) as a NET-evasion factor in vivo and in vitro. Piglets were intranasally infected with S. suis strain 10 or an isogenic ssnA mutant. DNase and NET-formation were analyzed in cerebrospinal fluid (CSF) and brain tissue. Animals infected with S. suis strain 10 or S. suis 10ΔssnA showed the presence of NETs in CSF and developed similar clinical signs. Therefore, SsnA does not seem to be a crucial virulence factor that contributes to the development of meningitis in pigs. Importantly, DNase activity was detectable in the CSF of both infection groups, indicating that host nucleases, in contrast to bacterial nuclease SsnA, may play a major role during the onset of meningitis. The effect of DNase 1 on neutrophil functions was further analyzed in a 3D-cell culture model of the porcine blood–CSF barrier. We found that DNase 1 partially contributes to enhanced killing of S. suis by neutrophils, especially when plasma is present. In summary, host nucleases may partially contribute to efficient innate immune response in the CSF.


Introduction
Bacterial meningitis is a life-threatening disease [1] in humans and pigs that can be caused by Streptococcus suis [2,3]. The innate immune system immediately reacts against invading bacteria by recruiting neutrophils to the site of infection. Neutrophils are the major immune cell inside the cerebrospinal fluid (CSF) during bacterial meningitis [4]. In addition to phagocytosis, neutrophil granulocytes are known to release neutrophil extracellular traps (NETs), which consist of a DNA backbone decorated with several antimicrobial components [5] that serve to entrap and kill various pathogens. The host itself is known to produce DNases to recycle NET-structures and to avoid tissue-damaging effects of NETs during infectious or non-infectious diseases [6][7][8].
Various bacteria have evolved DNases as NET-evasion factors to escape from NET-entrapment and killing. For S. suis two NET-evading nucleases have been described: EndAsuis and SsnA. The membrane-anchored EndAsuis only contributes to NET-evasion. The released SsnA has additionally been shown to efficiently reduce antimicrobial effects of NETs to improve survival of S. suis in the presence of NETs [9,10]. Furthermore, SsnA is active in CSF during S. suis meningitis [2]. However, the question of whether SsnA contributes to the development of meningitis in piglets as the natural host in vivo has not been investigated so far.
Generally, the specific role of nucleases during bacterial infections is controversially discussed in the literature. A recent publication discussed the infusion of DNase 1 into the CSF [11] as a treatment opportunity during pneumococcal meningitis. The authors observed an enhanced bacterial clearance in the CSF of infected rats by increased phagocytosis and NET degradation when treated with DNase 1. In contrast to this protective effect of DNase 1 in meningitis, it was found that the respiratory tract pathogens Actinobacillus pleuropneumoniae and Haemophilus influenzae survive even better in the presence of neutrophils treated with DNases, because degraded NETs release nutrients as NAD, which enhance bacterial growth [12].
In this study, we investigated the role of the bacterial nuclease SsnA and the role of the host DNase 1 during S. suis meningitis in vivo and in vitro.

SsnA Only Slightly Increases S. suis Virulence in Pigs
The DNase SsnA is known as a major NET-evasion factor of S. suis [10]. To investigate the influence of the bacterial DNase SsnA during S. suis meningitis, we conducted two animal experiments. The first was performed as a survival experiment to examine the influence of the DNase on mortality over a period of 13 days. The second experiment was performed to study the immune response, as well as bacterial load, during the onset of meningitis (early phase). In the survival experiment, slightly higher internal body temperatures occurred in the S. suis 10-infected group than in the 10∆ssnA-infected group ( Figure 1A). In the S. suis 10-infected group, three out of nine animals died and in the 10∆ssnA infected group one out of nine animals died ( Figure 1B and Table S1). The pathohistological score and re-isolation of infection strains was almost similar in both infection groups (Tables S2 and S3) [13]. In the second experiment, we investigated the host-pathogen interaction during the onset of infection. Piglets were either infected with both infection strains or remained uninfected, and were euthanized after 48 or 96 h post-infection. Prior to the experiment, blood was taken and used for assaying the killing of S. suis 10 and 10∆ssnA. Both infection strains showed similar survival rates in this assay ( Figure S1). After intranasal infection, symptoms of CNS disorders were detected in four of 28 infected animals (S. suis 10: one pig; 10∆ssnA: three pigs). A similar morbidity as in the initial survival experiment was found. The internal body temperature of infected animals increased significantly in both groups after infection and a slightly higher body temperature was detected in the S. suis 10-infected group post-infection ( Figure S2 and Table S4). In piglets showing clinical signs of CNS disorders, the infection strains were re-isolated from the brain and in high numbers from the CSF (1.9 × 10 6 to 3.7 × 10 7 CFU/mL). The brain lesion score in this experiment was comparable between both infection groups (Tables S5 and S6). In summary, comparison of morbidity and mortality of both infection strains in both animal experiments leads to the conclusion that SsnA does not contribute to the development of meningitis in piglets.

