Intranasal vaccination with a recombinant protein CTA1-DD-RBF protects mice against hRSV infection

Human respiratory syncytial virus (hRSV) infection is a major pediatric health concern worldwide. Despite more than half a century of efforts, there is still no commercially available vaccine. In this study, we constructed and purified the recombinant protein CTA1-DD-RBF composed of a CTA1-DD mucosal adjuvant and prefusion F protein (RBF) using Escherichia coli BL21 cells. We studied the immunogenicity of CTA1-DD-RBF in mice. Intranasal immunization with CTA1-DD-RBF stimulated hRSV F-specific IgG1, IgG2a, sIgA, and neutralizing antibodies as well as T cell immunity without inducing lung immunopathology upon hRSV challenge. Moreover, the protective immunity of CTA1-DD-RBF was superior to that of the RBF protein, as confirmed by the assessment of serum-neutralizing activity and viral clearance after challenge. Compared to formalin-inactivated hRSV (FI-RSV), intranasal immunization with CTA1-DD-RBF induced a Th1 immune response. In summary, intranasal immunization with CTA1-DD-RBF is safe and effective in mice. Therefore, CTA1-DD-RBF represents a potential mucosal vaccine candidate for the prevention of human infection with hRSV.

The successful development of a mucosal vaccine depends on the use of safe and effective adjuvants. The nontoxic mucosal adjuvant CTA1-DD comprises the CTA1 subunit of cholera toxin (CT) and two immunoglobulin (Ig) binding domains (DD) of staphylococcal protein A (SpA). CTA1 is an adenosine diphosphate (ADP)-ribosyltransferase 35 . SpA contains five highly homologous extracellular domains in tandem, designated domains E, D, A, B, and C. The domain D consist of three a-helical structures 36 . CTA1-DD has been shown to bind to all immunoglobulins through the DD domain, including Ig on the B cell surface 37 . To date, the CTA1-DD mucosal adjuvant has been demonstrated to be effective against many pathogens, including Ebola virus, influenza virus, and HIV [38][39][40] . Combining the neutralizing peptides and CTA1-DD into one protein confers strong protective immunity against Helicobacter pylori and influenza virus in mice 41,42 . A substantial amount of preclinical data have confirmed the efficiency and safety of CTA1-DD, which is one of the most promising mucosal adjuvants to date 43 . Although previous research suggests that the F-ctxA2B fusion protein consists of residues 412-524 of the hRSV F protein and that the ctxA2B subunit of CT can provide partial protection 44 , no research has been performed on the CTA1-DD mucosal adjuvant used in hRSV vaccines until now.
In the present study, we constructed and purified the CTA1-DD-RBF protein using E. coli BL21 and tested its immunogenicity and protective efficacy in mice. Compared with RBF, CTA1-DD-RBF induced higher levels of neutralizing serum antibodies and inhibited virus replication in the lungs of immunized mice. Intranasal vaccination induced a Th1-biased immune response, while formalin-inactivated RSV (FI-RSV) induced Th2biased immune responses. This study demonstrates that the uncleaved CTA1-DD-RBF protein expressed in E. coli elicits protective immunity in mice without enhancing the disease.

Results
Production, purification and conformation of the CTA1-DD-RBF protein. CTA1-DD-RBF was purified by HisTrap FF columns, and the tags were removed by overnight digestion with thrombin at 26 °C. The protein (without a 6 × His-tag) was then further purified using a Superdex 200 column. To verify the efficacy of thrombin cleavage, CTA1-DD-RBF was further purified using a HisTrap FF column after digestion by thrombin. SDS-PAGE analysis indicated that the purified CTA1-DD-RBF (the protein that passed through the HisTrap FF column) was ≥ 90% pure, and a 90 kDa band was identified (Fig. 1B). Nearly 60% of the CTA1-DD-RBF proteins passed through the HisTrap FF column after digestion by thrombin, indicating that the efficacy of thrombin cleavage was high. The thrombin site is native to the pET28a plasmid, and thrombin cleaves and removes 17 amino acids, including the 6 × His-tag. However, 17 amino acids remained on the CTA1-DD-RBF protein. After chromatographic purification and ultrafiltration, the endotoxin concentration was 4.22 EU per mg CTA1-DD-RBF.
