Serum Antibodies to Pathogenic Actinomycetes in the Normal Human Population

Zbl. Bakt. 274, 398-405 (1990) © Gustav Fischer Verlag, StuttgartfNew York

Serum Antibodies to Pathogenic Actinomycetes in the Normal Human Population * HANS-PETER WElL and KLAUS PETER SCHAAL Institut fur Medizinische Mikrobiologie und Immunologie, Universitat Bonn (Direktor: Prof. Dr. K. P. Schaal), D-5300 Bonn

With 3 Figures· Received April 20, 1990· Accepted May 5, 1990

Summary Sera of persons without known actinomycotic infection (n = 153) were tested for antibodies reacting with antigenic preparations of different Actinomyces spp. and Nocardia spp. By using an enzyme immunoassay, 16% of all of the sera analyzed reacted significantly with antigens of A. viscosus and 2% those of with A. naeslundii. The antigens detected by these antibodies were of low molecular mass (14-32 kDa) and showed a uniform reaction pattern in the immunoblot analysis. Multiple bands coccurred with a difference of approximately 2 kDa in size suggesting the presence of repetitive units. Analogous antibodies were not observed with A. israelii serovar 1 and serovar 2 antigens, which were only insignificantly bound by the sera. The antibodies demonstrated were most probably acquired during episodes of periodontal disease or gingivitis in which A. viscosus is etiologically involved. Antibodies against nocardial antigens were not demonstrated in significant proportions by enzyme immunoassay and immunoblot analysis. Zusammenfassung In der vorliegenden Arbeit wurden 153 Seren von Probanden ohne bekannte Aktinomyzeteninfektion auf Antikorper gegen humanpathogene Aktinomyzeten untersucht. Hierzu wurden Antigenpraparationen von je vier klinisch re!evanten Actinomyces spp. und Nocardia spp. hergestellt und in einer quantitativen enzymimmunologischen Bestimmungsmethode (EIA) und in der Immunoblotanalyse eingesetzt. 16% der untersuchten Seren waren im EIA mit A. viscosus deutlich positiv, 2% regierten deutlich mit A. naeslundiiAntigenen. Die positiven Seren zeigten im Immunoblot ein rege!magiges Bandenmuster mit niedermolekularen Substanzen (14-32 kDa). Die Unterschiede von jeweils 2 kDa in der molekularen Masse weisen auf eine Substanz hin, die aus repetitiven Einhe.iten dieser Groge besteht. Ein ahnliches Reaktionsmuster ergab sich auch gegenuber A. naeslundii. Analoge Antikorper waren mit A. israelii-Antigenen nicht nachweisbar. Die gegen A. viscosus gerichteten Antikorper sind wahrscheinlich bei fruher abgelaufenen Gingivitiden oder Periodon-

* Dedicated to Prof. Dr. Dr. h. c. G. Pulverer on the occasion of his 60th birthday.

Serum Antibodies to Pathogenic Actinomycetes in the Human Population

399

titiden entstanden, an deren Genese dieses Bakterium atiologisch beteiligt ist. Antikorper gegen Nocardia spp. wurden weder in der Immunoblotanalyse noch im EIA in signifikanter Hohe nachgewiesen. Introduction Traditionally, the microbiological diagnosis of actinomycotic infections in man has mainly been based upon special cultural techniques with subsequent identification of the actinomycetes isolated (15). A serological test would be helpful in supporting the classical bacteriological methods. Several approaches were described in the past (2, 3, 17), but their diagnostic potential was questioned because of unspecificity and comparatively low sensitivity. Holmberg et al. (7) demonstrated an antibody response in patients with infections by Actinomyces israelii. The precipitin test had a sensitivity of 86% and sera of healthy subjects did not react. Nevertheless, information on the cross reactivity of the antibodies with other actinomycetes was not given. In addition reports on the high prevalence of "natural" antibodies directed against A. viscosus (10, 11), a common member of the indigenous oral microflora, in sera of apparently healthy persons may complicate a specific serological diagnosis. Promising studies on antibody response and its use for diagnosing systemic nocardiosis were presented by Sugar et al. (18) and by Angeles et al. (1). They isolated a 55 kDa protein from the culture filtrate of Nocardia asteroides, with which the serum of patients with culture-proven nocardiosis specifically reacted; but only insufficient data on the reactivity of sera of healthy controls were given. In this investigation we analyzed 153 sera of persons without apparent actinomycotic infection by an enzyme immunoassay and an immunoblotting technique. Antigens were prepared from four Actinomyces spp. and four Nocardia spp., all potentially involved in human infectious diseases. Materials and Methods Sera. Sera of 153 persons of both sexes without known actinomycotic infection were analyzed. The age of the persons tested ranged from 14 to 87 years, with a mean age of 46 years. Strains. The following strains were used for antigen preparation:

A. israelii serovar 1 A. israelii serovar 2 A. viscosus A. naeslundii N. farcinica N. asteroides N. brasiliensis N. otitidiscaviarum

A 34 A 19 A 84 A 336 D 455 D 444 D 133 D 254

IMMIB IMMIB IMMIB IMMIB IMMIB IMMIB IMMIB IMMIB

All strains were taken from the collection of actinomycetes at the Institute of Medical Microbiology and Immunology, University of Bonn (IMMIB). Antigen preparation. Each of the Actinomyces spp. was streaked on ten brain heart infusion agar plates and incubated under anaerobic conditions for seven days at 36°C. The cells were harvested by scraping the colonies from the plates and were washed thrice in icecold phosphate-buffered saline (PBS). 27 Zbl. Bakt. 274/3

400

H.-P. Weil and K. P. Schaal

A 500 ml culture with brain heart infusion medium was inoculated with each of the Nocardia spp. tested and was incubated for four days at 36°C in a rotary shaker (120 rpm). The cells were harvested by centrifugation (5000 x g, 10 min) and washed thrice in ice-cold PBS. All cell preparations were further processed in the same way. Cells were disrupted in a MSK homogenizer (Braun, Melsungen) for 5 min with extensive cooling. Unbroken cells and cell debris were removed by centrifugation (7500 x g, 15 min). The supernatants were dialyzed against 50 mM NH4 HC0 3 overnight at 4°C, lyophilized and used as antigen preparations. Enzyme immunoassay. The following method was used to measure specific IgG activities in the serum samples. Briefly, microtiter plates (type 29A, Dynatech) were incubated with 100 !ll antigen preparation (100 !lg/mI3.5 mM NaHC03/Na2C03, pH 9.6) per well overnight at 4°C. The plates were washed thrice with PBS-Tween 20 (0.5%, v/v). The diluted sera (1 : 200, v/v) were incubated for 1 h at room temperature and specific IgG was detected by rabbit-anti-human-IgG-antiserum coupled to horse-radish peroxidase (1: 1000, v/v, Sigma). All dilutions were done with PBS-bovine serum albumin (BSA, 0.5%, w/v, Sigma). As enzyme substrate served 2,2-azino-di-(3-ethyl-benzthiazoline sulfonic acid-6) diammonium salt (ABTS, Boerhinger) as described in (19). After 30 min of incubation the optical density was determined at 405 nm with a Titertek Multiscan spectrophotometer (Flow). Immunoblot analysis. The PAGE was performed as described by Laemmli (9), 12% gels were used throughout the experiments. After electrophoresis the separated antigens were transferred onto nitrocellulose membranes (0.2 !lm pore size, Schleicher and Schull) within 1 hand 100 V at 4°C. These were blocked with skim milk (1 %, w/v) overnight at 4°C. Diluted serum samples (1 : 200, v/v) were incubated for 1 h at room temperature. Protein A coupled to horse-radish peroxidase (1 : 500, v/v, Sigma) was used for the detection of bound IgG antibodies. IgM-ciass antibodies were demonstrated by incubation with a rabbit-antihuman-IgM-peroxidase conjugate (1: 500, v/v, Sigma). The blots were finally stained with o-dianisidine-tetrahydrochloride (0.5 mg/ml PBS) and H 20 2.

Results Antibodies reactive to Actinomyces spp. could be demonstrated by the enzyme immunoassay (Fig. 1). Using a cut-off value of ~E30 min = 0.300, 15% (n = 23) of the sera tested had elevated antibody levels against A. viscosus and 2 % (n = 3) against A. naeslundii, whereas no significantly elevated antibody levels were seen with both A. israelii antigens. Sera with high levels of antibodies against A. viscosus did also react with A. naeslundii and to a lesser extent with A. israelii. There was no correlation of the positive sera with both sex or age. These sera were further analyzed by the immunoblot method. A strong and typical reaction was always seen with A. viscosus: multiple bands occurring in the molecular mass range of 14-32 kDa with 2 kDa difference in size were detected (Fig. 2). These bands were also present in the blots with A. naeslundii except for one serum (4075, Fig. 3). This typical staining pattern was never seen with A. israelii, but cross reactions did occur in the high molecular mass range. Besides the reactivity to the low molecular mass substances no constant binding to certain antigens could be observed within the Actinomyces spp. analyzed. Antibodies of the IgM-class against the low molecular mass antigens of A. viscosus and A. naeslundii were not demonstrated by the immunoblot assay, essentially no IgMactivity was seen to any Actinomyces sp.

Serum Antibodies to Pathogenic Actinomycetes in the Human Population Aelm. 'lrollii Sirovor 2

I.e 1m. i.rOllii Sirovor 1 n 1t.Q 120

140

89%

120

~

100

100

80

80

60

60

~

40

401

40

20

11%

20

A

0·01

I

02-03

0

"E30mln

0-01 01-02

6E30mln

01 -0 2

Actm

Actm na.slundli

VISCOSUS

UO

140

120

120

100

100

77%

80

80 1.1%

60

60

300;,

40

40

18%

11. 1/.

20

0'01

I

02-0 J

01-02

20

I

04-05

OJ-04

I

05-06

0 "EJOm,"

3% 0-0.1

I

2%

Q2-0.3

01-0.2

I

6E30 mi.

03-04

Fig. 1. Results of the enyzme immunoassay with antigen preparations of Actinomyces spp. and sera of the normal human population (n = 153).

Significantly elevated antibody levels against Nocardia spp. could not be observed with the enzyme immunoassay (Fig. 3). Only a small percentage (3-5%) had a slight increase of reactivity. There were only faint bands without any characteristic pattern on the immuno-blotted membranes when these potentially positive sera were probed, so that unspecific reactions are most probable. Discussion Elevated levels of antibodies directed against A. viscosus in sera of healthy individuals were observed in earlier investigations (11). Up to 30% of the sera tested contained antibodies when analyzed by immunoprecipitation methods. An explanation for the

402

H.-P. Wei! and K. P. Schaal

occurrence of these antibodies is the sensitization of these apparently healthy persons during episodes of periodontal disease or gingivitis in which A. viscosus is suspected to be one of the causative microorganisms (13). We could demonstrate high amounts of antibodies in 15% of all of the sera analyzed, but inclusion of sera with moderately elevated levels (cut-off-value: LlE30 min = 0.200) against A. viscosus even resulted in a percentage of 29% positive sera, which is in good agreement with other studies. The demonstration of these antibodies understandably depends upon the sensitivity of the test system applied. Even in some sera with background antibody activity in the enzyme immunoassay we could observe weak but typical reactions with the low molecular mass substances. Thus, IgG antibodies to A. viscosus and A. naeslundii can frequently be demonstrated and are of no specific diagnostic value, in that they are most probably due to earlier infections of the peridontium or of other oral tissue. However, these antibodies may exert protective activities in the development of caries, periodontal disease and/or gingivitis (5). The determination of specific IgM-antibodies, which we could not detect in our serum samples may help to establish a serological test for A. viscosus and A. naeslundii infections. The chemical nature of the low molecular mass antigens detected by immunoblot analysis has not been clarified in this study. But the typical size distribution with 2 kDa differences suggests that these substances are fragments of a polymer with repetitive units as present in the cell walls of Gram-positive bacteria. We could not stain this substances with Coomassie blue in the gel, so that a non-proteinaceous antigen is probable. Levine and Movafagh (10) isolated a substance of 2 kDa from a cell wall preparation of A. viscosus ATCC 19246 which inhibited the precipitation with positive human sera. This substance contained several amino acids including ornithine but no lysine, which is present in the peptide subunit of the peptidogylcan of A. viscosus (16).

