Oxidative killing of encapsulated and nonencapsulated Streptococcus pneumoniae by lactoperoxidase-generated hypothiocyanite

Streptococcus pneumoniae (Pneumococcus) infections affect millions of people worldwide, cause serious mortality and represent a major economic burden. Despite recent successes due to pneumococcal vaccination and antibiotic use, Pneumococcus remains a significant medical problem. Airway epithelial cells, the primary responders to pneumococcal infection, orchestrate an extracellular antimicrobial system consisting of lactoperoxidase (LPO), thiocyanate anion and hydrogen peroxide (H2O2). LPO oxidizes thiocyanate using H2O2 into the final product hypothiocyanite that has antimicrobial effects against a wide range of microorganisms. However, hypothiocyanite’s effect on Pneumococcus has never been studied. Our aim was to determine whether hypothiocyanite can kill S. pneumoniae. Bactericidal activity was measured in a cell-free in vitro system by determining the number of surviving pneumococci via colony forming units on agar plates, while bacteriostatic activity was assessed by measuring optical density of bacteria in liquid cultures. Our results indicate that hypothiocyanite generated by LPO exerted robust killing of both encapsulated and nonencapsulated pneumococcal strains. Killing of S. pneumoniae by a commercially available hypothiocyanite-generating product was even more pronounced than that achieved with laboratory reagents. Catalase, an H2O2 scavenger, inhibited killing of pneumococcal by hypothiocyanite under all circumstances. Furthermore, the presence of the bacterial capsule or lytA-dependent autolysis had no effect on hypothiocyanite-mediated killing of pneumococci. On the contrary, a pneumococcal mutant deficient in pyruvate oxidase (main bacterial H2O2 source) had enhanced susceptibility to hypothiocyanite compared to its wild-type strain. Overall, results shown here indicate that numerous pneumococcal strains are susceptible to LPO-generated hypothiocyanite.

Introduction Streptococcus pneumoniae (Spn) is a leading cause of bacterial infections such as otitis media, pneumonia, septicemia and meningitis [1,2]. Colonization can occur at any point in a person's life but occurs most commonly in young children where Spn prevalence reaches over 50% in hosts 2-3 years old [3]. Worldwide, Spn is a major cause of infant mortality with 1.2 million deaths reported every year [2,4]. Current pneumococcal vaccines target the capsular polysaccharide of Spn, but these vaccines only provide serotype-specific protection against less than one third of circulating serotypes [5]. Spn infections can also be controlled with antibiotics, but widespread antibiotic use has led to accelerated antibiotic resistance in Spn [6]. These challenges have led to the need for novel therapeutics and a better understanding of Spn interactions with the host.
The airway epithelium represents one of the largest physical and immune barriers against airborne microbes such as Spn [7]. Lactoperoxidase (LPO) is a heme peroxidase found in the airway surface liquid (ASL) where it performs its antimicrobial activity [8]. The LPO-based antimicrobial system requires two other components to function properly. First, LPO needs a source of H 2 O 2 to catalyze the reaction. In the human airways, H 2 O 2 is mainly supplied by the NADPH oxidase Dual oxidase 1, Duox1 [9][10][11]. LPO then uses H 2 O 2 to oxidize the pseudohalide thiocyanate (SCN -) which is abundantly present in the ASL into the antimicrobial ion hypothiocyanite (OSCN -) [9,12]. The LPO-based system has previously been shown to be an effective in vitro neutralizer of a wide variety of viruses [13][14][15] and bacteria [8,16]. Interestingly, even though Spn represents an enormous health burden, the effectiveness of the LPObased system against Spn has not been tested so far.
Due to the relevance of Spn in public health combined with the emergence of antibiotic resistance, the LPO-based system could provide valuable insight and a possible new therapeutic option for management of Spn infections. We hypothesized that the LPO-based system is effective at killing Spn in vitro. We found that both encapsulated and nonencapsulated pneumococci were susceptible to OSCNmediated killing in a cell-free experimental system.

