Selected Characteristics of Pathogenic and Nonpathogenic Strains of ...

Vol. 25, No. 4

JOURNAL OF CLINICAL MICROBIOLOGY, Apr. 1987, p. 738-740 0095-1137/87/040738-03$02.00/0 Copyright © 1987, American Society for Microbiology

Selected Characteristics of Pathogenic and Nonpathogenic Strains of Bacteroides gingivalis DANIEL GRENIERt AND DENIS MAYRAND* Groupe de Recherche en Ecologie Buccale, Université Laval, Quebec, GIK 7P4 Quebec, Canada Received 11 August 1986/Accepted 19 December 1986

Strains of Bacteroides gingivalis were compared for the presence of properties associated with pathogenicity. Some strains were infectious in pure culture in an in vivo model (guinea pig), and all but one of these were more collagenolytic than those which failed to cause lesions in guinea pigs. However, other factors seem to be necessary for the induction of an infection in this animal model.

The collagenolytic activity of the strains was determined by the method of Gisslow and McBride (3). The collagenase assay consisted of mixing 100 ,uI of bacteria (ca. 2 x 109 cells) with 200 puI of ['4C]collagen solution plus 100 RI of buffer (5 mM CaCl2, 50 mM Tris [pH 7.2]) and 100 ,uI of cysteine (5 mM in buffer). The assays were done in microtubes incubated at 37 or 25°C. To limit the action of nonspecific proteases on the labeled collagen, the incubation period was limited to 2 h (8). At the end of the incubation period, 100 ,uI of the assay mixture was transferred to a second microtube containing 50 puI of HCl (2 N) and 50 pul of phosphotungstic acid (0.04 N). The microtubes were agitated with a vortex mixer and left at room temperature for 10 min before centrifugation. The supernatant (100 ptI) was counted by liquid scintillation spectrometry. Nonspecific protease activity found in culture supernatants (4-day culture) was determined by measuring activity against Azocoll (Sigma Chemical Co.) as already described (9). The stability of the nonspecific proteolytic activity associated with the supernatant was evaluated by using a 13-day culture or by heating at 55°C for 30 min (7). The cytotoxic activity of B. gingivalis culture supernatants was measured against Vero cells (5). The eucaryotic cells were grown for 2 days, and the medium was replaced by bacterial filtrates which had their pH adjusted to 7.2. The plates were incubated for 2 days in air containing 5% C02 at 37°C and 98% humidity. After the incubation period, the liquid was decanted, and the Vero cells were fixed with methanol and subsequently stained with 10% Giemsa. Cytotoxic activity of culture supernatants was also measured against guinea pig leukocytes by the methods described by Gadeberg and 0rskov (2). Polyacrylamide gel electrophoresis of soluble proteins was done by the methods and procedures described by Moore et al. (11); this technique allows the detection of only major differences between the bacterial strains. Virulence assays showed that 6 of 14 strains (W83, BH18/10, 22B4, RB46D-1, RB24M-2, and RB22D-1) were pathogenic in pure culture (Table 2). Att the infections were transmissible to a second animal. The minimum number of viable bacteria required to cause the infections ranged from 4 x 108 to 7 x 109/ml (Table 2). The induced lesions were similar for the six pathogenic strains and were characterized by a necrotic abscess of up to 20 ml of exudate containing altered erythrocytes, some leukocytes, and large quantities of bacteria. The other B. gingivalis strains always gave negative results (or, at most, a slight redness at the injection site), even when inoculated in great numbers. Strain 19A4,

