- Email: [email protected]
Failure of phenotypic characteristics to distinguish between carrier and invasive isolates of Staphylococcus epidermidis
Journal
of Hospital
Failure
Infection
(1991)
17, 107-l
15
of phenotypic characteristics to distinguish between carrier and invasive isolates of Staphylococcus epidermidis M. J. Souto,
Departamento
C. M. Ferreirbs
and M. T. Criado
de Microbiologia y Parasitologia, Facultad Universidad de Santiago de Compostela, Spain Accepted for publication
de Farmacia,
3 September1990
Summary: Thirty carrier and 29 invasive Staphylococcus epidermidis isolates were analysed for production of slime, extracellular enzymes and antibiotic resistance. Evaluation of slime production was by two methods which gave differing results, but both showed no difference between the two groups of isolates. Exoenzyme production by the two groups of isolates was also similar, and the only differences were shown in resistance to some antibiotics; carrier isolates were more resistant to novobiocin and cephalothin, but resistance to chloramphenicol, tunicamycin, teicoplanin, penicillin, ampicillin and tetracycline was similar in both groups. Cluster analysis based on exoenzyme production showed no consistent difference between invasive and carrier isolates, with the lowest similarity coefficient over 88% (except for one isolate). Consequently, none of the characteristics studied was useful in differentiating potentially invasive isolates. Introduction
Interest in epidemiological markers that might prove useful for the identification and control of infections caused by Staphyloccocus epidermidis has increased during recent years. This microorganism, along with other coagulase-negative staphylococci (CNS), is a normal saprophyte on the skin and anterior nares of healthy individuals. Staphylococcus epidermidis is one of the CNS most frequently isolated from infections and it has been increasingly associated with nosocomial infections of prosthetic valves and hosts joints and vascular devices,’ with infections in immunocompromised and in patients subjected to aggressive treatments such as continuous ambulatory peritoneal dialysis.2 The ubiquitous presence of S. epidermidis in humans, in the skin and nares,3 and its behaviour as a nosocomial pathogen, makes it necessary to investigate methods that can be effective for the investigation of specific strains, for example study of their biochemical and antimicrobial resistance profiles. Slime production has been proposed as a tool for measuring the Correspondence
to: Prof.
C. Ferreiros.
019%6701/91/020107+09$03.00/0
0
107
1991
The
Hospital
Infection
Socwry
108
M. J. Souto et al.
clinical significance of CNS,4 and the production of exoenzymes or synergistic haemolysis’ have been suggested as being characteristic of virulence. Owing to the large number of carriers of this microorganism among hospital personnel, it is important to know if patients are infected by their own cutaneous microbial flora (normally they are compromised hosts) or by specific nosocomial strains with a characteristic pathogenic potential. Our aim in this work was to study the properties cited above in two groups of S. epidermidis isolates; one group isolated from clinical samples and the other from healthy individuals, not associated with the hospital environment in an attempt to determine the possible clinical significance of these properties and their possible use for the differentiation of strains in epidemiological studies. Material
and methods
Bacterial isolates A total of 59 isolates of S. epidermidis was studied. Thirty were isolated from healthy carriers in the Bacteriology Laboratory (Direction de Salud, La Cor&a, Spain), and 29 clinical isolates were obtained mainly from blood cultures in the Hospital General de Galicia, Spain. Identification of each isolate was by the usual techniques: Gram staining, production of catalase and coagulase negativity, and inability to ferment mannitol. Identification was confirmed with the API Staph Gallery (API Systems S.A., Montalieu Vercien, France). Biotype was determined according to the scheme described by Parisi. All isolates were maintained at 4°C on brain-heart agar (Biolife, Italy), and sub-cultured monthly. For prolonged storage the organisms were kept at - 20°C in skimmed milk. Slime production Slime production was determined by the method described by Christensen et a1.7 The organisms were grown in 3 ml trypticase soy broth in polystyrene tubes for 18 h at 37°C. The tubes were emptied and immediately stained for 60 s with 0.1% Alcian Blue. Results were recorded as ‘ + + ‘, ‘ + ‘, ‘ f ’ or ‘ - ’ depending on the degree of staining of the tube wall. Slime production was also determined by slide agglutination with Concanavalin A (ConA) according to Ludwicka et aZ.* Twenty-five ~1 of a bacterial suspension (5 X lo9 bacteria ml-‘) in phosphate-buffered saline (PBS) was mixed with 25 ~1 of serial twofold dilutions of a 1% (w/v) solution of ConA (Serva Feinbiochemica, Heidelberg, NY) in PBS. After 1 min at 20°C the results were recorded as the highest dilution of ConA causing agglutination of the bacteria. Extracellular enzymes The production of acid and alkaline phosphatase, caseinase and a- and P-haemolysins was determined
lipase, gelatinase, by the methods
Phenotypic
characteristics
in S. epidermidis
109
described by Smibert & Krieg. 9 For the lipase determination, Tween 20, Tween 40 and Tween 60 were added independently to the basal medium. Extracellular production of DNase was detected on deoxyribonuclease test medium (Biolife, Italy). Antimicrobial susceptibility The following antimicrobial agents were tested: penicillin, ampicillin, streptomycin, novobiocin, tunicamycin, tetracycline, clindamycin, teicoplanin, gentamicin, methicillin, vancomycin, rifampicin, kanamycin, chloramphenicol, cephalothin and erythromycin. For each isolate, minimal inhibitory concentrations (MICs) were determined by a standard macrodilution method.” Using a Steers replicator, 1O4colony forming units (cfu) of each isolate were deposited onto the surface of Mueller-Hinton agar plates containing serial twofold dilution of the antibiotics (from an initial concentration of 256 I.tg ml-‘). All tests were performed in duplicate. Each MIC was recorded after both 24 and 48 h of incubation at 37°C. MIC,, and MIC,, were calculated as the concentrations inhibiting 50% and 90% of isolates respectively. Statistical methods The SPSS/PC+ statistical package (SPSS Inc. and Microsoft Corp.) was used. Grouping of the strains was by cluster analysis using the unweighed pair group method of cluster analysis (UPGMA) and arithmetic averages of Euclidean distances.”
