Oral Peptide Vaccine against Hookworm Infection: Correlation of Antibody Titers with Protective Efficacy

Approximately 0.4 billion individuals worldwide are infected with hookworm. An effective vaccine is needed to not only improve the health of those affected and at high risk, but also to improve economic growth in disease-endemic areas. An ideal anti-hookworm therapeutic strategy for mass administration is a stable and orally administered vaccine. Oral vaccines are advantageous as they negate the need for trained medical staff for administration and do not require strict sterility conditions. Vaccination, therefore, can be carried out at a significantly reduced cost. One of the most promising current antigenic targets for hookworm vaccine development is the aspartic protease digestive enzyme (APR-1). Antibody-mediated neutralization of APR-1 deprives the worm of nourishment, leading to reduced worm burdens in vaccinated hosts. Previously, we demonstrated that, when incorporated into vaccine delivery systems, the APR-1-derived p3 epitope (TSLIAGPKAQVEAIQKYIGAEL) was able to greatly reduce worm burdens (≥90%) in BALB/c mice; however, multiple, large doses of the vaccine were required. Here, we investigated a variety of p3-antigen conjugates to optimize antigen delivery and establish immune response/protective efficacy relationships. We synthesized, purified, and characterized four p3 peptide-based vaccine candidates with: (a) lipidic (lipid core peptide (LCP)); (b) classical polymeric (polymethylacrylate (PMA)); and (c) novel polymeric (polyleucine in a branched or linear arrangement, BL10 or LL10, respectively) groups as self-adjuvanting moieties. BL10 and LL10 induced the highest serum anti-p3 and anti-APR-1 IgG titers. Upon challenge with rodent hookworms, the highest significant reduction in worm burden was observed in mice immunized with LL10. APR-1-specific serum IgG titers correlated with worm burden reduction. Thus, we provide the first vaccine-triggered immune response-protection relationship for hookworm infection.


Introduction
Hookworms are hematophagous parasites that infect half a billion individuals globally [1,2]. Hookworm infection is one of the most widespread but neglected tropical diseases. Its effects cause substantial economic losses to developing countries, in particular [3][4][5]. Despite the high efficacy of anthelmintic drugs, however, hookworms are developing resistance [6] and they do not prevent reinfection which is common in epidemic areas [7]. Therefore, an effective anthelmintic vaccine is urgently needed.
Early hookworm vaccines for veterinary purposes utilized whole parasite-based approaches. However, recombinant forms of larval antigens, e.g., Ancylostoma secretory protein-2 (ASP-2), caused strong allergenic responses in vaccinated human subjects [8,9]. this study, we aimed to employ lower doses to distinguish subtle differences between new vaccine candidates and their protective efficacies in mice.
We investigated whether novel vaccine candidates, designed to resist intestinal proteolysis, could trigger strong protective immune responses against hookworms with fewer doses. To facilitate comparison between the applied delivery systems, all vaccine candidates were produced in branched structure, although the polyleucine conjugate was also produced in linear structure ( Figure 1). All candidates carried the APR-1 B-cell epitope (p3), a CD4-T cell epitope (p25), a solubilizing lysine-based unit (SKKKK), and a hydrophobic adjuvanting moiety. Following immunization of BALB/c mice with 4 × 100 µg doses of the vaccine candidates, antibody production, in vitro neutralization of APR-1 by anti-vaccine antibodies, and efficacy against challenge infection in vivo were investigated. Finally, we attempted to determine the correlation between immune responses and worm and egg burden reductions.
Vaccines 2021, 9, x FOR PEER REVIEW 3 of 16 this study, we aimed to employ lower doses to distinguish subtle differences between new vaccine candidates and their protective efficacies in mice. We investigated whether novel vaccine candidates, designed to resist intestinal proteolysis, could trigger strong protective immune responses against hookworms with fewer doses. To facilitate comparison between the applied delivery systems, all vaccine candidates were produced in branched structure, although the polyleucine conjugate was also produced in linear structure ( Figure 1). All candidates carried the APR-1 B-cell epitope (p3), a CD4-T cell epitope (p25), a solubilizing lysine-based unit (SKKKK), and a hydrophobic adjuvanting moiety. Following immunization of BALB/c mice with 4 × 100 µg doses of the vaccine candidates, antibody production, in vitro neutralization of APR-1 by anti-vaccine antibodies, and efficacy against challenge infection in vivo were investigated. Finally, we attempted to determine the correlation between immune responses and worm and egg burden reductions.
Reverse-phase HPLC (Shimadzu Corp, Kyoto, Japan) with a Vydac C4 column was used for analysis and purification of the synthesized compounds. Attenuated total reflectance-Fourier transform infrared spectrophotometry (ATR-FTIR, Spectrum, PerkinElmer, Beaconsfield, UK) was used to evaluate the compounds' secondary structures. Particle size distribution of the self-assembled nanoparticle compounds was evaluated using photon correlation spectroscopy (Zetasizer Nano, Malvern instruments, Worcestershire, UK). A fluorescence plate reader was used for evaluating in vitro APR-1 enzyme neutralization (BMG CLARIOStar ® fluorescent plate reader, BMG LABTECH, Figure 1. Peptide antigen (p3-P25) and vaccine candidate (LCP, BL 10 , LL 10 and PMA) structures: p3 is an APR-1 B-cell epitope (TSLIAGPKAQVEAIQKYIGAEL), P25 is a T-helper epitope (KLIP-NASLIENCTKAEL), and SKKKK is a solubilizing moiety.
Reverse-phase HPLC (Shimadzu Corp, Kyoto, Japan) with a Vydac C4 column was used for analysis and purification of the synthesized compounds. Attenuated total reflectance-Fourier transform infrared spectrophotometry (ATR-FTIR, Spectrum, PerkinElmer, Beaconsfield, UK) was used to evaluate the compounds' secondary structures. Particle size distribution of the self-assembled nanoparticle compounds was evaluated using photon correlation spectroscopy (Zetasizer Nano, Malvern instruments, Worcestershire, UK). A fluorescence plate reader was used for evaluating in vitro APR-1 enzyme neutralization (BMG CLARIOStar ® fluorescent plate reader, BMG LABTECH, Cary, NC, USA). OriginPro 2020 Vaccines 2021, 9, 1034 4 of 16 software (OriginLab Corporation, Northampton, MA, USA) was used for mathematical data analyses.

