INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Anthrax is a well-known disease and was one of the first to be described in association with its causative organism, Bacillus anthracis (18). Although the disease is well characterized, it is only in recent years that we have begun to understand the molecular basis of anthrax. The principal virulence factor of B. anthracis is a multicomponent toxin secreted by the organism that consists of three separate gene products designated protective antigen (PA), lethal factor (LF), and edema factor (EF). The genes encoding these toxin components (pag, lef, and cya, respectively) are located on a 184-kb plasmid designated pXO1, carried by all strains of B. anthracis (26). PA (735 amino acids [aa]; M_r, 82,684) is a single-chain protein which binds to an as yet unidentified receptor on the cell surface and subsequently undergoes furin-mediated cleavage to yield a 63-kDa receptor-bound product (8, 16, 21, 31). The 63-kDa PA fragment forms a heptameric complex on the cell surface which is capable of interacting with either the 90-kDa LF protein or the 89-kDa EF protein, which is subsequently internalized (27, 31). LF (776 aa; M_r, 90,237) is a zinc metalloprotease that cleaves several isoforms of mitogen-activated protein kinase kinase (Mek1, Mek2, and MKK3), thereby disrupting signal transduction events within the cell and eventually leading to cell death (6, 30). The EF protein (767 aa; M_r, 88,808) is a calmodulin-dependent adenylate cyclase that causes deregulation of cellular physiology, leading to clinical manifestations that include edema (19). The LF protein combines with PA to form what is referred to as lethal toxin (Letx), which is considered to be the primary factor responsible for the lethal outcome of anthrax infection.

One of the earliest successful vaccines was an attenuated strain of B. anthracis used by Louis Pasteur to vaccinate sheep against anthrax (29). The current Food and Drug Administration-approved anthrax vaccine in the United States is produced from the culture supernatant fraction of the V770-NP1-R strain of B. anthracis and consists principally of PA adsorbed onto aluminum hydroxide. Protection against anthrax infection is associated with a humoral immune response directed against PA (14, 15). Some evidence suggests that EF and LF may also contribute to specific immunity (15, 24, 32), although these components have not been formulated into a subunit vaccine. At this time, there is significant interest in the development of a more highly defined anthrax vaccine and numerous efforts directed toward that goal are in progress. In this regard, in recent years there has been substantial interest in the development of DNA-based vaccines for genetic immunization due to the potential advantages associated with this approach (5, 25). With respect to DNA-based immunization against anthrax, it was demonstrated that one can obtain a protective response to an Letx challenge by immunization with a plasmid encoding the 63-kDa protease-cleaved fragment (PA₆₃) of PA (9). In the present study, our goals were to extend those observations and to explore whether DNA-based immunization against the LF gene product would contribute to or provide protection against an Letx challenge. In addition, we sought to explore whether combined immunization with genes encoding PA and LF would provide additional protection against the effects of Letx. In order to establish a baseline response for future epitope mapping considerations, we chose to utilize the minimum PA and LF structures which could form a functional binding complex while eliminating the metalloprotease function of LF. Therefore, these experiments were carried out using the gene fragment encoding PA₆₃, which is capable of binding to the PA receptor and to LF, and the gene fragment encoding LF4 (aa 1 to 254), which contains the N-terminal one-third of the LF antigen but lacks the domain associated with the LF metalloprotease function yet retains the ability to bind to PA₆₃ (2,12).

