Vol. 30, No. 9
JOURNAL OF CLINICAL MICROBIOLOGY, Sept. 1992, p. 2451-2455
0095-1137/92/092451-05$02.00/0 Copyright C 1992, American Society for Microbiology
Enzyme Immunoassay for Q Fever: Comparison with Complement Fixation and Immunofluorescence Tests and Dot Immunoblotting R. COWLEY,* F. FERNANDEZ, W. FREEMANTLE,
AND
D. RUTFTER
Bacterial Diagnosis Group, Division of Pathology, Public Health Laboratory Service, Centre for Applied Microbiology and Research, Porton Down, Salisbury, Wiltshire SP4 OJG, United Kingdom Received 7 February 1992/Accepted 18 June 1992
Enzyme-linked immunosorbent assays (ELISA) for the detection of specific immunoglobulin G (IgG) and IgM antibodies were developed by using purified CoxieUla burnetii cells. Variables, including type of microtiter plate, blocking agent, incubation conditions, antigen stability, and substrate type, were examined to achieve optimal ELISA performance. The reliabilities of the assay systems were compared with those of complement fixation (CF) and enhanced immunofluorescence (ElF) tests with 600 human serum samples from defined clinical cases of Q fever, routine samples, and serum specimens from farmers. ELISA and EIF test results agreed in all cases. Dot immunoblotting was also used to test some of these sera and gave a rapid, qualitative result, which agreed with ELISA and EIF test results in all cases. No instances were found in which both ELISA and EIF test results were negative and the CF test result was positive. However, approximately 5% of the sera were positive by ELISA and the EIF test while the CF test result was either negative or unreadable because of serum anticomplementary activity. We conclude that dot immunoblotting is a useful screening test, whereas ELISA and the EIF test are both rapid and sensitive tests when used for the serodiagnosis of Q fever and should be considered to be replacements for the CF test. suffers from the tendency for sera from patients with Q fever
Q fever is primarily a zoonosis, causing inapparent infection in domestic livestock worldwide (8). Human infection is due mainly to aerosol transmission of the causative organism Coxiella bumetii from infected animals. Exposure may lead to inapparent infection; if overt disease develops, it usually is similar to an atypical pneumonia, with chills, malaise, and severe headache. In approximately 2% of cases (17), the chronic form occurs. The chronic form may appear as endocarditis or hepatitis, requires long-term antibiotic therapy or surgery, and can be fatal (12). Human Q fever is associated mainly with rural populations, veterinary or abattoir workers, and with travellers (22) and has an apparently low incidence rate, with approximately 150 cases per year in England and Wales being reported to the Communicable Disease Surveillance Centre (21). This is likely to be an underestimation due in part to incorrect diagnosis, especially of the acute form, and also to unreliable laboratory serological methods. Diagnosis of Q fever in humans can be made via isolation of the causative organism from clinical specimens by using animal inoculation or tissue culture (23), but this is hazardous and demanding for most clinical laboratories. Traditionally the complement fixation (CF) test has been used to detect antibody response to this disease in the laboratory. Two distinct types of antibody are produced in a temporal manner; phase II antibodies appear early in infection, while phase I antibodies arise later and usually only to a relatively low level, except in the chronic form of the disease, in which they tend to become greatly elevated. However, in the majority of cases only a single convalescent-phase serum sample is available for testing and, as antibodies to C. burnetii can persist with titers decreasing at different rates, current Q fever may be difficult to diagnose. The CF test
*
to be anticomplementary. Also, the test is tedious and relatively slow, taking 2 days to complete, and usually relies on the accurate visual interpretation of erythrocyte hemoly-
sis (2). Other methods of laboratory serodiagnosis which have been described for Q fever include agglutination (11, 14) and immunofluorescence (1, 7, 13, 19) tests and the enzyme-linked immunosorbent assay (ELISA) (4, 15, 24, 27). The last two methods are reported to detect types or classes of antibody different from those detected by the CF test (10) and are able to determine the presence of specific immunoglobulins in test sera. In Q fever, immunoglobulin G (IgG) levels increase early in infection and can persist for many years, while the presence of specific IgM may indicate active or recent infection (6, 10). It has also been reported that the presence of specific humoral C. bumetii IgA in sera from patients with endocarditis is diagnostic (18). Therefore, laboratory detection of specific anti-C. burnetii immunoglobulins may be useful as a means of distinguishing the different forms of Q fever. We have developed ELISA and enhanced immunofluorescent test and dot immunoblotting techniques for the detection of specific Q fever antibodies and have compared these methods with the CF test by using 600 human serum samples from various sources.
MATERIALS AND METHODS C. burnetii. A standard egg-adapted phase II strain of C. burnetii was obtained from the Public Health Laboratory Service Division of Microbiological Reagents and Quality Control laboratory and grown in eggs to provide the antigen used for ELISA, the enhanced immunofluorescence (EIF) test, dot immunoblotting (DIB), and the CF test. Antigen preparation. The yolk sacs of 5-day-old chicken embryos were inoculated with live C. burnetii organisms,
Corresponding author. 2451
2452
COWLEY ET AL.
incubated at 35°C for up to 12 days, and then harvested. These were homogenized in 5 ml of sterile saline for 1 min. Formalin was added to a final concentration of 0.5%, and the suspension was incubated at 37°C for 5 days and then stored at 4°C until required. C. burnetii cells were then separated by a series of centrifugations. Briefly, the inactivated egg yolk sac suspension was centrifuged at 6,000 x g for 60 min. The deposit was resuspended to its original volume in phosphatebuffered saline (PBS) containing formalin at a 0.25% final concentration and homogenized for 2 min. These centrifugation and homogenization steps were repeated three times except that the suspension was finally centrifuged at 550 x g for 10 min. The supernatant fluid from this was removed and centrifuged at 6,000 x g for 60 min. The deposit was then resuspended in PBS, and 10-ml volumes were carefully layered onto 25 ml of sucrose-metrizoate (30% [wt/vol] sucrose, 7.6% [vol/vol] metrizoate [Schering Chemical Co.] in sterile distilled water) and centrifuged at 6,000 x g for 60 min. The deposits were washed four times by resuspending in PBS and further centrifuging at 6,000 x g for 60 min. The deposit was resuspended to a final concentration of 5% of the original yolk sac suspension. The final preparation was examined for purity by electron microscopy. Human sera. Human sera were kindly made available by N. Lightfoot, Public Health Laboratory, Newcastle, United Kingdom, T. Coleman, Public Health Laboratory, Hereford, United Kingdom, and P. Morgan-Capner, Public Health Laboratory, Preston, United Kingdom. A total of 430 serum samples were collected from farmers in Hereford and Somerset. Thirty-eight serum samples were from patients having clinically defined Q fever confirmed by the demonstration of an increasing C. burnetii phase II antibody titer by the CF test; 132 serum samples were selected from specimens sent to the Centre for Applied Microbiology and Research for Rickettsia and Coxiella testing, 44 of which were positive for various Rickettsia antibodies. Normal sera used as negative controls in ELISA, the EIF and CF tests, and DIB tests were from individuals with no clinical history suggestive of Q fever. Complement fixation. The CF test for antibodies specific for phase I and phase II C. burnetii antigens was the standard Public Health Laboratory Service method, using antigens and control sera (available from Public Health Laboratory Service DMRQC, Colindale). Serial doubling dilutions of sera were allowed to react overnight with phase I and phase II antigens together with guinea pig complement (Behring). Complement not fixed by antigen-antibody complexes was detected by the lysis of antibody-sensitized sheep erythrocytes. The end-point titer was taken at the dilution which showed 50% hemolysis relative to the control titration of the complement. A positive CF result was indicated by a titer of 1/10 or higher. Immunofluorescence. The EIF test was based on the standard fluorescent antibody methods (5, 18, 20) with the addition of biotinylated antibody conjugate (Vector Laboratories) and avidin-biotin labelled with a fluorochrome (Vector Laboratories) to enhance the fluorescence signal. In brief, purified whole-cell C. burnetii antigen was diluted to give approximately 103 to 104 fluorescing organisms per microscopic field (100 x oil immersion objective) and spotted onto 15-well microscope slides (Flow Laboratories). After being air dried, the slides were fixed with ice-cold acetone for 10 min and stored with desiccant at -20°C until required. Serial doubling dilutions of sera to be tested, ranging from 1/40 to 1/2,560, were prepared in PBS, pH 7.2. Ten microliters of each dilution was spotted onto the slide wells and