NETs are Formed in CSF and Consist of PR-39 at the Onset of Meningitis
In a previous study we analyzed pigs from a commercial herd with severe CNS disorders when suffering from S. suis meningitis and detected S. suis entrapped in NETs [2]. Furthermore, NETs were found in the CSF of experimentally S. suis-infected pigs with severe CNS disorders [2,14]. Here we were interested if NETs are also formed in the CSF at the onset of S. suis meningitis and analyzed CSF at defined time-points (48 and 96 h) post-infection. Indeed, we detected NETs in CSF in both infection groups (Figure 2A), indicating that SsnA does not completely destroy NETs in the CSF of infected piglets. Next we analyzed the amount of free DNA, described as a marker for NETs [15], to quantify the NET amount in the two infection groups ( Figure S3). In general, and independent of the infection group, a higher amount of free DNA was detected in CSF compared to serum. Furthermore, NETs were released upon infection by infiltrating neutrophils in CSF in both infection groups in animals with CNS disorders, and no difference was detectable between S. suis 10-and 10ΔssnAinfected piglets. Confirming our own previously published data on pigs from a commercial herd with severe CNS disorders [2], the antimicrobial peptide (AMP) PR-39 was detected in high amounts inside neutrophils in the CSF near the nuclei and in NET-fibers, together with DNA-histones ( Figure  2B). AMPs are known to stabilize NETs against DNases and induce NETs in human and mouse neutrophils [16,17]. Interestingly, already in the onset of meningitis we could quantify a high amount of PR-39 and additional cationic antimicrobial peptides such as PMAP-23 or PMAP-37 in the CSF of piglets with clinical signs of CNS disorder ( Figure 2C and Figure S4). These data indicate that during the onset of meningitis, NETs are formed as an innate immune response against S. suis invasion and are stabilized by host peptides such as PR-39 against degradation by S. suis nuclease.

NETs Are Formed in CSF and Consist of PR-39 at the Onset of Meningitis
In a previous study we analyzed pigs from a commercial herd with severe CNS disorders when suffering from S. suis meningitis and detected S. suis entrapped in NETs [2]. Furthermore, NETs were found in the CSF of experimentally S. suis-infected pigs with severe CNS disorders [2,14]. Here we were interested if NETs are also formed in the CSF at the onset of S. suis meningitis and analyzed CSF at defined time-points (48 and 96 h) post-infection. Indeed, we detected NETs in CSF in both infection groups (Figure 2A), indicating that SsnA does not completely destroy NETs in the CSF of infected piglets. Next we analyzed the amount of free DNA, described as a marker for NETs [15], to quantify the NET amount in the two infection groups ( Figure S3). In general, and independent of the infection group, a higher amount of free DNA was detected in CSF compared to serum. Furthermore, NETs were released upon infection by infiltrating neutrophils in CSF in both infection groups in animals with CNS disorders, and no difference was detectable between S. suis 10-and 10∆ssnA-infected piglets. Confirming our own previously published data on pigs from a commercial herd with severe CNS disorders [2], the antimicrobial peptide (AMP) PR-39 was detected in high amounts inside neutrophils in the CSF near the nuclei and in NET-fibers, together with DNA-histones ( Figure 2B). AMPs are known to stabilize NETs against DNases and induce NETs in human and mouse neutrophils [16,17]. Interestingly, already in the onset of meningitis we could quantify a high amount of PR-39 and additional cationic antimicrobial peptides such as PMAP-23 or PMAP-37 in the CSF of piglets with clinical signs of CNS disorder ( Figure 2C and Figure S4). These data indicate that during the onset of meningitis, NETs are formed as an innate immune response against S. suis invasion and are stabilized by host peptides such as PR-39 against degradation by S. suis nuclease.

NET-Markers Are Present in Meninges, but No NET-Fibers Are Detectable in Brain Tissue
In the case that S. suis is not efficiently eliminated from the CSF by the host, it has the ability to infect the meninges [18]. Accordingly, S. suis was re-isolated from brain swap and CSF of a pig with meningitis (pig4009, Figure 3A). To protect the host against further pathogen dissemination into the brain, neutrophils infiltrate into the meninges. We wondered whether NETs could be detected in meninges and the choroid plexus, which is described as one of the entry sites of S. suis into the CNS [19][20][21]. By immunofluorescence staining of three different brain regions that contain meninges or choroid plexus ( Figure S5), we visualized S. suis and NET markers ( Figure 3).