Bacteria
Spn strains EF3030 (encapsulated serotype 19F) [17], EF3030 Δcap (isogenic, nonencapsulated mutant strain) [18], MNZ41 (nonencapsulated) [19], TIGR4 (encapsulated serotype 4) [20], TIGR4ΔspxB (isogenic mutant deficient in pyruvate oxidase) [21] and TIGR Δcap (isogenic mutant deficient in capsule formation generated by the same strategy as EF303 Δcap) [18], D39 (encapsulated serotype 2) and its isogenic, capsule-free mutant, D39 Δcap [22] were inoculated on sheep blood agar plates (BAP) and incubated at 37˚C in 5% CO 2. After incubation, bacteria were collected and harvested by centrifugation at 10,000g for 5 minutes, washed twice with Hank's balanced salt solution (HBSS), and suspended in HBSS. Bacterial density was then determined by measuring optical density (OD) at 600 nm. The bacterial density was set to 0.6, which is representative of 10 9 CFU/mL Spn that was confirmed by performing serial dilutions, plating bacteria on BAP and counting colonies. Bacteria were prepared this way for both, bacterial killing and bacteriostatic measurements. The identities of the Spn strains were confirmed by 16S rRNA Gene Sequencing (Genewiz, South Plainfield, NJ, USA). Optochin-sensitivity of the Spn strains used was also confirmed for each experiment using BD BBL™ Taxo™ P Discs (Fisher Scientific, Pittsburgh, PA, USA). and 0.1 U/mL glucose oxidase. The reaction volume was set to 120 μL. Catalase (700 U/mL) was also used when indicated to inhibit the system by scavenging H 2 O 2 . The components were assembled in a sterile 96-well microplate in triplicates with the bacteria being added last at a maximal concentration of 5x10 5 CFU/ml. The plates were then placed in a 37˚C incubator with 5% CO 2 . After 6 hours of incubation, 40 μL was spread onto BAP in triplicate and incubated at 37˚C with 5% CO 2 . After 24 hours, the colonies were counted and CFU/mL was determined. Agar plates exposed to only the assay medium without Spn were always used to ensure that no potential contaminants were detected. A time 0 condition was also counted to make sure that bacterial death was due to OSCNand not related to an unknown variable, and that no significant changes in bacterial numbers were observed in samples containing only bacteria during the duration of the experiments. All the reagents were ordered from Sigma-Aldrich (St. Louis, MO, USA) unless stated otherwise.

Bacteriostatic activity measured by a microplate-based growth assay
The bacteriostatic activity of OSCNwas measured by a microplate-based assay described previously [23]. Briefly, the components (mentioned in the cell-free assay) were assembled in a sterile 96-well plate with the bacteria being added last. Bacterial growth was measured in a microplate spectrophotometer [Eon (BioTek Instruments Inc., Winooski, VT, USA) or Varioskan Flash (Thermo Scientific, Rochester, NY)] on the basis of following increases in OD as a measure of bacterial density. This method enables fast and very reproducible measurement of bacterial growth [23]. Spn strains were grown at 37˚C for 14 hours, and OD at 600 nm was measured every 3 minutes. Each sample was run in triplicate. The time required for the positive control (Spn alone) to reach an OD of 0.4 (exponential growth phase) was used as the reference point for all other conditions, and OD values of other samples were compared to this. All the Spn strains used in this work were tested individually for their suitability for this method.