The binding kinetics of CTA1-DD-RBF, RBF and Post-RBF with 5C4 were identified by BLI on the Octet RED96e instrument. The KD (M) for CTA1-DD-RBF and the 5C4 antibody was 4.78E-11 ( Supplementary  Fig. 1), which was lower than that for RBF and the 5C4 antibody (1.26E-09). The KD (M) for Post-RBF and 5C4 was 1.27E-07 ( Supplementary Fig. 1). The OD450 measured in sandwich ELISA shown that CTA1-DD-RBF and 5C4 antibody have better binding ability, compared with RBF or post-RBF ( Supplementary Fig. 2). These results suggested that CTA1-DD-RBF could bind the 5C4 antibody with a higher affinity than RBF and Post-RBF, which indicated that the CTA1-DD-RBF protein exists in the prefusion conformation. Negative stain electron microscopy of the CTA1-DD-RBF protein showed that the protein was highly homogenous (Fig. 1C).

CTA1-DD-RBF vaccination induced specific humoral immune responses and neutralizing antibodies.
We evaluated the hRSV-specific humoral immune response by indirect ELISA on day 14 (14 days after the initial immunization) and day 35 (14 days after the immunization boost). Mice immunized with CTA1-DD-RBF exhibited a stronger humoral immune response than mice immunized with the RBF protein alone ( Fig. 2A). We also found significant increases in antibody titers at day 35 compared with those at day 14.
The plaque reduction neutralization test showed that the neutralization activity of serum from FI-RSV-and RBF-immunized mice was lower than that of serum from CTA1-DD-RBF-immunized mice (Fig. 2B).

CTA1-DD-RBF reduced hRSV replication and induced mucosal immune responses. All mice
were challenged with 1 × 10 5 PFU of hRSV A Long 14 days after the booster immunization, and the hRSV titers in the lungs were detected at 4 days after hRSV infection. PBS-and FI-RSV-vaccinated mice had substantial amounts of hRSV in their lungs, with values reaching nearly 10 5 PFU/g. In contrast, CTA1-DD-RBF vaccination reduced hRSV replication and dramatically decreased the viral titer in the lungs. As shown in Fig. 3A, compared with FI-RSV immunization, intranasal immunization significantly reduced the viral load after hRSV infection (p < 0.05).
At 14 days after the second immunization, sIgA in the BALF was detected by ELISA. In the CTA1-DD-RBF group, the BALF sIgA antibody levels were high and were significantly higher than those in the RBF group ( Fig. 3B) (p < 0.005). Taken together, these results indicated that CTA1-DD-RBF promoted the production of sIgA and that intranasal vaccination induced specific humoral, cellular and mucosal immune responses.
To evaluate the type of cellular immune response (Th1 or Th2), we investigated the numbers of IL-4-and IFN-γ-secreting cells among spleen cells stimulated with RBF. The number of IL-4-secreting cells in the FI-RSV group was higher than those in the CTA1-DD-RBF and RBF groups (p < 0.005) (Fig. 5B). Furthermore, high numbers of IFN-γ-secreting cells were induced in the CTA1-DD-RBF group (Fig. 5A).
To further assess the immune response, we investigated the titers of IgG1 and IgG2a in serum collected 14 days after the immunization boost. Compared to the FI-RSV group, the CTA1-DD-RBF and RBF groups exhibited a relatively balanced humoral immune response (Fig. 5C). In conclusion, intranasal vaccination induced a Th1-biased immune response, in contrast to the FI-RSV vaccine. Fig. 6, after hRSV challenge, mice immunized with FI-RSV had severe alveolitis, interstitial pneumonitis and bronchiolitis (p < 0.001). In contrast, we observed only slight lung infiltration in CTA1-DD-RBF-and RBF-immunized mice. These results demonstrate that intranasal vaccination reduces lung injury after hRSV infection and does not enhance disease in mice.

