Progress towards the Development of a NEAT Vaccine for Anthrax II: Immunogen Specificity and Alum Effectiveness in an Inhalational Model

Bacillus anthracis is the causative agent of anthrax disease, presents with high mortality, and has been at the center of bioweapon efforts. The only currently U.S. FDA-approved vaccine to prevent anthrax in humans is anthrax vaccine adsorbed (AVA), which is protective in several animal models and induces neutralizing antibodies against protective antigen (PA), the cell-binding component of anthrax toxin. However, AVA requires a five-course regimen to induce immunity, along with an annual booster, and is composed of undefined culture supernatants from a PA-secreting strain.

ABSTRACT

Bacillus anthracis is the causative agent of anthrax disease, presents with high mortality, and has been at the center of bioweapon efforts. The only currently U.S. FDA-approved vaccine to prevent anthrax in humans is anthrax vaccine adsorbed (AVA), which is protective in several animal models and induces neutralizing antibodies against protective antigen (PA), the cell-binding component of anthrax toxin. However, AVA requires a five-course regimen to induce immunity, along with an annual booster, and is composed of undefined culture supernatants from a PA-secreting strain. In addition, it appears to be ineffective against strains that lack anthrax toxin. Here, we investigated a vaccine formulation consisting of recombinant proteins from a surface-localized heme transport system containing near-iron transporter (NEAT) domains and its efficacy as a vaccine for anthrax disease. The cocktail of five NEAT domains was protective against a lethal challenge of inhaled bacillus spores at 3 and 28 weeks after vaccination. The reduction of the formulation to three NEATs (IsdX1, IsdX2, and Bslk) was as effective as a five-NEAT domain cocktail. The adjuvant alum, approved for use in humans, was as protective as Freund’s Adjuvant, and protective vaccination correlated with increased anti-NEAT antibody reactivity and reduced bacterial levels in organs. Finally, the passive transfer of anti-NEAT antisera reduced mortality and disease severity, suggesting the protective component is comprised of antibodies. Collectively, these results provide evidence that a vaccine based upon recombinant NEAT proteins should be considered in the development of a next-generation anthrax vaccine.

INTRODUCTION

Bacillus anthracis is a Gram-positive, encapsulated, and spore-forming bacterium that is the causative agent of anthrax disease. While B. anthracis can cause disease through cutaneous and gastrointestinal routes, it is the organism’s ability to cause inhalational anthrax that is of particular concern. When inhaled, as few as 7.6 × 10⁵ B. anthracis spores can cause lethal disease in nonhuman primates (1). Due to the generic nature of the early symptoms, such as fever, headache, and muscle aches, infection can be difficult to diagnose, often resembling the flu (2, 3). Without proper antibiotic treatment with ciprofloxacin and doxycycline, inhalational anthrax causes 90% mortality (2, 4). Even with the recommended administration of antibiotics over sixty days, the mortality of inhalational anthrax remains at approximately 50% (2). The low dosage and high mortality make aerosolized B. anthracis spores an ideal tool for bioterrorism. In a 1993 U.S. COTA review, it was estimated that 100 kg of aerosolized spores dropped over Washington, DC, would cause greater mortality than a hydrogen bomb (5). In addition to the organism’s lethality, the spores have remarkable durability. Spores in the soil or animal products can remain viable for decades, and effective sterilization requires either overnight chemical treatment with glutaraldehyde formaldehyde or 30 min of heat treatment at 121°C (6). Even in the face of the intense antibiotic regimen described above, B. anthracis spores have been observed surviving in the lungs for over 75 days in nonhuman primates (7). Spores have also been observed to germinate and cause fatal disease even weeks after initial exposure (2, 7, 8). Due to this persistence, there is a need to develop a long-lasting form of protection, such as a vaccine.

There is currently one FDA-licensed vaccine against B. anthracis, termed anthrax vaccine adsorbed (AVA), which was developed in the 1950s. Since then, its composition has remained mostly the same, consisting of a cell-free filtrate of the avirulent, nonencapsulated B. anthracis strain V770-NPI-R culture in combination with aluminum hydroxide adjuvant (8, 9). This vaccine is highly protective in a wide variety of animal models (10, 11) but has some important limitations. The vaccination regimen requires five doses over eighteen months to confer protection, followed by an annual booster (4). In addition, the vaccine’s composition is poorly characterized. Although there is substantial evidence that the key component of AVA is the protective antigen (PA) subunit of anthrax toxin, the vaccine contains, on average, 176 antigens with approximately 80% similarity between batch preparations (12–17). The identity of most of these proteins remains unknown (12, 18). Finally, while the mechanism of AVA action is due to the neutralization of anthrax toxin, it has been shown that B. anthracis strains lacking toxin can cause anthrax even in vaccinated rabbits (19). These points highlight the need for a defined vaccine that targets a novel pathogenic mechanism and confers lasting immunity against replicating bacilli.

