ABSTRACT
Vaccination by anthrax protective antigen (PA)-based vaccines requires multiple immunization, underlying the need to develop more efficacious vaccines or alternative vaccination regimens. In spite of the vast use of PA-based vaccines, the definition of a marker for protective immunity is still lacking. Here we describe studies designed to help define such markers. To this end we have immunized guinea pigs by different methods and monitored the immune response and the corresponding extent of protection against a lethal challenge with anthrax spores. Active immunization was performed by a single injection using one of two methods: (i) vaccination with decreasing amounts of PA and (ii) vaccination with constant amounts of PA that had been thermally inactivated for increasing periods. In both studies a direct correlation between survival and neutralizing-antibody titer was found (r2 = 0.92 and 0.95, respectively). Most significantly, in the two protocols a similar neutralizing-antibody titer range provided 50% protection. Furthermore, in a complementary study involving passive transfer of PA hyperimmune sera to naive animals, a similar correlation between neutralizing-antibody titers and protection was found. In all three immunization studies, neutralization titers of at least 300 were sufficient to confer protection against a dose of 40 50% lethal doses (LD50) of virulent anthrax spores of the Vollum strain. Such consistency in the correlation of protective immunity with anti-PA antibody titers was not observed for antibody titers determined by an enzyme-linked immunosorbent assay. Taken together, these results clearly demonstrate that neutralizing antibodies to PA constitute a major component of the protective immunity against anthrax and suggest that this parameter could be used as a surrogate marker for protection.
Anthrax is a zoonotic disease caused by the spore-forming bacterium Bacillus anthracis. It most commonly occurs in wild and domestic mammals but can also occur in humans exposed to infected animals or tissue from infected animals. The increased concern over the potential use of anthrax spores in bioterrorism (1, 10) has boosted a surge in research related to protection against the disease.
A major factor in the virulence of B. anthracis is the secreted toxin complex comprising two toxins: the lethal toxin and the edema toxin (7, 25, 28). These toxins have distinct biochemically active components, lethal factor (LF) in the lethal toxin and edema factor (EF) in the edema toxin, yet they have a common component, protective antigen (PA) (for a recent review, see reference21). PA binds to the cell surface receptor, where it is proteolyticaly activated, creating a site for LF or EF binding. Once assembled, the toxin complex can be internalized and transported into the cell cytoplasm, where the toxigenic activity is expressed (11, 20, 26).
Consistent with the central role of PA in anthrax pathophysiology, vaccination with PA-based vaccines is commonly used to induce protective immunity (reviewed in references 39 and 40). The acellular vaccines licensed for human use are based on sterile culture supernatants of attenuated B. anthracis adsorbed on alum hydroxide (vaccine used in the United States) or precipitated with alum phosphate (vaccine used in the United Kingdom), and they contain various amounts of PA as well as lesser quantities of LF and EF (3, 32, 40). The partially defined composition of these vaccines as well as the requirement for six immunizations followed by annual boosters, underscores the need for the development of new, improved anthrax vaccines. This need has led in recent years to vast efforts in the design of shorter vaccination regimens (31), together with the development of improved cell-free PA vaccines (17, 19, 24, 36) and novel live attenuated vaccines (2, 5, 16) in which anti-PA protection plays a key role.
Development of new vaccines for anthrax is hampered by the difficulty in demonstrating their effectiveness in preventing disease in humans. In previous decades, anthrax infections were prevalent among mill workers, and these workers actually served as a target population for evaluating the efficacy of an earlier version of the acellular U.S. vaccine (4). At present, controlled efficacy studies in humans are not readily available. Clinical studies for anthrax vaccine evaluation nowadays rely mainly on determination of seroconversion and antibody titers to specific antigens, yet data on the correlation between such titers in human vaccinees and protection against exposure are unavailable.
