ABSTRACT
Immunization with Plasmodium sporozoites can elicit high levels of sterile immunity, and neutralizing antibodies from protected hosts are known to target the repeat region of the circumsporozoite (CS) protein on the parasite surface. CS-based subunit vaccines have been hampered by suboptimal immunogenicity and the requirement for strong adjuvants to elicit effective humoral immunity. Pathogen-associated molecular patterns (PAMPs) that signal through Toll-like receptors (TLRs) can function as potent adjuvants for innate and adaptive immunity. We examined the immunogenicity of recombinant proteins containing a TLR5 agonist, flagellin, and either full-length or selected epitopes of the Plasmodium falciparum CS protein. Mice immunized with either of the flagellin-modified CS constructs, administered intranasally (i.n.) or subcutaneously (s.c.), developed similar levels of malaria-specific IgG1 antibody and interleukin-5 (IL-5)-producing T cells. Importantly, immunization via the i.n. but not the s.c. route elicited sporozoite neutralizing antibodies capable of inhibiting >90% of sporozoite invasion in vitro and in vivo, as measured using a transgenic rodent parasite expressing P. falciparum CS repeats. These findings demonstrate that functional sporozoite neutralizing antibody can be elicited by i.n. immunization with a flagellin-modified P. falciparum CS protein and raise the potential of a scalable, safe, needle-free vaccine for the 40% of the world's population at risk of malaria.
INTRODUCTION
Plasmodium sporozoite vaccine studies in the 1960s and '70s provided the first demonstration that sterile immunity for malaria could be elicited in murine models and in human volunteers (1–4). Immune sera from the protected hosts identified the circumsporozoite (CS) protein covering the sporozoite surface as the target of antibody-mediated immunity (5, 6), and more recent studies have confirmed the CS protein as one of the major targets of protective immunity elicited by sporozoite immunization (7). A recent phase III clinical trial of a Plasmodium falciparum CS subunit vaccine, termed RTS,S, represents the first malaria vaccine to reach testing for commercial licensure (8, 9). The results demonstrate a reduction of clinical disease in 31% of neonates aged 6 to 12 weeks and in 56% of infants aged 5 to 17 months. While the endpoint in these trials was clinical disease, these results support efforts to develop more efficacious second-generation malaria vaccines to elicit sterile immunity for individuals living in areas where malaria is endemic, as well as for travelers to those regions.
A critical factor in the development of subunit vaccines is the need for strong adjuvants to stimulate innate immune responses required to initiate adaptive cellular and humoral immunity. Immunogenicity of the RTS,S vaccine is adjuvant dependent, and a formulation in alum, the aluminum adjuvant present in most licensed vaccines, elicited suboptimal protection against sporozoite challenge (10, 11). The current RTS,S adjuvant formulation was derived empirically and is comprised of a mixture of monophosphoryl lipid A (MPL) and QS21, a purified fraction of the detergent saponin, in an oil-in-water emulsion (AS02) or a liposome (AS01) formulation (12, 13).
In contrast to empirically defined adjuvants which lack a mechanistic rationale, pathogen-associated molecular patterns (PAMPs) provide well-defined protein, lipid, or other molecular moieties that are known to stimulate cytokine/chemokine production by innate immune cells (14). A number of natural and chemically synthesized Toll-like receptor (TLR) agonists, such as bacterial cell wall lipopeptides and CpG motifs of bacterial DNA, have been used as adjuvants (15, 16). Bacterial flagellin, an agonist of TLR5, is one of the limited number of protein PAMPs (17) and thus can be expressed readily in Escherichia coli, using standard scale-up and quality control methods, for production of low-cost vaccines for developing countries. Multiple studies have demonstrated the adjuvant properties of flagellin for enhancing T and B cell responses to bacterial, viral, and parasite antigens (18–23). Recent phase I trials demonstrated safety and immunogenicity of an influenza vaccine produced by VaxInnate, Inc., which is comprised of a recombinant fusion protein of flagellin and the globular head of hemagglutinin (HA) (24–26). The clinical trials demonstrated induction of virus-neutralizing antibodies in young (18 to 49 years) and elderly (>65 years) volunteers following intramuscular (i.m.) immunization with the flagellin-modified HA vaccine.
In the present study, we examined the immunogenicity and protective efficacy of a recombinant P. falciparum CS protein modified with either a full-length flagellin of Salmonella enterica serovar Typhimurium, i.e., FljB (STF2), or a truncated flagellin protein (STF2Δ) in which the hinge region containing immunodominant flagellin B cell epitopes was removed. The P. falciparum CS in these constructs was either a nearly full-length CS protein or a triepitope sequence, T1BT*, containing well-defined T and B cell epitopes of P. falciparum CS that are recognized by human and murine immune cells. Immunization of mice with either flagellin-modified CS construct, administered via the intranasal (i.n.) or subcutaneous (s.c.) route, elicited systemic malaria-specific immune responses of similar magnitudes and specificities. However, significant levels of sporozoite-neutralizing antibodies were elicited only by i.n. immunization, with >90% reductions in parasite levels in sporozoite invasion assays in vitro. Mice immunized with flagellin-modified CS administered i.n., but not s.c., had >90% reductions in liver parasite burdens in vivo following challenge with Plasmodium-infected mosquitoes.
This is the first description of a needle-free flagellin-modified P. falciparum CS protein vaccine that can elicit functional sporozoite-neutralizing antibodies. Optimization of an i.n. malaria vaccine would provide significant cost and safety advantages in resource-poor areas of the world, which experience the majority of the 250 to 500 million infections and ∼1 million deaths each year caused by the Plasmodium parasite.
MATERIALS AND METHODS
Flagellin-modified P. falciparum circumsporozoite protein.Two types of flagellin-modified fusion proteins containing the P. falciparum CS protein were expressed in E. coli (Fig. 1). The STF2.(T1BT*) constructs contained full-length flagellin from Salmonella enterica serovar Typhimurium FljB (STF2) fused to T1BT*, a triepitope malaria sequence representing multiple immunodominant T and B cell epitopes of P. falciparum CS protein that were identified using sera and cells from P. falciparum sporozoite-immunized volunteers (27–29) (Fig. 1B). The T1 and B epitopes are located in the repeat region of the P. falciparum CS protein (NF54 isolate), while the universal T cell epitope, T*, is located in C-terminal amino acids (aa) 326 to 345 (Fig. 1A). Constructs containing T1BT* as either a single copy [STF2.(T1BT*)1X] or four copies [STF2.(T1BT*)4X] were compared for immunogenicity.
