Mediterranean Spotted Fever

Transcription

1 JOURNAL OF CLINICAL MICROBIOLOGY, Sept. 1989, p /89/ $02.00/0 Copyright (O 1989, American Society for Microbiology Vol. 27, No. 9 Line Blot and Western Blot Immunoassays for Diagnosis of Mediterranean Spotted Fever DIDIER RAOULTt AND GREGORY A. DASCH* Rickettsial Diseases Division, Infectious Diseases Department, Naval Medical Research Institute, Bethesda, Maryland Received 12 January 1989/Accepted 11 May 1989 The line blot, a new immunoassay in which antigens are placed on nitrocellulose as narrow lines, was evaluated for its sensitivity and specificity rela to the microimmunofluorescence assay for the diagnosis of Mediterranean spotted fever (MSF). The line blot assay was only slightly less sensi and less specific than the microimmunofluorescence assay for detection of immunoglobulin M (IgM) or IgG in 100 serum specimens from 42 patients with MSF. No line blot reactions were observed among 50 control serum specimens from febrile patients with other illnesses. The line blot assay was largely group reac for spotted fever rickettsiae, but 26% of the posi serum specimens also cross-reacted by IgM with Rickettsia typhi. Western immunoblotting was used to characterize the antigenic components recognized by 19 MSF serum specimens. For both IgM and IgG, lipopolysaccharide was the cross-reac group antigen, whereas the high-molecular-weight speciesspecific protein antigens (SPAs) were the only reac proteins. Rela to the other nine rickettsiae, Rickettsia bellii was unique both in exhibiting no SPA reactions and in having a lipopolysaccharide with a predominantly high-molecular-weight distribution. Although most of the 19 MSF serum specimens examined by Western blotting exhibited preferential reactivity to SPAs of two strains of R. conorii and weaker reactions to the other rickettsiae, 2 serum specimens exhibited SPA reactions consistent with typhus infections. In comparison with other assays, the line blot and Western blot immunoassays have advantages which may permit an improvement in the general availability and commercialization of assays for the serodiagnosis of rickettsial infections. Mediterranean spotted fever (MSF) is a tick-transmitted rickettsiosis caused by Rickettsia conorii. The disease is endemic in southern Europe, Africa, the Middle East, and the Indian subcontinent. For reasons which are poorly understood, the incidence of human disease has increased dramatically in Spain, France, Italy, and Israel during the past decade (21, 32, 40). MSF has often been considered to be a mild, self-limiting, febrile exanthem not requiring treatment (25, 37). Indeed, mild undiagnosed illness or asymptomatic cases may account for some of the widespread 7 to 26% seroprevalence rates found in some endemic areas (8, 10, 20, 22, 33, 36). However, fulminant disease may occur in 1 to 6% of patients treated for MSF, and this form can be fatal (5, 9, 38, 39). Clinical misdiagnosis of MSF is still common and may result in delayed treatment or use of inappropriate antibiotic regimens. Confirmatory diagnosis of MSF is generally based on serologic procedures (6, 13, 35) which are very similar to those used for typhus or another spotted fever rickettsiosis, Rocky Mountain spotted fever (11, 12, 15-17, 24, 26-28). However, the current serodiagnostic tests for MSF all have requirements or shortcomings which limit their usefulness. The Proteus OX19 and OX2 (Weil-Felix) febrile agglutination test is not very sensi; it is not rickettsia specific, and the antibodies detected are largely of the immunoglobulin M (IgM) class and are produced for only a short time following infection (11, 16, 17, 35). The rickettsial complement fixation test is very insensi in the first 3 weeks postonset and is rather labor intensive; it cannot be used with anticomplementary sera, it is strongly IgG dependent, and it requires * Corresponding author. t Present address: Centre National de Reférénce des Rickettsioses, Centre Hospitalier et Universitaire de la Timone, Marseille Cedex 5, France substantial quantities of antigen (17, 24, 27). Rickettsial microagglutination assays are also insensi and require large amounts of antigen, and the species-specific corpuscular antigens are difficult to prepare and stabilize (17, 24, 26, 27). The latex agglutination assay for MSF is sensi but detects IgM preferentially, uses a group-reac rickettsial antigen, and cannot be used with lipemic sera (12, 14, 15, 31). Rickettsial passive hemagglutination tests react like the latex agglutination tests but suffer from substantial problems associated with storage of the less stable carrier, the erythrocyte (17, 27, 31). The microimmunofluorescence (MIF) assay can be used as an immunoglobulin class-specific test and is sensi, but it requires a fluorescence microscope, which may not be available in developing countries or in field situations (17, 24, 26-28, 31-35). In the recently developed immunoperoxidase assay, the fluorescence microscope is replaced by a light microscope (29). Neither the MIF assay nor the immunoperoxidase assay requires much antigen, but the storage and handling of antigen are inconvenient, and when human sera are used, neither assay is species specific among the rickettsiae (13, 14, 26, 28). For confirmation of the rickettsial species causing an infection, either multiple MIF or immunoperoxidase assay titers of the sera against different species are required (26, 28) or, preferably, the sera must be absorbed with different antigens and then titers must be determined. Neither method is convenient, and the quantities of antigen needed for absorption are prohibi for general use. We believe that there is a need for new serodiagnostic assays for MSF which avoid some of the limitations of the current tests. The major purpose of this work was to evaluate the sensitivity and specificity of a line blot assay for MSF. The line blot is an assay derived from the dot immunoassay which permits the very rapid simultaneous evaluation of the class-specific reactions of a serum against as many as 45

