IL-33/ST2 Axis Plays a Protective Effect in Streptococcus pyogenes Infection through Strengthening of the Innate Immunity

Group A Streptococcus (GAS) causes invasive human diseases with the cytokine storm. Interleukin-33 (IL-33)/suppression of tumorigenicity 2 (ST2) axis is known to drive TH2 response, while its effect on GAS infection is unclear. We used an air pouch model to examine the effect of the IL-33/ST2 axis on GAS-induced necrotizing fasciitis. GAS infection induced IL-33 expression in wild-type (WT) C57BL/6 mice, whereas the IL-33- and ST2-knockout mice had higher mortality rates, more severe skin lesions and higher bacterial loads in the air pouches than those of WT mice after infection. Surveys of infiltrating cells in the air pouch of GAS-infected mice at the early stage found that the number and cell viability of infiltrating cells in both gene knockout mice were lower than those of WT mice. The predominant effector cells in GAS-infected air pouches were neutrophils. Absence of the IL-33/ST2 axis enhanced the expression of inflammatory cytokines, but not TH1 or TH2 cytokines, in the air pouch after infection. Using in vitro assays, we found that the IL-33/ST2 axis not only enhanced neutrophil migration but also strengthened the bactericidal activity of both sera and neutrophils. These results suggest that the IL-33/ST2 axis provided the protective effect on GAS infection through enhancing the innate immunity.


Introduction
Group A Streptococcus (GAS; Streptococcus pyogenes) infection causes a wide range of human diseases, extending from noninvasive pharyngitis, impetigo, and scarlet fever to invasive cellulitis, necrotizing fasciitis, and streptococcal toxic shock syndrome. GAS also induces non-purulent complications, including rheumatic fever and acute glomerulonephritis. GAS produces several virulence factors to impede the host immune system-mediated opsonophagocytosis and killing, and these virulent factors include M protein, fibronectinbinding proteins, streptococcal pyrogenic exotoxin B (SPE B), immunoglobulin-degrading enzymes, and leukocidal toxins [1]. Clinical observations have indicated that patients who suffered with GAS invasive infections, such as necrotizing fasciitis, as well as streptococcal

IL-33/ST2 Axis Played a Protective Effect on GAS Infection
Intra-air pouch infection of GAS caused invasive necrotizing fasciitis-like symptoms in mice [19,20]. Surveys of both exudates of air pouches and sera from GAS-infected mice found that the expression of IL-33 significantly increased after GAS infection (Supplementary Figure S1). However, the effect of the IL-33/ST2 axis on GAS infection is unclear. To address this question, we used GAS to infect IL-33or ST2-knockout (KO) mice via the intra-air pouch route. As shown in Figure 1, GAS infection (3 × 10 8 CFU/mouse) resulted in 83% (five out of six mice) survival in WT mice for 13-days observation post-infection. However, GAS infection caused 33% (two out of six mice) and 0% (0 out of six mice) survival in IL-33-KO mice and ST2-KO mice, respectively. In IL-33and ST2-KO mice, most mice died within 2 days after infection ( Figure 1). From an examination of the skin lesions, quantitated by an exudates-absorbed method, around the air pouches, we found that all ST2-KO mice showed skin lesions within 2 days after infection and the lesion areas were 85 ± 28 mm 2 (n = 6); these lesion areas were larger than those of IL-33-KO mice (58 ± 5 mm 2 ; n = 6, p < 0.01) and WT mice (29 ± 15 mm 2 ; n = 6, p < 0.01) after GAS infection. The histological examination of skin lesions is shown in Figure 2. The skin lesions of GAS-infected ST2-KO mice appeared to show the most severe damage among the three groups, including the structure of the epidermis, dermis, and subcutaneous fat on the skin around the air pouch, which were severely damaged in GAS-infected mice (Figure 2d-f). These results indicate that the IL-33/ST2 axis provided a protective role in the GAS-infected air pouch model.

IL-33/ST2 Axis Played a Protective Effect on GAS Infection
Intra-air pouch infection of GAS caused invasive necrotizing fasciitis in mice [19,20]. Surveys of both exudates of air pouches and sera from GA found that the expression of IL-33 significantly increased after GAS infectio tary Figure S1). However, the effect of the IL-33/ST2 axis on GAS infectio address this question, we used GAS to infect IL-33-or ST2-knockout (KO intra-air pouch route. As shown in Figure 1, GAS infection (3 × 10 8 CFU/m in 83% (five out of six mice) survival in WT mice for 13-days observation However, GAS infection caused 33% (two out of six mice) and 0% (0 out o vival in IL-33-KO mice and ST2-KO mice, respectively. In IL-33and ST2mice died within 2 days after infection (Figure 1). From an examination of t quantitated by an exudates-absorbed method, around the air pouches, w ST2-KO mice showed skin lesions within 2 days after infection and the le 85 ± 28 mm 2 (n = 6); these lesion areas were larger than those of IL-33-KO m n = 6, p < 0.01) and WT mice (29 ± 15 mm 2 ; n = 6, p < 0.01) after GAS infec logical examination of skin lesions is shown in Figure 2. The skin lesions o ST2-KO mice appeared to show the most severe damage among the three ing the structure of the epidermis, dermis, and subcutaneous fat on the s air pouch, which were severely damaged in GAS-infected mice ( Figure 2 sults indicate that the IL-33/ST2 axis provided a protective role in the G pouch model. Figure 1. Increase in GAS-induced mortality in IL-33-KO and ST2-KO mice. Grou 33-KO and ST2-KO mice were inoculated via the intra-air pouch route with 3 × 10 8 S cells per mouse. The animals were observed every day for a total of 13 days. Intra-a of PBS would cause no mice death in WT, IL-33-KO and ST2-KO mice. Survival c pared for significance using the logrank test on the GAS-infected ST2-KO group infected B6 group (p = 0.0005, p < 0.01), and on the GAS-infected IL-33-KO group infected B6 group (p = 0.0417, p < 0.05). Figure 1. Increase in GAS-induced mortality in IL-33-KO and ST2-KO mice. Groups of six WT, IL-33-KO and ST2-KO mice were inoculated via the intra-air pouch route with 3 × 10 8 S. pyogenes NZ131 cells per mouse. The animals were observed every day for a total of 13 days. Intra-air pouch injection of PBS would cause no mice death in WT, IL-33-KO and ST2-KO mice. Survival curves were compared for significance using the logrank test on the GAS-infected ST2-KO group vs. the GAS-infected B6 group (p = 0.0005, p < 0.01), and on the GAS-infected IL-33-KO group vs. the GAS-infected B6 group (p = 0.0417, p < 0.05).