(A) Schematic diagram of P. falciparum CS protein showing locations of the T1 and B cell epitopes within the central repeat region and the universal Th cell epitope, T*, in the C terminus. (B) Schematic diagrams of E. coli-expressed recombinant proteins comprised of full-length flagellin (STF2) combined with P. falciparum T1BT* epitopes as either a single copy [STF2.(T1BT*)1X] or four copies [STF2.(T1BT*)4X]. A second type of construct, STF2Δ.CS, was comprised of a truncated flagellin protein (STF2Δ) and nearly full-length P. falciparum CS protein.
A second type of flagellin-modified protein, STF2Δ.CS, contained a nearly full-length P. falciparum CS that lacked only the N-terminal 23-aa signal sequence and the C-terminal 14 aa, containing the putative glycosylphosphatidylinositol (GPI) anchor sequence, fused to the C terminus of a truncated form of the S. enterica serovar Typhimurium flagellin (STF2Δ) from which the hypervariable hinge region (aa 170 to 415) was deleted to reduce the size of the chimeric fusion protein (Fig. 1B). A flexible linker sequence (GAPVDPASPW) was inserted into the hinge region of the truncated flagellin protein to facilitate interactions between the N- and C-terminal regions required for binding to TLR5 (30). The STF2Δ truncated flagellin was previously shown to function as a potent TLR5 agonist and adjuvant in murine studies of a West Nile virus vaccine (20).
The E. coli-expressed recombinant proteins were purified by column chromatography or by use of nickel chelating columns, as previously described (19–21). Endotoxin levels were quantified using a Limulus amoebocyte lysate (LAL) assay, and all constructs contained <0.01 endotoxin unit (EU)/μg. Western blots and an enzyme-linked immunosorbent assay (ELISA) using monoclonal antibody (MAb) 2A10, specific for P. falciparum CS repeats, were used to confirm the purity and antigenicity of the flagellin-modified recombinant proteins (see Fig. S1 in the supplemental material).
TLR5 agonist activity of the flagellin-modified CS proteins was assessed using a murine cell line, RAW 264.7, transfected with human TLR5 (hTLR5), as previously described (19). Briefly, hTLR5-transfected RAW 264.7 and untransfected RAW 264.7 cells were cultured in 96-well plates at 3 × 104 to 5 × 104 cells/well in Dulbecco's modified Eagle's medium (DMEM) (Cellgro, Fisher Scientific, Pittsburgh, PA) supplemented with 10% fetal calf serum (FCS) (HyClone, Logan, UT). Cells were treated for 5 h with serial dilutions of flagellin-modified CS or flagellin only, and levels of tumor necrosis factor alpha (TNF-α) in the supernatant were measured by ELISA (Invitrogen, Carlsbad, CA).
Dendritic cells (DC). (i) Murine DC.Bone marrow-derived dendritic cells (BMDC) were isolated from femurs and tibias of naive BALB/c or C57BL/6 mice by flushing with complete high-glucose DMEM with 10% FCS, penicillin-streptomycin, and β-mercaptoethanol. Cells were centrifuged at 300 × g for 5 min, and the pellet was resuspended in red blood cell lysis buffer at room temperature (RT) for 5 min. Washed cells were cultured in complete DMEM supplemented with 30% conditioned medium obtained from an Ag8.653 cell line transfected with the murine granulocyte-macrophage colony-stimulating factor (GM-CSF) gene (31), kindly provided by A. Rodriguez, New York University. Cells were cultured at 37°C and 5% CO2 in complete medium containing GM-CSF, which was changed every other day.
An immature myeloid DC line, D1, derived from spleens of C57BL/6 mice (32, 33), was provided by A. Rodriguez, New York University. Cells were grown in high-glucose DMEM supplemented with 10% FCS, penicillin-streptomycin, and 30% conditioned medium from transfected Ag8.653 cells expressing murine GM-CSF.
(ii) hDC.Human DC (hDC) were derived from monocytes purified from peripheral blood mononuclear cells (PBMC) following differentiation in vitro to myeloid CD11c+ CD14− DC by stimulation with human interleukin-4 (IL-4) and GM-CSF cytokines (R&D Systems, Minneapolis, MN), as previously described (34). The immature DC were cryopreserved and stored at −80°C until used. To obtain hDC for phenotypic assays, thawed cells were incubated overnight and stimulated the following day with IL-4 and GM-CSF (R&D Systems).
(iii) DC phenotypes.Subset and maturation markers were measured in a FACSCalibur flow cytometer (Becton, Dickinson, Franklin Lakes, NJ). Cells were preincubated on ice for 30 min with Fc block (CD16/32; BD Biosciences, San Diego, CA) diluted 1:200 in fluorescence-activated cell sorter (FACS) buffer (3% FCS in phosphate-buffered saline [PBS]). Cells were then incubated on ice for 30 min with a 1:100 dilution of fluorophore-conjugated antibodies: CD40-fluorescein isothiocyanate (CD40-FITC) (BD Biosciences), IAd-phycoerythrin (IAd-PE) or IAb-PE (BD Biosciences), CD86-peridinin chlorophyll protein/Cy5.5 (CD86-PerCP/Cy5.5), and CD11c-allophycocyanin (CD11c-APC) (both from Biolegend, San Diego, CA). For FACS analysis, cells were gated on CD11c, and levels of CD40, major histocompatibility complex class II (MHC II), and CD86 were analyzed with FlowJo software. Controls included cells stimulated with E. coli K-12-derived lipopolysaccharide (LPS) (InVivogen, San Diego, CA) or a maturation-inducing cytokine cocktail, in the case of hDC.
Uptake of flagellin-modified CS protein in vitro.Murine BMDC or human DC were plated into 8-well Permanox Lab-Tek chamber slides at 2 × 105 cells per well and incubated with 10 or 100 μg/ml STF2Δ.CS or STF2.(T1BT*)4X at 4°C for 30 min to allow synchronized binding, followed by 30 min or 24 h at 37°C to allow endocytic uptake. Cells were fixed in situ with 4% paraformaldehyde in PBS for 15 min at RT, washed with a fixation/permeabilization solution (Biolegend), and incubated in the same solution for 10 min at RT. Cells were permeabilized with 0.25% Triton X-100 in PBS. The presence of internalized CS protein in permeabilized, fixed DC was determined by labeling with biotinylated MAb 2A10, specific for P. falciparum CS repeats, followed by washing with 6% bovine serum albumin (BSA) in PBS and incubation for 30 min on ice with streptavidin-conjugated 585-nm Qdots (Invitrogen, Life Technologies, Grand Island, NY). Cells were then washed three times with FACS buffer and acquired in a FACSCalibur flow cytometer, as described above.