J. CLIN. MICROBIOL.
allowed to react at room temperature in a humid chamber for 30 min. After three washes in PBS, 0.01 ml of biotinylated goat anti-human IgG or IgM was added to each well and the slides were incubated for 30 min at room temperature. After being washed for 30 min in PBS, 0.01 ml of avidin-biotin fluoroisothiocyanate conjugate (Nordic Immunology) was spotted onto the slide wells and incubated as described previously. Following three washes in PBS and one in distilled water, slides were mounted in PBS-glycerol (Citifluor Ltd.) and examined at a magnification of x 1,000 by using a Nikon Diaphot-TMD microscope with a UV source. The antibody titer was evaluated qualitatively by reference to positive and negative controls on the same slide and was defined as the inverse of the highest dilution showing unequivocal fluorescent staining of the Coxiella cell membrane. A positive EIF result was indicated by a titer of 1/100 or higher. ELISA. An indirect ELISA method (9) was used. The optimal concentration of antigen was determined to be 4 jig per well by chessboard titration and also by reference to an estimation of the total protein (bicinchoninic acid; Pierce). The wells of flat-bottomed, sterile 96-well microtiter plates (ProBind; Falcon) were coated with 0.1 ml of antigen suspension at the optimal concentration in coating buffer (0.04 M sodium carbonate, 0.06 M sodium bicarbonate in sterile distilled water, pH 9.6). The optimal time and temperature of incubation during coating of the plates were found to be 4°C for at least 8 h. For convenience, plates were usually left overnight. Microtiter plates, coated with antigen, were either used immediately or else thoroughly air dried and then sealed in plastic film and stored at 4°C until required. For the assay, plates were washed three times (Wellwash 4; Denley) with 0.3-ml volumes of PBS (pH 7.2) containing 0.1% (vol/vol) Tween 20 (PBST) and 0.1-ml volumes of doubling dilutions of serum starting from 1/100 in PBST containing 5% heat-inactivated horse serum (PBST-HS) (SeraLab) were added to the wells. The plates were incubated at 37°C in humid conditions for 60 min. Each well was washed three times with PBST, and 0.1 ml of anti-human IgG or IgM (1/4,000 in PBST-HS) conjugated to horseradish peroxidase (HRP) (Nordic) was added to each well and incubated at 37°C in humid conditions for 60 min. Following a further three washes in PBST, 0.05 ml of tetramethylbenzidine (Sigma) substrate in sodium acetate buffer, pH 4.8, and 0.03% hydrogen peroxide were added to each well. The reaction was stopped after 15 min of incubation at 370C by the addition of 0.025 ml of 2 M sulfuric acid to each well. The optical density (OD) of each well was determined at 450 nm in a microELISA plate reader (Anthos 2001; Denley). Eight wells per plate served as serum-free controls, eight wells contained sera from known positive cases of Q fever and acted as positive controls, and eight wells contained pooled normal human sera which served as negative controls. The net absorbance (OD450) value was calculated for each well by subtracting the ODs of control blank wells. The end-point value for each serum titration was defined as the inverse of the highest dilution that resulted in an OD450 three times higher than the mean OD450 of the negative control serum at an equivalent dilution. A positive ELISA result was indicated by a titer of 1/100 or higher. The presence of specific IgM was assayed as described above, with the exception that each serum sample was pretreated with goat anti-human IgG antibody (MastSorb; Mast Laboratories) as directed by the manufacturer to remove any possible interference from rheumatoid factor (25, 26).
VOL. 30, 1992
DIB. The DIB method was based on an ELISA and used nitrocellulose (Hybond-C; Amersham) or nylon (Hybond-N; Amersham) as the solid phase support. The methods used for each support membrane were identical except where indicated. A membrane rectangle (10 by 12 cm) was placed onto a piece of filter paper (Whatman 3MM) cut to the same size. This was secured in a dot blot manifold (12 by 8 wells; Schleicher & Schuell) and maintained in this position until substrate addition. Nylon membranes were prewetted with 0.05 ml of PBST. Purified C. bumetii antigen (0.005 ml of a 1/400 dilution) was applied to each well in coating buffer, pH 9.6, under a vacuum. The membrane was allowed to air dry for 15 min prior to blocking with 0.05 ml of PBST-HS per well for 15 min. For the test, 0.005-ml doubling dilutions of each serum, starting from 1/100, were added. After incubation at room temperature for 15 min, membranes were washed by the addition of three separate 0.1-ml volumes of PBST per well. Anti-human IgG or IgM HRP conjugate (Nordic Immunology) (0.005 ml per well) at 1/4,000 in PBST-HS was then added, and, after incubation at room temperature for 15 min, the membranes were washed as described above. They were then removed from the manifold and immersed in DAB substrate (0.5 mg of 3,3'-diaminobenzidine tetrahydrochloride [Sigma] per ml in 50 mM Tris buffer [pH 7.2], 0.01% [vol/vol] hydrogen peroxide, 0.04% [vol/vol] nickel chloride solution). After incubation at room temperature for 15 min, the membrane was removed from the manifold and rinsed once in distilled water. A positive reaction was indicated by a titer of 1/100 or higher and was assessed by visual comparison with negative and positive control sera at the same dilution on the same membrane. ELISA and DIB conjugate titration. The optimum concentration of anti-human IgG or IgM conjugated to HRP was assayed by serial titration against human IgG or IgM, respectively. Briefly, the human immunoglobulin (10 ng per well) was bound to the wells of a sterile microtiter tray overnight at 4°C in coating buffer, pH 9.6. Plates were washed three times in PBST, and then 0.1 ml each of anti-human IgG or IgM conjugate dilution was added to the wells. Following incubation at 37°C for 60 min and five washes in PBST, bound immunoglobulin was detected by the addition of tetramethylbenzoate substrate as for the indirect ELISA. The reaction was stopped with 2 M sulfuric acid, and the plates were scanned at 450 nm. The optimum concentration of the conjugate was taken at the dilution which gave an OD of 1.0. Estimation of total protein. The total protein concentration of purified C. burnetii antigen was estimated by using a comparative microplate assay bicinchoninic acid; (Pierce). Bovine serum albumin (BSA) standards (0.05, 0.10, 0.15, 0.20, and 0.25 mg/ml) were prepared in PBS. C. burnetii antigen was diluted 1/10 to 1/640 by serial doubling dilution in a microtiter tray (flat bottomed; Falcon). Reagents were added according to the manufacturer's instructions. The tray was sealed and incubated at 60°C in a water bath for 30 min. The resulting color change was assessed by using a microplate reader at 562 nm, and the concentration of the antigen was calculated by reference to the BSA readings at the same OD.