2 2074 RAOULT AND DASCH antigens (in duplicate) but which uses minimal amounts of both antigen and antibody (D. Raoult and G. A. Dasch, submitted for publication). It was compared against MIF, which is the best present standard for MSF diagnoses. Western immunoblotting was used to identify the individual antigenic components being recognized by the MSF sera. Finally, we also evaluated whether the Western blot and line blot immunoassays exhibited any species specificity for MSF infection by testing MSF sera against six different spotted fever group rickettsial species, several unclassified spotted fever group strains from countries close to the Mediterranean basin, and R. typhi. MATERIALS AND METHODS Sera. Sera were obtained during the summers of 1982 through 1984 in Marseille, France, where MSF is endemic. A total of 42 MSF cases were selected among patients with febrile exanthems compatible with MSF; 29 of these patients were examined directly by one of us (D.R.) (37). All of these patients had at least two of the three major symptoms of MSF (fever, rash, and eschar), and all exhibited seroconversion or fourfold rises in titers against R. conorii by MIF. Of the 42, 12 also had posi immunofluorescence in skin biopsies with antisera against R. conorii (30). One hundred serum specimens encompassing the disease from day 4 through 1 year postinfection were used in this study (14 single serum specimens and 86 specimens from 28 other individuals). Fifty control specimens were selected from other patients who had acute febrile illness during the summer but who were clinically not compatible with a diagnosis of MSF and were not exhibiting seroconversion against R. conorii by MIF. The control patients were diagnosed as follows: infectious mononucleosis (n = 10), brucellosis (n = 3), cytomegalovirus (n = 7), Chlamydia trachomatis (n = 5), rubella (n = 6), hepatitis B (n = 5), measles (n = 6), typhoid (n = 2), and mumps (n = 6). MIF. The MIF assay was conducted as described previously (31, 34, 35). R. conorii Moroccan (ATCC VR 141) was grown in Vero cells. Heavily infected cultures were harvested and placed on slides, air dried, and fixed with acetone. Serum dilutions were applied for 30 min at 37 C in a humid chamber and then treated with specific fluorescein isothiocyanate-conjugated goat anti-human IgG (-y)- and IgM (,u)-specific conjugates (Bio-Merieux-France). Rickettsial antigens. R. typhi Wilmington was grown in the yolk sac of embryonated chicken eggs and purified of host cell contaminants by differential centrifugation and isopycnic banding in Renografin density gradients (42). Spotted fever strains were obtained as follows: R. rickettsii R (ATCC VR 891) and R. conorii Moroccan (ATCC VR 141), American Type Culture Collection, Rockville, Md.; Ethiopian spotted fever strain C84360 and R. slovaca B, W. Burgdorfer and M. Peacock, Rocky Mountain Laboratory, National Institute of Allergy and Infectious Diseases, Hamilton, Mont.; R. akari Hartford (H5564) and R. australis Phillips (H2292), J. M. Spielman, Harvard School of Public Health, Boston, Mass.; Israeli tick typhus, C. L. Wisseman, Jr., University of Maryland School of Medicine, Baltimore; R. sibirica 232 and Thai tick typhus 118, M. Bozeman, Bureau of Biologics, Food and Drug Administration, Bethesda, Md.; R. bellii , C. Pretzman, Vector Borne Diseases Unit, Ohio Department of Public Health, Columbus; and R. conorii Marseille 2, new isolate by Raoult and Dasch from Marseille MSF patient blood sample no. 2. R. bellii was grown in irradiated (3,000 R) Vero cell cultures, whereas all J. CLIN. MICROBIOL. of the other spotted fever strains were grown in irradiated mouse L929 monolayer cultures. The spotted fever rickettsial antigens were prepared as follows. Crude yolk sac-grown seeds (105 to 107 PFU) were diluted in 10 ml of brain heart infusion (GIBCO Laboratories, Grand Island, N.Y.). An inoculum of 2 ml was placed in each of five T175 (Nunc, Roskilde, Denmark) tissue culture flasks containing 107 cells (irradiated 1 to 4 days previously) from which the tissue culture medium had been removed. The flasks were rocked gently on a platform rocker for 60 min, the inoculum was aspirated, and 35 ml of minimal essential medium with 5% fetal bovine serum, 2 mm L- glutamine, and 1% (vol/vol) nonessential amino acid stock added. When 20 to 100 rickettsiae per attached cell were observed by phase microscopy in the fluid phase (5 to 10 days postinfection), the tissue culture medium was decanted into 50-ml conical centrifuge tubes, which were centrifuged at 1,000 rpm for 10 min in an IEC PRJ centrifuge to pellet crude cell debris. The supernatant fluids were removed, the rickettsiae were pelleted at 17,500 x g for 15 min, and the clarified supernatant was carefully