Immunization and challenge.Immunogenicity and protective efficacy of the flagellin-modified CS proteins were tested in C57BL/6 and BALB/c mice (3 to 6 mice/group) (Jackson Laboratories, Bar Harbor, ME). TLR5 knockout (KO) mice were kindly provided by Andrew Gewirtz, Emory University, Atlanta, GA. Mice were housed in AAALAC-approved animal facilities according to NYU School of Medicine IACUC-approved protocols. Mice were immunized s.c. or i.n. with 4 to 6 doses of flagellin-modified CS (10 to 50 μg/dose) at 2-week intervals. For i.n. immunization, 10 μl was placed in each naris of anesthetized mice. Mice were used at ≥8 weeks of age, as nasopharynx-associated lymphoid tissue (NALT) organogenesis does not begin until after birth and reaches maturation at ∼8 weeks of age (35). Controls included mice immunized with full-length or truncated flagellin without CS or with unmodified T1BT* peptide. Sera were collected 14 days after each immunization and stored at −20°C until used.
To test protective efficacy, mice were challenged by exposure to the bites of 5 to 15 mosquitoes infected with PfPb, a transgenic P. berghei rodent parasite expressing P. falciparum CS repeats (36, 37). The mechanism of immunity was investigated by depletion of T cells by injection of 100 μg anti-CD4 (GK 1.5) and anti-CD8 (2.43) MAbs on three consecutive days prior to challenge. Mice were dissected at 40 h postchallenge, and parasite 18S rRNA levels in liver extracts were quantified by real-time quantitative PCR (qPCR) (38). Briefly, livers were placed in TRI reagent (Molecular Research Center, Inc., Cincinnati, OH) and homogenized with a PowerGen 125 tissue homogenizer (Fisher Scientific, Pittsburgh, PA). Total RNA was extracted from the homogenate by use of chloroform-isoamyl alcohol (24:1), the aqueous phase was separated, and RNA was precipitated with isopropyl alcohol. RNA pellets were washed with 70% ethyl alcohol, dried, resuspended in diethyl pyrocarbonate (DEPC) H2O, and dissolved by heating at 44°C for up to 1 h. Total RNA was converted into cDNA by using reverse transcriptase (Applied Biosystems, Carlsbad, CA), and qPCR was performed using primers specific for the P. berghei 18S rRNA gene and SyBr green (Qiagen, Valencia, CA). The number of 18S rRNA copies in each sample was based on a standard curve generated with known amounts of plasmid 18S cDNA. Inhibition of >90% of 18S rRNA copy numbers compared to that in naive challenged mice was considered significant, based on previous studies using intravenous injection of known numbers of sporozoites which demonstrated that a log reduction in sporozoite numbers is required to give a 1-day delay in the prepatent period (1, 39–41).
Serologic assays. (i) ELISA.IgG titers were determined using 96-well plates coated with P. falciparum CS repeat peptide (T1B)4, recombinant full-length (STF2) or truncated (STF2Δ) flagellin protein, or flagellin-modified CS protein STF2.(T1BT*)4X or STF2Δ.CS. Results for individual sera are expressed as geometric mean titers (GMT) with upper and lower 95% confidence intervals (CI). A >4-fold difference in GMT was considered significant. IgG subtypes were determined using CS repeat peptide-coated wells and an ELISA kit containing an enzyme-labeled MAb specific for murine IgG subtypes (Southern Biotech, Birmingham, AL).
(ii) Sporozoite assays.Reactivity with native CS on P. falciparum sporozoites was measured by an indirect immunofluorescence assay (IFA) using 2-fold serum dilutions starting at 1:80. The endpoint titer was the last dilution giving unequivocal positive fluorescence. Reactivity with viable sporozoites was measured by the circumsporozoite precipitin (CSP) reaction (42), in which 2-fold dilutions of sera were incubated with viable PfPb sporozoites expressing P. falciparum CS repeats. Formation of a terminal CSP reaction was determined by phase microscopy, and the endpoint taken was the last serum dilution giving positive CSP reactions on 2 sporozoites among a total of 20 sporozoites counted.
(iii) TSNA.Functional anti-repeat antibodies were measured in an in vitro transgenic sporozoite neutralization assay (TSNA) as previously described (37, 43). Briefly, 2 × 104 transgenic PfPb sporozoites, in which the P. berghei CS repeats have been replaced with P. falciparum CS repeats, were incubated for 40 min on ice with a 1:5 dilution of immune or naive serum. The preincubated PfPb sporozoites were then added to confluent human HepG2 hepatoma cells and cultured at 37°C and 5% CO2. After 48 h, cells were lysed and homogenized using QIAshredder columns (Qiagen), and total RNA was extracted using a PureLink RNA minikit (Life Technologies, Grand Island, NY) according to the manufacturer's instructions. The level of parasites in the cultures was measured by real-time qPCR, as described above. Results were expressed as numbers of copies of P. berghei 18S rRNA contained in the cell extracts. A >90% reduction in 18S rRNA copy number compared to cultures receiving PfPb sporozoites in medium or naive serum was considered significant.
Cellular assays. (i) ELISPOT assay.Th1- and Th2-type cell responses were measured by gamma interferon (IFN-γ) and IL-5 ex vivo enzyme-linked immunosorbent spot (ELISPOT) assays using splenocytes obtained 7 to 14 days after the last immunization. For expanded ELISPOT assays, spleen cells were cultured in vitro for 7 days with T1BT* peptide (10 μg/ml) prior to assay. Cellular responses measured in the expanded ELISPOT assay reflect the presence of malaria-specific memory T cells (44).
For ELISPOT assays, washed cells were added at 4 × 105/well in ELISPOT plates coated with either anti-mouse IFN-γ or anti-mouse IL-5 (BD Biosciences). Malaria peptides were added at 10 μg/ml, and flagellin-modified CS or flagellin recombinant proteins were used at 1 μg/ml. Irradiated naive splenocytes were added as a source of antigen-presenting cells. The plates were incubated at 37°C and 5% CO2 for 40 h and developed according to the manufacturer's instructions. The number of spot-forming cells (SFC) was determined using an automated reader (CTL Technologies, Amarillo, TX), and results are expressed as mean numbers of SFC/106 cells ± standard deviations (SD).
(ii) Cytometric bead arrays.Supernatants from ELISPOT cultures were collected and stored at −20°C. Cytokines in the supernatants were measured using cytometric bead arrays (CBAs) according to the manufacturer's instructions (BD Biosciences). Bead fluorescence was measured in a FACSCalibur flow cytometer (Becton, Dickinson) and converted to cytokine concentrations by using CBA Analysis (version 1.4.2) or FCAP software from BD Biosciences.
Confocal microscopy.DC cultures were analyzed with an inverted Leica TCS SP2 AOBS confocal system equipped with a Ludin chamber for control of temperature and CO2 concentration. Data were acquired with Leica confocal software; Imaris 7.4 (Bitplane, Saint Paul, MN), Image-Pro Plus (Media Cybernetics, Bethesda, MD), AutoDeBlur (Media Cybernetics, Bethesda, MD), and NIH ImageJ were used for further image processing and deconvolution. Confocal figure panels were assembled in Adobe Photoshop.
