Receptor-enhanced Immunogenicity of Anthrax Protective Antigen is Primarily Mediated by Capillary Morphogenesis Protein-2

Abstract

Anthrax protective antigen (PA), the receptor-binding component of anthrax toxin, elicits toxin-neutralizing antibodies which provide protection against anthrax disease. PA binds to two mammalian receptors, capillary morphogenesis protein-2 (CMG2) and tumor endothelial marker-8 (TEM8). We previously observed that binding of PA to its receptors plays a role in eliciting a strong toxin-neutralizing antibody response. In this study, we examined the roles that individual receptors play in mediating the toxin-neutralizing antibody response. Mice immunized with PA that binds preferentially to CMG2 elicited a toxin-neutralizing antibody response similar to that elicited by wild-type PA, whereas the antibody response elicited by PA that binds preferentially to TEM8 was significantly lower. Also, the toxin-neutralizing antibody response elicited by wild-type PA in CMG2-null mice was found to be significantly lower than that induced in CMG2-sufficient mice, further supporting a predominant role for the CMG2 receptor in mediating a protective antibody response to PA.

1. Introduction

Anthrax, a disease caused by Bacillus anthracis, is associated with significant morbidity and mortality. Disease progression is mediated by anthrax toxin [1]. Vaccines prevent anthrax by inducing antibodies capable of neutralizing the action of the toxin, thereby halting the disease process [2-4]. Because neutralizing antibodies play a pivotal role in protection against disease, an understanding of the factors that lead to an optimal neutralizing antibody response is important for vaccine design.

Anthrax toxin is composed of a receptor-binding subunit known as protective antigen (PA) which can combine with two different enzymatically active subunits, lethal factor (LF) and edema factor (EF), forming lethal toxin (PA + LF) and edema toxin (PA + EF), respectively [5]. The toxin acts when PA binds to cells and is proteolytically cleaved to a 63-kDa form which then oligomerizes and binds LF and/or EF. The holotoxin is then endocytosed. When the toxin is exposed to the acidic environment of the endosome, PA forms a pore through which LF and EF pass to enter the cytoplasm of the cell, their site of catalytic action. LF is a metalloprotease that cleaves MAP-kinase kinases and regulator components of the PI3K pathway [6]. EF is a potent adenylate cyclase [7]. Most anthrax vaccines are based primarily on PA, which by itself is nontoxic and non-catalytic but is capable of inducing toxin neutralizing antibodies which have been demonstrated to correlate with protection against disease [2-4].

Recently, we demonstrated that loss of the ability of PA to bind to its cellular receptors results in a substantial decrease in its immunogenicity. In that work, we found that wild-type PA elicited robust neutralizing antibody titers in mice; however, a mutant PA that was incapable of binding to cellular receptors elicited neutralizing titers that were over an order of magnitude lower than those elicited by wild-type PA [8]. We further showed that PA capture by immune cells, such as antigen-presenting cells (APCs), through PA’s normal receptor uptake pathway yielded a significantly higher neutralizing antibody response as compared to the response seen when PA was engineered to bind to and enter cells by an alternate receptor pathway, i.e., the receptor pathway normally utilized by cholera toxin. Of note, while the neutralizing antibody response was higher when PA utilized its own receptor pathway as compared to the alternate pathway, the total antibody response, which includes non-neutralizing antibodies, was similar for the two capture pathways, indicating that the neutralizing antibody response was particularly sensitive to the specific pathway used by immune cells to capture PA [8].

As a first step in understanding the mechanisms leading to this receptor-mediated enhancement of the neutralizing (protective) antibody response to PA, we investigated the identity of the cellular receptor(s) involved in this process. Two cellular receptors for PA have been identified, tumor endothelium marker-8 [TEM8, also known as anthrax toxin receptor 1 (ANTXR1)] and capillary morphogenesis protein-2 [CMG2, also known as anthrax toxin receptor 2 (ANTXR2)] [5]. CMG2 has been identified as the major receptor mediating lethality of anthrax toxin [9], while the identity of the receptor that mediates enhanced immunogenicity of PA is not known. In this study, we used receptor-specific mutants of PA as well as CMG2-null mice to elucidate the identity of the major receptor mediating enhanced immunogenicity of PA.

2. Methods

2.1. Ethics statement

Animal procedures were performed at the facilities of the U.S. Food and Drug Administration, Silver Spring, MD and the National Institute of Allergy and Infectious Diseases (NIAID) at the National Institutes of Health (NIH), Bethesda, MD. Both facilities are accredited by the Association for Assessment and Accreditation of Laboratory Animal Care International (AAA-LAC). The animals at both facilities were handled in strict accordance with the recommendations outlined in the National Research Council’s Guide for the Care and Use of Laboratory Animals as well as with the Public Health Service Policy on Use of Lab Animals, and the Animal Welfare Act. The animal protocols were approved by the Institutional Animal Care and Use Committee of either the U.S. Food and Drug Administration (200127) or NIAID (LPD8E) depending on where the studies were conducted.

