Gene, 115 (1992) 219-222 © 1992 Elsevier Science Publishers B.V. All rights reserved. 0378-1119/92/$05.00
219
GENE 06407
Immunodiagnosis of nocardiosis* (Nocardia; antigen purification; monoclonal antibodies; serological diagnosis; leprosy; affinity chromatography; Mycobacterium; Western blot; circulating antigen)
P. Boirona and D. Stynenb a
Unit$ de Mycologie, Institut Pasteur, F 75724 Paris Cedex 15 (France), and b Sanofi/Diagnostics Pasteur. B 3600 Genk (Belgium) Tel. (32-
11)384252 Receivedby J.C. Ensign:23 August 1991 Accepted: 30 December 1991 Receivedat publishers:31 January 1992
SUMMARY
A specific immunodominant 54-kDa antigen was purified from a culture filtrate of Nocardia asteroides by immunoaffinity chromatography. The chromatography column was prepared with immunoglobulin G obtained from sera from patients with lepromatous leprosy. Unbound solutes consisted of specific, partially purified N. asteroides antigens, primarily a 54-kDa band, accompanied by two others of 31 and 62 kDa. The Western blot (immunoblot) technique was applied to detecting the immunologic response to nocardiae in the serum of nocardiosis patients. Each of the serum samples from immunosuppressed or immunocompetent patients infected with N. asteroides reacted with the 54-kDa band, and two reacted with the 31- and 62-kDa bands. There was no reaction to either the 54- or the 31-kDa antigen with all serum samples obtained from patients with tuberculosis, except for one, with all serum samples obtained from patients with leprosy, or with all sera obtained from healthy controls. The partially purified 54-kDa antigen, specific for N. asteroides, was used as the immunogen to generate monoclonal antibodies (mAbs) and two mAbs were selected. As determined by Western blot, both mAbs reacted with the 54-kDa band. Using indirect immunofluorescence or enzyme immunoassay with whole N. asteroides microorganisms, the mAbs did not react with N. asteroides cells. No cross-reactivity with mycobacterial antigens, either culturefiltrate antigens or tuberculin, was exhibited with any of the two mAbs. These mAbs are candidates to be used for the development of a sensitive and specific diagnostic test for nocardiosis.
INTRODUCTION
Nocardiae are Gram +, acid-fast variable, branching, illamentous bacteria. Opportunistic nocardial disease, usually caused by N. asteroides (accounting for 80-90% ofthe Correspondenceto: P. Boir,~n,Unit6 de Myeologie,lnstitut Pasteur, 28 rue du Dr Roux, F 75724 Paris Cedex 15 (France) Tel, (33-1)45688000, ext. 7252; Fax (33-1)45688420. * Presented at the International Symposiumon Biologyof Actinomycetes, University of Wisconsin, Madison, Wl (USA) 11-16 August 1991. Abbreviations: A49O, absorbance at 490 nm; EIA, enzyme immunoassay; IgG, immunoglobulinG; mAb, monoclonal antibody; M., Mycobacterium;
N., Nocardia.
cases), N. otitidiscaviarum and N. farcinica, has been associated with patients with an impaired host-defense mechanism (Shadomy and Warren, 1988), including steroid therapy, cytotoxins, immunosuppression, leukemia, lymphoma, neutropenia, but infection in normal hosts has also been reported (Rosendale et al., 1988). Infection may occur by inhalation or by traumatic implantation in the skin. Three distinct clinical syndromes may evolve: primary pulmonary and systemic, primary cutaneous, and primary subcutaneous. Primary pulmonary infection may be subclinical or pneumonic, chronic or acute, and may become systemic by hematogeneous spread. Infection frequently disseminates to the central nervous system and soft tissues and, less commonly, to other organs, such as the kidney.
