- Email: [email protected]
Molecular characterization and LD50 identify virulent isolates of Staphylococcus epidermidis from adult sepsis
Diagnostic Microbiology and Infectious Disease 36 (2000) 75– 80
Molecular characterization and LD50 identify virulent isolates of Staphylococcus epidermidis from adult sepsis Betty L. Herndona,*, Hanna Bialkowska–Hobrzanskab, Lawrence Dalla,1 a
b
Section of Infectious Diseases, University of Missouri–Kansas City School of Medicine, Kansas City, MO 64108 Molecular Biology Diagnostic Laboratory, Lawson Research Institute, University of Western Ontario, London, Ontario N6A 4V2, Canada Received 9 July 1999; received in revised form 13 September 1999; accepted 25 September 1999.
Abstract Staphylococcus epidermidis plays an important role in infections of patients with implanted prosthetic devices. The exact clinical significance of recovered S. epidermidis from clinical specimens is difficult to assess, as they are inhabitants of the normal skin. In this study, 11 adults with clinical sepsis and blood cultures that grew only S. epidermidis were the host population. Bacterial virulence in vivo was determined by using the mouse LD50 assay where the intravenous lethality was determined for each patient isolate. Bacterial dose (CFU ⫻ 109) that produced lethality in 50% of the animals at 12 h was the value used for comparison. Restriction fragment length polymorphism (RFLP) analysis of chromosomal DNA by pulsed-field gel electrophoresis (PFGE) was used for identification of individual strains and their clonal organization. Confirmation of species assignment was done by RFLP analysis of 16S ⫹ 23S rRNA gene regions (ribotyping). Plasmid profile analysis was also conducted. Four of 11 blood isolates from adults with S. epidermidis sepsis had indistinguishable or closely related DNA patterns and were considered clone A. The same clone was previously seen to account for the majority of sepsis in a neonatal intensive care unit. There were significant differences in virulence characteristics of the S. epidermidis isolates. Clone A isolates produced lethality by LD50 in mice at a dose averaging 2.35; clone B isolate at a dose of 2.54, and the remaining isolates, representing six distinct clones, were lethal to mice at significantly larger doses (3.51-5.17, average 4.16). These data suggest that individual clones of S. epidermidis isolated from septic adults have detectable differences in virulence as defined by an animal bioassay, and the more virulent clone is widespread. © 2000 Elsevier Science Inc. All rights reserved.
1. Introduction Staphylococcus epidermidis blood infections have increased as the use of implanted biopolymers has increased, and in this setting they act as pathogens with biofilm production, adherence to biomaterials and virulence in vivo (Deighton et al., 1996; Huebner and Kropec, 1995; Herndon et al., 1995a). Laboratory methods have been sought to determine if bacterial strains within this species differ in virulence (van der Mei et al., 1997; Heinzelmann et al., 1997). An ideal measure would reflect the clinical picture and prognosis and separate commensal S. epidermidis isolates from pathogenic– behaving isolates. Slime production and polysaccharide/adhesion have been documented as important factors in some infections (Patrick et al., 1995; * Corresponding author. Tel: ⫹1-816-235-1904; fax ⫹1-816-2355194. E-mail address: [email protected] (B. L. Herndon) 1 Present address: Midwest Hospital Specialists, Kansas City, MO.
McKenney et al., 1994). Clinical strains of S. epidermidis were found to elaborate toxins with biological and biochemical properties resembling alpha, beta, and delta toxins of S. aureus (Gemmell and Thelestam, 1981), and the presence of sequences related to those toxin genes of S. aureus have been recently identified in clinical strains of S. epidermidis (Bialkowska–Hobrzanska and Zochodne, unpublished). Furthermore, there has recently been detected in S. epidermidis an accessory gene regulator (agr), analogous to a regulator gene of S. aureus which is involved with virulence-associated factors (van Wamel et al., 1998). These findings all provide evidence for virulence potential of S. epidermidis. The purpose of this study was to examine whether clinical isolates of S. epidermidis recovered from severe cases of adult sepsis could be differentiated by their genetic characteristics, determined by chromosomal DNA RFLP-PFGE analysis, and by virulence, measured by mouse LD50 assay. We evaluated blood isolates from 11 adult patients with S. epidermidis sepsis. Clinical signs of each infection were
0732-8893/00/$ – see front matter © 2000 Elsevier Science Inc. All rights reserved. PII: S 0 7 3 2 - 8 8 9 3 ( 0 0 ) 0 0 1 1 4 - 5
76
B.L. Herndon et al. / Diagnostic Microbiology and Infectious Disease 36 (2000) 75– 80
reviewed and compared to genetic characteristics of isolates. Virulence of each isolate was quantified by intravenous LD50 in mice. RFLP-PFGE analysis of chromosomal DNA was used for identification of individual strains and their clonal organization.
immunosuppression). The score was based on a modified format of the APACHE II scale (Knaus et al., 1985), a classification validated in intensive care units that we have used previously to score the severity of infection (Dall et al., 1994).
2. Materials and Methods
2.4. Species identification by rDNA RFLP typing (ribotyping)
2.1. Bacterial strains The strains were selected from a population of 921 blood cultures positive for S. epidermidis over two years (1991– 1993) in a Midwestern medical-school-affiliated hospital bacteriology laboratory. Coagulase-negative isolates were identified as S. epidermidis based on biochemical characteristics determined by an automated system (Microscan Walkaway-40; Baxter Diagnostics, West Sacramento, CA USA). The following criteria were used for selection of S. epidermidis for this study: 1) Isolates from multiple blood collection sites must have grown in 4/4 or 4/6 blood collection bottles; 2) isolates must have grown in under 48 h in media incubated both aerobically and anaerobically; 3) the patient’s blood must not grow other organisms; and 4) each isolate must be unique as demonstrated by Western blot against a polyclonal anti-staphylococcal antiserum. Eleven isolates that met all selection criteria were examined in this study. 2.2. Virulence determination: LD50 Murine assays for bacterial virulence were performed with clearance from the University Animal Care and Use committee. The log10 CFU intravenous dose that was fatal to 50% of the dosed animals at 12 h was determined for each strain of S. epidermidis (IV LD50). Five to 20 male Swiss–Webster mice averaging 20 g were used to establish the intravenous (IV) LD50 at each dose level ranging from 1 ⫻ 108 to 8 ⫻ 109 CFU. S. epidermidis were grown in tryptic soy broth, and were suspended in saline to a final density of 1 ⫻ 109 CFU/mL. The required bacterial number, achieved by centrifugal concentration or dilution of this saline suspension, was always administered in 0.2 mL volume via tail vein. Animals were observed for 1 h, then checked at 12 h and daily for 2 weeks. Twelve hours deaths were utilized for LD50 determinations reported here. Mortality calculations were performed with the multivariate module of Statistica (Statsoft, Tulsa, OK, USA).
RFLP analysis of ribosomal RNA genes (ribotyping) was conducted as previously described (Bialkowska–Hobrzanska et al., 1990a). Briefly, total DNA was digested with ClaI restriction endonuclease. The restriction fragments were separated on agarose gel, transferred to a membrane and were hybridized with end-labeled 16⫹23S rRNA from Escherichia coli. Numbers and mobilities of rDNA fragments were visually compared on autoradiograms. The degree of similarity was calculated between rDNA patterns based on arbitrary definition. Strains displaying ⱖ70% similarity with the Staphylococcus epidermidis ATCC 14990 type strain were classified as this species (Bialkowska– Hobrzanska, 1990b). 2.5. RFLP-PFGE analysis A procedure of Goering et al. (1995) was used with the following modifications. Bacterial cells resuspended in TEN (100-mM Tris pH 7.5, 100-mM EDTA, 150-mM NaCl) buffer were cast at a density of 2.5 ⫻ 109 CFU into 1% SeaPlaque agarose plugs. The agarose plugs were incubated in lysis buffer supplemented with 2 mg/mL lysozyme (Sigma, St. Louis MO, USA) and 20 U/mL of lysostaphin (Sigma) for 4 h at 37°C. Subsequently, the plugs were incubated in DNA extraction buffer (50-mM Tris pH 8, 100-mM EDTA, 0.1% SDS, 100 g/mL proteinase K [Sigma]) overnight at 50°C. For restriction digestion, slices of agarose plugs containing DNA were extensively washed with TE buffer at 50°C followed by equilibration with the appropriate restriction endonuclease buffer and incubation with 50 U of SmaI (Boehringer–Mannheim) overnight at 25°C. Electrophoresis was performed by using the CHEFDRIII system (Bio–Rad, Richmond, CA, USA) at previously described conditions (Goering et al., 1995). 2.6. Plasmid profile analysis The cleared lysate procedure of total DNA purification (Bialkowska–Hobrzanska, 1990a) was employed for plasmid profile analysis.
2.3. Patient infection severity score Patient data were reviewed with approval from the University IRB. The patients from whose blood the S. epidermidis strains were isolated were scored for illness severity by using a form that included clinical signs of sepsis (immature PMNs, fever, BP, etc.) and contributing factors (age,
3. Results Both molecular typing and virulence analysis have been performed on bacteria isolated from blood cultures of 11 adults with S. epidermidis sepsis, and the severity of illness
B.L. Herndon et al. / Diagnostic Microbiology and Infectious Disease 36 (2000) 75– 80
77
Table 1 Virulence and genetic characteristics of S. epidermidis blood isolates from adult sepsis
Fig. 1. Chromosomal DNA RFLP analysis by using SmaI and PFGE: S. epidermidis blood isolates from septic adults. Lanes, S. epidermidis isolates and DNA pattern types are: (1) 2049, A1; (2) 1687, A2; (3) 4492, A2; (4) 1968, A1; (5) 1455, B1; (6) 4112, C1; (7) 2504, D1; (8) 1103, E1; (9) 4230, F1; (10) 2381, G1; (11) 1016, H1.
score of these patients compared to the bacterial analysis scores. Fig. 1 shows RFLP-PFGE analysis of chromosomal DNA of S. epidermidis isolates from septic adult patients. Four of 11 isolates examined had highly related chromosomal DNA RFLP-PFGE patterns, and were classified into clone A. The remaining seven isolates represented seven distinct clones. The intravenous LD50 in mice of each isolate and the molecular type of that isolate are shown in Table 1. The LD50, the dose in CFU ⫻ 109 which is fatal to half the mice, is inversely related to virulence. The values range from 1.50 to 5.17. Four of the seven most virulent isolates (by LD50 in mice, 1.50-3.37) were in clonal group A. The four isolates of clonal group A represent highly related strain variants (Fig. 1). Plasmid profile analysis of the S. epidermidis isolates correlated very well with patterns of chromosomal DNA (Table 1). The more virulent clonal group A isolates had an identical plasmid profile pattern (a3), comprising DNA bands in the following positions of linear DNA standards: 2.3, 4.4, 8.7, 13.5, 25, 27, and 44 kb. Fig. 2 graphs clinical signs in patients infected with clonal group A and clone B strains of S. epidermidis and compares this scoring to clinical signs in patients infected with other molecular types of S. epidermidis. Patients were
Isolate Code
Virulence (LD50)
2049 1687 4492 1968 1455 4112 2504 1103 4230 2381 1016
1.50 2.25 2.30 3.37 2.54 3.56 5.01 4.17 5.17 3.51 3.54
Species ID (rDNA Type)
Chrom DNA RFLP-PFGE Type
Plasmid Content
Clonal Group
S.epi(2) S.epi(2) S.epi(2) S.epi(2) S.epi(11) S.epi(2) S.epi(2) S.epi(1) S.epi(1) S.epi(5) S.epi(2)
A1 A2 A2 A1 B1 C1 D1 E1 F1 G1 H1
a3 a3 a3 a3 h3 d4 np b4 f4 e2 a4
A A A A B C D E F G H
The criteria for classification of S. epidermidis isolates based on chromosomal DNA RFLP-PFGE analysis were previously described (Tenover et al, 1995). Indistinguishable isolates with identical DNA patterns were considered to represent the same strain. Closely related isolates with similar DNA patterns that included 2 to 3 band difference were considered to represent closely related strains and designated A1 and A2, representing the same clonal group A. Unrelated isolates with distinct DNA RFLPPFGE patterns differing from strains of clonal group A by more than 7 band difference were considered distinct strains and designated type B, C, etc. The lethality to mice of clonal group A and B isolates was significantly greater than that of the other isolates of S. epidermidis, suggesting their greater virulence, *p ⫽ 0.0027.
scored for illness severity by using a form that included both elevation or depression from norm. Scoring suggests that all patients were similarly ill (Table 2), and a recent review of
Fig. 2. Clinical signs in patients with S. epidermidis sepsis. Adult patients from whose blood the S. epidermidis strains were isolated were scored for illness severity using a form that included clinical signs of illness: (% immature PMN, temperature, blood pressure, etc.) as well as contributing factors (age, immunosuppression). Both elevations and depressions from norm are incorporated in the score. Mean scores of clonal A and clone B groups were not significantly different from scores of patients infected with the other strains.
78
B.L. Herndon et al. / Diagnostic Microbiology and Infectious Disease 36 (2000) 75– 80
Table 2 S. Epidermidis blood isolates grouped by genetic identitya Bacterial Clone
Mouse LD50 b
Clone A Clone B Other Clones
Patient Score
Mean
S.D.
2.355 2.54 4.16
0.77 — 0.76
Mean 23.25 24.0 22.17
a Clone A, representing 4 of 11 human S. epidermidis blood isolates and clone B, representing one isolate, correlate with a significantly greater virulence to mice (reduced LD50). The patients from which the S. epidermidis were isolated show similar illness (APACHE II score), all were very ill. b Clone A group, either including or not including the clone B isolate, is significantly more lethal to mice than the other clones, *p ⫽ 0.0065 (A only) and **p ⫽ 0.0027 (A and B).
clinical outcomes reveals mortality rates in excess of 50% with an association with indwelling intravascular devices. The S. epidermidis blood isolates are grouped by genetic identity in Fig. 3. Box and whisker plots show the significant difference in mouse LD50 (Fig. 3A) when administered S. epidermidis of clonal group A and clone B isolates compared to other isolates. Clonal group A and clone B have a significantly greater virulence potential, with a lethality score of 2.39 compared to the other S. epidermidis isolates that have a mean score of 4.16. The difference between groups is significant, p ⫽ 0.0027, unpaired t-test. Fig. 3B, APACHE II scoring, demonstrates the similarities of the patient host illness. APACHE II scoring for patients infected with clonal groups A and B, 23.4 ⫾ 3.13, was not significantly different from patients with other isolates 22.17 ⫾ 4.99, p ⫽ 0.93.
4. Discussion We have titrated virulence in S. epidermidis isolates taken from bacteremic adults by the mouse LD50, and have demonstrated the presence of a common genetic pattern in 36% of the more lethal isolates. Our study also suggests that the presence of small plasmids, often known to correlate with antibiotic resistance, also occurred in the low LD50 isolates. Whether these common plasmids contribute to a higher pathogenicity and persistence of S. epidermidis strains in blood awaits further investigation. Antibiotic resistance measures on these strains were performed only as needed clinically; these organisms were all susceptible to vancomycin. Studies of the genetic structure of natural populations of human pathogens has indicated that most bacterial species are clonal in structure, with relatively few clones in each species and fewer yet associated with disease (Selander and Musser, 1990). Clonal structure of S. epidermidis species has been recently demonstrated (Bialkowska–Hobrzanska et al., 1993). This study has identified two clones of S. epi-
dermidis (clones A, B) isolated from bacteremic adults that expressed greater virulence by lethality to mice when compared to virulence shown by the remaining six clones (this study) and several others previously characterized by Herndon et al. (unpublished). It is of importance that S. epidermidis clone A, common in bacteremic adults in the Midwestern United States, was also the predominant clone seen in the neonatal intensive care unit in the St. Joseph’s Health Centre, London, Ontario (Herndon et al., 1995b; Bialkowska–Hobrzanska et al., 1993). These findings support the concept of the existence of invasive as well as commensal clones of S. epidermidis that can be distinguished by their chromosomal DNA RFLP patterns. Chromosomal DNA RFLP-PFGE is presently considered a gold standard method for specific identification of individual strains within bacterial species and interpretation of interstrain relationships (Arbeit et al., 1990, Tenover et al., 1995, Tenover et al., 1997). Eleven S. epidermidis isolates selected for this study from very ill bacteremic adults had eight different RFLP-PFGE patterns; four of the 11 isolates had identical or highly related RFLP-PFGE patterns A1-A2. Although multilocus sequence typing offers an attractive alternative approach for differentiation of bacterial isolates from invasive disease and healthy carriers (Maiden et al., 1998) it requires further evaluation with a large number of bacterial pathogens and comparison of its discriminatory power to the RFLP-PFGE method. Virulence of coagulase-negative staphylococci (CNS) has been often studied in animal models with foreign body infections, reflecting the commonest clinical presentation, although recent work has shown that a foreign body is not required for the expression of virulence by S. epidermidis (Deighton et al., 1996). The LD50 has long been the criterion by which dose-effect of drugs and chemicals is measured, and has also been utilized to study bacterial virulence (Herndon et al., 1995; Khurana et al., 1991; Halavalkar and Barrow, 1993). The intravenous LD50 provided a highly sensitive approach for the detection of S. epidermidis virulence in cases of human sepsis (Herndon et al., 1995b; Dall et al., 1994). With as few as 6 to 8 animals per dosage group, reproducible 12-h survival charts furnished individual bacterial profiles that were more uniform than in vitro virulence assays (Herndon et al., 1993). Furthermore, necropsy of test animals dying acutely showed strain-mediated signs that predominated in the animals administered the more virulent clones: 76% of clone A and B mice had hemorrhagic lungs at death, whereas 16.7% of mice administered other clonal groups had this finding. Inflamed Peyers patches occurred only in Clone A mice at 12-h necropsy. Because route of inoculation is suggested to be important to virulence, it is notable that the mouse LD50 virulence model used IV administration of the human isolates, reflecting the probable venous site of entry of these bacteria into the human hosts. Additionally, the IV LD50 was chosen for S. epidermidis virulence determination over the intraperitoneal (IP) route which is occasionally used in the bacterial LD50.
B.L. Herndon et al. / Diagnostic Microbiology and Infectious Disease 36 (2000) 75– 80
79
Fig. 3. Box and whisker plots of S. epidermidis blood isolates grouped by genetic identity, Clone A and clone B (n ⫽ 5) and other clones (n ⫽ 6). A) Mouse LD50: The groups are significantly different, p ⫽ 0.0027. B) APACHE II scoring: The groups are similarly ill.
By the IV route, bacteria are rapidly cleared from the circulation and sequestered into the liver and spleen with host phagocytes. The differences in relative virulence by IV administration in an immunocompetent host can accurately reflect the capability of a host’s primary defense. On the other hand, when bacteria are given IP, they are phagocytosed poorly from the peritoneal cavity and can multiply extracellularly with the possibility of an aberrant doseresponse. The lack of correlation between murine and molecular virulence and the human host APACHE scores was not unexpected. Because the APACHE score is a physiologic profile that reflects the host response to infection, a large number of unmeasured variables exist that could influence that score. Further, these patients varied in pre-existing phenomena (sickle cell disease, previous chemotherapy, presence of very long-term central lines). Any of these could have influenced the APACHE numbers and overshadowed the input of the virulence of the infecting organism. The cytotoxicity of S. epidermidis clone A has been
assayed and compared to that of several distinct clonal groups isolated in the NICU (London, Ontario) environment (Bialkowska–Hobrzanska et al., 1994). In the cytotoxicity assay, a S. epidermidis clone A was found to induce leakage of 3H-uridine from labeled human embryonic lung fibroblasts at 54% that of S. aureus, whereas a control S. epidermidis ATCC 35983 strain released less than 9% radioactive label. Membrane-damaging toxin determinants specifying alpha, beta, and delta toxins analogous to those of S. aureus were identified in the S. epidermidis clone A (Bialkowska–Hobrzanska and Zochodne, unpublished). Delta toxin activity as well as a functional accessory gene regulator-like locus were recently demonstrated in the majority of tested S. epidermidis strains (van Wamel et al., 1998). In conclusion, the LD50 in the mouse, a standard of toxicity, has been applied to S. epidermidis isolates from bacteremic adults, demonstrating a broad profile of virulence. A molecular typing system was employed to further evaluate the isolates, classifying individual strains and
80
B.L. Herndon et al. / Diagnostic Microbiology and Infectious Disease 36 (2000) 75– 80
clones within the S. epidermidis species. Four of 11 S. epidermidis isolates showed highly related chromosomal DNA RFLP-PFGE patterns, thus suggesting their common ancestry. These isolates were significantly more virulent by the LD50 mouse assay than were the remaining strains. These data suggest that individual clones exist within S. epidermidis that have measurable differences in virulence. Given the known inter and intraspecies genetic exchange among staphylococci, one can speculate that this ubiquitous and long-considered nonpathogenic group of microorganisms have at some point acquired virulence-associated determinants. Control of these pathogenic genetic traits awaits defining their phenotypic character.
Acknowledgments The authors thank R.F. Fletcher from the Molecular Biology Diagnostic Laboratory, Lawson Research Institute, London, Ontario for improving a protocol for RFLP-PFGE analysis of staphylococci, and V. Harry for conducting molecular typing.
References Arbeit, R. D., Arthur, M., Dunn, R., Kim, C., Selander, R. K., & Goldstein, R. (1990). Resolution of recent evolutionary divergence among E. coli from related lineages: the application of pulsed-field electrophoresis to molecular epidemiology. J Infect Dis 161, 230 –235. Bialkowska–Hobrzanska, H., Harry, V., Jaskot, D., & Hammerberg, O. (1990a). Typing of coagulase-negative staphylococci by Southern hybridization of chromosomal DNA fingerprints using a ribosomal RNA probe. Eur J Clin Microbiol Inf Dis 9, 588 –594. Bialkowska–Hobrzanska, H., Jascot, D., & Hammerberg, O. (1990b). A method for DNA restriction endonuclease fingerprinting of coagulasenegative staphylococci. J Microbiol Meth 12, 41– 49. Bialkowska–Hobrzanska, H., Jaskot, D., & Hammerberg, O. (1993). Molecular characterization of the coagulase-negative staphylococcal surface flora of premature neonates. J Gen Micro 139, 2939 –2944. Bialkowska–Hobrzanska, H., Harry, V. S., Gregson, D., & Hammerberg, O. (1994). Development of modified quantitative cytotoxicity assay for coagulase-negative staphylococci. American Society for Microbiology 94th General Meeting, ASM Abstracts: B98. p 46. Dall, L., Wang, J., Herndon, B., & Suvarna, P. (1994). Bacterial characterization in Staphylococcus epidermidis septicemia. J Lab Clin Med 124, 802– 807. Deighton, M. A., Borland, R., & Capstick, J. A. (1996). Virulence of Staphylococcus epidermidis in a mouse model: significance of extracellular slime. Epidemiol Infect 117, 267–280. Gemmell, C. G. & Thelestam, M (1981). Toxinogenecity of clinical isolates of coagulase-negative staphylococci towards various animal cells. Acta Path Microbiol Scand Sect B, 89, 417– 421. Goering, R. V., Fey, P. D., & Goldstein, F. W. (1995). Usefulness of pulsed-field gel electrophoresis in the epidemiological analysis of Staphylococcus aureus isolates with decreased susceptibility to teicoplanin Eur J Clin Microbiol Infect Dis 14(suppl 1), S3–5.
Halavalkar, H. & Barrow, P. A. (1993). The role of a 54-kb plasmid in the virulence of strains of S. enteritidis of phage type 4 for chickens and mice. J Med Microbiol 38, 171–176. Heinzelmann, M., Herzig, D. O., Swain, B., Mercer–Jones, M. A., Bergamini, T. M., Polk, H. C. Jr. (1997). Phagocytosis and oxidative burst response of planktonic S. epidermidis RP62A and its non-slime producing variant in human neutrophils. Clin Diag Lab Immunol 4, 705– 710. Herndon, B., Dall, L., & Barnes, W. (1993). In vitro assays of S. epidermidis characteristics and outcome in an endocarditis model. Can J Infect Dis 4, 203–208. Herndon, B., Dall, L., & Suvarna, P. (1995a). Contribution of the host to test results in assays of Staphylococcus epidermidis. J Clin Lab Analysis 9, 81– 88. Herndon, B., Bialkowska–Hobrzanska, H., & Dall, L. (1995b). Characterization of S. epidermidis isolates from septic patients by LD50 and restriction endonuclease fingerprinting: a pattern of virulence. American Society for Microbiology 95th General Meeting, ASM Abstracts: B-64. p 176. Huebner, J. & Kropec, A. (1995). Cross infections due to coagulasenegative staphylococci in high risk patients. Zentralblatt fuer Bakteriol 283, 169 –174. Khurana, S., Ganguly, N. K., Panigrahi, D., Khullar, M., & Walia, B. N. S. (1991). The correlation of biochemical virulence factors and enterotoxin production with LD50 of S. typhimurium in mice. Microbiologica 14, 113–117. Knaus, W. A., Draper, E. A., Wagner, D. P., Zimmerman, J. E. (1985). APACHE II: a severity of disease classification system. Crit Care Med 13, 818 – 829. Maiden, M. C., Bygraves, J. A., Feil, E., Morelli, G., Russell, J. E., Urwin, R., Zhang, Q., Zhou, J., Zurth, K., Caugant, D. A., Feavers, I. M., Achtman, M., & Spratt, B. G. (1998). Multilocus sequence typing: a portable approach to the identification of clones within populations of pathogenic microorganisms. Proc Natl Acad Sci USA 95, 3140 –3145. McKenney, D., Hubner, J., Muller, E., Wang, Y., Goldmann, D. A., & Pier, G. B. (1998). The ica locus of Staphylococcus epidermidis encodes production of the capsular polysaccharide/adhesion. Infec Immun 66, 4711– 4720. van der Mei, H. C., van de Beit–Gritter, B., Reid, G., Bialkowska– Hobrzanska, H., & Busscher, H. J. (1997). Adhesion of coagulasenegative staphylococci grouped according to physico-chemical surface properties. Microbiol 143, 3861–3870. Patrick, C. C., Hetherington, S. V., Roberson, P. K., Henwick, S., & Sloas, M. M. (1995). Comparative virulence of Staphylococcus epidermidis isolates in a murine catheter model. Pediatr Res 37, 70 –74. Selander, R. K. & Musser, J. M. (1990). Population genetics of bacterial pathogens In (I. C. Gunsalus, Ed.) The Bacteria Vol. xi, New York: Academic Press. pp. 11–36. Tenover, F. C., Arbeit, R. D., Goering, R. V., Mickelsen, P. A., Murray, B.E., Persing, D. H., & Swaminathan, B. (1995). Interpreting chromosomal DNA restriction patterns produced by pulsed-field gel electrophoresis: criteria for bacterial strain typing. J Clin Microbiol 33, 2233– 2239. Tenover, F. C., Arbeit, R. D., & Goering, R. V. (1997). How to select and interpret molecular strain typing methods for epidemiological studies of bacterial infections: a review for health care epidemiologists. Molecular Typing Working Group of the Society for Healthcare Epidemiology of America. Infect Control Hosp Epidemiol 18:426 – 439. Van Wamel, W. J., van Rossum, G., Verhoef, J., Vandenbroucke–Grauls, C.M., & Fluit, A. C. (1998). Cloning and characterization of an accessory gene regulator (agr)-like locus from Staphylococcus epidermidis. FEMS Microbiol Lett 163, 1–9.
Molecular characterization and LD50 identify virulent isolates of Staphylococcus epidermidis from adult sepsis Betty L. Herndona,*, Hanna Bialkowska–Hobrzanskab, Lawrence Dalla,1 a
b
Section of Infectious Diseases, University of Missouri–Kansas City School of Medicine, Kansas City, MO 64108 Molecular Biology Diagnostic Laboratory, Lawson Research Institute, University of Western Ontario, London, Ontario N6A 4V2, Canada Received 9 July 1999; received in revised form 13 September 1999; accepted 25 September 1999.
Abstract Staphylococcus epidermidis plays an important role in infections of patients with implanted prosthetic devices. The exact clinical significance of recovered S. epidermidis from clinical specimens is difficult to assess, as they are inhabitants of the normal skin. In this study, 11 adults with clinical sepsis and blood cultures that grew only S. epidermidis were the host population. Bacterial virulence in vivo was determined by using the mouse LD50 assay where the intravenous lethality was determined for each patient isolate. Bacterial dose (CFU ⫻ 109) that produced lethality in 50% of the animals at 12 h was the value used for comparison. Restriction fragment length polymorphism (RFLP) analysis of chromosomal DNA by pulsed-field gel electrophoresis (PFGE) was used for identification of individual strains and their clonal organization. Confirmation of species assignment was done by RFLP analysis of 16S ⫹ 23S rRNA gene regions (ribotyping). Plasmid profile analysis was also conducted. Four of 11 blood isolates from adults with S. epidermidis sepsis had indistinguishable or closely related DNA patterns and were considered clone A. The same clone was previously seen to account for the majority of sepsis in a neonatal intensive care unit. There were significant differences in virulence characteristics of the S. epidermidis isolates. Clone A isolates produced lethality by LD50 in mice at a dose averaging 2.35; clone B isolate at a dose of 2.54, and the remaining isolates, representing six distinct clones, were lethal to mice at significantly larger doses (3.51-5.17, average 4.16). These data suggest that individual clones of S. epidermidis isolated from septic adults have detectable differences in virulence as defined by an animal bioassay, and the more virulent clone is widespread. © 2000 Elsevier Science Inc. All rights reserved.
1. Introduction Staphylococcus epidermidis blood infections have increased as the use of implanted biopolymers has increased, and in this setting they act as pathogens with biofilm production, adherence to biomaterials and virulence in vivo (Deighton et al., 1996; Huebner and Kropec, 1995; Herndon et al., 1995a). Laboratory methods have been sought to determine if bacterial strains within this species differ in virulence (van der Mei et al., 1997; Heinzelmann et al., 1997). An ideal measure would reflect the clinical picture and prognosis and separate commensal S. epidermidis isolates from pathogenic– behaving isolates. Slime production and polysaccharide/adhesion have been documented as important factors in some infections (Patrick et al., 1995; * Corresponding author. Tel: ⫹1-816-235-1904; fax ⫹1-816-2355194. E-mail address: [email protected] (B. L. Herndon) 1 Present address: Midwest Hospital Specialists, Kansas City, MO.
McKenney et al., 1994). Clinical strains of S. epidermidis were found to elaborate toxins with biological and biochemical properties resembling alpha, beta, and delta toxins of S. aureus (Gemmell and Thelestam, 1981), and the presence of sequences related to those toxin genes of S. aureus have been recently identified in clinical strains of S. epidermidis (Bialkowska–Hobrzanska and Zochodne, unpublished). Furthermore, there has recently been detected in S. epidermidis an accessory gene regulator (agr), analogous to a regulator gene of S. aureus which is involved with virulence-associated factors (van Wamel et al., 1998). These findings all provide evidence for virulence potential of S. epidermidis. The purpose of this study was to examine whether clinical isolates of S. epidermidis recovered from severe cases of adult sepsis could be differentiated by their genetic characteristics, determined by chromosomal DNA RFLP-PFGE analysis, and by virulence, measured by mouse LD50 assay. We evaluated blood isolates from 11 adult patients with S. epidermidis sepsis. Clinical signs of each infection were
0732-8893/00/$ – see front matter © 2000 Elsevier Science Inc. All rights reserved. PII: S 0 7 3 2 - 8 8 9 3 ( 0 0 ) 0 0 1 1 4 - 5
76
B.L. Herndon et al. / Diagnostic Microbiology and Infectious Disease 36 (2000) 75– 80
reviewed and compared to genetic characteristics of isolates. Virulence of each isolate was quantified by intravenous LD50 in mice. RFLP-PFGE analysis of chromosomal DNA was used for identification of individual strains and their clonal organization.
immunosuppression). The score was based on a modified format of the APACHE II scale (Knaus et al., 1985), a classification validated in intensive care units that we have used previously to score the severity of infection (Dall et al., 1994).
2. Materials and Methods
2.4. Species identification by rDNA RFLP typing (ribotyping)
2.1. Bacterial strains The strains were selected from a population of 921 blood cultures positive for S. epidermidis over two years (1991– 1993) in a Midwestern medical-school-affiliated hospital bacteriology laboratory. Coagulase-negative isolates were identified as S. epidermidis based on biochemical characteristics determined by an automated system (Microscan Walkaway-40; Baxter Diagnostics, West Sacramento, CA USA). The following criteria were used for selection of S. epidermidis for this study: 1) Isolates from multiple blood collection sites must have grown in 4/4 or 4/6 blood collection bottles; 2) isolates must have grown in under 48 h in media incubated both aerobically and anaerobically; 3) the patient’s blood must not grow other organisms; and 4) each isolate must be unique as demonstrated by Western blot against a polyclonal anti-staphylococcal antiserum. Eleven isolates that met all selection criteria were examined in this study. 2.2. Virulence determination: LD50 Murine assays for bacterial virulence were performed with clearance from the University Animal Care and Use committee. The log10 CFU intravenous dose that was fatal to 50% of the dosed animals at 12 h was determined for each strain of S. epidermidis (IV LD50). Five to 20 male Swiss–Webster mice averaging 20 g were used to establish the intravenous (IV) LD50 at each dose level ranging from 1 ⫻ 108 to 8 ⫻ 109 CFU. S. epidermidis were grown in tryptic soy broth, and were suspended in saline to a final density of 1 ⫻ 109 CFU/mL. The required bacterial number, achieved by centrifugal concentration or dilution of this saline suspension, was always administered in 0.2 mL volume via tail vein. Animals were observed for 1 h, then checked at 12 h and daily for 2 weeks. Twelve hours deaths were utilized for LD50 determinations reported here. Mortality calculations were performed with the multivariate module of Statistica (Statsoft, Tulsa, OK, USA).
RFLP analysis of ribosomal RNA genes (ribotyping) was conducted as previously described (Bialkowska–Hobrzanska et al., 1990a). Briefly, total DNA was digested with ClaI restriction endonuclease. The restriction fragments were separated on agarose gel, transferred to a membrane and were hybridized with end-labeled 16⫹23S rRNA from Escherichia coli. Numbers and mobilities of rDNA fragments were visually compared on autoradiograms. The degree of similarity was calculated between rDNA patterns based on arbitrary definition. Strains displaying ⱖ70% similarity with the Staphylococcus epidermidis ATCC 14990 type strain were classified as this species (Bialkowska– Hobrzanska, 1990b). 2.5. RFLP-PFGE analysis A procedure of Goering et al. (1995) was used with the following modifications. Bacterial cells resuspended in TEN (100-mM Tris pH 7.5, 100-mM EDTA, 150-mM NaCl) buffer were cast at a density of 2.5 ⫻ 109 CFU into 1% SeaPlaque agarose plugs. The agarose plugs were incubated in lysis buffer supplemented with 2 mg/mL lysozyme (Sigma, St. Louis MO, USA) and 20 U/mL of lysostaphin (Sigma) for 4 h at 37°C. Subsequently, the plugs were incubated in DNA extraction buffer (50-mM Tris pH 8, 100-mM EDTA, 0.1% SDS, 100 g/mL proteinase K [Sigma]) overnight at 50°C. For restriction digestion, slices of agarose plugs containing DNA were extensively washed with TE buffer at 50°C followed by equilibration with the appropriate restriction endonuclease buffer and incubation with 50 U of SmaI (Boehringer–Mannheim) overnight at 25°C. Electrophoresis was performed by using the CHEFDRIII system (Bio–Rad, Richmond, CA, USA) at previously described conditions (Goering et al., 1995). 2.6. Plasmid profile analysis The cleared lysate procedure of total DNA purification (Bialkowska–Hobrzanska, 1990a) was employed for plasmid profile analysis.
2.3. Patient infection severity score Patient data were reviewed with approval from the University IRB. The patients from whose blood the S. epidermidis strains were isolated were scored for illness severity by using a form that included clinical signs of sepsis (immature PMNs, fever, BP, etc.) and contributing factors (age,
3. Results Both molecular typing and virulence analysis have been performed on bacteria isolated from blood cultures of 11 adults with S. epidermidis sepsis, and the severity of illness
B.L. Herndon et al. / Diagnostic Microbiology and Infectious Disease 36 (2000) 75– 80
77
Table 1 Virulence and genetic characteristics of S. epidermidis blood isolates from adult sepsis
Fig. 1. Chromosomal DNA RFLP analysis by using SmaI and PFGE: S. epidermidis blood isolates from septic adults. Lanes, S. epidermidis isolates and DNA pattern types are: (1) 2049, A1; (2) 1687, A2; (3) 4492, A2; (4) 1968, A1; (5) 1455, B1; (6) 4112, C1; (7) 2504, D1; (8) 1103, E1; (9) 4230, F1; (10) 2381, G1; (11) 1016, H1.
score of these patients compared to the bacterial analysis scores. Fig. 1 shows RFLP-PFGE analysis of chromosomal DNA of S. epidermidis isolates from septic adult patients. Four of 11 isolates examined had highly related chromosomal DNA RFLP-PFGE patterns, and were classified into clone A. The remaining seven isolates represented seven distinct clones. The intravenous LD50 in mice of each isolate and the molecular type of that isolate are shown in Table 1. The LD50, the dose in CFU ⫻ 109 which is fatal to half the mice, is inversely related to virulence. The values range from 1.50 to 5.17. Four of the seven most virulent isolates (by LD50 in mice, 1.50-3.37) were in clonal group A. The four isolates of clonal group A represent highly related strain variants (Fig. 1). Plasmid profile analysis of the S. epidermidis isolates correlated very well with patterns of chromosomal DNA (Table 1). The more virulent clonal group A isolates had an identical plasmid profile pattern (a3), comprising DNA bands in the following positions of linear DNA standards: 2.3, 4.4, 8.7, 13.5, 25, 27, and 44 kb. Fig. 2 graphs clinical signs in patients infected with clonal group A and clone B strains of S. epidermidis and compares this scoring to clinical signs in patients infected with other molecular types of S. epidermidis. Patients were
Isolate Code
Virulence (LD50)
2049 1687 4492 1968 1455 4112 2504 1103 4230 2381 1016
1.50 2.25 2.30 3.37 2.54 3.56 5.01 4.17 5.17 3.51 3.54
Species ID (rDNA Type)
Chrom DNA RFLP-PFGE Type
Plasmid Content
Clonal Group
S.epi(2) S.epi(2) S.epi(2) S.epi(2) S.epi(11) S.epi(2) S.epi(2) S.epi(1) S.epi(1) S.epi(5) S.epi(2)
A1 A2 A2 A1 B1 C1 D1 E1 F1 G1 H1
a3 a3 a3 a3 h3 d4 np b4 f4 e2 a4
A A A A B C D E F G H
The criteria for classification of S. epidermidis isolates based on chromosomal DNA RFLP-PFGE analysis were previously described (Tenover et al, 1995). Indistinguishable isolates with identical DNA patterns were considered to represent the same strain. Closely related isolates with similar DNA patterns that included 2 to 3 band difference were considered to represent closely related strains and designated A1 and A2, representing the same clonal group A. Unrelated isolates with distinct DNA RFLPPFGE patterns differing from strains of clonal group A by more than 7 band difference were considered distinct strains and designated type B, C, etc. The lethality to mice of clonal group A and B isolates was significantly greater than that of the other isolates of S. epidermidis, suggesting their greater virulence, *p ⫽ 0.0027.
scored for illness severity by using a form that included both elevation or depression from norm. Scoring suggests that all patients were similarly ill (Table 2), and a recent review of
Fig. 2. Clinical signs in patients with S. epidermidis sepsis. Adult patients from whose blood the S. epidermidis strains were isolated were scored for illness severity using a form that included clinical signs of illness: (% immature PMN, temperature, blood pressure, etc.) as well as contributing factors (age, immunosuppression). Both elevations and depressions from norm are incorporated in the score. Mean scores of clonal A and clone B groups were not significantly different from scores of patients infected with the other strains.
78
B.L. Herndon et al. / Diagnostic Microbiology and Infectious Disease 36 (2000) 75– 80
Table 2 S. Epidermidis blood isolates grouped by genetic identitya Bacterial Clone
Mouse LD50 b
Clone A Clone B Other Clones
Patient Score
Mean
S.D.
2.355 2.54 4.16
0.77 — 0.76
Mean 23.25 24.0 22.17
a Clone A, representing 4 of 11 human S. epidermidis blood isolates and clone B, representing one isolate, correlate with a significantly greater virulence to mice (reduced LD50). The patients from which the S. epidermidis were isolated show similar illness (APACHE II score), all were very ill. b Clone A group, either including or not including the clone B isolate, is significantly more lethal to mice than the other clones, *p ⫽ 0.0065 (A only) and **p ⫽ 0.0027 (A and B).
clinical outcomes reveals mortality rates in excess of 50% with an association with indwelling intravascular devices. The S. epidermidis blood isolates are grouped by genetic identity in Fig. 3. Box and whisker plots show the significant difference in mouse LD50 (Fig. 3A) when administered S. epidermidis of clonal group A and clone B isolates compared to other isolates. Clonal group A and clone B have a significantly greater virulence potential, with a lethality score of 2.39 compared to the other S. epidermidis isolates that have a mean score of 4.16. The difference between groups is significant, p ⫽ 0.0027, unpaired t-test. Fig. 3B, APACHE II scoring, demonstrates the similarities of the patient host illness. APACHE II scoring for patients infected with clonal groups A and B, 23.4 ⫾ 3.13, was not significantly different from patients with other isolates 22.17 ⫾ 4.99, p ⫽ 0.93.
4. Discussion We have titrated virulence in S. epidermidis isolates taken from bacteremic adults by the mouse LD50, and have demonstrated the presence of a common genetic pattern in 36% of the more lethal isolates. Our study also suggests that the presence of small plasmids, often known to correlate with antibiotic resistance, also occurred in the low LD50 isolates. Whether these common plasmids contribute to a higher pathogenicity and persistence of S. epidermidis strains in blood awaits further investigation. Antibiotic resistance measures on these strains were performed only as needed clinically; these organisms were all susceptible to vancomycin. Studies of the genetic structure of natural populations of human pathogens has indicated that most bacterial species are clonal in structure, with relatively few clones in each species and fewer yet associated with disease (Selander and Musser, 1990). Clonal structure of S. epidermidis species has been recently demonstrated (Bialkowska–Hobrzanska et al., 1993). This study has identified two clones of S. epi-
dermidis (clones A, B) isolated from bacteremic adults that expressed greater virulence by lethality to mice when compared to virulence shown by the remaining six clones (this study) and several others previously characterized by Herndon et al. (unpublished). It is of importance that S. epidermidis clone A, common in bacteremic adults in the Midwestern United States, was also the predominant clone seen in the neonatal intensive care unit in the St. Joseph’s Health Centre, London, Ontario (Herndon et al., 1995b; Bialkowska–Hobrzanska et al., 1993). These findings support the concept of the existence of invasive as well as commensal clones of S. epidermidis that can be distinguished by their chromosomal DNA RFLP patterns. Chromosomal DNA RFLP-PFGE is presently considered a gold standard method for specific identification of individual strains within bacterial species and interpretation of interstrain relationships (Arbeit et al., 1990, Tenover et al., 1995, Tenover et al., 1997). Eleven S. epidermidis isolates selected for this study from very ill bacteremic adults had eight different RFLP-PFGE patterns; four of the 11 isolates had identical or highly related RFLP-PFGE patterns A1-A2. Although multilocus sequence typing offers an attractive alternative approach for differentiation of bacterial isolates from invasive disease and healthy carriers (Maiden et al., 1998) it requires further evaluation with a large number of bacterial pathogens and comparison of its discriminatory power to the RFLP-PFGE method. Virulence of coagulase-negative staphylococci (CNS) has been often studied in animal models with foreign body infections, reflecting the commonest clinical presentation, although recent work has shown that a foreign body is not required for the expression of virulence by S. epidermidis (Deighton et al., 1996). The LD50 has long been the criterion by which dose-effect of drugs and chemicals is measured, and has also been utilized to study bacterial virulence (Herndon et al., 1995; Khurana et al., 1991; Halavalkar and Barrow, 1993). The intravenous LD50 provided a highly sensitive approach for the detection of S. epidermidis virulence in cases of human sepsis (Herndon et al., 1995b; Dall et al., 1994). With as few as 6 to 8 animals per dosage group, reproducible 12-h survival charts furnished individual bacterial profiles that were more uniform than in vitro virulence assays (Herndon et al., 1993). Furthermore, necropsy of test animals dying acutely showed strain-mediated signs that predominated in the animals administered the more virulent clones: 76% of clone A and B mice had hemorrhagic lungs at death, whereas 16.7% of mice administered other clonal groups had this finding. Inflamed Peyers patches occurred only in Clone A mice at 12-h necropsy. Because route of inoculation is suggested to be important to virulence, it is notable that the mouse LD50 virulence model used IV administration of the human isolates, reflecting the probable venous site of entry of these bacteria into the human hosts. Additionally, the IV LD50 was chosen for S. epidermidis virulence determination over the intraperitoneal (IP) route which is occasionally used in the bacterial LD50.
B.L. Herndon et al. / Diagnostic Microbiology and Infectious Disease 36 (2000) 75– 80
79
Fig. 3. Box and whisker plots of S. epidermidis blood isolates grouped by genetic identity, Clone A and clone B (n ⫽ 5) and other clones (n ⫽ 6). A) Mouse LD50: The groups are significantly different, p ⫽ 0.0027. B) APACHE II scoring: The groups are similarly ill.
By the IV route, bacteria are rapidly cleared from the circulation and sequestered into the liver and spleen with host phagocytes. The differences in relative virulence by IV administration in an immunocompetent host can accurately reflect the capability of a host’s primary defense. On the other hand, when bacteria are given IP, they are phagocytosed poorly from the peritoneal cavity and can multiply extracellularly with the possibility of an aberrant doseresponse. The lack of correlation between murine and molecular virulence and the human host APACHE scores was not unexpected. Because the APACHE score is a physiologic profile that reflects the host response to infection, a large number of unmeasured variables exist that could influence that score. Further, these patients varied in pre-existing phenomena (sickle cell disease, previous chemotherapy, presence of very long-term central lines). Any of these could have influenced the APACHE numbers and overshadowed the input of the virulence of the infecting organism. The cytotoxicity of S. epidermidis clone A has been
assayed and compared to that of several distinct clonal groups isolated in the NICU (London, Ontario) environment (Bialkowska–Hobrzanska et al., 1994). In the cytotoxicity assay, a S. epidermidis clone A was found to induce leakage of 3H-uridine from labeled human embryonic lung fibroblasts at 54% that of S. aureus, whereas a control S. epidermidis ATCC 35983 strain released less than 9% radioactive label. Membrane-damaging toxin determinants specifying alpha, beta, and delta toxins analogous to those of S. aureus were identified in the S. epidermidis clone A (Bialkowska–Hobrzanska and Zochodne, unpublished). Delta toxin activity as well as a functional accessory gene regulator-like locus were recently demonstrated in the majority of tested S. epidermidis strains (van Wamel et al., 1998). In conclusion, the LD50 in the mouse, a standard of toxicity, has been applied to S. epidermidis isolates from bacteremic adults, demonstrating a broad profile of virulence. A molecular typing system was employed to further evaluate the isolates, classifying individual strains and
80
B.L. Herndon et al. / Diagnostic Microbiology and Infectious Disease 36 (2000) 75– 80
clones within the S. epidermidis species. Four of 11 S. epidermidis isolates showed highly related chromosomal DNA RFLP-PFGE patterns, thus suggesting their common ancestry. These isolates were significantly more virulent by the LD50 mouse assay than were the remaining strains. These data suggest that individual clones exist within S. epidermidis that have measurable differences in virulence. Given the known inter and intraspecies genetic exchange among staphylococci, one can speculate that this ubiquitous and long-considered nonpathogenic group of microorganisms have at some point acquired virulence-associated determinants. Control of these pathogenic genetic traits awaits defining their phenotypic character.
Acknowledgments The authors thank R.F. Fletcher from the Molecular Biology Diagnostic Laboratory, Lawson Research Institute, London, Ontario for improving a protocol for RFLP-PFGE analysis of staphylococci, and V. Harry for conducting molecular typing.
References Arbeit, R. D., Arthur, M., Dunn, R., Kim, C., Selander, R. K., & Goldstein, R. (1990). Resolution of recent evolutionary divergence among E. coli from related lineages: the application of pulsed-field electrophoresis to molecular epidemiology. J Infect Dis 161, 230 –235. Bialkowska–Hobrzanska, H., Harry, V., Jaskot, D., & Hammerberg, O. (1990a). Typing of coagulase-negative staphylococci by Southern hybridization of chromosomal DNA fingerprints using a ribosomal RNA probe. Eur J Clin Microbiol Inf Dis 9, 588 –594. Bialkowska–Hobrzanska, H., Jascot, D., & Hammerberg, O. (1990b). A method for DNA restriction endonuclease fingerprinting of coagulasenegative staphylococci. J Microbiol Meth 12, 41– 49. Bialkowska–Hobrzanska, H., Jaskot, D., & Hammerberg, O. (1993). Molecular characterization of the coagulase-negative staphylococcal surface flora of premature neonates. J Gen Micro 139, 2939 –2944. Bialkowska–Hobrzanska, H., Harry, V. S., Gregson, D., & Hammerberg, O. (1994). Development of modified quantitative cytotoxicity assay for coagulase-negative staphylococci. American Society for Microbiology 94th General Meeting, ASM Abstracts: B98. p 46. Dall, L., Wang, J., Herndon, B., & Suvarna, P. (1994). Bacterial characterization in Staphylococcus epidermidis septicemia. J Lab Clin Med 124, 802– 807. Deighton, M. A., Borland, R., & Capstick, J. A. (1996). Virulence of Staphylococcus epidermidis in a mouse model: significance of extracellular slime. Epidemiol Infect 117, 267–280. Gemmell, C. G. & Thelestam, M (1981). Toxinogenecity of clinical isolates of coagulase-negative staphylococci towards various animal cells. Acta Path Microbiol Scand Sect B, 89, 417– 421. Goering, R. V., Fey, P. D., & Goldstein, F. W. (1995). Usefulness of pulsed-field gel electrophoresis in the epidemiological analysis of Staphylococcus aureus isolates with decreased susceptibility to teicoplanin Eur J Clin Microbiol Infect Dis 14(suppl 1), S3–5.
Halavalkar, H. & Barrow, P. A. (1993). The role of a 54-kb plasmid in the virulence of strains of S. enteritidis of phage type 4 for chickens and mice. J Med Microbiol 38, 171–176. Heinzelmann, M., Herzig, D. O., Swain, B., Mercer–Jones, M. A., Bergamini, T. M., Polk, H. C. Jr. (1997). Phagocytosis and oxidative burst response of planktonic S. epidermidis RP62A and its non-slime producing variant in human neutrophils. Clin Diag Lab Immunol 4, 705– 710. Herndon, B., Dall, L., & Barnes, W. (1993). In vitro assays of S. epidermidis characteristics and outcome in an endocarditis model. Can J Infect Dis 4, 203–208. Herndon, B., Dall, L., & Suvarna, P. (1995a). Contribution of the host to test results in assays of Staphylococcus epidermidis. J Clin Lab Analysis 9, 81– 88. Herndon, B., Bialkowska–Hobrzanska, H., & Dall, L. (1995b). Characterization of S. epidermidis isolates from septic patients by LD50 and restriction endonuclease fingerprinting: a pattern of virulence. American Society for Microbiology 95th General Meeting, ASM Abstracts: B-64. p 176. Huebner, J. & Kropec, A. (1995). Cross infections due to coagulasenegative staphylococci in high risk patients. Zentralblatt fuer Bakteriol 283, 169 –174. Khurana, S., Ganguly, N. K., Panigrahi, D., Khullar, M., & Walia, B. N. S. (1991). The correlation of biochemical virulence factors and enterotoxin production with LD50 of S. typhimurium in mice. Microbiologica 14, 113–117. Knaus, W. A., Draper, E. A., Wagner, D. P., Zimmerman, J. E. (1985). APACHE II: a severity of disease classification system. Crit Care Med 13, 818 – 829. Maiden, M. C., Bygraves, J. A., Feil, E., Morelli, G., Russell, J. E., Urwin, R., Zhang, Q., Zhou, J., Zurth, K., Caugant, D. A., Feavers, I. M., Achtman, M., & Spratt, B. G. (1998). Multilocus sequence typing: a portable approach to the identification of clones within populations of pathogenic microorganisms. Proc Natl Acad Sci USA 95, 3140 –3145. McKenney, D., Hubner, J., Muller, E., Wang, Y., Goldmann, D. A., & Pier, G. B. (1998). The ica locus of Staphylococcus epidermidis encodes production of the capsular polysaccharide/adhesion. Infec Immun 66, 4711– 4720. van der Mei, H. C., van de Beit–Gritter, B., Reid, G., Bialkowska– Hobrzanska, H., & Busscher, H. J. (1997). Adhesion of coagulasenegative staphylococci grouped according to physico-chemical surface properties. Microbiol 143, 3861–3870. Patrick, C. C., Hetherington, S. V., Roberson, P. K., Henwick, S., & Sloas, M. M. (1995). Comparative virulence of Staphylococcus epidermidis isolates in a murine catheter model. Pediatr Res 37, 70 –74. Selander, R. K. & Musser, J. M. (1990). Population genetics of bacterial pathogens In (I. C. Gunsalus, Ed.) The Bacteria Vol. xi, New York: Academic Press. pp. 11–36. Tenover, F. C., Arbeit, R. D., Goering, R. V., Mickelsen, P. A., Murray, B.E., Persing, D. H., & Swaminathan, B. (1995). Interpreting chromosomal DNA restriction patterns produced by pulsed-field gel electrophoresis: criteria for bacterial strain typing. J Clin Microbiol 33, 2233– 2239. Tenover, F. C., Arbeit, R. D., & Goering, R. V. (1997). How to select and interpret molecular strain typing methods for epidemiological studies of bacterial infections: a review for health care epidemiologists. Molecular Typing Working Group of the Society for Healthcare Epidemiology of America. Infect Control Hosp Epidemiol 18:426 – 439. Van Wamel, W. J., van Rossum, G., Verhoef, J., Vandenbroucke–Grauls, C.M., & Fluit, A. C. (1998). Cloning and characterization of an accessory gene regulator (agr)-like locus from Staphylococcus epidermidis. FEMS Microbiol Lett 163, 1–9.