ABSTRACT
Bacteremia is a hallmark of invasive Streptococcus suis infections of pigs, often leading to septicemia, meningitis, or arthritis. An important defense mechanism of neutrophils is the generation of reactive oxygen species (ROS). In this study, we report high levels of ROS production by blood granulocytes after intravenous infection of a pig with high levels of S. suis-specific antibodies and comparatively low levels of bacteremia. This prompted us to investigate the working hypothesis that the immunoglobulin-mediated oxidative burst contributes to the killing of S. suis in porcine blood. Several S. suis strains representing serotypes 2, 7, and 9 proved to be highly susceptible to the oxidative burst intermediate hydrogen peroxide, already at concentrations of 0.001%. The induction of ROS in granulocytes in ex vivo-infected reconstituted blood showed an association with pathogen-specific antibody levels. Importantly, inhibition of ROS production by the NADPH oxidase inhibitor apocynin led to significantly increased bacterial survival in the presence of high specific antibody levels. The oxidative burst rate of granulocytes partially depended on complement activation, as shown by specific inhibition. Furthermore, treatment of IgG-depleted serum with a specific IgM protease or heat to inactivate complement resulted in >3-fold decreased oxidative burst activity and increased bacterial survival in reconstituted porcine blood in accordance with an IgM-complement-oxidative burst axis. In conclusion, this study highlights an important control mechanism of S. suis bacteremia in the natural host: the induction of ROS in blood granulocytes via specific immunoglobulins such as IgM.
INTRODUCTION
Streptococcus suis colonizes the mucosal surfaces of healthy pigs but can also become an invasive pathogen causing meningitis, septicemia, arthritis, endocarditis, and pneumonia in growing piglets (1). The species S. suis is very heterogeneous, comprising 29 serotypes and more than 700 sequence types (2). The prevalence of serotypes depends on the geographical region. In Europe, most invasive S. suis isolates belong to serotypes 9, 2, and 7 (3). Despite pigs being the main hosts, S. suis can also cause disease in other mammals and is a zoonotic pathogen, having so far been responsible for more than 1,600 human cases worldwide (4, 5). During infection, S. suis encounters neutrophils as the first line of defense not only in the bloodstream but also in infected tissue. Whereas numerous virulence factors, such as the polysaccharide capsule or the d-alanylation of lipoteichoic acid, have been investigated with regard to their potential to prevent phagocytosis by neutrophils (6–8), few studies have been conducted to elucidate the mechanisms by which these important immune cells actually kill S. suis. Specific antibodies and complement have been shown to promote the killing of S. suis in whole blood or by isolated neutrophils (6, 9). Neutrophils are known to produce reactive oxygen species (ROS) in a process called the oxidative or respiratory burst (10). The oxidative burst is regarded as a rapid and extremely effective antimicrobial defense mechanism of the innate immune system against invading pathogens (11, 12). Catalase-negative streptococci are generally sensitive to ROS but have developed evasion mechanisms to circumvent the oxidative burst attack, as described previously for Streptococcus pneumoniae and S. pyogenes (13–15). The first enzyme in the oxidative burst cascade is the NADPH oxidase (NOX), which is a multicomponent enzyme complex with a catalytic core (gp91phox and p22phox), located in the cell membrane or the membrane of phagosomes, vesicles, and specific granules, and with regulatory subunits (Rac2, p67phox, p47phox, and p40phox) located in the cytosol of resting neutrophils. Once the neutrophil is stimulated, the cytosolic NADPH oxidase subunits translocate to the membrane-located core units to activate the NADPH oxidase. The active enzyme complex converts NADPH to NADP+ and 2O2−. These oxygen radicals are immediately transformed to hydrogen peroxide by the superoxide dismutase (SOD). The final bactericidal products of the oxidative burst, the hypohalites OCl− and SCN−, are generated by the activity of the myeloperoxidase and kill bacteria by oxidizing their proteins, DNA, and lipids (10). ROS exert their bactericidal activity both within the phagolysosome and also extracellularly when they are released by bursting neutrophils (16). Whereas the binding of immune complexes (ICs) or of antibody-labeled antigen to Fc receptors and integrin receptors can stimulate the oxidative burst directly, cytokine receptor, Toll-like receptor, or G-protein-coupled receptor binding only primes neutrophils and requires additional signaling for robust burst induction (16). ROS production has more far-reaching consequences than the direct killing of bacteria by oxygen radicals since it can also induce so-called neutrophil extracellular trap (NET) formation, the release of granules, and the production of proinflammatory cytokines (16). This combination leads to the increased killing of invading pathogens. Knowledge of the oxidative burst response to S. suis is very limited. Previous studies focused on the role of superoxide dismutase A (sodA) and NADH oxidase (NOX) of S. suis serotype 2 in susceptibility to oxidative stress and survival within murine macrophages and reported susceptibility to ROS intermediates only for the respective deletion mutants (17, 18).
In this study, we investigated the oxidative burst response of porcine granulocytes to different S. suis strains representing the three most common European serotypes to elucidate whether the oxidative burst is involved in the killing of S. suis. The role of IgM and complement in oxidative burst induction and the killing of S. suis was analyzed using a defined hyperimmune serum.
RESULTS
ROS production by blood granulocytes of an experimentally infected piglet with high S. suis-specific antibody levels and limited bacteremia.Since the oxidative burst in neutrophils is an important strategy of the immune system to attack invading pathogens, we wanted to investigate whether the oxidative burst is induced in vivo in blood granulocytes following intravenous (i.v.) infection of two pigs with S. suis cps2 strain 10. Blood samples were drawn at different time points to assess the time course of ROS production and bacterial loads. Oxidative burst analysis in granulocytes was performed directly in blood samples by staining with ROS-sensitive dihydrorhodamine 123 (DHR123). Oxidative burst measurements in the blood of pig A revealed no clear increase in the frequency of rhodamine 123-positive (Rho123+) granulocytes at 0.5 h or 13 h postinfection. Only a slight increase from 2.8 to 6.2% Rho123+ cells was detectable at 16 h postinfection (Fig. 1A). At the same time, pig A developed progressive bacteremia, starting with 3.4 × 104 CFU/ml blood (30 min postinfection) and reaching 5.2 × 105 CFU/ml blood at 19 h postinfection (Fig. 1B). The pig developed a fever of 40.7°C at 10 h postinfection. In contrast, pig B did not show clinical signs of S. suis disease but had a robust oxidative burst response in blood at 13 h postinfection (21.0% Rho123+ granulocytes), representing an almost 7-fold increase in comparison to the sample taken prior to infection (preinfection, 3.1% Rho123+ cells) (Fig. 1A). At 16 h postinfection, the oxidative burst response decreased to 8.7% Rho123+ granulocytes. Moreover, S. suis could be isolated from the blood of this piglet at all time points, but the CFU count in the blood of this piglet was remarkably lower than for the first piglet, with a maximum count of 3.8 × 104 CFU/ml at 16 h postinfection (Fig. 1B).
The oxidative burst of porcine granulocytes in response to S. suis infection in vivo corresponds to lower bacterial burdens in blood at later time points of infection. Two piglets were intravenously infected with 3 × 108 CFU of S. suis strain 10. Blood samples were drawn immediately before infection and at 0.5 h, 13 h, 14 h, 16 h, and 19 h postinfection. (A) The oxidative burst of porcine granulocytes was determined by DHR123 (Rho123) staining and flow cytometry. The Rho123 signal is depicted on the x axis. At each time point, PMA stimulation was used as a positive control. SSC, side scatter; FSC, forward scatter. (B) The CFU in blood were determined by plating of serial dilutions and plotted alongside the Rho123 signal. (C) Anti-S. suis strain 10 IgM and IgG antibody levels were determined in the serum of both piglets and are listed as relative ELISA units (EU) per milliliter.