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Construction of PA and LF expression plasmids. The eucaryotic expression plasmid pCI (Promega, Inc., Madison, Wis.) was used in this study for the expression of truncated versions of the PA and LF proteins. The gene fragment encoding aa 175 to 764 of the PA protein was PCR amplified using the plus-strand primer 5'-ACA AGT CTC GAG ACC ATG GTT CCA GAC CGT GAC-3' and the minus-strand primer 3'-CTC TAT CCT ATT CCA TTA AGA TCT ACT AAA-5', with pYS2 as a template (33, 35). Included in the primer sequences are XhoI and XbaI restriction sites (underlined), respectively. The PA gene fragment expressed in this study corresponds to the biologically active, protease-cleaved PA₆₃ fragment of the full-length 83-kDa protein (8). The PCR product was digested with XhoI and XbaI and ligated into the pCI vector, which had been cut with the same two restriction enzymes. The plasmid construct pCLF4 encodes aa 10 to 254 of LF, which constitutes the PA binding domain. This plasmid was constructed from a PCR-amplified fragment using the primers 5'-GT CAT GGT CTA GAA ACC ATG CAC GTA AAA GAG-3' and 3'-TTG CTT GTT CTT TAT ATT TAG ATA TCA GAT CTG CAT-5', which incorporate XbaI cleavage sites (underlined). The XbaI-digested PCR and pCI plasmid fragments were ligated to form the pCLF4 plasmid used in this study. Neither of the resulting plasmid constructs, pCPA and pCLF4, contains a signal sequence for secretion of the expressed gene product. All plasmids were purified from Escherichia coli DH5 using Endo-free plasmid preparation kits (Qiagen) and were dissolved before use in phosphate-buffered saline (0.15 M NaCl, 0.01 M Na phosphate, pH 7.3).

Protein preparations. PA, LF, and LF_E687C (LF7) used in this study were expressed and purified as previously described (20, 28). LF_E687C is the full-length, enzymatically inactive LF protein containing the indicated amino acid substitution within the zinc-binding active site (17).

DNA vaccination. One-micrometer-diameter gold particles were coated with purified plasmid DNA according to the instructions of the manufacturer (Bio-Rad Laboratories, Richmond, Calif.). Separate groups of female BALB/c mice at 4 to 5 weeks of age (Jackson Laboratories, Bar Harbor, Maine) were immunized intradermally in the abdomen via biolistic particle injection (Helios gene gun; Bio-Rad Laboratories) on days 0, 14, and 28 with approximately 1 µg of plasmid DNA-coated gold particles for each injection. Immunization groups included mice injected with the same microparticles coated with pCPA, pCLF4, a 1:1 mixture of the pCPA and pCLF4 plasmids, or, as a vector control, the pCI plasmid. For the prime-boost immunization experiments, groups of BALB/c mice were first immunized twice with plasmid DNA as described above and then with a third and final boost of purified antigen (12.5 µg) (PA and/or LF7) emulsified in Freund's incomplete adjuvant (1:1 ratio of adjuvant to protein, vol/vol). The protein immunizations were administered intramuscularly. Blood samples were obtained 2 weeks following each vaccination, and the sera were pooled and stored at 20°C until analyzed.

Mouse macrophage protection assay. The cytotoxicity of the purified Letx was established using a previously described macrophage cytotoxicity assay (28, 34). For the protection assay, J774A.1 mouse macrophage cells were placed in flat-bottom 96-well microtiter plates at a concentration of 6 × 10⁴ cells/well in Dulbecco's modified Eagle's medium (Sigma, St. Louis, Mo.) with 7% fetal bovine serum, 4.5 g of glucose per liter, and 2 mM L-glutamine and incubated for 24 h at 37°C. Serum from a pCLF4-immunized New Zealand White rabbit was serially diluted and incubated with LF protein for 1 h to allow neutralization to occur. Following this incubation, the LF-anti-LF mixture was added to PA protein to achieve a final concentration of 3 µg of Letx per ml. This preparation was incubated at room temperature for 1 h prior to being added to the cells, which were then incubated for an additional 7 h at 37°C. At the end of the incubation, 100 µl of 0.5-mg/ml MTT (3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyltetrazolium bromide; Sigma) per well was added, followed by a 1-h incubation. Cells that survive exposure to Letx are able to oxidize MTT to an insoluble purple pigment, thus providing a proportional measure of the viability of the cells. At the end of the incubation period, the culture supernatant was aspirated, 50 µl of a solution containing 0.5% (wt/vol) sodium dodecyl sulfate and 25 mM HCl in 90% (vol/vol) 2-propanol was added, and the suspension was vortexed. The A₄₅₀ was determined using a microplate reader (Bio-Tek Instruments, Inc.).

In vivo protection experiment. Plasmid-immunized BALB/c mice that had received a total of three injections were challenged with purified Letx 2 weeks following the third and final injection. The challenge was conducted by tail vein injection of a previously mixed combination of purified PA and LF proteins (60 µg of PA and 25 to 30 µg of LF per mouse), the equivalent of approximately 5 50% lethal doses (LD₅₀) of Letx.