discarded. The small rickettsial pellets were then pooled, diluted in 40 ml of K36 buffer (42), repelleted at 17,500 x g, and finally suspended in 0.5 or 1.0 ml of water and frozen at -20 C. Protein concentrations were determined with a modified Lowry reagent (Pierce Chemical Co., Rockford, 111.). Line blot immunoassay. Antigen was applied to nitrocellulose sheets as a line with a dip pen point (Speedball pens, round gothic letter style, B6; Hunt Manufacturing Co., Statesville, N.C.), using a ruler as a guide. The antigen solution was prepared by mixing two parts of the rickettsial antigen, adjusted to 1 mg of protein per ml, with one part of phosphate-buffered saline and one part of the Laemmli solubilizer (4% sodium dodecyl sulfate, 10% 2-mercaptoethanol, 0.05% bromphenol blue, M Tris hydrochloride [ph 6.8], 25% glycerol) at room temperature (18). Lines of individual rickettsial antigens were placed 0.5 cm apart. Nonspecific binding sites were blocked by placing the airdried nitrocellulose sheet in TBS (10 mm Tris hydrochloride [ph 7.5], 250 mm NaCl, 0.01% Merthiolate) with 5% nonfat dry milk for 1 h. Sheets were washed in TBS and then in water, air dried, and cut perpendicularly to the lines in 5-mm strips. The strips were then placed individually in 25 wells of milled acrylic incubation trays (Bio-Rad Laboratories, Richmond, Calif.) and incubated with 2 ml of sera which had been diluted 1:200 in TBS with 3% milk (TBSM). After 2 h at 30 C, the strips were rinsed in water and washed twice for 10 min in TBS. A 2-ml portion of horseradish peroxidaseconjugated, affinity-purified rabbit antibodies to human IgG (-y-chain specific, diluted to 1:1,000 in TBSM), or IgM (,u-chain specific, diluted to 1:3,000 in TBSM) (Calbiochem- Behring, La Jolla, Calif.) was then incubated with the strips for 2 h at 30 C. Finally, the strips were rinsed in water and washed three times for 10 min each in TBS. Bound enzyme was detected by reaction with 0.015% 4-chloro-1-naphthol % hydrogen peroxide in 16.7% methanol in TBS for 10 min. The strips were thoroughly washed in water and air dried between sheets of filter paper, and the reactions were observed and recorded. Western blot immunoassay. Rickettsial antigens were solubilized at room temperature in 50% Laemmli solubilizer (see above) (18). Polyacrylamide gel electrophoresis of the antigens was carried out on 8 to 16% acrylamide (acrylamide/bisacrylamide ratio of 40:0.8) linear gradient separating gels (0.75 mm by 10 cm by 14 cm [width]) with 5% stacking gels (2 cm below comb), using the discontinuous

3 VOL. 27, 1989 BLOT IMMUNOASSAYS FOR MEDITERRANEAN SPOTTED FEVER 2075 *: I. * -S E *..& ;=.,.s... a bcde* f ghblj k Im f.-.. m, FIG. 1. IgM line blot reactions against spotted fever group rickettsiae of 13 serum specimens from patients with MSF. The specimens are numbered a to m. The antigens are numbered 1 to 10 as follows: 1, R. rickettsii R; 2, R. conorii Marseille 2; 3, R. conorii Moroccan; 4, R. sibirica 232; 5, R. slovaca B; 6, Israeli tick typhus; 7, Ethiopian SF C84360; 8, R. akari Hartford; 9, R. australis Phillips; and 10, R. bellii Tris buffer system of Laemmli (18). Two groups of 10 samples of 20,ul (10 to 15,ug of protein) were applied per gel. Twelve gels were run simultaneously in a gel slab unit (no. SE700; Hoefer Scientific Instruments, San Francisco, Calif.) at 8 ma per gel at 4 C with prechilled buffer; 2 F1d of 1% methyl green in 20% glycerol was applied to lane 11 (2 h before completion of electrophoresis) as a marker to permit accurate separation of the two sets of antigens on the nitrocellulose after the electroblot transfer had been effected. Resolved antigens were transferred to nitrocellulose (pore size, 0.45,um; BA 85; Schleicher & Schuell, Inc., Keene, N.H.) at 4 C in a TE50 Transphor (Hoefer) at 0.5 A for 4 h with a 25 mm NaH2PO4 (ph 7.5) buffer. After air drying, nonspecific sites on the nitrocellulose were blocked for 2 h with TBS containing 5% nonfat dry milk. The sheets were again air dried and then cut into two sheets of 10 identical lanes by using the methyl green marker. Immunodetection was conducted precisely as described for the line blot immunoassay, with the same reagents but using 25-ml volumes per half gel sheet. Lipopolysaccharide (LPS) antibody reactions were identified by their characteristic multiband washboard patterns and resistance to proteinase K digestion (2, 7, 14) and by the cross-reactivity of a subset of these antibodies with purified LPSs of Proteus spp., Legionella spp., and typhus rickettsiae (41; G. A. Dasch, unpublished observations). Protein reactions were localized to single bands stainable with Coomassie blue and were proteinase K sensi. Their molecular weights corresponded to those of proteins studied previously with monoclonal antibodies and identified as the major species-specific protein antigens of the rickettsiae (1-4, 7, 19, 23). RESULTS Line blot immunoassay. The line blot immunoassay was evaluated with a serum dilution of 1:200 to minimize the volume of serum required (20,u1 for both IgM and IgG assays) but particularly to permit direct comparison with the Western blot assay, which requires a larger volume (0.25 ml) of serum. With well-characterized MIF-posi sera, this serum dilution gave strong posi violet line blot reactions on a white background (Fig. 1, lanes c to i). No false-posi reactions were detected among the 50 control serum