S. suis-specific IgM and IgG levels in the sera of the two piglets before infection were measured by an enzyme-linked immunosorbent assay (ELISA). Pig A had S. suis-specific IgM and IgG levels (46.7 and 46.5 ELISA units [EU]/ml, respectively) that were substantially lower than the S. suis-specific IgM and IgG levels of pig B (55.4 and 276.4 EU/ml, respectively) (Fig. 1C).
In summary, induction of ROS was shown in vivo in blood granulocytes of a pig experimentally infected with S. suis. Based on the high levels of S. suis-specific antibodies and the controlled bacteremia in this pig, we postulated a crucial role for the antibody-induced oxidative burst of blood granulocytes in the control of S. suis bacteremia.
Growth of S. suis is impaired in the presence of hydrogen peroxide.Based on the in vivo data, we hypothesized that S. suis is susceptible to hydrogen peroxide, an important intermediate of the oxidative burst of phagocytic cells. We therefore investigated the susceptibility of four S. suis strains representing three different serotypes and an unencapsulated mutant to hydrogen peroxide. The growth of all five investigated S. suis strains was severely impaired in the presence of hydrogen peroxide concentrations of ≥0.01% (Fig. 2). A hydrogen peroxide concentration of 0.001% led to delayed growth of all investigated S. suis strains. The unencapsulated S. suis mutant 10cpsΔEF showed a susceptibility to hydrogen peroxide comparable to those of the four wild-type (wt) strains. Since all strains were susceptible to hydrogen peroxide, we tested if they possessed the sodA gene, which had previously been described to play a role in defense against oxidative stress (17). PCR analysis revealed that all five investigated S. suis strains were sodA positive (see Fig. S1A in the supplemental material). Furthermore, quantitative PCR (qPCR) of reverse-transcribed RNA confirmed that all four investigated wt strains of serotypes 2, 7, and 9 as well as the unencapsulated mutant 10cpsΔEF expressed sodA mRNA (Fig. S1B). We thus conclude that the ROS hydrogen peroxide impairs the growth of different S. suis cps types at concentrations as low as 0.001% despite sodA mRNA expression.
Growth of S. suis is impaired in the presence of hydrogen peroxide. S. suis strains 10, A3286/94, 16085/3b, and 13-00283-02, representing three different serotypes (cps2, cps9, and cps7), as well as a capsule-deficient S. suis mutant (10cpsΔEF) were grown in THB medium. After 2 h, hydrogen peroxide was added at the indicated concentrations (percent, weight per volume). The OD600 was measured every hour and is depicted on the y axis. Data from one representative experiment out of four are shown.
The oxidative burst of porcine granulocytes in response to S. suis strain 10 is induced in the presence of S. suis-specific antibodies.Granulocytes express both high- and low-affinity Fc receptors that are involved in the recognition and phagocytosis of antibody-opsonized pathogens but also in the induction of ROS production (19). To test our hypothesis that S. suis-specific antibodies are associated with the induction of the oxidative burst in blood granulocytes, we performed in vitro assays to define the factors involved in more detail. We therefore investigated the oxidative burst response of porcine granulocytes to S. suis in the presence of various antibody concentrations by using previously characterized porcine sera. As shown in Fig. 3A, the oxidative burst level in reconstituted blood infected with S. suis strain 10 was significantly higher in the presence of hyperimmune serum than in the presence of serum containing moderate anti-S. suis IgM and IgG antibody levels (22% ± 3.96% versus 5.36% ± 1.25% Rho123+ cells). The oxidative burst level was also significantly higher upon S. suis infection of reconstituted blood containing moderate antibody serum levels (5.36% ± 1.25% Rho123+ cells) than in colostrum-deprived piglet serum (CDS) (0.45% ± 0.13% Rho123+ cells). Noteworthy, the high oxidative burst response in blood reconstituted with hyperimmune serum depended on the presence of S. suis (22% ± 3.96% infected versus 5.11% ± 4.27% uninfected Rho123+ cells). An oxidative burst was not detectable in blood reconstituted with CDS containing only natural antibodies, even in the presence of S. suis (0.45% ± 0.13% Rho123+ cells). In fact, values were comparable to those for uninfected samples supplemented with CDS (0.61% ± 0.11% Rho123+ cells) and significantly lower than the oxidative burst response to S. suis strain 10 in the presence of hyperimmune serum. Thus, ROS induction in blood granulocytes was not detectable in the absence of S. suis-specific antibodies.
Antibody-mediated induction of the oxidative burst in porcine granulocytes in response to S. suis. (A) S. suis strain 10 was incubated with three sera, serum from colostrum-deprived piglets (CDS) (free of IgG), serum containing moderate levels of IgM and IgG antibodies against S. suis strain 10 (moderate Ab serum), and anti-S. suis strain 10 hyperimmune serum. Subsequently, the whole-blood cell pellet of healthy donor piglets was added, and ROS production in blood granulocytes was measured by flow cytometry (n = 3). Bars and error bars represent means and standard deviations, and significant differences are indicated (**, P < 0.01; ***, P < 0.001). (B) Anti-S. suis IgM and IgG antibody levels in the chosen experimental sera, listed as ELISA units. The prime-booster hyperimmune serum was defined to include 100 ELISA units of IgG and 100 ELISA units of IgM. Note that ELISA units for the different Ig classes are therefore not comparable.
We therefore conclude that specific antibodies are crucial for the induction of an oxidative burst in porcine blood granulocytes in response to S. suis.
Inhibition of the serum-induced oxidative burst in porcine granulocytes leads to increased survival of S. suis.We next wanted to find out if the oxidative burst response to S. suis in the presence of hyperimmune serum is also involved in the killing of S. suis. We therefore used the NADPH oxidase inhibitor apocynin (Apo) in oxidative burst and bactericidal assays. Control experiments confirmed that apocynin alone does not influence the growth of S. suis in Todd-Hewitt broth (THB) at the concentration used (Fig. S2A). The immune sera used for the assays described below were previously tested for their ability to mediate the killing of different S. suis strains. As shown in Fig. 4A, the oxidative burst responses to S. suis strain 10, 10cpsΔEF, A3286/94, 16085/3b, and 13-00283-02 in the presence of hyperimmune sera were significantly reduced by the addition of apocynin. The inhibition of the oxidative burst by apocynin led to significantly increased survival factors (SFs [calculation of the values is described in the legend for Fig. 4B]) for S. suis strains 10 (cps2), 16085/3b (cps9), and 13-00283-02 (cps7), as demonstrated in Fig. 4B. The bacterial survival factors were approximately 3-fold higher in these strains in the presence of apocynin, indicating an important bactericidal role of ROS generated in blood granulocytes by NADPH oxidase activity. The addition of apocynin also resulted in increased bacterial survival of the unencapsulated strain 10cpsΔEF and the cps9 wt strain A3286/94, although the differences were smaller and not significant. Noteworthy, the mean survival factor for wt strain 10 was lower than the respective value for its unencapsulated mutant 10cpsΔEF in the absence of apocynin (Fig. 4B), and the inhibitory effect of apocynin on ROS generation was greater in the case of the cps2 wt strain 10 (Fig. 4A). These results suggest that an important portion of the antibodies in this hyperimmune serum, raised against cps2 strain 10 inducing ROS in response to strain 10, was directed against capsular polysaccharides. As capsular polysaccharides are T lymphocyte-independent antigens inducing mainly IgM antibodies, we specifically focused on IgM in the course of this study.
Inhibition of NADPH oxidase by apocynin leads to a decreased oxidative burst in porcine granulocytes and increased survival of S. suis in blood reconstituted with hyperimmune sera, without affecting granulocyte viability. (A) Oxidative burst of granulocytes in porcine blood in response to in vitro infection with different S. suis serotypes with or without the NADPH oxidase inhibitor apocynin. The oxidative burst induced by S. suis and hyperimmune sera was set at 100% (control [ctr]). The reduction in the oxidative burst by the addition of apocynin (n = 5 to 6) is depicted as a percentage of the oxidative burst without the inhibitor. (B) Survival factors of S. suis serotypes 2, 9, and 7 in the presence of hyperimmune sera with or without the addition of the NADPH oxidase inhibitor apocynin as the quotients of CFU per milliliter after 1 h and CFU per milliliter directly after in vitro infection with S. suis (n = 5 to 6). (C) The viability of granulocytes after the bactericidal assay (B) was measured by flow cytometry by staining with the fixable viability dye eF506 (n = 3). Bars and error bars represent means and standard deviations. Statistical analyses between controls and the respective apocynin-treated samples were done by an unpaired t test (n = 5 to 6/group) (*, P < 0.05; **, P < 0.01; n.s., not significant).