The commercial 1 st line™ immune support product
1 st line™ is an over-the-counter product that is marketed as an immune supplement (distributed by Profound Products). This product uses a proprietary technology to keep OSCNstable for a longer period of time allowing for a better antimicrobial effect. To our knowledge, this product is the only commercially available product producing OSCN -. We tested this product in conjunction with our previously described cell-free system. Briefly, 0.1 g of LPO was reconstituted in 25 mL of HBSS. 750 μL of H 2 O 2 solution was added to a 15 mL conical tube followed by the addition of 700 U/mL of catalase. The solution was incubated for 10 minutes to allow catalase to scavenge all H 2 O 2 present. 12.5 mL of LPO solution was then added to each tube and mixed. Following this step, 750 μL of SCNwas also added to each sample and mixed thoroughly. Finally, 750 μL poly aluminum chloride was administered to each solution and mixed well. Samples were then incubated for 30 minutes at room temperature allowing the generation of OSCN -. By the end of this incubation time, the solution separates into two distinct phases. The top, clear phase containing OSCNwas used for experiments while the pelleted precipitate was discarded.

Bacterial H 2 O 2 production
Generation of H 2 O 2 in bacterial suspension was measured by the ROS-Glo TM luminescence kit following the manufacturer's instructions (Promega Corporation, Madison, WI, USA). This sensitive assay enables specific and direct detection of low amounts of H 2 O 2 . TIGR4 wildtype or ΔspxB bacteria (5x10 6 /ml) suspended in HBSS buffer were incubated at 37 ºC for 30 minutes, followed by centrifugation to collect supernatants for analysis of H 2 O 2 production. Ros-GLo TM reagent was added to bacterium-free supernatants and luminescence was read using a Varioskan Flash microplate luminometer (Thermo Scientific, Rochester, NY). The assay was run in triplicates. Results are expressed as relative luminescence units (RLU).

Statistical analysis
Significance among multiple samples was calculated using One-way or Two-Way ANOVA followed by Tukey's or Sidak's multiple comparison post-hoc tests. Significance between two samples was calculated using Mann-Whitney's test. Statistical analysis was performed using Prism 6 for Windows version 6.07 software. � , p<0.05; �� , p<0.01; ��� , p<0.001.

LPO-derived hypothiocyanite kills a diverse array of Spn strains
To explore the effects of OSCNon Spn, we used our previously established cell-free in vitro system that generates OSCNat levels comparable to those measured in human airways [26]. This system utilizes the enzyme glucose oxidase, which oxidizes its substrate glucose to produce D-gluconolactate [27]. H 2 O 2 is a byproduct of the reaction and allows us to mimic the nature of H 2 O 2 production in vivo by Dual oxidases [8,26]. Previously, we have successfully used this experimental system to show that LPO-generated OSCNinactivates a wide range of influenza strains [13,27]. Using this H 2 O 2 /LPO/SCNcell-free system, we wanted to determine the effectiveness of OSCNagainst Spn. Physiologically relevant levels of each component of the system were utilized: 400 μM SCN - [28] and 6.5 μg/ml LPO [11]. H 2 O 2 production by glucose oxidase was set to a rate of 0.01 U/ml, which is similar to what is seen in primary normal human bronchial epithelial (NHBE) cells by Duox [13,29]. We tested both encapsulated (TIGR4, EF3030) and nonencapsulated (MNZ41) Spn strains, and we observed that the cellfree H 2 O 2 /LPO/SCNsystem effectively killed all three strains of Spn ( Fig 1A). Fig 1B shows that OSCNis only produced when all components of the cell-free system are added. OSCNis generated reproducibly to achieve a final OSCNconcentration of 41.2 ± 4.2 μM (mean ± S.E. M., n = 4) ( Fig 1B). Since H 2 O 2 alone is capable of killing Spn [30], we exposed EF3030 Spn to the same levels of glucose oxidase without SCNor LPO to ensure our results were OSCNmediated. The results show that Spn survival is not impaired by H 2 O 2 alone (Fig 1C).