IL-33/ST2 Axis Affected Bacterial Burdens, Cell Infiltration and Cell Viability in Air Pouches of GAS-Infected Mice
To examine the inflammatory conditions within the air pouches among differe mice, we chose a lower infective dose (2 × 10 8 CFU/mouse) to treat mice. At 24 h po infection, the highest bacterial burdens were found within the air pouches of GAS-infect ST2-KO mice, followed by in GAS-infected IL-33-KO mice, with the lowest burden bei found in the GAS-infected WT mice. The bacterial burdens in the group consisting of W mice with an intra-air pouch with sST2/Fc fusion proteins (sST2) added 30 min before t infection, were similar to those of the GAS-infected ST2-KO mice. Applying recombina mouse IL-33 (rIL-33) 30 min before infection significantly decreased the bacterial burd in GAS-infected IL-33-KO mice, when compared with GAS-infected IL-33-KO ones. Mo over, the bacterial burdens in the air pouches at 48 h post-infection were significantly creased, compared with 24 h post-infection, regardless of whether they were GAS-infect WT, IL-33-KO, or ST2-KO mice, and the pattern of bacterial burdens among differe groups were similar to that of 24 h post-infection (Figure 3a).
To further examine the infiltrating cell numbers within the air pouches, we fou that the infiltrating cell numbers were significantly decreased in GAS-infected IL-33-K ST2-KO mice and WT mice which were given sST2 proteins at 24 h post-infection. Provi ing rIL-33 into the air pouches of IL-33-KO mice would restore infiltrating cell numbers 24 h post-GAS infection (Figure 3b). This result indicates that infiltrating cells bearing S receptors were the major effector cells at the early stage in the GAS-infected air pou model. At 48 h post-infection, there were no differences in infiltrating cell numbers amo different groups, possibly owing to the presence of different chemotactic molecules, oth than IL-33, which were released within the air pouch at the late stage, and worked gether in the recruitment of effector cells into the air pouches.
Besides bacterial burdens and infiltrating cell numbers, we also tested the cell viab ity of infiltrating cells in the air pouches after GAS infection. We found that the cell v

IL-33/ST2 Axis Affected Bacterial Burdens, Cell Infiltration and Cell Viability in Air Pouches of GAS-Infected Mice
To examine the inflammatory conditions within the air pouches among different mice, we chose a lower infective dose (2 × 10 8 CFU/mouse) to treat mice. At 24 h post-infection, the highest bacterial burdens were found within the air pouches of GAS-infected ST2-KO mice, followed by in GAS-infected IL-33-KO mice, with the lowest burden being found in the GAS-infected WT mice. The bacterial burdens in the group consisting of WT mice with an intra-air pouch with sST2/Fc fusion proteins (sST2) added 30 min before the infection, were similar to those of the GAS-infected ST2-KO mice. Applying recombinant mouse IL-33 (rIL-33) 30 min before infection significantly decreased the bacterial burden in GAS-infected IL-33-KO mice, when compared with GAS-infected IL-33-KO ones. Moreover, the bacterial burdens in the air pouches at 48 h post-infection were significantly increased, compared with 24 h post-infection, regardless of whether they were GAS-infected WT, IL-33-KO, or ST2-KO mice, and the pattern of bacterial burdens among different groups were similar to that of 24 h post-infection ( Figure 3a).
To further examine the infiltrating cell numbers within the air pouches, we found that the infiltrating cell numbers were significantly decreased in GAS-infected IL-33-KO, ST2-KO mice and WT mice which were given sST2 proteins at 24 h post-infection. Providing rIL-33 into the air pouches of IL-33-KO mice would restore infiltrating cell numbers at 24 h post-GAS infection (Figure 3b). This result indicates that infiltrating cells bearing ST2 receptors were the major effector cells at the early stage in the GAS-infected air pouch model. At 48 h post-infection, there were no differences in infiltrating cell numbers among different groups, possibly owing to the presence of different chemotactic molecules, other than IL-33, which were released within the air pouch at the late stage, and worked together in the recruitment of effector cells into the air pouches. infection (Figure 3c). Based on the results mentioned above, we suggest that the absence of the IL-33/ST2 axis worsens the GAS infection through decreasing infiltrating cells, as well as their cell viability, at the early stage, and further enhancing bacterial burdens in the air pouches after GAS infection.  were inoculated via the intra-air pouch route with 2 × 10 8 S. pyogenes NZ131 cells per mouse. In one group, IL-33-KO mice were treated with rIL-33 (1 µg per mouse) via the intra-air pouch route at 30 min before GAS infection and 24 h post-GAS infection. In the other group, WT mice were treated with sST2 fusion protein (1 µg per mouse) via the intra-air pouch route at 30 min before GAS infection and 24 h post-GAS infection. At 24 h or 48 h post-infection, exudates of the air pouch were collected, and bacterial numbers (a), the numbers (b) as well as the cell viability (c) of infiltrating cells were counted, as described in Materials and Methods. The results are expressed as the mean ± standard deviation. The infiltrating cells in the air pouches of WT, IL-33-KO and ST2-KO mice given PBS only, without GAS infection, were approximately (1.5 ± 0.3) × 10 4 cells/mL. The infiltrating cells in the air pouch of WT mice given sST2 were approximately (2.5 ± 0.4) × 10 4 cells/mL, while infiltrating cell numbers of B6 mice given rIL-33 were approximately (1.5 ± 0.5) × 10 6 cells/mL. Treatment groups were compared for significance using ANOVA. * p < 0.05, ** p < 0.01, *** p < 0.001 compared with the GAS-infected WT group.
Besides bacterial burdens and infiltrating cell numbers, we also tested the cell viability of infiltrating cells in the air pouches after GAS infection. We found that the cell viability of infiltrating cells was significantly decreased in IL-33-KO mice, ST2-KO mice, and WT mice given sST2 proteins post-GAS infection, when compared with that of GAS-infected WT mice at 24 h post-infection. Adding rIL-33 into the air pouches of IL-33-KO mice would restore the cell viability of infiltrating cells at 24 h post-GAS infection. As time went on, the cell viability of infiltrating cells among different groups kept falling at 48 h post-infection ( Figure 3c). Based on the results mentioned above, we suggest that the absence of the IL-33/ST2 axis worsens the GAS infection through decreasing infiltrating cells, as well as their cell viability, at the early stage, and further enhancing bacterial burdens in the air pouches after GAS infection.