RESULTS Assay optimization. Electron microscopy of the purified C. bumetii antigen preparation showed the presence of trace particulate matter among the mainly intact Coxiella cells.
Q FEVER ELISA
2453
TABLE 1. Comparison of ELISA and the EIF and CF tests for serodiagnosis of Q fever by using 600 human serum samples result' Assay resulta Assay
of ~~~~~~~serumNo.samples
ELISA .......................................... EIF -.
401
CF -.
ELISA+ ..........................................3 EIF +.
CFFELISA
..........................................
74
CF A/C . ELISA + ................ ......................... EIF +. CF +.
96
ELISA ..........................................26 EIF +. CF A/C .
ELISA CF +
..........................................
0
..............................................................
a -, negative; +, positive; A/C, result unreadable because of anticomplementary activity in serum.
The total protein content of the cell suspension was estimated to be between 35 and 40 mg/ml. The optimum antigen coating concentration for ELISA of 0.004 mg per well was determined by chessboard titration. The optimum concentration of antihuman antibody conjugated to HRP for ELISA were determined to be 1/4,000 for both IgG and IgM. The so-called "edge effect" in ELISA microtiter plate tests, in which differences may occur in binding of the sample to the outer wells (16) probably because of a thermal gradient across the plate, was not apparent in our ELISA. Plates were incubated singly without lids on small shelves in a fan-assisted incubator to improve circulation of air. Also, the ELISA antigen-coated plates were stable, showing no loss of antibody binding capability over an 8-month period. The substrate tetramethylbenzidine was found to give a higher positive/negative OD ratio than that obtained with o-phenylenediamine, with a lower background OD, and tetramethylbenzidine is also reported to be nonmutagenic (3). Nonspecific binding of undesirable proteins was reduced by the addition of 5% heat-inactivated horse serum in the presence of Tween 20 as a blocking step. Comparison of assays. In total, 600 human serum samples were compared by ELISA and the EIF and CF tests. The results of this comparison are shown in Table 1. Approximately 66% were negative by all assays, 12% were negative by ELISA and the EIF test where anticomplementary activity made the CF test result unreadable, 16% were positive by all three assays, and 4% were found to be positive by ELISA and the EIF test where anticomplementary activity made the CF test result unreadable. No cases were found in which the CF test result was positive and both ELISA and EIF test results were negative. ELISA and the EIF test gave 100% agreement. One hundred serum samples were also assayed by DIB; the results of this assay were evaluated visually relative to the appearance of both positive and negative control serum dots at the same dilution and are shown in
2454
J. CLIN. MICROBIOL.
COWLEY ET AL.
TABLE 2. Comparison of DIB with ELISA and the EIF and CF tests for serodiagnosis of Q fever by using 100 human serum samples Negative PositiveDIB Results
Assay result'a
No.samples of serum
ELISA + EIF + CF +
50
50
0
ELISA + EIF + CF A/C
10
10
0
ELISA + EIF + CF -
3
3
0
ELISAEIF CF A/C
7
0
7
ELISA EIF CF -
30
0
30
a +, positive; -, negative; A/C, result unreadable because of anticomplementary activity in serum.
comparison with the ELISA and the EIF and CF test results in Table 2. DIB gave 100% agreement with ELISA and EIF test results. For DIB, the best results were obtained with nylon membranes, which were also sturdier than the nitrocellulose. In all, 125 IgG positives were found. These either were known to be positive or were from specimens received at the Centre for Applied Microbiology and Research for Q fever serology. Only 4 serum samples from the 125 IgG positives were found to contain specific IgM by ELISA, the EIF test, and DIB. The CF test result was also positive for all four serum samples. These were all from known cases of acute Q fever. DISCUSSION Antibodies against C. bumetii are highly specific, and no cross-reactions with other microorganisms are known to occur. As antigen detection methods are currently specialized techniques that may require hazardous and relatively lengthy procedures, serological tests are more important in the diagnosis of Q fever than of other diseases. The results of this comparative study show that the CF test, which is widely used in diagnostic laboratories, fails to detect a small percentage of cases or else can be unreliable because of the presence of anticomplementary substances in the test serum. The inability of the CF test to produce a reliable and early indication of clinical disease has been noted by many other workers. This may be due to the recognition of different antibodies by different tests. For instance, Peter et al. (19) showed that primary Q fever could not be detected by the CF test during the onset of illness, whereas indirect immunofluorescence gave a positive result for approximately 50% of the sera in their study. This may well be the case with the three serum samples in this study which were positive by ELISA and the EIF test but were negative by the CF test. Also, the criteria for a positive CF test result include at least a fourfold increase in titer to phase II antigen between the