220 N. brasiliensis is more virulent and often has caused disease in apparently immunocompetent individuals (Smego and Gallis, 1984). It may be considered essentially a primary pathogen, most often responsible for actinomycetomas and seen almost exclusively in Central and South America. Additionally, there is now considerable evidence that nocardiosis is transmissible, and epidemics have occurred in a renal transplant unit (Badour et al., 1986). Nocardiosis is difficult to recognize on the basis of clinical, radiological, or histological findings. A definitive diagnosis depends on obtention of appropriate clinical samples, which may require the use of invasive procedures, and on the isolation and identification of a Nocardia sp., a procedure which can take two to three weeks (Gordon, 1985). A high level of suspicion on the part of the physician, who should always alert the laboratory when nocardiae are suspected, and of experience on the part of the laboratory personnel, are essential for early detection. Therefore, tests for the demonstration of delayed cutaneous hypersensitivity and for humoral antibodies for the diagnosis of nocardiosis have been investigated as alternatives to classical microbiological laboratory procedures. It is now well-known that members of the genera Nocardia and Mycobacterium have many properties in common, as attested by similarities of morphological, cultural, chemotaxonomic, and particularly serological characteristics (Nesterenko et al., 1988). Nocardiae possess several polysaccharidic antigens, principally arabinomannan and arabinogalactan, which are shared with mycobacteria, corynebacteria and rhodococci. Ribosomal antigens cross-react with ribosomal antigens of mycobacteria and corynebacteria (Lamb et al., 1978). There are also many crossreactions with Streptomyces spp. Consequently, a number of skin test antigens that have been produced were not clinically useful, and most of the serological tests that have been developed have met with limited success because cross-reaction or lack of sensitivity of nocardial antigens with sera from cases of tuberculosis and leprosy was noted with all methodologies. Moreover, reactions occur frequently within a normal healthy population. Whole-cell and filtrate antigens were used in immunodiffusion tests (Blumer and Kaufman, 1979; Humphreys et al., 1975). These tests were 45-47% sensitive. 7he use of reactions of identity, with serum from a human nocardiosis case as reference, was more specific and more predictive than the presence of any band, but less sensitive. Culture filtrate antigens were also used to develop a complement fixation test, with rabbit reference antiserum (Shainhouse et al., 1978). An overall sensitivity of 81% was found. These data supported the suggestion that infected individuals can be identified but that a high degree of crossreactivity seriously hinders the specificity of these tests. Moreover, becar~se of the relatively poor sensitivity of these
assays, they were unable to detect antibody response in immunosuppressed patients. In addition to the methods used for detection, considerable variation of the antigenic composition of N. asteroides could be attributed to the variable cultural conditions in these studies. EXPERIMENTAL AND DISCUSSION
(a) Partial purification of 54-kDa antigen Sugar et al. (1985) have documented the presence of immunodominant N. asteroides antigens. One of these proteins has been shown to be a highly specific marker for identifying patients infected with N. asteroides, N. brasiliensis or N. otitidiscaviarum (Angeles and Sugar, 1987). Positive antibody titers were found using EIA in 91 ~ of the patients with nocardiosis. None of the hospitalized control patients or the patients with tuberculosis had positive titers. We take advantage of the high degree of cross-reactivity between nocardiae and mycobacteria to partially purify the 54-kDa Nocardia-specific antigen by immunoaftinity chromatography (Boiron and Provost, 1990). The medium used for the crude antigenic production was prepared from Bennett's broth medium (Jones, 1949) according to a modification of the double dialysis method of Edwards (1972), as to eliminate interference from extraneous proteins in the analysis of N. asteroides antigens. For partial purification of 54-kDa antigen, the double dialyzed antigen was eluted through an affinity chromatography column. The chromatography column was prepared with resin coupled to IgG obtained from patients with lepromatous leprosy (Boiron and Provost, 1990). Unbound solutes consisted of specific, partially purified N. asteroides antigens, primarily a 54-kDa band, accompanied by two others of 31 and 62 kDa, The previous study of Sugar et al. (1985) also reported the presence of bands of about 24 and 65 kDa in culture filtrates which were consistently more heavily stained than were the other bands.