Absence of IL-33/ST2 Axis Enhanced The Expression of Interleukin-6 and Interleukin-1β but Did Not Affect the Balance of T H 1 and T H 2 Responses
Reports indicate that the IL-33/ST2 axis is associated with enhancing T H 2 response in different models [8]. We examined the expression of cytokines in the air pouches after GAS infection. We tested the expression of proinflammatory cytokines, including TNF-α, interleukin-17A (IL-17A), IL-6 and IL-1β, T H 1 cytokines such as interleukin-12 (IL-12) and interferon-γ (IFN-γ), T H 2 cytokines such as interleukin-4 (IL-4) and IL-13, and antiinflammatory cytokine interleukin-10 (IL-10) [21,22]. As shown in Table 1, we found that T H 2 cytokines, both IL-4 and IL-13, in all groups were under the minimal detection limit (5 pg/mL) using the cytometric bead array (CBA) kit. T H 1 cytokines, IL-12 and IFN-γ, and inflammatory cytokine IL-17A showed no obvious expression in the air pouch after GAS infection. The expression of proinflammatory cytokine IL-6 and IL-1β, but not TNF-α, was significantly enhanced in GAS-infected IL-33-KO mice, ST2-KO mice, and WT mice given sST2/Fc fusion proteins. The expression of IL-10 was slightly decreased in IL-33-or ST2-deficient mice, but there were no significant differences among different groups. These results indicate that the IL-33/ST2 axis inhibited the local inflammatory response but did not affect the balance of T H 1 and T H 2 in GAS-infected air pouches.
Groups of four B6, IL-33-KO mice and ST2-KO mice were inoculated via the intra-air pouch route with 2 × 10 8 S. pyogenes NZ131 cells per mouse. In one group, IL-33-KO mice were given rIL-33 (1 µg per mouse) via the intra-air pouch route at 30 min before GAS infection. In the other group, B6 mice were treated with recombinant sST2/Fc fusion protein (1 µg per mouse) via the intra-air pouch route at 30 min before GAS infection. In control groups, B6 mice were given rIL-33 (1 µg per mouse) or sST2/Fc fusion protein (1 µg per mouse) only via the intra-air pouch route. At 24 h post-infection, exudates of the air pouch were collected, and cytokine expressions were examined using a cytometric bead array (CBA) kit. The results were expressed as the mean ± standard deviation. The statistical analysis was performed using repeated-measures ANOVA. * p < 0.05 compared with results for GAS-infected WT mice. a The minimal detection concentration of the cytokine using a CBA kit.

