Interaction between Human Polymorphonuclear Leukocytes and

85

Interaction between Human Polymorphonuclear Leukocytes and Streptococcus milleri Group Bacteria Anna Wanahita,1 Elizabeth A. Goldsmith,1 Daniel M. Musher,1,3,4 Jill E. Clarridge III,2,3,5 Jose Rubio,6 Bhuvaneswari Krishnan,2,5 and JoAnn Trial1,6

1

Medical and 2Laboratory Services, Veterans Affairs Medical Center, and Departments of 3Medicine, 4Molecular Virology and Microbiology, 5Pathology, and 6Immunology, Baylor College of Medicine, Houston, Texas

Because Streptococcus milleri group (SMG) bacteria—Streptococcus constellatus, Streptococcus intermedius, and Streptococcus anginosus—exhibit a striking propensity to cause abscesses, the interaction of these organisms with human polymorphonuclear leukocytes (PMNL) was examined. After incubation in pooled normal human serum, SMG stimulated less chemotaxis than did Staphylococcus aureus, in contrast to viridans streptococci, which caused greater chemotaxis than did S. aureus. PMNL ingested greater numbers of SMG and viridans streptococci than S. aureus but killed these organisms more slowly and less completely. Relative resistance to killing by PMNL is expected in organisms that cause abscesses, and inhibition of chemotaxis may contribute to pathogenicity, because delayed arrival of PMNL gives a head start to proliferating bacteria. This study helps explain the capacity of SMG to cause abscesses. It is unclear, however, why viridans streptococci, bacteria that rarely produce abscesses, share some of these same properties.

The Streptococcus milleri group (SMG) consists of 3 species of bacteria (Streptococcus constellatus, Streptococcus intermedius, and Streptococcus anginosus) that share a remarkable propensity to cause abscesses [1– 3]. We recently showed [3] that the great majority of clinical isolates of S. constellatus and S. intermedius are obtained from patients in whom these organisms have caused abscesses. S. intermedius tended to cause pleuropulmonary infections, central nervous system abscesses, and abscesses of deep soft tissues, whereas S. constellatus tended to cause odontogenic, soft tissue, pleuropulmonary, and intra-abdominal abscesses. A smaller proportion of S. anginosus isolates are associated with abscesses (generally involving soft tissue, the pleural space, and the peritoneal cavity), which reflects the more frequent isolation of this species from blood, infected soft tissues, and urine. We hypothesized that the remarkable capacity to cause abscesses might reflect a unique interaction between SMG organisms and human polymorphonuclear leukocytes (PMNL). We reasoned that their capacity to stimulate PMNL migration (chemotaxis) and/or to resist ingestion or killing by PMNL might be involved. Accordingly, we undertook the present study to investigate this interaction; we compared isolates of Staphylococcus aureus, a well-recognized abscess-forming bacteri-

Received 24 May 2001; revised 29 August 2001; electronically published 6 December 2001. Reprints or correspondence: Dr. Daniel Musher, Infectious Disease Section, Veterans Affairs Medical Center, 2002 Holcombe Blvd., Bldg. 10, Rm. 4B-370, Houston, TX 77030 ([email protected]). The Journal of Infectious Diseases 2002;185:85–90 q 2002 by the Infectious Diseases Society of America. All rights reserved. 0022-1899/2002/18501-0011$02.00

um, and viridans streptococci, a relatively avirulent bacterial group. Materials and Methods Bacteria