Results All isolates were identified as S. epidermidis, biotype I. Table I shows the results obtained for slime production in both carrier and invasive isolates by the methods of adherence to polystyrene (method 1) and agglutination with ConA (method 2). The methods were not equivalent in determining slime production, as tested by calculation of the Kendall’s z coefficient of concordance (z= -0.21) on the results obtained for the two groups of isolates. Neither of the two methods showed significant differences in slime production between the two groups of isolates (carrier and invasive) when tested by the Mann-Whitney U-test (P=O*SS for method 1; P=O*38 for method 2). Figure 1 shows the percentages of isolates of each group producing the different exoenzymes tested. There were no differences between carrier and invasive isolates (P>O*OS, X*-test) and none of the enzymes could be associated specifically with either of the groups. Table II shows the MIC,, for each antibiotic tested. MICs of chloramphenicol, and MIC,, ampicillin and tetracycline were tunicamycin, teicoplanin, penicillin, similar in the two groups of isolates. Carrier isolates were more resistant to
110
M. J. Souto et al. Table I. Slime production of Staphylococus
epidermidis
Invasive isolates
isolates
Carrier isolates
Method 1
Method 2
Isolate
Method 1
Method 2
E
-+
64 16
Nl
*f
328
s4
32 2 0
!G: N7 N15 N16 N17 N18 N20 N21 N23 N24 N33 N34 N35 N37 N38 N39 N41 N42
+ ++ ++ ++ f ++ ++ + + + + + + + +
32
El S18 s20 S23 S26 S27 S28 S29 s31 S32 s33 534 s35 S36
f + ++ ++ ++ + ++ ++ + + -t i + ++
S38 s37 s39 540 s41 S42 s43 s44
++ ++ + ++ ++ It
1:8 128 16 4 4 32 4
N44 N43 N45 N47 N51 N53 NS5 N57
+ &
+
: 8 8 8 4 64 8
8
N6.5 N61
-+
6:
Isolate
:z
545 Method staining; Method
+
; 0 : 32 4 8 16 1:: ; 16
+
+ + +
1: 64 4 2 32 32 16 8 4 8 4 1: 16 128 64
1, Adherence to polystyrene results graded ‘ - ‘: not seen; ‘ f ’ weak staining; ‘ + ’ intermediate ‘ + + ’ strong staining. 2, ConA agglutination: figures represent highest dilution observed for each isolate.
novobiocin and cephalothin, and invasive ones were more resistant to the remainder of the antibiotics. Figure 2 shows the clusters formed by invasive and carrier isolates when enzymatic profiles were subjected to cluster analysis by a UPGMA clustering method. Only one isolate showed a similarity coefficient less than clusters were connected at 86% similarity. 86%, and the 16 remaining Clusters at 100% similarity often contained both invasive and carrier isolates. Discussion
All S. epidermidis studied in this work, both clinical and carrier isolates, were characterized by the API Staph System, which offers a good correlation with conventional methods,12 and has also been used for the biotyping of coagulase-negative staphylococci.4
Phenotypic
FAC
F/iL
characteristics
TiO
TiO
in S. epidermidis
T&O
GiL
DiA
CiS
111
HiL
HiE
Enzymes
Figure 1. Percentages of isolates producing enzymes in carrier (solid bars) and invasive (cross-hatched bars) Staphylococcus epidermidis groups. FAC, acid phosphatase; FAL, alkaline phosphatase; T20, Tween-20 hydrolase; T40, Tween-40 hydrolase; T60, Tween-60 hydrolase; GEL, gelatinase; DNA, deoxyribonuclease; CAS, caseinase; HAL, a-haemolysin; HBE, P-haemolysin.