The Predominant Effector Cells in GAS-Infected Air Pouch Model Were Neutrophils
Previous results indicate that the absence of the IL-33/ST2 axis increased bacterial burdens, decreased both the cell viability and the numbers of infiltrating cells, and worsened the local inflammatory response in GAS-infected air pouches. We hypothesized that the 7 of 16 infiltrating cells played an important role in controlling the local bacterial infection. The cells bearing ST2 receptors are NK cells, NKT cells, mast cells, macrophages, dendritic cells, neutrophils, eosinophils, basophils, ILC2 cells, T H 2, and T reg [6][7][8]17,18,23]. To examine the effector cells in the air pouches after GAS infection, we distinguished the populations of infiltrating cells using the staining of specific cell markers by flow cytometry. As shown in Table 2, the predominant infiltrating cells in the air pouches after GAS infection were Gr1 + or CD11b + cells (88~96%), regardless of whether the mice were WT or KO, indicating that the major effector cells were neutrophils or monocytes [24,25]. Moreover, the microscopic observation of infiltrating cells confirmed that most effector cells were neutrophils (Supplementary Figure S2). No obvious T cells (CD3 + ) and NK cells (NK1.1 + ) were recruited into the air pouches after GAS infection. Although there were approximately 3~4% of B cells (CD19 + cells) in the air pouches, no statistically significant differences were found among different groups. However, the ILC2 cell ( [26] population was less than 1% within the GAS-infected air pouch, and there were no statistical differences among different groups. Based on the results mentioned above, we suggest that neutrophils were the predominant effector cells in GAS-infected air pouches. Table 2. Analysis of cell populations of infiltrating cells in the air pouches of mice after GAS infection. Groups of four B6, IL-33-KO mice and ST2-KO mice were inoculated via the intra-air pouch route with 2 × 10 8 S. pyogenes NZ131 cells per mouse. In one group, IL-33-KO mice were given rIL-33 (1 µg per mouse) via the intra-air pouch route at 30 min before GAS infection. In the other group, B6 mice were treated with recombinant sST2/Fc fusion protein (1 µg per mouse) via the intra-air pouch route at 30 min before GAS infection. In control groups, B6 mice were given rIL-33 (1 µg per mouse) or sST2/Fc fusion protein (1 µg per mouse) only via the intra-air pouch route. At 24 h post-infection, exudates of the air pouch were collected, and the cell populations of infiltrating cells were examined using the specific cell marker staining. The results were expressed as the mean ± standard deviation. The infiltrating cells in control groups were counted and analyzed by the specific cell marker staining. The predominant infiltrating cell populations in both control groups were also Gr1 + or CD11b + cells.

IL-33/ST2 Axis Affected Neutrophil Migration in the In Vitro Model
To examine whether the IL-33/ST2 axis affected neutrophil migration, we collected neutrophils from the bone marrow of WT, IL-33-KO, and ST2-KO mice, and then detected them using the transwell assay. CXCL2 induced migration of neutrophils from WT mice, but not from IL-33-KO or ST2-KO mice, owing to the absence of the IL-33/ST2 axis which decreased CXCR2 expression on neutrophils [18]. Neutrophil migration induced by rIL-33 was more effective in WT neutrophils than in IL-33-KO neutrophils, although these findings were not statistically significant. However, rIL-33 had no recruitment effect on neutrophils from ST2-KO mice, indicating that the IL-33/ST2 axis clearly affected neutrophil migration ( Figure 4). Mimicking in vivo infection, we used GAS as a stimulant. We found GAS itself was able to attract neutrophil migration regardless of whether the neutrophils were from WT, IL-33-KO, or ST2-KO mice, even though there are differences in migration efficiency among the three groups, indicating that there were other chemoattractants, distinct from the IL-33/ST2 axis, which could attract ST2-deficient neutrophil migration during infection (Figure 4). Methods, and were stimulated by different stimulants using the transwell assay. The activity was examined by migrated cell numbers per HPF. Five HPFs were selected and each group and the results are expressed as the mean ± standard deviation. Treatment g compared for significance using ANOVA. * p < 0.05, ** p < 0.01, *** p < 0.001 compare relative groups as shown in the figure.