(b) Diagnosis of nocardiosis by Western blot (immunoblot) assay The Western blot technique was applied to detect the immunologic response to Nocardia. A particular pattern of immunological reactivity in patients with infection due to N. asteroides (reactivity of the 54-kDa band with or without reactivity to the 31- and 62-kDa bands) (Table I) was demonstrated, Nocardiosis is encountered with increasing frequency in high-risk immunocompromised hosts, who often produce little antibody during systemic infections. Current serological methods, including EIA, failed to reveal any serological response to Nocardia antigens in such patients (Gordon, 1985). The high sensitivity of the biotin-avidin system, because of the amplification due to multiple-site binding,
221 TABLE I
TABLE II
Analysis of sera from 35 patients with nocardiosis, leprosy, or tuberculosis and from seven healthy individuals by Western blot technique
Characterization of mAbs to Nocardia asteroides mAb a
Disease"
Number of patients
Number of patients positive with proteins of.. b 31kDa
Nocardiosis Non-immunocompromised lmmunosuppressed Leprosy Tuberculosis None
5 10 8 5 7
2 0 0 0 0
54kDa
5 10 0 1 0
62kDa
2 0 1 0 1
" Of the 15 patients with nocardiosis, five were without other discernible disease; ten were under immunosuppressive therapy and had documented pulmonary or disseminated infection. In all instances, the serum samples were obtained at the time of bacteriological diagnosis. " Proteins consisted of specific, partially purified IV. asteroides antigens, obtained according to the procedure described by Boiron and Provost (1990). Immune-assay was performed by Western blotting using biotinovidin system (Boiron and Provost, 1990).
improves the detection of humeral response, and a positive serological response became evident in these immunosuppressed hosts.
(c) mAbs against the 54-kDa antigen The production and characterization of mAbs against the 54-kDa Nocardia antigen was undertaken, as a step toward the standardization of reagents for immunodiagnosis of nocardiosis. A LOU/C rat was immunized with the partiaily purified 54-kDa antigen as immunogen. The spleen cells were then fused with rat myeloma cells IR9835, using polyethylene glycol 4000 as the fusing agent. After the initial screening, two clones of hybridoma were selected for further characterization, cloning by limiting dilution and ascites production. They were called EB-NA 1 and EB-NA2. In an indirect EIA, EB-NAI had a higher avidity for the partially purified 54-kDa preparation than EB-NA2. The avidity constants were 10S/M and 4 x 107/ M, respectively (Table II). Three methods were used to assess anti-54-kDa antigen mAb activity: Western blot assay, indirect immunofluorescence, and EIA on whole Nocardia cells (Boiron and Provost, 1988). Using the Western blot technique on partially purified 54-kDa antigen, only the 54-kDa band was revealed by both mAbs. The 54-kDa antigen was apparently not exposed to microbial surface, as evidenced from indirect immunofluorescence experiments that were negative with all mAbs tested. This was confirmed by lack of reaction of mAbs by EIA on whole Nocardia cells.
EIA
Avidity
(/4490) b
constantc
Indirect immunefluorescenced
EIA/whole cell e (.449o)
l0 s
m
10 7
m
0.15 0.18
(l/M) EB-NAI EB-NA2
1.50 1.30
"4x
" Screening of mAbs was carried out according to standard EIA procedures (Conroy et al., 1991) using partially purified antigen and cell culture supernatants as primary antibodies. Two clones were selected for further characterization, cloning by limiting dilution and ascites production. The mAbs were purified from the ascitic fluid by allotype specific affinity chromatography on an affinity matrix consisting ofmAb MARK3 (Biosys, Compi~gne, France) coupled to Sepharose-4B beads (Pharmacia, Uppsala, Sweden) (Bazin et al., 1984). Both mAbs were of the leG2 isotype, as determined by an indirect double sandwich EIA (Manouvriez et al., 1990). b Reactivity of mAbs was determined according to standard EIA procedures (Conroy et ai., 1991). Absorbance was read at 490 nm. The avidities of the selected monoclonai antibodies were compared using an indirect EIA procedure (Van Heyningen and Van Heyningen, 1987). The K value (in I/M, molarity) was calculated as the reverse of the average mAb concentration (in M) of three assays at half maximum binding. d Indirect immunofluorescence test was performed following the procedure ofJim6nez et al. (1990). Immunofluorescencetests were negative (see section e). e The EIA procedure on whole Nocardia cells adsorbed on microtitration plates was previously described (Boiron and Provost, 1988). Absorbance was read at 490 nm.