Statistics.Statistical analyses used GraphPad Prism, version 6.00, for Windows (GraphPad Software, La Jolla, CA). Pairwise comparisons used Student's t test or the nonparametric Mann-Whitney rank sum test, while multiple comparisons used analysis of variance (ANOVA) or the Kruskal-Wallis test with the Dunn procedure. Reductions of >90% of parasite rRNA copy numbers were considered significant in both the in vitro and in vivo assays. The chi-square test (2 × 2 contingency tables) was used to compare the numbers of mice in groups with >90% inhibition. In all tests, P values of <0.05 were considered statistically significant.
RESULTS
Flagellin-modified P. falciparum CS functions as a TLR5 agonist.Two types of flagellin-modified P. falciparum CS recombinant proteins (Fig. 1), comprised of either full-length flagellin and immunodominant CS epitopes [STF2.(T1BT*)n] or truncated flagellin and a nearly full-length CS protein [STF2Δ.CS], were analyzed for TLR agonist activity, immunogenicity, and protective efficacy. Both STF2.(T1BT*)n containing full-length flagellin (Fig. 2A) and STFΔ.CS containing truncated flagellin (Fig. 2B) effectively stimulated RAW 264.7 cells transfected with human TLR5. The cytokine levels elicited by STF2Δ.CS stimulation of hTLR5-transfected cells were comparable to levels obtained following stimulation with STF2Δ only, indicating that the presence of the ∼25-kDa CS protein at the C terminus of the truncated flagellin did not interfere with flagellin-mediated TLR5 signaling (Fig. 2B). The lack of stimulation of untransfected RAW 264.7 cells (Fig. 2A and B, open symbols), which express TLR4, is consistent with the low endotoxin levels (<0.01 EU/μg) and high levels of purity of the flagellin-modified CS constructs (see Fig. S1 in the supplemental material).
TLR5 signaling by flagellin-modified CS constructs. Human TLR5-transfected RAW 264.7 cells (closed symbols) or untransfected cells (open symbols) were stimulated with 10-fold dilutions of (A) STF2.(T1BT*)1X or a positive-control protein, flagellin-modified ovalbumin (STF2.OVA), or (B) STF2Δ.CS, and levels of TNF-α in culture supernatants were measured by ELISA. (C) Flow cytometry of TLR5 expression in human monocyte-derived DC that were immature, matured with a cytokine cocktail, or stimulated with STF2 flagellin (10 μg/ml). (D) Flow cytometry of TLR5 expression in murine D1 cells, an immature DC line.
Flagellin-modified P. falciparum CS enhances murine and human DC maturation.Given the ability to trigger cytokine production in hTLR5-transfected cells (Fig. 2A and B), the flagellin-modified CS constructs were examined for the ability to stimulate murine and human DC. TLR5 was detected by flow cytometry on a subset of human monocyte-derived DC (hDC), with approximately 18% of CD11c+ hDC reacting with anti-TLR5 antibody (Fig. 2C). The percentage of TLR5-positive cells did not change when the hDC cultures were stimulated with flagellin or with maturation medium, suggesting that the levels of TLR5 expressed on the hDC were not modified during DC maturation and may represent a subset of the CD11c+ DC population.
In contrast to the human DC, murine BMDC stained only weakly with anti-TLR5 antibody (data not shown), as noted in previous studies (45). However, TLR5 was detected on a murine immature DC line, D1, with approximately 40% of the cells staining positive (Fig. 2D).
Signaling through TLR5 leads to upregulation of costimulatory molecules required for DC-T cell interactions that initiate adaptive immune responses (23, 46). To determine if flagellin interaction with TLR5 induced maturation of human DC, CD86 expression on monocyte-derived hDC stimulated with STF2Δ.CS, or with unmodified CS without flagellin, was measured by confocal microscopy and flow cytometry. Increased CD86 expression was detected on hDC incubated with STF2Δ.CS compared to cells incubated with PBS (Fig. 3A). Recombinant CS protein alone, without flagellin, elicited only low levels of CD86 in hDC, indicating that the flagellin moiety of the modified CS was required for upregulation of CD86.
DC maturation following stimulation with flagellin-modified CS constructs. (A) Confocal microscopy of monocyte-derived human DC expression of costimulatory molecule CD86 (green) following incubation with unmodified CS protein or STF2Δ.CS. The DC nuclei were stained with Hoechst stain (blue). (B) Percentages of CD86+ human DC after 6 or 24 h of stimulation with the flagellin-modified CS construct STF2.(T1BT*)4X or STFΔ.CS or with full-length or truncated flagellin (STF2 and STF2Δ). Means with SD for replicates are shown. Data are representative of two independent experiments. (C) Percentages of murine D1 cells expressing costimulatory and MHC II molecules following stimulation with 10 μg/ml of flagellin or flagellin-modified CS proteins. Means with SD for replicates are shown. Data are representative of three independent experiments.
The expression of CD86 on hDC increased with time of incubation. After 6 h of incubation with STF2Δ.CS or STF2.(T1BT*)4X, 10 to 20% of hDC were CD86+, with levels increasing to approximately 40% in 24-h cultures (Fig. 3B). The truncated flagellin protein (STF2Δ) was as effective as full-length flagellin (STF2) in stimulating upregulation of costimulatory molecules on the hDC. The similar percentages of CD86-positive cells and similar mean fluorescence intensities (MFI) (data not shown) in cultures stimulated with flagellin-modified CS constructs or flagellin only indicate that the presence of the large CS protein or multiple T1BT* copies did not alter the interaction with TLR5 on hDC.
The flagellin-modified CS constructs also increased murine DC expression of costimulatory molecules. Incubation of D1 cells with STF2Δ.CS or STF2.(T1BT*)4X stimulated upregulation of CD40 and CD86, as well as increasing the percentage of cells expressing MHC II molecules (Fig. 3C). Incubation with either full-length flagellin (STF2) or truncated flagellin (STF2Δ) elicited similar increases, indicating that upregulation of costimulatory molecules on D1 cells was induced by the flagellin moiety, as found with the hDC.
DC uptake of flagellin-modified CS proteins.In addition to expression of costimulatory molecules, DC must also efficiently internalize antigens to process and present peptides to MHC II-restricted CD4+ T helper cells. To assay uptake of flagellin-modified CS, murine BMDC were incubated with either STF2.(T1BT*)4X or STF2Δ.CS, followed by permeabilization and staining with MAb 2A10, specific for P. falciparum CS repeats. Intracellular CS was detected in murine BMDC incubated with either of the flagellin-modified CS constructs (Fig. 4A). Cells incubated with STF2 flagellin only did not stain with MAb 2A10, confirming the specificity of antibody staining for malaria CS protein.