The ability of some CNS to produce a mucoid substance called slime was recognized two decades ago. The composition of this slime is still not well known, although mannose and galactose are considered to be the main sugar constituents of the glycosidic molecule. l3 The slime allows bacteria to Table
II. Susceptibility
of Staphylococcus MI%,
Novobiocin Cephalothin Gentamicin Kanamycin Erythromycin Vancomycin Streptomycin Methicillin Chloramphenicol Tetracycline Ampicillin Penicillin Rifampicin Clindamycin Tunicamycin Teicoplanin * I, Invasive
isolates;
C, carrier
epidermidis
@g ml-‘)
isolates to different
antibiotics
MIC,,
(pg ml-‘)
I”
C*
I
0.1 I.0 1.0 256.0 256.0 4.0 16.0 16.0 128.0 32.0 I.0 I.0 1.0 64.0 I.5 4.0
1.0 I.0 I.0 4.0 I-O 3.0 4.0 1.0 128.0 64.0 I.0 I.0 0.1 128.0 2.0 4.0
0.1 I.0 128.0 256.0 256.0 4.0 256.0 64-O 256.0 256.0 4.0 16.0 256.0 256.0 44.0 8.0
isolates.
C 1.0 1.9 1.0 16.0 1.0 4.0 256.0 56.5 256.0 256.0 2.0 4.0 0.1 128.0 2.0 8.0
112
M. J. Souto et al. Distance
O-
II
III
IV V
Figure 2. Dendrogram showing clusters formed by invasive and carrier Staphylococcus epidermidis isolates according to their enzymatic profiles. Distances were calculated by the UPGMA method and are based on the Euclidean distances between individuals.
adhere to several surfaces and colonize them, but only recently has slime production been considered as a parameter for the discrimination between strains.4 Some authors argue that, in combination with other parameters, slime production is a useful discrimination marker for the evaluation of pathogenic potential of isolates and for epidemiological investigation of infections due to CNS.r4 Our results are not in agreement with these observations because the two methods used for evaluation of slime production did not differentiate between clinical and carrier isolates. This indicates that, at least with our isolates, slim’e production is a poor predictor of virulence. Another point of
Phenotypic
characteristics
in S. epidermidis
113
interest in this respect is the lack of correlation found in our isolates between slime production measured by ConA agglutination and adherence to polystyrene; this contrasts with previously published data.8x’5 Although our conclusions are drawn from studies with a relatively limited number of strains, these authors worked with only one S. epidermidis strain and this makes it impossible to generalize their results. Differences in slime detection by the two methods employed could be explained by differences in composition of this layer between strains (ConA mainly reacts with mannose and galactose); this remains to be investigated. It has also been suggested that slime can protect staphylococcal cells from the host’s defence mechanisms and from the action of antibiotics;‘3F’6 however our data show that this is not a universal finding. We have found a greater resistance to antibiotics in invasive isolates, regardless of their slime production. On the other hand, the conditions (especially the duration of the experiments) for the assays of resistance to antibiotics did not allow the staphylococcal cells to be embedded in the slime they could produce. This higher level of resistance of invasive isolates is in agreement with the results of other workers, although it has been shown that the antibiotic susceptibility patterns can vary, not only between isolates obtained from different geographic locations,i7 but also from time to time when susceptibility is re-determined.” This is due, in part, to the lack of genetic stability, especially of plasmids; this makes susceptibility patterns less useful as epidemiological markers. Of our carrier isolates 83% had MICs <4 for methicillin whereas 96% of the invasive isolates had MICs B 8. This suggests that susceptibility to this antibiotic could be used, at least with our isolates, as a valid differential marker. This is contrary to the results of other workers.” As has been shown to be the case for Staphylococcus UUY~US~~~~’ the ability to produce disease depends on a combination of factors and whether the host is resistant to infection. Staphylococcus epidermidis produces some virulence factors similar to those produced by S. aureus; these can be used to determine the significance of certain biotypes in infections in immunocompromised hosts. Slime is not the only extracellular product that can contribute to virulence, and some soluble exoproteins, produced during growth, could act as virulence factors. Studies cited by Gemme showed significant differences in the production of extracellular enzymes (notably haemolysin, DNase, fibrinolysin) between isolates obtained from healthy carriers and infected patients. In our study, we have not found such differences with any of the enzymes tested, which indicates that enzyme production alone cannot be used to differentiate pathogenic potential of different isolates. However, as with antibiotic susceptibility, enzyme production may show important differences depending on geographic origin of the isolates. The absence of differentiating characteristics was confirmed by the application of techniques of cluster analysis. As can be seen in the
114
M. J. Souto et al.
dendrograms, there were no groups in which isolates were only carrier or only invasive, and the similarity between all isolates is over 88%, making it impossible to discriminate isolates only on the basis of these characteristics. In our opinion the pathogenicity of S. epidermidis is not inherent in phenotypic properties of the isolates but is more probably related to the immune status of the infected patient.
