reagents in a rapid extraction of antigen procedure to free heat-stable ... eliminated by each of the six extraction reagents. ..... Baker, C. J., and M. A. Rench. 1983 ...
Vol. 20, No. 5
JOURNAL OF CLINICAL MICROBIOLOGY, Nov. 1984, p. 981-984
0095-1137/84/110981-04$02.00/0
Improved Detection of Bacterial Antigens by Latex Agglutination After Rapid Extraction from Body Fluids L. PATRICK SMITH,l.2* KENNETH W. HUNTER, JR.,2'3 VAL G. HEMMING,2 AND GERALD W. FISCHER2
Infectious Diseases Division, Naval Hospital,1 and Departments of Pediatrics and Preventive Medicine2 and Biometrics,3 Uniformed Services University of the Health Sciences, Bethesda, Maryland 20814 Received 29 December 1983/Accepted 4 May 1984
Nonspecific agglutination of antibody-coated latex particles, unrelated to the presence of specific bacterial antigens, is a major difficulty with commercial latex particle agglutination tests. Rheumatoid and other factors are known to interfere with latex tests. We studied the use of six chelating, reducing, and anticoagulatory reagents in a rapid extraction of antigen procedure to free heat-stable antigens of Haemophilus influenzae type b and group B streptococcus which had been added to human sera. We also screened sera for the incidence of nonspecific agglutination from the three following groups: 123 patients with positive serology tests, 112 hospitalized patients, and 87 blood donors. The rapid extraction of antigen procedure involved a 1:4 dilution of the sera with each of the six reagents, incubation at 100°C for 3 min, and centrifugation at 13,000 x g for 5 min. Two commercial latex kits were tested (Bactigen and Wellcogen). Nonspecific agglutination was entirely eliminated by each of the six extraction reagents. Sera from 52% of the patients with positive serology tests, 29% of the hospitalized patients, and 28% of the blood donors showed nonspecific agglutination with Bactigen before extraction. Nonspecific agglutination was eliminated in all but one sample after the rapid extraction of antigen procedure. This simple, rapid extraction procedure eliminated nonspecific reactions in cerebrospinal fluids and amniotic fluids and reduced this problem in urines and sera with each commercial kit used on clinical specimens.
The immunochemical detection of bacterial antigens has become routine in many laboratories. Various methods exist, but not all are equally applicable since they vary in sensitivity, specificity, ease of performance, and cost of analysis. Latex agglutination (LA) and coagglutination tests for cerebrospinal fluid (CSF) are among the commercial products which have appeared in the last 2 years. Whether the antibodies are attached to protein A of staphylococci or coated on spheres of latex, nonspecific aggregation (NSA) of these antibody-coated particles, unrelated to the presence of specific bacterial antigen, is a significant problem (13, 23). Rheumatoid and other nonspecific factors have been noted to interfere with LA tests (3, 9, 12, 18, 19, 25), and several methods have been reported to be helpful in removing interfering factors (6, 7). In a previous study, we reported that serum and amniotic fluid could not be used directly with Wellcogen Strep B reagents and that a modification to the method of the manufacturer was necessary (11). We also reported that the use of the serum buffer in the original Bactigen Haemophilus influenzae type b kit did not completely eliminate NSA (L. P. Smith, K. W. Hunter, and G. W. Fischer, Abstr. Third Int. Symp. Rapid Methods Autom. Microbiol. 1981, abstr. no. 144). Daum et al. (5) evaluated the Bactigen kit and reported that 18.4% of fresh sera and 15.5% of specimens heated at 56°C for 30 min gave inconclusive results. In this study, we examined six chelating, reducing, and anticoagulatory agents for their usefulness in a rapid extraction technique to free heat-stable H. influenzae type b and group B streptococcal antigens from human serum samples. Human sera were also tested to assess the incidence of NSA with commercial LA bacterial antigen detection reagents. Those which induced NSA were then extracted and retested to examine the effectiveness of the
