... H. YOLKEN,1* DAWN DAVIS,' JERRY WINKELSTEIN,2HAROLD RUSSELL,3 AND JOHN E. SIPPEL4 ..... Fossieck, B., Jr., R. Craig, and P. Y. Paterson. 1973.
Vol. 20, No. 4
JOURNAL OF CLINICAL MICROBIOLOGY, OCt. 1984, p. 802-805 0095-1137/84/100802-04$02.00/0
Copyright © 1984, American Society for Microbiology
Enzyme Immunoassay for Detection of Pneumococcal Antigen in Cerebrospinal Fluid ROBERT H. YOLKEN,1* DAWN DAVIS,' JERRY WINKELSTEIN,2 HAROLD RUSSELL,3 AND JOHN E. SIPPEL4 Eudowood Division of Infectious Diseases' and Eudowood Division of Immunology,2 Department of Pediatrics, The Johns Hopkins Hospital, Baltimore, Maryland 21205; Respiratory and Special Pathogens Laboratory Branch, Division of Bacterial Diseases, Centers for Disease Control, Atlanta, Georgia 303333; and Naval Biosciences Laboratory, Naval Supply Center, Oakland, California 946254 Received 25 April 1984/Accepted 29 May 1984
A solid-phase immunoassay utilizing horse antiserum against the C polysaccharide of Streptococcus pneumoniae and biotinylated rabbit antibodies to type-specific pneumococcal polysaccharides was developed to detect pneumococcal antigens in human body fluids and in broth cultures. Pneumococcal antigen could be detected in broth cultures of serotypes of S. pneumoniae containing as little as 102 to 103 organisms per ml. The assay system detected pneumococcal antigen in all 25 cerebrospinal fluid specimens obtained from patients with documented pneumococcal meningitis. There were no positive reactions noted in specimens from patients infected with Neisseria meningitidis group A or from patients without evidence of bacterial infection. The solidphase enzyme immunoassay utilizing these reagents is a sensitive and specific assay for the immunodetection of a wide range of pneumococcal antigens.
saccharide to formulate an efficient enzyme immunoassay for the detection of pneumococcal antigens in CSF obtained from patients with pneumococcal meningitis. MATERIALS AND METHODS Broth cultures of 16 serotypes of S. pneumoniae were obtained from clinical isolates or the American Type Culture Collection, Rockville, Md. The organisms were grown in stationary broth cultures (19), using brain heart infusion medium to concentrations of 106 organisms per ml (as determined by plating on blood agar), diluted, and tested in the immunoassay system without further processing. CSF specimens obtained between 1979 and 1983 at the time of hospital admission from 25 children and adults with meningitis caused by S. pneumoniae and 25 patients with meningitis due to Neisseria meningitidis group A were provided by Nabil Guirgis, Naval Medical Research Unit 3, Cairo, Egypt. The specimens were streaked on blood and chocolate agar plates, incubated in an atmosphere of approximately 10% carbon dioxide (candle jar) at 37°C, and examined after 24 and 48 h. Alpha-hemolytic gram-positive diplococci were confirmed to be S. pneumoniae by testing for optochin susceptibility, using disks. Oxidase-positive gram-negative diplococci were identified as Neisseria species by sugar fermentation tests and were placed in serogroups by slide agglutination tests. Twenty-three of the pneumococcal specimens and 19 of the meningococcal CSF samples were culture positive. The remaining specimens were diagnosed by detection of specific antigen or by coagglutination. The CSF specimens were immediately frozen at -70°C, transported to the United States on dry ice, and stored at -70°C until testing. CSF specimens from patients without bacterial meningitis were obtained from children at The Johns Hopkins Hospital who were undergoing diagnostic lumbar puncture for suspected meningitis. Antibody to pneumococcal C polysaccharide prepared in a horse was obtained from M. Heidelberger, Columbia University School of Medicine, New York, N.Y. Pooled rabbit antiserum to pneumococcal polysaccharide (Omni serum) was obtained from the Statens Seruminstitut, Copenhagen,
Streptococcus pneumoniae is a major cause of systemic bacterial infections in humans (1, 11). The potentially serious nature of systemic pneumococcal disease requires the early diagnosis of infection to ensure the institution of appropriate antimicrobial chemotherapy (2, 5, 16, 17). Thus there is a need for a rapid diagnostic assay capable of detecting and identifying pneumococcal antigens in body fluids, especially in patients with sickle cell disease and other immunodeficiency states. One problem in the formulation of assays for the direct detection of pneumococcal antigens in body fluids is the fact that the capsular polysaccharide of S. pneumoniae can exist in multiple antigenic types (4, 7, 12). Available antisera to S. pneumoniae are generally prepared against purified type-specific polysaccharides which are antigenically distinct (8). In preliminary attempts to develop a solidphase immunoassay system for the detection of S. pneumoniae, we found that antisera directed at individual serotypes of capsular polysaccharides could be used to detect the homologous serotypes but could not efficiently detect heterologous serotypes of S. pneumoniae. Furthermore, the use of pooled reagents on the solid phase did not result in a sensitive assay system for the detection of pneumococcal antigens. Also, earlier studies showed that a double-antibody sandwich enzyme-linked immunosorbent assay, using the same pooled rabbit antipneumococcal serum for both capture and detection, was unable to effectively differentiate pneumococcal cerebrospinal fluid (CSF) specimens from specimens obtained from meningococcal meningitis patients (18a). The C polysaccharide of S. pneumoniae is a cholinecontaining cell wall surface antigen derived from teichoic acid (14, 17, 19). Since C polysaccharide antigenic activity is widely prevalent in different serotypes of S. pneumoniae, antisera to the C polysaccharide can react with a wide range of pneumococcal organisms. We used antiserum to C poly-
