Immunofluorescence Assay for Identification of Legionella

Transcription

1 JOURNAL OF CLINICAL MICROBIOLOGY, Nov. 1983, p /83/ $2./ Copyright C 1983, American Society for Microbiology Vol. 18, No. 5 Comparison of Slide Agglutination Test and Direct Immunofluorescence Assay for Identification of Legionella Isolates W. LANIER THACKER,l* HAZEL W. WILKINSON,' AND ROBERT F. BENSON2 Division ofbacterial Diseases' and Biological Products Program,2 Center for Infectious Diseases, Centers for Disease Control, Atlanta, Georgia 3333 Received 23 June 1983/Accepted 5 August 1983 It is technically impractical for many clinical laboratories to use the direct immunofluorescence assay for identifying and serogrouping clinical isolates of Legionella. We compared the results obtained with the direct immunofluorescence assay with the results of a simple and less-demanding slide agglutination test for identifying 15 serogroups representing seven Legionella species. The slide agglutination test was in complete agreement with the direct immunofluorescence assay, and the serogroup to which 64 clinical isolates of Legionella belonged was correctly identified. With polyvalent, pooled antisera and absorbed, serogroupspecific antisera, the slide agglutination test is a useful alternative to the direct immunofluorescence assay in the diagnosis of Legionella infections and for studying the serological relationships of Legionella-like organisms. The direct immunofluorescence assay (DFA) is used to identify the species and serogroup of Legionella strains either in infected tissue of patients with legionellosis or in isolated cultures (7). At present, the Centers for Disease Control (CDC) routinely tests specimens with DFA conjugates for six Legionella species: Legionella pneumophila serogroups 1 to 6 (5, 1, 14-16), Legionella bozemanii (4, 8), Legionella dumoffii (4, 8), Legionella gormanii (8, 17), Legionella micdadei (11), and Legionella longbeachae serogroups 1 and 2 (2, 13). Although the number of conjugates required per test was reduced by using three polyvalent reagents that contain four conjugates each, the time and cost necessary to complete the identification are considerable and likely to be attempted only in reference laboratories or in laboratories of relatively large hospitals. Furthermore, the number of known Legionella species and serogroups has continued to increase. This in turn increases the diagnostic work load of even reference laboratories to an unacceptable level. For example, two new serogroups of L. pneumophila (1, 3) and three new species, Legionella jordanis (6), Legionella wadsworthii (9), and Legionella oakridgensis (18), were recently described, and many other Legionella-like organisms (LLOs) have been isolated that do not react with currently used DFA conjugates. An alternative, less time-consuming test is clearly needed. In previous studies (9, 2), we showed that a slide agglutination test (SAT) with hyperimmune rabbit serum could be used to determine the serogroup of stock strains of L. pneumophila, L. bozemanii, L. dumoffii, L. gormanii, and L. micdadei. The SAT was simple to perform, required neither conjugated antiserum nor microscope, and gave identical test results to those obtained with DFA. The main disadvantage of the test was that it could be used only with an isolated culture, which often was not obtained from clinical specimens. However, the commercial availability of Legionella media should help overcome this problem, especially if selective media are improved. Accordingly, we compared the results obtained from testing 64 blind-coded isolates of Legionella with the SAT and the DFA. MATERIALS AND METHODS Antisera for SAT. The Legionella species, serogroup, and strain designation of each antigen used to prepare antisera for the SAT and the homologous titer of each antiserum are listed in Table 1. Antisera were prepared with whole cell vaccines by the method of McKinney et al. (15) with slight modifications. For each antigen, the 2-day growth from five bufferedcharcoal-yeast extract (BYCE) agar plates (15 by 1 mm) was suspended in 3 ml of phosphate-buffered saline (PBS), ph 7.6, containing 1% Formalin. The suspension was allowed to sit overnight at room temperature to kill the bacteria. It was then filtered through sterile gauze pads to remove residual medium particles. The filtrate was centrifuged at 1, x g for 15 min at room temperature, and the cell pellet was suspended in PBS containing.5% Formalin. The cell 1113

