Iron oxide nanoparticles as a clinically acceptable delivery platform for a recombinant blood-stage human malaria vaccine

Abstract

This study explored the novel use of iron oxide (IO) nanoparticles (<20 nm) as a vaccine delivery platform without additional adjuvants. A recombinant malaria vaccine antigen, the merozoite surface protein 1 (rMSP1), was conjugated to IO nanoparticles (rMSP1-IO). Immunizations in outbred mice with rMSP1-IO achieved 100% responsiveness with antibody titers comparable to those obtained with rMSP1 formulated with a clinically acceptable adjuvant, Montanide ISA51 (2.7×10⁻³ vs. 1.6×10⁻³; respectively). Only rMSP1-1O could induce significant levels (80%) of parasite inhibitory antibodies. The rMSP1-IO was highly stable at 4°C and was amenable to lyophilization, maintaining its antigenicity, immunogenicity, and ability to induce inhibitory antibodies. Further testing in nonhuman primates, Aotus monkeys, also elicited 100% immune responsiveness and high levels of parasite inhibitory antibodies (55–100% inhibition). No apparent local or systemic toxicity was associated with IO immunizations. Murine macrophages and dendritic cells efficiently (>90%) internalized IO nanoparticles, but only the latter were significantly activated, with elevated expression/secretion of CD86, cytokines (IL-6, TNF-α, IL1-b, IFN-γ, and IL-12), and chemokines (CXCL1, CXCL2, CCL2, CCL3, CCL4, and CXCL10). Thus, the IO nanoparticles is a novel, safe, and effective vaccine platform, with built-in adjuvancy, that is highly stable and field deployable for cost-effective vaccine delivery.—Pusic, K., Aguilar, Z., McLoughlin, J., Kobuch, S., Xu, H., Tsang, M., Wang, A., Hui, G. Iron oxide nanoparticles as a clinically acceptable delivery platform for a recombinant blood-stage human malaria vaccine.

With only a handful of clinically approved vaccine adjuvants available (1, 2), the development of new adjuvants has not been keeping pace with the increasing demand for their use in vaccine formulations. The fact that adjuvants often influence the quality of the immune responses in different ways (1) indicates that there is no single adjuvant formulation that can be universally effective for all vaccines. Thus, new and alternative strategies need to be explored to expand the portfolio of vaccine adjuvants and delivery platforms. One strategy makes use of particle-mediated delivery systems, such as micro- and nanoparticles, in an attempt to improve immunogenicity through targeted antigen delivery and/or presentation (3). Among the different types of particles being evaluated are lipid polymers (e.g., PLGA, PGA, PLA; refs. 4–7), virus-like particles (VLPs; refs. (8, 9), immune-stimulating complexes (ISCOMs; refs. 10, 11), chitosan (12–14), and inorganic particles (15). Several vaccines, such as the hepatitis B vaccine and the human papilloma virus vaccine, have already been developed utilizing these platforms (16, 17). Other examples of experimental use of particles for vaccine delivery are the self-assembling polypeptide-based nanoparticles (SAPNs) for peptide sporozoite malaria vaccine (18), nanolipoproteins for vaccines against West Nile encephalitis (19), and poly(propylene sulfide) nanoparticles for intranasal delivery of peptide antigens to enhance mucosal immune responses (20). A recent immunogenicity study of a malaria circumsporozoite peptide epitope vaccine, presented on self-assembling peptide nanofibers, shows promising results in developing the nanofibers as a multi-epitopes vaccine platform (21).

Few attempts have been made to develop solid inorganic nanoparticles as vaccine platforms, although they have been tested for drug delivery (15). Previous studies indicate that particle size is an important factor in determining the ability of the nanoparticles to achieve effective delivery of the payload (22). Solid nanoparticles of sizes <20 nm can behave as true solutions, which may facilitate rapid dispersion and tissue penetration to reach immunological sites and organs. These can be positive attributes for vaccine delivery. Iron oxide (IO; Fe₂O₃) nanoparticles differing in size (15–180 nm) and surface modifications are currently employed in several medical applications (23). Dextran-coated superparamagnetic iron oxide (SPIO) nanoparticles are used as MRI contrast agents. Another SPIO particle of a smaller size is undergoing clinical trials for lymph node metastasis detection (24). In addition to MRI agents, IO nanoparticles are being used for drug delivery/targeting and hyperthermia (23). The excellent safety profiles, prior uses in drug delivery, and low cost of production of IO have prompted us to examine whether these nanoparticles are effective vaccine delivery platforms. The IOs being evaluated in this study are small (<20 nm), stable, water soluble, and consist of an amphiphilic polymer coating. The surface layer contains carboxylate groups that are readily available for conjugation with proteins, peptides, and DNA (25–27).

We used a recombinant malaria vaccine antigen, the Plasmodium falciparum merozoite surface protein 1–42 (MSP1–42; refered to here as rMSP1), as a model immunogen to evaluate IO nanoparticles as an adjuvant-free vaccine delivery vehicle. MSP1–42 is found on the surface of the invading merozoites during the erythrocytic stage of the malaria life cycle (28, 29) and is one of the most promising and most studied malaria vaccine candidates (30–34). Protective immunity to malaria infections has been correlated with parasite inhibitory antibodies specific for MSP1–42 (32, 33, 35–39). In this study, outbred mice and Aotus monkeys were immunized with rMSP1 conjugated to IO (rMSP1-IO). Results showed that rMSP1-IO was as effective in enhancing immunogenicity as rMSP1 administered with a clinically acceptable adjuvant, Montanide ISA51. Moreover, rMSP1-IO induced parasite inhibitory antibodies in more than one animal species. Preliminary toxicity studies in mice and monkeys showed no significant deviations from normal values. Equally significant is the finding that the rMSP1-IO formulation was very stable in solution and was also amenable to lyophilization with no loss in antigenicity and immunogenicity. Lastly, we investigated the effects of IO uptake by dendritic cells and macrophages as the possible mode of action in enhancing vaccine-induced immune responses; and provided evidence that the IO nanoparticles have built-in immunomodulating properties.