NET-Markers Are Present in Meninges, but No NET-Fibers Are Detectable in Brain Tissue
In the case that S. suis is not efficiently eliminated from the CSF by the host, it has the ability to infect the meninges [18]. Accordingly, S. suis was re-isolated from brain swap and CSF of a pig with meningitis (pig4009, Figure 3A). To protect the host against further pathogen dissemination into the brain, neutrophils infiltrate into the meninges. We wondered whether NETs could be detected in meninges and the choroid plexus, which is described as one of the entry sites of S. suis into the CNS [19][20][21]. By immunofluorescence staining of three different brain regions that contain meninges or choroid plexus ( Figure S5), we visualized S. suis and NET markers ( Figure 3). NET-markers are present in brain tissue of piglets with meningitis in situ, but no NET-fibers are detectable. (A) In the staining of formalin-fixed brain tissue from pigs out of survival experiments, NET markers and S. suis were visualized, but no NET fibers. * cortex cerebri and # inflammation in meninges (blue = counterstaining of DNA (DAPI); green = DNA/histone-1-complexes (NETs); red = S. suis (white arrow); iso = isotype control; 4009 = animal with meningitis from this survival experiment study; 9899, 9884 and 9807 = pigs with meningitis from other survival experiment study [2,14] (scale bar upper panel = 50 µm; lower panel = 20 µm). (B) Representative hematoxylin-eosin (HE) stained tissues of meningitis show neutrophils invading from vessels into the meninges. This picture was observed in both infection groups of the early phase experiment. NET markers (green = DNA/histone-1-complexes (NETs) and yellow = neutrophil elastase) and Streptococci (red) were found in brain tissue of these animals in both infection groups; however, as in the survival experiment, no NET fibers were detectable. Representative pictures are shown; white arrows = S. suis. The pictures for both infection groups present slides that were sliced successively (scale bar HE, isotype and S. suis 10 = 50 µm, S. suis 10ΔssnA = 20 µm). Respective isotype controls were used to adjust the settings.
In the case of the early phase experiment, we analyzed serial cuts, and DNA-histone and elastase as NET-markers, as well as S. suis, could be detected in both infection groups linked to neutrophils in the meninges. Importantly, no NET fibers were detectable. Hematoxylin-eosin (HE)-staining identified neutrophils in the choroid plexus and meninges only in animals with CNS disorders Figure 3. NET-markers are present in brain tissue of piglets with meningitis in situ, but no NET-fibers are detectable. (A) In the staining of formalin-fixed brain tissue from pigs out of survival experiments, NET markers and S. suis were visualized, but no NET fibers. * cortex cerebri and # inflammation in meninges (blue = counterstaining of DNA (DAPI); green = DNA/histone-1-complexes (NETs); red = S. suis (white arrow); iso = isotype control; 4009 = animal with meningitis from this survival experiment study; 9899, 9884 and 9807 = pigs with meningitis from other survival experiment study [2,14] (scale bar upper panel = 50 µm; lower panel = 20 µm). (B) Representative hematoxylin-eosin (HE) stained tissues of meningitis show neutrophils invading from vessels into the meninges. This picture was observed in both infection groups of the early phase experiment. NET markers (green = DNA/histone-1-complexes (NETs) and yellow = neutrophil elastase) and Streptococci (red) were found in brain tissue of these animals in both infection groups; however, as in the survival experiment, no NET fibers were detectable. Representative pictures are shown; white arrows = S. suis. The pictures for both infection groups present slides that were sliced successively (scale bar HE, isotype and S. suis 10 = 50 µm, S. suis 10∆ssnA = 20 µm). Respective isotype controls were used to adjust the settings.
In the case of the early phase experiment, we analyzed serial cuts, and DNA-histone and elastase as NET-markers, as well as S. suis, could be detected in both infection groups linked to neutrophils in the meninges. Importantly, no NET fibers were detectable. Hematoxylin-eosin (HE)-staining identified neutrophils in the choroid plexus and meninges only in animals with CNS disorders ( Figure 3B and Figure S6). In good correlation to findings in the CSF, we found PR-39-positive neutrophils in the choroid plexus. Nevertheless, PR-39 was only detectable inside neutrophils and no NET fibers as shown in CSF were detectable ( Figure 4A and Figure S7). In animals without symptoms of CNS disorders, we could neither detect PR-39 signals nor neutrophils in the brain. Fibers seen in HE-staining are not positive for DNA-histone, elastase or PR-39, and may therefore consist of fibrin instead of NETs.
( Figure 3B and Figure S6). In good correlation to findings in the CSF, we found PR-39-positive neutrophils in the choroid plexus. Nevertheless, PR-39 was only detectable inside neutrophils and no NET fibers as shown in CSF were detectable ( Figure 4A and Figure S7). In animals without symptoms of CNS disorders, we could neither detect PR-39 signals nor neutrophils in the brain. Fibers seen in HE-staining are not positive for DNA-histone, elastase or PR-39, and may therefore consist of fibrin instead of NETs.
Based on this finding that NET fibers occur in CSF but not in the meninges of both infection groups, we conclude that the bacterial DNase SsnA does not play a major role in the destruction of NETs in this immune-privileged compartment. Thus, we assume that the host itself seems to be responsible for modulating the formation and destruction of NETs by the production of host DNases.  Based on this finding that NET fibers occur in CSF but not in the meninges of both infection groups, we conclude that the bacterial DNase SsnA does not play a major role in the destruction of NETs in this immune-privileged compartment. Thus, we assume that the host itself seems to be responsible for modulating the formation and destruction of NETs by the production of host DNases.

Host DNases Are Active during S. suis Meningitis in Brain Tissue, CSF and Serum
Host DNases eliminate NETs to avoid tissue damage [8]. To prove our hypothesis of the involvement of host DNases during S. suis meningitis, we analyzed brain tissue by immunofluorescence microscopy. More DNase 1 signaling was detected in the choroid plexus of animals with clinical meningitis than in those animals without clinical meningitis ( Figure 4A).
To generate quantitative data, we analyzed serum and CSF for DNase activity ( Figure 4B,C). In serum was almost no DNase activity in the control group, but significantly higher DNase activity was detectable in both infection groups ( Figure 4B), independent of whether S. suis 10 or 10∆ssnA was used to infect the animals. Similar in CSF no nuclease activity was detected in the control group ( Figure 4C) but significantly higher DNase activity was detected 96h post-infection in both the CNS. S. suis 10 infected animals sacrificed 96h post-infection had a significantly higher DNase activity than the group euthanized after 48h ( Figure 4C and Figure S8). High amounts of specific DNase 1 were detected in animals with clinical signs of CNS disorders 96 h post-infection in CSF by ELISA ( Figure S9).
The in vivo experiments indicate a minor influence of the bacterial DNase on the host-pathogen interaction during S. suis meningitis. To understand the role of host DNases, we conducted in vitro experiments.