To ensure that the increased survival of S. suis by the addition of apocynin was not due to a toxic effect of the inhibitor on phagocytic granulocytes, we investigated the viability of these cells subsequent to the bactericidal assay. The viability of porcine granulocytes was not reduced by the addition of the NADPH oxidase inhibitor, neither in samples with S. suis nor in samples with serum only (Fig. 4C). Viability dye staining results were validated using additional blood samples that were stressed by heat treatment (Fig. S2B). To demonstrate a direct impact of apocynin on granulocytes, we additionally conducted an assay with isolated granulocytes in 20% hyperimmune serum to analyze the oxidative burst and bacterial survival. We observed a reduction of the oxidative burst and an increase in bacterial survival (Fig. S2C) by apocynin treatment, as shown in Fig. 4A and B for the reconstituted blood assays. This shows that oxidative burst induction in granulocytes does not require additional cellular interactions and that apocynin has a direct impact on granulocytes.
Taken together, these experiments demonstrate that the immune serum-induced oxidative burst in blood granulocytes is an important part of the immune response leading to the killing of S. suis.
Complement is involved in oxidative burst induction of porcine granulocytes in response to S. suis.To investigate the involvement of complement in the induction of the oxidative burst and killing of streptococci, we used vaccinia virus complement control protein (VCP), which was previously described to be an effective complement inhibitor in porcine blood (9, 20). Bactericidal and oxidative burst assays with S. suis strain 10 were conducted in the presence of VCP, apocynin, or the combination of both inhibitors. In these assays, porcine blood cells were resuspended in hyperimmune serum raised against S. suis cps2 strain 10. As shown in Fig. 3, this hyperimmune serum induced high frequencies of Rho123+ granulocytes (oxidative burst rate) in the presence of S. suis strain 10 (23.81% ± 4.72% [n = 6]) in comparison to the control without S. suis (5.065% ± 2.26% [n = 6]) (Fig. 5A). In infected samples, VCP treatment significantly reduced the frequencies of Rho123+ granulocytes (11.17% ± 3.60% [n = 6]). As expected, apocynin also reduced the oxidative burst (7.58% ± 2.50% [n = 3]). The combination of VCP and apocynin led to highly significantly lower frequencies of Rho123+ granulocytes (3.33% ± 2.31% [n = 6]). However, the bacterial survival factors increased significantly only in the presence of apocynin (4.03 ± 1.73 for Apo [n = 3] versus 0.02 ± 0.01 for the control [n = 6]) or by the combination of VPC and apocynin (4.84 ± 2.63 [n = 6]). Survival factors were only marginally increased by VCP alone (0.15 ± 0.11 [n = 6]), as demonstrated in Fig. 5B. From these results, we conclude that complement is involved in ROS induction. However, in the presence of complete hyperimmune serum (containing S. suis-specific IgG and IgM), the killing of streptococci also involves complement-independent mechanisms.
Complement partially mediates oxidative burst induction but is not crucial for the killing of S. suis strain 10 in blood reconstituted with hyperimmune serum raised against cps2 strain 10. Hyperimmune serum raised against S. suis strain 10 was used to reconstitute porcine blood subsequently infected with S. suis strain 10 to analyze the oxidative burst (shown as percentages of Rho123+ granulocytes) (A) and bacterial survival (shown as survival factors) (B) in the untreated control sample (ctr) or in the presence of the complement inhibitor VCP (100 μg/ml), the NADPH oxidase inhibitor apocynin (Apo) (1.5 mM), or a combination of both. Survival factors were determined after 2 h at 37°C after in vitro infection with 2 × 106 CFU/ml S. suis strain 10. To investigate the S. suis-induced oxidative burst, the same sample treatments were conducted in the absence of S. suis (n = 3). PMA (0.1 μg/ml) was used as a positive control in oxidative burst experiments. Bars and error bars represent means and standard deviations. For statistical analysis, the Kruskal-Wallis test with Dunn’s multiple-comparison test was used (n = 6). All S. suis in vitro-infected groups (black brackets) or only the subgroups of controls versus VCP and VCP plus Apo (gray brackets) were included in statistical analyses (*, P < 0.05; **, P < 0.01; ***, P < 0.001).
We observed a stronger inhibitory effect of apocynin in the cps2 wt- than in the 10cpsΔEF-infected sample (Fig. 4A). Since capsule-specific antibodies are primarily of the IgM isotype, this observation encouraged us to investigate the role of IgM in the induction of ROS. To eliminate the effects of IgG, hyperimmune serum against S. suis strain 10 was depleted of total IgG. The removal of IgG from serum was verified by affinity chromatography (Fig. S3). As the IgG depletion process goes along with a high dilution and loss of serum components, the IgG-depleted hyperimmune serum was subsequently concentrated and supplemented with 10% CDS. As shown in Fig. 6A, we observed an S. suis-induced oxidative burst in the absence of IgG (8.52% ± 2.63% for the control with S. suis strain 10 [n = 6] versus 3.16% ± 0.85% for the control with no S. suis [n = 3]) but at a level lower than that with complete hyperimmune serum. In the absence of IgG, the impact of VCP on the oxidative burst and on bacterial survival was stronger than that in complete hyperimmune serum. VCP treatment reduced the oxidative burst rate >4.5-fold (1.88% ± 0.59% [n = 6]) (Fig. 6A). Apocynin treatment alone again reduced the oxidative burst significantly (0.99% ± 0.25% [n = 6]), as did the combination of both inhibitors (1.06% ± 0.31% [n = 6]). Interestingly, the inhibition of complement in IgG-depleted hyperimmune serum had a significant effect on bacterial survival: whereas the untreated control still mediated bacterial killing (survival factor, 0.09 ± 0.10 [n = 6]), VCP treatment resulted in a significantly increased survival factor (3.99 ± 2.38 [n = 6]) (Fig. 6B). Similarly, treatment with apocynin also led to a significant increase of the survival factor (4.29 ± 1.14 [n = 6]). In contrast to complete serum, the combination of apocynin and VCP increased bacterial survival immensely in IgG-depleted serum (17.38 ± 4.56 [n = 6]). This additive effect of both inhibitors in the absence of IgG suggests that besides the complement-induced oxidative burst axis, killing activities of complement and the oxidative burst also work independently of each other (Fig. 6B).
Influence of complement on the oxidative burst (A) and the survival of cps2 S. suis strain 10 (B) in porcine blood reconstituted with IgG-depleted hyperimmune serum. IgG-depleted hyperimmune serum raised against cps2 strain 10, supplemented with 10% CDS, was used to reconstitute porcine blood samples subsequently infected with S. suis. The oxidative burst (shown as percentages of Rho123+ granulocytes) (A) and bacterial survival (shown as survival factors) (B) were analyzed in the presence of the complement inhibitor VCP (100 μg/ml), the NADPH oxidase inhibitor apocynin (Apo) (1.5 mM), or a combination of both as well as in an untreated sample (ctr). Bacterial survival factors were determined after 2 h at 37°C after in vitro infection with 2 × 106 CFU/ml S. suis strain 10. To determine S. suis-specific mechanisms of oxidative burst induction, the same sample treatments were conducted in the absence of S. suis (n = 3). PMA (0.1 μg/ml) was used as a positive control in oxidative burst experiments. Bars and error bars represent means and standard deviations. For statistical analysis, the Kruskal-Wallis test with Dunn’s multiple-comparison test was used (n = 6). All S. suis in vitro-infected groups (black brackets) or only the subgroups of controls versus VCP and VCP plus Apo (gray brackets) were included in statistical analyses (*, P < 0.05; **, P < 0.01; ***, P < 0.001).
