Background
Streptococcus pneumoniae is a gram-positive encapsulated coccus that generally colonizes the upper respiratory tract in humans without symptoms [
1]. However, it may cause community-acquired pneumonia and invasive infections owing to mucosal translocation, such as bacterial meningitis, bacteremia, and otitis media [
2,
3]. Therefore,
S. pneumoniae remains the most common acute pneumonia-causing pathogen in infants and the elderly worldwide [
4,
5]. Not only is it a common cause of primary bacterial pneumonia,
S. pneumoniae also frequently causes secondary bacterial pneumonia following influenza virus infection, thus becoming the main cause of high mortality in adults [
6,
7].
Considering
S. pneumoniae’s impact on morbidity and mortality, healthcare providers promote vaccination to prevent pneumococcal infection. PPSV23 and conjugate vaccines PCV10 and PCV13 are currently available for use as vaccines [
8,
9]. The unconjugated polysaccharide PPSV23 vaccine provides higher coverage of serotypes against pneumococci than the other vaccines, but it cannot be administered to infants owing to their underdeveloped immune system. The conjugated vaccine has reduced the occurrence of invasive pneumococcal diseases since PCV7 was introduced in 2000 [
10,
11]; however, cases of pneumococcal infections increased because of non-vaccine serotypes [
12,
13]. In fact, invasive diseases attributable to serotypes that are included in the current polysaccharide vaccines may still threaten the protective effect [
14]. Notably, capsular polysaccharide vaccines are less effective against acute otitis media and non-bacteremic pneumococcal pneumonia in adult populations [
15,
16], possibly due to capsule shedding in response to the epithelium during mucosal infection [
17]. Therefore, efforts are being made to develop vaccines containing highly conserved, immunogenic protein antigens, which are selected by immunoscreening with the patients’ sera [
18] or by reverse vaccinology [
19], among other approaches.
PrtA, a cell wall-associated protein, was screened from convalescent patient serum after
S. pneumoniae infection [
20] and was identified as a serine protease. The amino terminal of PrtA, containing catalytic domains, was highly conserved among 78 clinical pneumococcal isolates displaying 22 different serotypes, including the D39 strain [
21]. The deletion of the
prtA in
S. pneumoniae D39 reduced mortality at 36 h after intraperitoneal infection [
21] and alleviated lung inflammation at 48 h after intranasal infection [
22]. Furthermore,
prtA expression in
S. pneumoniae could be induced during epithelial cell contact, pneumococcal bacteremia, and meningitis in infected mice [
23]. Although the immunogenicity, conservation, and virulence of PrtA have been reported, the efficacy of PrtA as a vaccine against pneumococcal infection has not been studied.
The Th17 response is considered effective against
S. pneumoniae infection. The intratracheal administration of recombinant IL-17A can stimulate the local release of MIP-2 and IL-1β, leading to the recruitment of polymorphonuclear leukocytes to the lungs [
24]. Th17 response also triggers the mucosal epithelium to generate anti-microbial peptides, which facilitate the elimination of mucosal pathogens [
25]. Mice lacking the IL-17A receptor, but not IFN-γ or IL-4 receptors, demonstrated decreased protection against
S. pneumoniae [
26]. Thus, screening for Th17-based antigens is a feasible approach to select vaccine candidates that could reduce
S. pneumoniae colonization [
27]. Adjuvants provide an alternative approach to support ideal vaccine candidates and develop appropriate cellular immunity.
Curdlan is a linear and nonionic β-1,3-glucan isolated from the bacterium
Alcaligenes faecalis. Due to its non-toxicity [
28,
29] and heat-based gel-forming capabilities [
30], curdlan has been approved as a food additive by the U.S. Food and Drug Administration. Curdlan is an agonist of Dectin-1, activating dendritic cells to induce Th17 differentiation [
31], along with being a strong inducer of Th17 response [
32]. Therefore, curdlan has been successfully adopted as a vaccine adjuvant against
Pseudomonas aeruginosa pulmonary infection [
33], and lethal Candida infection [
34], among other infections [
35]. In the present study, we investigated the efficacy of PrtA immunization combined with curdlan to prevent
S. pneumoniae infection. We used BALB/c mice to deliver the vaccine via the nasal route to trigger both local mucosal and systemic immune responses, we also checked the vaccines’ capability to reduce the severity of pneumonia as well as the incidence of invasion.
