Protective efficacy of the chimeric Staphylococcus aureus vaccine candidate IC in sepsis and pneumonia models

Staphylococcus aureus causes serious sepsis and necrotic pneumonia worldwide. Due to the spread of multidrug-resistant strains, developing an effective vaccine is the most promising method for combating S. aureus infection. In this study, based on the immune-dominant areas of the iron surface determinant B (IsdB) and clumping factor A (ClfA), we designed the novel chimeric vaccine IsdB151-277ClfA33-213 (IC). IC formulated with the AlPO4 adjuvant induced higher protection in an S. aureus sepsis model compared with the single components alone and showed broad immune protection against several clinical S. aureus isolates. Immunisation with IC induced strong antibody responses. The protective effect of antibodies was demonstrated through the opsonophagocytic assay (OPA) and passive immunisation experiment. Moreover, this new chimeric vaccine induced Th1/Th17-skewed cellular immune responses based on cytokine profiles and CD4+ T cell stimulation tests. Neutralisation of IL-17A alone (but not IFN-γ) resulted in a significant decrease in vaccine immune protection. Finally, we found that IC showed protective efficacy in a pneumonia model. Taken together, these data provide evidence that IC is a potentially promising vaccine candidate for combating S. aureus sepsis and pneumonia.

Animals and bacterial strains. BALB/c and C57BL/6 mice (7-to 8-week-old females) were purchased from Beijing HFK Bioscience Limited Company (Beijing, People's Republic of China). IL-17A gene knockout (IL-17AKO) mice (C57BL/6 background) were kindly provided by Richard A. Flavell (Yale University School of Medicine, New Haven, CT, USA). Female New Zealand white rabbits (weighing 2.00 ± 0.20 kg) were provided by TengXin Company (Chongqing, People's Republic of China). The animals were maintained under specific-pathogen-free (SPF) conditions. S. aureus strain MRSA 252 was purchased from the American Type Culture Collection (Manassas, VA, USA). The WHO2 strain was provided by Hong Zhou (Third Military Medical University, China). The three other clinical S. aureus isolates were collected from three different hospitals in China (Supplemental Table 1). The bacteria were cultured in tryptic soy broth at 37 °C for 6 h, centrifuged at 5000 × g for 5 min, washed with phosphate-buffered saline (PBS) 3 times, and diluted with PBS to an appropriate cell concentration using spectrophotometry at 600 nm. PCR amplification. The complete genome of strain MRSA 252 was used as the PCR template. IsdB  was generated in our previous work 30 . The ClfA  gene was amplified by PCR primers P3 and P4 (Supplemental Table 2). For the IsdB 151-277 -TACGTTCCGGTTGATGTT-ClfA  construct, the first round of PCR was performed using the primers P1 and P2 (Supplemental Table 2) to generate IsdB 151-277 -TACGTTCCGGTTGATGTT. Similarly, the primers P3 and P4 were used to generate TACGTTCCGGTTGATGTT-ClfA  . By employing the first-round PCR products as the templates, the second round of PCR was performed using the primers P1 and P4. BamHI and NotI sites were introduced at the beginning and end of the PCR products through the primers to obtain the amplified genes.
Expression and purification of recombinant proteins. Briefly, the PCR product was cloned into pGEX-6p-2 (Sangon Biotech, shanghai) and transformed into the Escherichia coli strain Xl1-blue (Biovector Scienc Lab, Beijing). Escherichia coli strain Xl1-blue containing the recombinant plasmid was induced with isopropyl-b-D-1-thiogalactopyranoside (IPTG) at a final concentration of 0.1 mM for GST fusion protein expression. We purified the GST-tagged proteins from the cleared lysates by Capto MMC (GE), and the GST tag was cleaved by preScission protease (GE). Then, we removed the endotoxin from the protein eluate using Triton X-114 phase separation as described elsewhere 31 . The resulting protein was analysed by gel-filtration using the Superdex TM 200 10/300GL column (GE). Protein purity was determined using sodium dodecyl sulphate-polyacrylamide gel electrophoresis (SDS-PAGE) and using high-performance liquid chromatography (HPLC) with a C3 column. The concentration of the resulting protein was determined using the bicinchoninic acid (BCA) method (Pierce). The endotoxin content was detected using the tachyplens ameboyto lysate assay (Houshiji cod Inc., Xiamen, China).