Animal models, including guinea pigs (22, 37), rhesus monkeys (9, 19), and rabbits (41) developed for studying anthrax pathogenesis and vaccine efficacy, can be used to examine the correlation between immune response and protection. A systematic study correlating survival after challenge and antibody titers in animals could provide reliable surrogate markers, which in turn could provide a basis for evaluating the protective potential of antibody titers in humans.
In the present study we have used two assays for anti-PA antibody titers and evaluated their effectiveness in providing correlates for protection of guinea pigs against challenge with virulent anthrax spores. To this end we have used three experimental systems: (i) immunization of guinea pigs with various PA vaccine dilutions, (ii) immunization of guinea pigs with PA vaccine that was partially inactivated by incubation at 40°C for different periods, and (iii) passive immunization of guinea pigs with various amounts of hyperimmune sera. Our data demonstrate that anti-PA neutralizing-antibody titers per se appear to be a reliable surrogate marker for protective immunity.
MATERIALS AND METHODS
Production and purification of PA and vaccine formulation. B. anthracis strain V770-NP1-R (ATCC 14185) were anaerobically grown as described previously (5). After 24 h of growth, the bacteria were removed by microfiltration (pore size, 0.2 μm), while the PA-containing supernatant was concentrated by ultrafiltration (30,000 molecular weight cutoff) and dialyzed against 20 mM phosphate buffer (pH 8.0). PA was purified by Q-Sepharose chromatography essentially as described previously (33). This chromatography also yielded purified LF used in the neutralization assays. PA and LF were each analyzed by sodium dodecyl sulfate-polyacylamide gel electrophoresis (Coomassie blue staining), and each exhibited a single band with a molecular weight of ∼80,000. Cross-contamination between LF and PA in each of these preparations was lower than 0.1% as estimated by specific enzyme-linked immunosorbent assays (ELISAs).
The PA vaccine was prepared by adsorption of the purified PA at a final concentration of 50 μg/ml to alum hydroxide gel (Alhydrogel; Superfos Biosector, Denmark) at a concentration of 1.7 mg of Al/ml of vaccine. More than 98% of the added PA was adsorbed instantly as determined by measuring the residual protein concentration in the supernatant. After adsorption, NaCl was added to a final concentration of 0.9% (wt/vol).
Immunization and challenge of guinea pigs.Vaccine preparations were evaluated in female Hartley guinea pigs (weighing 220 to 250 g each) obtained from Charles River Laboratories. The animals were cared for according to the 1997 NIH guidelines for the care and use of laboratory animals; all experimental protocols were approved by the IIBR Animal Use Committee.
Groups of 10 to 30 guinea pigs were immunized by a single subcutaneous (s.c.) injection of 0.5 ml of either the PA vaccine or a dilution of the PA vaccine. At the indicated time (14 or 28 days after vaccination), the guinea pigs were challenged by intradermal injection of 0.1 ml of B. anthracis Vollum strain spore suspension containing 2,000 spores (40 50% lethal doses, estimated by the method of Reed and Muench [34]) and survival was monitored for 10 days.
Passive immunization.Hyperimmune serum was collected from guinea pigs immunized by triple injections (0, 2, and 4 weeks) of the PA vaccine. Various doses of the hyperimmune serum pool were administered intramuscularly to groups of naive guinea pigs. At 1 day after serum transfer, some of the guinea pigs from each group were used for determination of actual antibody titers in circulation while the rest of the animals were challenged as described above.
Serological tests. (i) ELISA for Anti-PA antibody.Antibody titers were determined by direct ELISA in 96-well microtiter plates (Nunc), using PA as the capture antigen and alkaline phosphatase goat anti-guinea pig immunoglobulin G (IgG) conjugate (Sigma, Israel) as the detecting reagent. The plates were coated with 5 μg of purified PA per ml (50 μl/well) in 50 mM NaHCO3 buffer (pH 9.6) and subsequently blocked with TSTA buffer (50mM Tris [pH 7.6], 142 mM sodium chloride, 0.05% sodium azide, 0.05% Tween 20, 2% bovine serum albumin). Tested sera were subjected to twofold serial dilutions, and the plates were incubated for 2 h at 37°C. The plates were then washed and developed with the detecting antibody conjugate, using p-nitrophenyl phosphate (Sigma) as the substrate. The absorbance at 405 nm was determined with a SPECTRA microplate reader. The end point was defined as the highest dilution exhibiting absorbance higher than 2 standard deviations above the negative control (normal guinea pig). Antibody titers were expressed as the reciprocal end-point dilution.