References DS, Massanari RN, Pfaller MA, Bale MJ, Streed SA, Hierholzer WJ Jr. 1. Davenport Usefulness of a test for slime production as a marker for clinically significant infections with coagulase-negative staphylococci. J infect Dis 1986; 153: 332-339. DG. Peritonitis. In: Nolph KD, Ed. Peritoneal Dialysis. 2. Vas SI, Low DE, Oreopulus The Hague: Martinus Nijhoff Publishers 1981; 344-365. 3. Kloos WE. Natural populations of the genus staphylococcus. Ann Rev Microbial 1980; 34: 559-592. GD, Parisi JT, Bisno AL, Simpson WA, Beachey EH. Characterization of 4. Christensen clinically significant strains of coagulase-negative staphylococci. J Clin Microbial 1983;
18: 258-269. 5. Wadstrtim T, Rozgonyi F. Virulence determinants of coagulase-negative staphylococci. In: Mardh PA, Schleifer T, Eds. Coagulase-Negative Staphylococci. Sweden: Almquist & Wiksell International 1986; 123-l 30. JT. Coagulase-negative staphylococci and the epidemiological typing of 6. Parisi,
Staphylococcus epidermidis. Microbial
Rev 1985; 49: 126-I 39.
CD, Simpson WA, Bisno AL, Beachey EH. Adherence of slime-producing 7. Christensen strains of Staphylococcus epidermidis to smooth surface. Infect Immun 1982; 37: 318-326. A, Uhlenbruck G, Peters H et al. Investigation on extracellular slime 8. Ludwicka substance produced by Staphylococcus epidermidis. Zentralbl Bakteriol Mikrobiol Hyg 1984: 258: 256-267. RM, Krieg NR. General characterization. In: Gerhardt P, Murray RGE, 9. Smibert Costilow, RN et al. Manual of Methods for General Bacteriology. American Society for
Microbiology.
Washington 198 1; 409-482.
10. Washington procedures. 11. 12. 13.
14. 15.
JA, Sutter VL. Dilution susceptibility test: agar and macro-broth dilution In: Lennette EH, Balows A, Hansler WJ, Truant JP, Eds. Manual of Clinical Microbiology. Washington: American Society for Microbiology 1980; 453-458. Pimental RA. Ordination and Cluster Analysis. In: Pimentel RA, Ed. Morphometrics. Iowa: Kendal/Hunt Publishing Co. 1979; 133-174. Kloos WE, Jorgensen JH. Staphylococci. In: Lennette A, Balows A,, Hausler WJ Jr, Shadomy HJ, Eds. Manual of Clinical Microbiology. 4th edn. Washington: American Society for Microbiology 1985; 145-153. Peters G, Shumacher-Pedrau F, Jansen B, Bey M, Pulverer G. Biology of Staphylococcus epidermidis extracellular slime. In: Pulverer G, Quie PG, Peters G, Eds. Pathogenicity and Clinical SigniJcance of Coagulase-negative Staphylococci. Stuttgart: Gustav Fischer Verlag 1987; 15-3 1. Christensen GD, Simpson WA, Bisno AL, Beachey EH. Experimental foreign body infections in mice challenged with slime-producing Staphylococcus epidermidis. Infect Immun 1983; 40: 407-419. Ludwicka A, Locci R, Jansen B, Peters G, Pulverer G. Microbial colonization of prosthetic devices. V. Attachment of coagulase-negative staphylococci and “slime” production on chemically pure synthetic polymers. Zentralbl Bakteriol Mikrobiol Hyg I
Abt Orig B 1983; 177: 527-532. 16. Gray, ED, Peters G, Verstegen M, Regelman WE. Effects of extracellular slime substance from Staphylococcus epidermidis on the cellular immune response. Lancet 1984; 1: 365-367. G, Damen G, Neugebauer M. Antibiotic resistance of Staphylococcus albus. 17. Pulverer Med Microbial Immunol 1972; 158: 32-43.
Phenotypic
characteristics
in S. epidermidis
115
18. Hebert GA, Cooksey RC, Clark NC, Hill BC, Jarris WR, Thornsberry C. Biotyping coagulase-negative staphylococci. J Clin lllicrobiol 1988; 26: 1950-1956. 19. Mickelsen PA, Plorde JJ, Gordon KP et ~2. Instability of antibiotic resistance in a strain of Staphylococcus epidermidis isolated from an outbreak of prosthetic valve endocarditis. J Infect IX 1985; 152: 50-58. 20. Males BM, Rogers WA Jr, Parisi JT. Virulence factors of biotypes of Staphylococcus epidermidis from clinical sources. J Clin Microbial 1975; 1: 256-261. 21. Gemmell CG. Pathogenicity of coagulase-negative staphylococci with respect to the nature of the host response. Zentralbl Bakteriol Mikrobiol Hyg A 1987; 266: 52-59. 22. Gemmell CG. Extracellular toxins and enzymes of coagulase-negative staphylococci. In: Easmon CSF, Adlam C, Eds. Staphylococci and Staphylococcal Infections. London: Academic Press 1983; 809427.