2 1114 THACKER, WILKINSON, AND BENSON J. CLIN. MICROBIOL. TABLE 1. Legionella antisera used in the SAT Immunogen Serum Species Serogroup Strain titer L. pneumophila 1 Knoxville 1 1:32a L. pneumophila 2 Togus 1 1:8 L. pneumophila 3 Bloomington 2 1:32 L. pneumophila 4 Los Angeles 1 1:8 L. pneumophila 5 Cambridge 2 1:64 L. pneumophila 6 Houston 2 1:32 L. pneumophila 7 Chicago 8 1:64 L. pneumophila 8 Concord 3 1:64 L. bozemanii 1 WIGA 1:8 L. dumoffii 1 NY-23 1:64 L. gormanii 1 LS-13 1:8 L. micdadei 1 TATLOCK 1:8 L. longbeachae 1 Long Beach 4 1:8 L. Iongbeachae 2 Tucker 1 1:16 L. jordanis 1 BL-54 1:32 a The titer was expressed as the serum dilution that gave 4+ agglutination within 3 sec. suspension was adjusted to 4 IU of turbidity (12) which corresponded to approximately 4 x 19 bacterial cells per ml. Young adult New Zealand white rabbits were injected with 2 ml of equal volumes of cell suspension and incomplete Freund adjuvant (Sigma Chemical Co., St. Louis, Mo.) intracutaneously in approximately 2 sites along the shaved back of each rabbit. After 31 days, they were injected with 2 ml of equal volumes of cell suspension and incomplete Freund adjuvant in two deep-muscle injections in the hind quarters. After 7 days, they were injected intramuscularly with 2 ml of cell suspension without adjuvant. After 7 days, 5 ml of blood were taken from the ear artery, and 1 ml of cell suspension was injected intravenously. Again, after 7 days, 5 ml of blood was withdrawn from the ear artery, and 1 ml of cell suspension was injected intravenously. The rabbits were exsanguinated 7 days later. For the SAT, antiserum titers were determined with heated, formalinized antigens that were prepared with the stock strains for each homologous Legionella serogroup (Table 1; protocol described below). Twofold dilutions of antisera were made in PBS, ph 7.6. The serum titer was defined as the highest dilution which caused 4+ (strong) agglutination of homologous antigen within 3 sec. A working dilution, one twofold dilution lower than the serum titer, was prepared for each serogroup and tested against all of the Legionella antigens. If cross-reactions were observed, they were removed by absorption with the cross-reacting strain. The working dilutions of antisera were stored at 4 C where they were stable for at least 4 months. Absorption of antisera for SAT. Legionella cells for absorption were grown for 48 h at 35 C on BCYE agar plates; harvested in PBS, ph 7.6, containing 1% Formalin; and killed as described above. The cells were then washed twice in PBS by centrifugation. For each absorption, cells and undiluted antiserum were mixed at a 1:5 ratio (vol/vol); the mixture was incubated at 37 C for 2 h and then placed at 4 C overnight. The absorbed antiserum was separated from the cells by centrifugation. If subsequent tests showed that the cross-reaction was still present, the absorption procedure was repeated. SAT. Sixty-four isolates from patients with Legionella infections that were submitted to CDC for diagnostic testing were stored on BCYE slants at -65 C until tested with the SAT in the blind by one of us. The cultures were allowed to thaw, inoculated onto BCYE agar slants (13- by 1-mm tubes), and incubated for 2 days at 35 C. The bacterial growth from each slant was suspended in.8 ml of 1% (vol/vol) neutral Formalin. The antigens were then heated in a boiling water bath for 15 min to produce a smooth suspension of cells. To perform the SAT, 1 drop (approximately.25 ml) of antigen was mixed with 1 drop of antiserum on a glass microscope slide (5 by 75 mm) with an applicator stick. The slide was rocked for 3 sec, and the reactions were scored on a scale from 1' (barely visible) to 4+ (strong agglutination). DFA test. DFA conjugates for L. jordanis and L. pneumophila serogroups 7 and 8 were provided by William Bibb, CDC. Other DFA conjugates and positive control antigens were supplied by the Biological Products Program, CDC. The conjugates were absorbed with Legionella cells of heterologous species and serogroups to ensure that they were specific. The Legionella isolates were stained and examined by DFA as described by Cherry