DC uptake of flagellin-modified CS constructs. (A) FACS histogram of murine BMDC following 24 h of incubation with flagellin or flagellin-modified CS (10 μg/ml). Intracellular CS was detected by labeling permeabilized cells with biotinylated MAb 2A10 followed by streptavidin-conjugated Qdots. (B) MAb 2A10-positive murine BMDC measured after incubation for various times with flagellin-modified CS constructs or flagellin-only controls (STF2 or STF2Δ). Means with SD for replicates are shown. Data are representative of three independent experiments. (C) Dose-dependent increase in MAb 2A10 staining of human DC following 24 h of incubation with flagellin-modified CS constructs or flagellin-only controls. Means with SD for replicates are shown. Data are representative of three independent experiments.
Uptake of flagellin-modified CS by murine BMDC was time and dose dependent. After incubation with 10 μg/ml STF2Δ.CS, intracellular CS could first be detected in BMDC at 0.5 to 6 h of incubation, with the intensity of staining increasing after 24 h (Fig. 4B). Cells incubated with a log lower concentration of STF2Δ.CS (1 μg/ml) gave lower levels of intracellular CS (data not shown). Only low levels of MAb 2A10 stain were detected in cells incubated with T1BT* peptide without flagellin (data not shown), suggesting that the presence of flagellin in the fusion protein enhanced antigen uptake, as reported for other flagellin fusion proteins (45).
Human DC also internalized the flagellin-modified CS after 24 h of incubation with 10 μg/ml of STF2Δ.CS (Fig. 4C). Lower levels of intracellular CS were detected following incubation of cells with STF2.(T1BT*)4X, which may reflect enhanced detection by the MAb due to the presence of a larger number of repeats in the full-length CS protein (41 repeats) than in the (T1BT*)4X construct (24 repeats).
Humoral immunity elicited by flagellin-modified P. falciparum CS constructs.To assess immunogenicity, antibody responses were measured in C57BL/6 and BALB/c mice injected s.c. with the flagellin-modified CS constructs. In preliminary studies, a flagellin-CS fusion protein containing four copies of the T1BT* sequence, STF2.(T1BT*)4X, induced higher anti-CS repeat antibody titers with faster kinetics than constructs containing only a single copy of the malaria epitopes (see Fig. S2 in the supplemental material). Further studies therefore compared the immunogenicity of STF2.(T1BT*)4X to that of STF2Δ.CS, which contains the entire repeat region (41 tetramer repeats) as well as multiple CD4+ and CD8+ T cell epitopes in the C terminus of CS.
C57BL/6 and BALB/c mice immunized s.c. with STF2Δ.CS or STF2.(T1BT*)4X developed antibodies with similar fine specificities and peak GMT in ELISA (Table 1). IgG titers in both strains of mice were highest against the respective immunogens, with peak GMT of >105 following s.c. immunization with either STF2.(T1BT*)4X or STF2Δ.CS. High titers of anti-flagellin antibodies (GMT > 104), were also detected following immunization with either of the flagellin-modified CS proteins (data not shown). In contrast, antibody to the CS repeat peptide was a log lower, with peak titers of 2 × 103 to 3 × 103 in both strains of mice. Overall, there was no significant difference (>4-fold) in ELISA titers elicited by s.c. immunization with the different flagellin-modified CS constructs, indicating that both full-length and truncated flagellin provided similar adjuvant properties for induction of antibody to malaria CS repeats.
Immunogenicity of flagellin-modified CS proteins
The antibodies elicited in BALB/c and C57BL/6 mice immunized with either of the flagellin-modified CS constructs also reacted with native CS protein on P. falciparum sporozoites. A maximum IFA titer of 103 to 104 was detected in pooled immune sera, consistent with the magnitude of the anti-repeat antibodies measured by ELISA. These antibodies also reacted with viable sporozoites and elicited circumsporozoite precipitin (CSP) titers when immune sera were incubated with PfPb sporozoites (data not shown). The CSP reaction reflects the shedding of antibody-cross-linked surface CS protein (42, 47), demonstrating that the antibodies elicited with flagellin-modified CS proteins effectively recognized native CS protein on the viable sporozoite.
Route of immunization.S. enterica serovar Typhimurium is a bacterial pathogen that invades the host through the gut mucosa. To determine if flagellin-modified CS was immunogenic when delivered via a mucosal route, C57BL/6 mice were immunized i.n., and levels of antibody were compared to those in s.c. immunized mice. Mice immunized i.n. developed a systemic IgG response comparable in magnitude and fine specificity to that with s.c. immunization (Table 1). Antibody to the STF2Δ.CS immunogen increased with increasing i.n. doses, with a peak GMT of 105 obtained post-third dose, similar to the magnitude and kinetics of the antibody response following s.c. immunization (Fig. 5A). No significant increase (>4-fold) was obtained following additional i.n. or s.c. booster immunizations. A large proportion of antibodies elicited by either the i.n. or s.c. route were specific for flagellin, with peak GMT of ∼104. All mice also developed antibodies to CS repeats following i.n. immunization, although, as found with s.c. immunization, the peak GMT was 103, a log lower than that of anti-flagellin antibodies.
Kinetics and fine specificities of antibody responses in mice immunized with flagellin-modified CS constructs. (A) Kinetics of IgG response of C57BL/6 mice (3 mice/group) immunized s.c. or i.n. with STF2Δ.CS (50 μg/dose), as measured by ELISA using CS repeat peptide, flagellin (STF2Δ), or STF2Δ.CS as the coating antigen. Results shown are geometric mean titers (GMT) with 95% confidence intervals. (B) IgG GMT in tlr5−/− mice and WT controls (3 mice/group) immunized s.c. or i.n. with four doses of STF2.(T1BT*)4X (10 μg/dose). Following four immunizations, all WT controls seroconverted. In contrast, none of the tlr5−/− mice immunized s.c. had detectable antibodies to CS repeats or flagellin, with one mouse having detectable antibodies to the immunogen. Two of three tlr5−/− mice immunized i.n. had detectable antibodies to CS repeats, with one mouse having detectable antibodies to flagellin and immunogen. The tlr5−/− mice that did not seroconvert were assigned a value of 20 for calculation of GMT.
To confirm that the immunogenicity of the flagellin-modified CS constructs was mediated through TLR signaling, mice lacking TLR5 were immunized i.n. or s.c. with STF2.(T1BT*)4X. Following priming and boosting by either the i.n. or s.c. route, only a single tlr5−/− mouse had antibody to the immunogen, and none had antibody to flagellin, while all wild-type (WT) mice had seroconverted to both the immunogen and flagellin (data not shown). Additional doses of STF2.(T1BT*)4X did not compensate for the lack of TLR5 in the tlr5−/− mice (Fig. 5B). After a total of four immunizations, delivered either i.n. or s.c., the GMT of antibody specific for immunogen, flagellin, or CS repeats in the tlr5−/− mice was 2 logs lower than corresponding titers in WT mice, indicating that development of malaria-specific antibodies as well as flagellin-specific antibodies was TLR5 dependent. The presence of low levels of antibody in the hyperimmunized TLR5 KO mice may reflect inefficient signaling through cytosolic inflammasomes or other innate immune pathways (48).