MATERIALS AND METHODS

Mouse and nonhuman primates

Outbred Swiss Webster (SW) mice and C57Bl/6 mice (female, 6–8 wk old) were obtained from Charles River Laboratory (Wilmington, MA, USA). Aotus lemurinus trivirgatus, karyotype II and III, adult monkeys (1 female and 3 males, ages 12–15 yr) were colony born and raised at the University of Hawaii Nonhuman Primate Facility. Use of all animals was approved by the University of Hawaii Institutional Animal Care and Use Committee.

rMSP1

A truncated version of MSP1–42, rMSP1, based on the FUP (Falciparum Uganda Palo-Alto) strain was expressed in Drosophila cells (40) and purified by affinity chromatography (41). Figure 1A shows the SDS-PAGE profile of the purified protein. The rMSP1 has been shown to induce parasite inhibitory antibodies (42).

Figure 1.

Purification and conjugation of rMSP1 recombinant protein to IO nanoparticles. A) SDS-PAGE gel of purified rMSP1 protein. Lane 1: molecular marker; lane 2: purified rMSP1 recombinant protein. B) Agarose gel electrophoresis of unconjugated IO nanoparticles (lane 1) and rMSP1-conjugated IO nanoparticles (lane 2).

Conjugation of rMSP1 to IO nanoparticles

The rMSP1-IO conjugates were prepared using N-hydroxysulfosuccinimide sodium salt (sulfo-NHS) and 1-ethyl-3-(3-dimethylaminopropyl) carbodiimide (EDC) covalent coupling chemistry. IOs with carboxyl groups on the surface (5 mg/ml) were activated by incubating with sulfo-NHS (molar ratio 2000:1) and EDC (molar ratio 2000:1) for 5 min in borate buffer (pH 7.4), after which 2 mg of rMSP1 was added, vortexed thoroughly, and incubated for 2 h at room temperature. Following incubation, the reaction was quenched by adding 5 μl of Ocean's quenching buffer (Ocean NanoTech, Springdale, AR, USA), mixed, and incubated for 10 min at room temperature. The rMSP1-IO conjugates were then purified/separated by using a SuperMag separator (Ocean NanoTech) for 10–24 h.

The rMSP1-IO conjugates and unconjugated IOs were evaluated by agarose (1.5%) gel electrophoresis in tris-acetate-EDTA (TAE) buffer (pH 8.5). For each well, 20 μl of IO samples at 100 nM were mixed with 5 μl of 5×TAE loading buffer: 5×TAE, 25% (v/v) glycerol, and 0.25% (w/v) Orange-G at pH 8.5. The samples were resolved at 100 V for 30 min (PowerPak Basic; Bio-Rad, Hercules, CA, USA) and then imaged with two exposures using a gel imaging system (Alpha Imager HP 2006; Alpha Innotech, San Leandro, CA, USA; Fig. 1B).

Lyophilization of IO nanoparticles

Cryoprotectant, trehalose, was added to the purified and separated rMSP1-IO nanoparticles, and the solution was frozen at −70°C. The rMSP1-IO was then lyophilized with the Labconco Freezone 2.5 (Labconco Corp., Kansas City, MO, USA). Lyophilized rMSP1-IO was rehydrated with deionized water to obtain a concentration of 5 mg/ml. Analysis of the same batch of material before and after lyophilization at identical concentrations was evaluated by agarose (1%) gel electrophoresis in TAE buffer (pH 8.5) and imaged as described above (see Fig. 5A).

Figure 5.

Antigenicity and immunogenicity of lyophilized rMSP1-IO. A) Agarose gel electrophoresis of nonlyophilized rMSP1-IO nanoparticles (lane 1) and lyophilized rMSP1-IO nanoparticles (lane 2). B) Antigenicity of lyophilized rMSP1-IO nanoparticles. ELSIA titration curves of nonlyophilized rMSP1-IO (open circles) and lyophilized rMSP1-IO nanoparticles (solid circles) against MSP1-42 specific monoclonal antibody, MAb 5.2. Straight line represents OD reading of MAb 5.2 reactivity to native MSP1-42 at a coating concentration of 0.4 μg/ml. C) Antibody response against MSP1-19 in SW mice immunized with nonlyophilized and lyophilized rMSP1-IO via the i.p. route. Results of tertiary bleed are shown. D) IL-4 (solid circles) and IFN-γ (open circles) responses as determined by ELISPOT of splenocytes from mice immunized with nonlyophilized and lyophilized rMSP1-IO. Horizontal lines indicate mean SFU.

Antigenicity of rMSP1 conjugated to IO nanoparticles

The antigenicity of the conjugated rMSP1 was evaluated by ELISA reactivity with MAb5.2, a conformational sensitive monoclonal antibody specific for MSP1-19 (43). Freshly prepared rMSP1-IO was stored at 4°C and aliquots were sampled at 0, 6, 12, and 18 mo for analyses. In addition, rMSP1-IO samples taken before and after lyophilization were also analyzed. Serial dilutions of each aliquot sample of the rMSP1-IO were used for coating ELISA plates. MAb 5.2 was used at a concentration of 0.2 μg/μl in 1% yeast extract and 0.5% BSA in BBS. Horseradish peroxidase (HRP)-conjugated anti-mouse antibodies (H & L chain specific; Kirkgaard and Perry Laboratories, Gaithersburg, MD, USA) at a dilution of 1:2000 were used as the secondary antibody conjugate. Color development was made using the peroxidase substrates, H₂O₂ and 2.2′-azinobis(3-ethylbenzthiazolinesulfonic acid)/ABTS (Kirkgaard and Perry Laboratories). Optical density (OD) was determined at 405 nm. ODs for each serial dilution were plotted, and the levels of reactivity were compared to the reactivity of MAb 5.2 against serially diluted, unconjugated rMSP1.

Immunizations with rMSP1-IO

Groups of outbred SW mice (n=6) were immunized with rMSP1-IO via the intraperitoneal (i.p.), intramuscular (i.m.), and subcutaneous (s.c.) routes. The injection volume for the i.p. and s.c. routes was 100 μl/dose (16 μg/dose), and for the i.m. route was 20 μl/dose (5 μg/dose). SW mice were also immunized with rMSP1-IO preparations before and after lyophilization via the i.p. route (100 μl/dose, 16 μg/dose). In addition, mice were immunized via the i.p. route with rMSP1 emulsified in either complete Freund's adjuvant (CFA), incomplete Freund's adjuvant (IFA), or Montanide ISA51 (43). Mice were immunized 3 times at 21-d intervals, as described previously (44). Sera were obtained through tail bleeds on the 14th day after each immunization.