ELISA for anti-PA and anti-LF antibodies. Titers of antibodies to PA and LF were determined by enzyme-linked immunosorbent assay (ELISA). Briefly, Immulon 4 96-well plates (Dynatech Laboratories, Inc., Chantilly, Va.) were coated at 4°C overnight with 100 ng of purified PA or LF7 protein dissolved in 0.1 M carbonate buffer, pH 9.6. Plates were washed with TBS (Tris-buffered saline; 50 mM Tris-HCl, 0.15 M NaCl, pH 7.3) and blocked with 1% (wt/vol) bovine serum albumin in TBS. Serum samples were serially diluted in TBS containing 0.05% Tween 20 and added to the plates. All incubations were carried out at 37°C for 1 h. Anti-mouse immunoglobulin G (IgG) conjugated to horseradish peroxidase (Amersham Life Science, Arlington Heights, Ill.) was added as the secondary antibody. The presence of bound antibody was detected following a 30-min incubation in the presence of ABTS [2,2'-azinobis(3-ethylbenzthiazolinesulfonic acid)] substrate (Zymed, South San Francisco, Calif.), and absorbance was read at 405 nm using a Bio-Rad model 550 plate reader. Antibody titers were defined as the highest dilution in serum that resulted in an absorbance value two times greater than that of a serum sample from a nonimmune control, with a minimum value of 0.05. Antibody isotypes were determined in a similar manner, except that anti-mouse IgG1 or anti-mouse IgG2a conjugated to alkaline phosphatase was used as the secondary antibody (Zymed Laboratories). Antibody quantitation was done by ELISA analysis using a standard curve with purified IgG1 and IgG2 antibody reagents.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Immunization with plasmids encoding portions of PA or LF. This study utilized the pCI mammalian expression vector (Promega), in which the human cytomegalovirus immediate-early enhancer-promoter region drives strong, constitutive expression of the incorporated gene (Fig. 1). In this study we chose to express only partial sequences of the PA and LF genes, as shown in Fig. 1. The pCPA plasmid expresses a truncated version of the PA gene product, aa 175 to 735, which is PA₆₃ lacking the furin cleavage site (aa 164 to 167), yet is fully functional in vivo (8). The pCLF4 plasmid expresses a truncated form of LF, aa 10 to 254, which lacks the catalytic domain of LF yet retains PA₆₃ binding activity and is therefore capable of interacting with the truncated form of PA expressed from pCPA (1).

View larger version (20K): [in this window] [in a new window]   FIG. 1.   Plasmid vectors used in this study. Plasmid pCI (Promega, Inc.), a eucaryotic expression vector, was used to express aa 10 to 254 of B. anthracis LF protein and aa 175 to 764 of B. anthracis PA protein. CMV, cytomegalovirus; SV40, simian virus 40.

View larger version (13K): [in this window] [in a new window]   FIG. 2.   Titers of antibodies to purified LF protein (A) or PA (B) in the sera of BALB/c mice immunized with pCPA, pCLF4, or a combination of pCPA and pCLF4.

View larger version (11K): [in this window] [in a new window]   FIG. 3.   Serum antibody titers in BALB/c mice immunized against PA with pCPA (PA), pCPA and pCLF4 (PA/LF), and pCPA and pCLF4 and boosted with PA and mutant LF proteins on day 28 (PA/LF pb) (A) or against LF with pCLF4 (LF), pCLF4 and pCPA (PA/LF), and pCPA and pCLF4 and boosted with PA and mutant LF protein on day 28 (PA/LF pb) (B).

Plasmid immunization results in a protective response. Small groups of BALB/c mice were immunized three times with pCPA, pCLF4, a 1:1 combination of pCPA and pCLF4, or the plasmid vector (pCI). In an effort to determine whether DNA-based immunization alone can provide protection against exposure to Letx, these mice were challenged with 5 LD₅₀ of Letx administered intravenously. The results of this challenge study are presented in Table 1, where it can be seen that all animals immunized with a plasmid containing the gene for either PA or LF survived. Control mice succumbed to the Letx challenge within hours. These results demonstrate that DNA-based immunization alone can provide a protective response to exposure to the lethal anthrax toxin.