specimens from febrile patients without MSF (nega control reactions were entirely comparable to the nonreac MSF J TABLE 1. Results of the line blot assays on 100 serum specimens from 42 patients with serologically confirmed MSF No. of posi reactions Antigen tested Sera Patients IgM IgG IgM IgG R. conorii Marseille R. conorii Moroccan R. akari Hartford Ethiopian SF C R. rickettsii R R. australis Phillips Thai tick typhus R. sibirica Israeli tick typhus R. bellii R. typhi Wilmington serum reactions shown in Fig. 1, lanes a, b, and j to m). The pen point left a nega relief on the nitrocellulose, which made location of individual antigen lines easy even with nega reactions; these occasionally appeared as weak reactions in photographs but were in fact readily distinguished by their lack of color (Fig. 1, lane m). The 3,300 line blot results which were obtained with the 100 MSF and 50 control serum specimens are summarized in Tables 1 to 3. IgM reactions were detected against all 10 spotted fever rickettsiae in 41 of the 69 serum specimens which exhibited any IgM reactivity (Table 1). In contrast, only 29 of the 67 serum specimens posi for IgG reacted against all 10 rickettsiae. Although the largest number of posi IgM or IgG reactions were obtained against the two R. conorji isolates as expected, most of the sera reacted with antigenically distant spotted fever species such as R. akari and Thai tick typhus. This suggested that the test detects primarily group-reac antigen rather than species-specific antigens. The greater differences in IgG reactivity among the individual spotted fever group species were due in part to the fact that these acute-phase sera exhibited very strong IgM reactions, whereas the IgG reactions were weaker and more difficult to see. It was not determined whether these IgG reactions could be enhanced by prior removal of the IgM. In addition, some of the apparent species differences in reactivity probably reflected variations in the quality of the individual spotted fever antigens which were applied. On the other hand, the MSF sera reacted much less frequently with the R. typhi antigens in the IgM assay and only once in the IgG line blot (Table 1). Since highly purified R. typhi antigen was used, its lack of reactivity was not due to dilution with host cell proteins, as might be true for some spotted fever antigens with low reactivity. Consequently, with the antigens used here, the line blot assay appears to be primarily group specific but with 26% (18 of 69 serum specimens) cross-reactivity with typhus group in the IgM assay. Comparison of MIF and Une blot assays. Results obtained by line blot with R. conorii Marseille 2 and those previously obtained with R. conorii Moroccan are compared in Tables 2 and 3. The Marseille 2 and Moroccan strains had indistinguishable protein profiles on Coomassie blue-stained polyacrylamide gels (data not shown) and identical antigenic profiles by Western blotting analysis with a large number of animal and human sera (Fig. 2 shows two examples). By using MIF as the true measure of the reactivity of the sera and a cutoff for posi reactions of.32, 3 false-posi and 9 false-nega IgM line blot reactions were obtained

4 2076 RAOULT AND DASCH TABLE 2. Comparison of results by line blot and MIF assays with R. conorii antigens on 100 serum specimens from patients with MSF and 50 serum specimens from febrile control patients No. of specimens or patients at indicated MIF titer Line blot IgM IgG reactions Line Line < >64 blot < >64 blot Serum results Line blot posi Line blot nega MIF Patient results Line blot posi Line blot nega MIF with 100 MSF serum specimens (Table 2). No false-posi IgM or IgG line blot reactions were obtained with the 50 control specimens. With IgG line blot, 5 false-posi and 16 false-nega reactions were obtained. The MSF line blot data were analyzed rela to the MIF data by single serum and by patient for sensitivity, specificity, and efficiency in IgM and IgG class reactions (Table 4). Although the line blot was sufficiently specific, it was less sensi than MIF for both IgM and IgG when the minimal cutoff of.32 was used. This may be an unfair comparison, since the line blot data were obtained at a 1:200 dilution. However, although the apparent sensitivity of both the IgM and IgG line blot assays was better when the line blot analysis was compared with a more comparable MIF cutoff of >64, more apparent falseposi reactions were detected and the assay specificity declined, without any overall effect on the assay efficiency (Table 4). When seroconversions of 28 paired MSF serum specimens were examined, 2 false-nega and 3 falseposi results were obtained in both the IgM and IgG line blot assays rela to MIF (Table 3). With these data, both the IgM and IgG line blot assays again exhibited lower sensitivity and specificity than the MIF assay did (Table 4). Western blot immunoassay. With the exceptions of R. typhi and R. bellii LPS patterns, only subtle differences in the Western blot reaction patterns of the rickettsiae were observed in either the IgM or IgG assay (Fig. 2, Table 5). The R. typhi LPS pattern had a somewhat different molecular weight distribution from that of the eight spotted fever group rickettsiae (Fig. 2) and exhibited the least frequent (often weak) IgM reaction and only the rare IgG reaction TABLE 3. Comparison of results by line blot and MIF assays with R. conorii