*
Corresponding author. 802
VOL. 20, 1984
Denmark. Avidin covalently linked to peroxidase was obtained from Vector Laboratories, Sunnyvale, Calif. The immunoglobulin G fraction of the pooled rabbit antiserum to the pneumococcal capsular polysaccharide was prepared by ammonium sulfate precipitation and ion-exchange chromatography, utilizing the method of Fahey and Terry (6). The immunoglobulin G fraction was labeled with biotin by reaction with an N-hydroxysuccinimide ester as previously described (3). The enzyme immunoassay for S. pneumoniae antigen was performed by a modification of previously described enzyme immunoassays, utilizing biotin-labeled antibodies (21). The optimal dilution of all reagents was determined by checkerboard titration. Alternate wells in polyvinyl microtiter plates (Dynatech Laboratories, Inc., Alexandria, Va.) were coated with 100-,ul aliquots of horse antibody to the C polysaccharide or with nonimmune horse serum. After incubation overnight at 4°C, the microtiter plates were either used immediately or stored at 4°C. Immediately before use, the microtiter plates were washed five times with phosphatebuffered saline containing 0.05% Tween 20, and a 25-,ul aliquot each of broth culture, CSF, or control was diluted 1:4 in phosphate-buffered saline with 0.05% Tween 20 and added to the microtiter plates. After incubation for 1 h at 37°C, the microtiter plates were washed as described above, and a 100,ul aliquot of a 1:400 dilution of biotinylated, pooled rabbit antiserum to pneumococcal capsular polysaccharide was added to the wells. After another incubation for 1 h at 37°C, the microtiter plates were washed and peroxidase-labeled avidin diluted 1:10,000 in phosphate-buffered saline with 0.05% Tween 20 was added. After incubation for 30 min at 37°C, the plates were washed, and o-phenylenediamineperoxidase substrate, prepared as previously described, was added. After incubation for approximately 15 min at room temperature, the amount of color generated by the interaction of bound peroxidase avidin and substrate was measured at a wave-length of 492 nm in a microplate colorimeter (Dynatech MR580). For each specimen or dilution of broth culture, a specific activity was calculated by subtracting the optical density generated in wells coated with nonimmune horse serum from that generated in wells coated with horse antibody to C polysaccharide (23). For broth cultures, a dilution was considered to be positive if it yielded an optical density which was 3 standard deviations greater than the optical density of uninoculated broth. In the case of CSF, a specimen was considered positive for S. pneumoniae if it yielded an optical density value which was 3 standard deviations greater than the mean of negative CSF obtained from patients without evidence of bacterial infection. RESULTS Initial studies were directed at determining the sensitivity of the immunoassay system for the detection of purified pneumococcal antigen. The assay system utilizing antibody to the pneumococcal C polysaccharide could detect as little as 3 ng of a pool of purified pneumococcal polysaccharides. Additional studies were directed at the detection of pneumococcal organisms in broth cultures grown in brain heart infusion medium. For all 16 serotypes tested, a dilution corresponding to approximately 103 organisms per ml yielded a specific activity which was at least 3 standard deviations greater than the activity of uninoculated medium (Table 1). In addition, dilutions of types 1, 2, 3, 8, 11, 13, 16, and 37 corresponding to 102 organisms per ml were positive for pneumococcal antigen by this criterion. Broth cultures containing 105 organisms of the following types did not give
DETECTION OF PNEUMOCOCCAL ANTIGEN IN CSF
803
TABLE 1. Reactivity of pneumococcal organisms in an enzyme immunoassay utilizing anti-C-substance antibody (OD,pecimens ODbIanks)
1 2 3 4 7 8 9 11 12 13 14 16 21 25 37 42
lo4b 510 542 556 493 526 531 589 569 512 524 551 471 436 375 490 492
X
1,000"
lo,
102
267 227 552 195 206 280 204 363 316 287 314 174 124 116 262 200
33 34 241 10 22 75 10 88 81 32 20 44
10 26 54 20
" The mean t standard deviation of negative control specimens containing nonpneumococcal organisms was 13 + 06 U. Underlined values represent the minimum dilution which was 3 standrd deviations greater than the negative controls. OD, Optical density. bOrganisms per milliliter.