In summary, complement is strongly involved in the IgG-independent induction of the oxidative burst in blood granulocytes of an S. suis-immunized pig. Our results suggest that the role of complement for bacterial killing is especially important in the absence of or at low levels of antigen-specific IgG.
Cleavage of IgM by rIdeSsuis_h reduces the oxidative burst and increases the survival of the S. suis cps2 strain in porcine blood in a partially complement-independent manner.To further characterize the role of IgM and the possible IgM-complement-oxidative burst axis that might mediate the killing of S. suis, we used IgG-depleted hyperimmune serum but now in combination with IdeSsuis, a protease highly specific for porcine IgM, derived from S. suis itself (21). Noteworthy, previous bactericidal assays showed that the expression of IdeSsuis by S. suis is not sufficient to cleave most of the IgM in porcine blood (9). Consequently, we conducted bactericidal and oxidative burst experiments with IgG-depleted hyperimmune serum after cleaving IgM with a recombinant IdeSsuis homologue (rIdeSsuis_h) (21), a truncated variant of the enzyme containing the active center. To reveal accidental side effects caused by potential contamination with endotoxin due to the expression of recombinant IdeSsuis_h in Escherichia coli, we used rIdeSsuis_h that was heat inactivated at 95°C (rIdeSsuis_h_95°C) as a negative control as well as the loss-of-function point-mutated variant rIdeSsuis_h_C195S (9). Complete cleavage of IgM in IgG-depleted hyperimmune serum through the addition of rIdeSsuis_h was proven by Western blotting, as demonstrated in Fig. S4. The killing of S. suis strain 10 in the bactericidal assay with blood reconstituted with the IgG-depleted hyperimmune serum was primarily mediated via IgM (Fig. 7A, left), as shown by the addition of rIdeSsuis_h. Survival factors of S. suis strain 10 in IgG-depleted serum (control, 0.03 ± 0.02 [n = 3]) were considerably increased by the addition of rIdeSsuis_h (26.08 ± 5.69 [n = 3]) but not by the addition of the nonfunctional rIdeSsuis variants rIdeSsuis_h_95°C (0.01 ± 0.002 [n = 3]) and rIdeSsuis_h_C195S (0.01 ± 0.01) (Fig. 7A, left). Simultaneously, the oxidative burst response to S. suis strain 10 (control, 7.55% ± 2.16% [n = 3]) was significantly reduced by the addition of the IgM protease rIdeSsuis_h (2.05% ± 0.25% [n = 3]), in contrast to the addition of rIdeSsuis_h_95°C (6.81% ± 1.30% [n = 3]) and rIdeSsuis_h_C195S (8.62% ± 1.50% [n = 3]) (Fig. 7B, left). Additionally, complement inactivation (heat inactivation for 30 min at 56°C) decreased the oxidative burst regardless of the integrity of IgM (Fig. 7B, middle). These results demonstrate the relevance of the IgM-complement-oxidative burst axis as a mechanism for the killing of S. suis in the blood of a bacterin-immunized piglet, independent of IgG. Interestingly, in bactericidal assays using heat-inactivated IgG-free serum, treatment with rIdeSsuis_h also resulted in a substantial increase of the bacterial survival factor. This indicates that the killing activity of IgM does not depend solely on complement (Fig. 7A, middle).
Impact of IgM on killing of S. suis (A) and induction of the oxidative burst in granulocytes (B) in porcine blood reconstituted with IgG-depleted hyperimmune serum. Shown are bacterial survival factors (A) and the oxidative burst responses of granulocytes (B) in S. suis strain 10-infected porcine blood reconstituted with IgG-depleted hyperimmune serum (raised against cps2 strain 10) pretreated with PBS as a control (ctr) or with the IgM protease IdeSsuis_h or its nonfunctional control rIdeSsuis_h_95°C (heat inactivated) or rIdeSsuis_h_C195S. Particular serum treatments were used without further treatment (left), were heat inactivated (56°C for 30 min) (middle and left) for complement inactivation, or were additionally treated with the NADPH oxidase inhibitor apocynin (1.5 mM) (right). The oxidative burst was measured by flow cytometry and is depicted as a percentage of Rho123+ cells of all granulocytes. Bars and error bars represent means and standard deviations, and significant differences are indicated. Statistical significances were calculated using the Kruskal-Wallis test with Dunn’s multiple-comparison test (n = 3 to 6) (*, P < 0.05; **, P < 0.01).
To investigate whether the complement-independent bactericidal activity of IgM is still mediated via the oxidative burst, samples with the NADPH oxidase inhibitor apocynin were included. Inhibition of the oxidative burst led to a substantial increase of the bacterial survival factor of strain 10 in blood reconstituted with IgG-depleted and heat-inactivated serum, as shown in Fig. 6B. However, treatment of reconstituted IgG-depleted, complement-inactivated, apocynin-treated blood with rIdeSsuis_h again led to a higher survival factor for S. suis strain 10 than for the respective samples preincubated with nonfunctional rIdeSsuis constructs, although these differences were not significant (Fig. 7A, right). This suggests that the killing or at least inhibition of proliferation of S. suis by IgM is not mediated only via the complement system or the oxidative burst.
In summary, for blood reconstituted with IgG-depleted serum of a bacterin-vaccinated piglet, we showed IgM-mediated killing of S. suis in association with a pronounced induction of ROS. Complement inactivation and/or IgM degradation in this serum led to an overall reduction in ROS production corresponding to increased survival factors for S. suis. These results are in accordance with an IgM-complement-oxidative burst axis mediating the killing of S. suis, although IgM-mediated but complement-independent mechanisms were also observed (see Fig. 10).
Phagocytosis of S. suis in hyperimmune serum is primarily linked to ROS production and reduced by complement inactivation.In order to investigate the phagocytosis of S. suis in the context of ROS production in granulocytes, we used S. suis labeled with CellTrace far-red (FR) fluorescent dye in combination with DHR123 in assays with reconstituted blood to quantify complete phagocytosis or the subset showing both phagocytosis and ROS production (Fig. 8A, left).
Influence of complement and NADPH oxidase inhibition on the rate of phagocytosis of porcine granulocytes in blood reconstituted with hyperimmune serum. (A) Blood reconstituted with hyperimmune serum raised against cps2 strain 10 was infected with CellTrace far-red (FR)-labeled S. suis strain 10, and samples were analyzed by flow cytometry to estimate the oxidative burst response (Rho123 signal) within phagocytic granulocytes. (B) Complete phagocytosis rate depicted as a percentage of S. suis FR-positive granulocytes. Serum was untreated, heat inactivated (56°C for 30 min) to inhibit complement, and/or treated with the NADPH oxidase inhibitor apocynin (1.5 mM). Bars and error bars represent means and standard deviations, and significant differences are indicated. Statistical significances were calculated using the Kruskal-Wallis test with Dunn’s multiple-comparison test (n = 7) (*, P < 0.05; ***, P < 0.001).