Methods
Escherichia coli strains DH5α and BL21 (DE3) were used for plasmid cloning and protein expression, respectively. The strains harboring plasmids were grown in Luria–Bertani medium supplemented with kanamycin (50 μg/ml). S. pneumoniae D39 (NCTC7466) was purchased from the National Collection of Type Cultures (London, UK), and S. pneumoniae TIGR4 (ATCC BAA-334) was obtained from American Type Culture Collection (Manassas, VA, USA). S. pneumoniae strains were cultured at 37 °C in 5% CO2 in Todd–Hewitt broth supplemented with 0.5% yeast extract (THY) until mid-log phase for the bacterial challenge, and on sheep-blood agar plates for bacterial load examination after infection.
Construction of expression vector
The gene encoding a PrtA fragment (amino acids 144–1041) was cloned into plasmid pET29a using the primer pair F1-AATCGAGCTCCTATCCAATC and R1-TGAGCCTCGAGAGGATTTCC to construct pET-PrtA105-His. The underlined primer sequences were modified to obtain appropriate restriction sites. To generate recombinant PrtA with dual tags, another primer pair, F2-AGGGTACCGTATTCATGTCC and R2-GCAGATCGTCAGTCAGTCAC, was used to clone DNA encoding glutathione S transferase (GST) from pGEX-4 T-1, which was inserted into pET-PrtA105-His to construct pET-GST-PrtA105-His.
Purification of recombinant proteins
E. coli BL21 (DE3) was transformed with an expression vector. Protein expression was induced in the exponential phase [optical density (OD) at 600 nm = 1] using 1 mM isopropyl β-D-1-thiogalactopyranoside (IPTG). At 3 h after induction, the bacteria were harvested and resuspended in lysis buffer [20 mM Tris (pH 8.0), 5 mM imidazole, 500 mM NaCl, 10% glycerol, 1% Triton X-100, 1 mM phenylmethylsulfonyl fluoride, 1× protease inhibitor cocktail (Roche)], followed by sonication on ice. The recombinant protein PrtA105-His (PrtA1) was purified using Ni-NTA resin (GE Healthcare) according to the manufacturer’s instructions. The other recombinant protein, GST-PrtA105-His (PrtA2), was further purified with glutathione-Sepharose® 4B (Millipore). Lipopolysaccharide (LPS) contamination was reduced by using 0.1% Triton X-114 in the washing step [
36]. The residual LPS was less than 1 EU/μg protein, as determined by a ToxinSensor™ Gel Clot Endotoxin Assay Kit (GenScript).
Mice
BALB/c mice used in the immunization and challenge experiments were purchased from the National Laboratory Animal Center, NARLabs, Taiwan.
Mice immunization
Male mice (3–4 weeks old) were treated intranasally with 8 μg recombinant PrtA1 combined with 0.2 mg curdlan (Sigma) in 20 μl phosphate-buffered saline (PBS) once a week for 3 weeks under anesthesia (intramuscular injection of 40 μl PBS containing Zoletil 50 and 1% Rompun®). One week after the last immunization, mice were restrained to collect saliva before sacrificing them to collect nasal washes, sera, bronchoalveolar lavage fluid (BALF), and spleens.
For PPSV23 vaccination, Pneumovax®23 (Merck & Co., Inc.) vaccines were 10− 1 diluted in saline, and 100 μl was injected intraperitoneally into five-week-old mice. Booster immunizations were administered after 2 weeks.
Identification of CD4+ T cell subsets
CD4+ T cell subsets (Th1 [IFN-γ-producing CD4+ T cells], Th2 [IL-4-producing CD4+ T cells], and Th17 [IL-17A-producing CD4+ T cells]) were identified by intracellular cytokine staining. To estimate the total effector T cell ratio, splenocytes (2 × 106 cell/ml) were seeded in 24-well plates and stimulated with phorbol 12-myristate13-acetate (PMA) (20 ng/ml) and ionomycin (1 μM) for 6 h while incubating with 4 μM monensin. The antibodies used were FITC-anti-IFN-γ (XMG1.2), PE-anti-IL-4 (11B11), APC-anti-CD4 (RM4–5) (eBioscience), and PE-anti-IL-17A (TC11-18H10) (BD Biosciences). Intracellular staining was performed after treatment with permeabilization buffer (eBioscience) containing 0.1% saponin according to the instructions provided by eBioscience, and the effector T cell ratio was analyzed by flow cytometry using a FACSCalibur™ (BD Biosciences).
PrtA-specific cytokine responses
Splenocytes were prepared and seeded at 4 × 105/well using ten wells of a 96-well plate. Recombinant PrtA2 was added at 10 μg/ml and incubated for 48 h. Then, the supernatants were pooled and stored at − 80 °C until use. The cytokine levels were measured using a DuoSet® ELISA (enzyme-linked immunosorbent assay) Development System (R&D Systems).