DNase 1 Has No Impact on Transmigration of Neutrophils through a Choroid Plexus Epithelial Cell Layer
The blood-CSF barrier is described for S. suis as an entry site [21,22]. We investigated the role of host DNase 1 on host-pathogen interaction using a model of the porcine blood-CSF barrier [23]. This model was adapted for S. suis 10 infection and neutrophil transmigration ( Figure 5A). The barrier integrity was confirmed by transepithelial electrical resistance (TEER) measurement and dextran flux ( Figure S10). Neutrophils spontaneously transmigrated through this barrier and interleukin 8 (IL8) enhanced the transmigration of neutrophils ( Figure 5B). In contrast to IL8, S. suis 10 or DNase 1 or a combination of the two, did not increase the transmigration rate under the chosen 4-h incubation time ( Figure 5B and Figure S11). There was no efficient elimination of bacteria detectable whether DNase 1 was present or not ( Figure 5C).
As a further optimization of this model to mimic the in vivo situation as close as possible, we used whole blood in the upper blood-compartment and porcine CSF in the lower CSF-compartment instead of medium ( Figure 6). The barrier integrity ( Figure S12) and NET formation were confirmed using this optimized model ( Figure 7A).  As a further optimization of this model to mimic the in vivo situation as close as possible, we used whole blood in the upper blood-compartment and porcine CSF in the lower CSF-compartment instead of medium ( Figure 6). The barrier integrity ( Figure S12) and NET formation were confirmed using this optimized model ( Figure 7A). Again, using this optimized model, the transmigration rate of neutrophils into the IL8 stimulated CSF-compartment was not affected by DNase 1 ( Figure 7B). However, the number of transmigrated neutrophils seems to depend on the individual blood donor, i.e., characteristic or number of neutrophils present in the blood.  Again, using this optimized model, the transmigration rate of neutrophils into the IL8 stimulated CSF-compartment was not affected by DNase 1 ( Figure 7B). However, the number of transmigrated neutrophils seems to depend on the individual blood donor, i.e., characteristic or number of neutrophils present in the blood. To confirm the influence of DNase 1, we have chosen a more sensitive in vitro system and performed a neutrophil killing assay with the phagocytosis-sensitive S. suis capsule mutant (10cpsΔEF) [24,25]. Using immunofluorescence staining, neutrophils were confirmed to phagocytose and to release NETs in the same experiment ( Figure 8A). Without plasma (absence of opsonization factors) a tendency for less growth inhibition in the presence of DNase 1 was detected, indicating NET-mediated killing in the absence of plasma. In the presence of 10% autologous plasma a significant decrease of bacteria was detected with DNase 1 compared to neutrophils without DNase 2.6. DNase 1 Partially Improves Killing of S. suis by Neutrophils of Individual Animals in CSF As DNase 1 did not increase neutrophil transmigration, we analyzed if DNase 1 might influence the antimicrobial activity of the neutrophils as previously shown [11]. Interestingly, in 50% of the blood donors an increased killing of S. suis 10 was detectable in the presence of DNase 1; however, this difference was not statistically significant ( Figure 7C,D). A statistical correlation between transmigrated neutrophils and bacteria revealed the following phenomenon: the more neutrophils migrated through the barrier, the fewer bacteria survived in the CSF. Importantly, this correlation was only significant in the presence of DNase 1 ( Figure 7E,F). Thus, we assume that DNase 1 might partially enhance the killing activity of neutrophils.
To confirm the influence of DNase 1, we have chosen a more sensitive in vitro system and performed a neutrophil killing assay with the phagocytosis-sensitive S. suis capsule mutant (10cps∆EF) [24,25]. Using immunofluorescence staining, neutrophils were confirmed to phagocytose and to release NETs in the same experiment ( Figure 8A). Without plasma (absence of opsonization factors) a tendency for less growth inhibition in the presence of DNase 1 was detected, indicating NET-mediated killing in the absence of plasma. In the presence of 10% autologous plasma a significant decrease of bacteria was detected with DNase 1 compared to neutrophils without DNase 1 treatment ( Figure 8B). These data confirm that DNase 1 contributes to enhancement of neutrophil antimicrobial activity, as previously shown [11], but only in the presence of autologous plasma. 1 treatment ( Figure 8B). These data confirm that DNase 1 contributes to enhancement of neutrophil antimicrobial activity, as previously shown [11], but only in the presence of autologous plasma. Immunofluorescence staining of NETs and intra-and extracellular S. suis. NET production and phagocytosis exist in parallel during neutrophil killing assay. With plasma, less bacteria can be seen than without plasma. More Streptococci are extracellular than intracellular as the single channels visualize (blue = DNA (Hoechst); magenta = DNA/histone1-complexes (NETs, white arrowheads); red and green = extracellular S. suis; green = intracellular S. suis (white arrows)). Representative pictures are shown (scale bar = 50 µm, zoom = 10 µm). (B) The effect of DNase 1 on killing of S. suis 10cpsΔEF was proven with a neutrophil killing assay with or without 10% plasma of the blooddonating pig. After 2 h incubation with plasma, an increased killing of S. suis was detectable with DNase 1. Without plasma, the CFU/mL was significantly higher than with plasma. The best killing effect was detectable with plasma and DNase 1. Data shown as mean ± SD, n = 5 independent experiments. Statistical analysis: paired students t-test (*p < 0.05, **p < 0.01).