Scientific RepoRts | 6:20929 | DOI: 10.1038/srep20929 Immunisation of the mice and S. aureus infection. For active immunisation, purified proteins were dissolved in PBS and emulsified 1:1 (volume ratio) in AlPO 4 (Pierce), Al(OH) 3 (Pierce), or Freund's adjuvant (Sigma). Mice were intramuscularly injected with 100 μ L of the emulsion containing 20 μ g protein, isovolumetric PBS plus adjuvant, or PBS alone as the control on days 0, 14, and 21. On day 28, the mice were exsanguinated, and serum samples were collected for the enzyme-linked immunosorbent assay (ELISA).
For the pneumonia model, C57BL/6 mice were anaesthetised with isoflurane (RWD Life Science) and inoculated with 4 × 10 8 CFUs of a MRSA 252 suspension in the naris. Then, the mice were monitored for 7 days after infection. Thermalert TH-5 (Physitemp) was used for determination of the intrarectal temperature of infected mice.
Quantitative bacteriology in organs. Lungs, spleens, and kidneys from corresponding animals were removed, weighed, and homogenised in 2 mL of PBS. Peripheral blood was collected in heparin anticoagulant tubes. All samples used for quantitative cultures were grown on tryptic soy broth for 24 hours at 37 °C.
Histological analysis. Kidney tissues were obtained 3 days post-infection from the sepsis model mice, and lung tissues were collected 1, 3, and 7 days post-infection from the pneumonia model mice. The organs were fixed with 10% formalin and embedded in paraffin.
A double-blind histological analysis was performed to determine the severity of each section of lung or kidney. Each lung section from the pneumonia model mice was given a score of 0-4 (no abnormality to most severe) according to established criteria 32 . Each kidney section from the sepsis model mice was given a score of 0-4 (0: no abnormality; 1: area of renal tubular interstitial lesion < 5%; 2: 5%-25%; 3: 25%-75%; and 4: > 75%).
ELISA for specific antibodies. The ELISA was performed as previously described 30 . Serum samples were used as the primary antibodies. The secondary antibodies were HRP-conjugated goat anti-mouse IgG (Southern Biotech, Birmingham, AL, USA), anti-IgG1, or anti-IgG2a (Sigma). The titres were diluted and monitored at OD450 or defined as the highest dilution that yielded an absorbance value of more than twice the value of the blank control.
Opsonophagocytic assay. The assay was performed based on a method described by Burton and Nahm 33 .
Briefly, MRSA 252 was used as the target strain, and the HL60 cell line (ATCC, CCL-240) was used as the opsonophagocytic cells. HL60 cells were maintained in L-glutamine-containing IMDM medium (HyClone) supplemented with 20% heat-inactivated FBS (HyClone), 100 U/mL penicillin, 100 μ g/mL streptomycin, and 0.25 μ g/mL amphotericin (Mediatech) at 37 °C in 5% CO 2 . To differentiate the granulocytes, HL60 cells (5 × 10 5 cells/mL) were cultured in medium with 0.8% N,N-dimethylformamide (Sigma) for 4 days. Trypan blue tests were performed to verify HL60 cell viability, and results ≥ 90% were considered acceptable for the opsonophagocytic assay (OPA). The assay was performed in 96-well round-bottom plates. Each well (volume, 80 μ L) contained 4 × 10 5 HL60 cells, 10 3 CFUs of MRSA 252, vaccinated rabbit antisera, and 3-to 4-week-old infant rabbit serum as a complement source. Then, the microtitre plates were shaken on a mini-orbital shaker (700 rpm) for 45 minutes in an incubator (37 °C with 5% CO 2 ). For termination, the microtitre plates were placed on ice for 20 minutes. Next, 20 μ L of the reaction mixture from each well was spotted onto tryptic soy broth. The samples were plated in duplicate, and the killing effect was defined as a reduction in CFUs after overnight growth. Control samples were incubated with normal rabbit serum (NRS). The percentage of opsonophagocytic killing was determined by subtracting the number of colonies that survived the test assay from the number of CFUs on the NRS control.
