Use of an enzyme-linked immunosorbent assay to measure antigenaemia during acute plague*

Bulletin of the World Health Organization, 62 (3): 463-466 (1984) Use of an enzyme-linked immunosorbent assay to measure antigenaemia during acute plague* JAMES E. WILLIAMS,' MARY K. GENTRY, CAROL A. BRADEN,3 FLORA LEISTER, & ROBERT H. YOLKEN5 An enzyme-linked immunosorbent assay (ELISA) was developed to measure concen- trations of the specific Fl antigen of the plague bacillus in biological fluids. The assay employed a monoclonal antibody to capture the antigen. Sensitivity of the assay was 0.4 ng ofFl antigen. ELISA-inhibition was used to confirm the specificity of the reactions. This assay detected Fl antigen in two of ten sera from patients with acute bubonic plague and indicated that antigenaemia in man during plague may reach levels of4-8 yg of Fl antigen per ml of serum. The probability for a correct serodiagnosis ofplague was improved when the patients' sera were tested for both antibody and antigen. Two patients with antigenaemia did not :'ave antibody, while two patients with antibody lacked antigenaemia. Rapid diagnosis of a fulminating disease with a strong epidemic potential, such as plague, improves the chances for recovery of the patient through early use of appropriate chemotherapy and for prevention of new cases by the timely application of effective control measures. Enzyme-immunoassays that detect antigens of disease agents have proved to be a useful alternative for the rapid diagnosis of many infections (1). This report describes an enzyme-linked immuno- sorbent assay (ELISA) to detect the specific Fraction 1 (Fl) envelope antigen of the plague bacillus and dis- cusses its application in the serodiagnosis of plague. MATERIALS AND METHODS Sera from bubonic plague patients Ten sera from Vietnamese patients with bubonic plague, confirmed by isolation of the plague bacillus, were examined. The sera were collected during acute disease, but chemoprophylaxis for plague may have * From the Departments of Hazardous Microorganisms and Biological Chemistry, Walter Reed Army Institute of Research, Washington, DC 20307, USA, and the Eudowood Division of Infec- tious Diseases, Department of Pediatrics, The Johns Hopkins University School of Medicine, Baltimore, MD 21205, USA. l Chief, Department of Hazardous Microorganisms. 2 Research Chemist, Department of Biological Chemistry. 3 Medical Laboratory Specialist, Department of Hazardous Microorganisms. 4 Laboratory Technician, Department of Pediatrics. 5 Assistant Professor of Pediatrics, The Johns Hopkins Uni- versity. been instituted in some patients prior to obtaining the serum. Penicillin (100 units/ml) and streptomycin (100 Ag/ml) were added to the sera to prevent growth of plague bacilli, and subsequent plating of the sera on blood agar confirmed that viable plague bacilli were not present. The sera, collected in 1974, were re- frigerated or frozen until tested for antibody (in 1975) or antigen (in 1982). Passive haemagglutination (PHA) tests for antibody The sera were examined for antibody to the Fl antigen using the PHA microtitration procedure recommended by WHO for the serodiagnosis of plague (2). ELISA procedure for antigen A noncompetitive, indirect procedure (1) was used to measure the Fl antigen in the sera. Microtitration platesa were sensitized with 50 Al/well of mouse ascitic fluid containing anti-Fl monoclonal antibody 3G8b diluted 1:1000 with phosphate-buffered saline, pH 7.2 (PBS). The plates were sealed with tape and stored at 4 °C for 3-28 days. When a plate was used, the tape and sensitizing solution were removed, and a Immulon 2 flat bottom plates, Dynatech Laboratories, Inc., Alexandria, VA 22314, USA. b Production and characterization of this and other monoclonal antibodies for plague diagnosis will be reported in another paper. However, Ouchterlony double diffusion tests have demonstrated that monoclonal antibody 3G8 precipitates the Fl antigen produced by plague strains isolated in the USA, Peru, South Africa, Namibia, Madagascar, Yemen, India, Pakistan, USSR, Burma, Vietnam and Java. 4419 -463 464 J. E. WILLIAMS ET AL. the plate was washed with PBS containing 0.5 ml/ litre of Tween 20 (PBST).c Twofold dilutions in PBST (50 1A/well) of a standard solution of Fl antigen (positive control), a negative control serum, and the patients' sera were prepared directly in the plate and incubated for 30 min at 37 °C on an orbital shaker (3).d Specimen dilutions were aspirated, and the plate was rinsed twice with PBST. Then 50 Al/well of a second antibody solution (anti-plague rabbit serum globulin; a 1:100 dilution of the globulin fraction from I ml of serum in PBST) was added and incu- bated for 15 min at 37 °C with shaking. After wash- ing,' 50 pI/well of horseradish peroxidase-conju- gated antibody to rabbit IgG (H & L)e at a 1:300 dilution in PBST was added and incubated for 15 min at 37 °C with shaking. After washingc and drying the plate, 150 1l/well of fresh indicator solution (9 parts of 0.8 g/litre of 5-aminosalicylic acid, pH 6.0, and 1 part of 0.5 ml/litre H202) was added and