Discussion
Because hRSV infects via the respiratory tract, vaccines administered nasally may offer more protective benefits than those delivered intramuscularly. It is important to select the right adjuvant to tailor the required immune response. CTA1-DD has been shown to enhance immunogenicity when mixed or combined with relatively poor vacinnes delivered mucosally 41,42,45 . While CTA1-DD was previously shown to be nontoxic and to not www.nature.com/scientificreports/ cause inflammation at the application site 46 , it has not been studied in human clinical trials. Proving the safety of mucosal adjuvants is difficult. For example, a mucosal adjuvant derived from E. coli (LTK63) is considered to be safe in macaques, Guinea pigs and mice 47 , but the intranasal administration of LTK63 has been shown to induce transient facial nerve paralysis (Bell's palsy) in humans 48 . Therefore, more thorough toxicological and immunological analyses may be required before CTA1-DD can be studied in clinical trials. Numerous vaccine candidates are currently under clinical investigation, including live attenuated vaccines, viral vector vaccines, and subunit vaccines 31,[49][50][51] . Nonreplicating vaccines (subunit or inactivated vaccines) should be avoided in hRSV-naïve young children 52 . Live attenuated vaccines and viral vector vaccines tend to    www.nature.com/scientificreports/ be administered intranasally to children 8 . Enhanced pulmonary histopathology was observed in cotton rats immunized intramuscularly with purified F protein 53 . However, the results might depend on the immune dose 54 and conformation of the protein 55 . Mice immunized intramuscularly with CTA1-DD-RBF had slightly severe alveolitis, interstitial pneumonitis and bronchiolitis. Multiple in vitro expression systems, including Escherichia coli, insect cells (Sf9) and mammalian cells, have been investigated for production of the F protein or peptide as an immunogenic hRSV subunit vaccine 11,44,56 . This study is the first to report that the recombinant CTA1-DD-RBF protein can be successfully expressed in E. coli. The E. coli protein expression system is one of the simplest and most cost-effective systems and is suitable for the large-scale production of recombinant proteins 57 . BLI analyses herein showed that the KD (M) for CTA1-DD-RBF and the 5C4 antibody (specific to pre-F) was lower than that for RBF and 5C4 (1.26E-09). These results suggested that CTA1-DD-RBF could bind the 5C4 antibody with higher affinity than the Post-RBF protein.
Cellular immune strength, mucosal IgA and serum neutralizing antibody titers were negatively correlated with disease severity 58 . Intranasal vaccination of mice with CTA1-DD-RBF induced hRSV-specific humoral, cellular and mucosal immune responses. The abilities of CTA1-DD-RBF to induce neutralizing serum antibodies and inhibit virus replication in the lungs were superior to those of RBF. Due to the major role of T cell responses in hRSV clearance 59 , we investigated the number of IFN-γ-secreting cells among spleen cells stimulated with F protein 60 , and intranasal immunization with CTA1-DD-RBF significantly increased the number of IFN-γsecreting cells.
The ability of vaccines administered via the noninvasive mucosal route to induce mucosal antibody responses in addition to systemic antibody responses has been explored 61 . hRSV infects people through mucosal sites, and the mucosal antibody sIgA provides protection at sites of viral entry 62 . A vaccine that elicits a long-lasting hRSV-specific sIgA response is more protective than one that produces only systemic antibodies 63 . Therefore, we evaluated mucosal sIgA in BALF after the nasal administration of the CTA1-DD-RBF protein. At 14 days after the second immunization, the high antigen-specific sIgA in the CTA1-DD-RBF group was significantly higher than that in the RBF group. These results indicated that intranasal immunization with the mucosal adjuvant CTA1-DD could induce better mucosal sIgA antibodies than that with RBF.