Passive immunisation. To generate rabbit polyclonal antibodies, an immunisation procedure was performed on rabbits similar to the procedure performed on mice. On day 28, the serum was pooled and the IgG fraction was obtained. The IgG concentration was determined by the BCA method. One day before infection, 5 mg of antibody per mouse was passively immunised via the intraperitoneal (i.p.) route (n = 10). Survival was monitored for 12 days.
Splenocyte stimulation test. The amounts of IFN-γ and IL-17A in the cell culture supernatants were determined based on a previously described method 30 . Spleens were aseptically removed, and cells were suspended at a concentration of 2 × 10 6 cells/mL in complete media (RPMI 1640 with 10% FBS). The cells were stimulated with or without 10 μ g/mL of IsdB, ClfA 33-213 , or IC protein at 37 °C for 5 days. The supernatants were collected, and the amounts of IFN-γ and IL-17A were determined by ELISA using mouse IFN-γ ELISA and IL-17A ELISA kits (Biolegend), respectively.
Statistical analysis. Data are presented as the means ± SD or means ± SEM. The non-parametric log rank test was used for determining the differences in the survival rates. The Student's t-test was used for determining the antibody titers. Otherwise, the non-parametric Mann-Whitney test was used for analysis the bacterial burdens and the severity score of kidneys and lungs. GraphPad Prism 5.0 (GraphPad Software) was used for data analyses, and a p-value < 0.05 was considered significant.

Results
Expression and purification of IsdB, ClfA 33-213 , and IC. We amplified IsdB 151-277 and ClfA 33-213 by PCR and combined them using the "YVPVDV" linker ( Fig. 1a). Recombinant ClfA  and IC were expressed in Escherichia coli strain Xl1-blue in the soluble fraction. We purified IsdB 35-629 in our previous work 30 ; this peptide was used as the control for the individual IsdB component in this study. As demonstrated by SDS-PAGE in (e) BALB/c mice (n = 10) were challenged by an intravenous injection of WHO2 (5 × 10 8 CFUs) and clinical S. aureus isolates (3 × 10 8 CFUs) after immunisation. Survival rates were monitored for 12 days. Compared with animals receiving the AlPO 4 , the significance of the protective immunity generated by the various antigens was measured with a log rank test. The asterisks represent a statistically significant difference ( * P < 0.05, ** P < 0.01, *** Scientific RepoRts | 6:20929 | DOI: 10.1038/srep20929 Fig. 1b, the masses of the recombinant proteins IsdB, ClfA  , and IC were in accordance with their predicted molecular masses (98, 19, and 36 kDa, respectively).

Validation of the optimal adjuvant and protective effect of IC in a lethal S. aureus sepsis model.
To choose an optimal adjuvant for the IC immunisation strategy, we immunised mice with the same dose of IC plus different adjuvants (Freund's adjuvant, AlPO 4 , and Al(OH) 3 ). The Freund's adjuvant group showed significantly high titres in the serum antibody detection test after the first two immunisations; however, after the third boost, AlPO 4 induced titres nearly the same as those detected for Freund's adjuvant (Fig. 1c). Because Freund's adjuvant is not recommended for use in humans due to its side effects, we chose AlPO 4 as the optimal adjuvant in the following experiments. Mice were immunised with IsdB, ClfA  , and IC plus AlPO 4 as the adjuvant, and the protective effects against S. aureus infection were evaluated in a sepsis model. The animals vaccinated with the IsdB, ClfA 33-213 , or IC antigens exhibited higher survival rates (55%, 50%, and 85% 12 days post-infection, respectively) compared with the AlPO 4 group (15% survival) and the PBS group (5% survival). The significance of the protective effects compared with those of AlPO 4 was assessed using a log rank test (P IsdB = 0.0151; P ClfA33-213 = 0.0423; and P IC < 0.0001). The chimeric vaccine yielded a higher protective effect than IsdB or ClfA 33-213 alone (Fig. 1d, P IC-IsdB = 0.0498 and P IC-ClfA33-213 = 0.0260). Therefore, vaccination with IC can generate a relatively ideal protective effect in a lethal S. aureus sepsis model.