incubated for 30 min at 37 °C with shaking. The absorbance values at 492 nm were recorded with a photometer.! ELISA-inhibition on controls and sera under test ELISA-inhibition was performed concurrently with ELISA on each dilution of serum to confirmn the specificity of the reaction and on dilutions of the control antigen solution to establish the test's sensi- tivity. The inhibition test differed from ELISA only in that the PBST used to prepare the dilutions con- tained 3G8 anti-Fl mouse ascitic fluid at a 1:500 dilution. Competition for Fl antigen between the monoclonal antibody in solution and that bound to the plastic plate produced a lower absorbance value for the inhibition test compared to the ELISA. The ELISA and the inhibition tests on a given serum or antigen dilution were conducted side-by-side in the microtitration plate and read by the same photocell of the photometer! A similar inhibition procedure has been used to confirm the specificity of ELISA for plague antibody (4, 5). RESU LTS ELISA detected 8 ng of Fl antigen per ml (0.4 ng/ well) at the I1Vo probability level of statistical signifi- cance (P = 0.01) and 4 ng of Fl antigen per ml Three rinses were followed by a wash in which the PBST was left in the wells for 5 minutes. d Mini-Orbital shaker for microtitration plates, Bellco Glass, Inc., Vineland, NJ 08360, USA. e Catalog No. 141506, Kirkegaard & Perry Laboratories, Inc., Gaithersburg, MD 20879, USA. f Titertek Multiskan photometer, Flow Laboratories, Inc., McLean, VA 22102, USA. This instrument has eight photoelectric cells and calculates the percentage decimals of maximum ab- sorbance. 1.2 1.0 .8 .2 (57%) (33%) (21%) (10%) (6%) (4%) (4%) (3%) 64 32 16 8 4 2 0.5 0.25 0 MANOGRAMS OF FRACTION 1 ANTIGEN FR K Fig. 1. Sensitivity of ELISA for measurement of the Fl antigen of the plague bacillus. Parentheses show the percentage difference, from an absorbance value of 1.2, between ELISA (o) and ELISA-inhibition (*) reactions for given concentrations of antigen. ELISA values at zero nanograms of antigen (saline blanks) represent the eight photocells of the photometer, and the bar shows the range of variation at ± 3 standard deviations from the mean. (0.2 ng/well) at the 507o probability level of statistical significance (P = 0.05). The difference from maxi- mum absorbance between ELISA and the inhibition test was 13% at the 1% probability level (Fig. 1). However, a difference of 20%o between ELISA and ELISA-inhibition was chosen as an endpoint in the tests of sera to simplify endpoint determination. In effect, the chance for false positives was further reduced (P < 0.01). Fl antigenaemia was detected in the sera from two bubonic plague cases (Table 1). Antigenaemia was at the levels of 4-8 yg/ml. The sera were tested three times, and measurements of titre never varied more than ± 1 dilution. These patients with antigenaemia did not have a detectable serum titre of Fl antibody. Sera from eight patients, including two with anti- body, were negative for antigenaemia. ELISA values for negative sera diluted 1:4 to 1:512 were always within 100o of the absorbance value of unreacted indicator solution. DISCUSSION Confirmation of a clinical diagnosis of plague has relied on bacteriological isolation of the bacillus, demonstration of an increase in specific antibody in sera, or a positive fluorescent antibody test. Isolation of the plague bacillus takes several days to achieve, antibody usually appears at significant titre only about a week after infection, and the fluorescent anti- body test is somewhat limited because it detects only MEASUREMENT OF PLAGUE ANTIGENAEMIA Table 1. Titres of antibody and antigen in acute sera from patients with bubonic plague Antibody to Concentration Group No. Fl antigen Fl antigen of Fl antigen by PHA test' by ELISA" in serum Patients with 2 < 1:4 1:512 4 Fg/ml antigenaemia < 1:4 1:1024 8 AgIml Patients with antibody 2 1:64 < 1:4 Not detected 1:128 < 1:4 Patients without 6 < 1:4 < 1:4 antibody and antigenaemia Control " < 1:4 < 1:4 Values are highest serum dilutions with positive reactions. Pool of 10 sera from unimmunized persons residing in plague-free areas. bound antigen (e.g., encapsulated bacilli), with results that are difficult to quantify. Thus, the confir- mation of plague by these procedures frequently is achieved retrospective to the outcome for the patient, since most fatalities from plague occur within a week of infection. The ELISA described in this paper meets the requirement for a rapid diagnostic procedure that is definitive, sensitive, quantitative, and capable of measuring soluble antigen. It detects 0.4 ng of Fl, is rigidly controlled for specificity by ELISA-inhibition and is completed in 2j hours. Where available, probably radioimmunoassay could also be employed to satisfy these criteria (6). Antigenaemia was not present in all patients with acute plague. Only 2007o of their sera were positive for Fl antigen. Those patients with antigenaemia were probably experiencing a dissemination of the disease, via a bacteraemia that developed from a localized bubonic infection. However, bacteriological