Blood reconstituted with hyperimmune serum and infected with FR-labeled streptococci showed that 16.61% ± 4.73% of granulocytes were positive for engulfed fluorescent streptococci. The majority (89%) of these granulocytes (equal to 14.48% ± 4.58% of the total granulocytes) showed simultaneous ROS production (Fig. 8A and B). These data confirm ROS induction in association with the phagocytosis of S. suis in porcine blood granulocytes. However, subsets of granulocytes showed either exclusively phagocytosis (control, 2.13% ± 0.65%) or exclusively ROS production (control, 8.64% ± 5.90%). Consequently, the phagocytosis assay allows a more precise analysis of S. suis-specific ROS induction. Heat treatment of hyperimmune serum reduced the percentages of both granulocytes with engulfed streptococci and granulocytes with detectable phagocytosis and ROS induction, to similar extents (Fig. 8B). This indicates that complement-mediated phagocytosis is associated with ROS induction. Apocynin treatment reduced the percentage of granulocytes showing ROS induction in association with phagocytosis significantly (1.75% ± 0.76% versus 14.48% ± 4.58%). However, the complete phagocytosis rate was also reduced in apocynin-treated samples albeit to a lesser extent (7.9% ± 1.9% versus 16.61% ± 4.73%), as shown in Fig. 8A and B. To prove the specific reduction of ROS production by apocynin in porcine granulocytes, we additionally treated phorbol myristate acetate (PMA)-stimulated blood samples with apocynin. We detected a clear reduction of the Rho123 signal by apocynin treatment even when PMA, as a strong stimulus for ROS induction, was used (Fig. S2D). In samples reconstituted with complement-inactivated hyperimmune serum, the addition of apocynin led to a statistically significant decrease of complete phagocytosis (1.85% ± 1.04%) and particularly to a strong reduction in ROS-producing phagocytic cells (0.41% ± 0.35%) compared to the untreated control.
Our results indicate that the majority of apocynin-treated blood granulocytes having phagocytosed S. suis do not produce ROS (Fig. 8B). This supports the conclusion that apocynin is an effective inhibitor of the oxidative burst in granulocytes with demonstrated phagocytic activity. However, since apocynin treatment leads not only to reduced ROS production but also to reduced phagocytosis, the increase of bacterial survival by this inhibitor might be a result of a combination of reduced ROS production and phagocytosis. Recently, it was shown for porcine alveolar macrophages that NADPH oxidase-dependent ROS generation has a positive effect on phagocytosis via mitogen-activated protein kinase (MAPK) activation and that apocynin treatment efficiently reduces this ROS-mediated stimulation of phagocytosis (22). Whether this also applies to porcine granulocytes is not clear, but inhibition of phagocytosis through apocynin treatment has also been described for human granulocytes (23).
The detection of S. suis-specific phagocytosis in combination with ROS confirmed our above-described findings of complement-mediated induction of the oxidative burst.
IgM-mediated phagocytosis of fluorescently labeled S. suis cells is associated with ROS production and decreased bacterial survival.We further investigated IgM-mediated effects on phagocytosis, ROS induction, and bacterial survival using FR-labeled S. suis and a new preparation of IgG-depleted hyperimmune serum. Noteworthy, fluorescently labeled streptococci showed significantly reduced survival factors in blood reconstituted with CDS (Fig. S5), indicating diminished viability, which makes it difficult to compare bacterial survival data with the results of the experiments with unlabeled bacteria described above. Using IgG-depleted hyperimmune serum for bactericidal assays with reconstituted blood, we observed reductions of phagocytosis and the oxidative burst in comparison to blood reconstituted with complete hyperimmune serum (Fig. 9). Cleavage of IgM with IdeSsuis or treatment with VCP significantly reduced the phagocytosis rate from 4.6% ± 2.8% to 0.2% ± 0.2% and 0.4% ± 0.2%, respectively. This indicates that phagocytosis in the absence of specific IgG is mainly driven by IgM and complement. Since the phagocytosis rates were not significantly different between control samples with cleaved IgM, VCP-treated samples containing intact IgM, and VCP-treated samples with cleaved IgM, we conclude that IgM-mediated phagocytosis is mainly driven by complement activation. The increased phagocytosis in the presence of intact IgM and active complement was associated with a percentage of ROS-positive granulocytes that was at least 30-fold higher than that in the samples with cleaved IgM or complement inhibition (2.6% ± 1.9% versus 0.02% ± 0.02% and 0.08% ± 0.06%, respectively) (Fig. 9A). Determination of bacterial survival confirmed the IgM-mediated killing of streptococci under these conditions, as IdeSsuis treatment led to an increase of the survival factor from 0.1 ± 0.2 (control) to 2.7 ± 1.1 (IdeSsuis), indicating efficient killing and bacterial proliferation, respectively (Fig. 9B). Inhibition of complement by VCP treatment resulted in only a moderate increase of the bacterial survival factor to 0.2 ± 0.1, which is significantly lower than the bacterial survival factor in blood with cleaved IgM.
Impact of IgM and complement on phagocytosis and ROS production of porcine granulocytes in blood reconstituted with IgG-depleted hyperimmune serum. Fluorescently labeled S. suis strain 10 bacteria were used to infect porcine blood reconstituted with IgG-depleted hyperimmune serum (raised against cps2 strain 10) pretreated with PBS as a control (PBS) or with the IgM protease IdeSsuis_h to analyze the phagocytosis rate and oxidative burst within phagocytic granulocytes (A) and bacterial survival (B). Serum was left without inhibitor (ctr), treated with VCP (100 μg/ml) for complement inactivation, and/or treated with the NADPH oxidase inhibitor apocynin (1.5 mM). The phagocytosis rate and oxidative burst response of granulocytes were measured by flow cytometry and are depicted as percentages of S. suis FR-positive and Rho123+ cells of all granulocytes, respectively. Bars and error bars represent means and standard deviations, and significant differences are indicated. Statistical significances were calculated using the Kruskal-Wallis test with Dunn’s multiple-comparison test (n = 6) (marked with brackets) and the Wilcoxon test for each group of samples versus the IdeSsuis-treated equivalent (without brackets) (*, P < 0.05; **, P < 0.01; ***, P < 0.001). n.d., not determined.
Apocynin treatment reduced the frequency of ROS-positive granulocytes containing S. suis 37-fold, from 2.6% ± 1.9% to 0.07% ± 0.09%, whereas the phagocytosis rate was reduced only 5.8-fold, from 4.6% ± 2.9% to 0.8% ± 0.6%. A significant increase in the bacterial survival factor to 1.0 ± 0.2 was observed in apocynin-treated samples, whereas the combination of VCP and apocynin treatment resulted in a survival factor of 0.4 ± 0.2. However, as the killing of S. suis under these conditions was IgM dependent and apocynin-treated samples had significantly higher bacterial survival factors, we conclude that IgM-dependent ROS induction is important for the killing of S. suis. This might include a putative direct effect of ROS generation on the phagocytosis rate.
DISCUSSION
The killing of invading bacteria via the oxidative burst of phagocytic cells is an important defense mechanism of innate immunity (10, 24). However, knowledge of the susceptibility of S. suis to ROS and the contribution of the oxidative burst to the killing of this important pathogen remains limited. As shown in this study, S. suis is highly susceptible to oxidative stress. We demonstrate the antibody-mediated induction of the oxidative burst in blood granulocytes with a partial contribution from complement. The involvement of the oxidative burst in the killing of S. suis was demonstrated by using an NADPH oxidase inhibitor.
Although streptococci are catalase negative, alternative antioxidant systems have been described for S. pneumoniae and S. pyogenes (13, 14). Whether S. suis is sensitive to H2O2, a potent bactericidal intermediate of the oxidative burst, was investigated in a previous study. Using H2O2 concentrations of up to 0.04%, Y. Tang et al. concluded that S. suis wt serotype 2 strains are not impaired in growth by the addition of hydrogen peroxide to brain heart infusion broth (17). Our results, on the other hand, demonstrate a clear impact of as little as 0.001% hydrogen peroxide on the growth of S. suis in THB, added after a 2-h growth phase without hydrogen peroxide. Using virulent S. suis strains of three different serotypes, we show that susceptibility to hydrogen peroxide is not limited to a certain strain or serotype, whereas previous studies were conducted only with strains of serotype 2 (17, 18).