of Hospital
Failure
Infection
(1991)
17, 107-l
15
of phenotypic characteristics to distinguish between carrier and invasive isolates of Staphylococcus epidermidis M. J. Souto,
Departamento
C. M. Ferreirbs
and M. T. Criado
de Microbiologia y Parasitologia, Facultad Universidad de Santiago de Compostela, Spain Accepted for publication
de Farmacia,
3 September1990
Summary: Thirty carrier and 29 invasive Staphylococcus epidermidis isolates were analysed for production of slime, extracellular enzymes and antibiotic resistance. Evaluation of slime production was by two methods which gave differing results, but both showed no difference between the two groups of isolates. Exoenzyme production by the two groups of isolates was also similar, and the only differences were shown in resistance to some antibiotics; carrier isolates were more resistant to novobiocin and cephalothin, but resistance to chloramphenicol, tunicamycin, teicoplanin, penicillin, ampicillin and tetracycline was similar in both groups. Cluster analysis based on exoenzyme production showed no consistent difference between invasive and carrier isolates, with the lowest similarity coefficient over 88% (except for one isolate). Consequently, none of the characteristics studied was useful in differentiating potentially invasive isolates. Introduction
Interest in epidemiological markers that might prove useful for the identification and control of infections caused by Staphyloccocus epidermidis has increased during recent years. This microorganism, along with other coagulase-negative staphylococci (CNS), is a normal saprophyte on the skin and anterior nares of healthy individuals. Staphylococcus epidermidis is one of the CNS most frequently isolated from infections and it has been increasingly associated with nosocomial infections of prosthetic valves and hosts joints and vascular devices,’ with infections in immunocompromised and in patients subjected to aggressive treatments such as continuous ambulatory peritoneal dialysis.2 The ubiquitous presence of S. epidermidis in humans, in the skin and nares,3 and its behaviour as a nosocomial pathogen, makes it necessary to investigate methods that can be effective for the investigation of specific strains, for example study of their biochemical and antimicrobial resistance profiles. Slime production has been proposed as a tool for measuring the Correspondence
to: Prof.
C. Ferreiros.
019%6701/91/020107+09$03.00/0
0
107
1991
The
Hospital
Infection
Socwry
108
M. J. Souto et al.
clinical significance of CNS,4 and the production of exoenzymes or synergistic haemolysis’ have been suggested as being characteristic of virulence. Owing to the large number of carriers of this microorganism among hospital personnel, it is important to know if patients are infected by their own cutaneous microbial flora (normally they are compromised hosts) or by specific nosocomial strains with a characteristic pathogenic potential. Our aim in this work was to study the properties cited above in two groups of S. epidermidis isolates; one group isolated from clinical samples and the other from healthy individuals, not associated with the hospital environment in an attempt to determine the possible clinical significance of these properties and their possible use for the differentiation of strains in epidemiological studies. Material
and methods
Bacterial isolates A total of 59 isolates of S. epidermidis was studied. Thirty were isolated from healthy carriers in the Bacteriology Laboratory (Direction de Salud, La Cor&a, Spain), and 29 clinical isolates were obtained mainly from blood cultures in the Hospital General de Galicia, Spain. Identification of each isolate was by the usual techniques: Gram staining, production of catalase and coagulase negativity, and inability to ferment mannitol. Identification was confirmed with the API Staph Gallery (API Systems S.A., Montalieu Vercien, France). Biotype was determined according to the scheme described by Parisi. All isolates were maintained at 4°C on brain-heart agar (Biolife, Italy), and sub-cultured monthly. For prolonged storage the organisms were kept at - 20°C in skimmed milk. Slime production Slime production was determined by the method described by Christensen et a1.7 The organisms were grown in 3 ml trypticase soy broth in polystyrene tubes for 18 h at 37°C. The tubes were emptied and immediately stained for 60 s with 0.1% Alcian Blue. Results were recorded as ‘ + + ‘, ‘ + ‘, ‘ f ’ or ‘ - ’ depending on the degree of staining of the tube wall. Slime production was also determined by slide agglutination with Concanavalin A (ConA) according to Ludwicka et aZ.* Twenty-five ~1 of a bacterial suspension (5 X lo9 bacteria ml-‘) in phosphate-buffered saline (PBS) was mixed with 25 ~1 of serial twofold dilutions of a 1% (w/v) solution of ConA (Serva Feinbiochemica, Heidelberg, NY) in PBS. After 1 min at 20°C the results were recorded as the highest dilution of ConA causing agglutination of the bacteria. Extracellular enzymes The production of acid and alkaline phosphatase, caseinase and a- and P-haemolysins was determined
lipase, gelatinase, by the methods
Phenotypic
characteristics
in S. epidermidis
109
described by Smibert & Krieg. 9 For the lipase determination, Tween 20, Tween 40 and Tween 60 were added independently to the basal medium. Extracellular production of DNase was detected on deoxyribonuclease test medium (Biolife, Italy). Antimicrobial susceptibility The following antimicrobial agents were tested: penicillin, ampicillin, streptomycin, novobiocin, tunicamycin, tetracycline, clindamycin, teicoplanin, gentamicin, methicillin, vancomycin, rifampicin, kanamycin, chloramphenicol, cephalothin and erythromycin. For each isolate, minimal inhibitory concentrations (MICs) were determined by a standard macrodilution method.” Using a Steers replicator, 1O4colony forming units (cfu) of each isolate were deposited onto the surface of Mueller-Hinton agar plates containing serial twofold dilution of the antibiotics (from an initial concentration of 256 I.tg ml-‘). All tests were performed in duplicate. Each MIC was recorded after both 24 and 48 h of incubation at 37°C. MIC,, and MIC,, were calculated as the concentrations inhibiting 50% and 90% of isolates respectively. Statistical methods The SPSS/PC+ statistical package (SPSS Inc. and Microsoft Corp.) was used. Grouping of the strains was by cluster analysis using the unweighed pair group method of cluster analysis (UPGMA) and arithmetic averages of Euclidean distances.”