acute and convalescent stages of the disease (18), and, as this may take several weeks, rapid diagnosis of Q fever by this method is not possible. In comparison with the CF test, both ELISA and the EIF test proved to be more sensitive, rapid, and reliable, and neither test showed any significant cross-reactions with sera from known cases of illness caused by various rickettsiae. The presence of specific IgM can be taken as an indicator of acute Q fever by using phase II antigen. In our ELISA for IgM antibodies, sera were pretreated with a commercially available antihuman IgG antibody (MastSorb; Mast Laboratories) where necessary. This removed the majority of IgG from the test serum, preventing IgM autoantibodies (e.g., rheumatoid factor) from binding to their Fc region and thereby giving false positive results, especially in the chronic stages of Q fever. Removal of the IgG from the test serum in this way also prevents physical competition with the less avid IgM molecules for antigen binding sites. This is essential to avoid false-positive results in a system that does not discriminate in the final antibody-antigen reaction. The major conclusion from this work is that the immunofluorescence and enzyme-linked assays are equally effective in determining the presence of specific anti-C. bumnetii antibodies, and both are superior to the CF test. We therefore suggest that the fluorescent technique be adopted as a rapid and reliable means of screening relatively small numbers of sera. This should be especially easy for laboratories already using indirect immunofluorescence for measuring levels of antibody to other infectious agents such as legionellae and Treponema pallidum. ELISA for the detection of specific IgG and IgM using phase II antigen has also been shown to be a sensitive, rapid, and reliable means of screening large numbers of sera for epidemiological studies or as a reference method for the early and rapid detection of Q fever. ACKNOWLEDGMENT This work was supported by the Ministry of Defence, United Kingdom. REFERENCES 1. Ascher, M. S., G. S. Horwith, M. F. Thornton, J. R. Greenwood, and M. A. Berman. 1983. A rapid immunofluorescent procedure for serodiagnosis of Q fever in mice, guinea pigs, sheep and humans. Diagn. Immunol. 1:33-38. 2. Berge, T. O., and E. H. Lennette. 1953. Q-fever: a study of serological relationships among strains of Coxiella burnetii. Am. J. Hyg. 57:144-169. 3. Bos, E. S., A. A. van der Doelen, N. van Rooy, and A. H. W. M. Schuurs. 1981. 3,3',5,5'-Tetramethylbenzidine as an Ames test negative chromogen for horse-radish peroxidase in enzymeimmunoassay. J. Immunoassay 2:187-204. 4. Cracea, E., S. Constantinescu, A. Dumitrescu, M. Stefanescu, and G. Szegli. 1983. ELISA in the Q fever diagnosis. Arch. Roum. Pathol. Exp. Microbiol. 42:283-288. 5. Davis, G. E., and H. R. Cox. 1938. A filter-passing infectious agent isolated from ticks. I. Isolation from Dermacentor andersoni, reactions in animals and filtration experiments. Public Health Rep. 53:2259-2267. 6. Doller, G., P. C. Doller, and H.-J. Gerth. 1984. Early diagnosis of Q fever: detection of immunoglobulin M by radioimmunoassay and enzyme immunoassay. Eur. J. Clin. Microbiol. 3:550553. 7. Edlinger, E. 1985. Immunofluorescence serology: a tool for prognosis of Q-fever. Diagn. Micrcobiol. Infect. Dis. 3:343. 8. Elisberg, B. L., and F. M. Bozeman. 1979. The rickettsiae, p. 1061-1108. In E. H. Lennette and N. J. Schmidt (ed.), Diagnostic procedures for viral, rickettsial, and chlamydial infections,
Q FEVER ELISA
VOL. 30, 1992
5th ed. American Public Health Association, Washington, DC. 9. Engvall, E., and P. Perimann. 1972. Enzyme-linked immunosorbent assay (ELISA). III. Quantitation of specific antibodies by enzyme-linked anti-immunoglobulin in antigen coated tubes. J. Immunol. 109:129. 10. Field, P. R., J. G. Hunt, and A. M. Murphy. 1983. Detection and persistence of specific IgM antibody to Coxiella bumetii by enzyme linked immunosorbent assay: a comparison with immunofluorescence and complement fixation tests. J. Infect. Dis. 148:477-487. 11. Fiset, P., R. A. Ormsbee, R. Silberman, M. Peacock, and S. H. Spielman. 1969. A microagglutination technique for detection and measurement of rickettsial antibodies. Acta Virol. 13:60-66. 12. Fiset, P., and T. E. Woodward. 1982. Q-fever, p. 435-448. In S. A. Evans and H. A. Feldman (ed.), Bacterial infections of humans. Plenum Press, New York. 13. Hunt, J. G., P. R. Field, and A. M. Murphy. 1983. Immunoglobulin responses to Coxiella bumnetii (Q fever): single-serum diagnosis of acute infection, using an immunofluorescence technique. Infect. Immun. 39:977-981. 14. Kazar, J., R. Brezina, S. Schramek, A. Palanova, and B. Tvrda. 1981. Suitability of the microagglutination test for detection of post infection and post vaccination Q-fever antibodies in human sera. Acta Virol. 25:235-240. 15. Lang, G. H. 1988. Serosurvey of Coxiella burneti infection in dairy goat herds in Ontario. A comparison of two methods of enzyme-linked immunosorbent assay. Can. J. Vet. Res. 52:3741. 16. Oliver, D. G., A. H. Sanders, R. D. Hogg, and J. W. Hellman. 1981. Thermal gradients in microtitration plates. Effects on enzyme-linked immunoassay. J. Immunol. Methods 42:195. 17. Palmer, S. R., and S. E. J. Young. 1982. Q fever endocarditis in England and Wales, 1975-1981. Lancet ii:1148-1150. 18. Peacock, M. G., R. N. Philip, J. C. Williams, and R. S. Faulkner.
19.
20.
21.
22.
23.
24.
25. 26. 27.
2455
1983. Serological evaluation of Q fever in humans: enhanced phase I titers of immunoglobulins G and A are diagnostic for Q fever endocarditis. Infect. Immun. 41:1089-1098. Peter, O., G. Dupuis, W. Burgdorfer, and M. Peacock. 1985. Evaluation of the complement fixation and indirect immunofluorescence tests in the early diagnosis of primary Q fever. Eur. J. Clin. Microbiol. 4:394-396. Philip, R. N., E. A. Casper, R. A. Ormsbee, M. G. Peacock, and W. Burgdorfer. 1976. Microimmunofluorescence test for the serological study of Rocky Mountain spotted fever and typhus. J. Clin. Microbiol. 3:51-61. Public Health Laboratory Service Communicable Disease Report. 1991. Respiratory tract infections, England and Wales: laboratory reports, weeks 90/51-91/02. Public Health Laboratory Service Communicable Dis. Rep. 1:12. Raoult, D. 1991. Rickettsioses in the Mediterranean-a problem for travellers. Rev. Med. Microbiol. 2:115-120. Raoult, D., G. Vestris, and M. Enea. 1990. Isolation of 16 strains of Coxiella burnetii from patients by using a sensitive centrifugation cell culture system and establishment of the strains in HEL cells. J. Clin. Microbiol. 28:2482-2484. Roges, G., and E. Edlinger. 1986. Immunoenzymatic test for Q-fever. Diagn. Microbiol. Infect. Dis. 4:125-132. Vejtorp, M. 1980. The interference of IgM rheumatoid factor in enzyme-linked immunosorbent assays of rubella IgM and IgG antibodies. J. Virol. Methods 1:1-9. Vejtorp, M. 1980. Solid phase anti-IgM ELISA for detection of rubella specific IgM antibodies. Acta Pathol. Microbiol. Immunol. Scand. Sect. B 89:123-128. Williams, J. C., L. A. Thomas, and M. G. Peacock. 1986. Identification of phase-specific antigenic fractions of Coxiella burnetii by enzyme-linked immunosorbent assay. J. Clin. Microbiol. 24:929-934.