Discussion
The aim of our study was to investigate the role of the bacterial nuclease SsnA and the host DNase 1 during S. suis meningitis in vivo and in vitro. Bacteria developed nucleases in evolutionary biology for phage defense [26]. Since the principle of defense by DNA destruction was very effective, it can be hypothesized that it was also adopted by mammalian cells or perhaps even transferred from bacteria into these cells by endosymbiosis; for example, mitochondrial endonuclease G, which has (A) Immunofluorescence staining of NETs and intra-and extracellular S. suis. NET production and phagocytosis exist in parallel during neutrophil killing assay. With plasma, less bacteria can be seen than without plasma. More Streptococci are extracellular than intracellular as the single channels visualize (blue = DNA (Hoechst); magenta = DNA/histone1-complexes (NETs, white arrowheads); red and green = extracellular S. suis; green = intracellular S. suis (white arrows)). Representative pictures are shown (scale bar = 50 µm, zoom = 10 µm). (B) The effect of DNase 1 on killing of S. suis 10cps∆EF was proven with a neutrophil killing assay with or without 10% plasma of the blood-donating pig. After 2 h incubation with plasma, an increased killing of S. suis was detectable with DNase 1. Without plasma, the CFU/mL was significantly higher than with plasma. The best killing effect was detectable with plasma and DNase 1. Data shown as mean ± SD, n = 5 independent experiments. Statistical analysis: paired students t-test (* p < 0.05, ** p < 0.01).