antigens on 28 paired serum specimens from patients with MSF No. of MIF seroconversions Line blot IgM IgG reactions Present Absent Line blot Present Absent Line blot Present Absent MIF TABLE 4. Statistical evaluation of reactivity of line blot assay for MSF on the basis of correlation of MIF and line blot results Population' Immuno- Parameter measures at indicated MIF cutoff e globulin Sensitivityb Specificity Efficiency' -32 >64.32 >64.32 >64 Single sera IgM IgG Patients IgM IgG Seroconversion IgM IgG adata of Tables 2 and 3. b True line blot posis/mif posis. C True line blot negas/mif negas. d True line blot posis plus negas/total results. (Table 5). R. bellii LPS was also less frequently detected than the other spotted fever group rickettsiae (Table 5), particularly by the IgG reaction (Fig. 2B), and exhibited a distinc high-molecular-weight bias in its molecular weight distribution (Fig. 2A). Protein reactions were restricted almost exclusively to the high-molecular-weight species-specific protein antigens (SPAs), which appear to be dominant protein immunogens in both typhus and spotted fever group rickettsiae (1-4, 7, 19). Although such protein reactions were detected with nearly the same frequency as LPS reactions by either IgM or IgG (Table 5), both the number of protein cross-reactions (Table 5) and the strength of the cross-reactions (Fig. 2) were more specific than found for LPS. R. typhi exhibited substantially fewer protein reactions than were found with LPS in the IgM class but more by IgG. In contrast, R. bellii, which does not appear to have a protein corresponding to the typhus and spotted fever group SPAs, had no IgM or IgG protein reactions (Table 5). e A J. CLIN. MICROBIOL. B FIG. 2. Western blot reactivity against rickettsiae in sera from patients with MSF. (A) IgM reaction of patient no. 1. (B) IgG reaction of patient no. 2. The asterisks indicate the SPA reaction detected at 120 kilodaltons. The brackets indicate the position of the LPS washboard at 20 to 50 kilodaltons. Lanes (antigens): 1, R. typhi; 2, R. bellii; 3, R. conorii Marseille 2; 4, R. conorii Moroccan; 5, Israeli tick typhus; 6, R. rickettsii; 7, R. sibirica; 8, Thai tick typhus; 9, R. akari; 10, R. australis.

5 VOL. 27, 1989 BLOT IMMUNOASSAYS FOR MEDITERRANEAN SPOTTED FEVER 2077 TABLE 5. Summary of IgM and IgG Western blot reactions of 19 serum specimens from patients with MSF No. of specimens reacting Antigen tested IgM IgG LPS Protein protein LPS or LPS Prt* LPS proten or R. conorii Marseille R. conorii Moroccan R. rickettsii R il R. akari Hartford il 3 15 R. australis Phillips il 2 il Thai tick typhus il 9 12 R. sibirica il Israeli tick typhus 17 il R. bellii R. typhi Wilmington Among the spotted fever group rickettsiae, the frequency and strength of protein reactions were greatest against both R. conorii strains, somewhat weaker with comparable reactions against a group including R. rickettsii, Thai tick typhus, R. sibirica, and Israeli tick typhus, and least with with the genetically most distant R. akari and R. australis (Table 5). DISCUSSION From our results we can draw three major conclusions about the line blot assay for MSF. First, the line blot assay has sensitivity and specificity similar to those of the MIF assay, but it uses a much lower concentration of serum. In our experience, as noted by Hechemy et al. (12), a weak reaction at a titer of 32 to 64 is not specific, since in southern France about 16% of blood donors have MIF reactions at this level (33, 36). Only 2% of blood donors were posi at an MIF titer of 128, and none were posi at an MIF titer of 256. If an MIF IgM titer of 128 is presumed diagnostic of past disease, and assuming there are no false-posi fine blot reactions at a serum dilution of 1:200, the line blot is actually more sensi with the 100 serum specimens than the MIF assay is; the assays are equally sensi if an MIF titer of 64 is considered diagnostic, and the line blot is slightly less sensi if a titer of 32 is used. With IgG the trends are similar, but there are slightly fewer differences between the two assays, particularly when analyzed by patient. Whether nonspecific reactions at higher concentrations of serum or greater antigen deposition would permit the line blot to greatly exceed the MIF standard for sensitivity without a large loss in specificity is unknown at present. Second, the present line blot assay is no more species specific than most of the other rickettsial tests are. LPS is a major immunogen for humans in all the rickettsioses; it is known to be a group antigen for typhus and spotted fever group rickettsiae, and it also contains determinants which cross-react between the two groups and with Proteus strains OX2 and OX19 and Legionella bozemanii (41, 43; D. Raoult and G. A. Dasch, manuscripts in preparation). The intrarickettsial LPS cross-reactivities are clearly demonstrated by the Western blot analyses reported here. Consequently, it is hardly surprising that a line blot assay in which unfractionated cells are used would be largely group reac. The Western blot data suggest that separate line blot assays based on purified LPS and SPAs might permit both completely group-reac and more specific tests to be done. Results of preliminary experiments with typhus group rickettsiae, which used proteinase K-digested whole cells as a source of enriched LPS and ultracentrifuged hypotonic shock extracts as the