A. lemurinus trivirgatus monkeys were likewise immunized with rMSP1-IO, 0.5 ml/dose (80 μg antigen/dose), via the i.m. route. Immunizations were administered 3 times at 21-d intervals, alternating the right and left thigh. Sera were collected 21 d after the last immunization for ELISAs and parasite growth inhibition assays (33).

MSP1-specific antibody assays

Mouse and monkey sera were assayed for anti-MSP1 antibodies (MSP1–42 and MSP1–19 specific) by direct binding ELISA, as described previously (33, 45). The MSP1–19 and MSP1–42 used for coating ELISA plates were expressed in yeast (46) and in baculovirus (41), respectively. Plates were coated with these antigens at a concentration of 0.4 μg/ml. Sera were serially diluted in 1% yeast extract, 0.5% BSA in borate-buffered saline (BBS). HRP-conjugated anti-mouse antibodies (H & L chain specific; Kirkgaard and Perry Laboratories) were used as the secondary antibody conjugate at a dilution of 1:2000; and HRP-conjugated anti-Aotus antibodies, graciously provided by Hawaii Biotech Inc. (Aiea, HI, USA), were used at a dilution of 1:16,000. Color development was performed by using the peroxidase substrates, H₂O₂ and 2.2′-azinobis(3-ethylbenzthiazolinesulfonic acid)/ABTS (Kirkgaard and Perry Laboratories). OD was determined at 405 nm. Endpoint titers were calculated using the serum dilutions that gave an OD reading of 0.2, which is >4-fold of background absorbance using preimmune mouse or monkey serum samples.

IFN-γ and IL-4 ELISPOT assays

ELISPOT assays of splenocytes from immunized mice were performed according to methods previously described (47). Briefly, 96- well PVDF plates (Millipore Inc., Bedford, MA, USA) were coated with 10 μg/ml of monoclonal antibodies (mAbs) against IFN-γ (R4-642) and 5 μg/ml of mAb against IL-4 (11B11) (BD Biosciences, San Diego, CA, USA), and incubated overnight at room temperature. Plates were washed with phosphate-buffered saline (PBS) and blocked with 10% fetal bovine serum (FBS) in DMEM for 60 min. Mouse spleens were harvested, and single-cell suspensions of splenocytes were prepared, as described previously (47). Purified splenocytes were plated at 0.5 × 10⁶, 0.25 × 10⁶, and 0.125 × 10⁶ cells/well, and rMSP1 expressed in Drosophila cells (20 μg/ml) was added to each well as the stimulating antigen. Positive-control wells were incubated with 5 ng/ml of phorbol myristate acetate (PMA) and 1 ng/ml of ionomycin (Sigma-Aldrich, St. Louis, MO, USA). Plates were incubated at 37°C in 5% CO₂ for 48 h. Wells were washed and incubated with biotinylated mAb against IFN-γ at 2 μg/ml (XMG1.2), or mAbs against IL-4 at 1 μg/ml (BVD6-24G2) (BD Biosciences), followed by the addition of peroxidase-conjugated streptavidin (Kirkgaard and Perry Laboratories) at a dilution of 1:800. Spots were developed with a solution consisting of 3,3′-diaminobenzidine tetrahydrochloride (DAB; Sigma-Aldrich, 1 mg/ml) and 30% H₂O₂ (Sigma-Aldrich) and enumerated microscopically. Data are presented as spot-forming-units (SFU) per million plated splenocytes.

In vitro parasite growth inhibition assay

The ability of the mouse and monkey sera generated by immunizations with rMSP1-IO to inhibit parasite growth was determined using an in vitro assay, as described previously (48, 49). For testing of mouse serum samples, immunoglobulins from pooled mouse serum samples from each group were purified as described previously (48). Briefly, antibodies were purified by ammonium sulfate precipitation followed by dialysis using the Amicon Ultra-10 Centrifugal Filter units (Millipore) with a molecular mass cutoff of 100 kDa. Purified antibody samples were reconstituted to original serum volume with RPMI 1640 medium and were used at a 20% serum concentration. For testing monkey samples, individual serum samples were heat inactivated, absorbed with normal red blood cells (RBCs), and used at a 30% final serum concentration (49). Inhibition assays were performed using sorbitol synchronized parasite cultures (3D7 strain) as described previously (50). Synchronized parasite cultures at a starting parasitemia of 0.2% and 0.8% hematocrit were incubated in antibody or serum samples for 72 h with periodic mixing. Culture parasitemias were determined microscopically by Giemsa staining of thin blood smears, and the degree of parasite growth inhibition was determined by comparing the parasitemias of immune sera with the corresponding preimmune sera, as described previously (51, 52).

Toxicity studies on IO-immunized mice and Aotus monkeys

Mice were divided into 4 groups (n=6) receiving escalating doses of IO nanoparticles ranging from 1.1 to 4.4 mg via i.p. injection; with one control group receiving no IO nanoparticles (saline control). Mice in each group were immunized 3 times, at 21-d intervals, with their corresponding IO doses. All mice were bled prior to IO immunizations and after the third IO immunization. Blood samples were directly placed into i-STAT EG6+ cartridges and measured by the i-STAT Blood Analyzer (Abbott Diagnostic Laboratories, Abbott Park, IL, USA) for hematocrit, hemoglobin, blood urea nitrogen, anion gap, carbon dioxide, and potassium (Supplemental Table S1). Aotus monkeys immunized with rMSP1-IO were bled prior to and after the third booster immunization for blood chemistry tests. Whole-blood and serum samples were sent to IDEXX Laboratories Inc. (Westbrook, ME, USA), where blood chemistry tests were performed. A detailed panel of tests performed is listed in Supplemental Table S2.

Isolation of murine dendritic cells and macrophages

Immature bone marrow cells were isolated from 12-to 14-wk-old C57Bl/6 mice, as described previously (53). Stromal cells were purified by passage through a cell strainer to remove bones and debris. RBC lysis buffer consisting of 0.15 M NH₄Cl, 10 mM KHCO₃, and 0.1 mM EDTA was used to remove RBCs. After washings, bone marrow cells were plated in 6-well plates (Cell Star, Monroe, NC, USA) at a density of 10⁶ cells/ml together with either GM-CSF (Peprotech Inc., Rocky Hill, NJ, USA) at a concentration of 20 ng/ml or with M-CSF (eBioscience, San Diego, CA, USA) at a concentration of 10 ng/ml. After 24 h, cell cultures were incubated in RPMI 1640 with GM-CSF for an additional 8 d for differentiation into bone marrow dendritic cells (BMDCs) or incubated for an additional 6 d in DMEM with M-CSF for differentiation into macrophages (54). On d 8, BMDCs in suspension were transferred to new plates and used as the cell source for all subsequent experiments (55). Experiments were also performed using macrophages from d 6 cultures (54).