We used several S. aureus strains to test the broadness of the protective effect against S. aureus. We collected several clinical S. aureus isolates from different hospitals throughout China (Beijing, Guangzhou, and Chongqing), and all three of the isolates expressed both IsdB and ClfA (Supplementary Table S1). Then, we validated the protective effects against WHO2 and the three clinical S. aureus isolates. As shown in Fig. 1e, immunisation with IC induced a broad protective effect against WHO2 and the three clinical isolates. These data indicated that this novel vaccine yielded broad protective effects against several S. aureus strains in a sepsis model. Immunisation with IC significantly reduced the bacterial burdens in different organs and decreased pathological damage to the kidneys. The bacterial burdens in the blood, spleens, and kidneys were calculated 1 and 3 days post-infection. In the blood, the bacterial burdens were much lower in the IC group 1 day post-infection compared with IsdB (P = 0.0370) and ClfA 33-213 (P = 0.0060) (Fig. 2a). However, there was no significant difference in these three vaccinated groups 3 days post-infection (Fig. 2d). In the kidneys and spleens, the elimination of S. aureus was enhanced in the IC group within 3 days post-infection compared with the IsdB and ClfA groups (Fig. 2a-f). These results showed that immunisation with IC protected against S. aureus infection by reducing S. aureus colonisation in organs and preventing a direct attack on the organs by the bacteria. Consistent with the survival rates, immunisation with IC induced a higher protective effect against S. aureus by reducing the bacterial burden.
Based on our histological analysis, the kidneys from the mice in the recombinant vaccine immunised groups showed less inflammatory cell infiltration and renal abscesses after infection compared with the AlPO 4 group (Fig. 2g). Furthermore, the glomeruli in the IC group showed less damage, with fewer inflammatory cells detected in the kidneys. In contrast, the kidneys in the AlPO 4 group harboured bacterial abscesses with several large foci of staphylococci (Fig. 2g). Taken together, the results demonstrate that the severity of the kidney damage was significantly lower in the IC group than in the IsdB and ClfA 33-213 groups (Fig. 2h).
Active immunisation with IC induced strong antibody responses. Next, we determined the antigen-specific antibody responses in the sera. The recombinant proteins plus AlPO 4 produced strong antibody responses. The total titres of IgG and the IgG subgroups (IgG1 and IgG2a) in the IC-immunised mouse group were much higher compared with those in the mice immunised with IsdB or ClfA   (Table 1).
To determine whether antibodies were generated against both components in the IC-immunised group, we analysed antigen-specific IgG in the sera of the IC group (Table 2). There were no significant differences in the levels of antibodies targeting IsdB and ClfA induced by IC compared with those generated by immunisation with the single components alone. These results show that immunisation with IC alone can generate strong antibody responses against both IsdB and ClfA.

Opsonophagocytic assay and passive immunisation test corroborated the importance of antibodies in vaccine protective efficacy.
To evaluate the efficacy of the IC-specific antibodies, an OPA that measured antibody-and complement-mediated bacterial killing was performed in vitro. The killing assay was performed using immune cells (neutrophils) that play a critical role in the host clearance of S. aureus. In the presence of HL60 phagocytic cells and complement, rabbit serum raised against the three proteins exhibited opsonophagocytic activity against S. aureus. As shown in Fig. 3a, the percentages of S. aureus killed in the IC group (70.1%), IsdB group (56.3%), and ClfA 33-213 group (51.8%) were higher than those killed in the AlPO 4 group (17.5%). The IC-specific antibodies were more efficient than the IsdB-specific (P = 0.0300) and ClfA 33-213 -specific antibodies alone (P = 0.0035). These results indicated that the IC-specific antibodies were more effective in killing S. aureus; thus, the antibody response may functionally regulate phagocytosis and killing of S. aureus by innate immune cells.