PrtA-specific antibody titers
PrtA-specific isotype antibody titers were measured in the sera, saliva, nasal washes, and BALF, as described previously [
37,
38]. In brief, a 96-well plate was coated with 100 ng/50 μl PrtA2 in PBS and incubated overnight at 4 °C. After blocking with a 50 μl blocking buffer [0.5% bovine serum albumin (BSA), 0.05% Tween-20, and 1 mM EDTA in PBS], 50 μl of twofold serially diluted samples were added and incubated at room temperature for 2 h. Following three washes, horseradish peroxidase-labeled goat anti-mouse IgG or rat anti-mouse IgA (Southern Biotech) was added, and the plate was incubated at room temperature for 2 h. The color was developed using tetramethylbenzidine (Science Products, Inc.). The endpoint titer was expressed as the reciprocal log
2 of the last dilution that gave an OD of 0.1 or 0.2 at 450 nm for IgG and IgA, respectively.
To analyze the titers of PrtA-specific IgG subclasses, modified mouse Ig isotyping ELISA was used. Slightly, diluted rat anti-mouse Ig isotype antibodies (1 × 10− 3; IgG1, IgG2a, IgG2b, and IgG3) (Affymetrix, eBioscience) were used to detect the levels of PrtA-bound IgG subclasses in 96-well plates as described earlier. HRP-labeled donkey anti-rat Ig antibodies (Jackson ImmunoResearch) were pre-adsorbed with 2% nonimmune mouse sera to minimize the cross-reaction.
Detection of PrtA on pneumococcal surfaces using vaccinated mouse sera
Bacteria cultured from stock for less than 13 h were diluted 1:50 in THY broth and cultured until the mid-log phase. Next, 1 ml bacterial broth was centrifuged and resuspended in filtered PBS containing 1% BSA. PrtA was detected using vaccinated mouse sera (1:20 dilution) by incubating at 37 °C for 20 min. The mouse sera collected from adjuvant-treated littermates were used as controls. The pneumococcal surface-bound Ig was detected using FITC-anti-mouse IgG antibody (eBioscience) and was analyzed by flow cytometry using a FACSCalibur™. Anti-cell wall polysaccharide (CWPS) antisera (SSI Diagnostica) were diluted 1:100 to interact with D39 cells as another control. Surface-bound Ig was identified using FITC-anti-rabbit IgG antibody (BD).
S. pneumoniae D39 cells were labeled with 0.5 mg/mL FITC, as described previously [
39,
40], and were resuspended in PBS. The whole-cell immunofluorescence assay was performed according to the method published by Jose et al. [
41] with slightly modification. The cells dried on the round coverslips were first fixed with 4% paraformaldehyde for 20 min and then washed by PBS for three times. After blocking with 3% BSA in PBS, FITC-labeled D39 cells were incubated with 1/150 diluted antisera at room temperature for 1 h. The pneumococcal surface-bound Ig was identified using DyLight™ 649-conjugated anti-mouse IgG antibody (1:150) (BioLegend) or Cy5-anti-rabbit IgG antibody (Abcam) (1:150), and then was visualized using immunofluorescence microscopy with DeltaVision™ Elite.
Opsonophagocytosis assay
The opsonophagocytosis assay was performed according to the method published by Martinez et al. [
42], by using murine RAW264.7 macrophages. An aliquot of 20 μl of bacterial suspension containing 8 × 10
6 CFU FITC-labeled D39 cells was mixed in a 96-well round bottom plate with 1:1 serially diluted antiserum from the vaccinated mice. The plate was incubated at 37 °C for 30 min with shaking at 200 rpm. Opsonized bacterial cells were then co-cultivated with RAW264.7 macrophages in 24-well plates, that were activated with 100 nM PMA for 3 days. The opsonophagocytosis was conducted at 37 °C for 30 min with shaking at 200 rpm. The bacterial cells were then washed twice with cold PBS, and RAW264.7 macrophages were detached with 2% EDTA in PBS. Extracellular FITC was quenched with 20 μg/ml trypan blue followed by flow cytometry analysis by using FACSCalibur™.