Gram-negative anaerobic bacteria may play an important role in the development of periodontal diseases. Bacteroides gingivalis has been associated with adult chronic periodontitis (21, 22). Several studies have described the role this species plays in various types of experimental infections (4, 9, 17). Indeed, the inclusion of B. gingivalis in several different mixed-culture inocula resulted in the development of necrotic lesions, organisms from which were transmissible. In contrast, omission of B. gingivalis from the mixedculture inoculum failed to induce lesions. On the other hand, a limited number of B. gingivalis strains have been pathogenic when injected into guinea pigs or mice in pure culture (6, 18). The pathogenic potential of this species may result from the presence of a variety of virulence factors (16). The goal of our study was to determine whether it was possible to detect differences among some putative virulence factors diversely exhibited by B. gingivalis strains. A total of 14 human strains of B. gingivalis were tested (Table 1). All were grown in a Trypticase (BBL Microbiology Systems)-yeast extract medium described by Sawyer et al. (15); the requirement for the two growth factors (hemin and vitamin K1) was determined. In all cases, the cultures were grown in an anaerobic chamber at 37°C to late exponential phase, concentrated by centrifugation (10,000 x g, 15 min), and suspended in phosphate-buffered saline supplemented with sodium thioglycolate (0.05%) to a concentration corresponding, after a dilution of 1:10, to a McFarland no. 10 standard. At this time, viable counts were done on ToddHewitt agar supplemented with hemin and vitamin K1. Dilutions of this bacterial suspension were made in the modified phosphate-buffered saline, and 0.5 ml was then injected into the groin of each of three Hartley guinea pigs (weight, 180 to 220 g). The animals were examined daily for 2 weeks, and the lesions were inspected and characterized as follows: -, no infection; +, localized abscess, less than 2 cm in diameter; + +, localized abscess, 2 cm or more in diameter; + + +, necrotic abscess, not fatal; + + + +, massive infection resulting in the death of the animal in less than 3 days. Only strains which exhibited + + or greater reaction were considered virulent. We verified that abscess fluid could induce similar lesions by injecting material aspirated from a lesion into a second animal. Guinea pigs injected with killed B. gingivalis or uninoculated medium were used as controls. *

Corresponding author.

t Present address: Department of Microbiology, University of British Columbia, Vancouver, British Columbia V6T 1Z7, Canada.

738

VOL. 25, 1987

NOTES

TABLE 1. B. gingivalis strains and sources No. of Strain a

W83 BH18/10 33277 381 6/26 23A4 22B4

19A4 LB13D-3 RB22D-1, RB24M-2, RB46D-1 HW11D-5, HW24D-1

Source

in vitro transfers

Clinical specimen Human periodontal pocket Human gingival sulcus Human periodontal pocket Human periodontal pocket Human periodontal pocket Human gingival sulcus Human periodontal pocket Human periodontal pocket Human periodontal pocket

>30 >30 >30 >30 >30 15 15 15

Human periodontal pocket

5

S 5

Strains listed together were from the same patient. Approximate number of transfers on solid media from isolation to injection. a

b

for example, was injected at a concentration which was 200 times greater than that of strain BH18/10 without detectable effect. The relative collagenolytic and proteolytic activities of all strains are also shown in Table 2. Except for 22B4, the pathogenic strains had a higher collagenolytic activity than the nonvirulent strains. The nonspecific proteolytic activity of certain avirulent strains was also higher than that found for the pathogenic strains, but no correlation could be established between protease activity and infectivity. In contrast with the results obtained for Bacteroides nodosus (7), it was impossible to distinguish the pathogenic strains from the nonpathogenic strains of B. gingivalis in terms of the stability of the proteolytic enzymes. Cytotoxicity experiments revealed that all B. gingivalis strains had the same effect on Vero cells. The affected cells exhibited long filamentous tendrils and a thickening of the cellular membranes. In extreme cases, the cells were rounded and often free in the medium. On the other hand, culture supernatants of all

TABLE 2. Characteristics of B. gingivalis strains Requirement

Strain

Concna

RB22D-1 RB24M-2 RB46D-1 W83 BH18/10 22B4 381 HW24D-1

x x x x x

4.0 6.5 5.0 4.5 5.0 7.0 7.0 4.0 19A4 1.0 1.6 33277 1.2 23A4 6/26 7.2 LB13D-3 1.5 HW11D-5 1.0

x

108 + + + + 108 + + + + 108 + + + + 109 + + + + 108 109

x 1010 x 1010

x 1011 x 1010 x 1010 x x x

Collagen- ProteoVitamin olytic lytic Hemmn K1 activity activity

Infectivity H

1010 1010 1010

+++ +++ + + -

-

+ + + + + + + + + + +

-

+ + +

+ + +

+ + + + +

-

42

49 42 45 51 28 18 19 20 28 14 22 30 17

1.16 0.47 0.38 0.97 0.81 1.30 1.13 1.14 1.34 1.33 0.57

0.91 0.61 1.35

a Minimum number of cells needed to infect in the case of a necrotic infection or maximum number of cells tested in the case of negative or light abscess formation. b Expressed as percentage of total radioactivity added in the assay (2 h); bacterial cells from early stationary phase. Under identical conditions, commercial collagenase degraded 55% of the labeled substrate in 2 h. C Optical density at 520 nm (4 h); supernatants from a 4-day culture.