source of purified SPAs (3, 4; Raoult and Dasch, submitted), suggest that such refined line blot assays may be possible for the spotted fever group rickettsiae as well. Third, the line blot assay is far superior to the other tests in its ability to screen reactions of many sera against a large number of antigens with a minimum of effort and negligible equipment. Indeed, it is entirely possible to screen each serum specimen against the whole-cell antigens, purified LPS, and purified SPAs of 15 rickettsiae at the same time. Other bacteria or additional rickettsiae can also be included if duplicate assays are run on separate strips. Some other potential benefits of the line blot assay were not evaluated in the present study. For example, the line configuration of the assay should be highly suited to automated preparation of test antigen strips on nitrocellulose and to automated densitometric quantification or possibly even bar code reading of the results. The major current impediment to further development and widespread use of the assay may be difficulties in production and standardization of the antigens used in the assay. This problem is hardly unique to the line blot assay, since it is also encountered with the other rickettsial immunodiagnostic assays. We are currently developing and evaluating monoclonal antibodies against both cross-reac LPS determinants and species-specific epitopes of SPAs as reagents for standardizing the antigen quantity and quality. It is hoped that current efforts to clone and sequence the SPAs of rickettsiae (23; M. E. Dobson and G. A. Dasch, submitted for publication) will permit the use of recombinant protein products or synthetic peptides corresponding to species-specific regions as antigens in the line blot assay. Since only 19 of the 77 MSF serum specimens exhibiting MIF IgM or IgG titers of >32 (generally ones with defini MIF titers) were evaluated by Western blotting, it was not possible to make a full quantita evaluation of the rela sensitivity of Western and line blot assays. However, possibly because the immunoreac antigens are separated from nonreac material and concentrated in narrow bands in the Western blotting procedure, whereas they remain unfractionated in the line blot, more species were detected and the results were more defini by Western blotting than by line blotting for both IgM and IgG. In addition to confirming the LPS nature of the cross-reactions observed with these sera, the primary use of the Western blot was in permitting the first analyses of the reactivity of human MSF sera with the SPAs of a large number of spotted fever group rickettsiae. Studies with monoclonal antibodies and hyperimmune animal antisera have shown that these proteins have both cross-reac and species-restricted and, in some cases, even strain-restricted epitopes (1-4, 7, 19). The majority of MSF sera reacted most strongly against the homologous R. conorii SPAs (Fig. 2B; Table 5), but weaker reactions indica of cross-reac SPA epitopes were also present. One major exception is worth noting further. In Fig. 2A, a reaction completely compatible with that of sera from patients with murine typhus was found. The typhus SPA reacted more strongly than any of the spotted fever group rickettsiae. This observation cannot be considered unequivocally diagnostic by itself, since some normal individuals have such IgM SPA reactions. However, posi IgG SPA reactions have been found only in bona fide clinical cases of murine or epidemic typhus and are almost invariably accompanied by both IgM and IgG LPS Western blot reactions (G. A. Dasch et al., manuscripts in preparation).

6 2078 RAOULT AND DASCH Consequently, at least 2 of the 19 serum specimens may have actually been from typhus infections (Table 5) rather than MSF, while the other serum specimens have reactions with R. typhi which are more consistent with the weak reactions observed against epitopes shared on the SPAs and LPS of both typhus and spotted fever group rickettsiae (compare Fig. 2A and B) (Raoult and Dasch, in preparation). The Western blot may be a suitable alterna to current methods for establishing the rickettsial species causing a rickettsiosis, but this possibility requires much more detailed study. In conclusion, the line blot assay appears to be most useful for large-scale screening of sera when quantita titers are not needed and when tests against a large number of agents are required. Although nitrocellulose sheets for rickettsial line blot and Western blot assays could be prepared easily in quantity in a centralized laboratory and disseminated to individual medical offices or clinical laboratories for use, these assays would be unlikely to replace current assays where they are now readily available in a timely clinical setting. However, most use of the current tests is strictly retrospec and they are not widely done, and so these new tests may have sufficient advantages in simplicity, cost, and speed that they might be welcomed clinically. Indeed, the exceptional stability of the antigens on the nitrocellulose (data not shown) and the simplicity of their storage are particularly advantageous in situations in which such assays are performed infrequently. Posi presump tests by line blot or Western blot or both could then be confirmed by a reference method or sent to a reference laboratory. The format of the line blot assay may also be particularly useful for screening the many antigens that might be considered in cases with nonspecific or atypical clinical presentations. ACKNOWLEDGMENTS This investigation was supported by the Naval Medical Research and Development Command, research task 62770A 3M A870.AQ.120. We thank Emilio Weiss for his helpful comments on the manuscript and Walter G. Sewell and Richard Grays for their technical assistance. LITERATURE CITED 1. Anacker, R. L., R. H. List, R. E. Mann, and D. L. Wiedbrauk Antigenic heterogeneity in high- and low-virulence strains of Rickettsia rickettsia revealed by monoclonal antibodies. Infect. Immun. 