In order to test if the polysaccharide capsule of S. suis protects against oxidative burst intermediates, we included a nonencapsulated strain in our study. We observed that the unencapsulated S. suis mutant 10cpsΔEF was not more susceptible to hydrogen peroxide than the wt strain 10. Whereas the role of the polysaccharide capsule in the susceptibility of S. suis to ROS had not been investigated to date, the influence of the capsule on burst induction was studied for S. pneumoniae. Barbuti et al. showed that the presence of a capsule in pneumococci does not alter the extent of ROS production by human neutrophils (25). We generated similar results by comparing oxidative burst levels induced by an S. suis serotype 2 wt strain and an isogenic capsule-deficient mutant (raw data not shown).
Previous studies investigating the oxidative burst response to S. suis were conducted using murine macrophages or field-infected animals whose cells were stimulated with fMLP (N-formyl-Met-Leu-Phe) (17, 26). Here, we investigated for the first time the induction of the oxidative burst in porcine blood granulocytes in response to S. suis in the context of defined humoral factors. The experimental conditions of our study are of substantial biological relevance, as septicemia and meningitis in pigs following bacteremia are predominant pathologies of S. suis infections in the field. Consecutive measurements of ROS and associated specific bacterial loads revealed for the first time the induction of the oxidative burst in vivo in blood granulocytes of a pig following an experimental challenge. Although conclusive evidence from the animal experiment is limited, as it included only two animals, the results are in accordance with the concept that S. suis-specific antibodies lead to the induction of ROS in blood granulocytes, which in turn contributes to the killing of streptococci during bacteremia. In any case, the results demonstrate that two pigs of the same age and herd may exhibit substantial differences in ROS production during bacteremia. This was reason enough for us to focus on the biological role of ROS in the control of bacteremia in pigs with specific antibodies.
The use of the NADPH oxidase inhibitor apocynin revealed that inhibition of the oxidative burst leads to increased survival of S. suis strains belonging to different serotypes. Apocynin acts on the NADPH oxidase subunit p47phox that is present only in the NOX2 isoform, and as it needs preactivation by myeloperoxidase, it inhibits the oxidative burst in phagocytic leukocytes only (27, 28). Lu et al. (22) showed that NOX2-based NADPH oxidase-dependent ROS generation is responsible for increased phagocytosis of porcine alveolar macrophages. Additionally, Almeida et al. (23) described a link between ROS production and phagocytosis in granulocytes. Accordingly, our data also suggest a reduction in the percentage of phagocytic granulocytes after inhibition of NADPH oxidase activity with apocynin. Thus, increased survival of S. suis under these conditions might be a combined result of decreased phagocytosis and increased bacterial survival within phagolysosomes of granulocytes incapable of producing ROS under the influence of apocynin. From these results, we conclude that the oxidative burst of porcine granulocytes is indeed an important defense mechanism contributing to the killing of S. suis. In addition, we demonstrate that there is no derogatory effect of apocynin on the viability of isolated porcine granulocytes (see Fig. S2B in the supplemental material).
Human granulocytes express both high- and low-affinity Fc receptors that are involved in the recognition and phagocytosis of antibody-opsonized pathogens (19). It is known that the activation of Fc receptors on neutrophils directly induces high levels of ROS (16), and some studies suggest that only Fc receptor-mediated phagocytosis can induce ROS production (29). Our data show a distinct impact of antibody concentrations in serum on oxidative burst levels. However, the depletion of total IgG from hyperimmune serum does not result in a complete abrogation of the oxidative burst. As the concentration of serum and the addition of CDS (free of IgG) were conducted after IgG depletion, differences between original and IgG-depleted hyperimmune sera cannot be attributed solely to the absence or presence of IgG. Nevertheless, the higher level of ROS induction mediated by the original hyperimmune serum (compare Fig. 5A and Fig. 6A) is in accordance with IgG-FcγR-mediated oxidative burst induction in granulocytes (Fig. 10, pathway 1).
Schematic overview of mechanisms inducing the oxidative burst and mediating killing of S. suis by blood granulocytes. Mechanisms inducing the oxidative burst in porcine granulocytes and the killing of S. suis as well as inhibitors of the different pathways are depicted. Inhibitors used in this study to block certain players in the ROS induction cascade are shown as black blind-ended arrows, whereas pathways observed in this study are indicated by blue arrows or as dotted blue arrows in case of supposed pathways. Binding of IgG to S. suis can induce the oxidative burst via direct IgG-Fcγ receptor binding (pathway 1) or via complement (2). IgM induces the oxidative burst via complement (3). The complement-induced oxidative burst is potentially mediated via complement receptor (CR) signaling (4). Complement-mediated killing, independent of ROS production, could also be observed (5) and presumably occurs extracellularly. However, we observed IgM-mediated killing of S. suis independent of a complement-induced oxidative burst (6). Note that FcμR is absent from phagocytic cells (34). The activation of NADPH oxidase (NOX) ultimately leads to the production of bactericidal hydrogen peroxide (H2O2) and hypohalites such as hypochlorite (OCl−) that mediate the killing of S. suis (7) in the suggested phagosomal compartment (depicted as a pink-shaded trapeze-shaped extension).
The results obtained with IgG-depleted hyperimmune serum clearly show that other mechanisms besides IgG might induce an oxidative burst and killing of streptococci in porcine blood of a bacterin-immunized piglet. Similar results with regard to the role of IgG-independent oxidative burst induction were obtained by Nilsson et al., who showed that blocking of Fc receptors on human neutrophils prevented neither the oxidative burst nor the killing of S. pyogenes (30). Besides IgG, complement can also induce the oxidative burst in neutrophils, which possess five distinct complement receptors (31). For group A Streptococcus, it was demonstrated that the activation of human complement receptor 3 (CR3 or CD11b/CD18) via iC3b is required for ROS production and killing of S. pyogenes by human neutrophils (30). Similarly, we show in this study that inhibition of complement in blood significantly reduced the oxidative burst. Since porcine neutrophils express the CR3 orthologue wCD11R3, it is likely that the process of complement-mediated burst induction in porcine granulocytes (Fig. 10, pathway 2) resembles the process in human cells (32). However, inhibition of complement in blood reconstituted with complete hyperimmune serum including high IgG levels showed marginal effects on the killing of S. suis.
The killing of S. suis could still be observed in blood reconstituted with IgG-depleted serum, but in this case, it was complement dependent, as shown by complement inhibition via VCP. However, this effect was less pronounced in experiments using fluorescently labeled streptococci. As IgM induces the classical complement pathway (33), we investigated whether IgM-induced complement activation (Fig. 10, pathway 3) leads to the subsequent induction of the oxidative burst (Fig. 10, pathway 4), here designated the IgM-complement-oxidative burst axis. The existence and relevance of this pathway were confirmed by the reconstitution of blood with IgG-depleted and IgM protease-treated hyperimmune serum. These treatments result in a reduction of the oxidative burst and an increase in bacterial survival. However, our results suggest that IgM also possesses antibacterial activity against S. suis independently of complement and the oxidative burst (Fig. 7A and Fig. 9B). It is conceivable that the aggregation of S. suis by IgM pentamers is involved in the oxidative burst- and complement-independent antimicrobial activity of IgM since aggregation might prevent bacterial proliferation (Fig. 10, pathway 6). This might even be the reason why the survival factors of fluorescently labeled streptococci were lower in samples with combined apocynin and VCP treatment than in samples treated with apocynin alone (Fig. 9B). However, we do not know why this is contrary to the findings obtained for unlabeled bacteria shown in Fig. 6B, and we can only speculate that this might be related to the different viability of fluorescently labeled streptococci in porcine blood (Fig. S5). In any case, the complement-independent killing mechanism of IgM is unlikely to be due to direct IgM-Fc receptor-mediated signaling since phagocytic cells do not possess an Fcμ receptor (34). Instead, IgM might mediate antibacterial activity and complement-independent burst induction via immune complexes (ICs). Lucisano et al. and Furriel et al. showed that IgM-ICs can bind to rabbit polymorphonuclear cells and induce the oxidative burst and lysosomal enzyme release (35, 36). This mode of action does not require the phagocytosis of IgM-ICs, and it is conceivable that S. suis-IgM-ICs also contribute to oxidative burst induction and killing of S. suis. Accordingly, the percentages of Rho123+ granulocytes following complement inhibition were lower in IgM-cleaved samples, but the differences were rather small (Fig. 6B and Fig. 9B). For pneumococci, it has been shown that capsular polysaccharide-specific human monoclonal IgM has the ability to reduce the number of CFU even in the absence of phagocytes or complement and that this correlates with IgM aggregates (37).