Comparison of prime-boost and DNA-only immunizations. In a separate set of experiments we investigated the ability to enhance titers of antibodies to PA and LF using the prime-boost method. This method consists of priming the immune system with a series of three plasmid-based immunizations followed by a final booster immunization with the protein antigen. In Fig. 3 it can be seen that coadministration of the pCPA and pCLF4 plasmids followed by a final protein booster immunization with the recombinant PA and LF7 antigens produced a substantially higher endpoint titer against either PA or LF at the same time point than the antibody titers resulting from DNA-based immunization alone. It was also observed that a consistently higher antibody titer formed against the LF antigen regardless of the immunization regimen used.

Neutralization of Letx using anti-LF4 serum. To determine whether anti-LF antibodies in the sera of pLF4-vaccinated animals could protect against the effects of Letx challenge, an in vitro neutralization assay was performed using murine J774A.1 cells, which are sensitive to Letx. As shown in Fig. 4, a 1:4 dilution of serum from a rabbit immunized two times with pCLF4 and then with a single protein boost with recombinant LF7 conferred 100% protection against the cytotoxic effects of Letx. The results of this assay confirm the macrophage cytotoxicity of the Letx preparations used in this study and demonstrate that the anti-LF antibody inhibits Letx cytotoxicity by blocking the incorporation of LF into the cell.

View larger version (7K): [in this window] [in a new window]   FIG. 4.   Neutralization of anthrax toxin by rabbit anti-LF4 antibody. Various dilutions of anti-LF4 serum were preincubated with recombinant LF () for 1 h. The mixture was added to J774A.1 cells in the presence of Letx for 7 h, and cell viability was measured. , absence of MTT; , negative Letx control.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

It was previously demonstrated that DNA-based expression vectors encoding PA of B. anthracis are immunogenic and can produce a protective response to Letx challenge in a mouse model system (9). In the present study, we have not only repeated these observations concerning PA but also demonstrated that it is possible to produce protective immunity against anthrax Letx by DNA-based immunization using a truncated form of the LF gene which produces an N-terminal fragment of LF lacking metalloproteinase activity. Furthermore, we observed a distinct synergistic effect associated with coimmunization using plasmids encoding fragments of the PA and LF genes, which results in significantly higher antibody responses to both PA and LF. In this study we demonstrated that a significant antibody response is generated using DNA-based immunization alone and that the levels of antibody produced are sufficient to protect animals against a Letx challenge that is five times the LD₅₀.

Of potential significance, it is worth noting that, in the previous DNA immunization study, PA was expressed in a form that may be secreted from the cell but that, in the present study, the expressed PA and LF proteins were not expected to be secreted. In vitro transfection experiments using human UM449 cells demonstrated the expression of these antigens but did not provide evidence of their release from the cell (data not shown). In spite of the apparent lack of secretion of PA and LF, a substantial antibody response was generated. This result is not surprising in view of the study previously reported by Haddad et al. (13). That study demonstrated that, while attachment of a secretion signal sequence resulted in differential intracellular targeting of an encoded antigen, the presence of a signal sequence had no effect on B-cell recognition of the antigen and the subsequent production of an antibody response. Thus, it is apparently not necessary to attach a secretion signal sequence to PA or LF in order to obtain a protective antibody response to Letx. We have recently reported similar findings in a DNA-based immunization study using Pseudomonas aeruginosa exotoxin A (4). At this time it is not clear how the truncated PA and LF proteins may exit the cell for interaction with B cells.