Absence of IL-33/ST2 Impaired the Bactericidal Activity of Sera and Neutrophils
The bactericidal activity of sera and neutrophils were also examined. The lected from different mice, were incubated with GAS for 2 h, and then the remn ria were counted. The sera, pretreated at 56 °C for 30 min, lost the bactericida regardless of whether it was collected from WT, IL-33-KO, or ST2-KO mice. Th cidal activity of sera, pretreated at 4 °C for 30 min, was significantly impaired in and ST2-KO mice, compared with that of WT mice ( Figure 5). To exclude the ef serum, the neutrophils, collected from the bone marrow, were incubated with G out being sera-opsonized, and then the bactericidal activity of the neutrophils w ined. The phagocytosis activity of neutrophils from different mice was first as the short-term infection of GAS. Neutrophils were infected with GAS at a mul infection (MOI) of 0.01 for 30 min, and then extracellular bacteria were killed medium containing gentamicin. The remnant intracellular bacteria were count terpreted as an indicator of the phagocytosis activity of neutrophils. As shown 6a, there were no significant differences in phagocytosis among the three grou trophils. In addition, neutrophils collected from WT, IL-33-KO, and ST2-KO m infected with GAS at a MOI of 0.01 for 4 h in an antibiotic-free RPMI medium remnant bacteria were counted, and the bactericidal activity of the neutrophils w Methods, and were stimulated by different stimulants using the transwell assay. The chemotaxis activity was examined by migrated cell numbers per HPF. Five HPFs were selected and counted in each group and the results are expressed as the mean ± standard deviation. Treatment groups were compared for significance using ANOVA. * p < 0.05, ** p < 0.01, *** p < 0.001 compared with the relative groups as shown in the figure.
2.6. Absence of IL-33/ST2 Impaired the Bactericidal Activity of Sera and Neutrophils The bactericidal activity of sera and neutrophils were also examined. The sera, collected from different mice, were incubated with GAS for 2 h, and then the remnant bacteria were counted. The sera, pretreated at 56 • C for 30 min, lost the bactericidal activity, regardless of whether it was collected from WT, IL-33-KO, or ST2-KO mice. The bactericidal activity of sera, pretreated at 4 • C for 30 min, was significantly impaired in IL-33-KO and ST2-KO mice, compared with that of WT mice ( Figure 5). To exclude the effect of the serum, the neutrophils, collected from the bone marrow, were incubated with GAS, without being sera-opsonized, and then the bactericidal activity of the neutrophils was examined. The phagocytosis activity of neutrophils from different mice was first assessed by the shortterm infection of GAS. Neutrophils were infected with GAS at a multiplicity of infection (MOI) of 0.01 for 30 min, and then extracellular bacteria were killed by RPMI medium containing gentamicin. The remnant intracellular bacteria were counted and interpreted as an indicator of the phagocytosis activity of neutrophils. As shown in Figure 6a, there were no significant differences in phagocytosis among the three groups of neutrophils. In addition, neutrophils collected from WT, IL-33-KO, and ST2-KO mice were infected with GAS at a MOI of 0.01 for 4 h in an antibiotic-free RPMI medium, then the remnant bacteria were counted, and the bactericidal activity of the neutrophils was examined. The result indicated that the bactericidal activity of the neutrophils was significantly weakened in IL-33-KO and ST2-KO mice (Figure 6b). Taken together, these results suggest that the innate immunity, including that of the serum and neutrophil defense, was defective in IL-33/ST2-deficient mice. Figure 5. Serum killing activity against GAS. Sera collected from WT, IL-33-KO and ST2-K were diluted 2-fold with PBS, then incubated at 4 °C or 56 °C for 30 min, and then were with 1.5 × 10 4 CFU of GAS (NZ131) for another 2 h. At the indicated time, bacteria were qua by plate counting. One of three experiments is represented, and the results are expresse mean ± standard deviation. Treatment groups were compared for significance using ANOV < 0.001, compared with the relative group as shown in the figure. Figure 6. The phagocytosis activity and the bactericidal activity of neutrophils collected from WT, IL-33-KO, and ST2-K mice. Bone marrow neutrophils (1.5 × 10 6 cells) from WT, IL-33-KO, and ST2-KO mice were infected with GAS NZ131 ce at MOI of 0.01 for 30 min at 37 °C using a rotating mixer. After that, the extracellular bacteria were eliminated using RP medium containing gentamicin at 37 °C for 30 min. (a) After several washes by RPMI medium, the cells were lysed a intracellular viable bacteria were quantified by plating on THY agar. The bacterial numbers presented on this timepo represent the phagocytosis of neutrophils. (b) Bone marrow neutrophils (1.5 × 10 6 cells) from WT, IL-33-KO, and ST2-K mice were infected with GAS NZ131 cells at MOI of 0.01 for 4 h at 37 °C using a rotating mixer. The remaining via bacteria were counted, representing the bactericidal activity of neutrophils. Treatment groups were compared for sign cance using ANOVA. * p < 0.05, ** p < 0.01, *** p < 0.001 compared with the relative group as shown in the figure.

Discussion
The incidences of invasive GAS infection are approximately 2 to 4 per 100,000 in developed countries and 12 to 83 per 100,000 people in developing countries. Th tality rate of streptococcal toxic shock syndrome may exceed 50%, despite antibi

Discussion
The incidences of invasive GAS infection are approximately 2 to 4 per 100,000 peopl in developed countries and 12 to 83 per 100,000 people in developing countries. The mor tality rate of streptococcal toxic shock syndrome may exceed 50%, despite antibiotic ad ministration and aggressive treatment [27]. Even though GAS vaccines based on th GAS's virulence factor, M protein, have been developed [28,29], the safety issue due t Figure 6. The phagocytosis activity and the bactericidal activity of neutrophils collected from WT, IL-33-KO, and ST2-KO mice. Bone marrow neutrophils (1.5 × 10 6 cells) from WT, IL-33-KO, and ST2-KO mice were infected with GAS NZ131 cells at MOI of 0.01 for 30 min at 37 • C using a rotating mixer. After that, the extracellular bacteria were eliminated using RPMI medium containing gentamicin at 37 • C for 30 min. (a) After several washes by RPMI medium, the cells were lysed and intracellular viable bacteria were quantified by plating on THY agar. The bacterial numbers presented on this timepoint represent the phagocytosis of neutrophils. (b) Bone marrow neutrophils (1.5 × 10 6 cells) from WT, IL-33-KO, and ST2-KO mice were infected with GAS NZ131 cells at MOI of 0.01 for 4 h at 37 • C using a rotating mixer. The remaining viable bacteria were counted, representing the bactericidal activity of neutrophils. Treatment groups were compared for significance using ANOVA. * p < 0.05, ** p < 0.01, *** p < 0.001 compared with the relative group as shown in the figure.