JOURNAL OF CLINICAL MICROBIOLOGY, Sept. 1992, p. 2451-2455
0095-1137/92/092451-05$02.00/0 Copyright C 1992, American Society for Microbiology
Enzyme Immunoassay for Q Fever: Comparison with Complement Fixation and Immunofluorescence Tests and Dot Immunoblotting R. COWLEY,* F. FERNANDEZ, W. FREEMANTLE,
AND
D. RUTFTER
Bacterial Diagnosis Group, Division of Pathology, Public Health Laboratory Service, Centre for Applied Microbiology and Research, Porton Down, Salisbury, Wiltshire SP4 OJG, United Kingdom Received 7 February 1992/Accepted 18 June 1992
Enzyme-linked immunosorbent assays (ELISA) for the detection of specific immunoglobulin G (IgG) and IgM antibodies were developed by using purified CoxieUla burnetii cells. Variables, including type of microtiter plate, blocking agent, incubation conditions, antigen stability, and substrate type, were examined to achieve optimal ELISA performance. The reliabilities of the assay systems were compared with those of complement fixation (CF) and enhanced immunofluorescence (ElF) tests with 600 human serum samples from defined clinical cases of Q fever, routine samples, and serum specimens from farmers. ELISA and EIF test results agreed in all cases. Dot immunoblotting was also used to test some of these sera and gave a rapid, qualitative result, which agreed with ELISA and EIF test results in all cases. No instances were found in which both ELISA and EIF test results were negative and the CF test result was positive. However, approximately 5% of the sera were positive by ELISA and the EIF test while the CF test result was either negative or unreadable because of serum anticomplementary activity. We conclude that dot immunoblotting is a useful screening test, whereas ELISA and the EIF test are both rapid and sensitive tests when used for the serodiagnosis of Q fever and should be considered to be replacements for the CF test. suffers from the tendency for sera from patients with Q fever
Q fever is primarily a zoonosis, causing inapparent infection in domestic livestock worldwide (8). Human infection is due mainly to aerosol transmission of the causative organism Coxiella bumetii from infected animals. Exposure may lead to inapparent infection; if overt disease develops, it usually is similar to an atypical pneumonia, with chills, malaise, and severe headache. In approximately 2% of cases (17), the chronic form occurs. The chronic form may appear as endocarditis or hepatitis, requires long-term antibiotic therapy or surgery, and can be fatal (12). Human Q fever is associated mainly with rural populations, veterinary or abattoir workers, and with travellers (22) and has an apparently low incidence rate, with approximately 150 cases per year in England and Wales being reported to the Communicable Disease Surveillance Centre (21). This is likely to be an underestimation due in part to incorrect diagnosis, especially of the acute form, and also to unreliable laboratory serological methods. Diagnosis of Q fever in humans can be made via isolation of the causative organism from clinical specimens by using animal inoculation or tissue culture (23), but this is hazardous and demanding for most clinical laboratories. Traditionally the complement fixation (CF) test has been used to detect antibody response to this disease in the laboratory. Two distinct types of antibody are produced in a temporal manner; phase II antibodies appear early in infection, while phase I antibodies arise later and usually only to a relatively low level, except in the chronic form of the disease, in which they tend to become greatly elevated. However, in the majority of cases only a single convalescent-phase serum sample is available for testing and, as antibodies to C. burnetii can persist with titers decreasing at different rates, current Q fever may be difficult to diagnose. The CF test
*
to be anticomplementary. Also, the test is tedious and relatively slow, taking 2 days to complete, and usually relies on the accurate visual interpretation of erythrocyte hemoly-
sis (2). Other methods of laboratory serodiagnosis which have been described for Q fever include agglutination (11, 14) and immunofluorescence (1, 7, 13, 19) tests and the enzyme-linked immunosorbent assay (ELISA) (4, 15, 24, 27). The last two methods are reported to detect types or classes of antibody different from those detected by the CF test (10) and are able to determine the presence of specific immunoglobulins in test sera. In Q fever, immunoglobulin G (IgG) levels increase early in infection and can persist for many years, while the presence of specific IgM may indicate active or recent infection (6, 10). It has also been reported that the presence of specific humoral C. bumetii IgA in sera from patients with endocarditis is diagnostic (18). Therefore, laboratory detection of specific anti-C. burnetii immunoglobulins may be useful as a means of distinguishing the different forms of Q fever. We have developed ELISA and enhanced immunofluorescent test and dot immunoblotting techniques for the detection of specific Q fever antibodies and have compared these methods with the CF test by using 600 human serum samples from various sources.
MATERIALS AND METHODS C. burnetii. A standard egg-adapted phase II strain of C. burnetii was obtained from the Public Health Laboratory Service Division of Microbiological Reagents and Quality Control laboratory and grown in eggs to provide the antigen used for ELISA, the enhanced immunofluorescence (EIF) test, dot immunoblotting (DIB), and the CF test. Antigen preparation. The yolk sacs of 5-day-old chicken embryos were inoculated with live C. burnetii organisms,
Corresponding author. 2451
2452
COWLEY ET AL.
incubated at 35°C for up to 12 days, and then harvested. These were homogenized in 5 ml of sterile saline for 1 min. Formalin was added to a final concentration of 0.5%, and the suspension was incubated at 37°C for 5 days and then stored at 4°C until required. C. burnetii cells were then separated by a series of centrifugations. Briefly, the inactivated egg yolk sac suspension was centrifuged at 6,000 x g for 60 min. The deposit was resuspended to its original volume in phosphatebuffered saline (PBS) containing formalin at a 0.25% final concentration and homogenized for 2 min. These centrifugation and homogenization steps were repeated three times except that the suspension was finally centrifuged at 550 x g for 10 min. The supernatant fluid from this was removed and centrifuged at 6,000 x g for 60 min. The deposit was then resuspended in PBS, and 10-ml volumes were carefully layered onto 25 ml of sucrose-metrizoate (30% [wt/vol] sucrose, 7.6% [vol/vol] metrizoate [Schering Chemical Co.] in sterile distilled water) and centrifuged at 6,000 x g for 60 min. The deposits were washed four times by resuspending in PBS and further centrifuging at 6,000 x g for 60 min. The deposit was resuspended to a final concentration of 5% of the original yolk sac suspension. The final preparation was examined for purity by electron microscopy. Human sera. Human sera were kindly made available by N. Lightfoot, Public Health Laboratory, Newcastle, United Kingdom, T. Coleman, Public Health Laboratory, Hereford, United Kingdom, and P. Morgan-Capner, Public Health Laboratory, Preston, United Kingdom. A total of 430 serum samples were collected from farmers in Hereford and Somerset. Thirty-eight serum samples were from patients having clinically defined Q fever confirmed by the demonstration of an increasing C. burnetii phase II antibody titer by the CF test; 132 serum samples were selected from specimens sent to the Centre for Applied Microbiology and Research for Rickettsia and Coxiella testing, 44 of which were positive for various Rickettsia antibodies. Normal sera used as negative controls in ELISA, the EIF and CF tests, and DIB tests were from individuals with no clinical history suggestive of Q fever. Complement fixation. The CF test for antibodies specific for phase I and phase II C. burnetii antigens was the standard Public Health Laboratory Service method, using antigens and control sera (available from Public Health Laboratory Service DMRQC, Colindale). Serial doubling dilutions of sera were allowed to react overnight with phase I and phase II antigens together with guinea pig complement (Behring). Complement not fixed by antigen-antibody complexes was detected by the lysis of antibody-sensitized sheep erythrocytes. The end-point titer was taken at the dilution which showed 50% hemolysis relative to the control titration of the complement. A positive CF result was indicated by a titer of 1/10 or higher. Immunofluorescence. The EIF test was based on the standard fluorescent antibody methods (5, 18, 20) with the addition of biotinylated antibody conjugate (Vector Laboratories) and avidin-biotin labelled with a fluorochrome (Vector Laboratories) to enhance the fluorescence signal. In brief, purified whole-cell C. burnetii antigen was diluted to give approximately 103 to 104 fluorescing organisms per microscopic field (100 x oil immersion objective) and spotted onto 15-well microscope slides (Flow Laboratories). After being air dried, the slides were fixed with ice-cold acetone for 10 min and stored with desiccant at -20°C until required. Serial doubling dilutions of sera to be tested, ranging from 1/40 to 1/2,560, were prepared in PBS, pH 7.2. Ten microliters of each dilution was spotted onto the slide wells and