Catalase prevents the antimicrobial action of OSCNon Spn
To further confirm our findings that OSCNis antimicrobial against Spn, we utilized a kinetic assay to measure bacterial growth in presence of an inhibitor of the H 2 O 2 /LPO/SCNcell-free system. Catalase is an enzyme found in almost all living organisms that catalyzes the decomposition of H 2 O 2 to water and oxygen [31]. The use of catalase eliminates H 2 O 2 and thereby OSCNin our cell-free system (Fig 2A), rendering the cell-free system nonfunctional while also ensuring that neither SCNnor LPO have an antimicrobial effect alone, independent of OSCNformation. Results shown in Fig 2B indicate that Spn bacteria exposed to OSCNhave inhibited bacterial growth compared to those treated with catalase or unexposed to OSCN -. Both Spn strains tested show the same trend, where OSCNtreatment significantly reduces bacterial growth (p<0.0001) and addition of catalase entirely rescues this effect (p<0.0001)

PLOS ONE
( Fig 2B). The nonencapsulated Spn strain, MNZ41, could not be tested in this assay because it did not grow in liquid cultures used under these experimental conditions. Taken together, these data show for the first time in an in vitro model that OSCNhas a catalase-sensitive, antimicrobial action against different strains of Spn.

LPO-derived OSCNinhibits Spn growth in a SCNconcentrationdependent manner
Physiological levels of SCNin the airways have been measured to be around 400 μM [28]. We decided to test SCNin a supra-and superphysiological concentration range between 40 μM-4 mM to determine if killing of one of the Spn strains, EF3030, can be enhanced. Our data show that supraphysiological levels (40 μM) of SCNkill Spn EF3030 in a robust manner (Fig 3 and  3B). Increasing the SCNconcentration showed a dose-dependent response, where bacterial killing continued to increase, all the way up to 4 mM of SCN - (Fig 3). From this data we conclude that the reported physiological SCNconcentration is sufficient to support the antibacterial activity of the LPO system against Spn.

Commercially available 1 st line™ effectively kills Spn via OSCN -
The H 2 O 2 /LPO/SCNcell-free system has proven to be effective at killing Spn in vitro. A drawback of this system is, however, that OSCNhas a very short life span (less than 30 minutes after it has been generated) requiring it to constantly be produced in vitro in order to test its effect on microbes. This is why we utilized glucose oxidase, not bolus-like addition of H 2 O 2 , to allow a steady production of H 2 O 2 and OSCN - [27]. As the next step, we took advantage of and tested a commercially available product that also generates OSCNand claims to keep it stable for much longer. This product, 1 st line™, utilizes a stabilizing molecule to allow OSCNto persist for over 12 hours. We compared the bacteriostatic effect of the 1 st line™ product on Spn EF3030 and TIGR4. The results demonstrate that OSCNgenerated by 1st line™ resulted in robust inhibition of Spn growth (Fig 4 and 4B). The addition of catalase during the generation of OSCNwith 1 st line™ also inhibited the antimicrobial action of OSCN - (Fig 4), similar to what was previously shown in our cell-free system (Fig 2). These results provide further evidence that OSCNis solely responsible for the inhibition of Spn growth.

The bacterial capsule is not protective against OCSN -
The bacterial capsule provides protection for Spn against threats of the environment including attacks by the immune system [32,33]. We next asked whether the capsule provides protection for Spn against OSCN -. To answer this question, we exposed isogenic, capsule-deficient strains of Spn D39, EF3030 and TIGR4, and their corresponding, encapsulated, parental wild-type counterparts to OSCNin the cell-free experimental system and measured bacterial killing by CFU counting. Results in Fig 5 show that the capsule-free mutants were also susceptible to OSCNon all three backgrounds, similar to their encapsulated control strains. Catalase was also partially or fully effective in preventing the bactericidal effect of OSCNagainst all strains tested ( Fig 5). Therefore, we conclude that the capsule provides no protection against the antipneumococcal action of OSCN -.