(ii) Neutralization test.Neutralizing-antibody titers were determined by virtue of their ability to prevent the PA- and LF-induced mortality of J774A.1 cells (American Type Culture Collection, Manassas, Va.) (35). Aliquots of 0.2-ml cell suspension (6 × 105 to 8 × 105cells/ml) were plated into 96-well cell culture plates (Nunc). Tested sera were twofold serially diluted in TSTA buffer. PA and LF at final concentrations of 5 and 2 μg/ml, respectively, were then added to the antiserum dilutions. After incubation for 1 h, 10 μl of each of the antiserum-toxin complex mixtures was added to 100 μl of J774A.1 cell suspension. The plates were incubated for 5 h at 37°C under 5% CO2, and cell viability was monitored by the MTT assay (27) (absorbance was measured at 540 nm). The end point was defined as the highest serum dilution exhibiting 0.025 absorbance unit above that of the corresponding identical dilution of the control normal serum. Neutralizing-antibody titers were expressed as the reciprocal end-point dilution.
Both the ELISA and the neutralization assay were performed in duplicate. Reproducibility was verified by the use of a negative control (normal serum) and a positive standard for each plate. In the ELISA, the values of the positive standard in the various plates were within 2 geometric standard deviations from a predetermined average. For the neutralization assay, the values of the positive standard were within a 1.3 geometric standard deviations from a predetermined average. The limit of detection in both assays was a titer of 50. For more information about the validity of the two immunoassays, see reference 13).
Tests were performed on pooled sera, except when specified. Equal amounts of serum collected from all animals within an experimental group were mixed to generate the pool. In some experiments the variability of antibody titers between individual animal sera comprising a serum pool was determined. When this was done, the geometric standard deviations observed in the ELISA and in the neutralization assay were not greater than 1.7 and 1.8, respectively.
RESULTS
The effect of PA vaccination on antibody titers and on the extent of protection against challenge.A multiple-immunization schedule is the most commonly used experimental system for evaluation of anthrax vaccines (18, 19, 22, 37, 38). Such schedules, which induce a high humoral response and confer full protection against anthrax spore challenge, are not suitable for devising correlates for protective immunity. In an attempt to overcome this problem, we have investigated the immune response following a single injection of PA vaccine (containing 25 μg of purified PA). The results presented in Fig. 1 show that 4 weeks after injection, both anti-PA IgG and neutralizing-antibody titers leveled off at titers of about 90,000 and 4,000, respectively. Guinea pigs challenged with 40 LD50 of Vollum spores were fully protected 2 weeks after vaccination, well before the time when maximal antibody titers were reached.
Kinetics of anti-PA (ELISA) and neutralizing-antibody titer development in guinea pigs following immunization. Guinea pigs were immunized with a single injection of 0.5 ml of PA vaccine containing 25 μg of purified PA. Periodically, anti-PA (ELISA) (▵) and neutralizing-antibody (□) titers were determined. Each point represents the average of duplicate determinations performed on a pool of sera derived from 10 guinea pigs.
To reduce the levels of antibodies against PA and achieve partial protection, guinea pigs were immunized with decreasing serial dilutions (1:2 to 1:32) of PA vaccine (Table 1). As expected, the reduction in PA dose led to a decrease in the anti-PA antibody titers (as measured both by ELISA and by the neutralization assay), with a concomitant decrease in the extent of protection against the lethal challenge. To examine the correlation between the survival rate and antibody titer, we have plotted these two parameters against each other (Fig. 2). Indeed, a clear correlation between neutralizing-antibody titers and the extent of protection was observed. Full protection could be reached when the neutralization titers were at least 300. The linear-regression plot (r2 = 0.92) depicted in the inset to Fig. 2indicates that 50% protection was achieved when the neutralizing-antibody titer was 80. When a similar analysis was conducted for anti-PA IgG titers (determined by ELISA), full protection was correlated with antibody titers higher than 2,500. However, in this case the rather low r2 value (0.56) obtained for the linear-regression line prohibited the determination of a valid value for 50% protection.
Active immunization of guinea pigs with PA vaccine: effect of vaccine dilution
Immune response induced by vaccination with serially diluted PA vaccine: correlation between neutralizing-antibody titer and protection. PA vaccine (PA concentration, 50 μg/ml) was serially diluted 1:2 to 1:32 in saline and injected (s.c. 0.5 ml) into guinea pigs. Two or four weeks following the immunization, protection and antibody titers were determined. Each point in the group represents the average of duplicate determinations performed on a pool of sera derived from 10 guinea pigs. A titer of 25 was assigned arbitrarily to serum pools with a titer below the detection limit of 50. (Inset) Linear regression was performed using data from experimental cohorts with neutralizing-antibody titers of 150 or less. The calculatedr2 value is 0.92.
It should be noted that the information presented above (Fig. 1 and 2) is based on antibody titers determined in pooled sera collected from half of the animals in each group (the rest of the animals were used for challenge). To find whether the correlates described above were also valid at the level of an individual animal, we have conducted an experiment in which each of the vaccinated animals was first used for serum collection and then (4 days later) challenged with Vollum spores. The results presented in Table 2 indicate that indeed the antibody titer in the animal prior to challenge could be of predictive value in evaluating survival. Further evaluation of the antibody titers in predicting protection in an individual vaccinee would obviously require a separate study.
Correlation between survival and titera for individual guinea pigs
The effect of thermal inactivation of PA vaccine on antibody titers and extent of protection.To further explore the interrelationship between levels of anti-PA antibodies and protective immunity, we have vaccinated animals with vaccine preparations subjected to various extents of heat inactivation. To this end, PA vaccine was incubated at 40°C for different periods and guinea pigs were immunized with a single injection of one of the various inactivated preparations. Fourteen days later, antibody titers and protection against challenge were determined. Not surprisingly, prolonged heat inactivation led to a decrease in the ability of the vaccine to induce anti-PA antibodies as determined by both ELISA and the neutralization assay. However, the decline in response to heat inactivation was more pronounced when antibodies were determined by the neutralization assay (Fig.3A). The ratio between the neutralizing-antibody titer and anti-PA antibody titer changed from 1:4 in guinea pigs immunized with noninactivated vaccine to about 1:20 in guinea pigs immunized with vaccine incubated for 6 to 33 days. Vaccinated animals were challenged with a lethal dose of anthrax spores, as described above, and the levels of protection were plotted against the reciprocal neutralizing-antibody titers (Fig. 3B). A clear correlation with survival to challenge was observed when neutralizing-antibody titers were between 25 and 125. When a linear-regression plot was generated (r2 = 0.95) , 50% protection could be correlated with a neutralizing-antibody titer of 65. Full protection was correlated with antibody titers of 300 and above. These values are in good agreement with the ones obtained when various doses of vaccine were used to induce partial protection.
Immune response induced by vaccination with heat-inactivated PA vaccine. (A) Effect of the duration of heat inactivation of the PA vaccine on its ability to generate anti-PA (ELISA) and neutralizing-antibody titers. Guinea pigs were immunized with a single injection of PA vaccine (containing 25 μg of purified PA) that was incubated at 40°C for various periods. Two weeks after immunization, the anti-PA (ELISA) (▵) and neutralizing-antibody (□) titers were determined. (B) Correlation between neutralizing-antibody titer and protection. Guinea pigs were immunized as described above. Two weeks following the immunization, the neutralizing-antibody titer and protection were determined. Each point in the group represents the average of duplicate determinations performed on a pool of sera derived from 10 guinea pigs. A titer of 25 was assigned arbitrarily to serum pools with a titer below the detection limit of 50. (Inset) Linear regression was performed using data from experimental cohorts with neutralizing-antibody titers of 150 or less. The calculatedr2 value is 0.95.
Attempts to determine in this experiment a correlation between survival and anti-PA IgG antibody titers, determined by ELISA, were again less successful. The r2 of the linear-regression curve was low, and calculation of titers for 50% protection was impractical. Full protection could be correlated with titers of at least 4,200, a value higher than the one observed in the experiment using the diluted vaccine (see above). This observation appears to reflect the fact that the ELISA for anti-PA IgG antibodies is affected to a lesser degree by the conformation of the PA immunogen and therefore implies that ELISA titers are probably less reliable as a surrogate marker for protection.
Evaluation of the protective potential of anti-PA antibodies by passive transfer.To further evaluate the protective value of the antibodies generated by immunization with the PA-based vaccine, homologous passive-transfer experiments were performed. Various doses of hyperimmune serum (volumes of 0.1 to 2.0 ml with a titer ranging from 2,400 to 24,000) obtained from guinea pigs (2 weeks after a schedule of three immunization with the PA vaccine) were introduced i.m. into guinea pigs. The anti-PA antibody titers in the recipient animals (titer of 50 to 3,200) remained stable for at least 1 week (result not shown). One day after serum transfer, the animals were challenged. As in the two previous experiments, correlation between circulatory antibody titers and protection was evaluated (Fig.4). Guinea pigs with neutralizing-antibody titers above 220 were fully protected. As a result of the complexity of the experimental system, ther2 value obtained in this case was rather low (r2 = 0.64) . However, calculations indicate that 50% protection was achieved at a titer of about 80, which is similar to that achieved by active immunization.
Passive immunization with PA hyperimmune serum: correlation between protection and immunogenicity. Pooled hyperimmune serum was obtained from guinea pigs immunized with PA vaccine as described in Materials and Methods. Various amounts of the serum were injected intramuscularly into naive guinea pigs. One day after the serum transfer, actual serum neutralizing-antibody titers and protection were determined. Each point in the graph represents results obtained from a group of 12 to 18 animals, half of which were used for the actual titer determination and the other half of which were used for the challenge. Neutralizing-antibody titers below 50 were calculated using linear extrapolation from the plot of the known injected amount of antibodies against the actual antibody titer in circulation. The inset shows the results of linear regression analysis performed using data from experimental cohorts with neutralizing antibody titers of less than 250. The calculatedr2 value is 0.64.
DISCUSSION
Cumulative information, gained over decades of research into the immunogenicity of anthrax vaccines, indicates that the presence of PA in cell-free vaccines or its production by live vaccine strains is required to confer effective protective immunity (2, 3, 5, 8, 12, 19, 29, 32, 36). Moreover, studies in which various vaccines were evaluated for their protective potential have indicated that elevation in the level of PA-specific antibodies was accompanied by increased survival. This was found in cell-free vaccines when various vaccine formulations were compared (8, 17, 37), in attenuated live vaccines where production levels of PA were altered (2, 5, 6), and also in passive immunization using anti-PA antibodies (23). Nevertheless, these studies did not provide quantitative correlates between anti-PA antibody level and protection, nor could they provide an estimate of the relative contribution of the PA-induced humoral response to the overall protective immunity against anthrax.
The present study was designed to allow for variation in levels of protection and to correlate these levels to anti-PA antibody titers. To eliminate potential contribution of other anthrax-derived components, we have used highly purified PA adsorbed to alum hydroxide as a vaccine. Partial protection was induced by two active-immunization methods as well as by passive immunization. In one of the active-immunization methods, decreasing amounts of the vaccine were administered to different guinea pig groups. In the other active-immunization method, the same amounts of the PA vaccine were injected into animals but, prior to injection, the vaccine was subjected to heat inactivation for various periods. Both vaccination methods promoted variable degrees of protection against Vollum challenge as well as variable titers of antibodies against PA. When we determined the anti-PA IgG antibody titers by ELISA, we were able to define a minimal titer which correlates with full protection; however, the actual values were different in the two methods used for active immunization. The minimal protection titer in the method using diluted vaccine was 2,500, while the corresponding titer in the method using heat-inactivated vaccine was 4,200. Titers for 50% protection could not be determined due to the low r2 obtained. These findings suggest that titers determined by ELISA have a limited value in predicting protective immunity.
On the other hand, when antibodies were determined by their ability to neutralize the cytotoxic effect of the lethal-toxin complex, a striking similarity was observed when the two active-vaccination experiments were compared. In effect, practically the same titer (titers of 65 to 80) was found to correlate with 50% protection in the two independent experimental immunization systems (Table3). Full protection in both methods of immunization correlated with titers of at least 300. This observation suggests that when the population of the partially heat-inactivated PA molecules is used for vaccination, only molecules that maintain intact functional domains play a role in promoting the production of antibodies that are of protective value. Moreover, these findings support the notion that antibodies involved in neutralization in vitro are also the ones involved in protection against the full course of infection. This is in agreement with previous observations in which monoclonal antibodies capable of neutralizing the cytotoxic effect were shown to delay death in guinea pigs (23).
Correlation between neutralizing-antibody titer and protection in various immunization protocols
Further support for the utilization of neutralizing-antibody titer as a valid surrogate marker for protection is provided by the passive-immunization experiment, in which anti-PA antibodies were transferred to naive guinea pigs. Here again, neutralizing-antibody titers of 220 and above were sufficient to confer full protection against lethal challenge and 50% protection was achieved by titers very similar to the ones calculated in the active-immunization experiments (Table 3).
Taken together, these results indicate that neutralizing-antibody titers can be used as a reliable correlate for protective immunity and that titers above 300 predict survival in guinea pigs. Moreover, it appears that PA-neutralizing antibodies at relatively low levels are sufficient for rescuing animals from experimental anthrax infection, underscoring the unique, but yet unresolved, role of PA-mediated toxicity in the pathogenesis of anthrax infection (for a review, see reference 15).
Although the humoral response to PA is a sufficient marker for protection, one cannot preclude the contribution of a cellular response in protective immunity against anthrax. Moreover, the central role of PA in anthrax protective immunity demonstrated here, as well as in previous studies, should not exclude the potential contribution of other antigens such as LF or spore constituents and somatic antigens, as suggested by studies with live attenuated vaccines (5, 16).
A correlation between protection and neutralization titers was recently observed (30) in another animal model (rabbit), suggesting that this phenomenon is not species specific. Application of such markers to human studies therefore appears to be feasible, even though similar experiments in nonhuman primates would be required to gain more confidence in the system.
In conclusion, the clear correlation between neutralizing antibodies and protective immunity demonstrated in this study indicates that neutralizing-antibody titers can serve as a reliable surrogate marker in evaluating novel vaccines in preclinical studies. Moreover, the development of such antibodies in human vaccinees would attest to the potency of an anthrax vaccine, and the actual titers in human serum could be instrumental in comparing the efficiencies of various novel vaccine formulations.
ACKNOWLEDGMENTS
We thank Lea Silberstein, Shlomo Shmaya, and Edith Lupu for their excellent technical assistance and Sara Cohen for critical review of the manuscript.
Notes
Editor: D. L. Burns
FOOTNOTES
- Received 30 November 2000.
- Returned for modification 3 January 2001.
- Accepted 30 January 2001.
- Copyright © 2001 American Society for Microbiology