J. CLIN. MICROBIOL.
allowed to react at room temperature in a humid chamber for 30 min. After three washes in PBS, 0.01 ml of biotinylated goat anti-human IgG or IgM was added to each well and the slides were incubated for 30 min at room temperature. After being washed for 30 min in PBS, 0.01 ml of avidin-biotin fluoroisothiocyanate conjugate (Nordic Immunology) was spotted onto the slide wells and incubated as described previously. Following three washes in PBS and one in distilled water, slides were mounted in PBS-glycerol (Citifluor Ltd.) and examined at a magnification of x 1,000 by using a Nikon Diaphot-TMD microscope with a UV source. The antibody titer was evaluated qualitatively by reference to positive and negative controls on the same slide and was defined as the inverse of the highest dilution showing unequivocal fluorescent staining of the Coxiella cell membrane. A positive EIF result was indicated by a titer of 1/100 or higher. ELISA. An indirect ELISA method (9) was used. The optimal concentration of antigen was determined to be 4 jig per well by chessboard titration and also by reference to an estimation of the total protein (bicinchoninic acid; Pierce). The wells of flat-bottomed, sterile 96-well microtiter plates (ProBind; Falcon) were coated with 0.1 ml of antigen suspension at the optimal concentration in coating buffer (0.04 M sodium carbonate, 0.06 M sodium bicarbonate in sterile distilled water, pH 9.6). The optimal time and temperature of incubation during coating of the plates were found to be 4°C for at least 8 h. For convenience, plates were usually left overnight. Microtiter plates, coated with antigen, were either used immediately or else thoroughly air dried and then sealed in plastic film and stored at 4°C until required. For the assay, plates were washed three times (Wellwash 4; Denley) with 0.3-ml volumes of PBS (pH 7.2) containing 0.1% (vol/vol) Tween 20 (PBST) and 0.1-ml volumes of doubling dilutions of serum starting from 1/100 in PBST containing 5% heat-inactivated horse serum (PBST-HS) (SeraLab) were added to the wells. The plates were incubated at 37°C in humid conditions for 60 min. Each well was washed three times with PBST, and 0.1 ml of anti-human IgG or IgM (1/4,000 in PBST-HS) conjugated to horseradish peroxidase (HRP) (Nordic) was added to each well and incubated at 37°C in humid conditions for 60 min. Following a further three washes in PBST, 0.05 ml of tetramethylbenzidine (Sigma) substrate in sodium acetate buffer, pH 4.8, and 0.03% hydrogen peroxide were added to each well. The reaction was stopped after 15 min of incubation at 370C by the addition of 0.025 ml of 2 M sulfuric acid to each well. The optical density (OD) of each well was determined at 450 nm in a microELISA plate reader (Anthos 2001; Denley). Eight wells per plate served as serum-free controls, eight wells contained sera from known positive cases of Q fever and acted as positive controls, and eight wells contained pooled normal human sera which served as negative controls. The net absorbance (OD450) value was calculated for each well by subtracting the ODs of control blank wells. The end-point value for each serum titration was defined as the inverse of the highest dilution that resulted in an OD450 three times higher than the mean OD450 of the negative control serum at an equivalent dilution. A positive ELISA result was indicated by a titer of 1/100 or higher. The presence of specific IgM was assayed as described above, with the exception that each serum sample was pretreated with goat anti-human IgG antibody (MastSorb; Mast Laboratories) as directed by the manufacturer to remove any possible interference from rheumatoid factor (25, 26).
VOL. 30, 1992
DIB. The DIB method was based on an ELISA and used nitrocellulose (Hybond-C; Amersham) or nylon (Hybond-N; Amersham) as the solid phase support. The methods used for each support membrane were identical except where indicated. A membrane rectangle (10 by 12 cm) was placed onto a piece of filter paper (Whatman 3MM) cut to the same size. This was secured in a dot blot manifold (12 by 8 wells; Schleicher & Schuell) and maintained in this position until substrate addition. Nylon membranes were prewetted with 0.05 ml of PBST. Purified C. bumetii antigen (0.005 ml of a 1/400 dilution) was applied to each well in coating buffer, pH 9.6, under a vacuum. The membrane was allowed to air dry for 15 min prior to blocking with 0.05 ml of PBST-HS per well for 15 min. For the test, 0.005-ml doubling dilutions of each serum, starting from 1/100, were added. After incubation at room temperature for 15 min, membranes were washed by the addition of three separate 0.1-ml volumes of PBST per well. Anti-human IgG or IgM HRP conjugate (Nordic Immunology) (0.005 ml per well) at 1/4,000 in PBST-HS was then added, and, after incubation at room temperature for 15 min, the membranes were washed as described above. They were then removed from the manifold and immersed in DAB substrate (0.5 mg of 3,3'-diaminobenzidine tetrahydrochloride [Sigma] per ml in 50 mM Tris buffer [pH 7.2], 0.01% [vol/vol] hydrogen peroxide, 0.04% [vol/vol] nickel chloride solution). After incubation at room temperature for 15 min, the membrane was removed from the manifold and rinsed once in distilled water. A positive reaction was indicated by a titer of 1/100 or higher and was assessed by visual comparison with negative and positive control sera at the same dilution on the same membrane. ELISA and DIB conjugate titration. The optimum concentration of anti-human IgG or IgM conjugated to HRP was assayed by serial titration against human IgG or IgM, respectively. Briefly, the human immunoglobulin (10 ng per well) was bound to the wells of a sterile microtiter tray overnight at 4°C in coating buffer, pH 9.6. Plates were washed three times in PBST, and then 0.1 ml each of anti-human IgG or IgM conjugate dilution was added to the wells. Following incubation at 37°C for 60 min and five washes in PBST, bound immunoglobulin was detected by the addition of tetramethylbenzoate substrate as for the indirect ELISA. The reaction was stopped with 2 M sulfuric acid, and the plates were scanned at 450 nm. The optimum concentration of the conjugate was taken at the dilution which gave an OD of 1.0. Estimation of total protein. The total protein concentration of purified C. burnetii antigen was estimated by using a comparative microplate assay bicinchoninic acid; (Pierce). Bovine serum albumin (BSA) standards (0.05, 0.10, 0.15, 0.20, and 0.25 mg/ml) were prepared in PBS. C. burnetii antigen was diluted 1/10 to 1/640 by serial doubling dilution in a microtiter tray (flat bottomed; Falcon). Reagents were added according to the manufacturer's instructions. The tray was sealed and incubated at 60°C in a water bath for 30 min. The resulting color change was assessed by using a microplate reader at 562 nm, and the concentration of the antigen was calculated by reference to the BSA readings at the same OD.