739

strains were not toxic to leukocytes as measured by trypan blue exclusion. Nutritional studies revealed no perfect correlation between the requirement for hemin or vitamin K1 and virulence. The requirement for vitamin K1 differs within the species B. gingivalis. The growth of pathogenic strain W83 was independent of the presence of vitamin K1. Strain W50, used in the study of McKee et al. (10), is also infective in pure culture and does not require vitamin K1. Analysis of polyacrylamide (8.5%) gel electrophoresis patterns indicated that, except for minor variations in the median section of the gel, the protein profiles were similar. No correlation for the presence of bands and the ability to cause monoinfection in the guinea pig could be established. B. gingivalis is frequently isolated from patients with advanced chronic periodontitis. It seems that the pathogenicity of this species varies from strain to strain. In fact, some strains can infect an animal model when injected in pure culture, whereas other B. gingivalis strains require the presence of helper species to produce an infection. In this study, several strains were highly infectious upon pureculture inoculation. The pathogenicity of these strains was not related to the source of isolation, because some virulent strains have been isolated from infected sites and some have been isolated from healthy periodontal sites. We also found that the strains from different sites of the same subject were similar with respect to virulence, as well as vitamin K1 requirement. These observations are in accordance with the study of Notten et al. (13). Their findings indicated that in the mouth of an individual, one B. gingivalis antibiotype predominates and that different patients harbor different B. gingivalis antibiotypes. The number of in vitro transfers did not seem to influence the pathogenicity of strains, suggesting the stability of the virulence factor(s). Kastelein et al. (6) have already demonstrated the pathogenicity of strain W83 in guinea pigs. Our results confirm theirs in that it took at least 109 CFU/ml to infect the animal. Strain BH18/10, which has already been shown to exhibit a large quantity of extracellular polysaccharide material (G. H. Bowden and A. H. Holthuis, J. Dent. Res. 62:179, abstr. no. 89, 1983), and the freshly isolated RB46D-1 and RB22D-1 were the three strains which produced infection with the lowest number of CFU per milliliter. The virulence of certain microorganisms can be related to their proteolytic activity; for example, Entamoeba histolytica (12) and Aeromonas salmonicida (14). The importance of proteolytic activity in black-pigmented Bacteroides species was recently emphasized. van Steenbergen and de Graaff (19), using different species, concluded that organisms which had the lowest proteolytic activity were also the least virulent in an animal model, whereas B. gingivalis (highest proteolytic activity) was the most virulent of blackpigmented Bacteroides species. Another study (M. E. Neiders, P. Chen, H. S. Reynolds, H. Suido, J. J. Zambon, and R. J. Genco, J. Dent. Res. 65:208, abstr. no. 351, 1986) showed that some B. gingivalis strains which were lethal in a mouse model exhibited a higher capacity to degrade synthetic and native substrates. Our results indicate that B. gingivalis strains can be separated into two groups. Pathogenic strains (except 22B4) demonstrated high collagenolytic and proteolytic activities, whereas nonpathogenic strains showed lower collagenolytic activity but a high proteolytic activity. McKee et al. (10) demonstrated that the virulence of B. gingivalis W50 was closely related to the availability of hemin. A strong proteolytic activity could thus allow the

740

NOTES

release of heme from iron-transporting plasma proteins (1) and subsequently favor the development of the infection. Other factors could be responsible for the initiation of an experimental infection. van Steenbergen et al. (20) showed that strain W83 was more resistant to phagocytosis than were nonvirulent strains. More recently, the presence of two serogroups within B. gingivalis was proposed (J. G. Fisher, J. J. Zambon, P. Chen, and R. J. Genco, J. Dent. Res. 65:816, abstr. no. 817, 1986). Serogroup A strains were less virulent than serogroup B strains, which were lethal in a mouse model. Our pathogenic strains would probably fall within serogroup B of Fisher et al. However, in contrast with our results, their pathogenic strains (serogroup B) were always isolated from patients with severe periodontitis, whereas their serogroup A strains (avirulent) were found in sites of healthy subjects. Our study thus confirmed the presence of two distinct groups within the B. gingivalis species: virulent and avirulent strains. Although virulence factors responsible for pathogenicity may be produced to some degree by all B. gingivalis strains, the amount or level of activity of these factors found in vivo could determine whether a strain is pathogenic in pure culture. We thank Gene Bourgeau for his help in the preparation of the

manuscript. D.G. was supported by a studentship from the Fonds de la Recherche en Santé du Québec. This work was supported by the Medical Research Council of Canada. LITERATURE CITED 1. Carlsson, J., J. F. Hôfling, and G. K. Sundqvist. 1984. Degradation of albumin, haemopexin, haptoglobin and transferrin, by black-pigmented Bacteroides species. J. Med. Microbiol.