Bacterial challenge of immunized mice
Bacteria cultured overnight were diluted 1:50 in THY broth and cultured until the mid-log phase. Following washing with PBS, and the OD at 580 nm was adjusted to 3.0 before aliquots were stored at − 80 °C. The bacterial counts were estimated after plating. Mice with complete vaccinations were challenged with S. pneumoniae D39 on day 14 after the last boost. The mice were injected with 2.5 × 104 CFU of S. pneumoniae D39 in 100 μl PBS via the retro-orbital venous sinus or anesthetized and infected with 1–3 × 107 CFU of D39 in 40 μl PBS intranasally. Bacterial counts in the blood were monitored at 1–3 days after infection. Blood was collected by submandibular bleeding using a lancet. At 1–3 days after infection, the mice were sacrificed to remove the lungs, which were homogenized to examine bacterial load during acute infection. The humane endpoint was applied to the survival experiment to reduce distress to the animals. We monitored the health of the mice every 12 h for 6 days after the bacterial challenge. If the mice were found to experience labored breathing, lethargy, or inability to ambulate, then they were killed instead of letting them suffer and progress to the experimental endpoint.
Statistics
Significant differences between groups were analyzed using Student’s t-test or Mann-Whitney test. The value indicated in the figures represents mean and standard deviation (SD). The survival rate was investigated using the Kaplan–Meier estimation, and the significance was evaluated by the log-rank test.
Discussion
Overall results indicate that on one hand PrtA/curdlan immunization activated antigen-specific antibody and IL-17A response in the mice, but on the other, it failed to protect against pneumococcal pneumonia and invasive infection. PrtA mediates human apolactoferrin cleavage to yield an N-terminal lactoferricin-like peptide that is more efficient than apolactoferrin as a bactericide against
S. pneumoniae [
44]. It suggested that human innate immunity provided defense against pneumococci with assistance from PrtA and considered PrtA inadequate as a vaccine candidate. It is unknown if mouse apolactoferrin is the PrtA substrate, but a
prtA mutation reduced the virulence of
S. pneumoniae in a murine pneumonia model [
21,
22,
45], thereby suggesting that it plays a pathogenic role.
PrtA was selected using convalescent-phase serum [
20] and was considered an immunogen. Zysk et al. could not confirm the immunoreactivity of PrtA based on 45 serum samples from patients with invasive pneumococcal diseases [
46]. However, they used part of the
prtA gene (nucleotides 330–1041) to express recombinant PrtA as an antigen and that it was highly hydrophobic indicated by TopPred 2 [
47]; thus it might not have been an ideal T or B cell receptor ligand in all patients. Another study also observed no significant difference in the anti-PrtA IgG titer in sera from acute and convalescent patients with pneumococcal bacteremia [
48]. They further indicated that anti-PrtA IgG titers were higher among healthy children with pneumococcal carriage [
49], thereby suggesting that higher antibody titers are insufficient against pneumococcal colonization. In the present study, immunization with recombinant PrtA generated anti-PrtA IgG that bound to the surface of D39 pneumococci but failed to mediate protection against pneumococcal pneumonia and blood stream invasion. It is well known that prior exposure to whole
S. pneumoniae provides immunity against pneumococcal colonization, which depends on CD4
+ T cells and IL-17A, a neutrophil-activating cytokine [
26,
50]. This suggests that the IL-17A response aids vaccine efficacy. Curdlan is an agonist of dectin-1, which activates Syk-CARD9 signaling in dendritic cells to induce Th17 differentiation [
32]. As mentioned earlier, we saw that PrtA-specific IL-17A response was induced in the mice with curdlan as an adjuvant, but it did not protect against pneumococcal pneumonia. Curdlan has been previously shown to facilitate a conserved protein vaccine, PopB, in preventing
P. aeruginosa-induced pneumonia [
33]. With the potency of Th17 stimulation, it is noteworthy to mention that PopB/curdlan could induce a 30-fold higher antigen-specific IL-17A response than PrtA/curdlan, as seen in the antigen-recall T cell activation assay. The distinct IL-17A responses induced by the vaccines used in this study, combined with curdlan, might be attributable to the innate properties of the immunogens along with the potency of being conjugated with curdlan. Curdlan retains antigen in the polymer microspheres via protein-polysaccharide interaction and is extensively used as an adjuvant. High retention of immunogens with curdlan-based support would be expected to provide ideal adjuvant activity. A high molecular weight tetanus anatoxin (150 kDa) has been previously shown to provide about 16-fold lower retention on curdlan than a low molecular weight lysozyme (14.6 kDa) [
51]. PrtA is 105 kDa in size which is self-evidently higher than the 40 kDa PopB, that the inefficiency of the IL-17A response induced by PrtA was probably attributable to the low retention on curdlan.
Although PopB/curdlan confers higher protection against pulmonary infection, 40% lethality is still observed within 5 days [
33], which might be correlated with an antibody response that is ineffective suppressing bacterial propagation.