rapid extraction methodology for the elimination offactors in the sera which caused NSA. The rapid extraction technique was then utilized to analyze various clinical specimens for the presence of these two bacterial antigens. MATERIALS AND METHODS Reagents. All reagents were prepared both in sterile, distilled water and in sterile, normal saline. Trichloroacetic acid (Sigma Chemical Co., St. Louis, Mo.), N-acetyl-Lcysteine (Eastman Kodak Co., Rochester, N.Y.), ethylene glycol-bis (,B-aminoethyl)N,N,N'N'-tetraacetic acid (EGTA) (Sigma), and EDTA (Sigma) were prepared to 0.05, 0.1, and 0.2 M. Concentrations of dithiothreitol (Kodak) were 0.03, 0.003, and 0.001 M. Polyanetholesulfonic acid (Sigma) concentrations were 1,000, 500, and 250 ,ug/ml. Each reagent preparation was adjusted to a pH of 7.2 to 7.4. Bacterial antigens. H. influenzae type b polyribosyl ribitol phosphate (PRP) capsular polysaccharide was kindly supplied by John Robbins, Bureau of Biologics, Bethesda, Md. Group-specific group B streptococcal (GBS) antigen was a gift from Max Moody, Burroughs Wellcome Co., Research Triangle Park, N.C. Human sera. Three groups were screened for the incidence of NSA by two commercial LA kits. A control group of 87 sera was obtained from healthy blood donors. Sera from 112 hospitalized patients submitted to the clinical chemistry laboratory were also screened. A total of 123 sera demonstrating the following positive serology tests constituted the third group: fluorescent treponemal antibody, rheumatoid factor, antinuclear antibodies, rapid plasma reagin, antinative DNA, thyroid antibodies, heterophile antibodies, and C-reactive protein. Nineteen sera, producing NSA by one or both commercial LA kits, were pooled to provide sufficient sera for the study. After extraction, the pool was NSA negative.
* Corresponding author.
981
982
J. CLIN. MICROBIOL.
SMITH ET AL.
Clinical specimens. A variety of body fluids were submitted to the Infectious Diseases laboratory during 1983 for antigen analysis: 103 urines, 62 sera, 47 CSF, and 21 amniotic fluids. Most of these specimens were examined by each commercial kit and, after extraction, were reexamined. Latex reagents. Bactigen H. infliuenzae type b was obtained from Wampole Laboratories, Cranbury, N.J. Wellcogen Strep B was obtained from Burroughs Wellcome Co. Reagents were warmed to room temperature and used according to the recommended procedures of the manufacturers. Bactigen sensitized latex (10 [LI) was placed in one circle, and 10 pl1 of negative control latex was placed into a separate circle on a clear serological slide. Fifty microliters of each test fluid was added and mixed with the latex suspensions, and after mechanical rotation (150 rpm) for 10 min, the slides were inspected in oblique light over a black background to detect particle agglutination. LA patterns were defined as follows: 0, smooth, unchanged appearance; 1+, fine granular background; 2+, small but definite clumps against a cloudy background; 3+, moderate clumps against a slightly cloudy background; and 4+. large clumps against a clear background. A positive result was defined as a 2 to 4+ pattern in the test latex with 0 to 1+ in the control latex. A negative result showed 0 to 1 + in each reagent. An inconclusive interpretation resulted from 2 to 4+ in both test and control latex reagents. Wellcogen test latex (20 Rl) was placed in one circle on a tile, and 20 [L1 of negative control latex was placed in a separate circle. Forty microliters of each test fluid was added and mixed with the latex reagents. Tiles were rocked 20 times per min on an Ames Aliquot Mixer for 3 min and examined at normal reading distance from the eyes. A positive test was defined as visible clumping of latex reagents. Failure of LA and persistence of a milky appearance was interpreted as a negative test. Controls. After serial dilutions of pooled sera, distilled water, and normal saline containing the bacterial antigens, the minimum antigen concentration detected by each kit, before and after extraction, was determined. Samples of pooled sera, distilled water, and normal saline were made up to contain 4.0 ng of PRP per ml or 320 pLg of GBS antigen per ml. REAP. The rapid extraction of antigen procedure (REAP) was as follows. Pooled human sera (50 [lI) containing either PRP or GBS antigen was placed in 1.5-ml plastic microcentrifuge tubes, and 150 [I1 of the reagent to be tested was added. Tubes were vortexed and placed for 3 min in a 100°C dry bath filled with aluminum shavings. After incubation, the tubes were centrifuged for 5 min at 13,000 x g in a Fisher Microfuge, and then the supernatants were removed and tested for the presence of antigen. RESULTS