51: Anacker, R. L., R. E. Mann, and C. Gonzales Reactivity of monoclonal antibodies to Rickettsia rickettsii with spotted fever and typhus group rickettsiae. J. Clin. Microbiol. 25: Dasch, G. A Isolation of species-specific protein antigens of Rickettsia typhi and Rickettsia prowazekii for immunodiagnosis and immunoprophylaxis. J. Clin. Microbiol. 14: Dasch, G. A., J. P. Burans, M. E. Dobson, F. M. Rollwagen, and J. Misiti Approaches to subunit vaccines against the typhus rickettsiae, Rickettsia typhi and Rickettsia prowazekii, p In L. Leive and D. Schlessinger (ed.), Microbiology American Society for Microbiology, Washington, D.C. 5. Devriendt, J., M. Staroukine, R. Amson, F. Crockaert, M. Dratwa, R. Karmali, and Y. Thoua Malignant Mediterranean spotted fever. Report of a case with multiple organ failure, hypocalcemia, and euthyroid sick syndrome. Ann. Intern. Med. 145: Edlinger, E Serological diagnosis of Mediterranean spotted fever. Ann. Inst. Pasteur Microbiol. 130A: Feng, H. M., D. H. Walker, and J. G. Wang Analysis of T-cell-dependent and -independent antigens of Rickettsia J. CLIN. MICROBIOL. conorii with monoclonal antibodies. Infect. Immun. 55: Garcia Curiel, A., and E. Najera Morrondo Estudio epidemiolégico de las rickettsiosis en la provincia de Sevilla, basado en las reacciones serolôgicas de inmunofluorescencia indirect. Rev. Sanid. Hig. Publica 58: Gear, J. H. S., G. B. Miller, H. Martins, R. Swanepoel, B. Wolstenholme, and A. Coppin Tick-bite fever in South Africa. The occurrence of severe cases on the Witwatersrand. S. Afr. Med. J. 63: Gross, E. M., R. A. Goldwasser, J. E. Bearman, I. Sarov, B. Sarov, V. Torok, and L. Naggan Rickettsial antibody prevalence in southern Israel: IgG antibodies to Coxiella burnetii, Rickettsia typhi, and spotted fever group rickettsiae among urban- and rural-dwelling and Bedouin women. Am. J. Trop. Med. Hyg. 32: Hechemy, K. E Laboratory diagnosis of Rocky Mountain spotted fever. N. Engl. J. Med. 300: Hechemy, K. E., R. L. Anacker, R. N. Philip, K. T. Kleeman, J. N. MacCormack, S. J. Sasowski, and E. E. Michaelson Detection of Rocky Mountain spotted fever antibodies by a latex agglutination test. J. Clin. Microbiol. 12: Hechemy, K. E., and E. E. Michaelson Recent advances in the immunoserology of rickettsioses. J. Med. Marseille 4: Hechemy, K. E., D. Raoult, C. Eisemann, Y. Han, and J. A. Fox Detection of antibodies to Rickettsia conorii with a latex agglutination test in patients with Mediterranean spotted fever. J. Infect. Dis. 153: Hechemy, K. E., and B. B. Rubin Latex-Rickettsia rickettsii test reactivity in seroposi patients. J. Clin. Microbiol. 17: Hechemy, K. E., R. W. Stevens, S. Sasowski, E. E. Michaelson, E. A. Casper, and R. N. Philip Discrepancies in Weil- Felix and microimmunofluorescence test results for Rocky Mountain spotted fever. J. Clin. Microbiol. 9: Kaplan, J. E., and L. B. Schonberger The sensitivity of various serologic tests in the diagnosis of Rocky Mountain spotted fever. Am. J. Trop. Med. Hyg. 35: Laemmli, U. K Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature (London) 227: Li, H., B. Lenz, and D. H. Walker Protec monoclonal antibodies recognize heat-labile epitopes on surface proteins of spotted fever group rickettsiae. Infect. Immun. 56: Mansueto, S., G. Tringali, R. Di Leo, M. Maniscalco, M. R. Montenegro, and D. H. Walker Demonstration of spotted fever group rickettsiae in the tache noire of a healthy person in Sicily. Am. J. Trop. Med. Hyg. 33: Mansueto, S., G. Tringali, and D. H. Walker Widespread, simultaneous increase in the incidence of spotted fever group rickettsioses. J. Infect. Dis. 154: Mansueto, S., G. Vitale, M. D. Miceli, G. Tringali, P. Quartararo, D. M. Picone, and C. Occhino A sero-epidemiological survey of asymptomatic cases of boutonneuse fever in western Sicily. Trans. R. Soc. Trop. Med. Hyg. 78: McDonald, G. A., R. L. Anacker, and K. Garjian Cloned gene of Rickettsia rickettsii surface antigen: candidate vaccine for Rocky Mountain spotted fever. Science 235: Newhouse, V. F., C. C. Shepard, M. D. Redus, T. Tzianabos, and J. E. McDade A comparison of the complement fixation, indirect fluorescent antibody, and microagglutination tests for the serological diagnosis of rickettsial diseases. Am. J. Trop. Med. Hyg. 28: Olmer, D., and J. Olmer Fievre boutonneuse. Fievre exanthematique du littoral mediterraneen. Masson, Paris. 26. Ormsbee, R., M. Peacock, R. Philip, E. Casper, J. Plorde, T. Gabre-Kidan, and L. Wright Antigenic relationships between the typhus and spotted fever groups of rickettsiae. Am. J. Epidemiol. 108: Philip, R. N., E. A. Casper, J. N. MacCormack, D. J. Sexton, L. A. Thomas, R. L. Anacker, W. Burgdorfer, and S. Vick A comparison of serologic methods for diagnosis of Rocky Mountain spotted fever. Am. J. Epidemiol. 105:56-67.