IO uptake by dendritic cells and macrophages

Unconjugated IO nanoparticles were introduced at a final concentration of 4 μM to the 8-d-old BMDCs or 6-d-old macrophages, and incubated for 24 h at 37°C. To first visualize the uptake of IO nanoparticles, BMDCs and macrophages were fixed with 4% paraformaldyhide (PFA) and stained with Prussian blue (Biopal, Worcester, MA, USA) according to the manufacturer's protocol (http://www.biopal.com/Molday%20ION.htm). The same cells were then stained for surface markers with biotinylated, anti-CD11c or anti-CD11b antibodies (eBioscience) at a dilution of 1:2000 for 1 h. After washings, cells were further incubated with streptavidin-QDots (1:2000; Ocean NanoTech), which have an emission wavelength of 620 nm, for an additional hour for cell type identification and purity assessment. Cells were imaged using a fluorescent microscope (Olympus ix71; Olympus, Tokyo, Japan) with a fluorescent cube containing the following filters: V-N41004 (λ_ex 560 and λ_em 585) and V-N41001 (λ_ex 480 and λ_em 535).

Dendritic cell and macrophage activation by IO

Unconjugated IO nanoparticles (5 mg/ml) were introduced to 7-d-old BMDCs (55) or 6-d-old macrophages for 24 h at 37°C. The cells were harvested and washed twice with FACS buffer (PBS with 2% FBS) and fixed with 0.25% PFA for 10 min on ice. Cells were separated by passing through a magnetic LD column (Miltenyi Biotec Inc., Auburn, CA, USA) to obtain an enriched population of cells that have taken up the IO nanoparticles (∼98%). BMDCs and macrophages were stained with one or more of the following fluorescent dye conjugated antibodies to the cell surface markers: APC-labeled anti-CD80, PE-labeled anti-MHC II, AlexaFluor488-labeled anti-CD11c, or AlexaFluor488-labeled anti-CD11b (eBiosciences), and PE-Cy7-labeled anti-CD86 (Invitrogen, Carlsbad, CA, USA). Labeled cells were analyzed using the FACSAria flow cytometer with FACSDiva software (Becton Dickinson, San Jose, CA, USA).

Cytokine gene expression by IO-stimulated dendritic cells and macrophages

BMDCs and macrophages (3×10⁶ cells) were stimulated with unconjugated IO (4 μM) or LPS (100 ng/ml), and RNA was extracted at 0, 3, 6, and 12 h using the RNeasy Kit (Qiagen, Valencia, CA, USA). RNA concentrations were measured and then reversed transcribed in 50-μl reactions using the iScript cDNA synthesis kit (Bio-Rad) following manufacturer's protocol. Real-time PCR reactions using iQ SYBR Green Supermix (Bio-Rad) were run on the MyiQ Single-Color Real Time Detection System (Bio-Rad). Primers for TNF-α, TGF-β, IL-12, IL-6, IFN-γ, and IL-1β were used at 10 nM (IDT, Coralville, IA, USA). Primer sequences are provided in Supplemental Table S3. Analysis of gene expression was performed by the ΔΔC_t method (56, 57). Briefly, each sample was normalized to an endogenous control, GAPDH, and fold change for each assayed gene was determined via the ΔΔC_t.

Multiplex assay for cytokine detection

Supernatants from IO- and LPS-stimulated BMDCs were tested for the presence of cytokines and chemokines over a 12 h period. Cytokines and chemokines were measured using the Milliplex MAP Mouse Cytokine/Chemokine 32-plex assay (Millipore) according to manufacturer's protocol. The following cytokines/chemokines were measured: eotaxin, G-CSF, GM-CSF, IFN-γ, IL-10, IL-12 (p40), IL-12 (p70), IL-13, IL-15, IL-17, IL-1α, IL-1β, IL-2, IL-2, IL-4, IL-5, IL-6, IL-7, IL-9, IP-10, KC-like, LIF, LIX, M-CSF, MCP-1, MIG, MIP-1α, MIP-1β, MIP-2, RANTES, TNF-α, and VEGF.

Data handling and statistics

SigmaPlot 10 (Systat Software, Inc., San Jose, CA, USA) and GraphPad Prism 4 (GraphPad, San Diego, CA, USA) were used to calculate the endpoint titers. The Mann-Whitney test was used to determine significant differences in antibody responses, and the expression of cell surface activation markers among the test groups. A value of P < 0.05 was considered statistically significant.

RESULTS

Antigenicity of rMSP1-conjugated IO nanoparticles

To determine whether rMSP1 was successfully conjugated to the IO nanoparticles, unconjugated and conjugated IOs were analyzed by agarose gel electrophoresis (Fig. 1B). The rMSP1-IO sample (Fig. 1B, lane 2) migrated as a single band and at a higher molecular mass than the unconjugated IO sample (Fig. 1B, lane 1), indicating that the conjugation process had successfully produced a homogeneous species of rMSP1-IOs. To evaluate whether the chemical conjugation process affected the antigenicity and stability of rMSP1, the reactivity of a conformation-dependent anti-MSP1–19 monoclonal antibody, MAb 5.2 with rMSP1-IO, was tested. MAb 5.2 strongly reacted with the rMSP1 conjugated to IO nanoparticles but did not recognize the unconjugated IO particles (Fig. 2A). As a reference, an OD reading of 1.3 was observed with MAb 5.2 incubated with unconjugated rMSP1–42 at a plating concentration of 0.4 μg/ml. This suggests that the antigenicity of the rMSP1 antigen was preserved during the conjugation process. The conjugated nanoparticles stored at 4°C were tested over a period of 18 months for any loss of antigenicity of the rMSP1. The rMSP1-IO was equally reactive with MAb 5.2 at 6, 12, and 18 mo postconjugation (Fig. 2B), demonstrating the stability of these conjugated IO nanoparticles.