An FI-RSV vaccine tested in the 1960s was shown to lead to ERD upon viral challenge 64 . FI-RSV immunization induces a Th2-biased immune response that results in pulmonary eosinophilia following hRSV challenge in multiple animal models 65 . Therefore, hRSV vaccines should be designed to reduce detrimental Th2-biased immune responses and to induce a Th1-type immune response. The serum titers of IgG1 and IgG2a at 14 days post boost immunization and the numbers of IFN-γ-and IL-4-secreting cells among spleen cells stimulated with RBF demonstrated that CTA1-DD-RBF administered mucosally induced a Th1-biased immune response. To further assess the immune response, we investigated the contents of Th1-type cytokines (IL-2, IL-12p70, IFN-γ) www.nature.com/scientificreports/ and Th2-type cytokines (IL-10, IL-4, IL-5) in BALF. Compared with the FI-RSV group, CTA1-DD-RBF and RBF administered intranasally elicited a Th1 cytokine response in BALF.
Previous studies have shown that immunization with 5 μg 66 or 3.2 μg 67 of FI-RSV prevented virus replication in the lungs at day 4 post challenge. However, in this study, the hRSV titers in the lungs of FI-RSV-immunized mice were not reduced compared with those in the lungs of PBS group mice, and the protective effect of FI-RSV may depend on the vaccine dose 54 . The FI-RSV vaccine is considered to be effective when the hRSV titer in the cell culture supernatant is 2 × 10 6 PFU/ml 68 or 4.6 × 10 5 PFU/ml 69 prior to formalin inactivation. However, the hRSV titer was only 6.5 × 10 4 PFU/ml prior to formalin inactivation in this study.
These studies have several limitations. First, assessment of cytokines in the BALF is not a great measure of local cellular immune responses in the lung. These limitations will be addressed by our future animal experiments, by flow cytometry after antigen stimulation and intracellular cytokine staining. Second, some measure of memory B and T cells, including local tissue-resident memory T cells in the lung would add significantly to the study. Third, Using i.n. immunization, the relative distribution between the respiratory tract and the gastrointestinal tract is heavily influenced by delivery volume and level of anesthesia 70 . 100 µl of the vaccine is a very high volume and is likely to result in pulmonary and maybe gastric delivery of the antigen in mice, especially under isoflurane anaesthesia. We should reduce protein immune volume in future animal experiments. Fourth, the viral load in the lungs is assessed at one time point post challenge. This would be more convincing if assessed at several time points.
In conclusion, we established a highly efficient prokaryotic expression system for CTA1-DD-RBF protein production that has great potential for expressing hRSV vaccines. Compared with RBF, CTA1-DD-RBF induced higher levels of neutralizing serum antibodies in immunized mice. Compared to FI-RSV mice, mice immunized with CTA1-DD-RBF had no signs of ERD upon hRSV challenge. This study provides insight into developing CTA1-DD-RBF as an effective and safe hRSV mucosal vaccine.

Materials and methods
Ethics statement. Female BALB/c mice (6-8 weeks) were purchased from SPF Biotechnology (Beijing).
All the protocols were performed in accordance with the guidelines of the Animal Experimental Ethics Committee of the National Institute for Viral Disease Control and Prevention (no. 20190610024). This study was carried out in compliance with the ARRIVE guidelines. The animals were housed under pathogen-free and temperature-controlled conditions at the Animal Center of the Chinese Center for Disease Control and Prevention. The mice were observed every day. The animals were lightly anesthetized with isoflurane for immunizations and blood draws and euthanized with carbon dioxide for terminal organ harvests. All methods were carried out in accordance with relevant guidelines and regulations.
Construction of expression plasmids. The RBF protein contains a sequence encoding hRSV F residues 26-105 and 146-513 (replacing two furin cleavage sites and P27 with a variable linker, GSGSG) and a T4 fibritin trimerization motif (foldon) at the C-terminus. To retain prefusion-specific neutralizing epitopes and increase the antigenic stability to heat inactivation, DS-Cav1 mutations (S155C and S290C, S190F, V207L) and a disulfide bond (A149C, Y458C) were incorporated into the sequence 71 . We expressed the pre-and postfusion F proteins, called RBF and Post-RBF, respectively, using Escherichia coli BL21 72  The binding kinetics of CTA1-DD-RBF, RBF and Post-RBF to the 5C4 antibody (specific to pre-F) were detected using biolayer interferometry (BLI) with the Octet RED96e instrument. The specific method is provided in the methods section of the Supplementary Materials. The RBF and Post-RBF proteins were kept in our laboratory and are described in the Supplementary Material.