et al. (7). Briefly, each of the 64 cultures was suspended in a volume of PBS containing 1% Formalin that was sufficient to give a light turbidity reading that corresponded to a McFarland no. 2 standard. The suspension was placed in wells of a microscope slide (25- by 75-mm acetoneresistant glass slide with 12 staggered wells, 5 mm in diameter; Cel-Line Associates, Inc., Newfield, N.J.) and allowed to air dry. After gentle heat fixation, the smears were covered with conjugate and incubated in a moist chamber for 2 min at room temperature. The slides were washed in PBS for 5 min, rinsed with distilled water, allowed to air dry, and then mounted with buffered glycerol, ph 9., and a cover slip (Corning no. 1; Corning Glass Works, Corning, N.Y.). The slides were examined with a Leitz Dialux 2 fluorescence microscope (Leitz/Opti-Metric Div. of E. Leitz Inc., Rockleigh, N.J.) equipped with an HBO- 1 mercury incident light source, the Leitz I-cube filter system, 4x dry and 5x oil objectives, and 6.3x binoculars. RESULTS The results of testing heated, formalinized antigens of each Legionella serogroup with homologous and heterologous antisera prepared for use in the SAT are presented in Table 2. Most of the antisera were serogroup specific without absorption. Expected and observed were the reciprocal cross-reactions between L. pneumophila serogroups 3 and 6 (16) and L. bozemanii and L. jordanis (6), between L. longbeachae serogroup 2 antiserum and L. Iongbeachae serogroup 1 antigen (2), and between L. pneumophila serogroup 3 antiserum and serogroup 2 antigen (14). Also expected were crossreactions among L. pneumophila serogroups 4, 5, and 8 (3). The cross-reaction of L. Iongbeachae serogroup 2 antiserum with L. jordanis

3 VOL. 18, 1983 SLIDE AGGLUTINATION TEST FOR LEGIONELLA SP ; l O4 O O ~ o Oz 3,r tj z ' 'Ir 't: 't: 'tz O -o Zfz -, " - > - CA~~~~~~~~~~~~~~~~~~~~~~C, '-C O~~~~~~ I O C1=,co = -+ Q 8. ~~A~~( ~ ~ ~..# C Co ~~~~~~~~~~~~~C4 tiil +>IIlv c x ~I l l o ~~~~~~~~~~~~~ O' ItI vi I It c - + It o t' 1+I+ +I+ + MCo CA + I,~ I+ ; z II I. I+ :, O QQ~~~~~~~~~~~~~~~~- II1++I It- I1I+ II I+II I' I I11., o o~~~~~~~~~~l 2 GO A+vSII I1I11+1R 1+I R II +I1 I+ o~ ' GO~~~~~~~~~~~~

4 1116 THACKER, WILKINSON, AND BENSON TABLE 3. Slide agglutination reactions of Legionella antigens with polyvalent antiserum pools Legionella Sero- Antiserum pool antigen group A B C D L. pneumophila 1 4+a b _ L. pneumophila L. pneumophila L. pneumophila 4 4+J - L. pneumophila 5-4+ L. pneumophila L. pneumophila 7-4+ L. pneumophila L. bozemanii L. dumoffii L. gormanii L. micdadei L. Iongbeachae L. Iongbeachae L. jordanis a Positive reactions scored on a scale from 1+ (barely visible) to 4+ (strong agglutination). Blocks indicate homologous reactions. b -, Negative reaction. antigens has not been previously reported in DFA studies but was observed with the SAT. All cross-reactions were removed by absorption with the appropriate Legionella strain without removing homologous agglutination reactions. To determine whether polyvalent, pooled antisera could be used to conserve laboratory time and test reagents, four pools were prepared with unabsorbed antisera, each at a final dilution that had given optimal agglutination reactions before pooling. As shown in Table 3, strong homologous agglutination and only the previously noted cross-reactions occurred with each polyvalent pooled reagent. These pools were then used to test 64 clinical isolates in the blind. Isolates that agglutinated with the polyvalent antisera were tested with the constituent monovalent antisera within each pool that gave a positive test result. Each strain showing multiple agglutination reactions was retested with absorbed antisera to identify the serogroup to which it belonged. As shown in Table 4, there was 1% agreement between the DFA and SAT results. DISCUSSION Clinical laboratories routinely use the DFA for identifying the species and serogroup of Legionella strains either in patient specimens or in isolated cultures (7). Its two main advantages are the capability of directly staining Legionella species in tissue or respiratory secretions and the fact that the antiserum conjugates can be highly diluted and thus conserved. The main disadvantages are the requirements of expensive J. CLIN. MICROBIOL. equipment and specially trained