One option is to construct vaccines from factors involved in the growth of bacteria over the course of infection, a concept commonly referred to as nutritional immunity. This tactic involves preventing the bacteria from acquiring the necessary nutrients to grow rather than neutralize the toxin or other effectors that are more directly involved in virulence. One promising avenue to pursue is to target proteins involved in iron acquisition. Iron plays crucial roles in energy production and DNA replication, and it serves as a cofactor for several enzymatic reactions (20). However, very little iron is freely available in circulation in mammals, with approximately 75% to 80% being sequestered inside heme proteins, which are themselves bound by hemoglobin and tucked away inside blood cells (21, 22). Bacteria have, by necessity, evolved systems to overcome these barriers. In Gram-positive bacteria, one common mechanism is the near-iron transporter (NEAT) domain (23, 24). This protein domain is highly conserved across multiple genera, including Staphylococcus, Streptococcus, Listeria, Clostridia, and Bacillus (25). In B. anthracis, five proteins contain NEAT domains (IsdX1, IsdX2, IsdC, Bslk, and Hal), and they collectively function to import host heme (26, 27). These protein domains possess a YXXXY heme binding sequence that is contained in 48% of predicted NEAT domains (28). The importance of this function and its widespread conservation make it an attractive subject for vaccination. Indeed, in 2005, Merck Pharmaceuticals generated the V710 vaccine, whose main component is the IsdB protein, a Staphylococcus aureus hemoglobin receptor. This vaccine proved moderately effective, providing a 20% or greater increased survival in BALB/c mice challenged with multiple clinical S. aureus isolates (29). However, research was abruptly discontinued after phase IIB clinical trials, which reported a higher-than-expected rate of adverse reactions (30). In 2016, Balderas et al. showed that vaccination with a combination of the five NEAT protein domains from B. anthracis conferred complete survival of A/J mice infected with a subcutaneous injection of B. anthracis Sterne spores (31). The study conducted by Balderas et al. was an invaluable proof-of-concept experiment. Specifically, it established that a combination of the 5 NEAT domains, in formulation with Complete Freund’s Adjuvant (CFA)/Incomplete Freund’s Adjuvant (IFA), could protect A/J mice against 10× the 50% lethal dose (LD₅₀) of injected B. anthracis Sterne strain spores. In addition, this protection was conferred with a reduced, 3-immunization-vaccine regimen and was correlated with both an anti-NEAT antibody response and reduced bacterial numbers of CFUs per gram in multiple organs.

This work aims to expand upon these efforts by examining the ability of the proposed NEAT vaccine to protect against inhalational anthrax and whether such a treatment can confer immunity several months after immunization. Furthermore, we determined whether the number of immunogens could be reduced and whether significant immunity could be achieved with the human treatment-approved alum adjuvant. Finally, we also assessed whether protection could be mediated by antibodies via passive immunization. Our findings suggest that a small number of recombinant NEAT protein antigens can generate a protective immune response that increases survival when faced with a lethal challenge of B. anthracis.

RESULTS

Assessment of a NEAT domain vaccine for protection from inhalational anthrax.

Building off the previously established discovery that immunization with a cocktail consisting of each of the five NEAT protein domains was protective against subcutaneous anthrax infection, we tested the ability of the 5-NEAT cocktail to also prevent inhalational anthrax (31). DNA encoding the individual NEAT domains of each of the heme transporters (IsdX1, IsdC, Bslk, and Hal, as well as a polypeptide of all five IsdX2 NEAT domains) were fused to the gene encoding glutathione S-transferase (GST) and expressed in Escherichia coli. Protein then was purified by GST-affinity chromatography, followed by the removal of the GST tag with thrombin and release of the NEAT polypeptide. Figure 1A shows an SDS-PAGE analysis of the purified, GST-free versions of these NEAT polypeptides and demonstrates a typical preparation, as used in the experiments presented throughout this report. Figure 1B illustrates the vaccination regimen used in this experiment. This includes subcutaneous immunization with the NEAT domains on days 0, 14, and 28. The initial immunization on day 0 utilized Complete Freund’s Adjuvant, while the following immunizations on days 14 and 28 were performed with Incomplete Freund’s Adjuvant. Mice were vaccinated with the 5-NEAT domain cocktail (IsdX1, IsdX2, IsdC, Bslk, and Hal) (N = 8), while control mice were vaccinated with only the adjuvant (N = 4). In this experiment, serum samples were collected on day 53 via mandibular bleed. On day 56, mice were anaesthetized with an intraperitoneal injection of Averton (tribromoethanol) and received an intranasal challenge of 5× the LD₅₀. Mice were observed for ten days following challenge. Of the mice immunized with the 5-NEAT vaccine, 62% survived the ten-day challenge period. This compares to a 25% survival rate for the control mice vaccinated with the adjuvant only (p = 0.31) (Fig. 1C). Upon death, the bacterial levels were measured in the heart, lungs, liver, spleen, and kidneys (Fig. 1D). While there was a trend for reduced bacterial levels with vaccination in each of the individual organs, the difference was not statistically significant (p > 0.35). However, combining the counts from all organs for both groups did yield a statistically significant difference (p = 0.0017), indicating that the vaccine reduced bacterial burden (Fig. 1E). Blood was collected from mice 3 days prior to challenge via mandibular bleed. Blood sera were analyzed by enzyme-linked immunosorbent assay (ELISA) to determine if there were differences between the vaccinated and control groups in their antibody responses against each individual NEAT domain (Fig. 1F). It was observed that the sera from vaccinated individuals contained a significantly higher level of anti-NEAT antibodies than the control mice for all NEATs in question (p < 0.0001), with the IsdC NEAT generating a slightly higher response than the other four NEATs.

FIG 1.