We confirmed that IC could induce functional antibodies that were effective in vitro. However, whether these antigen-specific proteins provided protective immunity in vivo remained unclear. To verify the protective role of antibodies in vivo, we performed passive immunisation. The mice passively immunised with IC showed higher protection (80% survival) compared with the IsdB (50% survival), ClfA 33-213 (30% survival), and AlPO 4 IgG-treated groups (0% survival) (Fig. 3b). Anti-IC antibodies showed higher protective effects compared with those in the ClfA 33-213 group and IsdB group. We conclude that anti-IC antibodies effectively protect mice against S. aureus infection.  Bacterial burdens in the blood, spleens and kidneys (n = 8) were calculated at 1 (a-c) and 3 days (d-f) postinfection. Representative results from one of three independent experiments were shown. Data were presented as box and whisters, and the medians were shown. Asterisks indicate significant differences between two groups ( * P < 0.05, ** P < 0.01, *** P < 0.001). (g) HE-stained kidneys from the AlPO 4 group and recombinant vaccine immunised groups at 3 days post-infection were shown. BALB/c mice were immunised with recombinant proteins. Three days post-infection (5 × 10 8 CFUs), the kidneys were collected, and representative histopathological sections were shown (magnification = 400×). The kidney tissue of mouse without infection were used as control. Arrowheads indicate Staphylococcal abscesses. (h) Severity scores of kidneys (n = 10) from the AlPO 4 group and recombinant vaccine immunised groups at 3 days post-infection were shown. Data were presented as scatter plots, and the means ± SEM were shown. Asterisks indicate significant differences between two groups ( * P < 0.05, ** P < 0.01). Cytokine responses and CD4 + T cell polarisation of splenocytes. Next, we evaluated antigen-specific cellular responses. The amounts of cytokines in the supernatants were measured after stimulation with recombinant proteins. Splenocytes from the vaccinated mice (IsdB, ClfA 33-213 , and IC) produced significantly more IFN-γ and IL-17A (Fig. 4a,b) when stimulated with the corresponding recombinant proteins than splenocytes from the AlPO 4 group. Furthermore, the splenocytes from the IC group produced the highest amount (Fig. 4a,b, P < 0.05), which showed that splenocytes from mice immunised with IC could produce more IFN-γ and IL-17A following antigen encounters.  Table 2. IgG titers of specific antibodies for recombinant proteins (IsdB and ClfA 33-213 ). * IgG titers (mean serum titers ± SD) in response to immunization of mice as determined by ELISA (n = 6 animals). # Specific antibody titers raised against proteins were not significantly different when immunizing antigens were administered individually or in combination. a P, anti-IsdB antibodies from the mice immunized with IsdB were compared to those from the mice immunized with IC; b P, anti-ClfA 33-213 antibodies from the mice immunized with ClfA 33-213 were compared with those from the mice immunized with IC. ND means not detected. Then, the splenocytes were harvested for T cell polarisation tests in vitro. After stimulation, the splenocytes in the IC group showed an increased ability to polarise toward the production of IFN-γ and IL-17A compared with the splenocytes in the IsdB and ClfA 33-213 groups (Fig. 4c). Although CD4 + T cells from mice immunised with IC were more likely to produce IL-17A than the IsdB group, this difference did not reach statistical significance (P = 0.1148) (Fig. 4e). The IC group had a higher number of IFN-γ + IL-17A − CD4 + T cells than the IsdB group (P = 0.0365) and ClfA 33-213 group (P < 0.0001, Fig. 4d). These results showed that immunisation with IC increased the potential of splenocytes to induce cytokine responses and polarise toward Th1/Th17 sub-group subsets.

Roles of IL-17A in IC vaccine efficacy.