Twenty-two SMG isolates (7 S. constellatus, 7 S. intermedius, and 8 S. anginosus ) from different patients were selected to reflect the clinical spectrum of disease caused by these 3 species (table 1) [3]. These isolates had been stored at 2 70 C after their original isolation in the Microbiology Laboratory of the Veterans Affairs Medical Center of Houston. Assignment to SMG was based on results of the API 20 Strep system (BioMe´rieux Vitek), with further speciation by polymerase chain reaction amplification and sequence analysis of a segment of the 16S rRNA gene [4]. A convenience sample of 8 viridans streptococci isolates was obtained from sputum cultures of 8 individual patients; these strains were speciated by the API 20 Strep system (BioMe´rieux Vitek) and were found to include Streptococcus mitis (5 strains), Streptococcus parasanguis (2 strains), and Streptococcus sanguis (1 strain). Two strains of S. aureus (ATCC 29213 and 25923) and 4 clinical isolates of S. aureus obtained from blood culture were also studied. Bacteria were cultured overnight in Todd-Hewitt broth supplemented with 5% yeast extract (Difco). Then they were collected by centrifugation, were washed in Hanks’ balanced salt solution (HBSS; Life Technologies), were collected by centrifugation, and were sonicated (Ultrasonic Cleaner; Fisher Scientific) for 5 min. Microscopic examination showed that individual cocci predominated; pairs and short chains persisted to a greater or lesser extent, depending on the isolate. Sonication was followed by an increase in colony-forming units. Bacteria were collected again by centrifugation and were resuspended to the desired concentration, as determined by serial dilution and plating of aliquots on 5% sheep blood agar (BBL; Becton Dickinson).

86

Wanahita et al.

Table 1. Sources of isolates in study of interaction between human polymorphonuclear leukocytes and Streptococcus milleri group bacteria. Bacteria

Abscessa

Blood

Urine

Colonizing

6 5 2 0

1 2 3 0

0 0 3 0

0 0 0 8

Streptococcus constellatus Streptococcus intermedius Streptococcus anginosus Viridans streptococci

NOTE. Data are no. of isolates. Including intracerebral, intrathoracic, intra-abdominal, and perirectal sites.

a

Human Serum and PMNL

Blood obtained from 6 healthy adults was allowed to clot at room temperature, after which it was promptly centrifuged. Serum was removed, was pooled for use in this study (pooled human serum [PHS]), and was stored in aliquots at 2 70 C until immediately before use. Bacteria were opsonized by incubating an estimated 108 cfu/mL in 40% PHS in HBSS for 30 min in a shaking water bath at 37 C. A single donor provided PMNL for all studies reported here. Chemotaxis Assay

PMNL. PMNL were isolated from whole blood by sedimentation in Histopaque 1119 and 1077 (Sigma). Red blood cells were lysed by brief exposure to 0.2% saline, after which PMNL were resuspended in HBSS with 0.1% gelatin, to yield 2 £ 106 cells/mL. Bacteria. Triplicate samples of opsonized bacteria were pipetted into wells in a multiwell plate that served 1 chamber of a modified Boyden apparatus (NeuroProbe). A filter was placed over the plate, and 50 mL of a suspension that contained 2 £ 106 PMNL/mL was pipetted onto the filter above each well. Preliminary studies showed that 50 min of incubation at 37 C maximized the difference between random and directed migration. Accordingly, after 50 min, the reaction was stopped. Excess cells were removed, after which filters were stained with Hema-Diff (StatLab Medical Products). Cells at the lower margin of the filter were counted under £100 magnification. Twenty fields were examined at random by 2 experienced observers who were blinded to the conditions under which each filter had been incubated. The result for each condition was calculated by averaging triplicate determinations. For each isolate, a range of 3-fold dilutions was studied, and a graph was constructed to extrapolate the attraction of PMNL at 107 cfu of bacteria/mL. Migration under each condition was reported as the percentage of cells at the margin after exposure to S. aureus. Migration of PMNL in response to HBSS (random migration) was subtracted before calculating actual migration in response to other stimuli.

JID 2002;185 (1 January)