Assessment of a 5-NEAT cocktail for protection against inhalational anthrax. (A) SDS-PAGE analysis of purified NEAT proteins, IsdX1, IsdX2, IsdC, Bslk, and Hal, and GST. (B) The vaccination scheme used in this experiment. (C) Mice were immunized with the 5-NEAT cocktail formulated in Complete Freund’s Adjuvant (CFA)/Incomplete Freund’s Adjuvant (IFA) (N = 8) or given CFA/IFA adjuvant alone (N = 4) and challenged with 5× LD₅₀ of B. anthracis Sterne spores. Survival was determined using a Gehan-Breslow-Wilcoxon comparison. (D and E) Box-and-whisker plots of the bacterial levels (CFUs/g) in the heart, lung, spleen, kidney, and liver (D) or the total of all organs (E) following necropsy. (F) ELISA analysis of sera from 5-NEAT-vaccinated animals using each individual NEAT domain as the capture antigen. p values were determined by a 2-way ANOVA. The error bars on bar graphs represent the standard deviations.

Assessment of long-term protection afforded by a NEAT domain cocktail against inhalational anthrax.

We sought to determine if subjects immunized with a cocktail of five NEAT domains would have maintained protection from anthrax when challenged 7 months after vaccination, thereby ensuring that any provided protection would last long enough to cover the late germination of inhaled spores. Mice were immunized with either the 5-NEAT cocktail (4 μg each, 20 μg total) (N = 3) in CFA/IFA (as in Fig. 1) or with a mock control consisting of only 1× phosphate-buffered saline (PBS) and adjuvant (N = 3). Following immunization, animals were challenged with 10× the LD₅₀ of B. anthracis Sterne 7 months after the third boost (Fig. 2A). After 8 days, two of the three animals in the adjuvant-only group had succumbed to anthrax, while all three animals in the NEAT-vaccinated group survived (Fig. 2B). To further test the strength of this protection, animals were challenged a second time, now with 30× the LD₅₀, and observed for an additional 12 days. Even after this additional challenge, all of the 5-NEAT-vaccinated mice survived, compared to none of the adjuvant-only control mice (p < 0.05). Upon the measurement of levels in individual organs, it was observed that while there was a trend between vaccination and reduced number of CFUs per gram, the difference was not statistically significant (p > 0.31) (Fig. 2C). However, just as observed for the results shown in Fig. 1, when the bacterial levels of all organs were combined and compared, the vaccinated group had statistically fewer bacteria (p = 0.0053) (Fig. 2D).

FIG 2.

Effect of NEAT vaccination on long-term protection from inhalational anthrax. (A) The vaccination scheme used in this experiment. (B) Mice were immunized with the 5-NEAT cocktail and CFA (N = 3) or CFA only (N = 3) and challenged after 7 months with 10× the LD₅₀ of B. anthracis strain Sterne spores, followed by a rechallenge with 30× the LD₅₀ 8 days later. Survival was determined using the Gehan-Breslow-Wilcoxon comparison. (C) Bar graph of the bacterial levels (CFUs/g) in the heart, lung, spleen, kidney, and liver. (D) Box-and-whisker plot of the total bacterial levels of all organs following necropsy. (E) ELISA analysis of mouse sera from the vaccinated groups using the purified NEAT domains as capture antigens. (F) ELISA analysis of mouse sera collected at 1 and 7 months postvaccination using the purified NEAT domains as capture antigens. p values were determined by a 2-way ANOVA. The error bars on bar graphs represent the standard deviations.

We next assessed the antibody response 7 months postimmunization and compared this to the antibody response 1 month postimmunization. Blood was collected via cardiac puncture from deceased animals, and sera were analyzed via ELISA to determine if antibodies against the NEAT domains were produced (Fig. 2E). It was observed that the immunized mice presented with a significantly higher anti-NEAT antibody response than the adjuvant-only control group (p < 0.0001). Vaccination with a cocktail of NEAT protein domains yielded strong antibody responses to all five NEATs. IsdX2, which itself contains five NEAT domains in the full-length polypeptide, showed the highest antibody reactivity (p < 0.0001 via Sidak’s multiple-comparison test). This test, combined with the survival rates observed in Fig. 2B, alleviated the concern that the prior anthrax challenge unduly affected the results of this experiment, as the sera collected from control mice did not possess a heightened anti-NEAT antibody response, and the remaining control animals did not survive this second challenge.

In addition, the antibody response from this study (7 months postimmunization) was compared to that of the sera attained 1 month postimmunization of previous animals treated with the 5-NEAT vaccine (Fig. 2F). After 7 months postimmunization, reactivity to all five NEAT domains was still similar in level to that from sera isolated 1 month postimmunization, although IsdC and Hal reactivity both showed a statistically significant reduction (p < 0.0001).

Assessment of the protective ability of individual NEAT immunogens against inhalational anthrax.

To determine if the protective function of the 5-NEAT cocktail observed in Fig. 1 and 2 could be maintained with reduced NEAT domain components, we tested if immunization with each NEAT domain would be efficacious against a lethal dose of B. anthracis. Murine hosts (N = 7 to 8) were immunized with 4 μg each NEAT domain (IsdX1, IsdX2, IsdC, Bslk, or Hal) in combination with Complete Freund’s Adjuvant for the first immunization and Incomplete Freund’s Adjuvant for the second and third, as described above. The control group consisted of animals vaccinated with only adjuvant (N = 4) (Fig. 3A). On day 53, sera were collected from the mice via mandibular bleed before challenge. All mice were challenged intranasally with 5× the LD₅₀ of Sterne spores and observed over a period of 10 days (Fig. 3B). Rather strikingly, each of the five NEAT domains, when used as individual immunogens, provided some level of protection against lethal challenge. The most significant protection came from three immunogens: the two secreted hemophores, IsdX1 and IsdX2, and the surface-anchored BslK (p = 0.0129). Upon measurement of the levels of bacilli, the NEAT-vaccinated groups showed a trend of reduced levels compared to those of the control (Fig. 3C). However, the difference was not statistically significant in any group. However, when organs were grouped together, there was a significant reduction (p > 0.21) in bacillus levels in those mice that were treated with IsdX1, IsdX2, or Bslk, the same three treatment groups with the highest rate of survival (Fig. 3D). As expected, and consistent with the previous two vaccinations described in this study, each experimental group’s sera reacted to its immunogen upon ELISA, indicating once again that antibodies are made against each of the NEAT domains (Fig. 3E). Interestingly, there was significant cross-reactivity between each of the NEATs. For example, anti-IsdX1 sera cross-react strongly with all NEATs other than Hal. This is not too surprising, considering the high degree of homology between the amino acid sequences of the NEAT domains. IsdX1 also showed impressive individual protection, as did IsdX2. Despite the high survival of IsdX2-vaccinated mice, few anti-IsdX2 antibodies cross-react to other NEATs. These results, along with the trend in survival, imply that IsdX1 and IsdX2 are key components for a reduced-component vaccine.

FIG 3.

Assessment of each NEAT domain in the protection from inhalational anthrax. (A) The vaccination scheme used in this experiment. (B) Mice were immunized with individual NEAT domains (IsdX1, IsdX2, IsdC, BslK, or Hal; 4 μg each) in CFA/IFA (N = 8 for each) or CFA only (N = 4) and challenged with 5× the LD₅₀ of Sterne spores 1 month after vaccination. Survival was determined using a Gehan-Breslow-Wilcoxon comparison. (C and D) Box-and-whisker plots of the bacterial levels (CFUs/g) in the heart, lung, spleen, kidney, and liver (C) or the total of all organs (D) following necropsy. (E) ELISA analysis of the sera from each NEAT immunization for each NEAT domain as the capture antigen, represented as a black-and-white heat map of absorbance values. p values were determined by 2-way ANOVA.

Effect of reducing the immunogen formulation to three NEAT domains in the protection against inhalational anthrax disease.

The results shown in Fig. 3 suggest that a cocktail of IsdX1, IsdX2, and BslK alone is enough to confer protection against anthrax disease; thus, we tested the hypothesis that the 3-NEAT cocktail (IsdX1, IsdX2, and BslK; total of 12 μg) (N = 10) would be as efficacious as the 5-NEAT cocktail (IsdX1, IsdX2, IsdC, BslK, and Hal; total of 20 μg) (N = 6) in protection from inhalational anthrax. Control mice (N = 10) were immunized with Complete Freund’s Adjuvant containing the protein GST, the tag from which the NEAT domains are purified by affinity chromatography. Animals then were challenged intranasally with B. anthracis Sterne spores. After 8 days, only one animal in the control group had succumbed to the infection; thus, we challenged them again with Sterne spores at approximately 30× the LD₅₀ to induce a higher rate of mortality, followed by the observation of animals for another 12 days (Fig. 4A). Compared to the control group, of which 30% survived, those mice vaccinated with the 5-NEAT cocktail demonstrated 66% survival (p = 0.091), and mice vaccinated with the 3-NEAT cocktail exhibited 70% survival (p = 0.44) (Fig. 4B). It was observed that while vaccination was associated with reduced bacterial levels in all organs, the difference was not statistically significant for any organ in any group (p > 0.359) (Fig. 4C). When organs are grouped together, though, there was a significant reduction in the average number of CFUs per gram in both 5-NEAT (p = 0.0008)- and 3-NEAT (p < 0.0001)-treated mice compared to levels for control mice (Fig. 4D). Sera were once again assessed for reactivity to NEAT domains by ELISA (Fig. 4E). The results show not only that the level of 3-NEAT sera was higher than that of control sera for reactivity against every NEAT (p < 0.0001) but also that the 3-NEAT sera demonstrated significantly higher (p < 0.0001) reactivity against every NEAT than against 5-NEAT sera, with the exception of IsdX2 (p = 0.0562). This included reactivity against IsdC and Hal, the two NEATs that were excluded from the 3-NEAT vaccine formulation, further suggesting that there are similar epitopes shared between all five NEAT domains. Survival and bacterial levels were also compared between the treated (all NEAT cohorts) and the GST control-vaccinated animals. Although the overall trend was toward protection, the differences were not significant (p = 0.156) (Fig. 4F). However, it was observed that there was a statistically significant reduction in the bacterial levels of the kidneys (p = 0.0045) and the liver (p = 0.0331) of vaccinated mice compared to those of mice only given GST (Fig. 4G).

FIG 4.

Assessment of a 3-NEAT cocktail for protection from inhalational anthrax. (A) The vaccination scheme used in this experiment. (B) Mice were immunized with either the 5-NEAT cocktail in CFA/IFA (N = 6), 3-NEAT cocktail (IsdX1, IsdX2, and Bslk) in CFA/IFA (N = 10), or GST with CFA/IFA (N = 10), challenged 1 month postimmunization with 10× the LD₅₀ of B. anthracis Sterne spores, and rechallenged 8 days later with 30× the LD₅₀ of spores. Survival was determined using a Gehan-Breslow-Wilcoxon comparison. (C and D) Box-and-whisker plots of the bacterial levels (CFUs/g) in the heart, lung, spleen, kidney, and liver (C) or the total of all organs (D) following necropsy. (E) An ELISA analysis of mouse sera using the NEATs as capture antigens for the three treatment groups. The asterisk indicates that X2 reactivity not significantly different (p > 0.05). (F) Same data as those observed in panel B but now plotted as treated (all NEAT cohorts grouped together) versus the control. (G) Box-and-whisker plot of the bacterial level of data observed in panel C but now plotted as treated (all NEAT cohorts grouped together) versus the control. p values were determined by 2-way ANOVA. The error bars on bar graphs represent the standard deviations.

Assessment of the efficacy of a cocktail of IsdX1, IsdX2, and BslK with alum in protection from inhalational anthrax disease.

To this point, the available data indicated that the 3-NEAT cocktail (containing IsdX1, IsdX2, and BslK) is as effective as the 5-NEAT cocktail in protecting from inhalational anthrax. Here, we made two additional modifications to help us understand the degree of efficacy of the NEAT vaccine in attenuating anthrax. The replicate number was increased, and alum approved for human treatment was used as the adjuvant. Murine hosts (N = 17) were immunized with 4 μg of each NEAT domain (12 μg total) formulated with a 1:2 ratio of alum to antigen. Control mice (N = 20) were immunized with 12 μg GST formulated with a 1:2 ratio of alum to antigen. Subjects then were challenged intranasally with 30× the LD₅₀ of B. anthracis Sterne spores and observed over ten days (Fig. 5A). At the end of the challenge, 65% of mice vaccinated with the 3-NEAT (IsdX1, IsdX2, and Bslk) vaccine survived. This contrasts with survival of only 10% of the control groups, which was a statistically significant increase in survival with vaccination (p = 0.0004) via the Genhan-Breslow-Wilcoxon test (Fig. 5B). This increase in survival was correlated with a statistically significant reduction (p < 0.0013) in the bacterial levels of all observed organs, save the lungs (p = 0.1065) (Fig. 5C). Grouping of organ data showed that there was a significant reduction (p < 0.0001) in bacterial burden in the vaccinated mice (Fig. 5D). ELISA analysis of sera from both groups indicated that vaccination with the three NEATs IsdX1, IsdX2, and Bslk yielded anti-NEAT antibodies that were not observed for the control vaccinated animals (p < 0.0001) (Fig. 5E).

FIG 5.

Assessment of a 3-NEAT cocktail with alum in protection against inhalational anthrax. (A) Schematic of the experimental scheme used in this study. (B) Mice were immunized with either the 3-NEAT cocktail (IsdX1, IsdX2, and BslK) in alum (N = 17) or GST in alum (N = 20) and then challenged 1 month after vaccination with 30× the LD₅₀ of Sterne spores. Survival data were determined using a Gehan-Breslow-Wilcoxon comparison. (C and D) Box-and-whisker plots of the bacterial levels (CFUs/g) in the heart, lung, spleen, kidney, and liver (C) or the total of all organs (D) following necropsy. (E) ELISA analysis of pooled sera comparing antibody reactivity against IsdX1, IsdX2, and Bslk. p values determined by 2-way ANOVA. The error bars on bar graphs represent the standard deviations.

Assessment of the level of protection from anthrax disease conferred by anti-NEAT antibodies.

The available data thus far indicate a relationship between the protection from anthrax conferred by NEAT domain immunization and the production of antibodies against those domains in vaccinated individuals. To determine if the anti-NEAT antibodies themselves were protective, we assessed the ability of antibodies raised against IsdX1, IsdX2, and Bslk to passively protect a murine host against a lethal dose of B. anthracis. Murine hosts (N = 10) to be tested for passive transfer protection were given a combination of anti-IsdX1, -IsdX2, and -Bslk antisera and then infected 24 hr later with B. anthracis Sterne spores (approximately 30× the LD₅₀) via the inhalational route. Control mice (N = 10) were given an equivalent volume of antisera generated against GST. Mice then were given a second dose of antisera 24 hr after infection (Fig. 6A). At the end of the ten-day observational period, 30% of the anti-NEAT antiserum-treated mice survived compared to no survivors in the group treated with the anti-GST antisera (Fig. 6B), a difference that was statistically significant (p = 0.0223), as calculated by the Genhan-Breslow-Wilcoxon test. The examination of the bacterial levels in murine hosts treated with anti-NEAT antibodies did show reduced bacterial levels in each of the five organs tested, but none of the differences were statistically significant from the control (p > 0.1056) (Fig. 6C). Grouping of the organ data, however, revealed there was a statistically significant reduction (p = 0.0021) in organ bacilli (Fig. 6D). Blood was collected via cardiac puncture from moribund animals, and the pooled sera were tested by ELISA and compared to the input antisera. Reactivity of the output sera to IsdX1, IsdX2, and Bslk protein domains was detected, albeit at low levels compared to that of input sera, to be significantly higher (p < 0.0001) than that of GST control mice (Fig. 6E). Collectively, these results show that at least some of the protection from anthrax disease is conferred by anti-NEAT antibodies, but the relationship between antibody titer and protection remains to be quantified.

FIG 6.