IgG subtypes in mice immunized with flagellin-modified CS.Mice immunized either i.n. or s.c. with STF2Δ.CS developed mixed Th1/Th2-type antibody responses. At a 1:80 serum dilution, high levels of IgG1 anti-repeat antibodies were detected, with lower levels of Th1-associated IgG2a/c antibodies (Fig. 6A). A predominance of the IgG1 subtype was also observed in the anti-flagellin antibody response, along with IgG2a/c and IgG2b, in both s.c. and i.n. immunized mice (Fig. 6B). Few or no IgG3 antibodies were observed to either CS repeats or flagellin.
IgG subtypes in mice immunized with flagellin-modified CS. Antibodies specific for CS repeats (A) or flagellin (B) were measured in pooled immune sera (1:80 dilution) from mice immunized i.n. or s.c. with four doses of STF2Δ.CS or flagellin (50 μg/dose). OD, optical density.
Similarly, mice immunized with STF2.(T1BT*)4X had only IgG1 and IgG2b subtype antibodies, with no detectable Th1-type IgG2a/c antibody (see Fig. S3 in the supplemental material). Only low levels of IgA to flagellin, but not to CS repeats, were detected in sera of i.n. immunized mice (data not shown).
Cellular immunity elicited by flagellin-modified P. falciparum CS constructs.To assess T cell responses, Th1-type (IFN-γ) and Th2-type (IL-5) cytokine-producing cells were enumerated by ELISPOT assays using spleen cells of mice immunized s.c. or i.n. with STF2Δ.CS. In ex vivo ELISPOT assays, the majority of IL-5- and IFN-γ-specific spot-forming cells (SFC) in the immune spleens were specific for the STF2Δ.CS immunogen or for the STF2Δ flagellin protein (see Fig. S4 in the supplemental material). Stimulation with the malaria T1BT* peptide elicited minimal or no IL-5- and IFN-γ-secreting T cells.
To enhance detection of malaria-specific T cells, immune spleen cells were cultured with the malaria peptide for 7 days prior to testing in an expanded ELISPOT assay. In peptide-expanded cultures, IL-5 SFC were detected upon restimulation with the T1BT* malaria peptide, with higher levels in mice immunized i.n. than in those immunized s.c. (Student's t test; P = 0.006) (Fig. 7A). The i.n. immunized mice also had higher levels of IL-5 SFC specific for the STF2Δ.CS immunogen (Student's t test; P = 0.002). Similar levels of malaria-specific IL-5 SFC were observed in mice immunized i.n. with STF2.(T1BT*)4x or STF2Δ.CS (Student's t test; P = 0.10) (Fig. 7B). IFN-γ-secreting SFC could not be evaluated due to high background levels in the expanded ELISPOT assay.
Flagellin-modified CS constructs elicit Th2-type cytokine responses. Spleen cells were obtained from mice (A) immunized s.c. or i.n. with four doses of STF2Δ-CS (50 μg/dose) or (B) immunized i.n. with five doses of either STF2.(T1BT*)4X or STF2Δ.CS (10 μg/dose). Malaria-specific cells were expanded in vitro with T1BT* peptide for 7 days prior to ELISPOT assay. Results are shown as numbers of spot-forming cells (SFC)/106 cells following restimulation with T1BT* peptide, flagellin (STFΔ), or the respective immunogen [STF2Δ-CS or STF2.(T1BT*)4X]. IL-5 (C) or IL-6 (D) in pooled cell culture supernatants were measured using multiplex CBA. Cytokine concentrations are shown in pg/ml, after subtraction (Δ) of levels in medium-only cultures.
To further explore T cell responses in the i.n. immunized mice, Th1-type (IL-2, IFN-γ, and TNF-α) or Th2-type (IL-4, IL-5 or IL-6, and IL-10) cytokines were measured in supernatants of the stimulated cells by using a cytometric bead array (CBA). Consistent with results obtained in the expanded ELISPOT assay, IL-5 was detected in supernatants following T1BT* restimulation of cells from mice immunized i.n. with STF2.(T1BT*)4X or STF2Δ.CS (Fig. 7C). Conversely, IL-6 levels were higher following stimulation with STF2.(T1BT*)4X or STF2Δ.CS than with the malaria T1BT* peptide (Fig. 7D). Th1-type cytokines were either not detected in the culture supernatants (IL-2 and IFN-γ) or not different from levels in medium controls (TNF-α), consistent with a predominantly Th2-type T cell response detected by ELISPOT assay.
Protective efficacy of flagellin-modified P. falciparum CS constructs.To determine whether immunization with the flagellin-modified CS constructs elicited functional antibodies, sera from immunized mice were tested for the ability to inhibit in vitro invasion of hepatoma cells by PfPb sporozoites, which are transgenic P. berghei parasites that express the P. falciparum CS repeat region (37, 43). The sporozoite-neutralizing activity in the immune sera was dose dependent. Serum obtained following three i.n. immunizations with either STF2.(T1BT*)4X or STF2Δ.CS was not inhibitory (0% or 42% inhibition, respectively) (Fig. 8A). Additional boosters increased sporozoite-neutralizing activity, with sera of mice immunized i.n. with 5 doses of STF2.(T1BT*)4X and STF2Δ.CS giving 98% and 96% inhibition, respectively, compared with naive serum. An additional 6th dose of immunogen did not significantly increase the level of inhibition, with STF2.(T1BT*)4X immune serum giving 95% inhibition and STF2Δ.CS immune serum giving 99% inhibition. Inhibition of >90% is considered significant, as a 90% reduction in sporozoite numbers is required to give a 1-day delay in the prepatent period following sporozoite challenge (39–41).