RESULTS Assay optimization. Electron microscopy of the purified C. bumetii antigen preparation showed the presence of trace particulate matter among the mainly intact Coxiella cells.
Q FEVER ELISA
2453
TABLE 1. Comparison of ELISA and the EIF and CF tests for serodiagnosis of Q fever by using 600 human serum samples result' Assay resulta Assay
of ~~~~~~~serumNo.samples
ELISA .......................................... EIF -.
401
CF -.
ELISA+ ..........................................3 EIF +.
CFFELISA
..........................................
74
CF A/C . ELISA + ................ ......................... EIF +. CF +.
96
ELISA ..........................................26 EIF +. CF A/C .
ELISA CF +
..........................................
0
..............................................................
a -, negative; +, positive; A/C, result unreadable because of anticomplementary activity in serum.
The total protein content of the cell suspension was estimated to be between 35 and 40 mg/ml. The optimum antigen coating concentration for ELISA of 0.004 mg per well was determined by chessboard titration. The optimum concentration of antihuman antibody conjugated to HRP for ELISA were determined to be 1/4,000 for both IgG and IgM. The so-called "edge effect" in ELISA microtiter plate tests, in which differences may occur in binding of the sample to the outer wells (16) probably because of a thermal gradient across the plate, was not apparent in our ELISA. Plates were incubated singly without lids on small shelves in a fan-assisted incubator to improve circulation of air. Also, the ELISA antigen-coated plates were stable, showing no loss of antibody binding capability over an 8-month period. The substrate tetramethylbenzidine was found to give a higher positive/negative OD ratio than that obtained with o-phenylenediamine, with a lower background OD, and tetramethylbenzidine is also reported to be nonmutagenic (3). Nonspecific binding of undesirable proteins was reduced by the addition of 5% heat-inactivated horse serum in the presence of Tween 20 as a blocking step. Comparison of assays. In total, 600 human serum samples were compared by ELISA and the EIF and CF tests. The results of this comparison are shown in Table 1. Approximately 66% were negative by all assays, 12% were negative by ELISA and the EIF test where anticomplementary activity made the CF test result unreadable, 16% were positive by all three assays, and 4% were found to be positive by ELISA and the EIF test where anticomplementary activity made the CF test result unreadable. No cases were found in which the CF test result was positive and both ELISA and EIF test results were negative. ELISA and the EIF test gave 100% agreement. One hundred serum samples were also assayed by DIB; the results of this assay were evaluated visually relative to the appearance of both positive and negative control serum dots at the same dilution and are shown in
2454
J. CLIN. MICROBIOL.
COWLEY ET AL.
TABLE 2. Comparison of DIB with ELISA and the EIF and CF tests for serodiagnosis of Q fever by using 100 human serum samples Negative PositiveDIB Results
Assay result'a
No.samples of serum
ELISA + EIF + CF +
50
50
0
ELISA + EIF + CF A/C
10
10
0
ELISA + EIF + CF -
3
3
0
ELISAEIF CF A/C
7
0
7
ELISA EIF CF -
30
0
30
a +, positive; -, negative; A/C, result unreadable because of anticomplementary activity in serum.
comparison with the ELISA and the EIF and CF test results in Table 2. DIB gave 100% agreement with ELISA and EIF test results. For DIB, the best results were obtained with nylon membranes, which were also sturdier than the nitrocellulose. In all, 125 IgG positives were found. These either were known to be positive or were from specimens received at the Centre for Applied Microbiology and Research for Q fever serology. Only 4 serum samples from the 125 IgG positives were found to contain specific IgM by ELISA, the EIF test, and DIB. The CF test result was also positive for all four serum samples. These were all from known cases of acute Q fever. DISCUSSION Antibodies against C. bumetii are highly specific, and no cross-reactions with other microorganisms are known to occur. As antigen detection methods are currently specialized techniques that may require hazardous and relatively lengthy procedures, serological tests are more important in the diagnosis of Q fever than of other diseases. The results of this comparative study show that the CF test, which is widely used in diagnostic laboratories, fails to detect a small percentage of cases or else can be unreliable because of the presence of anticomplementary substances in the test serum. The inability of the CF test to produce a reliable and early indication of clinical disease has been noted by many other workers. This may be due to the recognition of different antibodies by different tests. For instance, Peter et al. (19) showed that primary Q fever could not be detected by the CF test during the onset of illness, whereas indirect immunofluorescence gave a positive result for approximately 50% of the sera in their study. This may well be the case with the three serum samples in this study which were positive by ELISA and the EIF test but were negative by the CF test. Also, the criteria for a positive CF test result include at least a fourfold increase in titer to phase II antigen between the