Assessment of the ability of anti-NEAT antisera to confer protection against inhalational anthrax. (A) A schematic showing the passive immunization scheme used in this experiment. (B) Mice were treated on day −1 and +1 with 90 μl of antisera generated against IsdX1, IsdX2, and Bslk (N = 10) or against GST (N = 10) and then challenged on day 0 with 30× the LD₅₀ of Sterne spores. Survival data were determined using a Gehan-Breslow-Wilcoxon comparison. (C and D) Box-and-whisker plots of the bacterial levels (CFUs/g) in the heart, lung, spleen, kidney, and liver (C) or the total of all organs (D) following necropsy. (E) ELISA analysis of the input antisera and the output sera recovered from both groups and tested for reactivity against IsdX1, IsdX2, and Bslk or the GST control protein as the capture antigen. p values determined by 2-way ANOVA. The error bars on bar graphs represent the standard deviations.

DISCUSSION

In this study, we sought to build upon the report of Balderas et al. (31), which found that immunization with recombinant NEAT protein domains protected a murine host from subcutaneous anthrax. The pertinent findings of the current study are the following: (i) a cocktail of five NEAT protein domains was efficacious at preventing anthrax after a lethal intranasal challenge of strain Sterne spores, (ii) the 5-NEAT cocktail retained its protective effect for at least 7 months following the final immunization, (iii) a vaccine formulation consisting of three of the NEATs (IsdX1, IsdX2, and BslK) appears to be equally as effective as a formulation consisting of all five NEAT domains, (iv) the 3-NEAT vaccine is effective when formulated with alum approved for human treatment as the adjuvant, and finally, (v) anti-NEAT antisera are sufficient to attenuate disease when delivered via passive transfer. In addition, the vaccine treatment was shown to be efficacious within the sixty-day time frame of the postexposure antibiotic regimen. Collectively, these results build upon a previous report and support the further development of recombinant heme transporters as a potentially effective vaccine at preventing anthrax disease.

One avenue to investigate is the mechanism of protection induced by the immunogens reported here. In this capacity, there may be several ways a NEAT-based vaccine functions to attenuate anthrax. This includes the inhibition of the NEAT domains’ ability to take up heme (perhaps by blocking heme access to the heme-binding pocket), the induction of opsonophagocytosis as the anti-NEAT antibodies bind the surface of bacilli, the stimulation of complement-based killing via the antibody-induced membrane attack complex, or cooperation with natural killer cells for antibody-dependent cellular cytotoxicity (ADCC). However, the ADCC mechanism primarily clears host cells that are either cancerous or infected with a virus and is unlikely to be the mechanism observed here. The current data do not differentiate between the first three possibilities. Balderas et al. found that anti-NEAT antibodies added to a B. anthracis culture whose growth was dependent on taking up heme in an iron-depleted environment did not prevent iron-dependent growth (31). These results have been confirmed once more in this study (data not shown). Thus, it seems that antibodies generated by this vaccine formulation provide protection by stimulating cellular immunity via opsonophagocytosis, complement killing via the membrane attack complex, or both. Given that these studies were performed in A/J mice that lack the C5 component of complement yet protection was still achieved, the likely mechanism of action observed here is the induction of opsonophagocytic antibodies against growing bacilli. Future studies will test this hypothesis.

A second area of investigation is to determine the efficacy of this vaccine in protecting against other strains of B. anthracis or even other Gram-positive bacteria. These studies were performed with animal challenges against B. anthracis strain Sterne 34F2. This strain is useful for a number of reasons, principal of which is because such studies can be conducted under biosafety level 2 (BSL2) instead of the more challenging BSL3 setting. However, Sterne’s nonencapsulated nature means that it is more vulnerable to cell wall-targeting immune responses. If these antibodies are only efficacious when binding the cell wall-anchored or adjacent NEATs, then this protection may be reduced or eliminated when testing against strains where the capsule is present. For example, the vaccine proposed here should be tested against a nontoxigenic encapsulated B. anthracis strain, such as the Vollum strains previously used to demonstrate toxin-independent virulence (19). Future studies also should test against the fully virulent encapsulated and toxigenic Ames strain of B. anthracis. Despite these concerns, there is previous evidence supporting NEAT-based immunity against these pathogens, as guinea pigs infected with Ames bacilli have been observed to generate antibodies against IsdX1 and IsdX2 (32). In addition to B. anthracis, Gram-positive bacteria that utilize NEAT proteins include Staphylococcus aureus and multiple species of Streptococcus and Listeria pathogens (23, 24). If a vaccine generated against the B. anthracis NEAT domains can provide cross-protection against bacteria from a different genus, such a strategy of targeting conserved nutrient transporters may become more commonplace. Future work aims to provide answers to these questions.

Finally, a third avenue of development would be to enhance the formulation by including other antigens that target different pathogenic processes important in anthrax disease. The most striking would be to include the protective antigen subunit of anthrax toxin. Such a formulation could provide a powerful one-two punch in combating anthrax by targeting both the growing bacilli and the most substantial secreted virulence factor. While recombinant PA vaccines are in development, formulation with NEAT immunogens may prove to be the missing step to maximize protection.