Discussion
The aim of our study was to investigate the role of the bacterial nuclease SsnA and the host DNase 1 during S. suis meningitis in vivo and in vitro. Bacteria developed nucleases in evolutionary biology for phage defense [26]. Since the principle of defense by DNA destruction was very effective, it can be hypothesized that it was also adopted by mammalian cells or perhaps even transferred from bacteria into these cells by endosymbiosis; for example, mitochondrial endonuclease G, which has homologous nucleases in bacteria [27]. Homologous nucleases of SsnA also occur in other bacterial species such as Streptococcus mutans, Streptococcus pyogenes, Bacillus anthracis, Vibrio cholerae or Aeromonas hydophila [28]. In in vitro experiments, SsnA is described as a NET evasion factor and the strongest phenotype was found with human neutrophils [10]. Most S. suis field isolates found in inner organs of pigs express SsnA [28]; therefore, we analyzed if SsnA is necessary to cause meningitis in piglets in vivo. Two independent infection studies with piglets showed that the SsnA mutant was as virulent as the wild-type strain (Figure 1 and Figure S2). Several animals in both infection groups from the early phase experiment developed clinical signs of CNS disorders between 48 h and 96 h post-infection. In addition, S. suis strains were re-isolated from different inner organs of piglets without CNS disorders (Tables S3  and S6). In these piglets, bacteria breached the mucosal barrier and caused bacteremia. Thus, as piglets in both infection groups showed disease symptoms and developed meningitis (Tables S4 and S5), it may be concluded that SsnA does not play a role in the pathogenesis. It is also conceivable that the disadvantage of the lack of nuclease can be compensated by the presence of other virulence factors.
Independently of the S. suis DNase activity, we detected NETs inside the CSF already in the early phase of infection ( Figure 2) and were able to confirm findings from our previous study in the later phase of infection [2]. Neutrophils in CSF during the early phase of infection are able to release NETs containing PR-39 (Figure 2A,B). Human and murine NETs are stabilized against DNase degradation by the human cathelicidin LL-37 or the murine cathelicidin CRAMP [17,29]. In pigs PR-39 was also identified to protect DNA and NETs against degradation [2]. Here we showed in vivo that PR-39 in neutrophils is already present during transmigration through the choroid plexus epithelium in the early onset of meningitis. Furthermore, we confirmed a high amount of PR-39 in CSF as well as in neutrophils and NETs in the CSF compartment ( Figure 2). For LL-37, a loss of its antimicrobial function against Pseudomonas aeruginosa was described when bound to DNA fibers [30]. Further studies are needed to clarify if the function of PR-39 in NET-structures is to kill microbes or only to prevent their spread by stabilizing NETs, or both. Furthermore, the function of other peptides such as PMAP-23 or PMAP-37, which are present in infected piglets ( Figure S4), has not yet been investigated.
As S. suis and neutrophils invade and infiltrate the meninges, we confirmed NET markers in situ by immunofluorescence staining in meninges of pigs with CNS disorders (Figure 3). However, it is remarkable that no NET fibers, as seen in CSF, were detectable in situ in the meninges, not even in pigs infected with the 10∆ssnA strain. Based on our findings, we assumed that the host itself releases high amounts of DNases to destroy NETs and thereby reduces the possible tissue-damaging effect of NETs in the meninges [8,31]. Indeed, in the choroid plexus, DNase 1 was only found in animals with CNS disorders ( Figure 4A). Interestingly, DNase was detectable in the choroid plexus in the area of transmigrating neutrophils. Furthermore, animals with meningitis showed high amounts of DNase 1 in CSF ( Figure S9). A high DNase activity was also found in serum and in CSF independently of whether piglets were infected with S. suis 10 or 10∆ssnA (Figure 4). In humans, DNase 1 alone is not able to destroy NETs completely; other host nucleases like DNase 1L3 and TREX1 are additionally needed [32]. The question of whether this finding also applies to pigs needs further investigation, but is still technically challenging for pigs based on lack of available tools.
Recently, an enhanced bacterial clearance in the CSF of S. pneumoniae-infected rats by increased antimicrobial activity of neutrophils when treated with DNase 1 was observed [11]. To further investigate the effect of DNase 1 on neutrophil activities against S. suis, we adapted a model of the blood-CSF barrier [23], with integrity during infection and transmigration of neutrophils ( Figure S10). As no effect on neutrophil transmigration and killing of S. suis by adding DNase 1 was found in the classical model using cell culture medium and isolated neutrophils ( Figure 5), we established a physiologically-improved blood-CSF barrier model with porcine CSF and whole blood in the respective compartments ( Figure 6). To investigate if DNase 1 enhances the bacterial clearance in the CSF model, S. suis was directly applied with the same CFU/mL into the CSF. Therefore, S. suis was not transmigrating through the cell layer, and neutrophil transmigration was triggered with IL8. It is known that S. suis induces IL8 production in porcine choroid plexus epithelial cells [33], and an IL8 concentration that can be reached during meningitis [34] was added to the CSF compartment. In bacterial meningitis of children a significantly increased IL8 level was also described [35]. Efficient neutrophil transmigration was present in this optimized model, but the supplementation of DNase 1 did not cause any differences ( Figure 7B). However, in the case of approximately 50% of the blood donors, an increased killing of S. suis 10 was detectable in the presence of DNase 1 ( Figure 7D). Importantly, a statistical correlation between transmigrated neutrophils and bacterial killing was only confirmed when DNase 1 was added to the experimental setting ( Figure 7F). Thus, these data lead to the assumption that DNase 1 does not alter the transmigration of neutrophils but might contribute to enhanced killing activity. These data are in line with the recently published phenotype that DNase 1 triggers antimicrobial activity of neutrophils during Streptococcus pneumoniae CNS infections in rats. The authors proposed that intact NETs are not beneficial during acute pneumococcal meningitis and that pneumococcal strains with low DNase-activity remain unaffected by NETs [11]. Treatment with DNase 1 released the bacteria from the NETs, and an increased level of neutrophil phagocytosis was found at the same time. In our study, an increased killing by neutrophils was only confirmed for the phagocytic sensitive capsule mutant of S. suis in the presence of DNase 1 and autologous plasma ( Figure 8B). Without autologous plasma, the phenotype of increased killing changed in the other direction, assuming that NETs play a major role in the absence of plasma components and phagocytosis of S. suis is increased if NETs are degraded and opsonin factors are present. However, when using a capsulated S. suis 10 strain that exhibits high DNase activity, this phenomenon of DNase 1-mediated enhancement of neutrophil killing was not confirmed in our study and thus may depend highly on the bacterial strain and its virulence factors ( Figure 7C,D). As the combination of plasma and DNase led to the best antimicrobial effect of neutrophils, this could also explain the difference between the two cell culture models used in our study. Although in the cell culture with purified neutrophils and the absence of plasma the killing effect was not influenced by DNase 1, an effect was detectable with whole blood. It is conceivable that next to neutrophils, plasma components of the whole blood can pass through the barrier into the CSF compartment in high amounts when barrier integrity is lost [36,37]. One possible factor is the complement factor C1q. The factor itself has no DNase activity, but it can highly increase serum DNase 1 activity [38]. Interestingly, it was found that complement factors and common complement pathway factors were increased in CSF of patients with bacterial meningitis, compared to controls [39].
As a natural host of S. suis, we used blood from male and female pigs with differing genetic backgrounds and of different ages, which were kept under common husbandry conditions. We thus hypothesize that the immune status of each individual pig might influence the results and explain the high variability among the individual pigs during cell culture experiments. In contrast to our study, the in vivo study with pneumococci was conducted with adult male Sprague Dawley rats, thus originating from a closed outbreed colony [11]. It is likely that they were kept under laboratory animal housing conditions, as usual, and therefore that all rats reacted in a similar fashion. Therefore, in future studies the influence of different blood donors should be investigated more in detail, in order to reflect the variable situations in conventional husbandry conditions. This might be transferred to the variable conditions in humans as the main reason for high inter-individual differences in innate immune reactions.
In summary, our results indicate that during S. suis meningitis the effects of host DNases seem to be more relevant than the effects of bacterial DNase. S. suis SsnA does not contribute to the pathogenesis of meningitis in piglets. However, host nucleases such as DNase 1 may participate in an efficient innate immune response by partially triggering the killing of S. suis by neutrophils when plasma is present. By immunofluorescence staining, we were able to show that DNase 1 is produced in the choroid plexus during acute meningitis ( Figure 4A). Perhaps DNases are also stored in CSF for rapid deployment, eventually bound to actin to stay inactive until needed [40]. Whether treatment with DNase 1 is beneficial in S. suis meningitis, as discussed for pneumococcal meningitis [11] or used in cystic fibrosis [41][42][43], needs to be further investigated.
However, independently of nuclease production, both antimicrobial functions of neutrophils, phagocytosis, as well as NET-formation, are found in vivo in the CSF of infected piglets and may contribute to bacterial elimination. In our in vitro model S. suis was detectable extra-and intracellularly, as found in CSF of animals with CNS disorders ( Figure S13). Interestingly, the production of "vital" NETs that are still able to be chemotactic active and phagocytize, are actually controversially discussed in the literature in comparison to "suicidal" NET release [44,45]. This phenomenon might explain our findings, as well as Mohanty's findings that DNase treatment of neutrophils increases its antimicrobial functions by cleaving NET fibers and thereby supporting intracellular uptake of bacteria. Further studies should clarify the role of "vital" versus "suicidal" NETs in inhibiting the spread of S. suis in CSF and clarify the role of host DNases on neutrophil functions in greater detail to understand the underlying mechanisms.
Strains were cultivated out of frozen glycerol stocks on Columbia Agar with 7% Sheep Blood (Thermo Scientific™ PB5008A).
For the infection of piglets, bacterial strains were grown to early stationary phase in tryptic soy broth (TSB) without dextrose (Becton Dickinson, 286220, Franklin Lakes, NJ, USA) at 37 • C with 5% CO 2 . The 40 mL portions of the suspension were centrifuged and after discarding the supernatant bacteria were suspended in 3 mL PBS, leading to a bacterial concentration of 4 × 10 9 CFU/mL. The exact CFU/mL was determined by plating serial dilutions on blood agar plates (Columbia Agar with 7% Sheep Blood; Thermo Scientific™ PB5008A, Waltham, MA, USA).
For in vitro assays, working cryostocks of bacterial strains were generated by growing the bacterial strains to early stationary phase in TSB without dextrose (Becton Dickinson, 286220, Franklin Lakes, NJ, USA) at 37 • C with 5% CO 2 . Aliquots with 15% Glycerol (Sigma Aldrich, 13487-2, St. Louis, MO, USA) were frozen in liquid nitrogen and stored at −80 • C.