Synthesis of Peptide Vaccines
All peptides were synthesized using Fmoc solid-phase peptide synthesis (SPPS) (Figure 1). Resin was swollen in DMF, then put through cycles of double-deprotection with 20% piperidine in DMF (1 × 5 min, then 1 × 10 min) and double-couplings to resin, using activated-protected amino acids (4.2 Eq), HATU (4.0 Eq) and DIEA (6.2 Eq) dissolved in DMF, for each amino acid (2 × 30 min). When the peptide sequences were complete, the N-termini were acetylated using acetic anhydride and DIEA in DMF (0.5:0.5:9.0). The resin was dried overnight in a desiccator, then the peptide was cleaved off using a 20 mL/g resin solution of TIS and MilliQ water in TFA (0.25:0.25:9.5). After cleavage, the peptides were precipitated using cold diethyl ether; extracted using a mixture of acetonitrile, MilliQ water, and TFA at a ratio of 90:10:0.1 for hydrophobic peptides, p3-K(L 10 )-P25-SK 4  Synthesis of PMA. A mixture of p3-K (pentynoyl)-P25-SK 4 (6.0 mg, 1.2 mmole, 1.2 Eq) and azide-polymethyl acrylate polymer (2.4 mg, 1.0 mmole, 1.0 Eq) was dissolved in DMF (3 mL). Copper wire (50 mg) was added to the reaction mixture and the air was partially removed with nitrogen gas for 15-30 s. The reaction mixture was stirred at 50 • C for 14 h.
The PMA conjugate, p3-K(PMA)-P25-SK 4 (Figure 1), in DMF solution was allowed to form nanoparticles by controlled antisolvent nanoprecipitation via syringe pump at a flow rate of 0.1 mL/hour into deionized water (water volume 4x greater than DMF volume), with stirring at 1000 RPM. The nanoparticles were purified from excess free/unconjugated peptide and DMF by dialysis (10,000 MWCO) in deionized water several times over three days; the compound was then freeze-dried for later use. Controlled antisolvent nanoprecipitation was conducted to produce fresh PMA conjugate nanoparticles prior to each immunization to ensure uniformity in particle size. A freeze-dried peptide-polymer conjugate (PMA) sample (5 mg) was analyzed using elemental microanalysis. The determined nitrogen to carbon ratio (N/C) provided the conjugation extent between nitrogen-rich peptide (N/C = 0.2850) and polymethylacrylate polymer (nitrogen-free, except for the azide group, N/C = 0.0312). The extent of conjugation was determined by plotting calculated theoretical N/C ratios, ranging from 0% (free polymer, N/C = 0.0312) to 100% conjugation (conjugate N/C = 0.205), followed by interpolation using the measured/found N/C ratio  Figure S1). The synthesized PMA conjugate-measured N/C ratio (0.198) corresponded to a conjugation extent of 93%.