acute and convalescent stages of the disease (18), and, as this may take several weeks, rapid diagnosis of Q fever by this method is not possible. In comparison with the CF test, both ELISA and the EIF test proved to be more sensitive, rapid, and reliable, and neither test showed any significant cross-reactions with sera from known cases of illness caused by various rickettsiae. The presence of specific IgM can be taken as an indicator of acute Q fever by using phase II antigen. In our ELISA for IgM antibodies, sera were pretreated with a commercially available antihuman IgG antibody (MastSorb; Mast Laboratories) where necessary. This removed the majority of IgG from the test serum, preventing IgM autoantibodies (e.g., rheumatoid factor) from binding to their Fc region and thereby giving false positive results, especially in the chronic stages of Q fever. Removal of the IgG from the test serum in this way also prevents physical competition with the less avid IgM molecules for antigen binding sites. This is essential to avoid false-positive results in a system that does not discriminate in the final antibody-antigen reaction. The major conclusion from this work is that the immunofluorescence and enzyme-linked assays are equally effective in determining the presence of specific anti-C. bumnetii antibodies, and both are superior to the CF test. We therefore suggest that the fluorescent technique be adopted as a rapid and reliable means of screening relatively small numbers of sera. This should be especially easy for laboratories already using indirect immunofluorescence for measuring levels of antibody to other infectious agents such as legionellae and Treponema pallidum. ELISA for the detection of specific IgG and IgM using phase II antigen has also been shown to be a sensitive, rapid, and reliable means of screening large numbers of sera for epidemiological studies or as a reference method for the early and rapid detection of Q fever. ACKNOWLEDGMENT This work was supported by the Ministry of Defence, United Kingdom. REFERENCES 1. Ascher, M. S., G. S. Horwith, M. F. Thornton, J. R. Greenwood, and M. A. Berman. 1983. A rapid immunofluorescent procedure for serodiagnosis of Q fever in mice, guinea pigs, sheep and humans. Diagn. Immunol. 1:33-38. 2. Berge, T. O., and E. H. Lennette. 1953. Q-fever: a study of serological relationships among strains of Coxiella burnetii. Am. J. Hyg. 57:144-169. 3. Bos, E. S., A. A. van der Doelen, N. van Rooy, and A. H. W. M. Schuurs. 1981. 3,3',5,5'-Tetramethylbenzidine as an Ames test negative chromogen for horse-radish peroxidase in enzymeimmunoassay. J. Immunoassay 2:187-204. 4. Cracea, E., S. Constantinescu, A. Dumitrescu, M. Stefanescu, and G. Szegli. 1983. ELISA in the Q fever diagnosis. Arch. Roum. Pathol. Exp. Microbiol. 42:283-288. 5. Davis, G. E., and H. R. Cox. 1938. A filter-passing infectious agent isolated from ticks. I. Isolation from Dermacentor andersoni, reactions in animals and filtration experiments. Public Health Rep. 53:2259-2267. 6. Doller, G., P. C. Doller, and H.-J. Gerth. 1984. Early diagnosis of Q fever: detection of immunoglobulin M by radioimmunoassay and enzyme immunoassay. Eur. J. Clin. Microbiol. 3:550553. 7. Edlinger, E. 1985. Immunofluorescence serology: a tool for prognosis of Q-fever. Diagn. Micrcobiol. Infect. Dis. 3:343. 8. Elisberg, B. L., and F. M. Bozeman. 1979. The rickettsiae, p. 1061-1108. In E. H. Lennette and N. J. Schmidt (ed.), Diagnostic procedures for viral, rickettsial, and chlamydial infections,
Q FEVER ELISA
VOL. 30, 1992
5th ed. American Public Health Association, Washington, DC. 9. Engvall, E., and P. Perimann. 1972. Enzyme-linked immunosorbent assay (ELISA). III. Quantitation of specific antibodies by enzyme-linked anti-immunoglobulin in antigen coated tubes. J. Immunol. 109:129. 10. Field, P. R., J. G. Hunt, and A. M. Murphy. 1983. Detection and persistence of specific IgM antibody to Coxiella bumetii by enzyme linked immunosorbent assay: a comparison with immunofluorescence and complement fixation tests. J. Infect. Dis. 148:477-487. 11. Fiset, P., R. A. Ormsbee, R. Silberman, M. Peacock, and S. H. Spielman. 1969. A microagglutination technique for detection and measurement of rickettsial antibodies. Acta Virol. 13:60-66. 12. Fiset, P., and T. E. Woodward. 1982. Q-fever, p. 435-448. In S. A. Evans and H. A. Feldman (ed.), Bacterial infections of humans. Plenum Press, New York. 13. Hunt, J. G., P. R. Field, and A. M. Murphy. 1983. Immunoglobulin responses to Coxiella bumnetii (Q fever): single-serum diagnosis of acute infection, using an immunofluorescence technique. Infect. Immun. 39:977-981. 14. Kazar, J., R. Brezina, S. Schramek, A. Palanova, and B. Tvrda. 1981. Suitability of the microagglutination test for detection of post infection and post vaccination Q-fever antibodies in human sera. Acta Virol. 25:235-240. 15. Lang, G. H. 1988. Serosurvey of Coxiella burneti infection in dairy goat herds in Ontario. A comparison of two methods of enzyme-linked immunosorbent assay. Can. J. Vet. Res. 52:3741. 16. Oliver, D. G., A. H. Sanders, R. D. Hogg, and J. W. Hellman. 1981. Thermal gradients in microtitration plates. Effects on enzyme-linked immunoassay. J. Immunol. Methods 42:195. 17. Palmer, S. R., and S. E. J. Young. 1982. Q fever endocarditis in England and Wales, 1975-1981. Lancet ii:1148-1150. 18. Peacock, M. G., R. N. Philip, J. C. Williams, and R. S. Faulkner.
19.
20.
21.
22.
23.
24.
25. 26. 27.
2455
1983. Serological evaluation of Q fever in humans: enhanced phase I titers of immunoglobulins G and A are diagnostic for Q fever endocarditis. Infect. Immun. 41:1089-1098. Peter, O., G. Dupuis, W. Burgdorfer, and M. Peacock. 1985. Evaluation of the complement fixation and indirect immunofluorescence tests in the early diagnosis of primary Q fever. Eur. J. Clin. Microbiol. 4:394-396. Philip, R. N., E. A. Casper, R. A. Ormsbee, M. G. Peacock, and W. Burgdorfer. 1976. Microimmunofluorescence test for the serological study of Rocky Mountain spotted fever and typhus. J. Clin. Microbiol. 3:51-61. Public Health Laboratory Service Communicable Disease Report. 1991. Respiratory tract infections, England and Wales: laboratory reports, weeks 90/51-91/02. Public Health Laboratory Service Communicable Dis. Rep. 1:12. Raoult, D. 1991. Rickettsioses in the Mediterranean-a problem for travellers. Rev. Med. Microbiol. 2:115-120. Raoult, D., G. Vestris, and M. Enea. 1990. Isolation of 16 strains of Coxiella burnetii from patients by using a sensitive centrifugation cell culture system and establishment of the strains in HEL cells. J. Clin. Microbiol. 28:2482-2484. Roges, G., and E. Edlinger. 1986. Immunoenzymatic test for Q-fever. Diagn. Microbiol. Infect. Dis. 4:125-132. Vejtorp, M. 1980. The interference of IgM rheumatoid factor in enzyme-linked immunosorbent assays of rubella IgM and IgG antibodies. J. Virol. Methods 1:1-9. Vejtorp, M. 1980. Solid phase anti-IgM ELISA for detection of rubella specific IgM antibodies. Acta Pathol. Microbiol. Immunol. Scand. Sect. B 89:123-128. Williams, J. C., L. A. Thomas, and M. G. Peacock. 1986. Identification of phase-specific antigenic fractions of Coxiella burnetii by enzyme-linked immunosorbent assay. J. Clin. Microbiol. 24:929-934.