This 3-NEAT vaccine holds an interesting comparison to that used by AVA. AVA, as previously discussed, is a cell-free filtrate of the V770-NPI-R B. anthracis strain, and, as such, contains a wide range of components and potential antigens. Despite its lack of purity, the vaccine is undoubtedly effective. In nonhuman primates, AVA treatment results in 60% to 100% survival of subjects challenged with up to 840× the LD₅₀ (33). In New Zealand White rabbits, even 1:16 diluted AVA vaccination results in 96% survival, compared to 0% of the control (34). Finally, it was seen in mice, specifically the A/J strain used in this study, that vaccination with AVA induces 100% survival when challenged with 20× LD₅₀ inhalational Sterne strain 7702 spores (35). One concern, however, is that AVA is ineffective in protecting rabbits against nontoxigenic strains of B. anthracis (19). In addition, AVA requires a lengthy vaccine regimen of five immunizations over eighteen months, followed by an annual booster. The target NEAT antigens used in this vaccine are widely conserved across Gram-positive bacteria, and the protection observed here was obtained using a vaccine regimen that took only three immunizations in the 2 months before challenge. Thus, while the survival rate obtained with this 3-NEAT vaccine is not as high as that of previous AVA experiments, it seems worth investigating whether using this NEAT domain formulation, perhaps at higher doses or as one fused polypeptide, confer protection as effective as that of AVA while being more economical and safe to use.

A vaccine developed by Merck (V710 vaccine) targeting S. aureus consisted of the iron acquisition protein IsdB, which contains a NEAT domain, and showed strong success in protecting against S. aureus infection in mouse models as well as inducing antibodies in nonhuman primates (29). In two separate phase I studies, the vaccine proved to be safe for injection into humans, with no adverse events reported over the course of either study (36). However, V710 met with challenges during its phase IIb/III trial. In that trial, it was used as an experimental preventative against S. aureus infection in patients who were undergoing cardiothoracic surgery (30). It was observed that while those who received V710 generated a robust antibody response, they were significantly more likely to experience multiorgan failure. This resulted in a higher rate of S. aureus-related deaths in the V710-vaccinated group than in the placebo group (15/73 versus 4/96, respectively). A retrospective analysis determined that there was a correlation between decreased levels of interleukin-2 (IL-2) and IL-17a in patient blood samples, collected at hospital admission and prior to operation, and fatal S. aureus infection in V710-immunized, but not placebo-treated, patients (37). While a clear mechanism between the increased anti-IsdB antibody response and the fatal patient outcome was not established, it was hypothesized that the immunization of a host with decreased IL-2 presence could result in a misdirected or suboptimal Th-2 immune response. While the mechanism for this B. anthracis 3-NEAT vaccine has yet to be established, it may be based on opsonophagocytosis, while the IsdB vaccine protection has been associated with a strong Th-17 response (38). Nonetheless, it would be informative to test these ideas through the utilization of IL-2 and IL-17a knockdown mice.

A vaccine that clears the host of replicating bacilli would ensure that any spores that survive long enough to germinate would not have the opportunity to multiply before removal. This is especially important when considering the utility of a vaccine in a postexposure setting. Reports from rabbit and monkey studies indicate that spores germinate anywhere from days to months after inhalation (7). Thus, the current postexposure protocol utilizes a sixty-day antibiotic regimen. However, the long-term continuation of antibiotics is associated with adverse side effects, resulting in reduced compliance with antibiotic regimens. In fact, after the anthrax attacks in 2001, only 40% of the approximately 5,000 people potentially exposed to the mailed spores finished the full course of antibiotics (39). Thus, immunization with a vaccine that clears germinated spores and growing bacilli could be highly effective in protecting large populations where compliance is hard to maintain. A key way to test this vaccine would be to observe how well it performs in a postexposure treatment plan. The current immunization regimen places the onset of immune protection within the sixty-day antibiotic treatment recommended by the CDC after an exposure event. If this vaccine can provide significant protection in organisms previously challenged with Ames strain B. anthracis, even after an antibiotic regimen, then there will be a strong case for a NEAT-based vaccine, perhaps in combination with AVA, to be considered another option in combating anthrax disease.

MATERIALS AND METHODS

Protein purification from E. coli.

All NEAT protein domains were expressed as fusions bound to glutathione-S-transferase (GST), expressed by E. coli mutants, and purified as previously described by Balderas et al. (31). Collected protein domain elutions were examined for purity and concentration via SDS-PAGE and Bradford assay. Purified NEATs then were concentrated using gravity filtration (Vivaspin Turbo 15; Sartorious). All elutions then were examined by SDS-PAGE, and NEAT protein domain concentration was determined using the Bradford assay. In cases such as that of IsdX2, where multiple bands were observed, it was believed that, as previously observed, these lower-molecular-weight products were breakdown products of the target IsdX2, owing to their increase in production in the absence of protease inhibitors (32, 40).

Mouse strains.

All mice used in this study were female A/J mice (Jackson Laboratory) due to their deficiency in complement factor C5. This complement deficiency meant that the mice would experience fulminant disease when infected with Sterne strain B. anthracis, which does not possess the capsule that would protect it from complement and the membrane attack complex (MAC). This mouse strain has been well documented as an accepted model for these infections with the Sterne strain of B. anthracis (41, 42).

Vaccination studies.