Discussion
The incidences of invasive GAS infection are approximately 2 to 4 per 100,000 people in developed countries and 12 to 83 per 100,000 people in developing countries. The mortality rate of streptococcal toxic shock syndrome may exceed 50%, despite antibiotic administration and aggressive treatment [27]. Even though GAS vaccines based on the GAS's virulence factor, M protein, have been developed [28,29], the safety issue due to autoreactive rheumatic heart disease by repeat immunization with recombinant M protein needs to be reassessed [30]. Hence, the therapeutic strategy of GAS invasive infection needs to be explored. Based on the clinical observations, GAS invasive infection with a high mortality is associated with the cytokine storm [2][3][4]. In this study, we examined the T H 2 cytokine, IL-33, on a GAS-induced necrotizing fasciitis model. Our results indicate that the IL-33/ST2 axis provided a protective effect on the GAS-induced air pouch infection model, and its mechanism was through enhancing the innate immunity, including increasing the bactericidal activity of sera as well as neutrophils (Figures 5 and 6b), and enhancing neutrophil migration into the infected sites (Figure 3b). Within the air pouches, we found that inflammatory cytokines, instead of T H 1 or T H 2 cytokines, were significantly influenced by the IL-33/ST2 axis ( Table 1), indicating that the protective effect of the IL-33/ST2 axis in our model was independent of a type 2 response. This observation is similar to the experimental sepsis using the CLP model, inducing acute polymicrobial abdominal sepsis, in which IL-33 reduces mortality and improves bacterial clearance by mechanisms involved in functions of neutrophils, γδ T, and NK cells, but not T H 2 cytokines [18,[31][32][33]. However, the local environments and the virulence, as well as dosages of pathogens, seems to influence the outcomes caused by the IL-33/ST2 axis. ST2 deficiency improves the outcome of Staphylococcus aureus-induced septic arthritis using the intraarticular infection model [34]; in a systemic infection model induced by a lethal intravenous dose of S. aureus, IL-33 administration protects against lethality via ILC2-dependent type 2 cytokine production [35]; in post-influenza S. aureus pneumonia, IL-33 diminishes bacterial loads and mortality by an effect that does not require ILC2 cells [36]. Furthermore, IL-33 increases bacterial loads and lethality through induction of type 2 cytokines in Legionella pneumophila-induced pneumonia [37], while IL-33 improves local immunity during Klebsiella pneumoniae-induced pneumonia by a combined effect on neutrophils and inflammatory monocytes but not on B, T, NK and ILC2 cells [38].
In our air pouch infection model, we found that the local bacterial loads at the early stage played an important role in deciding the outcome. The IL-33/ST2 axis not only enhanced the number of infiltrating cells (Figure 3b), but also prolonged the survival of infiltrating cells (Figure 3c) at the early infective stage, and then accelerated the bacterial clearance ( Figure 3a) and further decreased the mortality (Figure 1). Comparing IL-33-KO mice with ST2-KO mice after GAS infection, we found that ST2-KO mice had fewer infiltrating cells and lower cell viability of infiltrating cells than those of IL-33-KO mice (Figure 3b,c), indicating that ST2 receptors expressed on cells played a decisive role in influencing the survival of infiltrating cells and cell migration at the beginning of infection. Based on cell marker staining by flow cytometry (Table 2) and microscopic observation of infiltrating cells (Supplementary Figure S2), we found that neutrophils, but not T, B, NK, ILC2 cells, were the predominant effector cells within the GAS-infected air pouches. Using the transwell assay, we found that the migration of neutrophils induced by rIL-33 seemed more effective in WT neutrophils than in IL-33-KO neutrophils, although differences were not statistically significant. However, rIL-33 had no recruitment effect on neutrophils of ST2-KO mice, indicating that the IL-33/ST2 axis clearly affected neutrophil migration ( Figure 4). However, we could not rule out that there may be other unidentified molecules, other than IL-33, that could bind to ST2 receptors to induce neutrophil migration during the early phase of infection. The directed neutrophil recruitment is orchestrated by chemoattractants, such as lipids, N-formylated peptides, complement molecules, anaphylotoxins and chemokines, produced during bacterial infection [39]. We found GAS itself was able to attract neutrophil migration using the transwell assay, regardless of whether the neutrophils were collected from WT, IL-33-KO, or ST2-KO mice, even though there are differences in migration efficiency among the three groups ( Figure 4). This indicates that there were other chemoattractants released by GAS which could attract ST2-deficient neutrophil migration during infection. This finding explains why the numbers of infiltrating cells within the air pouches showed no differences among WT, IL-33-KO, and ST2-KO mice at 48 h post-infection (Figure 3b).
In addition to neutrophil recruitment, we found that bacterial clearance in the infected areas was also significantly affected by the IL-33/ST2 axis (Figure 3a). The first defense at the beginning of the bacterial infection is dependent on the complement system and phagocytes. To examine the effect of the complement system, we collected sera from different strains of mice, and the bactericidal activity in vitro was further examined. The results showed that sera from IL-33-KO and ST2-KO mice had weaker bactericidal activity than those of WT mice, and the remnant bacterial numbers of the KO group was five times greater than in the WT group after in vitro infection ( Figure 5). Pretreatment of sera at 56 • C for 30 min destroyed the bactericidal activity of sera, regardless of whether from the WT, IL-33-KO, or ST2-KO group (Figure 5), indicating that the bactericidal activity of sera was complement-dependent. No literature indicates that the IL-33/ST2 axis affects the production of complement molecules. Hepatocytes can produce IL-33, and they also express ST2 receptors [40]. Whether the IL-33/ST2 axis affects complement production from hepatocytes needs further investigation. Reports have indicated that the virulence factors of GAS interfere with the bactericidal activity of complements [41][42][43][44]. As the production of complements was impaired in IL-33-KO and ST2 KO mice, GAS may defeat the innate immunity and rapidly replicate in KO mice. Bypassing the complement effect, the bactericidal activity of neutrophils was also examined in this study. The result indicates that the bactericidal activity of neutrophils collected from KO mice was weakened, especially in those from ST2-KO mice (Figure 6b). A previous report indicates that the phagocytosis and production of reactive oxygen species (ROS) of phagocytes are diminished in ST2deficient mice [45]. However, our results showed that there were no statistical differences in phagocytosis activity of neutrophils among WT and KO mice ( Figure 6a). As for identifying which of the bactericidal weapons of neutrophils were affected by the IL-33/ST2 axis, this needs further examination.
In this study, we first demonstrated that the IL-33/ST2 axis was involved in protecting the mice from GAS-induced air pouch infection. Its protective mechanism was through enhancing neutrophil migration and the bactericidal activity of sera and neutrophils at the early stage of infection. These effects accelerated the bacterial clearance, following a decrease in the local inflammation, and further decreased the mortality of mice. In our air pouch infection model, no B, T, NK, or ILC2 cells were involved, and the IL-33/ST2 axis distinctly influenced the innate immunity to clear GAS infection. Based on our results shown in Figure 3, adding rIL-33 into the air pouch of IL-33-KO mice before infection significantly enhanced the bacterial clearance. The local treatment with rIL-33 provided a new therapeutic strategy on GAS invasive infection.