The mechanism of action of OSCNis independent of lytA-mediated autolysis
Autolysis is a form of programmed cell death in bacteria including Spn that plays a roles in genetic exchange between bacterial cells, in eliminating damaged cells and has also been implicated in mediating the effects of antibiotics [34]. The lytA genes encodes a major autolysin (Nacetylmuramoyl-l-alanine amidase) in Spn that is a cell wall-degrading enzyme located in the cell envelope [35,36]. Based on these, we postulated that lytA-mediated autolysis could represent the anti-pneumococcal mechanism of action of OSCN -. To test this hypothesis, we exposed lytA-competent and lytA-deficient D39 Spn strains to OSCNin the cell-free system generated by 1st line™ and measured bacterial survival by colony counting. As the results in Fig  6 show, 1st line™ not only inhibits bacterial growth of Spn but also kills this bacterium in an

PLOS ONE
OSCNdependent manner. Fig 6 also shows that the lytA-deficient mutant was also killed efficiently by OSCNindicating that OSCNdoes not initiate lytA-mediated autolysis in Spn.

Spn-derived H 2 O 2 provides some protection against OSCNmediated killing
Interestingly, not only human cells but Spn itself is capable of producing H 2 O 2 [37]. Spn-generated H 2 O 2 could interfere with the killing effect of OSCN -. Spn-generated H 2 O 2 could provide additional H 2 O 2 for OSCNgeneration by LPO leading to improved bacterial killing or it could prime bacteria against oxidative stress resulting in impaired Spn killing by OSCN -. To explore these possibilities, we tested a mutant TIGR4 Spn strain (ΔspxB) deficient in pyruvate oxidase, the main H 2 O 2 producer in Spn [21], for susceptibility to OSCN -. As expected, the spxB-deficient TIGR4 strain had an H 2 O 2 generation that was reduced by 72.5±0.9% (mean ± SEM, n = 2) compared to the wild-type TIGR4 counterpart (Fig 7A). The TIGR4ΔspxB mutant and its parental strain were exposed to OCSNproduced by 1 st line™, and bacterial killing and growth were evaluated with both, CFU-based counting and microplate-based bacterial growth assays. The TIGR4ΔspxB mutant was also susceptible to the antimicrobial effect of OSCNthat was partially inhibited by catalase (Fig 7B). To better present the antibacterial action of OSCN -, we next defined and calculated "susceptibility to OSCN -" as the decrease in log 10 of the colony counts. CFU count results obtained at 6 hours of incubation reached the level of significance (p = 0.029) (Fig 7C). Overall, we conclude that pyruvate oxidase provides improvement in Spn survival following OSCNexposure.