7 VOL. 27, 1989 BLOT IMMUNOASSAYS FOR MEDITERRANEAN SPOTTED FEVER Philip, R. N., E. A. Casper, R. A. Ormsbee, M. G. Peacock, and W. Burgdorfer Microimmunofluorescence test for the serological study of Rocky Mountain spotted fever and typhus. J. Clin. Microbiol. 3: Raoult, D., C. De Micco, H. Chaudet, and J. Tamalet Serological diagnosis of Mediterranean spotted fever by the immunoperoxidase reaction. Eur. J. Clin. Microbiol. 4: Raoult, D., C. De Micco, H. Gallais, and M. Toga Laboratory diagnosis of Mediterranean spotted fever by immunofluorescent demonstration of Rickettsia conorii in cutaneous lesions. J. Infect. Dis. 150: Raoult, D., K. E. Hechemy, and H. Chaudet Sérologie de la fièvre boutonneuse méditerranéenne. Cinétique des anticorps détectés par trois méthodes: l'immunofluorescence indirecte, l'hémagglutination indirecte, et l'agglutination latex. Pathol. Biol. 33: Raoult, D., G. Lepeu, P. De Micco, J. L. San Marco, P. J. Weiller, H. Gallais, and P. Casanova Recrudescence de la fièvre boutonneuse méditerranéenne dans le Sud de la France. J. Med. Marseille 3: Raoult, D., D. Nicolas, P. De Micco, H. Gallais, and P. Casanova Aspects épidémiologiques de la fièvre boutonneuse méditerranéenne en Corse du Sud. Bull. Soc. Pathol. Exot. 78: Raoult, D., S. Rousseau, B. Toga, H. Chaudet, J. Tamalet, and P. De Micco Indirect fluorescent antibody test in Mediterranean spotted fever. Detection and kinetics of antibodies. Acta Virol. 29: Raoult, D., S. Rousseau, B. Toga, C. Tamalet, H. Gallais, P. De Micco, and P. Casanova Diagnostic sérologique de la fièvre boutonneuse méditerranéenne. Pathol. Biol. 32: Raoult, D., B. Toga, H. Chaudet, and C. Chiche-Portiche Rickettsial antibody in southern France: antibodies to Rickettsia conorii and Coxiella burnetii among urban, suburban and semi-rural blood donors. Trans. R. Soc. Trop. Med. Hyg. 81: Raoult, D., P. J. Weiller, A. Chagnon, H. Chaudet, H. Gallais, and P. Casanova Mediterranean spotted fever: clinical, laboratory and epidemiological features of 199 cases. Am. J. Trop. Med. Hyg. 35: Raoult, D., P. Zuchelli, P. J. Weiller, C. Charrel, J. L. San Marco, H. Gallais, and P. Casanova Incidence, clinical observations and risk factors in the severe form of Mediterranean spotted fever among patients admitted to hospital in Marseilles J. Infect. 12: Ruiz Beltràn, R., J. I. Herrero Herrero, A. M. Martin Sanchez, V. Vicente Garcia, F. Sanz Ortega, A. A. Mateos Sanchez, R. Querol Prieto, and J. de Portugal Alvarez Formas graves de fiebre exantematica mediterranea. Analisis prospec de 71 enfermos. Ann. Med. Intern. 2: Scaffidi, V Contemporaneità della recente espansione endemo-epidemica della febbre bottonosa in Italia ed in Israele. G. Mal. Infett. Parassit. 34: Sompolinsky, D., I. Boldur, R. A. Goldwasser, H. Kahana, R. Kazak, A. Keysary, and A. Pik Serological cross-reactions between Rickettsia typhi, Proteus vulgaris OX19 and Legionella bozemanii in a series of febrile patients. Isr. J. Med. Sci. 22: Weiss, E., J. C. Coolbaugh, and J. C. Williams Separation of viable Rickettsia typhi from yolk sac and L cell host components by Renografin density gradient centrifugation. Apple. Microbiol. 30: Westfall, H. N., R. A. Goldwasser, E. Weiss, and D. Hussong Prevalence of antibodies to Legionella species in a series of patients in Israel. Isr. J. Med. Sci. 22: Downloaded from on April 28, 2018 by guest