Bacterial Strain
The GAS strain NZ131 (type M49, T14) was a gift from Dr. D. R. Martin, New Zealand Communicable Disease Center, Porirua. The GAS was grown in Todd-Hewitt medium supplemented with 0.2% yeast extract (THY) (Difco Laboratories) for 12 h at 37 • C and then subcultured into fresh broth (1:50 [vol/vol]) for another 5 h to produce the log-phase bacteria. The bacterial concentration was determined with a spectrophotometer (Beckman Instruments) by measuring the optical density at 600 nm. For quantitating the exact bacterial concentration, the bacterial suspension was serially diluted with sterile phosphatebuffered saline (PBS), then 0.1 mL was poured on THY agar plates, and incubated at 37 • C overnight [46].

Mice
C57BL/6 (B6; WT) mice were purchased from the National Laboratory Animal Center in Taiwan. IL-33-KO and ST2-KO mice [47] were donated by Dr. Wei-Yu Chen, Kaohsiung Chang Gung Memorial Hospital, Taiwan, and were bred in the animal center at I-Shou University. All animals were maintained on standard laboratory chow and water ad libitum in the animal center at I-Shou University. The animals were raised and cared for in accordance with guidelines established by the Ministry of Science and Technology in Taiwan. All procedures, care and handling of the animals were reviewed and approved by the Institutional Animal Care and Use Committee at I-Shou University (ISU107026). The mice used in the experiments were 7 to 8-weeks-old (body weight 23 ± 1.0 g per mouse).

Air Pouch Model of GAS Infection and Evaluation of Skin Lesion
Groups of six WT, IL-33-KO, and ST2-KO mice were anesthetized by isoflurane inhalation and then injected subcutaneously with 1 mL of air to form an air pouch one day before the infection. One day later, 0.1 mL of bacterial suspension containing 3 × 10 8 S. pyogenes NZ131 cells was inoculated into the air pouch [48]. The animals were observed every day for a total of 13 days. Survival curves were then determined. At 48 h post-infection, the degrees of skin lesion were quantitated by an exudates-absorbed method, which is based on the premise that the more severe the skin damage, the more that the exudates will effuse. A Kimwipes (Kimberly-Clark Global Sales, Inc., Roswell, GA, USA) paper was attached to the air pouch area, and the wet area of the paper was then gauged [49]. The average lesion area in each group was generated by examination of skin lesions from six mice. After that, the skin lesion tissue around the air pouch was excised, fixed in 3.7% formaldehyde, and embedded in paraffin. The 5-µm-thick tissues were sliced and stained with hematoxylin and eosin. In some experiments, 0.1 mL of mouse rIL-33 (R&D Systems) (1 µg per mouse) and sST2 protein (R&D Systems) (1 µg per mouse) were injected into the air pouches of IL-33-KO mice and WT mice, respectively, at 30 min before GAS infection or 24 h post-infection.

Examination of the Inflammation and the Populations of Infiltrating Cells within the Air Pouches
Groups of four WT, IL-33-KO, and ST2-KO mice were inoculated with 0.1 mL of bacterial suspension containing 3 × 10 8 S. pyogenes NZ131 cells into the air pouch. At 48 h post-infection, the mice were anesthetized by isoflurane inhalation, then the blood was collected by the cardiac puncture, and the exudates of the air pouch were collected by injecting 1 mL of sterile PBS into the air pouch and aspirating the exudates by syringe with an 18-gauge needle. The blood and exudates were examined by a capture IL-33 enzymelinked immunoassay (ELISA) kit (R&D Systems). For evaluation of the inflammation after GAS infection in the groups of four WT, IL-33-KO, and ST2-KO mice, WT mice given sST2 proteins, and IL-33-KO mice given rIL-33, were inoculated with 0.1 mL of bacterial suspension containing 2 × 10 8 S. pyogenes NZ131 cells into the air pouch. At different times after infection, the exudates of the air pouch were collected. For quantitation of bacterial numbers within the air pouch, 10 µL of the exudates were lifted, underwent serial dilution, poured (0.1 mL) in THY agar plates, and incubated at 37 • C overnight. The number of CFU of S. pyogenes was then quantitated and expressed as the mean ± standard deviation. The remnant exudates were then centrifuged for 10 min at 1200 rpm. The sediment containing the infiltrating cells was collected, and the number and viability of the infiltrating cells were further determined by an eosin Y exclusion assay [49]. The number and the cell viability of infiltrating cells were then quantitated and expressed as the mean ± standard deviation. In addition, infiltrating cells were further stained by antibodies of different cell markers, including CD3ε monoclonal antibody (145-2C11) conjugated with APC (eBioscience 17-0031-83), CD19 monoclonal antibody (eBio1D3) conjugated with PE (eBioscience 12-0193-82), NK1.1 monoclonal antibody (PK136) conjugated with PE (eBioscience 12-5941-81), Gr1 (Ly-6G/Ly-6C) monoclonal antibody (RB6-8C5) conjugated with FITC (eBioscience 11-5931-85), CD11b monoclonal antibody (M1/70) conjugated with PerCP-Cy™5.5 (BD Pharmingen 550993). The percentages of the cell marker staining were expressed as the mean ± standard deviation. Moreover, the infiltrating cells were further assessed by the cytospin and Liu stain, and were then observed by microscope. The expression of cytokines, containing TNFα, IL-17A, IL-6, IL-1β, IL-12, IFN-γ, IL-4, IL-13 and IL-10, in the supernatants of exudates was also examined using a cytometric bead array kit (BD Biosciences). In this detection kit, the limit of cytokine detection was 5 pg/mL. The concentrations of cytokines in the air pouches of mice were quantitated and expressed as the mean ± standard deviation.