Removal of latex interferences. The effect of REAP on the pool of 19 human sera without bacterial antigens added is depicted in Table 1. All three concentrations of the six extraction reagents that were tested effectively eliminated the NSA of latex reagents observed with each of the commercial bacterial antigen detection kits. The only differences noted were the clarity of the supernatants and the amount of precipitated protein after the centrifugation step of the extraction procedure. In general, the most concentrated reagents produced the best pellets. A 1:4 dilution was the minimum concentration for prevention of irreversible serum gelatification after extraction with any of the six reagents. However, after a 1:4 dilution of sera with distilled water or
TABLE 1. Effect of REAP" on bacterial antigen detection by two commercial latex kits Welicogen Bactigen REAP
PRP results
GBS results
Pooled serah
Before After
NSA' Neg
NSA Neg
Spiked" pool
Before After
NSA Pos
NSA Pos
Distilled water and normal saline
Before
Neg
Neg
After
Neg
Neg
Before
Pos
Pos
Sample
Spiked' distilled water and normal saline
Pos Pos "Three concentrations of each of six chelating, reducing, and anticoagulatory reagents were used. Since no differences were noted, data are pooled
After
from each experiment. " Individually determined to produce NSA by one or both kits. screened for absence of PRP and GBS. and then pooled for this experiment. NSA means that both test aind control reagents were clumped, giving inconclusive interpretation of results. d Antigen concentration before REAP: 4.0 ng of PRP per ml, and 320 [kg of GBS per ml: after REAP: 1.0 ng of PRP per ml. and 80 ,ug of GBS per ml. "
normal saline was used, the supernatant remained gelatinous. The addition of the reagents without heat or with heat but without centrifugation failed to eliminate NSA of latex reagents. The maximum speed of a commercial laboratory centrifuge (ca. 1,650 x g) did not pellet the precipitated protein after 1 h. Hence, the combination of three steps, dilution with an extraction reagent, incubation at 100°C for 3 min, and high-speed centrifugation, were necessary steps for successful elimination of nonspecific latex interference factors in human sera. Detection of bacterial antigens. IThe minimum antigen concentration detected by the commercial kits in distilled water and in saline was 1.0 ng of PRP per ml by Bactigen and 80 .'g of GBS per ml by Wellcogen. Once it had been demonstrated that each reagent removed the latex interferences from the pooled sera, REAP was performed with each reagent on antigen-spiked control serum pools. No differences were noted in the effectiveness of the six reagents. The amount of bacterial antigen detected in the spiked sera after REAP equaled that detected in the distilled water and normal saline controls (Table 1). Incidence of NSA of latex reagents in human sera. Altogether, 24 of 87 (28%) sera from blood donors, 33 of 112 (29%) sera from hospitalized patients, and 64 of 123 (52%) positive serology sera that were tested with Bactigen H. influenzae type b reagents gave NSA reactions. After REAP, with only the 0.1 M EDTA reagent, just 1 of these 322 sera demonstrated persistent NSA (Table 2). Differences in the groups were highly statistically significant (P < 0.001, chi-square). Table 3 presents a breakdown of the positive serology specimens into eight tests and shows the percentage of NSA found in each test with the Bactigen reagents. Antigen detection in clinical specimens. Urine, serum, CSF, and amniotic fluid specimens demonstrated various degrees of NSA in the two commercial LA kits. Table 4 shows the results of our analyses. REAP reduced the incidence of NSA in urines and sera and eliminated NSA in CSF and amniotic fluids. NSA was more evident with Wellcogen than with Bactigen. After REAP, the number of specimens giving a positive test with each kit increased in all body fluids tested.
TABLE 2. Nonspecific LA" demonstrated by human sera % NSA After extraction
Before extraction
Specimen (no.)
Blood donors (87) Hospitalized (112) Positive serology (123) a Bactigen H. inflienzae type b kit.