In summary, this study was designed to investigate the oxidative burst responses of porcine granulocytes to different S. suis serotypes and elucidate in particular the oxidative burst-inducing factors as well as the role of the oxidative burst in the killing of S. suis. To conclude, in this study, we show for the first time the induction of the oxidative burst in porcine granulocytes in response to S. suis in porcine blood in vitro and in vivo and clearly demonstrate the involvement of the oxidative burst in the killing of S. suis in blood reconstituted with serum of a bacterin-vaccinated pig. The induction of the oxidative burst by S. suis was shown to be dependent on IgM antibodies as well as on complement, and our data suggest a likely role for IgG as well.
MATERIALS AND METHODS
Bacterial strains and growth conditions.S. suis was grown in Todd-Hewitt broth (THB) (product no. 249240; Becton, Dickinson), and glycerol stocks were prepared at the late exponential growth phase. For experimental infection of pigs, S. suis was grown in tryptic soy broth (TSB) without dextrose (product no. 286220; Becton, Dickinson) at 37°C with 5% CO2. The infection inoculum was prepared at the late exponential growth phase. S. suis strain 10 is a virulent muramidase-released protein-positive (MRP+), extracellular factor-positive (EF+), IdeSsuis-positive (IdeSsuis+) suilysin-positive cps2 strain that has been used for pathogenesis studies by different groups (21, 38, 39). S. suis 10cpsΔEF is an isogenic mutant of S. suis strain 10 deficient in capsule expression (38). S. suis A3286/94 is a virulent MRP+ suilysin-positive cps9 strain isolated from a pig with meningitis and has also been used in pathogenesis studies previously (40, 41). S. suis strain 16085/3b is a virulent MRP+ suilysin-positive cps9 strain (42). S. suis 13-00283-02 is a virulent MRP+ suilysin-negative cps7 strain (41). Escherichia coli strains harboring plasmids pET45bideSsuis_homologue and pET45bideSsuis_homologue_C195S were grown in Luria-Bertani medium (product no. X968.2; Carl Roth) at 37°C under constant shaking and with the addition of 100 μg/ml ampicillin (product no. K029.4; Carl Roth).
Experimental infection of piglets to measure the oxidative burst in in vivo blood samples.Two 8-week-old German Landrace piglets from a herd infected with numerous S. suis pathotypes were intravenously infected with 3 × 108 CFU of S. suis strain 10 under ketamine and azaperone anesthesia. At 12 h postinfection, the piglets were anesthetized again through the application of ketamine and azaperone, and anesthesia was maintained via inhalation of isoflurane. Blood samples were taken from the vena jugularis or arteria femoralis at defined time points (preinfection and 0.5 h, 13 h, 14 h, 16 h, and 19 h postinfection). CFU were determined in blood samples by plating of serial dilutions to monitor the status of bacteremia. Measurements of the oxidative burst in vivo were conducted using 0.1-ml heparinized whole-blood samples that were immediately stained with dihydrorhodamine 123 (DHR123) upon blood withdrawal. At each time point, an additional PMA-stimulated 0.1-ml blood sample (0.1 μg/ml) served as the positive control. Erythrocyte lysis, fixation, and sample analysis via flow cytometry were performed as described above for in vitro oxidative burst experiments. The animal experiment was approved by the Committee on Animal Experiments of the Lower Saxonian State Office for Consumer Protection and Food Safety under permit no. 33.8-42502-04-18/2879. Handling and treatment of animals were done in strict accordance with the principles of the European Convention for the Protection of Vertebrate Animals Used for Experimental and Other Scientific Purposes (43) as well as German animal protection law. Both piglets were euthanized 19 h after intravenous application of S. suis during inhalation anesthesia.
Growth of S. suis in the presence of hydrogen peroxide.Sixty milliliters of prewarmed THB medium was inoculated to an optical density at 600 nm (OD600) of 0.02 with cultures of S. suis strains 10, 10cpsΔEF, A3286/94, 16085/3b, and 13-00283-02 grown overnight for 10 h. After 2 h of incubation at 37°C, the cultures were divided into four tubes to continue growth in either THB alone or THB supplemented with hydrogen peroxide to render the following final concentrations of H2O2: 0.1%, 0.01%, and 0.001% (wt/vol). The OD600 was measured every hour for 8 h.
Expression and purification of recombinant His-tagged proteins.Expression of recombinant His-tagged IdeSsuis homologue (rIdeSsuis_h) and rIdeSsuis_h_C195S was performed as described previously (9).
Oxidative burst experiments.Oxidative burst detection in blood granulocytes of i.v. infected pigs was performed directly in blood samples without further addition of S. suis. In addition, assays of reconstituted whole blood were conducted to assess the role of specific serum components in ROS induction and the killing of bacteria. Therefore, plasma-derived blood donor cells were resuspended with defined sera from other pigs. In detail, heparinized whole blood from healthy donor piglets from a commercial pig farm in Saxony, Germany, was washed two times with 10 ml 0.9% sodium chloride. The withdrawal of blood was approved under permit no. N19/14 by the responsible authorities of the state of Saxony, Germany (Landesdirektion Sachsen). Washed blood cells were resuspended in 0.9% sodium chloride and diluted in selected hyperimmune sera at a ratio of 1:1. S. suis was added from frozen glycerol stocks at a concentration of 2 × 106 CFU/ml. Positive controls were incubated with 0.1 μg/ml PMA (product no. 79346-1MG; Sigma-Aldrich). Samples were incubated at 37°C for 15 min in a water bath after the addition of S. suis or PMA. Subsequently, DHR123 (5 μg/ml) (product no. D1054-2MG; Sigma-Aldrich) was added, and incubation at 37°C in a water bath was continued for another 10 min. ROS interaction leads to the oxidation of DHR123 to cationic rhodamine 123 (Rho123). Finally, erythrocytes were lysed two times using erythrocyte lysis buffer (0.155 M ammonium chloride, 10 mM potassium bicarbonate, 0.1 mM disodium EDTA [pH 7.2]), washed two times with phosphate-buffered saline (PBS), and fixed with 2% paraformaldehyde. Samples were measured immediately by flow cytometry (BD FACSCalibur) and analyzed by gating on granulocytes (FlowJo).
For experiments with the NADPH oxidase inhibitor apocynin (1.5 mM) (product no. 7884.1; Carl Roth), the inhibitor was added to reconstituted blood prior to the addition of S. suis and preincubated for 5 min at 37°C in a water bath. Complement inactivation of serum was achieved by treatment with VCP (vaccinia virus complement control protein) (100 μg/ml) (product no. GB-VCP250; Gene Balance Inc.) or by heat inactivation at 56°C for 30 min. For assays investigating the impact of IgM, serum was treated with 20 μg/ml of rIdeSsuis_h (21), the point-mutated variant rIdeSsuis_h_C195S (9), or heat-inactivated (10 min at 95°C) rIdeSsuis_h (rIdeSsuis_h_95°C), prior to reconstitution of blood by incubation at 37°C for 2.5 h on a rotator. To verify complete IgM cleavage, an anti-IgM Western blot assay under reducing conditions was conducted (see Fig. S4 in the supplemental material), as previously described (9).
Bactericidal assays.Bactericidal assays were always conducted in parallel with oxidative burst experiments, meaning that 0.2-ml samples containing reconstituted blood and S. suis at a concentration of 2 × 106 CFU/ml were divided in two immediately after the addition of bacteria. Whereas one half of the sample was used for oxidative burst experiments, the other half was used for bactericidal assays. CFU were determined by serial dilution at time zero and after incubation of samples at 37°C on a rotator for 1 or 2 h (indicated in the figures). The survival factor (SF) was calculated by dividing the CFU after the 1- or 2-h incubation period by the CFU at time zero.
