Medical Centre, Bedford Park, South Australia 5042. Received 6 June 1983/Accepted ..... Onderdonk, A. B., D. L. Kasper, R. L. Cisneros, and J. G.. Bartlett. 1977.
INFECTION AND IMMUNITY, Sept. 1983, p. 1373-1375 0019-9567/83/091373-03$02.00/0
Vol. 41, No. 3
Bactericidal Activity of a Granule Extract from Human Polymorphonuclear Leukocytes Against Bacteroides Species HENDRIK PRUUL,* BRUCE L. WETHERALL, AND PETER J. McDONALD
Unit of Clinical Microbiology, School of Medicine, Flinders University of South Australia and Flinders Medical Centre, Bedford Park, South Australia 5042
Received 6 June 1983/Accepted 22 June 1983
The microbicidal activity of an acetate extract of human polymorphonuclear leukocyte granules was tested against Bacteroides fragilis, Bacteroides vulgatus, Bacteroides distasonis and Bacteroides thetaiotaomicron. All strains tested were killed by the extract, and there were no significant differences between the different Bacteroides species.
The importance of Bacteroides species in a blood of healthy donors by the method of wide range of serious infections is now unequiv- B0yum, modified for large volumes (4). The final ocally established, and the prominence of Bac- cell suspension contained 95% PMN. The cells teroides fragilis as a cause of suppuration and were suspended in ice-cold sucrose (0.34 M) and sepsis suggests that this species of Bacteroides homogenized to greater than 85% disruption in a possesses unique virulence factors. However, Teflon glass homogenizer. The granule fraction the attributes of virulence constitute a set of was obtained by differential centrifugation, and complex issues; and very little is known of the an acetate extract was prepared from the granmechanisms by which obligate anaerobes are ules as described by Rest et al. (21). A 105-PMN dealt with by polymorphonuclear leukocytes amount yielded ca. 1 ,ug of granule protein. (PMN). PMNs maintain a formidable defense Portions (0.4 ml) were stored at -70°C and system against bacterial invaders. After phago- thawed immediately before use. Bacteroides strains were isolated from clinical cytosis, bacteria are exposed to potent antimicrobial mechanisms. These mechanisms are material, and identification to the species level broadly divided into two categories: the oxygen- was carried out according to established criteria dependent microbicidal systems and those not (7). The strains were maintained in an anaerobic requiring oxygen for activity (13). Within an glove box on blood agar and subcultured every 6 abscess, low oxygen tension restricts killing by to 8 weeks. For several strains, multiple subculPMN to the latter mechanisms; this may be the ture was avoided by storing primary isolates in reason for reduced killing of certain facultative skim milk at -70°C. Bactericidal assays were bacteria under anaerobic conditions (8, 15, 16, carried out in sterile Linbro tissue culture trays 17, 22, 26). However, the killing of anaerobic (Flow Laboratories, Inc., Hamden, Conn.). The bacteria, including Bacteroides species, is not bacteria were grown to the logarithmic phase in Schaedler broth and diluted through anaerobic impaired (3, 8, 15). The purpose of the present study was to saline solution; 2 x 104 to 4 x 104 bacteria were assess the effectiveness of the PMN oxygen- inoculated into 0.1 ml of aerobic granule extract independent bactericidal activity against several dilutions. These dilutions were prepared in trypspecies of Bacteroides and, as animal-passaged tone-saline solution, pH 5.0. The tissue culture B. fragilis has been reported to be more resistant trays were transferred into the anaerobic glove to opsonophagocytic killing than laboratory-pas- box and incubated for 2 h at 37°C. After incubasaged strains (23), to determine whether this tion, duplicate 20-,ul portions were spread onto finding could be explained by increased resist- blood agar plates. This procedure was carried ance of B. fragilis to intracellular killing mecha- out aerobically. The plates were incubated annisms. aerobically for 48 h, and the number of CFU The use of a neutrophil granule extract permit- (viable bacteria) were determined using an autoted direct study of the interaction of bacteria and matic colony counter (model 880; Artek Systems oxygen-independent microbicidal systems, with- Corp., Farmingdale, N.Y.). In some experiout interference of respiratory burst products ments, the time of exposure of bacteria to oxyand complexities inherent in the interpretation gen was minimized by carrying out all proceof phagocytic studies. dures under strictly anaerobic conditions, PMN were obtained from citrated venous
except for the final transfer onto blood agar
1373
1374
NOTES
INFECT. IMMUN.