Figure 2.

Antigenicity analysis of rMSP1 protein conjugated to IO nanoparticles. A) Antigenicity of rMSP1-conjugated IO nanoparticles. ELSIA titration curves of rMSP1 conjugated nanoparticles (open circles) and unconjugated nanoparticles (solid circles) against MSP1-42 specific monoclonal antibody, MAb 5.2. Straight line represents OD reading of MAb 5.2 reactivity to native MSP1-42 at a coating concentration of 0.4 μg/ml. B) Antigenicity of rMSP1-conjugated IO nanoparticles over an 18-mo period. ELISA titration curves of rMSP1 conjugated nanoparticles at 0 mo (open circles), 6 mo (solid circles), 12 mo (open squares), and 18 mo (solid triangles) postconjugation against MSP1-42 specific monoclonal antibody, MAb 5.2.

Immunogenicity of rMSP1-IO nanoparticles in SW mice

The immunogenicity of rMSP1-IO was compared to conventional adjuvants. SW mice were immunized with rMSP1 conjugated to IO nanoparticles, or formulated with CFA or Montanide ISA51. SW mice were specifically utilized in these studies in order to evaluate the immunogenicity of rMSP1-IO in a more genetically heterogeneous population. Immune sera were tested for antibodies against MSP1–19 by ELISA. Vaccine responders were defined as having an ELISA OD > 0.2 at a 1:50 serum dilution (42, 43). This was above the OD values observed for preimmune mouse sera. The rMSP1-IO induced an antibody response in all mice after 3 immunizations, resulting in a 100% response rate. The same response rate was observed with mice immunized with rMSP1-CFA. However, only 5 of 10 mice immunized with rMSP1-ISA51 responded, resulting in a 50% response rate (Fig. 3A).

Figure 3.

ELISA antibody response against MSP1-19 in SW mice immunized with rMSP1. A) Antibody titers of SW mice immunized with different adjuvant/delivery platforms (rMSP1-IO, rMSP1-CFA, rMSP1-ISA51). Results of the tertiary bleed are shown. B) Antibody response in mice vaccinated with rMSP1-IO via different immunization routes (i.p., i.m., s.c.). Results of tertiary bleed are shown. Significant differences in antibody titers among the vaccination groups are shown with P values (Mann-Whitney test).

Comparisons of antibody endpoint titers of tertiary bleeds among the 3 immunization groups show that rMSP1-IO induced a mean antibody titer of 2.7 × 10⁻³ (Fig. 3A), whereas the ISA51 formulation induced a lower mean antibody titer of 1.6 × 10⁻³ (P=0.012). The potent CFA formulation induced the highest mean antibody titer of 2.8 × 10⁻⁴; however, this level was not significantly higher than the rMSP1-IO group (Fig. 3A). Immunizations using CFA and ISA51 induced high and low responders within the group of immunized outbred mice, as reflected in the broad range of endpoint titers. Encouragingly, rMSP1-IO induced a more uniform antibody response (Fig. 3A).

Mice were also immunized with rMSP1-IO via the i.m. and s.c. routes. The mean antibody titers induced by i.m. immunization were higher than those induced by i.p. or s.c. immunizations (Fig. 3B), but the differences were not statistically significant (Fig. 3B). Immunizations via the i.m. and i.p. routes achieved a 100% response rate, whereas s.c. immunizations resulted in a 60% response rate (Fig. 3B).

Sera from rMSP1-IO immunized mice were also tested for their ability to inhibit parasite growth in vitro. Inhibition >50% was considered to be biologically significant (48, 58). As shown in Table 1, antibodies obtained from rMSP1-IO immunizations via the i.p. and i.m. routes significantly inhibited parasite growth at 80 and 74%, respectively. In comparison, antibodies from mice immunized with rMSP1 emulsified with CFA and ISA51 were both ineffective in inhibiting parasite growth (Table 1). In addition, IO immunization via the s.c. route was also ineffective at a 37% parasite growth inhibition (Table 1).

Table 1.

In vitro parasite growth inhibition of purified mouse anti-MSP1 antibodies

Pooled mouse purified antibody, tertiary bleed	Parasite growth inhibition (%)
rMSP1-IO, i.p.	80
rMSP1-IO, i.m.	74
rMSP1-IO, s.c.	37
rMSP1-CFA, i.p.	17
rMSP1-ISA51, i.p.	0
Lyophilized rMSP1-IO, i.p.	79

Values are means of 2 growth inhibition assays.

IO-immunized mice showed a predominant IL-4 cellular response

ELISPOT of splenocytes from mice immunized with rMSP1-IO via the i.m., s.c., and i.p. routes showed higher production of IL-4 as compared to IFN-γ (Fig. 4), indicative of a TH2 type response. The rMSP1/ISA51 and rMSP1-IO groups produced significantly higher IL-4 responses than that observed with rMSP1-CFA (Fig. 4A). Of the rMSP1-IO immunizations, i.p. and s.c. delivery were especially effective, and gave significantly higher IL-4 responses than i.m. injections (P=0.015 and P=0.014, respectively; Fig. 4A).

Figure 4.

MSP1-specific IL-4 and IFN-γ responses in SW mice immunized with rMSP1 in different delivery platform/adjuvants. IL-4 (A) and IFN-γ (B) responses as determined by ELISPOT of splenocytes from immunized mice. Horizontal lines indicate mean SFU. PMA/ionomycin-positive wells have mean SFU of 4 × 10⁸ cells. Significant differences in SFUs among the vaccination groups are shown with P values (Mann-Whitney test). *P < 0.05 vs. all other groups.

Antigenicity and immunogenicity of lyophilized rMSP1-IO nanoparticles

To determine whether the rMSP1-IO was successfully lyophilized, nonlyophilized and lyophilized/rehydrated rMSP1-IOs were analyzed by agarose gel electrophoresis (Fig. 5A). Both the nonlyophilized rMSP1-IO sample (Fig. 5A, lane 1) and the lyophilized/rehydrated rMSP1-IO sample (Fig. 5A, lane 2) migrated to the same position and the migrating bands had the same sharpness, indicating that lyophilization maintained the size and charge of rMSP1-IO nanoparticles. To evaluate whether the lyophilization process affected the antigenicity of rMSP1-IO, the reactivity of MAb 5.2 with lyophilized rMSP1-IO was tested. MAb 5.2 reacted equally well with the lyophilized rMSP1-IO as with the nonlyophilized rMSP1-IO, suggesting that the antigenicity was preserved during the lyophilization process (Fig. 5B).