laboratory personnel and the considerable time and expense required to test each specimen with at least 12 conjugates. The number of serogroups, and therefore reagents required to identify them, may continue to increase for an undetermined time, as there are LLOs being investigated that do not stain with existing conjugates and that may eventually be described as new Legionella species and serogroups. As shown in this study, the SAT could provide a less time-consuming test as a rapid screening test for organisms that are isolated in culture. Ir. contrast to the DFA, the SAT is rapid, simple, and requires no sophisticated equipment or training to perform. If the antisera are sufficiently potent to give titers of at least 1:8, polyvalent pools can be prepared and used as screening reagents. Monovalent antisera can then be used to identify the serogroup of the isolate. However, if therapy is the same for all Legionella species, identification only to the genus level with polyvalent antisera may be sufficient for routine diagnosis. Monovalent antisera could then be conserved for epidemiological studies in which identification of serogroup-specific antigens is useful. Antigenic diversity to the serogroup level has been reported for two species: L. pneumophila (1, 3, 1, 14, 16) and L. longbeachae (2, 13). Further antigenic diversity occurs within the L. pneumophila serogroups. Two major subgroups of L. pneumophila serogroup 1 have been reported in studies in which monoclonal antibodies were used (R. M. McKinney et al., personal communication). We detected the same antigenic patterns with the SAT. An antigen prepared with the Olda strain of L. pneumophila serogroup 1, which represents one of the monoclonal TABLE 4. SAT results of clinical isolates of Legionella spp. Species Serogroup No. of strainsa L. pneumophila 1 18 L. pneumophila 2 9 L. pneumophila 3 3 L. pneumophila 4 8 L. pneumophila 5 4 L. pneumophila 6 8 L. pneumophila 7 1 L. pneumophila 8 1 L. bozemanii 1 2 L. dumoffii 1 2 L. gormanii 1 1 L. micdadei 1 2 L. Iongbeachae 1 2 L. Iongbeachae 2 1 L. jordanis 1 2 a The 64 strains gave identical DFA results.

5 VOL. 18, 1983 SLIDE AGGLUTINATION TEST FOR LEGIONELLA SP subgroups, gave weak agglutination reactions with antisera prepared with the Knoxville 1 antigen which represents a second monoclonal subgroup. Therefore, Olda strain antisera should be included either with or instead of Knoxville 1 antiserum, as the Olda serum reacts strongly with all L. pneumophila serogroup 1 isolates. Additional antigenic diversity within serogroups was shown by the fact that some strains within a single serogroup showed cross-reactions with heterologous antisera, and others within the same serogroup did not. Notable examples were seen in L. pneumophila serogroups 4, 5, and 8. Such patterns of cross-reactivity are reminiscent of those that occur among Salmonella antigens. The lipopolysaccharide of Legionella species may also contain numerous antigenic determinants which contribute to various crossreactions with unabsorbed antisera. It is unlikely that flagellar antigens were also present in the SAT antigens, as the formalinized suspensions were heated in a boiling water bath to reduce nonspecific agglutination. In summary, the SAT, like DFA, can be used to identify the species and serogroups of primary Legionella isolates. The SAT is specific, simple, and inexpensive. Preliminary studies have shown that the time required to perform the SAT could be further reduced by preparing suspensions of suspected Legionella colonies directly from BCYE plates. Since incorporating the test into our laboratory, two Legionella species originally isolated from environmental specimens were rapidly and specifically identified in cultures from human specimens: L. jordanis from lung tissue and a new species from sputum that stained with a conjugate prepared with the LLO strain WO-44C. The former isolate gave a positive DFA test result with the L. bozemanii conjugate which was prepared before L. jordanis was recognized. The latter organism has been epidemiologically linked to Pontiac fever (manuscript in preparation). To our knowledge, this is the first evidence that L. jordanis and strains of the same species as LLO strain WO- 44C are etiological agents of human pneumonia. ACKNOWLEDGMENTS We thank William Bibb for providing some of the Legionella isolates and antisera used in this study. We also thank Joan Nagel for secretarial assistance. LITERATURE CITED 1. Bibb, W. F., P. M. Arnow, D. L. Dellinger, and S. R. Perryman Isolation and characterization of a seventh serogroup of Legionella pneumophila. J. Clin. Microbiol. 