Chemotaxis Assay of Mouse Neutrophils Isolated from Bone Marrow
WT, IL-33-KO, and ST2-KO mice were sacrificed, and femur and tibia from both hind legs were removed and cleared of remaining soft tissue attachments. After that, the extreme distal tip of each extremity was cut off and the bones flushed with HBSS medium (with calcium, magnesium, Gibco) containing 0.5% bovine serum albumin and 1% glucose (SigmaAldrich) and 20 mM HEPES (Merck-SigmaAldrich) at a pH of 7.2, using a 25G needle. After, the bone marrow was flushed on ice, then centrifuged at 1300 rpm for 10 min, and the sediment containing the cells was treated with the RBC lysing buffer (Sigma) to eliminate RBCs. The remaining cells were suspended in 2 mL of HBSS medium and treated on a three-layer Percoll (GE Heathcare) density gradient, 72%, 63%, and 50%, respectively, diluted in HBSS, and centrifuged for 25 min at 3000 rpm at 4 • C. Neutrophils were harvested from the 63%/72% interface after carefully removing the cells from the upper phases [50]. The cell preparations contained >90% neutrophils as determined by visual counting. The neutrophil populations with viability over 95%, as assessed by trypan blue exclusion, were used to conduct the chemotaxis assay. The hanging cell insert with 8-µm membrane (Millipore) was hung in a 24-well culture plate (Nunc) to examine the chemotaxis of neutrophils [18]. Neutrophils (1 × 10 6 cells/mL) were added in the hanging cell insert, and different stimulants, such as PBS, CXCL2 (100 ng/mL), mouse rIL-33 (50 ng/mL), or GAS NZ131 cells (2 × 10 6 CFU/mL), were added to the 24-well culture plate. After a 1 h culture, the hanging cell insert was removed, and the membrane was cut, fixed, and stained with 0.05% crystal violet (Sigma). The bottom layer of the membrane was examined by microscope, and the number of cells on each high-power field (HPF) was counted, and then cells of 5 HPFs from each group were quantitated and expressed as the mean ± standard deviation.

Serum Bactericidal Assay
The sera from WT, IL-33-KO, and ST2-KO mice were collected. The sera, diluted 2-fold with PBS, were incubated at 56 • C for 30 min to inactivate the complement system. GAS NZ131 cells (1.5 × 10 4 CFU in 100 µL PBS) were incubated with 5 µL of diluted sera, pretreated for 30 min at either 4 • C or 56 • C, and at 37 • C for 2 h using a rotating mixer. After that, the remaining viable bacteria were diluted with PBS, and then quantified by plating on THY agar plates. One of three experiments was represented, and the results were expressed as the mean ± standard deviation.

Phagocytosis and Bactericidal Assay of Neutrophils
Neutrophils from the bone marrow were separated using a Percoll (GE Heathcare) density gradient, as described previously. Neutrophils (1.5 × 10 6 cells in 0.2 mL RPMI medium, Gibco) from WT, IL-33-KO, and ST2-KO mice were infected with GAS (50 µL) at a MOI of 0.01 for 30 min at 37 • C using a rotating mixer. After that, cell suspensions were centrifuged at 6000 rpm for 10 min at 4 • C. The supernatants were discarded, and the sediment cells were cultured with RPMI medium containing gentamicin (50 µg/mL) at 37 • C for 30 min to kill extracellular bacteria. After several washes by RPMI medium, the cells were lysed, and intracellular viable bacteria were then quantified by plating on THY agar plates as described previously [48]. The bacterial numbers presented on this timepoint represent the phagocytosis of neutrophils.
Neutrophils (1.5 × 10 6 cells in 0.2 mL RPMI medium, Gibco) from WT, IL-33-KO, and ST2-KO mice were infected with GAS (50 µL) at a MOI of 0.01 for 4 h at 37 • C. After that, cell suspensions were centrifuged, the supernatants were collected, the sediment cells were lysed. The intracellular viable bacteria and remaining viable bacteria in the supernatants were further quantified by plating on THY agar plates as described previ-ously [48]. The bacterial numbers presented on this timepoint represent the bactericidal activity of neutrophils.

Statistics
The statistical analysis was conducted using Prism 5.0 software (GraphPad Software, San Diego, CA, USA). In Figure S1, treatment groups were compared for significance using the t-test. For the mouse survival in Figure 1, survival curves were compared for significance using the logrank test. The data shown in Figures 3-6 and the tables were compared for significance using ANOVA. Statistical significance was set at * p < 0.05, ** p < 0.01, ***p < 0.001.

Supplementary Materials:
The following are available online at https://www.mdpi.com/article/10.3 390/ijms221910566/s1, Figure S1: The IL-33 expression in GAS-infected B6 mice, Figure S2  Institutional Review Board Statement: The study was conducted according to the guidelines of the Ministry of Science and Technology in Taiwan, and approved by the Institutional Animal Care and Use Committee at I-Shou University (ISU107026, 2019/01/08).

Informed Consent Statement: Not applicable.
Data Availability Statement: The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.