The immunogenicity of the lyophilized rMSP1-IO was compared to nonlyophilized rMSP1-IO in SW mice. Lyophilized rMSP1-IO induced an antibody response in all mice, resulting in a 100% response rate, which was identical to the response rate of nonlyophilized rMSP1-IO (Fig. 5C). Comparisons of antibody endpoint titers between the two vaccination groups showed no significant differences (Fig. 5C). Likewise, ELISPOT of splenocytes from mice immunized with nonlyophilized and lyophilized rMSP1-IO showed no significant differences, both in terms of magnitude and IL-4 vs. IFN- γ responses (Fig. 5D). Most important, similar high levels of parasite growth inhibitory antibodies were induced by the lyophilized rMSP1-IO as compared to the nonlyophilized rMSP1-IO (Table 1).

Immunogenicity of rMSP1-IO nanoparticles in Aotus monkeys

All four Aotus monkeys immunized with rMSP1-IO produced anti-MSP1–42 and anti-MSP1–19 antibodies (Table 2), with endpoint titers specific for MSP1–42 ranged from 1/2800 to 1/29,000; and those specific for MSP1–19 ranged from 1:3000 to 1:24,000 (Table 2). Sera from Aotus monkeys immunized with rMSP1-IO were also evaluated for inhibition of parasite growth as above (49). All immunized monkeys produced significant levels of parasite growth inhibitory antibodies, ranging from 55 to 100% inhibition (Table 2). This level of inhibition is comparable to studies where Aotus monkeys were vaccinated and protected with MSP1–42-CFA (33). Due to the old age of these monkeys, they were not challenged with blood-stage parasites.

Table 2.

In vitro parasite growth inhibition of monkey anti-MSP1 serum samples

Monkey serum, tertiary bleed	Anti-MSP1-42 antibody titers	Anti-MSP1-19 antibody titer	Parasite growth inhibition (%)
Monkey 1	1:2800	1:3000	82
Monkey 2	1:29,000	1:24,000	100
Monkey 3	1:4500	1:10,000	55
Monkey 4	1:10,000	1:20,000	66

Toxicity studies showed no abnormalities in IO-immunized animals

Escalating doses of IO nanoparticles, up to 4.4 mg/injection, did not cause any abnormalities or changes in the blood chemistries in all four groups of mice tested after each of the three immunizations (Supplemental Table S1). Similarly, a more comprehensive test panel of blood chemistry levels in the Aotus monkeys after three rMSP1-IO immunizations revealed no significant deviations from normal ranges (Supplemental Table S2). Thus, immunization with IO nanoparticles did not have toxic systemic affects in either animal model.

Uptake of IO nanoparticles by dendritic cells and macrophages

IO nanoparticles were introduced to 7-d-old BMDC cultures and to 6-d-old macrophage cultures. BMDCs and macrophages both actively internalized the IO nanoparticles, as shown in Fig. 6. BMDCs were identified by staining for the surface marker, CD11c (Fig. 6A), and the presence of internalized IO particles was identified by Prussian Blue staining (Fig. 6B). Macrophages were identified by staining for the surface marker, CD11b (Fig. 6C), and the presence of internalized IO nanoparticles was also revealed by Prussian Blue staining (Fig. 6D). Approximately 89% of the BMDCs and ∼94% of the macrophages internalized the IO nanoparticles. Furthermore, no differences were found in the amount of IO nanoparticles taken up by BMDCs and by macrophages based on microscopic examination of these cell cultures.

Figure 6.

Uptake of IO nanoparticles by BMDCs and macrophages. A) Surface staining of BMDCs with QDot-labeled anti-CD11c (red). B) Localization of IO nanoparticles (blue) in the same BMDCs. C) Surface staining of macrophages with QDot-labeled anti-CD11b (red). D) Localization of IO nanoparticles (blue) in the same macrophages.

Dendritic cell and macrophage activation by IO nanoparticles

Unconjugated IO nanoparticles were introduced to immature BMDCs and macrophages, and the degree of activation was determined by cell surface expression of CD86 and CD80 using flow cytometry (43). Unstimulated, IO-stimulated, and LPS-stimulated dendritic cells were first gated for the presence of CD11c, and the CD11c⁺ cells were analyzed for the expression of the activation marker, MHC II, CD86, and CD80. IO-stimulated, CD11c⁺ dendritic cells (Fig. 7A, top center panel) were activated and showed an increase in expression of MHC II (Fig. 7A, middle center panel), CD86, and CD80 (Fig. 7A, bottom center panel). IO-stimulated dendritic cells had the highest percentage of MHC II marker (34%) and CD80 marker (28%) as compared to unstimulated dendritic cells (28 and 22%, respectively). However, these increases did not reach statistical significance (Fig. 7B). Percentages of CD86⁺ cells and CD80⁺/CD86⁺ cells were significantly higher than those observed for unstimulated dendritic cells, with P = 0.05 and P = 0.03, respectively (Fig. 7B). LPS-stimulated dendritic cells had significantly higher percentage of CD86⁺, and CD80⁺/CD86⁺ cells than IO-stimulated dendritic cells (P=0.05 and 0.04, respectively; Fig. 7B).

Figure 7.

Activation of BMDCs and macrophages by IO nanoparticles. A) Flow data from a representative experiment. BMDCs (1×10⁶ cells) were incubated with medium alone, IO nanoparticles, or LPS at 37°C for 24 h. Cells were stained for surface markers CD11c-FITC, MHC II-PE, CD80-APC, and CD86-PE-Cy7 and analyzed by flow cytometry. Top panels are gated for CD11c⁺ cells; middle panels are live gates of the top panels and are stained for MHC II; bottom panels are live gates of the top panels and are stained for CD80, and CD86. B) Summary of the activation marker expression from 3 BMDC-activation experiments. C) Summary of activation maker expression from 3 macrophage-activation experiments. *P ≤ 0.05 between unstimulated vs. IO-stimulated BMDCs for expression of CD86⁺ (P=0.05) and CD80⁺/CD86⁺ (P=0.03); **P ≤ 0.05 between IO-stimulated and LPS-stimulated dendritic cells for expression of CD86⁺ (P=0.05) and CD80⁺/CD86⁺ (P=0.04); ***P ≤ 0.05 between unstimulated and LPS-stimulated macrophages for expression of CD86⁺ (P=0.05) and CD80⁺/CD86⁺ (P=0.03).