Six-week-old female A/J mice, housed in groups of 3 to 5, were given three subcutaneous injections of 4 μg protein on days 0, 14, and 28. In cocktail vaccines, mice were given 4 μg of each of the 5-NEAT domains (IsdX1, IsdX2, IsdC, Bslk, and Hal), for a total of 20 μg, or of the 3-NEAT domains (IsdX1, IsdX2, and Bslk), for a total of 12 μg. The NEAT protein domains then were mixed 1:1 with Complete Freund’s Adjuvant (CFA) for the first injection and 1:1 with Incomplete Freund’s Adjuvant (IFA) for the second and third injections. Complete Freund’s Adjuvant, an oil emulsion containing mycobacteria, is well established for vaccine studies in mice. CFA is regularly used for the initial immunization, and IFA, lacking the mycobacterial component, is used for subsequent immunizations. In the final vaccination study described here (Fig. 5A), mice were vaccinated with a NEAT cocktail consisting of 4 μg each of IsdX1, IsdX2, and Bslk in a 1:2 volume ratio with alum for all three immunizations. Control groups were vaccinated with equivalent doses of GST.

B. anthracis strains and spore preparation.

The B. anthracis strain used in this study is Sterne strain 34F2. This strain lacks the poly-d-glutamic acid capsule. B. anthracis starter cultures were grown overnight in LB broth. Overnight cultures then were transferred to 30 ml modified G medium (no glucose) and incubated with shaking for a minimum of 36 hr to induce sporulation. Spores then were concentrated by centrifugation and washed a minimum of four times with sterile water. At the end of the spore generation process, and before each use of the spores, the preparations were heat shocked at 65°C for approximately 45 min to kill vegetative bacteria that may have germinated. Viable spores then were quantified by serial dilution and plating of samples on LB-agar plates.

Infections with B. anthracis.

On day 56 of the study, mice were anesthetized with an intraperitoneal injection of Averton. Once anesthetized, mice were infected intranasally with 5×, 10×, and/or 30× the calculated LD₅₀ of spores from the Sterne strain of B. anthracis (approximately 3.16 × 10⁵ spores). Spores were suspended in approximately 40 μl 1× PBS. Mice were monitored twice a day for a minimum of 10 days. At the end of the study, or when observed to be moribund, mice were euthanized and necropsied to collect heart, lungs, liver, kidneys, and spleen. Organs were weighed, homogenized, and plated to determine the number of bacteria or CFUs per gram. Moribundity was determined through the observation of multiple features, including activity level, posture, coat appearance, staggered movement, and hyperpnea. Blood was collected either via mandibular bleed 3 days prior to infection or by cardiac puncture posteuthanasia, where specified. Animals were handled in accordance with the Guide for the Care and Use of Laboratory Animals (43) and with approval by the Institutional Animal Care and Use Committee (protocol number AN-5177) at Baylor College of Medicine.

Passive transfer studies.

A/J mice were vaccinated with either IsdX1, IsdX2, Bslk, or GST (formulated with alum) as stated in the above-described immunization scheme. Upon the completion of the full vaccination time course, immunized mice were sacrificed and antisera harvested and pooled. Naive, 14-week-old female A/J mice then were treated with either the antisera generated against GST or with a combination of the antisera generated against IsdX1, IsdX2, and Bslk NEAT domains. Antiserum treatments were matched by volume between groups, with mice receiving 90 μl antisera via intraperitoneal injection 24 hr before and after intranasal challenge with 30× the LD₅₀ of Sterne spores (for a total of approximately 180 μl of antisera per mouse). Mice were monitored twice a day for ten days after infection. At the end of the study, or when animals became moribund, mice were euthanized and necropsied to collect heart, lungs, liver, kidneys, and spleen. Organs were weighed, homogenized, and plated to determine the number of bacterial CFUs per gram. Blood was collected via cardiac puncture and serum separated to observe antibody reactivity via indirect ELISA.

ELISA.

Indirect ELISAs were performed using ThermoFisher 96-well Nunc plates. Plates were coated with 100 μl 20-μg/ml protein solution in 1× PBS and allowed to bind overnight at 4°C. The following day, plates were washed once with 1× PBS–Tween 20 and then coated with 150 μl 5% milk solution in 1× PBS. Plates were incubated while rocking for 1 hr and then washed 3× with 1× PBS–Tween 20. Primary antibody was added via 10 μl serum diluted into 90 μl of 5% milk. All sera were diluted 10,000-fold before analysis by ELISA. Plates then were incubated with gentle rocking overnight at 4°C, followed by washing three times with 1× PBS–Tween 20. A volume of 100 μl secondary antibodies (anti-mouse IgG generated in rabbit conjugated to horseradish peroxidase), diluted 1:5,000 in 1× PBS, next was added to each well and the entire sample gently rocked at room temperature for 1 hr. Plates were washed 3 times with 1× PBS–Tween 20 before 100 μl TMB solution was added to the wells and allowed to incubate at room temperature for 5 to 10 min. The reaction was stopped by adding 50 μl 2 M H₂SO₄ to the wells. The absorbance of each well was measured at 450 nm using a BioTek Synergy HT plate reader. All experiments were performed with three replicates, from which a mean and standard deviation were calculated. In experiments that included a GST control, ELISA readouts were normalized to anti-GST reactivity.

Statistical methods.

All survival curves were compared using Genhan-Breslow-Wilcoxon curve comparison. Determining significant difference between individual organs in grouped organ data of numbers of CFUs per gram was performed using 2-way analysis of variance (ANOVA). Determining significant difference between combined organ data was performed using Welch’s t tests. Determining significant difference between individual anti-NEAT antibody reactivity from ELISA was performed using 2-way ANOVA.