Opsonophagocytosis assays.Opsonophagocytosis assays were conducted as previously described (44), with minor modifications. Briefly, 5 × 106 isolated porcine neutrophils resuspended in 500 μl RPMI supplemented with 20% (vol/vol) hyperimmune serum raised against cps2 strain 10 were infected with 2.5 × 106 CFU S. suis strain 10 to obtain a multiplicity of infection (MOI) of 0.05. The samples were incubated for 2 h at 37°C on a rotator in the presence or absence of 1.5 mM apocynin.
Generation of hyperimmune sera and depletion of IgG from serum.The generation of hyperimmune sera in piglets was approved by the Landesdirektion Sachsen under permit no. N01/16. The hyperimmune serum used for experiments with S. suis strains 10 and 10cpsΔEF was obtained from a pig that had been prime-booster vaccinated with an S. suis strain 10 bacterin. The depletion of total IgG from hyperimmune serum against S. suis strain 10 was conducted by protein G affinity chromatography as described in detail previously (9). The moderate-antibody serum was derived from a piglet that had been experimentally infected with strain 10 and had to be euthanized due to meningitis at 5 days postinfection (the corresponding animal experiment was approved under permit no. TVV11/16 by the ethics committee of the Landesdirektion Sachsen). Hyperimmune serum against S. suis serotype 9 was drawn from a piglet that was prime-booster immunized with a bacterin based on strain A3286/94. The serum was drawn 18 days after the booster vaccination. The convalescent-phase serum used in oxidative burst and bactericidal assays with S. suis cps7 strain 13-00283-02 was originally drawn from a piglet 14 days after experimental infection with S. suis cps9 strain 16085/3b and is known to mediate the killing of different cps types, including cps7 (the corresponding animal experiment was approved under permit no. TVV 28/16 by the ethics committee of the Landesdirektion Sachsen).
Anti-S. suis IgM and IgG ELISAs.IgM and IgG antibody levels against S. suis strain 10 were determined by ELISAs using plates coated with inactivated bacteria, as described previously (45). Moderate antibody levels were defined as lying within the range of 10 to 70 ELISA units. ELISA units were determined relative to reference serum that was defined as containing 100 ELISA units of both IgG and IgM.
Viability staining of granulocytes.The viability of granulocytes after bactericidal assays and subsequent erythrocyte lysis was proven by staining with a 1:500 dilution of eBioscience eF506 fixable viability dye (product no. 65-0866-14; Thermo Fisher Scientific) in PBS for 25 min. Afterwards, cells were washed, fixed, and measured via flow cytometry using an LSR Fortessa instrument (BD). As a positive control, an additional blood sample was stressed for 5 min at 95°C directly before staining with eF506 viability dye.
PCR for detection of the sodA gene of S. suis.For the detection of the sodA gene of S. suis, a colony PCR approach was used. Two colonies of each investigated S. suis strain (strains 10, 10cpsΔEF, A3286/94, 16085/3b, and 13-00283-02) were microwaved in 0.1 ml DNase- and RNase-free water (product no. T143.3; Carl Roth) at 100 W for 8 min, and 5 μl of this preparation served as the template for PCR with a total reaction mixture volume of 25 μl. Primers for the detection of the sodA gene were designed based on the sodA sequence of strain EA1832.92 (GenBank accession no. AB724057.1). The primer pair sodA_for (GCACCATGCAACTTATGTGGCAAATGC) and sodA_rev (CCTTCGCTGTTAACAACCAAFGAAAGCC), binding in a conserved region of the sodA gene, was used for the amplification of a 324-bp product. The PCR program using OneTaq DNA polymerase (product no. M0480S; NEB) was as follows: an initial denaturation step for 30 s at 94°C, followed by 30 cycles of 30 s at 94°C, 60 s at 60°C, and 30 s at 68°C and a final elongation step for 5 min at 68°C. Gel electrophoresis was performed with a 1.5% agarose gel.
Real-time PCR for detection of sodA transcripts in S. suis.Bacterial RNA was extracted from stationary-phase THB cultures of the different S. suis strains used in this study. In detail, a 10-ml culture was centrifuged (10 min at 2,600 × g at room temperature [RT]), and the supernatant was discarded. The pellet was stored at −80°C until RNA extraction. Isolation was conducted as previously described (46). After RNA extraction, the concentration and purity of the isolated RNA were determined using the Agilent 2100 bioanalyzer (RNA 6000 Pico kit; Agilent Technologies Inc., Santa Clara, CA, USA). Quantitative real-time PCR (qRT-PCR) of reverse-transcribed RNA was designed to analyze the expression of sodA and the housekeeping gene gyrB. The respective primers are listed in Fig. S1B. qRT-PCR was conducted with the AriaMX real-time PCR system (Agilent Technologies Inc., Santa Clara, CA, USA), as previously described (46). The following modified program was used: an initial denaturation step at 95°C for 20 min and 40 cycles of denaturation at 95°C for 25 s, annealing at 60°C for 30 s, and amplification at 72°C for 20 s. As negative controls, we included (i) water, (ii) a no-template control, and (iii) no reverse transcriptase to exclude contamination with DNA. Products were verified by melting-curve analysis and 2.0% agarose gel electrophoresis.
Combined phagocytosis and oxidative burst assays.The combined measurement of the oxidative burst and phagocytosis in a single blood sample was conducted essentially as described above for oxidative burst experiments. In order to be able to read out the oxidative burst and phagocytosis, living S. suis strain 10 bacteria were labeled with CellTrace far-red fluorescent dye (Thermo Fisher Scientific) and added to the blood samples at a concentration of 2 × 106 CFU/ml. All measurements using far-red-labeled bacteria were performed without the use of a quenching dye. Quenching of the extracellular signal with trypan blue was found to be unnecessary since the phagocytosis of carboxyfluorescein succinimidyl ester (CFSE)-labeled S. suis strain 10 bacteria was identical with and without the addition of trypan blue (data not shown).
Statistical analysis.Data were analyzed for normal distribution by the Shapiro-Wilk test. In the case of a normal distribution, unpaired two-tailed Student’s t test was used. In the case of not-normally distributed data, the nonparametric two-tailed Mann-Whitney or Kruskal-Wallis test with Dunn’s multiple-comparison test was applied. Significant outliers were calculated using GraphPad QuickCalcs (https://www.graphpad.com/quickcalcs/grubbs1/) and excluded from the analysis. A confidence interval of 95% was chosen for all analyzes. All figures and data in parentheses in the text represent the means and standard deviations (SD). Probabilities were considered as indicated in the figure legends. Flow cytometric data were analyzed using FlowJo_V10.
ACKNOWLEDGMENTS
We thank Hilde Smith (DLO, Lelystad, Netherlands) for providing S. suis strain 10 as well as its capsule-deficient mutant (10cpsΔEF) and Peter Valentin-Weigand (Institute for Microbiology, University of Veterinary Medicine Hannover, Germany) for S. suis strain A3286/94. We acknowledge Silke Lehnert (Institute of Immunology, Veterinary Faculty, University of Leipzig, Germany) for helping conduct oxidative burst experiments.
This project was funded by the German Federal Ministry of Education and Research within the Infect Control 2020 consortium (subproject 3 of VacoME to C.G.B. and G.A.), the German Research Platform for Zoonoses (HypoxiaInfect, DLR no. 01KI1819, provided to N.D.B. and M.V.K.-B.), and the German Research Foundation (DFG BA 4730/4-1 and KO 3552/7-1 to C.G.B. and M.V.K.-B., respectively).
FOOTNOTES
- Received 7 August 2019.
- Returned for modification 28 August 2019.
- Accepted 3 December 2019.
- Accepted manuscript posted online 16 December 2019.
Supplemental material is available online only.
- Copyright © 2020 American Society for Microbiology.