Discussion
While the LPO-based system has been shown to be effective at killing numerous species of bacteria and viruses in vitro, no report to date studied the interaction between this system and Spn. Spn first encounters epithelial cells at the apical surface of the nasal cavity, a region that is a hotbed of defense mechanisms. The LPO-based system kills microbes in the extracellular space before they enter the epithelium and establish infection [8]. We were able to demonstrate that OSCNkills Spn effectively. The ability of OSCNto kill a variety of microbes in the extracellular spaces makes it a very interesting innate immune mechanism to study.
It is possible that Spn stimulates one of the main cellular sources of H 2 O 2 , Duox1, since we had previously shown that bacterial ligands of P. aeruginosa participate in Duox1 activation [12]. Spn has been shown to trigger H 2 O 2 production in airway epithelial cells in a pneumolysin-independent but lytA-dependent manner [38]. We utilized the enzyme catalase, that converts H 2 O 2 into H 2 O and O 2 , to inhibit the production of OSCN -. This was necessary because Spn is a catalase-negative bacterium and is capable of producing its own H 2 O 2 [39]. Spnderived H 2 O 2 is sufficient to mediate bactericidal activity of other bacteria and to stimulate DNA damage and apoptosis in epithelial cells leading to tissue damage during infection [40]. It was found that in the absence of common antioxidant proteins, Spn utilizes pyruvate oxidase (SpxB) that has a dual role of creating H 2 O 2 , and protecting itself from oxidative damage [41]. Pyruvate oxidase, the main H 2 O 2 source in Spn [21], has been found to be important to initiate oxidative attacks on host cells [40]. Our data indicate that SpxB provides a moderate protection against the oxidative attack of OCSN -. Our results are in line with previous findings of other groups where SpxB was determined to be useful against oxidative stress of different origins [41][42][43]. H 2 O 2 generation by SpxB in Spn likely enhances its antioxidant capacity and thereby improves its defense against oxidative stress. Our results indicate a new, protective consequence of SpxB expression in Spn. The in vivo relevance of this finding remains to be confirmed in animal models.
Spn utilizes a wide spectrum of virulence factors, one of the most important ones being the polysaccharide capsule that forms the outermost layer of the bacteria [44]. This capsule provides protection against phagocytosis, complement components, mucus and spontaneous or antibiotic-induced autolysis [45]. Some Spn strains, however, do not possess a capsule [46,47]. Our results show that both encapsulated and nonencapsulated Spn strains are susceptible to OSCN -. Thus, it is likely that OSCNis able to penetrate the capsule and interact with the cell wall or other internal bacterial components. Interestingly, catalase rescued OSCNmediated Spn killing more efficiently when the capsule was absent, in case of TIGR4 and D39 (Fig 5). This difference seems to be strain-dependent as this effect was less pronounced in case of EF3030 (Fig 5). While the reason for this remains unclear, it could be related to differences in H 2 O 2 generation or in inhibition of catalase by the capsule or other microbial factors among tested strains.
OSCNhas a wide microbial target spectrum [8]. Its mechanism of action likely involves oxidative attack on one or more microbe-specific molecules or cellular mechanisms that essentially will lead to a bactericidal action. Previous studies have shown that OSCNis oxidizing Killing by OSCNgenerated via 1st line™ of wild-type and TIGR4 ΔspxB Spn strains was compared at 6 hrs via CFU counting (n = 4) and "susceptibility to OSCN -"was calculated as the decrease in log 10  bacterial sulfhydryls [48][49][50], thereby inhibiting bacterial respiration [51], but no further research has been published to support this possible mechanism of action.
The LPO-based system presents an interesting therapeutic target, since it is an effective antimicrobial innate mechanism that does not have many drawbacks due to its final product being nontoxic to host cells and its broad activity against a wide range of pathogens [52]. By testing the 1 st line™ product (alongside our cell-free method of OSCNgeneration), we observed the same efficient Spn killing results. While we experienced similar levels of killing of Spn when comparing our glucose oxidase system and the 1 st line™ product, there are some subtle differences that would likely affect the efficacy of in vivo studies. The greatest benefit to the 1 st line™ system is the stabilizing aspect of the compound. Allowing OSCNto persist for over 12 hours would likely increase the efficacy and potency of the system which could be a major advantage over the non-stabilized, natural OSCNanion. This also allows for higher concentrations of OSCNto be achieved without having toxicity issues.
While these are encouraging, the study presented here has technical limitations as it was conducted solely in an in vitro, cell-free experimental system. In vivo conditions are obviously much more complex and could represent new challenges to prove the antibacterial efficacy of OSCN -. Duox1-dependent H 2 O 2 production is regulated by several inflammatory and microbial stimuli in vivo and could be targeted by pneumococcal host evasion mechanisms not explored here. The in vivo life span of OSCNmight be influenced by several additional biological factors that could not be tested in the in vitro system studied here. The in vivo administration route and formulation to boost airway production of OSCNoffers numerous possibilities. Overall, these new questions will be answered in future studies using airway epithelial cultures, animal models and human patients.
We determined the effectiveness of OSCNagainst Spn in this proof-of-concept study as it had not been reported previously. We were able to demonstrate that OSCNeffectively kills both encapsulated and nonencapsulated strains of Spn in a cell-free system. We also successfully utilized a commercially available product, 1 st line™, to demonstrate the therapeutic potential of the LPO system in an in vitro model. The mechanism of anti-pneumococcal action of OSCNis independent of the bacterial capsule or lytA-mediated autolysis but is opposed by the bacterial oxidant generator SpxB. These studies warrant further research to elucidate the molecular mechanism of antibacterial action of OSCN -.