Results All isolates were identified as S. epidermidis, biotype I. Table I shows the results obtained for slime production in both carrier and invasive isolates by the methods of adherence to polystyrene (method 1) and agglutination with ConA (method 2). The methods were not equivalent in determining slime production, as tested by calculation of the Kendall’s z coefficient of concordance (z= -0.21) on the results obtained for the two groups of isolates. Neither of the two methods showed significant differences in slime production between the two groups of isolates (carrier and invasive) when tested by the Mann-Whitney U-test (P=O*SS for method 1; P=O*38 for method 2). Figure 1 shows the percentages of isolates of each group producing the different exoenzymes tested. There were no differences between carrier and invasive isolates (P>O*OS, X*-test) and none of the enzymes could be associated specifically with either of the groups. Table II shows the MIC,, for each antibiotic tested. MICs of chloramphenicol, and MIC,, ampicillin and tetracycline were tunicamycin, teicoplanin, penicillin, similar in the two groups of isolates. Carrier isolates were more resistant to
110
M. J. Souto et al. Table I. Slime production of Staphylococus
epidermidis
Invasive isolates
isolates
Carrier isolates
Method 1
Method 2
Isolate
Method 1
Method 2
E
-+
64 16
Nl
*f
328
s4
32 2 0
!G: N7 N15 N16 N17 N18 N20 N21 N23 N24 N33 N34 N35 N37 N38 N39 N41 N42
+ ++ ++ ++ f ++ ++ + + + + + + + +
32
El S18 s20 S23 S26 S27 S28 S29 s31 S32 s33 534 s35 S36
f + ++ ++ ++ + ++ ++ + + -t i + ++
S38 s37 s39 540 s41 S42 s43 s44
++ ++ + ++ ++ It
1:8 128 16 4 4 32 4
N44 N43 N45 N47 N51 N53 NS5 N57
+ &
+
: 8 8 8 4 64 8
8
N6.5 N61
-+
6:
Isolate
:z
545 Method staining; Method
+
; 0 : 32 4 8 16 1:: ; 16
+
+ + +
1: 64 4 2 32 32 16 8 4 8 4 1: 16 128 64
1, Adherence to polystyrene results graded ‘ - ‘: not seen; ‘ f ’ weak staining; ‘ + ’ intermediate ‘ + + ’ strong staining. 2, ConA agglutination: figures represent highest dilution observed for each isolate.
novobiocin and cephalothin, and invasive ones were more resistant to the remainder of the antibiotics. Figure 2 shows the clusters formed by invasive and carrier isolates when enzymatic profiles were subjected to cluster analysis by a UPGMA clustering method. Only one isolate showed a similarity coefficient less than clusters were connected at 86% similarity. 86%, and the 16 remaining Clusters at 100% similarity often contained both invasive and carrier isolates. Discussion
All S. epidermidis studied in this work, both clinical and carrier isolates, were characterized by the API Staph System, which offers a good correlation with conventional methods,12 and has also been used for the biotyping of coagulase-negative staphylococci.4
Phenotypic
FAC
F/iL
characteristics
TiO
TiO
in S. epidermidis
T&O
GiL
DiA
CiS
111
HiL
HiE
Enzymes
Figure 1. Percentages of isolates producing enzymes in carrier (solid bars) and invasive (cross-hatched bars) Staphylococcus epidermidis groups. FAC, acid phosphatase; FAL, alkaline phosphatase; T20, Tween-20 hydrolase; T40, Tween-40 hydrolase; T60, Tween-60 hydrolase; GEL, gelatinase; DNA, deoxyribonuclease; CAS, caseinase; HAL, a-haemolysin; HBE, P-haemolysin.
The ability of some CNS to produce a mucoid substance called slime was recognized two decades ago. The composition of this slime is still not well known, although mannose and galactose are considered to be the main sugar constituents of the glycosidic molecule. l3 The slime allows bacteria to Table
II. Susceptibility
of Staphylococcus MI%,
Novobiocin Cephalothin Gentamicin Kanamycin Erythromycin Vancomycin Streptomycin Methicillin Chloramphenicol Tetracycline Ampicillin Penicillin Rifampicin Clindamycin Tunicamycin Teicoplanin * I, Invasive
isolates;
C, carrier
epidermidis
@g ml-‘)
isolates to different
antibiotics
MIC,,
(pg ml-‘)
I”
C*
I
0.1 I.0 1.0 256.0 256.0 4.0 16.0 16.0 128.0 32.0 I.0 I.0 1.0 64.0 I.5 4.0
1.0 I.0 I.0 4.0 I-O 3.0 4.0 1.0 128.0 64.0 I.0 I.0 0.1 128.0 2.0 4.0
0.1 I.0 128.0 256.0 256.0 4.0 256.0 64-O 256.0 256.0 4.0 16.0 256.0 256.0 44.0 8.0
isolates.
C 1.0 1.9 1.0 16.0 1.0 4.0 256.0 56.5 256.0 256.0 2.0 4.0 0.1 128.0 2.0 8.0
112
M. J. Souto et al. Distance
O-
II
III
IV V
Figure 2. Dendrogram showing clusters formed by invasive and carrier Staphylococcus epidermidis isolates according to their enzymatic profiles. Distances were calculated by the UPGMA method and are based on the Euclidean distances between individuals.
adhere to several surfaces and colonize them, but only recently has slime production been considered as a parameter for the discrimination between strains.4 Some authors argue that, in combination with other parameters, slime production is a useful discrimination marker for the evaluation of pathogenic potential of isolates and for epidemiological investigation of infections due to CNS.r4 Our results are not in agreement with these observations because the two methods used for evaluation of slime production did not differentiate between clinical and carrier isolates. This indicates that, at least with our isolates, slim’e production is a poor predictor of virulence. Another point of
Phenotypic
characteristics
in S. epidermidis
113
interest in this respect is the lack of correlation found in our isolates between slime production measured by ConA agglutination and adherence to polystyrene; this contrasts with previously published data.8x’5 Although our conclusions are drawn from studies with a relatively limited number of strains, these authors worked with only one S. epidermidis strain and this makes it impossible to generalize their results. Differences in slime detection by the two methods employed could be explained by differences in composition of this layer between strains (ConA mainly reacts with mannose and galactose); this remains to be investigated. It has also been suggested that slime can protect staphylococcal cells from the host’s defence mechanisms and from the action of antibiotics;‘3F’6 however our data show that this is not a universal finding. We have found a greater resistance to antibiotics in invasive isolates, regardless of their slime production. On the other hand, the conditions (especially the duration of the experiments) for the assays of resistance to antibiotics did not allow the staphylococcal cells to be embedded in the slime they could produce. This higher level of resistance of invasive isolates is in agreement with the results of other workers, although it has been shown that the antibiotic susceptibility patterns can vary, not only between isolates obtained from different geographic locations,i7 but also from time to time when susceptibility is re-determined.” This is due, in part, to the lack of genetic stability, especially of plasmids; this makes susceptibility patterns less useful as epidemiological markers. Of our carrier isolates 83% had MICs <4 for methicillin whereas 96% of the invasive isolates had MICs B 8. This suggests that susceptibility to this antibiotic could be used, at least with our isolates, as a valid differential marker. This is contrary to the results of other workers.” As has been shown to be the case for Staphylococcus UUY~US~~~~’ the ability to produce disease depends on a combination of factors and whether the host is resistant to infection. Staphylococcus epidermidis produces some virulence factors similar to those produced by S. aureus; these can be used to determine the significance of certain biotypes in infections in immunocompromised hosts. Slime is not the only extracellular product that can contribute to virulence, and some soluble exoproteins, produced during growth, could act as virulence factors. Studies cited by Gemme showed significant differences in the production of extracellular enzymes (notably haemolysin, DNase, fibrinolysin) between isolates obtained from healthy carriers and infected patients. In our study, we have not found such differences with any of the enzymes tested, which indicates that enzyme production alone cannot be used to differentiate pathogenic potential of different isolates. However, as with antibiotic susceptibility, enzyme production may show important differences depending on geographic origin of the isolates. The absence of differentiating characteristics was confirmed by the application of techniques of cluster analysis. As can be seen in the
114
M. J. Souto et al.
dendrograms, there were no groups in which isolates were only carrier or only invasive, and the similarity between all isolates is over 88%, making it impossible to discriminate isolates only on the basis of these characteristics. In our opinion the pathogenicity of S. epidermidis is not inherent in phenotypic properties of the isolates but is more probably related to the immune status of the infected patient.