Unstimulated, IO-stimulated, and LPS-stimulated macrophages (CD11b⁺) were similarly analyzed for the activation markers as above. IO-stimulated macrophages did not significantly up-regulate any of the markers as compared to the unstimulated macrophages (Fig. 7C). However, LPS-stimulated macrophages expressed significantly higher levels of CD86 and CD80/CD86 than unstimulated cells (P=0.05 and 0.03, respectively; Fig. 7C).

IO uptake induced proinflammatory cytokine and chemokine production by BMDCs, but not macrophages

Immature BMDCs were exposed to IO nanoparticles over a 12-h period, and the expression of several cytokines, IL-6, IL-12, IL-1b, TNF-α, IFN-γ, and TGF-β, were monitored by RT-PCR. IO nanoparticles significantly increased the production of IL-6, TNF-α, IL1-b, IFN-γ, and IL-12 by >2-fold in BMDCs compared to baseline, i.e., 0 h (Fig. 8A). In particular, IL-6 and TNF-α were highly expressed (Fig. 8A). LPS-stimulated BMDCs induced significant expression of all cytokines assayed,, with the exception of TGF-β (Fig. 8B). In general, the cytokine expression profiles of LPS- and IO-stimulated BMDCs were similar.

Figure 8.

Cytokine expression of IO-stimulated BMDCs and macrophages. RT-PCR quantification of s6 different cytokine gene expressions in IO-stimulated BDMCs (A), LPS-stimulated BMDCs (B), IO-stimulated macrophages (C), and LPS-stimulated macrophages (D). Expression was monitored over a 12-h period. Data were normalized to GAPDH; fold changes were calculated based on 0 h samples.

Expression of cytokines in IO-stimulated, bone marrow-derived macrophages was much more transient and modest as compared to the IO-stimulated BMDCs (Fig. 8C). Significant levels of IL1-b, IL-6, and TNF- α were only detected in the first 3–6 h after IO stimulation (Fig. 8C), and they were at lower levels as compared to IO-stimulated BMDCs. The low levels of activation are not a result of an inherent defect of the cultured macrophages to respond to immune stimuli, since the cytokine expression profile of LPS-stimulated macrophages was similar to LPS-stimulated BMDCs (Fig. 8B, D).

To broaden our detection targets, a 32-plex Luminex assay was performed to test for cytokine and chemokine production. BMDCs stimulated with either IO nanoparticles or LPS were found to secrete cytokines (Fig. 9) and chemokines (Fig. 10) over a 12-h time course. In comparison to medium alone, IO-stimulated BMDCs produced higher levels of the proinflammatory cytokines, IL-1a, IL-1b, TNF-α, and IL-6 (Fig. 9). A number of chemokines were also produced as a result of IO stimulation, including CXCL1, CXCL2, CCL2, CCL3, CCL4, and CXCL10 (Fig. 10). Among them, CCL4 reached the same levels as LPS-stimulated BMDCs; and CCL3, CXCL10, and CCL2 reached levels close to those produced by LPS-stimulated BMDCs at 12 h (Fig. 10). In general, gradual increases in both cytokine and chemokine levels were observed over time with IO-stimulated BMDCs.

Figure 9.

Cytokine production by stimulated BMDCs. BMDCs (1×10⁶ cells) were incubated with medium alone (open squares), with IO nanoparticles (open circles), or with LPS (100 ng/ml; open triangles). Culture supernatants were assayed for the presence of the cytokines, IL-1a (A), IL-1b (B), TNF-α (C), and IL-6 (D) at 0, 3, 6, and 12 h by Luminex using the Milliplex MAP Mouse Cytokine/Chemokine 32 plex assay. Cell supernatants were measured in triplicates; only the cytokines with the highest expression are depicted.

Figure 10.

Chemokine production by stimulated BMDCs. BMDCs (1×10⁶ cells) were incubated with medium alone (open squares), with IO nanoparticles (open circles), or with LPS (100 ng/ml; open triangles). Culture supernatants were measured for the presence of the chemokines CXCL1 (A), CXCL2 (B), CCL3 (C), CCL4 (D), CXCL10 (E), and CCL2 (F) at 0, 3, 6, and 12 h by Luminex using the Milliplex MAP Mouse Cytokine/Chemokine 32 plex assay. Cell supernatants were measured in triplicates; only the chemokines with the highest expression are depicted.

DISCUSSION

Previously, we tested the use of solid, water-soluble, quantum dot (QD) nanoparticles as a vaccine delivery of rMSP1 (43). Although the QDs were effective in enhancing the immunogenicity of the rMSP1, the potential toxicity of the Cd-based core composition of the QD particles prohibits further studies as a clinically acceptable vaccine platform. The excellent safety profiles of IOs in clinical applications and low cost of production make these particles an attractive candidate for vaccine delivery. Our study here illustrates several key attributes, which further promote the IO nanoparticles as a novel, effective, and readily deployable vaccine platform.

Immunizations in mice demonstrated the effectiveness of these inorganic IO nanoparticles for the delivery of a recombinant blood stage malaria vaccine, rMSP1. Antibody levels induced by rMSP1-IO were equivalent to those induced by CFA, a highly potent but toxic adjuvant. They also surpassed the antibody responses induced by the low-toxicity adjuvant, Montanide ISA51. Of equal significance was the rMSP1-IO's ability to elicit a 100% response rate in outbred mice. The same degree of generalized responsiveness and antibody response could only be achieved with the toxic CFA adjuvant. As comparison, rMSP1-ISA51 induced only a 50% response rate (Fig. 3A).

The route of immunization is known to affect vaccine-induced immune responses (3, 59). Our results showed that the potency of rMSP1-IO as measured by antibody titers was independent of the delivery route (Fig. 3B). However, rMSP1-IO delivered via the s.c. route did not achieve a 100% response rate as observed with i.p. and i.m. injections. This may be a result of the anatomical differences with respect to the fate of the IO nanoparticles once injected. In addition to outbred mice, we further demonstrated the effectiveness of the rMSP1-IO formulation in a nonhuman primate model, Aotus monkeys, achieving a 100% response rate in the animals.