Peptide Vaccine Particle Size and Shape
Particle size distribution was evaluated by preparing a 0.2 mg/mL solution of each pure compound in PBS and measured using dynamic light scattering (DLS; Zetasizer Nano ZP) at 25 • C and an angle of 173 • .

Transmission Electron Microscopy
Compounds were evaluated by transmission electron microscopy using a JEM-1010 microscope (JEOL LTD, Tokyo, Japan) at 100 kV. The images were analyzed using AnalySIS ® software (Megaview III, Munster, Germany).

In Vitro Stability against Proteolysis
Enzyme 1000 U solution was prepared from bovine trypsin (10,000 U/mg, Sigma-Aldrich, Castle Hill, NSW, Australia) with 1 mM CaCl 2 in 1 mL PBS, as reported previously [39,40]. After preparation, 100 µL of enzyme solution was added to 400 µL of peptide solution (1.25 mg/mL) in PBS and incubated at 37 • C with shaking. To inhibit enzyme activity, 40 µL of TFA and acetonitrile mixture (1:2) was added to 60 µL aliquots of the mixture at each interval. The solution was analyzed by analytical HPLC with a calibration curve employing the compound's peak at its characteristic retention time to determine intact compound concentration.

Secondary Structure of Peptide Vaccines
The compounds' secondary structures were evaluated according to the methodology of Yang et al. [41]. Solutions of each compound in PBS (2 mg/mL) were prepared and placed on the diamond-attenuated total reflectance accessory of the ATR-FTIR, and absorbance spectra were collected from 500 and 4000 cm −1 , in triplicate, at 25 • C and 4 cm −1 resolution. The first and second derivative of the absorbance spectra were generated and analyzed using Originpro 2020 software. The second derivative secondary structure component-peak assignment was also detected and quantified using Originpro 2020 software.

Oral Immunization of BALB/c Mice
Six groups of eight (6-8-week-old) female BALB/c mice (Animal Resources Centre, Perth, WA, Australia) were employed for the study. Naïve sera, saliva and feces were collected from each mouse two days prior to immunization. The mice were orally immunized via oral gavage with either PBS as a negative control, p3-P25/cholera toxin subunit B (CTB) as a positive control, or one of our vaccine candidates: LCP, PMA, BL10, or LL10. Vaccine doses contained 100 µg of antigen in 100 µL of PBS. Four immunizations were given to each group, with two-week intervals between doses, over six weeks, followed by an infection challenge by the end of the eighth week. An infection challenge was performed two weeks after the last immunization. Sample processing was as follows: sera were diluted 10-fold in PBS; saliva samples were undiluted; fecal samples (0.1 g) were homogenized in 1 mL PBS containing 0.05 v/v% tween 20 and 0.05 w/v% sodium azide.

Serological and Mucosal Immunogenicity Assays
Enzyme-linked immunosorbent assays (ELISA) were used to evaluate antibody titers of p3-P25, LCP, PMA, BL 10 , and LL 10 using 96-well, flat bottom and high binding affinity plates (Sarstedt, Germany). The plates were coated with 100 µL of solution in carbonate buffer (pH 9.6) with either p3 epitope (1 µg/mL) or mutant APR-1 M74 (0.5 µg/mL) expressed as previously described [15]. The plates were incubated at 37 • C for 90 min. PBS containing 0.05% tween 20 (PBST) was used as wash buffer. The plates were washed using deionized water, followed by PBST, then dried by tapping. PBST with 0.5% skim milk was added to the plates at a volume of 135 µL for wells in the first column, and 100 µL to all remaining wells. Then, 15 µL of 10-fold diluted serum from each mouse was added to first-column wells, and the plates were mixed. Serial dilutions (3-fold) were done along plate rows using a multichannel pipette (Biotools Pty Ltd., Loganholme, QLD, Australia) and an auto-ELISA system (Viaflo ASSIST, Integra Biosciences, Hudson, NH, USA). For fecal samples, 100 µL of PBST with 0.5% skim milk was added to each well. 50 µL of fecal extract was added to first-columns wells. The plates were mixed and then serially diluted (3-fold) along the rows. The plates were incubated at 37 • C for 90 min, then washed using deionized water and wash buffer, and dried by tapping. Secondary antibody solution (100 µL, 1/4000 diluted in PBST with 0.5% skim milk) of either goat anti-mouse IgG-HRP (Bio-Rad, Gladesville, NSW, Australia), goat anti-mouse IgA-HRP (Life Technologies, Mulgrave, VIC, Australia), goat anti-mouse IgG1-HRP (Bio-Rad, Gladesville, NSW, Australia), or goat anti-mouse IgG2a-HRP (Bio-Rad, Gladesville, NSW, Australia) was added to the wells depending on antibody isotype or subclass of interest. The plates were incubated again at 37 • C for 90 min, then washed using deionized water and wash buffer, and dried by tapping. Finally, peroxidase o-phenylenediamine dihydrochloride (OPD) substrate was applied for 15 min. The reaction was stopped using 100 µL of 1 N sulfuric acid and the absorbance of each well was measured using a SpectraMax 250 microplate reader (Molecular Devices, San Jose, CA, USA) at 450 nm. The titer was described as the lowest dilution that produced an absorbance above the mean absorbance of control wells (containing naïve mouse serum that was sampled pre-immunization) plus 3 standard deviations.

Enzyme Neutralization
The enzyme neutralization assay was designed based on the strategy reported by Pearson et al. [13,15,24] but modified to suit immune mice sera instead of purified neutralizing antibodies. A functional APR-1 wild-type (10 ng/50 µL) solution, expressed as described previously [13,15,42], in 50 mM sodium acetate pH 3.5 buffer was incubated with MoCAc-GKPILFFRLK, hemoglobin-substitute pre-fluorescence substrate (1 µM final concentration), in the presence or absence of immune mouse serum in the wells of Corning ® costar white opaque fluorescence 96-well plates. The reaction volume was adjusted to 100 µL, using acetate buffer pH 3.5. The sera from four mice per group were pooled to result in two pooled serum samples per group. A total of 5 µL of each stock was added to the enzyme and allowed to bind/neutralize for 30 min. The fluorescent substrate was added immediately prior to commencement of fluorescence measurement using Spectra-Max 250 microplate reader (Molecular Devices, San Jose, CA, USA). The plate wells were temperature-controlled (37 • C) and measured for fluorescence intensity in 5-min cycles for 2 h, at an excitation wavelength of 330 nm. Signals were recorded at an emission wavelength of 390 nm using appropriate optic filters. The photomultiplier tube gain was set at 75%, with auto-adjusted focal height for optimum signal output. The kinetics of substrate proteolysis was monitored over the measurement period to compare relative inhibition efficacies of the different groups' sera. Pooled immune sera relative fluorescence units (RFU) were presented as time-integral percentage increases in RFU values from the onset of the experiment.

Challenge Infection
All life stages of N. brasiliensis were maintained in Sprague Dawley rats. From these, an L3 larval suspension was prepared and each mouse was infected subcutaneously in the scruff with 750 larvae, as recommended for rat-adapted larvae infectivity in mice [37]. The mice were euthanized 7 days post-infection, and their intestines were isolated to conduct intestinal adult worm counts. Colon fecal contents were collected to conduct egg counts as described previously [34,35,37]. Briefly, fecal contents were weighed (with a range of 100-200 mg), and 4-10 mL of saturated salt solution was added. The feces were homogenized and the egg suspension was loaded into 300 µL-chambers of a counting slide, placed under a microscope, and counted twice. The number of eggs per gram of feces was determined using Equation (1), where V is the total volume of salt solution added in mL, and W is the feces weight in grams. Intestinal worm burden was determined by isolating the intestines from cecum to pylorus. The intestine was cut longitudinally and placed in a petri dish containing 5 mL sterile PBS, then incubated at 37 • C for 1 h. Adult hookworms were counted using a dissection microscope at 2×-magnification.

Ethics Statement
All animal experiments were approved by the James Cook University (JCU) Animal Ethics Committee (ethics number A2571). The animals were housed at JCU, and allowed free access to pelleted food and water, in accordance with the Queensland Animal Care and Protection Act.

Secondary Structure
PMA nanoparticle size was similar to the light wavelength used in circular dichroism (CD) analysis and, therefore, caused CD spectrum flattening due to scattered radiation [43][44][45]. Therefore, we used ATR-FTIR to evaluate the secondary structure of all conjugates (Figures 4 and S2). The control peptide, p3-P25, adopted a predominantly random coil structure (60%) and α-helical (35%) conformation, with very minor β-sheet contribution (3%). Polyleucine-adjuvanted compounds also had the tendency to form random coil and helical structures; however, β-sheet contribution significantly increased to 7% and 6% for BL 10 and LL 10 , respectively. However, it is worth mentioning that LCP, PMA, BL 10 and LL 10 carry P25 (T-cell epitope) and adjuvanting moieties that affect the conformational content of each conjugate. This predicts potentially higher β-sheet contribution in p3 epitope conformation within the conjugates than for the whole conjugates. Furthermore, the quantitative secondary structure percentage values of non-polymeric peptide vaccines measured by circular dichroism spectropolarography, described previously [32], agree with ATR-FTIR results.

Secondary Structure
PMA nanoparticle size was similar to the light wavelength used in circular dichroism (CD) analysis and, therefore, caused CD spectrum flattening due to scattered radiation [43][44][45]. Therefore, we used ATR-FTIR to evaluate the secondary structure of all

Secondary Structure
PMA nanoparticle size was similar to the light wavelength used in circular dichroism (CD) analysis and, therefore, caused CD spectrum flattening due to scattered radiatio [43][44][45]. Therefore, we used ATR-FTIR to evaluate the secondary structure of a PMA, BL10 and LL10 carry P25 (T-cell epitope) and adjuvanting moieties that affect the conformational content of each conjugate. This predicts potentially higher β-sheet contribution in p3 epitope conformation within the conjugates than for the whole conjugates. Furthermore, the quantitative secondary structure percentage values of nonpolymeric peptide vaccines measured by circular dichroism spectropolarography, described previously [32], agree with ATR-FTIR results.

Enzymatic Stability In Vitro
The proteolytic stability of the peptide vaccines was investigated against bovine trypsin digestion at 37 °C ( Figure 5). PMA stability was significantly higher (p < 0.0001) than that of all other vaccine compounds, with only 10% degradation observed after 30 min of incubation with trypsin. LL10 was also relatively resistant to proteolysis (25% degradation at 30 min, p < 0.0001); however, only 15% of the conjugate remained intact after 60 min. BL10 and LCP were more rapidly degraded by trypsin; however, their degradation was still significantly slower than p3-P25.

Enzymatic Stability In Vitro
The proteolytic stability of the peptide vaccines was investigated against bovine trypsin digestion at 37 • C ( Figure 5). PMA stability was significantly higher (p < 0.0001) than that of all other vaccine compounds, with only 10% degradation observed after 30 min of incubation with trypsin. LL10 was also relatively resistant to proteolysis (25% degradation at 30 min, p < 0.0001); however, only 15% of the conjugate remained intact after 60 min. BL10 and LCP were more rapidly degraded by trypsin; however, their degradation was still significantly slower than p3-P25.
conformational content of each conjugate. This predicts potentially higher β-sheet contribution in p3 epitope conformation within the conjugates than for the whole conjugates. Furthermore, the quantitative secondary structure percentage values of nonpolymeric peptide vaccines measured by circular dichroism spectropolarography, described previously [32], agree with ATR-FTIR results.

Enzymatic Stability In Vitro
The proteolytic stability of the peptide vaccines was investigated against bovine trypsin digestion at 37 °C ( Figure 5). PMA stability was significantly higher (p < 0.0001) than that of all other vaccine compounds, with only 10% degradation observed after 30 min of incubation with trypsin. LL10 was also relatively resistant to proteolysis (25% degradation at 30 min, p < 0.0001); however, only 15% of the conjugate remained intact after 60 min. BL10 and LCP were more rapidly degraded by trypsin; however, their degradation was still significantly slower than p3-P25.

Immunogenicity of the Conjugates
Antibody titers were evaluated using ELISA for mouse fecal extracts, saliva samples and serum ( Figure 6). Mice immunized with LCP, PMA, BL10, and LL10 produced significantly higher p3-IgG titers compared to those in the PBS group ( Figure 6A). However, only some of these antibodies recognized the parent protein (APR-1). In addition, IgA and IgG levels in the saliva and fecal extracts were low, IgG2a levels in the serum were undetectable, and only mice immunized with LL10 and BL10 generated detectable mucosal antibody responses (Figures S3 and S4). No significant level of APR-1-specific antibody production was detected in fecal samples, most likely due to the sensitivity limitations of the method (detectability of such low titer levels). Sera from all immunized mice, except for those in the p3/CTB group, were able to significantly reduce APR-1 enzymatic activity compared to sera from PBS-immunized mice (Figure 7). The mice were monitored on daily basis between immunizations, and they did not exhibit any adverse events or clinical signs of illness or toxicity. addition, IgA and IgG levels in the saliva and fecal extracts were low, IgG2a levels in the serum were undetectable, and only mice immunized with LL10 and BL10 generated detectable mucosal antibody responses (Figures S3 and S4). No significant level of APR-1-specific antibody production was detected in fecal samples, most likely due to the sensitivity limitations of the method (detectability of such low titer levels). Sera from all immunized mice, except for those in the p3/CTB group, were able to significantly reduce APR-1 enzymatic activity compared to sera from PBS-immunized mice (Figure 7). The mice were monitored on daily basis between immunizations, and they did not exhibit any adverse events or clinical signs of illness or toxicity.  detectable mucosal antibody responses (Figures S3 and S4). No significant level of APR-1-specific antibody production was detected in fecal samples, most likely due to the sensitivity limitations of the method (detectability of such low titer levels). Sera from all immunized mice, except for those in the p3/CTB group, were able to significantly reduce APR-1 enzymatic activity compared to sera from PBS-immunized mice (Figure 7). The mice were monitored on daily basis between immunizations, and they did not exhibit any adverse events or clinical signs of illness or toxicity.  In vitro serum APR-1 enzyme neutralization test: integration of substrate-generated fluorescence (RFU) over 30 min per serum group. Fluorescent substrate digestion by APR-1 was inhibited by immune sera and compared to sera from the negative control group. Statistical analysis was performed using one-way ANOVA with Tukey's multiple comparison test (ns, non-significant, p > 0.05; (**) p < 0.01; and (***) p < 0.001).

Infection Challenge
Mice were euthanized 7 days post-infection, and hookworms in the intestines and eggs in the feces were counted (Figure 8). Vaccination with LL10 and LCP resulted in significantly reduced parasite and egg numbers after challenge infection; BL10 also produced significant protective immune responses according to non-parametric analysis. p > 0.05; (**) p < 0.01; and (***) p < 0.001).

Infection Challenge
Mice were euthanized 7 days post-infection, and hookworms in the intestines and eggs in the feces were counted (Figure 8). Vaccination with LL10 and LCP resulted in significantly reduced parasite and egg numbers after challenge infection; BL10 also produced significant protective immune responses according to non-parametric analysis.

Discussion
The immunogenicity of oral vaccines is affected by the interplay of three variables: oral delivery and uptake of nanoparticles, stability against enzymatic digestion, and intrinsic vaccine construct immunogenicity. Of the conjugates tested, only PMA demonstrated poor efficacy in the challenge study. Its conformational properties did not differ significantly from LCP, LL10 or BL10; however, PMA was very stable against enzymatic degradation and formed much larger particles than the other conjugates ( Figure 2). It is worthy to mention that the trypsin cleavage sites were present only in p3-P25 and this peptide was degraded almost completely in less than 30 min upon treatment with trypsin. All other conjugates contained this peptide and consequently possessed identical bovine trypsin cleavage sites. Thus, the enzymatic cleavage was inhibited by the presence of adjuvanting moieties and not by the reduced number of cleavage sites. Nanoparticles in the range of 20-50 nm are usually considered to be the most immunogenic [25]; thus, PMA's reduced efficacy might be correlated with its larger than optimal nanoparticle size. In other words, the larger particle size of PMA may have compromised its uptake by M-cells, delivery to underlying lymph nodes, and/or its intrinsic immunogenicity.
Since in vivo-derived data were available for each mouse, we sought to categorize mice according to worm burden reduction. Three categories were formed: 'high' (>75% worm burden reduction), 'moderate' (50-75% worm burden reduction), and 'low' (<50% worm burden reduction). The percentage of reduction was calculated in comparison to worm levels in control mice immunized with PBS ( Figure 9A). The same approach was applied to egg burden reduction ( Figure 9B). Low worm and egg burden reductions were not substantially different from the PBS group, while moderate and high reduction burdens were significantly different (p < 0.0001).

Discussion
The immunogenicity of oral vaccines is affected by the interplay of three variables: oral delivery and uptake of nanoparticles, stability against enzymatic digestion, and intrinsic vaccine construct immunogenicity. Of the conjugates tested, only PMA demonstrated poor efficacy in the challenge study. Its conformational properties did not differ significantly from LCP, LL 10 or BL 10 ; however, PMA was very stable against enzymatic degradation and formed much larger particles than the other conjugates ( Figure 2). It is worthy to mention that the trypsin cleavage sites were present only in p3-P25 and this peptide was degraded almost completely in less than 30 min upon treatment with trypsin. All other conjugates contained this peptide and consequently possessed identical bovine trypsin cleavage sites. Thus, the enzymatic cleavage was inhibited by the presence of adjuvanting moieties and not by the reduced number of cleavage sites. Nanoparticles in the range of 20-50 nm are usually considered to be the most immunogenic [25]; thus, PMA's reduced efficacy might be correlated with its larger than optimal nanoparticle size. In other words, the larger particle size of PMA may have compromised its uptake by M-cells, delivery to underlying lymph nodes, and/or its intrinsic immunogenicity.
Since in vivo-derived data were available for each mouse, we sought to categorize mice according to worm burden reduction. Three categories were formed: 'high' (>75% worm burden reduction), 'moderate' (50-75% worm burden reduction), and 'low' (<50% worm burden reduction). The percentage of reduction was calculated in comparison to worm levels in control mice immunized with PBS ( Figure 9A). The same approach was applied to egg burden reduction ( Figure 9B). Low worm and egg burden reductions were not substantially different from the PBS group, while moderate and high reduction burdens were significantly different (p < 0.0001).
We additionally analyzed the IgG titers generated by individual mice in correlation to worm burden reduction. High and medium worm burden reductions clearly correlated with APR-1-IgG titer levels when compared to mice with low reductions (p < 0.0001), while less significant differences (p < 0.01) were found when serum p3-IgG titers were used for the correlation ( Figure 9C). This clearly suggests that anti-APR-1 serum antibodies underpin the protection mechanism resulting in worm reduction. The difference in significance levels between p3-IgG and APR-1-IgG demonstrates that the quality of antibody recognition of native epitope conformation is a critical factor in the efficacy of the immune response generated.
Our categorization approach differentiated serum immune responses between vaccinated groups and pinpointed the most promising immune correlates of protection. In contrast, the correlation between mucosal antibody responses and worm burden reduction was, in general, not statistically significant (Supplementary Figure S5).
while less significant differences (p < 0.01) were found when serum p3-IgG titers were used for the correlation ( Figure 9C). This clearly suggests that anti-APR-1 serum antibodies underpin the protection mechanism resulting in worm reduction. The difference in significance levels between p3-IgG and APR-1-IgG demonstrates that the quality of antibody recognition of native epitope conformation is a critical factor in the efficacy of the immune response generated.
Our categorization approach differentiated serum immune responses between vaccinated groups and pinpointed the most promising immune correlates of protection. In contrast, the correlation between mucosal antibody responses and worm burden reduction was, in general, not statistically significant (Supplementary Figure S5). Figure 9. Correlation of antibody titers with protective efficacy for individual mice. The mice of all immunized groups were reshuffled according to their worm burden reduction; 'High' (>75%), 'medium' (>50%), and 'low' (<50%) worm burden reductions (A), to isolate the effective immune responses that distinguish the mice with superior protection. Egg burden reductions (B) corresponded with worm count categorization in panel (A). The correlation between worm burden reduction and serum anti-p3 IgG titers (p3-IgG) or serum anti-APR-1 IgG titers (APR-1-IgG) (C) among re-categorized/re-shuffled mice groups. Statistical analysis was performed using one-way Figure 9. Correlation of antibody titers with protective efficacy for individual mice. The mice of all immunized groups were reshuffled according to their worm burden reduction; 'High' (>75%), 'medium' (>50%), and 'low' (<50%) worm burden reductions (A), to isolate the effective immune responses that distinguish the mice with superior protection. Egg burden reductions (B) corresponded with worm count categorization in panel (A). The correlation between worm burden reduction and serum anti-p3 IgG titers (p3-IgG) or serum anti-APR-1 IgG titers (APR-1-IgG) (C) among recategorized/re-shuffled mice groups. Statistical analysis was performed using one-way ANOVA with Tukey's multiple comparison test (ns, non-significant, p > 0.05; (*) p < 0.05; (**) p < 0.01; (***) p < 0.001; and (****) p < 0.0001).
Immunizing mice groups with peptide vaccines resulted in different oral immunogenicity due to a variety of factors including their different self-assembled particle size (Figures 2 and 3), stability against intestinal digestion ( Figure 5), and intrinsic immunogenicity. The APR-1 neutralization capacity (Figure 7) of immune sera in vitro was shown to reduce worm burdens and egg counts in vivo in a similar manner (Figures 8 and 9). Furthermore, the efficacy of humoral responses agrees with earlier findings showing that humoral responses are the primary means of protection against adult hookworms, where protection was passively conveyed via the transfer of immune sera to naïve rats or dogs [46,47]. This holds especially true here as a neutralizing B-cell epitope (p3) was employed as the constant core antigen in all of our vaccine constructs. Moreover, serum anti-p3 IgG titers measured between immunizations did not decrease, suggesting a lack of vaccine-induced immune tolerance ( Figure S6). Further, mice sera did not generate anti-human cathepsin-D IgG titers or serum anti-APR-1 IgE titers ( Figure S7), thus peptide vaccines resulted in protection without compromising safety via allergic or autoimmune responses. The lack of anti-cathepsin-D antibodies was expected since the peptide epitope (p3) is not conserved in human or murine cathepsin-D ( Figure S8).
In summary, oral administration (4 × 100 µg) of self-assembled peptide nanovaccines of 10-25 nm in diameter induced antibody-based protective immune responses. These responses resulted in reduced intestinal worm burdens of 55-80% and fecal egg burden reductions of 55-85% in the BALB/c-Nb challenge infection model of human hookworm infection. Vaccine efficacy was reflected by a correlation between serum APR-1-specific IgG responses and worm and egg burden reductions. Linear polyleucine-adjuvanted compound (LL 10 ) showed improved resistance against trypsin digestion in vitro, self-assembled into small-sized nanoparticles (22 nm) and generated a protective anti-APR-1 IgG response that resulted in the highest reduction of worm burden (80%) in the BALB/c-Nb challenge model.