Sporozoite-neutralizing antibody and protection in mice immunized with flagellin-modified CS constructs. (A) TSNA was carried out using a 1:5 dilution of pooled sera (3 mice/group) obtained after the 3rd to 6th dose of STF2Δ-CS, STF2.(T1BT*)4X, or unmodified T1BT* peptide (10 μg/dose). Results measured by qPCR analysis of hepatoma cell extracts obtained at 48 h postinfection are shown as mean P. berghei 18S rRNA copy numbers. MAb 2A10, specific for P. falciparum CS repeats, and MAb 3D11, specific for P. berghei CS repeats, served as positive and negative controls, respectively. (B) TSNA was carried out using individual sera (1:5 dilution) from 3 mice/group immunized i.n. or s.c. with five doses of STF2Δ.CS (50 μg/dose). 18S rRNA copy numbers in cultures with i.n. immune sera, but not s.c. immune sera, were significantly reduced compared to naive control levels. (C) In vivo protection was measured in the mice used for panel B following challenge by exposure to the bites of 5 to 15 PfPb-infected mosquitoes. Results shown are mean parasite 18S rRNA levels in livers harvested at 40 h postinfection, as measured by qPCR. Consistent with in vitro TSNA, the liver parasite burden was reduced in i.n. immunized but not s.c. immunized mice compared to naive controls following challenge. *, P < 0.05 by the Kruskal-Wallis test; ns, nonsignificant.
The level of sporozoite-neutralizing activity in immune sera obtained after the 5th or 6th dose of STF2.(T1BT*)4X or STF2Δ.CS was comparable to that obtained with 25 μg/ml of MAb 2A10, an inhibitory antibody specific for P. falciparum CS repeats (6). No inhibition was found when sporozoites were preincubated with MAb 3D11, specific for P. berghei CS repeats, indicating that neutralizing activity was specific for P. falciparum CS repeats.
To determine whether the route of immunization was a factor in the induction of neutralizing antibodies, immune sera of mice immunized i.n. were compared to those of s.c. immunized mice by TSNA. Following immunization with five doses of STF2Δ.CS, the sera of mice immunized i.n. gave >90% inhibition, while immune sera from mice immunized s.c. gave a mean 69% reduction in parasite burden (Fig. 8B).
Although the transgenic PfPb parasites used in the TSNA express P. falciparum repeats, they remain rodent parasites that can efficiently infect mice and develop into normal liver- and blood-stage infections (36), thus allowing the measurement of in vivo protective immunity following challenge. Consistent with levels of inhibition observed in the in vitro TSNA, the rRNA copy number was significantly lower in the i.n. immunized mice than in naive mice, while the difference was not significant in s.c. immunized mice 40 h after challenge by the bites of PfPb-infected mosquitoes compared to naive challenge controls (Fig. 8C). All of the i.n. immunized mice (3/3 mice) had >90% reductions in liver-stage burdens (mean, 98%), while in the s.c. group, 0/3 mice had >90% inhibition (mean, 61%) (χ2; P = 0.014).
The correlation of in vitro and in vivo results suggested that the mechanism of immunity in the flagellin-CS-immunized mice is antibody mediated. The sporozoite-neutralizing activity in the sera of the i.n. immunized mice was antibody concentration dependent. In TSNA of immune sera obtained from mice immunized i.n. with five doses of STF2Δ.CS, there was a significant difference in 18S rRNA copy number observed at a 1:5 dilution (90% inhibition) compared to a 1:10 serum dilution (77% inhibition) (Fig. 9A). The level of inhibition at a 1:5 dilution correlated with the level of protection following challenge, with a mean 93% reduction of parasite burdens in the livers of immunized mice (Fig. 9B). Protection in vivo was confirmed to be antibody mediated, as depletion of CD4+ and CD8+ T cells prior to challenge did not abrogate immune resistance (Fig. 9B, hatched bars).
Antibody-mediated protection in mice immunized i.n. with flagellin-modified CS protein. Mice (n = 6 mice/group) were immunized i.n. with STF2Δ.CS or STF2 only (50 μg/dose). (A) TSNA was carried out using a 1:5 or 1:10 dilution of individual sera obtained after the fifth dose. Results shown are mean 18S rRNA copy numbers in 48-h extracts of HepG2 cell cultures as measured by qPCR. (B) In vivo protection measured in immunized mice that were nondepleted (3 mice/group; solid bars) or depleted of T cells (3 mice/group; hatched bars) by injection of anti-CD4 and anti-CD8 MAbs (100 μg of each MAb) for 3 days prior to challenge by exposure to the bites of PfPb-infected mosquitoes. Results shown are mean 18S rRNA copy numbers in liver extracts obtained at 40 h postchallenge, as measured by qPCR. *, P < 0.05; **, P < 0.01 by the Kruskal-Wallis test; ns, nonsignificant.
DISCUSSION
Subunit vaccines produced in standard bacterial expression systems provide numerous safety and cost advantages over attenuated or viral vector-based vaccines. A needle-free vaccine delivered by i.n. immunization would further reduce costs by eliminating the need for sterile syringes, trained medical personnel, and biohazard waste disposal. While frequently used to induce strong mucosal immunity, i.n. immunization also elicits systemic antibody responses comparable to those obtained by s.c. injection (Table 1) and therefore has the potential to target blood-borne pathogens such as Plasmodium parasites.
The TLR agonist properties of S. enterica serovar Typhimurium flagellin have been developed as adjuvants for bacterial, viral, and protozoan subunit vaccines which lack endogenous PAMPs required to stimulate innate immune responses that initiate adaptive humoral and cellular immunity (reviewed in reference 23). Previous studies demonstrated that S. enterica serovar Typhimurium type 1 (FLiC) or type 2 (FljB) flagellin, as well as flagellin from other Gram-negative bacteria, can be used as a vaccine adjuvant. The enhanced immunogenicity of STF2 flagellin fusion proteins compared to mixtures of flagellin and antigen (19, 20) is consistent with the requirement for colocalization of both the TLR agonist and antigen in the same phagosome for optimal immunogenicity (49, 50).
The construction of the flagellin-modified malaria constructs was based on flagellin-modified viral and bacterial vaccines which have been shown to elicit neutralizing antibodies to West Nile, dengue, and influenza viruses, as well as CD8+ T cell-mediated immunity to Listeria, in murine models (19–21). In recent phase I/II studies, an influenza vaccine comprised of the HA globular head inserted at both the hinge region and C terminus of flagellin was shown to be safe and efficacious when low doses were administered i.m. to elderly (>65 years) and young (18 to 49 years) volunteers (25, 26).
In the present study, CS protein modified with either full-length (STF2) or truncated (STF2Δ) flagellin was found to stimulate cytokine production by RAW cells transfected with human TLR5. The lack of response of untransfected RAW cells, which express TLR4, demonstrated that the purified constructs were free of bacterial LPS contaminants, consistent with the low levels of endotoxin detected by LAL assay. In additional studies, we found that C3H/HeJ mice, which are hyporesponsive to LPS due to a mutation in TLR4, developed peak antibody titers to CS repeats (GMT, 7,241) and flagellin (GMT, 97,420) after immunization with flagellin-modified CS protein (data not shown) that were comparable to those of BALB/c and C57BL/6 mice (Table 1), indicating that adjuvanticity was mediated by flagellin and not by minor bacterial contaminants from the E. coli expression system. Immunogenicity was dependent on flagellin stimulation through TLR5 signaling, as tlr5−/− mice developed minimal or no antibody responses following immunization with the flagellin-modified CS protein (Fig. 5B). The adjuvant activity of flagellin has been shown to require direct engagement of flagellin with TLR5 on CD11c+ DC (45). The flagellin-modified CS constructs increased expression of costimulatory molecules and MHC II molecules on a TLR5+ murine DC line (D1) and on monocyte-derived hDC. TLR5 staining of only a subpopulation of cells may reflect the presence of different DC subsets or activation states (51, 52).