which a final sonication was done for 5 min. Pilot studies showed that, if bacteria were not used the same day, the kinetics of phagocytosis shifted; accordingly, for the experiments reported here, bacteria were used on the same day that the staining process was completed. Phagocytosis. Aliquots that contained 5 £ 106 cfu of Texas red– labeled bacteria were added to 100 mL of heparinized blood (preservative-free heparin; ESI) to yield a final ratio of 10 bacteria to 1 PMNL. Tubes containing bacteria in blood were incubated for 60 min at 37 C in a shaking water bath. Ten microliters of fluorescein-labeled anti-CD16 (CD16-FITC; Beckman Coulter) was added, after which tubes were placed in a refrigerator for 30 min. Three milliliters of Dulbecco’s PBS lacking Mg or Ca (Life Technologies) was added, and cells were collected by centrifugation (200 g for 5 min at 4 C). Red blood cells were lysed by use of FACS lysing solution (Becton Dickinson), and cells were again collected by centrifugation, were washed, and were resuspended in 500 mL of PBS containing 1% paraformaldehyde. Flow cytometry. Flow cytometry (Epics XL-MCL; Beckman Coulter) was used to quantitate cell-associated bacteria, by modification of a method we reported elsewhere [5]. The instrument was standardized before each experiment by use of Flow-check Fluorospheres (Beckman Coulter) and Flow-set Fluorospheres (Beckman Coulter). Granularity (side scatter) was recorded on the X-axis, and particle size (forward scatter) was recorded on the Y-axis. The gate was set to show PMNL, both unassociated and associated with bacteria. Events from this gate were plotted on a histogram in which the X-axis represented CD16-FITC– labeled PMNL and the Y-axis represented Texas red– labeled bacteria. Determination of the mean number of organisms per PMNL was based on the ratio of the fluorescence intensity of cell-associated organisms to the mean fluorescence intensity of each bacterium, as assessed in a separate sample. PMNL-associated bacteria were determined by the ratio of CD16+ PMNL that ingested bacteria to the total number of CD16+ PMNL and were reported as a percentage.

Phagocytosis Assay

Labeling of bacteria. Opsonized bacteria were collected by centrifugation and were resuspended in carbonate buffer (pH 9) that contained 0.7 mg/mL Texas red sulfonyl chloride (Molecular Probes). This suspension was incubated on a shaking platform in the dark at 4 C for 18 h. Texas Red–labeled bacteria were washed 6 times and were resuspended in HBSS to yield 108 cfu/mL, after

Figure 1. Chemotaxis induced by Streptococcus milleri group bacteria (Streptococcus constellatus, Streptococcus intermedius, and Streptococcus anginosus) and viridans streptococci (Y-axis), shown as percentage of chemotaxis stimulated by reference strain Staphylococcus aureus ATCC 29213 (heavy vertical line, 100%). Error bars indicate SE.


S. milleri Group Bacteria and PMNL

Figure 2. Mean numbers of Streptococcus milleri group bacteria (Streptococcus constellatus, Streptococcus intermedius, and Streptococcus anginosus), Staphylococcus aureus, and viridans streptococci associated with polymorphonuclear leukocytes (PMNL) after incubation in whole blood at a 10:1 ratio of bacteria to PMNL for 60 min at 37 C. Error bars indicate SE.

Bacterial Killing by Human PMNL

Opsonized bacteria were added to freshly obtained blood from a human donor, to yield a ratio of bacteria to PMNL of 1:1, and this suspension was incubated at 37 C. Aliquots were removed at time 0 and after 30, 60, 90, and 120 min. Tenfold dilutions were made in distilled water, with ample vortexing to allow for disintegration of PMNL, and 10-mL aliquots were streaked onto blood agar for bacterial enumeration.

Statistics

Mean data (^ SE) are reported. The Kruskal-Wallis test was used to determine whether there were significant differences among the 3 species of SMG. The Mann-Whitney U test was used to make pairwise comparisons between species in phagocytosis and chemotaxis studies. Survival of bacteria in bactericidal assays was analyzed by the Kaplan-Meier method. To test for equality between curves, the log rank test, x2 probability was used. For all tests, a ¼ :05.

87

by opsonized SMG species ranged from 33:5% ^ 9:8% to 55:1% ^ 27:6% of that induced by S. aureus. Differences among the 3 SMG species were not significant (P ¼ :9, KruskalWallis test). The differences compared with S. aureus were significant for S. constellatus and S. anginosus (P , :05 for each comparison; P ¼ :07 for S. intermedius ). In contrast, chemotaxis induced by viridans streptococci greatly exceeded that induced by S. aureus (166:9% ^ 23:1%; P , :005). Phagocytosis. Pilot studies showed that the number of cellassociated bacteria and the proportion of PMNL that had .1 associated bacterium were similar for 2 ATCC strains and 4 clinical isolates of S. aureus. Each flow cytometry experiment thereafter included S. aureus ATCC 29213, .1 isolate of each species from SMG, and a viridans streptococcus. The number of organisms per PMNL and the percentage of PMNL with .1 cell-associated S. aureus were consistent from day to day and were similar to previous reports from this laboratory [5]. The mean (^ SE) number of organisms per PMNL for S. aureus was 4:7 ^ 2:4. Greater numbers of organisms per PMNL were found for SMG bacteria and viridans streptococci, with means ranging from 7:0 ^ 2:8 organisms/PMNL for S. intermedius to 14:3 ^ 2:8 organisms/PMNL for S. constellatus (figure 2). The differences compared with S. aureus were significant (P , :05) for S. constellatus, S. anginosus, and viridans streptococci (P ¼ :07 for S. intermedius ). The mean (^ SE) number of PMNL with cell-associated S. aureus was 77:6% ^ 7:7% (figure 3). After incubation with SMG organisms, a greater proportion of PMNL had cell-associated bacteria, ranging from 91:9% ^ 4:2% to 97:9% ^ 2:0%, with the differences being significant for S. constellatus and S. anginosus (P , :05 for each comparison); in the case of S. intermedius, the difference did not reach statistical significance

Results Chemotaxis. Preliminary experiments examined migration of PMNL in response to HBSS, PHS, unopsonized bacteria, and PHS after bacteria had been incubated in it and removed by centrifugation (spent serum). Identical results were obtained with S. aureus ATCC 29213 and a clinical S. aureus isolate after incubation in PHS; accordingly, ATCC 29213 served as the internal laboratory reference to be run in each assay and against which all other bacteria would be compared. Chemotaxis in response to pooled human serum, unopsonized bacteria, and spent serum was , 15% of that observed with opsonized S. aureus. As shown in figure 1, chemotaxis (mean ^ SE) induced

Figure 3. Percentage of polymorphonuclear leukocytes (PMNL) with . 1 cell-associated bacterium of Streptococcus milleri group bacteria (Streptococcus constellatus, Streptococcus intermedius, and Streptococcus anginosus), Staphylococcus aureus, or viridans streptococci after 1 h of incubation at 37 C. Error bars indicate SE.

88

Wanahita et al.


nearly identical, with the mean percentage killed at 120 min ranging from 65% to 71%. There was no killing of viridans streptococci under these conditions. Electron microscopic examination. To verify that the association between PMNL and bacteria, as shown in phagocytosis assays, was accompanied by actual bacterial ingestion, we examined PMNL after they had been incubated for 30 min at 37 C with opsonized bacteria at a ratio of 10 bacteria to 1 PMNL. Figure 5 shows representative results—in this case, phagocytosis of S. intermedius by PMNL. One isolate of each bacterial species described here was studied; ingestion was found to occur in each case (data not shown).

Figure 4. Killing of Streptococcus milleri group bacteria (Streptococcus constellatus, Streptococcus intermedius, and Streptococcus anginosus), Staphylococcus aureus, and viridans streptococci during incubation in whole blood for 120 min.

(P ¼ :1). For viridans streptococci, 91:8% ^ 6:1% of PMNL had ingested bacteria (P ¼ :03, vs. S. aureus ). Killing of bacteria by human PMNL. Time-dependent killing of S. aureus occurred during incubation with PMNL; at 120 min, 99% of organisms were killed (figure 4). Results were consistent from day to day. In contrast, killing of each of the SMG species proceeded at a substantially slower rate (P , :001, Kaplan-Meier method). Results for these 3 species were

Discussion To cause an abscess, bacteria must release substances that attract PMNL and, to a greater or lesser extent, resist ingestion and killing by PMNL. The present study examined the interaction of 3 species of SMG, a taxonomic group of bacteria that all share a remarkable propensity to cause abscesses, with human PMNL. We examined the chemotactic effect of these organisms, as well as their capacity to resist phagocytosis and killing by PMNL after opsonization in complement-rich, pooled human serum. The observation that SMG bacteria attracted significantly fewer PMNL than did S. aureus in an in vitro assay of chemotaxis might appear to be paradoxical. It should be noted,

Figure 5. Electron microscopic view (original magnification, £12,000) after incubating opsonized Staphylococcus intermedius with polymorphonuclear leukocytes (PMNL) for 30 min at a ratio of 10 bacteria per PMNL. Cytoplasmic projections are seen surrounding and about to ingest bacteria. Many bacteria are seen within cytoplasm, contained in phagocytic vacuoles. Some bacteria show signs of replicating; others have lost integrity, consistent with loss of viability. Most bacteria appear to be intact.


S. milleri Group Bacteria and PMNL

however, that S. aureus, with its well-recognized capacity to cause abscesses, has been shown to inhibit the ingress of PMNL in vivo [6] and in vitro [7], which is consistent with the even greater degree of chemotaxis after exposure to viridans streptococci. Leukocidin(s), probably identical with PantonValentine leukocidins that directly damage PMNL and/or red blood cells [8, 9], have been found to be responsible; these leukocidins are especially likely to be present in strains that cause furunculosis [10]. Inhibition of chemotaxis is thought to contribute to pathogenicity of S. aureus [11] by delaying the ingress of PMNL, thereby giving a head start to invading bacteria [6]. S. intermedius produces intermedilysin [12], a toxin that may share certain characteristics with staphylococcal leukocidins. Although the gene that encodes this toxin is not present in S. constellatus or S. anginosus [13], other toxins may exert a leukocidin-like effect in these species. In the case of SMG, the number of bacteria per PMNL and the number of PMNL that had cell-associated organisms after incubation exceeded that of S. aureus, an extensively studied bacterium that also causes abscesses. By contrast, a bacterial killing study showed that, after ingestion by PMNL, SMG bacteria were killed at a rate of only 3% of that of S. aureus. It is unclear why SMG are so highly resistant to PMNL killing. Seven or 8 distinct isolates were studied from each bacterial species, and some were studied on as many as 4 occasions, with remarkably uniform results. Relative resistance to killing by PMNL is an expected characteristic of organisms that cause abscesses. Because S. aureus is well known to survive relatively well within human leukocytes [14], the substantially better survival of SMG bacteria, at least during 2 h of study in vitro, is quite impressive. Organisms as diverse as Mycobacterium tuberculosis [15], Brucella abortus [16], and Bacteroides gingivalis [17] have been shown to produce substances that inhibit activation and/or discharge of granule enzymes into vacuoles within PMNL; a similar mechanism might be responsible in the case of SMG. Electron microscopic examination verified that streptococci were ingested by PMNL; although some cocci appeared to be disintegrating, the morphologic appearance of the bacteria inside phagocytic vacuoles was quite normal, which is consistent with a low degree of bacterial killing. Rakita et al. [18] suggested that the presence of large phagocytic vacuoles is associated with the failure of PMNL to kill Enterococcus faecalis. Although we detected large vacuoles after ingestion of SMG bacteria, which were not killed, similar-sized vacuoles were seen after ingestion of S. aureus, which were killed. An unexplained finding that emerged from these studies was the resistance of viridans streptococci to killing by PMNL. Mandell [19] found a single strain of viridans streptococci to be susceptible to PMNL-mediated killing, but our study used 8 separate clinical isolates, running each experiment .2 times, with consistent results. Electron microscopy showed that viri-

89

dans streptococci had been internalized, so the fact that they were not killed remains unexplained.

Acknowledgment

We thank Edward A. Graviss for statistical analysis.