Conclusions
In this study, we determined that APR-1-specific serum IgG is a significant immune correlate of protection against hookworm infection in vivo, in the absence of significant titers of mucosal/fecal APR-1-specific IgA/IgG. However, the generated IgG titers originated from gut-associated lymphoid tissue; therefore, local responses may have been stronger than peripheral (blood) responses, as these tissues are more concentrated around the infected intestinal sites. The IgG generated were dominated by the neutralizing IgG1 subclass, rather than IgG2, and protective efficacy did not significantly depend on antigenspecific salivary IgA/IgG. Moreover, we showed that fewer vaccine doses can still be effective; 4 × 100 µg oral immunizations with more stable, self-adjuvanted peptide, including LCP, LL 10 and BL 10 , were adequate to significantly reduce worm burden by 56-78% (p < 0.05) in mice after N. brasiliensis larval challenge. In this experiment, the positive control (p3-P25 + CTB) was not able to significantly reduce worm burden (reductions of only 35%; p > 0.05) at the vaccine dose employed. Therefore, despite lower in vitro stability against trypsin digestion, BL 10 resulted in similar serum titers to LL 10 . This suggests that BL 10 is more immunogenic and readily taken up by gut-associated lymphoid tissue, resulting in similar serum anti-p3-IgG titers despite its lower stability against digestion. LCP, BL 10 and LL 10 compounds were more stable against enzymatic degradation, and were significantly more protective, compared to the positive control (p3-P25 + CTB). Our data further support the development of oral vaccines targeting enzyme neutralizing epitopes of Na-APR-1 to control human hookworm infection in disease-endemic countries. The findings also have potential implications for the development of vaccines against other gastrointestinal helminth infections of humans and domestic livestock animals.