Prior studies have demonstrated that the levels of parasite-inhibitory anti-MSP1 antibodies correlate with natural and vaccine induced immunity (32, 33, 35–39). Thus, the presence of anti-MSP1 inhibitory antibodies in the rMSP1-IO vaccinated mice and monkeys were investigated as a measure of in vitro efficacy. The rMSP1-IO was highly effective in inducing parasite inhibitory antibodies in mice; whereas rMSP1-CFA and rMSP1-ISA51 were ineffective even though the rMSP1-IO and rMSP1-CFA formulations induced similar levels of antibodies (Fig. 3A). In previous immunogenicity studies of MSP1–42 using different adjuvants, we have shown that the specificity of the anti-MSP1 antibodies is influenced by adjuvant formulations (60–62). It is possible that rMSP1 delivered via IO nanoparticles induced antibodies that are much more focused toward parasite inhibitory epitopes than antibodies induced by rMSP1-CFA or rMSP1-ISA51, which was reflected in the differences in the degree of in vitro parasite inhibition observed (Table 1). More important, our results showed that in the nonhuman primates, the levels of parasite inhibitory antibodies induced by rMSP1-IO were not only significant (55–100% inhibition) but surpassed or matched those levels previously achieved in MSP1–42-vaccinated and protected monkeys (33).

Antigen-specific cellular responses, as analyzed by ELISPOT assays, revealed that IO immunizations via all three routes induced a more prominent IL-4 than IFN-γ response, which was similar in profile to immunizations with the conventional adjuvants, CFA and ISA51 (Fig. 4). IO immunizations via the i.p. and s.c. routes also induced significantly higher IL-4 responses than CFA. Although the relevance of this biased IL-4 response in malaria immunity is not clear, it is possible that a more skewed TH2 response would favor antibody production, which is critical in MSP1-specific immunity (32, 33, 35–39).

We began to investigate the immunological basis for the adjuvancy of the IO nanoparticles. Initially, it was thought that the rapid clearance of small (<20 nm) nanoparticles from the body (63, 64) would impede their use as a delivery platform for polypeptide antigens due to short half-life. However, the small size and their propensity to behave as a true solution may facilitate their dispersion and allow them to easily penetrate key immunological organs and immune cells, such as professional antigen-presenting cells (APCs; ref. 65). Indeed, our study showed that the IO nanoparticles were efficiently taken up by APCs, BMDCs and macrophages. Furthermore, uptake of IOs by BMDCs led to their activation, with increased expression of CD86, proinflammatory cytokines (IL-6, TNF-α, and IL-1β), and chemokines. The activation profiles as well as the levels of some of these immune mediators (i.e., CCL3 and CCL4) mimicked those produced via activation with LPS. It is likely that the combined effects of increased chemokine and proinflammatory cytokine production enhanced the migratory characteristics of these immune cells, resulting in more efficient antigen presentation and/or T-cell activation. These data indicate that the IO nanoparticles possess immunmodulating activities and function beyond the role of vaccine vehicle/antigen delivery to activate strong immune responses, specifically through dendritic cells.

Despite efficient uptake of IO nanoparticles, macrophages did not show levels of activation comparable to BMDCs, either in terms of the number of cytokines and chemokines produced or the duration of expression of these immune mediators. However, the fact that transient production of IL-1β, IL-6, and TNF-α were observed in these macrophages suggests that these cells may also play a role in immunoenhancements. We are currently investigating the molecular and biological basis for the differential activation of BMDCs and macrophages by the IO nanoparticles. The observed disparity in activation patterns may be the result of distinct mechanisms of nanoparticles uptake by the two cell types, with the possibility that once inside the cell, the fate of the nanoparticles in terms of subcellular location/translocation are different, leading to differences in their ability to activate or modulate the innate immune response pathways (66, 67). A better understanding of the mechanisms by which IO nanoparticles interact with APC populations is an important next step to further improve potency and efficiency. An empirical approach would be to devise strategies for specific targeting of the conjugated IOs to dendritic cells to increase the efficiency of IO uptake by these professional APCs, which may lead to stronger immunoenhancements.

In addition to its effectiveness as a vaccine platform with built-in adjuvancy, our study demonstrated other highly desirable attributes of the IO nanoparticles for use in vaccine formulations. First, rMSP1-IO was very stable in solution, with little or no loss of antigenicity over an 18-mo period. Especially significant is the finding that rMSP1-IO was amenable to lyophilzition, and the antigenicity and immunogenicity of rMSP1-IO were preserved during lyophilization and subsequent rehydration processes. The lyophilized rMSP1-IO induced comparable anti-MSP1–19 antibodies and cellular responses in mice as the nonlyophilized rMSP1-IO (Fig. 5C, D) and induced comparable high levels of parasite inhibitory antibodies (Table 1). These biophysical attributes make IO-conjugated vaccines highly suitable for field deployment in less ideal environments, whereby the vaccine formulations can be transported as powder form at room temperature and can be readily rehydrated (with water) prior to parenteral administrations.

Second, the IO nanoparticles were well tolerated as an injectable in nonhuman primate (Aotus) immunizations, where the immunized animals suffered no apparent systemic toxicity after a 3-dose regimen (Supplemental Table S2). This finding is further supported by a dose-escalating toxicity study in SW mice, which also revealed no significant systemic abnormalities that would indicate kidney or liver damage, despite the mice receiving 3 exceptionally large doses of IO nanoparticles every 21 d (Supplemental Table S1). We speculate that the small size of these nanoparticles (<20 nm) should facilitate their rapid clearance (63, 64), thereby reducing the toxic effects, if any. Preliminary examination of Prussian Blue-stained kidney sections of IO immunized mice did not identify accumulations of IO nanoparticles (data not shown). However, more sensitive systemic studies such as the use of fluorescent dyes will need to be performed in order to trace the actual clearance of the IO nanoparticles.

Results of this study provide strong evidence that a water-soluble, inorganic nanoparticle (<20 nm) with an IO core is an effective vaccine delivery platform with built-in adjuvancy and low toxicity. The stability of the IO-vaccine conjugate, in solution and as a lyophilized formulation, makes it ideal for field deployment. Further development of these nanoparticles for vaccine delivery may include optimization of parameters such as particle concentration and size, antigen conjugation methods, antigen release characteristics, and in vivo targeting methods.