The CS repeat region is known to contain a protective B cell epitope (5, 6). The magnitudes of the repeat-specific antibodies elicited by the flagellin-modified CS constructs, containing either full-length or minimal epitopes of CS, were similar (Table 1), indicating that the presence of only a small number of CS repeats and a T helper epitope was sufficient to elicit sporozoite-neutralizing antibodies. The systemic anti-repeat IgG antibody response following i.n. immunization was not significantly different from peak titers obtained by s.c. immunization, and the hierarchy of antibody responses remained the same: immunogen > flagellin > CS repeats (Fig. 5A). However, despite the similar kinetics, specificities, and magnitudes of antibody responses, immune sera of mice immunized i.n. had higher levels of sporozoite-neutralizing antibodies in vitro than those of s.c. immunized mice, and this correlated with higher levels of protection in vivo in i.n. immunized mice than in s.c. immunized mice (Fig. 8). Protection in vivo was mediated by antibody, as depletion of T cells from immunized mice prior to challenge did not reduce levels of inhibition (Fig. 9B).
The mechanism of the enhanced neutralizing activity of antibodies elicited by i.n. immunization compared to s.c. immunization remains to be defined. The affinities of anti-repeat antibodies elicited by i.n. or s.c. immunization were similar in elution assays with chaotropic NH4SCN (see Fig. S3 in the supplemental material). Both of the flagellin-modified CS constructs, STF2.(T1BT*)4X and STF2Δ.CS, administered either i.n. or s.c., elicited predominantly IgG1 (Th2-type) anti-repeat antibodies. Consistent with predominant IgG1 antibody responses, the majority of T cells elicited by the flagellin-modified CS constructs were of the Th2 type, based on production of IL-5 and IL-6, with minimal IFN-γ, as measured by ELISPOT assay and CBA (Fig. 7). Vaccines containing flagellin as adjuvant frequently elicit a mixed Th1/Th2-type response skewed toward Th2-type antibody subtypes in murine studies (19, 22, 53) and in volunteers immunized with a flagellin-modified flu vaccine (25, 26). Thus far, a role for Fc or immunoglobulin (Ig) subtypes in antibody-mediated sporozoite neutralization has not been demonstrated, as Fab of anti-repeat MAbs as well as MAbs of various IgG subtypes can neutralize sporozoite infectivity in vivo (54, 55). A requirement for high concentrations of anti-repeat antibody rather than a specific IgG subtype would be consistent with the immobilization of sporozoites, thereby preventing motility required for invasion of cells in the liver (56). Recent studies have also shown that anti-repeat antibodies immobilize sporozoites in the skin at the site of the mosquito bite, thereby preventing their egress into the circulation and transit to the liver (57).
The differences in the antibodies elicited by i.n. and s.c. immunizations may be related to the unique structure and cell populations found in the murine NALT (58, 59). We recently carried out confocal imaging of the NALT of mice immunized with flagellin-modified CS proteins (60). In contrast to lymph nodes, B cells predominate in the NALT, as found also in Peyer's patches of the gut-associated lymphoid tissue (GALT). The high levels of B cells persisted following a >2-fold increase in the overall size of the NALT in the flagellin-CS-immunized mice. A similar increase in NALT size was noted in mice immunized with flagellin alone, suggesting a role for TLR5 signaling in stimulating cell influx or multiplication within the NALT. Moreover, CD11c+ DC with transepithelial dendrites, as well as microfold (M) cells, were found in the NALT, which could potentially play a role in antigen capture and/or presentation, as found in GALT (61, 62). In humans, the lympho-epithelial structures of the Waldeyer's ring, considered the equivalent of the murine NALT (63), are enlarged in infants, the primary target population for an intranasal malaria vaccine. Although most standard childhood vaccines are given i.m. or s.c., an i.n. pediatric flu vaccine was recently approved by the FDA (64).
Our analysis of the immunogenicity of flagellin-modified CS recombinant proteins demonstrates that antibodies elicited by i.n. immunization, although of relatively modest titer, can neutralize sporozoite infectivity in vitro and in vivo. These results suggest that functional assays based on transgenic parasites expressing P. falciparum epitopes, rather than static serologic assays such as ELISA, may provide a more biologically relevant measurement of the sporozoite-neutralizing potential of vaccine-induced responses. These studies provide the first demonstration that an i.n. administered TLR5 agonist-modified P. falciparum CS protein can elicit protective antibody-mediated immune responses and support further efforts to investigate needle-free malaria vaccines. While the data are encouraging, the level of parasite inhibition observed in the flagellin-CS-immunized mice would not be expected to provide high levels of sterile immunity, as TSNA titers of 1:40 and higher have been obtained in previous murine studies using parenterally administered CS peptide-protein conjugates (65), while TSNA activities of immune sera from STF2Δ.CS-immunized mice did not exceed a 1:5 dilution (Fig. 9A). Based on analysis of NALT, optimization of adjuvant formulations to improve mucoadhesion and to specifically target NALT antigen-presenting cells (66–68) would be expected to improve the protective efficacy of i.n. delivered flagellin-CS vaccines. Delivery of optimized vaccine formulations i.n. would facilitate large-scale vaccination programs for the control, elimination, and possible eradication of P. falciparum, the most lethal of all human Plasmodium parasites, as well as the other human-infective Plasmodium species (69).
ACKNOWLEDGMENTS
VaxInnate is acknowledged as a source of flagellin and flagellin-modified CS proteins, and we thank A. Price and V. Nakaar for production and purification of the recombinant proteins. We thank Rita Altszuler for excellent technical assistance on all phases of the research and Leyda Cabrera for technical assistance on serological assays. We thank David O'Neill and Nina Bhardwaj for providing cryopreserved human dendritic cells.
This research was supported by NIH/NIAID grant R56 AI 083655 (E.N.) and NIH/NCRR grant S10 RR019288 (U.F.).
FOOTNOTES
- Received 1 March 2013.
- Returned for modification 21 March 2013.
- Accepted 5 September 2013.
- Accepted manuscript posted online 16 September 2013.
Supplemental material for this article may be found at http://dx.doi.org/10.1128/IAI.00263-13.
- Copyright © 2013, American Society for Microbiology. All Rights Reserved.