Fig. 2. Immunoblot analysis of three sera which were significantly positive in the enzyme immunoassay. Antigen preparations of the following Actinomyces spp. were analyzed: A. naeslundii (lane 1), A. viscosus (lane 2), A. israelii serovar 1 (lane 3) and A. israelii serovar 2 (lane 4).

Serum Antibodies to Pathogenic Actinomycetes in the Human Population t~o:::

Noe otilldl!:tCQ\llcrum

brasJ!I~I-j5.IS

96°"

95% 140

HO

120

120

100

100

80

eo

60

60

40

1.0

20

403

20

5% 0-01 01-0.2

1.% 0-0.1 01-0.2

.o. E 30mlfl

Noe a sterOl des

AEJOmin

Noe. f arc Inlea

97 %

91 % 11.0

120

120

100

100

80

BO

60

60

40

1.0

20

20

3% 0-01 01-02

3%

l:J.E30rriin

0

0-0.1 01-02

AEJOmin

Fig_ 3. Results of the enzyme immunoassay with antigen preparations of Nocardia spp. and sera of the normal human population (n = 153).

In our analysis the antibodies discriminate between A. viscosuslA. naeslundii and A. israelii serovar 1 and 2 although these three species have the same peptide subunit. Comparing the cell wall sugars, A. israelii contains only galactose whereas A. visocus and A. naeslundii both possess glucose, rhamnose, and 6-deoxytalose as characteristic components (16). Therefore, if cell wall components are the antigens detected by the human sera the carbohydrate moiety will be essential for antibody specificity. We cannot decide from our experiments if other cell wall components serve as the low molecular mass antigens. Lipoteichoic acids or analogous compounds are good candidates because they possess potent antigenic activity (8,21). In the sera of patients with acute gingivitis antibodies directed against lipoteichoic acids from Streptococcus mutans and Staphylococcus aureus were detected (21). The cell walls of actinomycetes do not contain lipoteichoic acids (4, 6), but do express a fatty-acid substituted heteropolysaccharide which cannot be distinguished serologically between the three

404

H.-P. Weil and K. P. Schaal

actinomycete species investigated (22), and is therefore most probably not the reacting substance. These results appear to indicate that the lack of cross reactivity with A. israelii should allow a serological diagnosis of infections by this actinomycete, although high titres of antibodies against other actinomycetes of the oral microflora are frequently present in sera of healthy persons. The analysis of antibody reactivity against nocardial antigens did show, that no relevant antibodies are present in the sera tested. In 2-5% of the samples slight elevations were measured, but these sera had significant titres against A. viscosus, so that minor cross-reactions may be responsible. We conclude from these results, that the diagnostic value of both enzyme immunoassay and immunoblot test are not adversely affected by the presence of frequently occurring, cross reacting antibody activity in sera of non-infected persons. Both methods will be applied to sera of patients with suspected and culture-proven actinomycotic infections in order to determine their suitability for serologically diagnosing invasive actinomycete infections.

Acknowledgements. We thank Dagmar Gierth and Monika Busch for expert technical assistance.