To evaluate the potential roles of IFN-γ and IL-17A in vaccine efficacy, neutralising antibodies targeting IFN-γ (α IFN-γ ) and/or IL-17A (α IL-17A) were administered to vaccinated mice. Bacterial burdens in blood, spleens, and kidneys were measured 3 days post-infection. The IC-vaccinated group displayed significantly lower bacterial colonisation versus the non-vaccinated controls. Neutralisation of IL-17A alone caused a significant increase in CFUs in the blood (P = 0.0017), spleens (P = 0.0043), and kidneys (P = 0.0018) in the vaccinated animals compared with the untreated vaccinated controls (Fig. 5a-c). Blocking both IFN-γ and IL-17A caused a significant increase in the mean organ bacterial burdens in the vaccinated mice compared with the control group, although the difference was not significant compared with the results of inhibition of IL-17A alone (Fig. 5a-c). In contrast, neutralisation of IFN-γ alone did not influence the bacterial burdens in the vaccinated mice (P > 0.05).
Furthermore, we used IL-17AKO mice to confirm the results from the IL-17A neutralisation tests. In the IL-17AKO-vaccinated group, the bacterial burdens were increased in blood (P = 0.0100), spleens (P = 0.0022), and kidneys (P = 0.0152) compared with those in the WT vaccinated group (Fig. 5d-f). The protective efficacy of the IC vaccine was still detected in the IL-17AKO mice because the bacterial burdens in the blood and kidneys in the IL-17AKO vaccinated group were much lower than those in the IL-17AKO unvaccinated group (Fig. 5d,f). Because IL-17A gene deficiency remarkably reduced the IC immune protective effect, we demonstrated that the strong IL-17A responses induced by IC played an important role in vaccine immune protection.

Immunisation with IC can effectively reduce lung damage in an S. aureus pneumonia model.
Finally, we evaluated the IC-induced immune protective effect in S. aureus pneumonia. In our study, the control group immunised with AlPO 4 alone showed serious symptoms, namely, lethargy, ruffled fur, hypothermic, and significant weight loss (Fig. 6a,b). In contrast, the body temperature and weight in the IC group were controlled within a relatively acceptable range. We calculated the bacterial burdens in the whole lung 1, 3, and 7 days post-infection and found that the bacterial burdens were significantly reduced compared with those in the AlPO 4 Figure 4. Cytokine responses and CD4 + T cell polarization of splenocytes. One week following the last booster, the splenocytes from vaccinated mice (n = 6) were incubated with corresponding antigen proteins (10 μ g/mL) for 5 days. The supernatants were harvested, and the cytokine levels of (a) IFN-γ , (b) IL-17A were determined. The means ± SEM were shown. Empty boxes representative without stimulation, black boxes representative splenocytes stimulate with corresponding proteins. Asterisks indicate significant differences. ( * P < 0.05, ** P < 0.01). (c) After 5 days incubation, the splenocytes were stimulated with PMA/ionomycin and Golgistop for 6 h. Cells were stained for CD3, CD4, IFN-γ , or IL-17, and was analysed by flow cytometry. The responses of Th1 (d) and Th17 (e) of the AlPO 4 group and recombinant vaccine immunised groups were presented as column bar graph and shown as means ± SEM. Results represent three independent experiments ( * P < 0.05, ** P < 0.01, *** P < 0.001). group (Fig. 6c, P 1day = 0.0004, P 3day = 0.0270, and P 7day = 0.0095). The lung exterior also showed less haemorrhaging in the IC group ( Supplementary Fig. S1), and pathological changes (Fig. 6d) and the lung severity score demonstrated that there was less damage compared with the AlPO 4 group (Fig. 6e). Therefore, immunisation with IC significantly decreased lung damage in the S. aureus pneumonia model.

Discussion
IsdB and ClfA are suitable potential candidates for use in the development of an effective S. aureus vaccine 12 because almost all S. aureus strains express these two surface proteins 14,34,35 . Due to the increasing spread of methicillin-resistant S. aureus strains, it is notable that S. aureus generally carry these two genes. However, vaccination with IsdB and ClfA alone did not induce ideal protective effects in a murine sepsis model (Figs 1d and 2).