1 0 0
28 29 52
DISCUSSION Accurate detection of bacterial antigen in body fluids by agglutination techniques has been hampered by rheumatoid and other nonspecific factors (12, 19). These interferences usually cause the control and test reagents to agglutinate. The interpretation of this reaction pattern is inconclusive; the bacterial antigen in question may or may not be present. A variety of methods have been reported to reduce this nonspecific activity in clinical specimens, including heat (1. 6, 7, 13, 16, 23, 24), EDTA (6, 7, 9, 11), N-acetyl-L-cysteine (26), EGTA (12), and dithiothreitol (2, 10, 22). Other investigators have reported the use of either trichloroacetic acid or EDTA to extract bacterial antigens (14, 15, 18), polyanetholesulfonic acid to precipitate serum protein (20, 21), and EDTA to enhance the sensitivity of agglutination reactions (12). After determining the high incidence of NSA in human sera with commercial latex reagents for bacterial antigen detection, we developed a simple, rapid method to remove these interferences. We found none of these reagents to be superior for detecting PRP or GBS antigens and chose EDTA as the reagent for the remainder of the study and for routine processing of clinical specimens. Previous reports showed that brief exposure to EDTA was a useful means of releasing cell wall lipopolysaccharide antigens from bacteria (15. 18), actually dissociating the resulting polymer into smaller subunits (4, 17). EDTA, an effective chelator of both calcium and magnesium, prevents serum complement activation better than EGTA (8). REAP did not adversely affect PRP or GBS antigens (Table 1). A 1:4 dilution was the minimum necessary to allow the precipitated protein to pellet after a >10.000 x g centrifugation. This dilution step also facilitated routine screening of small volumes of patient specimens submitted for antigen detection. TABLE 3. Nonspecific LA" of 106 frozen" human sera with positive serology results % NSA Postive
test'
CRP RPR FTA RA ANA DNA TA MONO
983
BACTERIAL ANTIGEN DETECTION
VOL. 20, 1984
No. tested
2 6 5 14
26 1 4 48
Before
After
extraction
extraction
0 100 100 93 50 100 100 38
0 0 0 0 0
0 0
0
Bactigen H. influenZae type b kit. Sera were frozen at -'20C. CRP, C-reactive protein; RPR. rapid plasma reagin; FTA. fluorescent treponemal antibody; RA. rheumatoid factor: ANA. antinuclear antibodies: MONO. heterophile antibodies.
TABLE 4. Analysis of clinical specimens submitted for antigen detection ~~% Positive %NSA Specimen
Urine (103)
NSA
REAP
(no.)
Before After
agglutination
Wellcogen'
Bactigen
Welicogen
4
63 18
15 22
4 12
Bactigen' 15
Serum (62)
Before After
17 4
29 0
10 21
5 8
CSF (47)
Before After
0 0
4 0
13 23
13 15
Amniotic fluid (21)
Before After
NDd ND
43 0
ND ND
43 86
a 0.1 M EDTA reagent used.
Bactigen H. infllenzae type b. 'Wellcogen Strep B. ND, Not done.
d
The incidence of NSA induced by Wellcogen Strep B while detecting GBS antigen in amniotic fluid, serum, and CSF was overcome by REAP. Urine NSA in Wellcogen Strep B and in Bactigen H. influenzae type b reagents was markedly decreased by REAP. Thus, rapid antigen extraction pretreatment of body fluids markedly reduced the incidence of NSA and allowed specimens from patients to be examined for the presence of bacterial polysaccharide antigens. ACKNOWLEDGMENTS We thank Sandra Sebers, Doris Gracy, Stacie Colwell, and Catherine Waters for excellent laboratory assistance. Portions of this work were supported in part by project CIRC 8206-1706 from the Naval Medical Command, Navy Department, Washington, D.C.. and by grant C08606 from the Uniformed Services University of the Health Sciences, Bethesda, Md.
LITERATURE CITED 1. Baker, C. J., and M. A. Rench. 1983. Commercial latex agglutination for detection of group B streptococcal antigen in body fluids. J. Pediatr. 102:393-395. 2. Behan, K. A., and G. C. Klein. 1982. Reduction of Brucelll species and Francisella titlarensis cross-reacting agglutinins by dithiothreitol. J. Clin. Microbiol. 16:756-757. 3. Bennett, J. E., and J. W. Bailey. 1971. Control for rheumatoid factor in the latex test for cryptococcosis. Am. J. Clin. Pathol. 56:360-365. 4. Cox, S. T., Jr., and R. G. Eagon. 1968. Action of ethylenediaminetetraacetic acid. tris (hydroxymethyl) aminomethane. and lysozyme on cell walls of Pseiudotinonais ieruiginost. Can. J. Microbiol. 14:913-922. 5. Daum, R. S., G. R. Siber, J. S. Kamon, and R. R. Russell. 1982. Evaluation of a commercial latex particle agglutination test for rapid diagnosis of Hauemoplilis influencae type b infection. Pediatrics 69:466-471. 6. Doskeland, S. O., and B. P. Berdal. 1980. Bacterial antigen detection in body fluids: methods for rapid antigen concentration and reduction of nonspecific reactions. J. Clin. Microbiol. 11:380-384. 7. Eng, R. H. K., and A. Person. 1981. Serum cryptococcal antigen determination in the presence of rheumatoid factor. J. Clin. Microbiol. 14:700-702. 8. Fine, D. P., S. R. Marney, D. G. Colley, J. S. Sergent, and R. M. Des Prez. 1972. C3 shunt activation in human serum chelated with EGTA. J. lmmunol. 109:807-809. 9. Gentry, L. O. I. D. Wilkinson, A. S. Lea, and M. F. Price. 1983. Latex agglutination test for detection of candida antigen in
984
10. 11.