References

1. Angeles, A. M. and A. M. Sugar: Identification of a common immunodominant protein in culture filtrates of three Nocardia species and use in etiologic diagnosis of mycetoma. J. Clin. Microbiol. 25 (1987) 2278-2280 2. Blenner, S. O. and C. Kaufman: Microimmunodiffusion test for nocardiosis. J. Clin. Microbiol. 10 (1979) 308-312 3. Boiron, P. and F. Provost: Enzyme immunoassay on whole Nocardia asteroides cells for human nocardiosis. Serodiagn. Immunother. Infect. Dis. 2 (1988) 445-452 4. Fischer, W.: Physiology of lipoteichoic acids in bacteria. Adv. Microb. Physiol29 (1988) 244-302 5. Haber, ]. and C. Grinnell: Analysis of the serum antibody responses to type 1 and 2 fimbriae in mice immunized with Actinomyces viscosus T14V. J. Periodont. Dis. 24 (1989) 81-87 6. Hamada, S., S. Tai, and H. D. Slade: Selective adsorption of heterophile polyglycerophosphate antigen from antigen extracts of Streptococcus mutans and other Grampositive bacteria. Infect. Immun. 14 (1976) 903-910 7. Holmberg, K., C. E. Nord, and T. Wadstrom: Serological studies of Actinomyces israelii: Taxonomic and diagnostic applications. Infect. Immun. 12 (1975) 398-403 8. Knox, K. W. and A.]. Wicken: Immunological properties of teichoic acids. Bact. Rev. 37 (1973) 215-257 9. Laemmli, U. K.: Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature (Lond.) 227 (1970) 680-685 10. Levine, M. and B. F. Movafagh: Analysis of the specificity of natural human antibody reactive to Actinomyces. Molec. Immunol. 23 (1986) 255-261 11. Levine, M., ]. A. Stober, and D. E. Parker: Human serum precipitins to oral bacteria related to dental caries. Arch. Oral BioI. 29 (1984) 191-194 12. Monefe/dt, K., T. Tollefsen, S. Assev, and G. Rolla: Increased serum IgG antibodies reactive with lipoteichoic acids in subjects with gingivitis. J. Periodont. Res. 22 (1987) 114-118

Serum Antibodies ro Pathogenic Actinomycetes in the Human Population

405

13. Moore, W. E. c.: Microbiology of periodontal disease. J. Periodont. Res. 22 (1987) 335-341 14. Pekovic, D. D. and E. D. Fillery: Identification of bacteria in immunopathological mechanisms of human periodontal diseases. ]. Periodont. Res. 19 (1984) 329-351 15. Schaal, K. P.: Laboratory diagnosis of actinomycete diseases. In: The Biology of the Actinomycetes (M. Goodfellow, M. Modarski, and S. T. Williams, eds.), pp. 425--456. Academic Press, London-New York-San Francisco (1984) 16. Schaal, K. P.: Genus Actinomyces Harz, 1877. In: Bergey's Manual of Systematic Bacteriology, Vol. 2., 8Th ed. (l'. H. A. Sneath, N. S. Mair, M. E. Sharpe, and J. G. Holt, eds.), pp. 1383-1418. Williams and Wilkins, Baltimore (1974) 17. Shain house, .f. Z., A. C. Pier, Jnd D. A. Stevens: Complement fixation test antibody for human nocardiosis. J. Clin. Microbiol. 8 (1978) 516-519 18. Sugar, A. M., G. K. Schoo/mk, and D. A. Stevens: Antibody response in human nocardiosis: identificatIOn of two immunodominant culture-filtrate antigens derived from Nocardia asteroides. J. Infect. Dis. 151 (1985) 895-901 19. Voller, A. and D. Bidwell: Enzyme-linked immunosorbent assay. In: Manual of Clinical Laboratory Immunology, yd ed. N. R. Rose, H. Friedman, and J. L. Fahey, eds.), pp. 99-109. American Society for Microbiology, WashingtonlDC (1986) 21. Wergeland, H. 1., 1. R. Haaheim, O. B. Natas, F. Wesenberg, and P. Oeding: Antibodies to staphylococcal peptidoglycan and its peptide epitopes, teichoic acid, and lipoteichoic acid III sera from blood donors and patients with staphylococcal infections. J. Clin. Microbiol. 27 (1989) 1286-1291 22. Wick en, A . .f., K. W. Broadv,.f. Evans, and K. W. Knox: New cellular and extracellular amphipathic antigen from Actinomvces viscosus NY1. Infect. Immun. 22 (1978) 615-616 Hans-Peter Weil, Professor Dr. Klaus Peter Schaal, Institut fur Med. Mikrobiologie und Immunologie der Universitiit Bonn, Sigmund-Freud-Str. 25, D-5300 Bonn 1