Experimental Infection of Piglets
All pigs used in this study were castrated male German Landrace piglets from a herd known to be free of S. suis 10 but not free of S. suis in general. Handling and treatment of animals was in strict accordance with the principles of the European Convention for the Protection of Vertebrate Animals Used for Experimental and Other Scientific Purposes, as well as the German Animal Protection Law. The animal experiments were approved by the Committee on Animal Experiments of the Lower Saxonian State Office for Consumer Protection and Food Safety (Niedersächsisches Landesamt für Verbraucherschutz und Lebensmittelsicherheit) under permit numbers 33.14-42502-04-12/0965 (last update 2013) and 33.12-42502-04-12/0991 (last update 10/2017). Post-infection, animals were monitored every eight hours for internal body temperature, food uptake, behavior, breathing, posture, motion sequence and lameness.

Survival Experiment
Eighteen 4-5-week-old pigs were divided into two groups. After pre-disposition with 1% acetic acid under anesthesia (azaperon 2 mg/kg body weight (Stresnil, Elanco GmbH, Cuxhaven, Germany), ketamine 10 mg/kg bodyweight (Ursotamin, Serumwerke Bernburg, Bernburg, Germany)), the animals in the respective groups were intranasally challenged with 1 × 10 9 CFU S. suis 10 or 10∆ssnA, as previous described [13]. The health status of the animals was monitored every 8 h. A piglet was classified as morbid if a body temperature of ≥40.2 • C or/and severe clinical signs of an acute disease were observed. In case of high fever (≥40.5 • C), apathy and anorexia persisting over 36 h, as well as in all cases of central nervous system dysfunction or clinical signs of acute polyarthritis, animals were euthanized for reasons of animal welfare. All surviving piglets were sacrificed 13 days post-infection (dpi). Histological samples for Figure 3A were taken from the same pigs as the already-published CSF samples [2]. All information about the experiment conducted under permission number 33.14-42502-04-12/0965 was published elsewhere [14].

Early Phase Experiment
Thirty-nine 8-week-old piglets were divided into three groups and infected with 6 × 10 9 CFU S. suis 10, 10∆ssnA or mock infected with PBS (uninfected) as described above.

Sample Collection
During anesthesia, before infection, swabs of the tonsils, heparin blood and serum samples were collected. Piglets in the survival experiment were sacrificed as described above, 13 days post-infection at the latest. In the early phase experiment, we sacrificed the first half of each group (uninfected 5/11, In both experiments the animals were anesthetized for euthanasia as already described and heparin blood and serum samples were collected. After euthanasia with T61 i.v., CSF samples were taken by puncturing of the cisterna magna [48]. During section swabs of the brain surface, the mitral valve, the pleura, the pericard and the peritoneum were taken for bacteriology. In addition, for bacteriology and histology, organ samples of the brain, liver, spleen, tonsils, heart, lung, as well as carpal and tarsal joints, were collected. Samples for histology were stored immediately in 10% buffered formalin.

Re-Isolation of Infection Strains
Organ and swab samples collected during section of the animals were spread onto blood agar plates, with tonsils additionally on selective agar for Streptococci (Oxoid, PB5049A, Waltham, MA, USA). Isolated S. suis strains were identified by matrix-assisted laser desorption ionization time-of-flight mass spectrometry (MALDI-TOF MS). Isolation of the challenge strain was confirmed by multiplex PCR detecting mrp, epf, sly, arcA, gdh, cps1, cps2, cps7 and cps9 [49], as well as PCR amplifying the ssnA gene, to verify the knock-out in the respective isolates. Primers used for ssnA detection are listed in Table S7 and cycling consisted of 30-times denaturation at 94 • C for 1 min, annealing at 58 • C for 45 s, elongation at 68 • C for 5 min and final elongation at 68 • C for 5 min.

Blood Survival Assay
Survival in the blood was determined by infecting 0.5 mL porcine whole blood with 3 × 10 5 CFU either S. suis 10 or 10∆ssnA for 2 h. During infection, the samples were rotated on a rotator at 8 rounds per minute in an incubator at 37 • C. At time-points 0 h and 2 h the CFU was determined by plating serial dilutions on blood agar plates (Thermo Scientific™ PB5008A, Waltham, MA, USA). The survival factor was calculated by dividing the number of CFUs after 2 h of incubation by the number of CFUs at 0 h.