IsdB consists of two different domains, namely, NEAT1 and NEAT2 36,37 . Although the biological function of these two domains has been extensively explored, the dominant immune responses are still unknown. The NEAT1 domain was sufficient to bind haemoglobin 37 and played a critical role in the heme-iron acquisition of S. aureus. Using epitope prediction, we inferred the existence of B-cell epitope enrichment regions in the NEAT1 domain. Therefore, we chose IsdB 151-277 (NEAT1) as our vaccine candidate. Prior to cloning the extra-cellular region of ClfA 40-559 , we inferred that amino acids 33 to 213 may be suitable for the design of a vaccine using epitope prediction (www.sbc.su.se). The use of an appropriate linker peptide is necessary to retain the activity of the peptide 38 . Therefore, we designed the YVPVDV linker, which maintained the original conformation of IsdB 151-277 and ClfA  . We linked these two immune-dominant areas together to construct the new chimeric vaccine "IC".
The critical role of antibodies in controlling S. aureus infection has been reported in vitro and in vivo [39][40][41][42] . Notably, immunisation with IC was efficacious and induced prominent antibody responses. First, immunisation with IC induced large quantities of total IgG, IgG1, or IgG2a compared with IsdB and ClfA   (Table 1), which showed balanced antibody production. Second, immunisation with IC alone generated similar antibody levels to those achieved by the single components ( Table 2), demonstrating that this chimeric vaccine could induce both IsdB and ClfA antibodies. This finding revealed that multivalent antigens in vaccines may be more effective in combating S. aureus. Nevertheless, the correlation between antibody responses and immune protective effects is still unclear 39,43 . Using the IC group as an example, our data provided evidence that the antibody titres correlated with the protective effects.
Passively administered anti-IsdB or anti-ClfA antibodies have demonstrated efficacy in rodent models 16 . In our study, we demonstrated that polyclonal antibodies to recombinant IC were effective in preventing bacteraemia and that this prevention was superior to that afforded by IsdB or ClfA polyclonal antibodies administered alone at the same dose (Fig. 3b). This result revealed that IC-specific antibodies play an important role in vivo and showed that the protective effect was at least partially antibody mediated. Additionally, the opsonophagocytic properties of antibodies were assessed in vitro. The results showed that IC-specific antibodies could promote S. aureus clearance in the presence of complement, thereby facilitating neutrophil killing (Fig. 3a). The above results revealed that polyclonal antibodies to recombinant IC directly influenced protection. An increasing amount of evidence has revealed that T cells play an important role in vaccine-mediated protective effects against pathogens 44 . Indeed, Th1 or Th17 polarised immune responses are closely related to the regulation and activation of neutrophils and macrophages 45,46 . In S. aureus infection, IL-17 mainly mediates the chemotaxis and activation of neutrophils 47 . IFN-γ is also a critical cytokine and increases the microbicidal activities of macrophages. Previous studies demonstrated that immunisation with IsdB or ClfA 40-559 induced IL-17 production, which was critical for the vaccine protective effect against S. aureus infection 15,23 . In contrast, IsdB N126-P361 , which contained NEAT1 and NEAT2, induced increased amounts of IFN-γ 48 . In this study, splenocytes from mice immunised with IC produced more IFN-γ and IL-17A than those from mice immunised with the single components (Fig. 4a,b). Our findings in this study were consistent with previous reports regarding the use of IsdB and ClfA in vaccine design that showed that IC played a combined role in inducing IFN-γ /IL-17A production. These results support the hypothesis that the Th1/Th17 response induced by IC may be a key cellular immune protective effect that affords protection against S. aureus.
Inhibition of IL-17A alone abrogated the vaccine protective effect, but neutralisation of IFN-γ alone did not influence the bacterial burdens (Fig. 5a-c). These results suggested that although IC could induce both IFN-γ and IL-17A production, the protective effect mainly depended on the Th17 pathway. Our current findings are consistent with recent reports that T cells and the Th17 pathway play critical roles in defence against S. aureus sepsis 9,10,49 . In IC-vaccinated mice, neutralisation of IL-17A damaged the clearance of S. aureus from the organs. However, the vaccinated group retained some protection despite IL-17A or IFN-γ neutralisation compared with the nonvaccinated controls. These results revealed that multiple mechanisms for immune protection were induced by immunisation with IC. For instance, our data demonstrated that IC induced a robust antibody response ( Table 1) that may confer protection in this model. Using IL-17AKO mice, we demonstrated the important role of IL-17A in immune protection (Fig. 5d-f) and showed that IL-17A contributed to vaccine efficacy. Immunisation with IC still induced protection in the IL-17AKO mice that mainly occurred through the functional antibody response. In conclusion, the high protection level induced by immunisation with IC was mainly dependent on antibody responses and the IL-17A immune pathway.