TABLE 1. Killing of Bacteroides species by granule extract Species
Number Mean Mean of LDa LDa strains L~aL~
LD90/LD50 ratio ai
1.76 5.6 9.9 25 B. fragilis 1.93 4.7 9.1 5 B. vulgatus 1.71 6.4 11 7 B. distasonis 1.94 5.2 10.1 8 B. thetaiotaomicron a Granule extract (,ug extract protein/ml) required to kill 50% or 90% of the bacterial inoculum (2 x 104 to 4 x 104 bacteria) after 2 h of incubation. Four to eight determinations were carried out for each strain.
these strains produced capsules after growth in Schaedler broth, as detected by the India ink wet-mount technique. The presence of capsular polysaccharide has been associated with the increased virulence of B. fragilis (18, 19). The great majority of B. fragilis strains are capable of synthesizing a capsular polysaccharide, whereas non-B. fragilis species possess such a capsule less frequently (1, 11, 14). The results of this study indicate that encapsulation of B. fragilis may not confer upon this species increased resistance to oxygen-independent killing mechanisms of human PMN. Facultative gram-negative bacteria are also killed by granule extracts (6, 21, 24, 25). Their susceptibility to oxygen-independent killing mechanisms of PMN varies greatly. The LD50 of clinical isolates of Escherichia coli range from 5 to 100 ,ug/ml (unpublished data), whereas the range for B. fragilis, given the same granule extract, is 3.3 to 8.7 ,ug/ml. This variation among facultative gram-negative strains of bacteria has been attributed to the structural organization of the bacterial cell surface, in particular, the lipopolysaccharide chemotype (6, 21, 24), and to differences in binding of the bactericidal proteins to the cell surface (25). With the granule extract as a probe for structural variability, the data suggests that there may be little variation among clinical isolates of Bacteroides, both between strains within species and between different species, in cell wall structures associated with conferring resistance to intracellular oxygen-dependent killing mechanisms. The importance of the oxygen-dependent bactericidal system of PMNs is well established (2, 5, 13). Chronic granulomatous disease is a clinical disorder of this bactericidal system in which
plates. Because similar results were obtained under either condition, the data presented are those obtained from experiments with less control of oxygen exposure. Between four and eight determinations of the 50% lethal dose (LD50) and the 90% lethal dose (LD%0) of the extract against each strain were carried out. The mean values for each species are given in Table 1. All strains were killed by the extract. The LD50 and LD90 values for the species were not significantly different (P > 0.05). To determine whether the bacterial populations were uniformly sensitive to the granule extract, the LDgaOLD5o ratio was calculated. A value of 1.8 indicates total uniformity of sensitivity between the first 50% and the next 40% of the population. The calculated values for the LD90/LD50 ratio given in Table 1 indicate that the four species of Bacteroides all consist of uniformly sensitive populations of bacteria, with no large subpopulations of bacteria which have increased resistance. An analysis of variance in LD50 values was carried out, both between species and within each species, and the results are given in Table 2. The within-species analysis demonstrates that TABLE 2. An analysis of variance in LD50 of individual strains differ significantly from each Bacteroides species other (p < 0.001). However, the between-speMean Variance cies analysis shows that the mean values for the p na squares ratio Species four species were not significantly different (p > of LD50rti 0.05). This indicates that the variation in sensitivity to the granule extract between the strains Within-species comparison of any one species of Bacteroides is as large as