Pico Green Quantification Assay
A Pico Green quantification assay (Quant-iT™ PicoGreen ® ; Invitrogen, Carlsbad, CA, USA) was performed to determine free DNA in CSF and serum, as described previously [12].

Histology of the Brain
Immediately after extraction, the entire brain was fixed in 10% formalin (buffered) for a maximum of 72 h. The regions of the corpus striatum, hippocampus and cerebellum, the latter two containing the choroid plexus, were embedded in paraffin and cut into 2-4 µm sections for hematoxylin-eosin (HE) staining and immunofluorescence staining.

Cytospin of CSF
At a maximum of 30 min after collection of CSF, 100 µL CSF of all non-infected animals and of animals without fever or altered CSF were transferred to CellView ® slides (Greiner bio-one 543079, Kremsmünster, Austria). From all animals with CNS disorders, fever or altered CSF, cells in CSF were counted. The cell number was adjusted with HBSS to 1 × 10 4 to 1 × 10 5 cells/100 µL and cells were added to the CellView ® slides. The slides were centrifuged at 370 g for 5 min at room temperature and fixed with a final concentration of 4% paraformaldehyde (PFA).

DNase Activity Test
To 50 µL serum or CSF of each pig, 1 µg deoxyribonucleic acid sodium salt from calf thymus (Sigma, D3664, St. Louis, MO, USA) was added. The samples were incubated at 37 • C for 20 h. After incubation, samples were run on an 1% electrophoresis gel containing Roti ® -GelStain (Roth, 3865.1, Karlsruhe, Germany) at 100 V for 30 min. Bands were detected with Biorad ChemiDoc MP (Hercules, California) and the degradation of the calf thymus DNA was classified into 4 groups between no (0) and total (3) degradation, by eye ( Figure S8D).

Cell Culture
Porcine choroid plexus epithelial cells (PCP-R) [50] were seeded on the underside of filter inserts with 3 µm pores (Greiner bio-one, 662631, Kremsmünster, Austria) and transmigration assays were performed as previously described [23], with the following modifications.

•
Cell Culture Model of the Porcine Blood-CSF Barrier ( Figure 5A): After three days' incubation, transepithelial electrical resistance (TEER) was measured with a Millicell ® ERS-2 (Millipore, Billerica, MA, USA) voltmeter to control the barrier density. Some filters were infected with S. suis 10 for one hour in the blood compartment. Before adding the freshly isolated porcine neutrophils, medium in the blood compartment was changed in all filters. Transmigration of neutrophils was performed for 4 h at 37 • C and 5% CO 2 . Serial dilutions of transmigrated bacteria were plated on blood agar (Columbia Agar with 7% Sheep Blood; Thermo Scientific™ PB5008A, Waltham, MA, USA) to determine the CFU/mL and 250 µL medium of the CSF compartment was fixed with a final concentration of 4% PFA (Paraformaldehyde 16% Science Services E15710, Munich, Germany) for flow cytometric analysis. TEER was measured again. The experiment was repeated three times on independent days.

Flow Cytometry
Two hundred microliters of PFA-fixed CSF samples from the cell culture experiments were measured in an Attune NxT Flow Cytometer (Invitrogen, Carlsbad, CA, USA) with forward and sideward scatter using a 488 nm laser. Data were analyzed with FlowJo TM version 10.6.1 (Ashland, OR, USA) software by drawing a gate to exclude the debris.

Statistical Analysis
Data were analyzed using Excel 2010 and 2016 (Microsoft, Albuquerque, NM, USA). Statistical analyses were performed with GraphPad Prism Version 8.0.1 (San Diego, CA, USA). A detailed description of tests used can be found in the figure legends, but the p-value has the same settings in all graphs (* p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001).  Table S1: Assessment of morbidity and mortality after intranasal infection in the survival experiment; Table S2: Scoring of fibrinosuppurative lesions of piglets in the survival experiment; Table S3: Re-isolation of the infection strains after intranasal infection of piglets in the survival experiment; Table S4: Assessment of morbidity after intranasal infection in the early phase experiment; Table S5: Scoring of brain lesions of piglets in the early phase experiment; Table S6: Re-isolation of the infection strains after intranasal infection of piglets in the early phase experiment; Table S7: Primers used for ssnA detection; Figure S1: No differences in survival of S. suis 10 and 10∆ssnA in blood of piglets before infection; Figure S2: Development of body temperature in the early phase experiment; Figure S3: Free DNA, as a marker for NETs, increases in CSF of animals with clinical meningitis but not in serum post-infection; Figure S4: ELISA of porcine antimicrobial peptides PMAP-23 and PMAP-37 in CSF; Figure S5: Regions of sagittal cuts from porcine brain for histology analysis; Figure S6: Single channels of immunofluorescence staining from Figure 3B; Figure S7: Single channels of immunofluorescence staining from Figure 4A; Figure S8: Gel electrophoresis gels of the DNase activity assay ( Figure 4C); Figure S9: Host nuclease DNase1 is detectable in CSF during S. suis meningitis; Figure S10: The barrier integrity of PCP-R cells was tight during S. suis 10 infection and neutrophil transmigration; Figure S11: Impact of IL8 on transmigration of isolated porcine neutrophils through a cell layer of porcine choroid plexus epithelial cells; Figure S12: The barrier integrity of PCP-R cells was tight during S. suis 10 infection and neutrophil transmigration in the more physiological cell culture system; Figure S13: More S. suis are found extracellularly than intracellularly in infected CSF of piglets with clinical meningitis.