2.2. Mutation and cloning of the pagA gene, gene expression and protein purification

Mutated pagA genes of B. anthracis encoding a PA mutant (I656S) that binds preferentially to CMG2 and a PA mutant (E654M, R659S and M662R) that binds preferentially to TEM8, that have been described and characterized previously [10], were synthesized at GenScript (NJ, USA). Wild-type PA, CMG2-specific PA mutant and TEM8-specific PA mutant genes were each cloned in a pET-22b plasmid carrying an N-terminal pelB signal sequence for periplasmic localization. These different PA constructs were expressed and purified in E. coli BL21(DE3) from the periplasmic content essentially as previously described [11]. Purified PA proteins were 95 to 99% pure as determined by SDS-PAGE.

2.3. Immunization studies

Groups of 10 mice (six-week-old female C57BL/6 mice) were immunized once intraperitoneally with 15 μg purified wild-type PA, CMG2-specific PA mutant and TEM8-specific PA mutant proteins in PBS (without any adjuvant) or injected with PBS alone. Mice were bled 28 days postimmunization. Alternatively, a total of 34 CMG2-null mice (CMG2^−/−) on a C57BL/6J background [9] and 30 CMG2-sufficient i.e., wild type; CMG2^+/+) were immunized subcutaneously with 5 μg of wild-type PA protein (without any adjuvant) on days 0 and 30. Mice were bled 2 weeks after the 2^nd immunization. Serum samples were analyzed individually using an enzyme-linked immunosorbent assay (ELISA) and a toxin-neutralizing antibody (TNA) assayw.

2.4. ELISA and TNA assay

Serum samples from different immunization studies were analyzed for the total anti-PA antibodies using the ELISA and for the toxin neutralizing antibodies by the TNA assay essentially as described previously [11] . For ELISA and TNA assays, a four-parameter logistic regression model was used to fit the data points generated when the absorbance was plotted against the reciprocal of the serum dilution. For ELISA, the inflection point of the curve, which indicates 50% response, was reported as the antibody titer. For TNA, the inflection point, which indicates 50% neutralization, was reported as the effective dilution at 50% inhibition (ED50) [12], which is defined as the reciprocal of the serum dilution at 50% inhibition.

2.5. Statistical analyses

Statistical analyses were conducted using GraphPad Prism software (version 6, GraphPad Software, La Jolla, CA). The conventional P = 0.05 was used as the level of significance. The data presented in Figure 1 passed the normality test and therefore an unpaired t-test was performed and geometric means are reported. The data presented in Figure 2 did not pass the normality test and therefore a non-parametric, Mann-Whitney test was performed and medians are reported.

Figure 1:

Antibody titers of mice immunized with wild-type PA and receptor-specific PA mutants. Three groups (10 mice in each group) were immunized with either wild-type PA, CMG2-specific PA mutant and TEM8-specific PA mutant (15 μg of antigen/mouse). Twenty-eight days postimmunization, the mice were bled, and serum samples were analyzed by the TNA assay and ELISA as described in Methods. (A) Total anti-PA antibody titers for mice as determined by ELISA. (B) TNA titers for mice, each symbol represents the TNA titer value (expressed as ED50) for an individual mouse. An unpaired t-test was used to determine statistical significance, P value is indicated. n.s., not significant. The horizontal line represents the geometric mean for the group of mice in each panel. The dotted line represents the limit of quantitation of the ELISA and TNA assay in their respective panels. This figure shows the results of one experiment and is representative of three independent immunization experiments.

Figure 2:

Antibody titers in CMG2-sufficient mice and CMG2-null mice immunized with wild-type PA. A total of 30 CMG2-sufficient mice (CMG2^+/+) and 34 CMG2-null mice (CMG2^−/− were immunized with wild-type PA (5 μg of antigen/mouse) on days 0 and 30. Two weeks after the 2^nd immunization, mice were bled, and serum samples were analyzed by the ELISA and TNA assay as described in Methods. (A) Total anti-PA antibody titers for each mouse as determined by ELISA (B) TNA titers for mice, each symbol represents the TNA titer value (expressed as ED50) for an individual mouse. Four independent immunization experiments were performed. This figure shows the analysis of the pooled data obtained for all four different immunization experiments. The Mann-Whitney test was used to determine statistical significance, P value is indicated. The horizontal line represents the median for the group of mice in each panel. The dotted line represents the limit of quantitation of the ELISA and TNA assay in their respective panels.