S. pneumoniae D39 is a highly invasive, encapsulated strain with the ability to mediate antiphagocytosis [
52]. Clinical PPSV23 vaccine offered 100% protection to mice infected with D39 due to optimal antibody binding, which results in complement deposition or opsonophagocytosis leading to effective pneumococcal clearance. Examination of the protective efficacy of PrtA immunization by determination of the bacterial load in the blood after intravenous infection revealed that the PrtA1/curdlan immunized group, but not the adjuvant control group, restrained bacteria proliferation during the first 2 days (Fig.
6a). However, it failed to protect against lethal bacteremia (Fig.
6b). We determined the potency of PrtA-antisera bound to pneumococcal surfaces and found only a portion of the pneumococcal population was conjugated (Fig.
3c, Additional file
1), which might be the point of vulnerability leading to immune escape and disease progression.
PrtA1/curdlan did not induce marked antigen-specific Th2 response, which was possibly attributable to Th1 skewing by curdlan modulation. Because Th1/2 imbalance could limit the antibody responses, we verified the efficacy of an alternative immunization with PrtA combined with traditional complete Freund’s adjuvant-incomplete Freund’s adjuvant (CFA-IFA), which was assumed to efficiently enhance global immune responses. When combined with CFA-IFA, PrtA immunization induced a considerably higher antigen-specific Th1 and Th2 response than did PrtA1/curdlan; Th17 induction was comparable with PrtA1/curdlan (Fig.
2, Additional file
2). Although CFA-IFA facilitated Th2 activation, PrtA-specific antibody titers did not improve in the sera or mucosal secretions compared with PrtA1/curdlan (Fig.
3, Additional file
3). Unlike the limited interaction of PrtA1/curdlan antisera with TIGR4 cells (Fig.
3c), the antisera from PrtA/CFA-IFA-immunized mice could successfully recognize D39 as well as TIGR4 cells (Additional file
3), possibly because of the T-cell receptor repertoire induced by CFA-IFA [
53]. However, the potency of antigen-specific antibody conjugation on the pneumococcal cell surface was still limited (Additional file
3). Moreover, it failed to improve bacterial clearance from the blood and could not stop the progression of bacteremia (Additional file
4).
The number of effective immunoglobulins from both PrtA-immunized antisera (PrtA/curdlan and PrtA/CFA-IFA) bound to pneumococcal cell surfaces was lower than that from PPSV23-immunized antisera (Additional file
1). This suggested that fewer epitopes on the pneumococcal cell surface were recognized by the PrtA-antisera than by PPSV23 antisera. Moreover, higher number of pneumococci were not conjugated by Ig in PrtA-antisera than in PPSV23 antisera, highlighting the imperfect reactivity of PrtA-antisera against the pneumococci due to population heterogeneity [
54,
55]. The inefficiency of PrtA/curdlan antisera was not different from that of PPSV23 antisera in the opsonophagocytosis assay (Fig.
4b). It seemed that the ratio of bacterial cells opsonized by either PrtA- or PPSV23 antisera versus phagocytes in the assay could be large enough to ignore the proportion of bacterial cells eluding Ig recognition. These escaped pneumococci might have continued to replicate; thereby, breaching the infectivity threshold. This is in accordance with the outcome of the systemic invasion murine model, which restrained bacteria load in day 2 but could not stop the bacterial propagation and disease severity to the end of experiment. The mechanism underlying the variable expression of PrtA on the pneumococcal surface remains unclear, and if elucidated, might partly explain the failure of PrtA immunization against
S. pneumoniae invasion.
Although the PrtA/curdlan vaccine did not suppress pneumococcal invasion, we found that it tended to reduce the bacterial load in the lung after 3 days of infection, in addition to seemingly restricting the bacterial propagation during the early phase of blood invasion, despite the limited potency of the antibody response. We have previously found that BABL/c mice might not be suitable for evaluating vaccine efficacy because they are more resistant to
S. pneumoniae infection than other mouse strains [
56]. Furthermore, a high dose of infection, which is needed to potentiate pneumococcal disease, could exceed the threshold of protection conferred by the PrtA/curdlan vaccine. This limitation might also partly explain the failure of the PrtA vaccine.
Global efforts are being made to develop efficient protein vaccine candidates against S. pneumoniae. An effective mucosal Th2 response and IgA titer, in addition to antigen-specific Th17 are essential to prevent pneumococcal colonization of the lungs. A high titer of antigen-specific IgG and sufficient potency to bind to the surface of all bacteria in population, are critical to prevent pneumococcal invasion. Thus, the variable expression of selected pneumococcal vaccine candidates needs to be considered in the field.