What also remains unclear at this point is how PA₆₃ and the LF4 antigen are processed within the cell in a manner that enables them to induce an enhanced antibody response when administered together. Nonetheless, we consistently observed a significantly greater titer of antibody to either PA or LF whenever the two genes were coadministered (Fig. 2 and 3). Although this observation was not supported in one instance by the data in Table 2 (comparison of PA-specific IgG levels between pCPA- and pCPA- or pCLF4-immunized groups), the data in Table 2 generally support the observation that coimmunization with the pCPA and pCLF4 plasmids produces an enhanced response. It is perhaps worth noting in this regard that the comparison between levels of PA-specific IgG (Table 2) and titers of antiserum to PA (Fig. 2 and 3) may not be directly comparable as determined in these separate assays. Nevertheless, an enhancement of the immune response following coimmunization with PA and LF has recently been reported by others (3). Of possible significance with regard to the enhanced antibody response observed following coimmunization with both the PA and LF4 genes is the reported mitogenic activity associated with LF (3, 10, 11, 32). Brossier et al. reports that the adjuvant effect requires binding activities between LF and PA but does not depend on the subsequent binding of the Letx complex to cells (3). In this regard, the ability of LF4 to bind to PA₆₃ would be consistent with the requirement for a PA-LF complex in the development of a synergistic immune response when these antigens are coadministered. This hypothesis would also predict that use of a mutated form of PA that is unable to bind to the cellular receptor would not diminish the synergistic immune response following coimmunization with the PA and LF antigens. Such experiments are in progress. A significant difference in the results of experiments reported in the present study is that when PA₆₃ and the LF4 antigen are first expressed within the cell, as under the conditions of these experiments, we also observe a synergistic immune response. In addition, the immune response to the LF antigen is generally greater than the response to PA, unlike the response reported by Brossier et al. when they used spore-based immunizations (3). Also, we find that intracellular expression of the LF4 antigen alone is sufficient to produce a significant immune response to the LF antigen without any requirement for previous binding to PA. Clearly, the synergism observed upon coimmunization with these genes does not depend on any metalloprotease activity associated with the LF antigen since a truncated form of the protein is expressed.

As expected, the majority of the immune response appears to be a TH2-type response, consistent with previously reported results of a study using plasmid-based immunization with the intact PA gene (9). This is clearly seen in the data presented in Table 2, where it can be seen that the predominant response is the production of IgG1 antibody in all cases. In the previous study plasmid immunizations were conducted via intramuscular injection, whereas in this study plasmids were delivered by means of a gene gun. This difference in results with various techniques is interesting in view of the fact that it has been reported that intramuscular vaccination with DNA generally produces a predominant TH1 response but that gene gun vaccination produces a TH2 response (7). However, it was noted by Gu et al. that both TH1 and TH2 cytokines were induced following DNA-based immunization with the intact PA gene (9).

It is noteworthy that we have been able to produce a protective response by immunizing with a truncated form of the LF antigen, and as far as we know, this is the first time protection following immunization against the LF antigen alone has been reported. Presumably, anti-LF antibodies that result following DNA-based immunization are capable of inhibiting the binding between PA and LF in vivo, thus preventing the formation of the Letx complex. This possibility is supported by the results depicted in Fig. 4, which demonstrate that anti-LF antibodies block the effect of Letx on cultured mouse macrophages. Also of interest is the greater immunogenicity of the LF4 protein. Once again, this is clearly observed in Fig. 2 and 3, as well as in the data presented in Table 2. In related experiments, we have observed that titers of anti-LF antibody remain at high levels for much longer periods of time than do titers of anti-PA antibody (data not shown). Collectively, these results support a possible role for an anti-LF antibody response in protection against anthrax infection. This possibility is a significant result in view of the fact that PA alone has been the primary target for vaccine studies since immunity against PA has been stated to be both necessary and sufficient for protection against anthrax infection (22). One implication of the findings presented in this report is that individuals who have been immunized with the current U.S. anthrax vaccine have undoubtedly produced an antibody response to the LF antigen, which is found in small quantities in the vaccine. The results presented in this paper suggest that the antibody response to LF following the current vaccine immunization series may comprise a significant part of the overall protective response to anthrax infection. This possibility is further supported by the fact that the LF antigen appears to be much more immunogenic and produces an immune response which lasts much longer than the response to the PA antigen.

An additional parameter investigated during this study was the use of a prime-boost strategy to enhance immunization against anthrax. This approach has gained considerable attention since several studies have reported that a combined DNA-protein immunization regimen results in the highest level of immune response overall. In the present study we are able to substantiate these other reports, since we found that the prime-boost approach significantly increases the overall level of antibody response to either PA or the LF antigen. Prime-boost antibody levels are substantially higher than those observed following DNA-only immunizations as observed in both Fig. 3 and Table 2. The results presented in this study indicate that it is feasible to use a DNA-based immunization strategy against anthrax and that any future vaccine against anthrax should consider incorporation of a mutated version of the LF antigen.