isolation was not attempted from the blood samples in ques- tion, so we cannot confirm this assumption. Patients with antigenaemia lacked antibody and vice versa, and the data indicate that an efficient strategy for serodiagnosis of plague is to test several sera, taken at various times, for both antigen and antibody. The sera examined in this study were about eight years old when tested by ELISA. Nevertheless, Fl antigen was detected at high levels (4-8 ug/ml) in two sera. This ELISA, therefore, appears to be valuable for the retrospective diagnosis of plague and could be used to test the sera from many clinically suspect but unconfirmed cases that have occurred over recent years in the United States, or elsewhere, so that more accurate assessments of the former incidence of this disease could be obtained. Other applications of this ELISA for the diagnosis of plague may include the detection of Fl antigen in bubo aspirates, in sputum from cases of suspect pneu- monic plague or asymptomatic carriers, in autopsy materials, and in rodent or flea tissues collected in plague surveillance and control programmes. In add- ition, this ELISA might be useful to determine the Fl content of plague vaccines for standardization of the product or to determine the potential immuno- genicity. Additional work is needed to evaluate these potentials. In order to achieve a full evaluation and utilization of ELISA, the procedure must be made suitable for use in the developing countries and in field situations. As the test requires special reagents (e.g., monoclonal antibody), the best alternative appears to be a "field kit" that could be supplied when and where a known or suspected plague emergency arises. We have initi- ated a series of studies on such a kit using lyophilized reagents. Shelf-life studies over six months now indi- cate that a kit is, indeed, feasible. ACKNOWLEDGEMENTS The sera examined in this study were collected by Dr T. Butler in 1974. This study was funded in part by contract no. NOI Al 92616 and grant no. NOI Al 02660 from the National Institute of Allergy and Infectious Diseases and by the Thrasher Research Fund, Salt Lake City, Utah. 465 466 J. E. WILLIAMS ET AL. RESUME TITRAGE PAR IMMUNO-ENZYMOLOGIE DES ANTIGENES SANGUINS DANS LES CAS DE PESTE AIGUt Une epreuve ELISA indirecte, n'utilisant pas le principe de competition, a et mise au point pour mesurer dans les liquides biologiques la concentration de P'antigene Fl specifique du bacille pesteux. L'epreuve est pratiquee sur plaque de microtitrage A l'aide d'un anticorps monoclonal fixant l'antigene. La sensibilite est de 0,4 ng d'antigene Fl (8 ng/ml) au seuil de signification de 1 Olo (P = 0,01). Une technique ELISA par inhibition a et utilisee pour confirmer la specificite des reactions. Dix echantillons de serum preleves sur des Vietnamiens atteints de peste bubonique confirmee par isolement du bacille ont et soumis A une double analyse: recherche d'antigenes dans le sang par ELISA et recherche d'anticorps anti-antigene Fl par hemagglutination passive. L'antigene a e decouvert chez deux patients A la concentration de 4-8 gg de Fl par ml de serum. L'anticorps n'a pas et decele. L'antigene etait absent chez deux autres patients presentant un titre d'anticorps de 1/64-1/128. Chez les 6 autres patients, on n'a mis en evidence ni l'anticorps ni l'antigene. En fait, les chances de poser correctment le diagnostic de la peste sont augmentees quand on recherche a la fois l'antigene et l'anticorps dans le serum des malades. Parmi les applications eventuelles de cette technique ELISA, il convient de noter la mise en evidence d'antigene Fl dans le serum, dans les ponctions de bubons, dans les crachats de cas suspects de peste pulmonaire ou de porteurs asymptomatiques, dans les prelevements necropsiques et dans les tissus de rongeurs et de puces recueillis dans le cadre des programmes de lutte et de surveillance antipesteux. Cette meme technique pourrait egalement etre utilisee pour determiner la teneur en Fl des vaccins antipesteux en vue de normaliser ces produits ou d'en estimer l'immunogenicite potentielle. REFERENCES 1. YOLKEN, R. H. Enzyme immunoassays for the detec- tion of infectious antigens in body fluids: current limit- ations and future prospects. Reviews of infectious diseases, 4: 35-68 (1982). 2. BAHMANYAR, M. & CAVANAUGH, D. C. Plague manual. Geneva, World Health Organization, 1976. 3. YOLKEN, R. H. & LEISTER, F. J. Investigation of enzyme immunoassay time courses: development of rapid assay systems. Journal of clinical microbiology, 13: 738-741 (1981). 4. WILLIAMS, J. E. & ROBINSON, D. M. Requirement to confirm the specificity of ELISA reactions. Trans- actions of the Royal Society of Tropical Medicine and Hygiene, 76: 280-281 (1982). 5. WILLIAMS, J. E. ET AL. Comparison of passive haemag- glutination and enzyme-linked immunosorbent assay for serodiagnosis of plague. Bulletin of the World Health Organization, 60: 777-781 (1982). 6. SOERGEL, M. E. ET AL. Solid-phase radioimmunoassay for detection of plague antigen in animal tissues. Journal of clinical microbiology, 16: 953-956 (1982).