significant reactions for pneumococcal antigens: Streptococcus sp. group A, Streptococcus sp. group B, Streptococcus sp. group D, Haemophilus influenzae type B, N. meningitidis type A, N. meningitidis type A, N. meningitidis type C, Staphylococcus aureus, Staphylococcus epidermidis, and Escherichia coli. A total of 25 CSF specimens were available from patients with documented meningitis due to S. pneumoniae. The mean specific activity of these species was 1.050 ± 0.283 optical density units. All 25 CSF specimens from patients infected with S. pneumoniae yielded specific activity values which were at least 3 standard deviations greater than the mean of the negative control CSF specimens. On the other hand, the mean specific activity of the specimens obtained from the patients infected with N. meningitidis was 0.079 + 0.05 (P < 0.001 compared with S. pneumoniae). None of the specimens available from patients infected with N. meningitidis group A or from uninfected patients had values which were positive by this criterion (Fig. 1).
DISCUSSION The sensitivity of an immunoassay system depends to a great extent upon the concentration and affinity of the immunoreactants utilized to bind antigen. In all solid-phase immunoassays, one of the components of the immunoassay system is immobilized onto a solid surface such as a plastic bead or a microtiter plate (21). Whereas solid-phase immunoassays offer a number of advantages in terms of convenience and rapidity, the sensitivity of solid-phase immunoasays is limited by the concentration of immunoglobulin which can be efficiently bound to the solid-phase surface. For microtiter plates in common use, the amount of immunoglobulin bound to the solid phase is limited to approximately 1 ,ug per cm2 of surface area (approximately 0.5 ,g per microtiter well from a standard 96-well plate) (10, 13, 15). This limitation would be expected to be particularly problematic when a reagent consisting of a pool of a number of different antisera directed at distinct determinants is utilized
J. CLIN. MICROBIOL.
YOLKEN ET AL.
804
> 1. 5
bacterial infection. In addition, whereas sufficient numbers of body fluids were not available for detailed study, preliminary studies indicate that the immunoassay system described above was capable of detecting pneumococcal antigens in more accessible body fluids such as blood and pleural fluid specimens from patients with disease due to S. pneumoniae. The efficiency of this assay system for the detection of S. pneumoniae in body fluids other than CSF should be
r-
I
0
examined in future studies.
1.0k
aIn
ACKNOWLEDGMENTS This work was supported in part by Public Health Service contract N01-Al-22680 from the National Institute of Allergy and Infectious Diseases and by the Thrasher Research Fund.
0
t -
LITERATURE CITED 1. Austrian, R., and J. Gold. 1964. Pneumococcal bacteremia with a special reference to bacteremic pneumococcal pneumonia. Ann. Intern. Med. 60:759-764. 2. Barrett-Conner, E. 1971. Bacterial infection and sickle cell anemia. An analysis of 250 infections in 166 patients and a review of the literature. Medicine 50:97-112.
4'
0.5 p (n
.
0.2 0.1
'.05
A
B
C
FIG. 1. Results of enzyme immunoassay for detection of pneumococcal antigens in CSF. (A) S. pneumoniae; (B) N. meningitidis type A; (C) no organism identified. OD450, Optical density at 450 nm.