References DS, Massanari RN, Pfaller MA, Bale MJ, Streed SA, Hierholzer WJ Jr. 1. Davenport Usefulness of a test for slime production as a marker for clinically significant infections with coagulase-negative staphylococci. J infect Dis 1986; 153: 332-339. DG. Peritonitis. In: Nolph KD, Ed. Peritoneal Dialysis. 2. Vas SI, Low DE, Oreopulus The Hague: Martinus Nijhoff Publishers 1981; 344-365. 3. Kloos WE. Natural populations of the genus staphylococcus. Ann Rev Microbial 1980; 34: 559-592. GD, Parisi JT, Bisno AL, Simpson WA, Beachey EH. Characterization of 4. Christensen clinically significant strains of coagulase-negative staphylococci. J Clin Microbial 1983;
18: 258-269. 5. Wadstrtim T, Rozgonyi F. Virulence determinants of coagulase-negative staphylococci. In: Mardh PA, Schleifer T, Eds. Coagulase-Negative Staphylococci. Sweden: Almquist & Wiksell International 1986; 123-l 30. JT. Coagulase-negative staphylococci and the epidemiological typing of 6. Parisi,
Staphylococcus epidermidis. Microbial
Rev 1985; 49: 126-I 39.
CD, Simpson WA, Bisno AL, Beachey EH. Adherence of slime-producing 7. Christensen strains of Staphylococcus epidermidis to smooth surface. Infect Immun 1982; 37: 318-326. A, Uhlenbruck G, Peters H et al. Investigation on extracellular slime 8. Ludwicka substance produced by Staphylococcus epidermidis. Zentralbl Bakteriol Mikrobiol Hyg 1984: 258: 256-267. RM, Krieg NR. General characterization. In: Gerhardt P, Murray RGE, 9. Smibert Costilow, RN et al. Manual of Methods for General Bacteriology. American Society for
Microbiology.
Washington 198 1; 409-482.
10. Washington procedures. 11. 12. 13.
14. 15.
JA, Sutter VL. Dilution susceptibility test: agar and macro-broth dilution In: Lennette EH, Balows A, Hansler WJ, Truant JP, Eds. Manual of Clinical Microbiology. Washington: American Society for Microbiology 1980; 453-458. Pimental RA. Ordination and Cluster Analysis. In: Pimentel RA, Ed. Morphometrics. Iowa: Kendal/Hunt Publishing Co. 1979; 133-174. Kloos WE, Jorgensen JH. Staphylococci. In: Lennette A, Balows A,, Hausler WJ Jr, Shadomy HJ, Eds. Manual of Clinical Microbiology. 4th edn. Washington: American Society for Microbiology 1985; 145-153. Peters G, Shumacher-Pedrau F, Jansen B, Bey M, Pulverer G. Biology of Staphylococcus epidermidis extracellular slime. In: Pulverer G, Quie PG, Peters G, Eds. Pathogenicity and Clinical SigniJcance of Coagulase-negative Staphylococci. Stuttgart: Gustav Fischer Verlag 1987; 15-3 1. Christensen GD, Simpson WA, Bisno AL, Beachey EH. Experimental foreign body infections in mice challenged with slime-producing Staphylococcus epidermidis. Infect Immun 1983; 40: 407-419. Ludwicka A, Locci R, Jansen B, Peters G, Pulverer G. Microbial colonization of prosthetic devices. V. Attachment of coagulase-negative staphylococci and “slime” production on chemically pure synthetic polymers. Zentralbl Bakteriol Mikrobiol Hyg I
Abt Orig B 1983; 177: 527-532. 16. Gray, ED, Peters G, Verstegen M, Regelman WE. Effects of extracellular slime substance from Staphylococcus epidermidis on the cellular immune response. Lancet 1984; 1: 365-367. G, Damen G, Neugebauer M. Antibiotic resistance of Staphylococcus albus. 17. Pulverer Med Microbial Immunol 1972; 158: 32-43.
Phenotypic
characteristics
in S. epidermidis
115
18. Hebert GA, Cooksey RC, Clark NC, Hill BC, Jarris WR, Thornsberry C. Biotyping coagulase-negative staphylococci. J Clin lllicrobiol 1988; 26: 1950-1956. 19. Mickelsen PA, Plorde JJ, Gordon KP et ~2. Instability of antibiotic resistance in a strain of Staphylococcus epidermidis isolated from an outbreak of prosthetic valve endocarditis. J Infect IX 1985; 152: 50-58. 20. Males BM, Rogers WA Jr, Parisi JT. Virulence factors of biotypes of Staphylococcus epidermidis from clinical sources. J Clin Microbial 1975; 1: 256-261. 21. Gemmell CG. Pathogenicity of coagulase-negative staphylococci with respect to the nature of the host response. Zentralbl Bakteriol Mikrobiol Hyg A 1987; 266: 52-59. 22. Gemmell CG. Extracellular toxins and enzymes of coagulase-negative staphylococci. In: Easmon CSF, Adlam C, Eds. Staphylococci and Staphylococcal Infections. London: Academic Press 1983; 809427.