References 1. Whiley RA, Beighton D, Winstanely TG, Fraser HY, Hardie JM. Streptococcus intermedius, Streptococcus constellatus, and Streptococcus anginosus (the Streptococcus milleri group): association with different body sites and clinical infections. J Clin Microbiol 1992;30:243–4. 2. Jacobs JA, Pietersen HG, Stobbringh EE, Soeters PB. Streptococcus anginosus, Streptococcus constellatus, Streptococcus constellatus, and Streptococcus intermedius: clinical relevance, hemolytic, and serologic characteristics. Am J Clin Pathol 1995;104:547–53. 3. Clarridge JE III, Attorri S, Musher DM, Hebert J, Dunbar S. Streptococcus intermedius, Streptococcus constellatus, and Streptococcus anginosus (“Streptococcus milleri group”) are of different clinical importance and not equally associated with abscess. Clin Infect Dis 2001;32:1511–5. 4. Clarridge JE, Osting C, Jalali M, Osborn J, Waddington M. Genotypic and phenotypic characterization of “Streptococcus milleri group” isolates from a Veterans Administration hospital population. J Clin Microbiol 1999;37:3681–7. 5. Bandres JC, Trial J, Musher DM, Rossen RD. Increased phagocytosis and generation of reactive oxygen products by neutrophils and monocytes of men with stage 1 human immunodeficiency virus infection. J Infect Dis 1993;168:75–83. 6. Cohn ZA. Determinants of infection in the peritoneal cavity. I. Response to and fate of Staphylococcus aureus and Staphylococcus albus in the mouse. Yale J Biol Med 1962;35:12– 28. 7. Weksler BB, Hill MJ. Inhibition of leukocyte migration by a staphylococcal factor. J Bacteriol 1969;98:1030– 5. 8. Loeffler DA, Schatt KA, Norcross NL. Use of 51Cr release to measure the cytotoxic effects of staphylococcal leukocidin and toxin neutralization on bovine leukocytes. J Clin Microbiol 1986;23:416– 20. 9. Prevost G, Cribier B, Couppie P, et al. Panton-Valentine leucocidin and g-hemolysin from Staphylococcus aureus ATCC 49775 are encoded by distinct genetic loci and have different biological activities. Infect Immun 1995;63:4121–9. 10. Lina G, Piemont Y, Godail-Gamot F, et al. Involvement of Panton-Valentine leukocidin–producing Staphylococcus aureus in primary skin infections and pneumonia. Clin Infect Dis 1999;29:1128–32. 11. Densen P, Mandell GL. Phagocyte strategy versus microbial tactics. Rev Infect Dis 1980;2:817– 38. 12. Nagamune H, Ohnishi C, Katsuura A, et al. Intermedilysin, a novel cytotoxin specific for human cells secreted by Streptococcus intermedius UNS46 isolated from a human liver abscess. Infect Immun 1996; 64: 3093– 100. 13. Nagamune H, Whiley RA, Goto T, et al. Distribution of the intermedilysin gene among the anginosus group streptococci and correlation between intermedilysin production and deep-seated infection with Streptococcus intermedius. J Clin Microbiol 2000;38:220– 6. 14. Melly MA, Thomison JN, Rogers DE. Fate of staphylococci within human leukocytes. J Exp Med 1960;112:1121–30. 15. Wadee AA, Clara AM. A 25-kilodalton fraction from Mycobacterium tuberculosis that inhibits hexose monophosphate shunt activity, lysozyme release, and H2O2 production: reversal by g interferon. Infect Immun 1989;57:864–9.

90

Wanahita et al.

16. Canning PC, Roth JA, Tabatabai LB, Deyoe BL. Isolation of components of Brucella abortus responsible for inhibition of function in bovine neutrophils. J Infect Dis 1985;152:913–21. 17. Yoneda M, Maeda K, Aono M. Suppression of bactericidal activity of human polymorphonuclear leukocytes by Bacteroides gingivalis. Infect Immun 1990;58:406–11.


18. Rakita RM, Vanek NN, Jacques-Palaz K, et al. Enterococcus faecalis bearing aggregation substance is resistant to killing by human neutrophils despite phagocytosis and neutrophil activation. Infect Immun 1999; 67: 6067–75. 19. Mandell GL. Bactericidal activity of aerobic and anaerobic polymorphonuclear neutrophils. Infect Immun 1974;9:337–41.