Immunization and challenge. The mice were randomly distributed into experimental groups (8 mice/ group) and were vaccinated intranasally or intramuscularly twice with a 3-week interval (Table 1). Briefly, the two groups of mice were vaccinated intranasally with 10 μg of CTA1-DD-RBF or RBF without adjuvant. One group of mice was administered 100 μl of phosphate buffered saline (PBS) twice as a placebo control. To appraise the safety of the mucosal vaccine, we immunized 8 mice with 100 μl of the FI-RSV vaccine via the intramuscular route. FI-RSV was constructed as described by Kim et al. 64  www.nature.com/scientificreports/ was diluted in 5 mL of MEM and subsequently mixed with alum adjuvant (4 mg/mL) at 10:1 (v/v). The alumadsorbed FI-RSV was collected by centrifugation at 1,000 rpm for 10 min, suspended in 1 mL of serum-free MEM, and stored at 4 °C. The mice were infected intranasally with hRSV A Long (1 × 10 5 PFU) in 100 μl under isoflurane on day 35. Serum samples were collected by centrifugation and kept at − 80 °C until use. For the collection of bronchoalveolar lavage fluid (BALF) and splenocytes, four mice from each group were sacrificed and blend thoroughly on day 35, while the other mice were sacrificed 4 days after the hRSV challenge.
Viral detection in the lungs and determination of antibody titers. The right lungs were removed from the mice aseptically at 4 days after the hRSV challenge. The titers of hRSV in the lungs were determined by the plaque test, and the serum neutralizing antibodies were detected by the plaque reduction neutralization test described by Garg et al. 73 . hRSV-specific antibodies (IgG, IgG2a and IgG1) were determined by ELISA using RBF (100 ng/well) as the coating antigen. Endpoint titers were calculated as the highest serum dilution that gave an optical density exceeding 2.1 times the background.
Detection of cytokines and sIgA in bronchoalveolar lavage fluid. BALF was collected by gentle injection of 1 ml of PBS into the lungs via the trachea with a syringe connected to the trachea 14 days after the second immunization. The hRSV-specific sIgA antibody titer in BALF was determined by ELISA using RBF protein (100 ng/well) as the coating antigen. BALF was added to 96-well plates (100 μl/well) for 1 h, and HRPconjugated goat anti-mouse sIgA was then diluted 5000 times and incubated with the sample for 1 h; the OD450 was measured using a microplate reader. The BALF levels of Th1-type cytokines (IL-2, IL-12p70, IFN-γ) and Th2-type cytokines (IL-10, IL-4, IL-5) were determined using ELISA kits (Dakewei, Beijing, China) according to the manufacturer's instructions.
ELISPOT assays. ELISPOT assays were used to evaluate the type of cellular immune response (Th1 or Th2).
Splenocytes were collected 14 days after the second immunization according to the manufacturer's instructions (Dakewe, Beijing, China). To evaluate the cellular immune responses, splenocytes were stimulated with the RBF protein (5 µg/mL).
Lung histopathology. The left lungs were fixed in 10% neutral buffered formalin at 4 days after the hRSV challenge and stained with hematoxylin and eosin (H&E) for histological evaluation under a Nikon Eclipse light microscope by a board-certified pathologist. Slides stained with H&E were evaluated for the following three parameters described by our previous study: bronchiolitis, alveolitis and interstitial pneumonitis 72 . The slides were evaluated in a blinded fashion using a 1-4 scoring system as follows: 0 (normal), 1 (mild inflammation), 2 (moderate inflammation), 3 (marked inflammation) and 4 (severe inflammation).
Statistical analysis. All data were analyzed using GraphPad Prism version 5. Differences among all groups were evaluated by one-way ANOVA, followed by a Newman-Keuls posttest. Differences were considered significant if P < 0.05.

Data availability
All data included in this study is available upon request by contact with the corresponding author.