Although the 54-kDa antigen was specific for Nocardia, we studied the specificity of mAbs against mycobacterial antigens (culture-filtrate antigen and tuberculin) using a dot immunobinding assay. As expected, no or weak crossreactivity was observed, suggesting that at least one of these mAbs may be used as a specific reagent. Antigenic factors displaying specificity for N. asteroides were characterized by EI-Zaatari etal. (1986) by the enzyme-linked immunoelectrotransfer blot technique adapted to isofocused polyacrylamide gels. The mAbs were produced against antigenic factors 1, 6 and 8; mAb against factor 8 cross-reacted with antigens ofM. intracellulare and M.fortuitum. The mAbs against factors 1 and 6 did not cross-react with cytoplasmic antigens of M. chelonae, M. intracellulore serotypes 4B and 8A, M. fortuitum, M. gordonae and M. kansasii. The relationship, if any, between antigens detected by these mAbs and 54-kDa antigen is unknown. The mAbs produced by Jim6nez et al. (1990) showed cross-reactivity with M. tuberculosis, M. intracellulare, M. scrofulaceum, M. kansasii and M. fortuitum, that may be viewed as undesirable for diagnostic purpose. The mAbs which reacted strongly with Nocardia were those that
222 showed the highest degree of cross-reactivity with Mycobacterium. This was not totally unexpected because both genera are antigenicaUy related (Rajki et al., 1982) and because the antigen used was unpurified whole-cell extracts. Moreover, the immunization of mice was m a d e with antigen in Freund's complete adjuvant. However, they recognized surface antigenic components of the Nocardia, as demonstrated by indirect immunofluorescence. One of our mAbs, EB-NA1, may have a significant diagnostic potential, as it could be assayed for the detection of the 54-kDa antigen of Nocardia that may be present in serum of patients with nocardiosis.
(d) Conclusions The role of the humeral response to infection with Nocardia spp. is not well understood. The use of a suitable antigen may further our understanding of the pathogenesis of nocardiosis. The 54-kDa antigen is a candidate to be used as a probe to study the humeral immune response to Nocardia. However, the immunosuppressed status of some patients may limit diagnostic procedures that rely on the detection of antibodies. The use of an antibody probe specific to the 54-kDa antigen could result in the development of a specific and sensitive serodiagnostic test for systemic nocardiesis based on detection of the circulating 54-kDa antigen. An increased sensitivity may be useful in detecting cases earlier in the course of the infection, since precocious and specific antibiotic therapy improves the prognosis. Such a reagent may be also of value in the characterization of the epitopes involved in host cellular reactions during invasive infection. REFERENCES Angeles, A.M. and Sugar, A.M.: Identificationof a common immunedominant protein in culture filtratesof three Nocardia species and use in etiologic diagnosis of mycetoma. J. Clin. Microbiol. 25 (1987) 2278-2280. Badour, L.M., Baselski,V.S., Herr, M.J., Christensen, G.D. and Bisno, A,L.: Nocardiosis in recipients of renal transplants: evidence for nosocomial acquisition. Am. J. Infect. Control 14 (1986) 214-217. Bazin, H., Cormont, F. and de Clercq, L.: Rat monoclonal antibodies,II. A rapid and efficientmethod ofpurificationfrom ascitic fluidor serum. J. Immunol. Methods 7 (1984)9-16. Blumer, S.O. and Kaufman, L.: Microimmunodiffusiontest for nocardi. osis. J. Clin. Microbiol. 10 (1979) 308-312. Boiron, P. and Provost, F.: Enzyme immunoassay on whole Nocardia asteroides cells for human nocardiosis. Serodiag. lmmunother. Infect. Dis. 2 (1988) 445-452. Boiron, P. aad Provost, F.: Use of partially purified 54-kilodalton antigen for diagnosisof nocardiosis by Western blot (immunoblot)assay. J. Clin. Mierobiol. 28 (1990) 328-33 I.