3.0. Results and Discussion

Identification of which cellular receptor for PA plays the major role in enhancing the ability of PA to induce a neutralizing antibody response would provide a better understanding of the immunological pathways that lead to robust protective antibody responses. To identify the receptor, we first utilized mutant forms of PA that bind preferentially to either CMG2 or TEM8 [10]. The CMG2-specific PA mutant that we used had a single substitution of serine for isoleucine at position 656 (I656S). The TEM8-specific PA mutant had three mutations: E654M, R659S and M662R. These mutant proteins were originally described by Chen et al [10] who extensively characterized their receptor specificity by examining their binding to Chinese hamster ovary (CHO) cells that selectively overexpressed either CMG2 or TEM8 receptor. We immunized groups of mice with either wild-type PA, the CMG2-specific PA mutant that preferentially binds to CMG2 or the TEM8-specific PA mutant that preferentially binds to TEM8. After 28 days, we examined the antibody responses induced by the different PA proteins. As seen in Figure 1, both the total anti-PA antibody titers and the toxin-neutralizing antibody response induced by the CMG2-specific PA mutant were not significantly different from that of wild-type PA. In contrast, both the total anti-PA antibody titers and the toxin-neutralizing antibody response to the TEM8-specific PA mutant were significantly lower than that of wild-type PA. By targeting PA to cells of the immune system, CMG2 plays a major role in generating both the total antibody response as well as the neutralizing antibody response to PA.

To further support this conclusion, we compared the total anti-PA antibody response and the toxin-neutralizing antibody of CMG2-null mice after immunization with wild-type PA to that of CMG2-sufficient mice. As seen in Figure 2, both the total anti-PA antibody titers and the toxin-neutralizing antibody response induced by wild-type PA in CMG2-sufficient mice were significantly greater than that elicited in the CMG2-null mice. These results are consistent with those generated using the receptor-specific mutant forms of PA and suggest that CMG2 plays the major role in mediating the total anti-PA antibody titers as well as the toxin-neutralizing antibody response to PA.

Recently, de Oliveira et al. [13] demonstrated that immunization of mice with wild-type PA complexed with the soluble domain of CMG2 resulted in markedly reduced immunogenicity of PA. Thus, blocking the ability of PA to bind to cellular CMG2 receptors decreases its immunogenicity. These results are consistent with and further support the findings and conclusions presented in our study.

Both CMG2 and TEM8 are ubiquitously expressed [9], however the relative abundance of CMG2 and TEM8 receptors can differ. For example, CMG2 is expressed at a higher level than TEM8 in the liver; conversely, TEM8 is expressed at a higher level than CMG2 in the lung [9]. Because the cell types mediating toxin lethality versus receptor-enhanced immunogenicity of PA could differ, the relative roles of CMG2 and TEM8 in receptor-enhanced immunogenicity of PA could not be predicted a priori. The work presented in this report demonstrates the greater importance of CMG2 compared to TEM8 acting as PA receptors in enhancing the toxin-neutralizing antibody response to PA. Previously, CMG2 expression was determined to be higher than that of TEM8 in a number of murine immune cell types that could contribute to the receptor-enhanced neutralizing antibody response, such as dendritic cells (DCs), macrophages, B cells, and CD4⁺ T cells [14]. Both this observation and the previously noted higher apparent affinity of PA for CMG2 versus TEM8 [9] might help to explain our findings.

The immune mechanisms that underlie the receptor-enhanced neutralizing antibody response to PA remain unknown. Shaw and Starnbach [15] previously demonstrated that binding of PA to its receptor on bone marrow-derived DCs enhanced activation of antigen-specific CD4⁺ T cells in vitro. Those investigators postulated that binding of PA to its receptor enhanced delivery of antigens to endosomal compartments of APCs which might increase antigen access to the MHC-II processing pathway, ultimately increasing presentation of antigen to CD4⁺ T cells. While those researchers were able to demonstrate enhanced activation in vitro, they did not see a similar PA receptor-induced activation of CD4⁺ T cells in in vivo studies, possibly due to the relatively low doses of PA given to the animals.

A receptor-induced increase in the presentation of PA antigen to T cells could explain, at least in part, the receptor-mediated increase in the antibody response to PA that we previously noted [8]. Binding of PA to its cellular receptor on antigen presenting cells might also increase the accessibility of PA to B cells, thus enhancing the antibody response, including the protective toxin-neutralizing response. Our demonstration that CMG2 plays the major role in the receptor-enhanced neutralizing antibody response to PA is a first, but important, step in understanding the characteristics of cellular receptors that can lead to enhanced neutralizing antibody responses to vaccine antigens and therefore to a better understanding of how neutralizing antibody responses might be optimized. Moreover, our results suggest that any mutant form of PA used as a vaccine antigen should retain residues critical for binding to CMG2 receptor.

4. Conclusion

Using mutant forms of PA that bind preferentially to either CMG2 or TEM8, the two cellular receptors for PA, as well as by using CMG2-null mice, we have identified CMG2 as being the major receptor mediating a receptor-dependent enhancement of the neutralizing antibody response to PA. These findings represent an important step in elucidating immune mechanisms that might be harnessed to optimize neutralizing antibody responses to vaccine antigens.