18:39-46. 2. Gadeberg, O. V., and I. 0rskov. 1984. In vitro cytotoxic effect of a-hemolytic Escherichia coli on human blood granulocytes. Infect. Immun. 45:255-260. 3. Gisslow, M. T., and B. C. McBride. 1975. A rapid sensitive collagenase assay. Anal. Biochem. 68:70-78. 4. Grenier, D., and D. Mayrand. 1983. Etudes d'infections mixtes anaérobies comportant Bacteroides gingivalis. Can. J. Microbiol. 29:612-618. 5. Grenier, D., and D. Mayrand. 1985. Cytotoxic effects of culture supernatants of oral bacteria and various organic acids of Vero cells. Can. J. Microbiol. 31:302-304. 6. Kastelein, P., T. J. M. van Steenbergen, J. M. Bras, and J. de Graaff. 1981. An experimentally induced phlegmonous abscess by a strain of Bacteroides gingivalis in guinea pigs and mice. Antonie van Leeuwenhoek J. Microbiol. 47:1-9. 7. Kortt, A. A., I. J. O'Donnell, D. J. Stewart, and B. L. Clark.

J. CLIN. MICROBIOL.

1982. Activities and partial purification of extracellular proteases of Bacteroides nodosus from virulent and benign

footrot. Aust. J. Biol. Sci. 35:481-489. 8. Mayrand, D., and D. Grenier. 1985. Detection of collagenase activity in oral bacteria. Can. J. Microbiol. 31:134-138. 9. Mayrand, D., and B. C. McBride. 1980. Ecological relationships of bacteria involved in a simple, mixed anaerobic infection. Infect. Immun. 27:44-50. 10. McKee, A. S., A. S. McDermid, A. Baskerville, A. B. Dowsett, D. C. Ellwood, and P. D. Marsh. 1986. Effect of hemin on the physiology and virulence of Bacteroides gingivalis W50. Infect. Immun. 52:349-355. 11. Moore, W. E. C., D. E. Hash, L. V. Holdeman, and E. P. Cato. 1980. Polyacrylamide slab gel electrophoresis of soluble proteins for studies of bacterial floras. Apple. Environ. Microbiol. 39:900-907. 12. Munoz, M. L., M. Rojkind, J. Calderon, M. Tanimoto, S. Arias-Negrete, and A. Martinez-Palomo. 1984. Entamoeba histolytica: collagenolytic activity and virulence. J. Protozool. 31:468-470. 13. Notten, F. J. W., F. H. M. Nieman, and F. H. M. Mikx. 1985. Antibiotypes of Bacteroides gingivalis assessed by antimicrobial disks and multivariate analysis. J. Clin. Microbiol. 22: 1020-1024. 14. Sakai, D. K. 1985. Loss of virulence in a protease-deficient mutant of Aeromonas salmonicida. Infect. Immun. 48:146152. 15. Sawyer, S. J., J. B. Macdonald, and R. J. Gibbons. 1962. Biochemical characteristics of Bacteroides melaninogenicus-a study of thirty-one strains. Arch. Oral Biol. 7:685-691. 16. Slots, J., and R. J. Genco. 1984. Black-pigmented Bacteroides species, Capnocytophaga species, and Actinobacillus actinomycetemcomitans in human periodontal disease: virulence factors in colonization, survival, and tissue destruction. J. Dent. Res. 63:412-421. 17. Socransky, S. S., and R. J. Gibbons. 1965. Required role of Bacteroides melaninogenicus in mixed anaerobic infections. J. Infect. Dis. 115:247-253. 18. Takazoe, I., M. Tanaka, and T. Homma. 1971. A pathogenic strain of Bacteroides melaninogenicus. Arch. Oral Biol. 16: 817-822. 19. van Steenbergen, T. J. M., and J. de Graaff. 1986. Proteolytic activity of black-pigmented Bacteroides strains. FEMS Microbiol. Lett. 33:219-222. 20. van Steenbergen, T. J. M., F. Namavar, and J. de Graaff. 1985. Chemiluminescence of human leukocytes by black-pigmented Bacteroides strains from dental plaque and other sites. J. Periodontal Res. 20:58-71. 21. White, D., and D. Mayrand. 1981. Association of oral Bacteroides with gingivitis and adult periodontitis. J. Periodontal Res. 16:259-265. 22. Zambon, J. J., H. S. Reynolds, and J. Slots. 1981. Blackpigmented Bacteroides spp. in the human oral cavity. Infect. Immun. 32:198-203.