17: Bibb, W. F., R. J. Sorg, B. M. Thomason, M. D. Hicklin, A. G. Steigerwalt, D. J. Brenner, and M. R. Wulf Recognition of a second serogroup of Legionella longbeachae. J. Clin. Microbiol. 14: Bissit, M. L., J.. Lee, and D. S. Lindquist New serogroup of Legionella pneumophila, serogroup 8. J. Clin. Microbiol. 17: Brenner, D. J., A. G. Steigerwalt, G. W. Gorman, R. E. Weaver, J. C. Feeley, L. G. Cordes, H. W. Wilkinson, C. Patton, B. M. Thomason, and K. R. L. Sasseville Legionella bozemanii sp. nov. and Legionella dumoffii sp. nov.: classification of two additional species of Legionella associated with human pneumonia. Curr. Microbiol. 4: Brenner, D. J., A. G. Steigerwalt, and J. E. McDade Classification of the Legionnaires' disease bacterium: Legionella pneumophila, genus novum, species nova, of the family Legionellaceae, familia nova. Ann. Intern. Med. 9: Cherry, W. B., G. W. Gorman, L. H. Orrison, C. W. Moss, A. G. Steigerwalt, H. W. Wilkinson, S. E. Johnson, R. M. McKinney, and D. J. Brenner Legionella jordanis: a new species of Legionella isolated from water and sewage. J. Clin. Microbiol. 15: Cherry, W. B., B. Pittman, P. P. Harris, G. A. Hebert, B. M. Thomason, L. Thacker, and R. E. Weaver Detection of Legionnaires disease bacteria by direct immunofluorescent staining. J. Clin. Microbiol. 8: Cordes, L. G., H. W. Wilkinson, G. W. Gorman, B. J. Fikes, and D. W. Fraser Atypical Legionella-like organisms: fastidious water-associated bacteria pathogenic for man. Lancet ii: Edelstein, P. H., D. J. Brenner, C. W. Moss, A. G. Steigerwalt, E. M. Francis, and W. L. George Legionella wadsworthii species nova: a cause of human pneumonia. Ann. Intern. Med. 97: England, A. C., III, R. M. McKinney, P. Skaliy, and G. W. Gorman A fifth serogroup of Legionella pneumophila. Ann. Intern. Med. 93: Hebert, G. A., A. G. Steigerwalt, and D. J. Brenner Legionella micdadei species nova: classification of a third species of Legionella associated with human pneumonia. Curr. Microbiol. 3: Maaloe, The international reference preparation for opacity-notes and description. Bull. W.H.O. 12: McKinney, R. M., R. K. Porschen, P. H. Edelstein, M. L. Bissett, P. P. Harris, S. P. Bondell, A. G. Steigerwalt, R. E. Weaver, M. E. Ein, D. S. Lindquist, R. S. Kops, and D. J. Brenner Legionella longbeachae sp. nov., another etiologic agent of human pneumonia. Ann. Intern. Med. 94: McKinney, R. M., L. Thacker, P. P. Harris, K. R. Lewallen, G. A. Hebert, P. H. Edelstein, and B. M. Thomason Four serogroups of Legionnaires' disease bacteria defined by direct immunofluorescence. Ann. Intern. Med. 9: McKhnney, R. M., B. M. Thomason, P. P. Harris, L. Thacker, K. R. Lewallen, H. W. Wilkinson, G. A. Hebert, and C. W. Moss Recognition of a new serogroup of Legionnaires disease bacterium. J. Clin. Microbiol. 9: McKinney, R. M., H. W. Wilkinson, H. M. Sommers, B. J. Fikes, K. R. Sasseville, M. M. Yungbluth, and J. S. Wolf Legionella pneumophila serogroup six: isolation from cases of legionellosis, identification by immunofluorescence staining, and immunological response to infection. J. Clin. Microbiol. 12: Morris, G. K., A. Steigerwalt, J. C. Feeley, E. S. Wong, W. T. Martin, C. M. Patton, and D. J. Brenner Legionella gormanii sp. nov. J. Clin. Microbiol. 12: Orrison, L. H., W. B. Cherry, R. L. Tyndall, C. B. Fliermans, S. B. Gough, M. A. Lambert, L. K. McDougal, W. F. Bibb, and D. J. Brenner Legionella oakridgensis: unusual new species isolated from cooling tower water. Appl. Environ. Microbiol. 45: Wilkinson, H. W., and B. J. Fikes Slide agglutination test for serogrouping Legionella pneumophila and

6 1118 THACKER, WILKINSON, AND BENSON atypical Legionella-like organisms. J. Clin. Microbiol. 11: Wilkinson, H. W., and B. J. Fikes Detection of cellassociated or soluble antigens of Legionella pneumophila J. CLIN. MICROBIOL. serogroups 1 to 6, Legionella bozemanii, Legionella dumoffii, Legionella gormanii, and Legionella micdadei by staphylococcal coagglutination tests. J. Clin. Microbiol. 14: