Specific, Sensitive, and Quantitative Enzyme-Linked Immunosorbent Assay for Human Immunoglobulin G Antibodies to Anthrax Toxin Protective Antigen

Discussion

The focus of this report is to describe the qualification and performance characteristics of an ELISA for anti-PA IgG antibodies and enhancement of its specificity by using a second-stage CI-ELISA. The application of these assays to the analysis of the antibody response following anthrax infection (Quinn CP et al., manuscript in preparation) and for serologic surveillance from clinical anthrax cases will be reported in detail elsewhere.^[13,14] Historically, if not identified and treated early, systemic anthrax in humans was invariably fatal. As a consequence, serologic assays have not featured prominently in the diagnosis of clinical anthrax and in some reports have been considered unreliable for early identification of the disease or for establishing a retrospective diagnosis.^[15] Serologic assays for anthrax have primarily been applied for the evaluation of immune responses to anthrax vaccines, in epidemiologic investigations of the disease in animals, and in confirmatory diagnosis of the various manifestations of anthrax in humans.^[16,17,18] A useful adjunct to serologic analysis of anthrax infection is the Anthraxin test,^[19] which elicits a localized delayed-type hypersensitivity reaction to intradermal injection of a complex uncharacterized extract from attenuated vegetative B. anthracis cells or edema fluid from B. anthracis-infected animals. Anthraxin has been reported to be very accurate for retrospective verification of anthrax in humans,^[20] to be applicable for up to 30 years after infection,^[15] and to have a qualitative positive correlation with anti-PA antibodies in the sera from human clinical anthrax cases as detected by ELISA.^[21] Anthraxin is not, however, approved by FDA as a diagnostic reagent in the United States.

For nearly 4 decades, anthrax serologic studies depended on the Ouchterlony agar gel diffusion test,^[16] which in turn replaced complement fixation tests and in vivo passive protection and neutralization tests.^[22] The development of an indirect (passive) microhemagglutination test^[23] was the next major progression in anthrax serologic testing. Based on the agglutination by serum antibodies of sheep erythrocytes sensitized with partially purified culture supernatants containing anthrax toxin PA (Factor II),^[24] this test provided greater sensitivity and speed than the agar gel diffusion technique. The microhemagglutination assay, however, was laborious to set up because antigen-coated erythrocytes had only a short shelf life, and variation in erythrocyte batches compromised reproducibility.^[25]

The microhemagglutination assay was replaced by the more sensitive and reproducible ELISA system, which has been applied in various formats including the immobilized antigen (direct) assay,^[25] an immobilized anti-PA antibody (antigen-capture) assay,^[26] and a competition ELISA,^[27] with, where reported, varying degrees of specificity and sensitivity.^[17,18,28] Turnbull et al.^[29] described using ELISA to confirm anthrax in humans and demonstrated that recipients of the licensed AVA could be distinguished from persons with natural infections on the basis of their lack of reactivity to the anthrax toxin lethal factor protein. Sirisanthana et al.^[17] in a serologic study of anthrax in northern Thailand and Harrison et al.^[18] in a serologic study of anthrax in Paraguay described use of both ELISA and Western blot to retrospectively evaluate seroconversion in cutaneous and oral-oropharyngeal anthrax. In both studies, separate ELISAs were applied for the detection of anti-toxin and anti-capsule antibody responses, and Western blot was used to enhance the specificity of serologic diagnosis.^[17,18] Turnbull et al.^[16] reported the detection of anti-PA antibodies in humans and animals in the Etosha National Park and also concluded that there is a residual antibody level in these populations in an area where the disease is endemic. However, the specificity and sensitivity of the ELISA used in that study were not reported, and the frequency of anti-PA antibodies in bacteriologically confirmed clinical cases was a maximum of 71%.^[16] The prevalence of true positive anti-PA antibody reactivity in the general human population therefore remains unknown, although our study of 515 non-anthrax-related control sera suggests that it is probably <7%.

The anti-PA ELISA developed and qualified at CDC before being applied in the anthrax emergency has a MDC of 0.06 µg/mL, an RDL of 0.09 µg/mL, and a reactivity threshold of 3.0 µg/mL anti-PA IgG. The reactivity threshold was adopted as the lower limit of quantification. The DSN of the assay is 97.6%, and the DSP is 94.2%. The CI-ELISA enhanced DSP to 100%. These results represent substantial improvements over the published sensitivities for anti-PA ELISAs of 72%^[17] and 91.7%.^[18] Although Harrison et al.^[18] reported a specificity of 100%, this was on a sample size of 18 controls, compared with the sample size reported here of 515 control sera (277 non-anthrax-related sera plus 238 NHANES controls). An additional important outcome of this study is the provision of a standard reference serum that can be used in a variety of serologic assays for the detection and quantification of anti-PA antibodies.

When evaluating the importance of a reactive serologic result, the prevalence of disease in the group of interest should first be considered. The assays reported in this study were primarily applied as a part of a panel of laboratory tests for the confirmation of clinical human anthrax in patients in whom the disease prevalence is expected to be high.^[30] The assays were also applied to serologic surveys of patients who may have been exposed to spores of B. anthracis, a group in which the disease prevalence may be expected to be low.^[13,14] For assays in which the specificity and sensitivity have been determined to be high (>90%), a reactive serum in a low-prevalence group has a much greater probability of being a false positive than it does in a high-prevalence group.^[31] Conversely, in the high-prevalence group nonreactive sera may not be indicative of the absence of disease. In practice, the 2.2% frequency of potential false positives reported here rarely presented a problem, and seroconversion was detectable by a ≥4-fold rise in anti-PA IgG concentration above the assay reactivity threshold or the acute-phase serum in all but three patients (data not shown), where paired samples with an appropriate time interval between them were available (0-7 days after symptoms for acute-phase sera and 14-28 days after symptoms for convalescent-phase sera).

The necessity for a rapid public health response meant that optimizing the number and timing of the patient serum collection was not always feasible or practicable. As a result, there was a high frequency of single (i.e., unpaired) sera from cases under investigation. To provide adequate analysis of these single sera and also for paired sera that were reactive but did not demonstrate changing levels of reactivity with time, a competitive rPA inhibition ELISA (CI-ELISA) was developed based on the qualified anti-PA ELISA. The objective of the CI-ELISA was to increase the DSP of the ELISA by reducing the incidence of false positives. The CI-ELISA effectively demonstrated the specificity of the ELISA format by using sera from AVA vaccinees and clinically confirmed anthrax cases.^[8]

Although very little published information supports the suggestion that antibiotic therapy can suppress the humoral immune response,^[32] anthrax infection studies in nonhuman primates have shown that early antibiotic treatment after a known challenge with B. anthracis spores abrogates an anti-PA antibody response.^[33] A plausible explanation for this is that early intervention in the infection process minimizes antigen presentation to the immune system. The implication, particularly for cutaneous anthrax in the context of a response to a bioterrorist attack, when antibiotic intervention is likely to be rapid and aggressive,^[2] is that serologic tests should not be used as the sole confirmatory tests for anthrax.

Emerging Infectious Diseases. 2002;8(10) © 2002 Centers for Disease Control and Prevention (CDC)

Cite this: Specific, Sensitive, and Quantitative Enzyme-Linked Immunosorbent Assay for Human Immunoglobulin G Antibodies to Anthrax Toxin Protective Antigen - Medscape - Oct 01, 2002.