solid-phase surface, as the amount of immunoglobulin bound to the solid phase directed at each component might be less than that required to efficiently bind antigen during the immunoassay procedures (21). Therefore, we utilized antisera to pneumococcal C polysaccharide, an antigen which is widely present in pneumococcal organisms of different antigenic types (9, 18, 20), as the solid-phase immunoreactant in an immunoassay system for the detection of pneumococcal antigens in CSF of patients infected with different strains of S. pneumoniae. Initial attempts were directed toward generating an assay system utilizing this antibody both as the solid phase and as the labeled second reagent. However, we found that the more efficient assay used antibody to C polysaccharide bound to the solid phase with pooled antisera directed at type-specific capsular polysaccharides in the liquid-phase reagent. Because the solidphase and liquid-phase antibodies are directed at different determinants, this system offers the potential advantage that the binding of antigen to the solid phase will not result in a decrease in the number of antigenic sites available to the liquid-phase antibody (21). We found that the most efficient assay used biotinylated immunoglobulin as the liquid-phase reagent for binding by reaction with peroxidase-linked avidin. This finding is consistent with recent reports describing the advantages of avidin-biotin immunoassay systems (22). on a
The
assay systems
described could detect between 102 and
103 organisms/ml of 16 serotypes of S. pneumoniae grown in broth culture. When applied to CSF specimens, the assay system could successfully detect pneumococcal antigens in all 25 specimens obtained from patients with documented pneumococcal meningitis. No positive reactions were noted in CSF specimens obtained from patients infected with N. meningitidis group A or from patients without evidence of
3. Bayer, E. A., and M. Wilchek. 1974. Avidin-biotin interactions. Methods Enzymol. 34:265-267. 4. Cone, L., and J. W. Uhr. 1964. Immunological deficiency disorders assoqiated with chronic lymphocytic leukemia and multiple myeloma. J. Clin. Invest. 43:2241-2248. 5. Cooksey, R. C., R. R. Facklam, and C. Thornsberry. 1978. Antimicrobial susceptibility patterns of Streptococcus pneumoniae. Antimicrob. Agents Chemother. 13:645-648. 6. Fahey, J. L., and E. W. Terry. 1973. Ion exchange chromotography and gel filtration of immunoglobulin, p. 71-76. In D. M. Weir (ed.), Immunochemistry, Blackwell Scientific Publications, London. 7. Fine, D. P. 1975. Pneumococcal type-associated variability in alternate complement pathway activation. Infect. Immun. 12:772-778.
8. Fossieck, B., Jr., R. Craig, and P. Y. Paterson. 1973. Counterimmunoelectrophoresis for rapid diagnosis of meningitis due to
Diplococcus pneumoniae. J. Infect. Dis. 127:106-109. 9. Gotschlich, E. C., and T. Y. Liu. 1967. Structural and immunological studies on the pneumococcal C-polysaccharide. J. Biol.
Chem. 242:463-470. 10. Herman, J. E., and M. F. Collins. 1976. Quantitation of immunoglobulin adsorption to plastics. J. Immunol. Methods
10:363-366. 11. Hodges, R. G., and C. M. MacLeod. 1946. Epidemic pneumococcal pneumonia. V. Final considerations of the factors underlying the epidemic. Am. J. Hyg. 44:237.
12. Larm, O., and B. Lindberg. 1976. The pneumococcal polysaccharides: a re-examination. Adv. Carbohydr. Chem. 33:295-322.
13. Lehtonen, 0. P., and M. C. ViUanen. 1980. Antigen attachment in ELISA. J. Immunol. Methods 34:61-70. 14. Liu, T.-Y., and E. C. Gotschlich. 1963. The chemical composition of pneumococcal C-polysaccharide. J. Biol. Chem. 238:1928-1934.
15. Place, J., and H. R. Schroader. 1982. The fixation of anti-HbsAg on plastic surfaces. J. Immunol. Methods 48:251-260. 16. Robinson, M. G., and R. J. Watson. 1966. Pneumococcal meningitis in sickle-cell anemia. N. Engl. J. Med. 274:1006. 17. Rytel, M. W. 1975. Rapid diagnostic methods in infectious diseases. Adv. Intern. Med. 20:37. 18. Schiffman, G., D. L. Bornstein, and R. Austrian. 1971. Capsulation of pneumococcus with soluble cell wall-like polysaccharide. II. Nonidentity of cell wall and soluble cell wall-like polysaccharides derived from the same and from different pneumococcal strains. J. Exp. Med. 134:600-617.
18a.Sippel, J. E., C. M. Prato, N. I. Girgis, and E. A. Edwards. 1984. Detection of Neisseria meningitidis group A, Haemophilus influenzae type b, and Streptococcus pneumoniae antigens in cerebrospinal fluid specimens by antigen capture enzyme-linked immunosorbent assays. J. Clin. Microbiol. 20:259-265. 19. Smith, M. R., and W. B. Wood, Jr. 1969. Heat labile opsonins
VOL. 20, 1984 I. Participation of complement. J. Exp. Med. 130:1209-1227. 20. Tomasz, A. 1981. Surface components of Streptococcus pneumoniae. Rev. Infect. Dis. 3:190-211. 21. Yolken, R. H. 1982. Enzyme immunoassays for the detection of infectious antigens in body fluids: current limitations and future prospects. Rev. Infect. Dis. 4:35-68.
to pneumococcus.
DETECTION OF PNEUMOCOCCAL ANTIGEN IN CSF
805
22. Yolken, R. H., F. J. Leister, L. S. Whitcomb, and M. Santosham. 1983. Enzyme immunoassays for the detection of bacterial
antigens utilizing biotin-labelled antibody and peroxidase biotinavidin complex. J. Immunol. Methods 56:319-327. 23. Yolken, R. H., and P. J. Stopa. 1979. Analysis of nonspecific reactions in enzyme-linked immunosorbent assay testing for human rotavirus. J. Clin. Microbiol. 10:703-710.