12.
13.
14.
SMITH ET AL.
patients with disseminated disease. Eur. J. Clin. Microbiol. 2: 122-128. Gordon, M. A., and E. W. Lapa. 1974. Elimination of rheumatoid factor in the latex test for cryptococcosis. Am. J. Clin. Pathol. 61:488-494. Hemming, V. G., W. T. London, L. P. Smith, B. L. Curfman, G. W. Fischer, and J. L. Sever. 1983. Detection of group B streptococcal antigens in amniotic fluid of rhesus mnonkeys. J. Clin. Microbiol. 17:1127-1131. Izakson, I., and S. A. Morse. 1981. Enhancement of coagglutination reactions of the Phadebact gonococcus test by ethylenediaminetetraacetate and ethylene glycol-bis(,B-aminoethyl ether)N,N-tetraacetate. J. Clin. Microbiol. 14:261-265. Jones, D. E., K. S. Kanarek, J. L. Angel, and D. V. Lim. 1983. Elimination of multiple reactions of the Phadebact Streptococcus coagglutination test. J. Clin. Microbiol. 18:526-528. Lancefield, R. C., and E. H. Freimer. 1966. Type-specific polysaccharide antigens of group B streptococci. J. Hyg. 64:
191-203. 15. Leive, L. 1965. Release of lipopolysaccharide by EDTA treatment of E. coli. Biochem. Biophys. Res. Commun. 21:290-292. 16. Newman, R. B., R. W. Stevens, and H. A. Gaafar. 1970. Latex agglutination test for the diagnosis of Haeinophilius inflienzae meningitis. J. Lab. Clin. Med. 76:107-113. 17. Olins, A. L., and R. C. Warner. 1967. Physiochemical studies on a lipopolysaccharide from the cell wall of Aztohbacter inelandii. J. Biol. Chem. 242:4994-5001. 18. Rogers, S. W., H. E. Gilleland, Jr., and R. G. Eagon. 1969. Characterization of a protein-lipopolysaccharide complex released from cell walls of Pseiudolm2onais aeruiiginosai by ethylene-
J. CLIN. MICROBIOL.
diaminetetraacetic acid. Can. J. Microbiol. 15:743-748. 19. Thirumoorthi, M. C., and A. S. Dajani. 1979. Comparison of staphylococcal coagglutination, latex agglutination, and counterimmunoelectrophoresis for bacterial antigen detection. J. Clin. Microbiol. 9:28-32. 20. Traub, W. H., and P. 1. Fukushima. 1978. Neutralization of human serum lysozyme by sodium polyanethol sulfonate but not by sodium amylosulfate. J. Clin. Microbiol. 8:306-312. 21. Traub, W. H., and B. L. Lowrance. 1970. Anticomplementary, anticoagulatory, and serum-protein precipitating activity of sodium polyanetholsulfonate. Applied Microbiol. 20:465-468. 22. Ward, J. I., G. k. Siber, D. W. Scheifele, and D. H. Smith. 1978. Rapid diagnosis of Hemophilus influenzae type b infections by latex particle agglutination and counterimmunoelectrophoresis. J. Pediatr. 93:37-42. 23. Wasilauskas, 1B. L. 1981. Staphylococcal coagglutination methods for rapid microbial detection and identification. Lab. Med. 12:411-414. 24. Webb, B. J., and C. J. Baker. 1980. Commercial latex agglutination test for rapid diagnosis of group B streptococcal infection in infants. J. Clin. Microbiol. 12:442-444. 25. Winchester, R. 1980. The nature of rheumatoid factors and their potential for interfering with assays. In K. McIntosh, C. Wilfert, M. Chernesky, S. Plotkin, and M. J. Mattheis (ed.), Summary of a workshop on new and useful techniques in rapid viral diagnosis, May 1980. National Institutes of Health, Bethesda, Md. (Suinmarized in J. Infect. Dis. 142:793-802.) 26. Yolken, R. H., and P. J. Stopa. 1979. Analysis of nonspecific reactions in enzyme-linked immunosorbent assay testing for human rotavirus. J. Clin. Microbiol. 10:703-707.