To date, most vaccines have evaluated the protective effect against a few laboratory strains. However, studies should focus on clinical S. aureus isolates because these strains show considerable differences both in the repertoire of surface proteins and sequence variation in their protein binding domains 12 . Immunisation with IC showed broad protection against the WHO2 strain and three clinical S. aureus isolates collected from Chongqing, Guangzhou, and Beijing hospitals (Supplemental Table 1). Overall, these results suggested that immunisation with IC was capable of protecting mice against a lethal challenge with different S. aureus strains.
In S. aureus pneumonia, active immunisation can significantly reduce lung bacterial burdens 1 day post-infection. In the control group, we found that the bacterial burdens continued to decrease over time but were still significantly higher compared with those in the IC group (Fig. 6c). While the severity of pneumonia was controlled in the IC group, the control group worsened over time (Fig. 6c). The most likely explanation for this result is that powerful antibody-mediated bacteria killing in the acute infection phase reduced the bacterial burden in the lung and thereby reduced lung damage. The temperature and weight loss were controlled in the IC group 1 day post-infection, on the contrary, significant temperature and weight loss happened in the AlPO 4 group (Fig. 6a,b). These results demonstrated that immunisation with IC mainly functioned in the acute infection phase by improving bacterial killing and protecting the lung from damage due to pneumonia. Many studies have shown that vaccines offer protection from pneumonia primarily by inducing antibody production [50][51][52] . Additionally, some studies have shown that the products of CD4 + T cells (i.e., IL-17A) are important in acute pneumonia 53 . In our previous section, we demonstrated that IC could induce strong humoral and cellular responses, both of which may play important roles in defending against S. aureus pneumonia.
Recently, Delfani et al. had shown that they combined with ClfA, IsdB and gamma hemolysin B (HlgB) fragments and constructed the multi-subunit fusion vaccine ClfA-IsdB-HlgB 54 . Immunisation with ClfA-IsdB-HlgB induced a protective effect in mice sepsis model, and it also induced strong antibody responses. Although ClfA-IsdB-HlgB and IC both contain the fragments of ClfA and IsdB, the truncated domains were different, and may induce different immune protective effects. In our paper, besides sepsis model, we used pneumonia model to evaluate the protective effect of IC, and immunisation with IC also conferred protection in all the five S. aureus strains. Thus, immunisation with IC shows broad protective effects in different models and different strains. Meanwhile, we evaluated the functional antibodies through OPA and passive immunisation, and detected Th17 immune pathway played a key role in combating against S. aureus.
In our previous research, we designed Hla H35L IsdB 348-465 as an S. aureus vaccine candidate based on the reverse vaccinology 30 . In the current study, IC chimeric vaccine was different from Hla H35L IsdB 348-465 in the properties of components, because IsdB and ClfA are both S. aureus surface proteins. Meanwhile, the IsdB domains in the IC and Hla H35L IsdB 348-465 were different: IC contained the NEAT1 domain of IsdB, whereas Hla H35L IsdB 348-465 contained the NEAT2 domain. Continuous B cell epitope prediction (http://tools.iedb.org/bcell/) indicated that NEAT1 domain may have advantages in antibody responses. Later, we combined IsdB 151-277 with ClfA 33-213 , and generated the novel chimeric vaccine IC. Strong antibody responses are the advantages of surface proteins, and interestingly, immunisation with IC also induced strong IL-17A responses. In conclusion, IC can be regarded as a novel S. aureus vaccine candidate. When combined with other antigens, such as Hla H35L IsdB 348-465 , IC may play a more important role in the fight against S. aureus infections.