Conroy, J.M., Stevens, R.W. and Hechenzy, K.E.: Enzyme immunoassay. In: Balows, A., Hausler Jr., W.J., Herrmann, K.L., lsenberg, H.D. and Shadomy, H.J. (Eds.), Manual of Clinical Microbiology, 5th ed. American Society for Microbiology,Washington, DC, 1991, pp. 87-92. Edwards, J.H.: The double dialysis method of producing farmer's lung antigen. J. Lab. Clin. Med. 79 (1972)683-688. EI-Zaatari, F.A., Reiss, E., Yakrus, A.M., Bragg, S.L. and Kaufman, L.: Monoclonal antibodies against isoelectrically focused Nocardia asteroides proteins characterized by the enzyme-linkedimmunoelectrotransfer blot method. Methods Diagn. Immunol.4 (1986) 97-106. Gordon, M.A.: Aeroblcpathogenic Actinomycetaceae.In: Lennette, E.H., Balows, A., Hausler Jr. W.J. and Shadomy, H.J. (Eds.), Manual of Clinical Microbiology,4til ed. American Society for Microbiology, Washington, D.C., 1985, pp. 249-262. Humphreys, D.W., Crowder, J.G. and White,A.: Serologicalreactions to Nocardia antigens. Am. J. Med. Sci. 269 (1975) 323-326. Jimtnez, T., Diaz, A.M. and ZIotnik, H.: Monoclonal antibodies to Nocardia asteroides and Nocardia brasiliensis. J. Ciin. Microbiol. 28 (1990) 87-91. Jones, K.L.: Fresh isolates of actinomycetcs in which the presence of sporogenous aerial myceliais a fluctuatingcharacteristic. J. Bacteriol, 57 (1949) 141-145. Lamb, R., Deviile,J. and Cocito, C.: Immunologicalrelatedness of ribo. seines from mycobaeteria,nocardiae and corynebacteria,and microorganisms in leprosy lesions. Infect. lmmun. 22 (1978) 540-547. Manouvriez, P., Nisei, F., Delaunay, T. and Bazin, H.: Isotyping of rat monoclonalantibodies. In: Bazin, H. (Ed.), Rat Hybridomas and Rat Monc,clonal Antibodies.CRC Press, Boca Raton, FL, 1990, pp. 127137. Nesterenko, O.A., Kvasnikov, E.I. and Nogina, T.M.: A proposal for includingmycolicacid-containingnoeardioformand coryneformbacteria into the familyMycobacteriacea¢Chestar 1897 Emend. of the order ActinomycetalesBuchanan 1917. Actinomycetes20 (1988) 93!16. Rajki, K., Brehmer,W., Hammer, H., Fischer, W., Daus, H. and Mauch, H.: Analysisof the soluble cytoplasmiccomponents of Mycobacteria and Nocardia by crossed immunoelectrofocusing.Zentralbl. Bakteriol. Mikrobiol. Hyg. Abt. l Orig, A 251 (1982) 389-398. Rosendale, D.E., Myers,C., Boyko, EJ. and Jafek, B.: Nocardiaasteroides cervical osteomyelitis in an immunocompetent host. Otolaryngol. Head Neck Surg. 99 (1988)334-337. Shadomy, HJ. and Warren, N.G.: Nocardiosis. In: Balows, A., Hausler Jr., W.J., Ohashi, M. and Turano, A. (Eds.), Laboratory Diagnosis of Infectious Diseases, Principles and Practice, Vol. I, Bacterial, Mycotic and Parasitic Diseases. Springer-Verlag, New York, 1988, pp. 671-677. Shainhouse, .I.Z., Pier, A.C. and Stevens, D,A.: Complement fixation antibody test for human nocardiosis. J. CUn. Microbiol. 8 (1978) 516-519. Smego Jr., R.A. and Gallis, H.A.: The clinical spectrum of Nocardia brasiliensis infection in the United States. Rev. Infect. Dis. 6 (1984) 164-180. Sugar, A.M., Schoolnik, G.K. and Stev¢,~, D.A.: Antibody response in human nocardiosis: identification of two immunodominant culturefiltrate antigens derived from Nocardia asteroides. J. Infect. Dis. 151 (i 985) 895-901. Van Heyningen,V. and Van HeyningenS.: Rankingthe affinitiesof monoclonal antibodies. In: Bartal, A.H. and Hirshaut, Y. (Eds.), Methods of Hybridoma. Humana Press, Clifton, NJ, 1987, pp. 399-411.