INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Heterogeneous expression of resistance among staphylococci is a well-known feature of methicillin resistance and has long been a major theme of investigation in these organisms (6). More recently, experimental studies have clearly documented that glycopeptide resistance can also be expressed heterogeneously by methicillin-resistant strains of both Staphylococcus aureus (1, 13, 25) and coagulase-negative staphylococci (26-28, 30); this heterogeneous resistance has been associated with failures of vancomycin therapy (13, 25, 26). Since earlier reports, however, selection of relatively stable glycopeptide-resistant subpopulations had been suggested as a probable explanation of therapeutic failures in infections due to S. haemolyticus in patients subjected to prolonged vancomycin treatment (23, 31).

Among coagulase-negative staphylococci, resistance to glycopeptide antibiotics (usually to teicoplanin more than to vancomycin) is expressed especially by species such as S. haemolyticus and S. epidermidis (4). Glycopeptide-resistant cells of clinical and laboratory-derived strains of S. haemolyticus and S. epidermidis have been shown to differ from their glycopeptide-susceptible counterparts in several features, including ultrastructural morphology (3, 11, 20, 24), glycopeptide-binding capacity (21, 28), membrane proteins (18), cell wall synthesis and composition (5), and susceptibility to cell wall-active antibiotics and enzymes (7, 9). Reduced and heterogeneous susceptibility to teicoplanin in methicillin-resistant coagulase-negative staphylococci appears to be on the increase in some hospital settings (28). The present study focused on the teicoplanin and vancomycin susceptibility profiles of 12 bloodstream isolates of S. haemolyticus and, by comparison, of six bloodstream isolates of other staphylococcal species. As regards S. haemolyticus, which was the object of our interest, it should be noted that (i) among clinical isolates of coagulase-negative staphylococci, this species is second in frequency only to S. epidermidis (14); (ii) it has been found to be highly prevalent in the hospital environment and on the hands of health care workers (19); (iii) it is regarded as an important nosocomial pathogen with a tendency to develop multiple resistances (10); (iv) isolates of this species acquired glycopeptide resistance even earlier than enterococci and other staphylococcal species (2, 8, 23, 33); (v) since early studies, S. haemolyticus was suggested to be unique among coagulase-negative staphylococci in being predisposed to develop glycopeptide resistance (24); and (vi) in cultures of this staphylococcal species, often more than in those of others, glycopeptides can, under laboratory conditions, select for clones for which the glycopeptide (especially teicoplanin) MICs are increased (3, 9, 12, 22, 23, 28, 31, 32).

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Bacterial strains. The test strains used in this study are listed in Table 1. Twelve S. haemolyticus strains were independently isolated from blood cultures between 1982 and 1997 in our and other clinical laboratories serving various hospitals in the Ancona area. Six blood culture isolates of other Staphylococcus species (three of S. aureus, two of S. epidermidis, and one of S. simulans), for all of which the teicoplanin MICs were 4 µg/ml, were isolated between 1982 and 1998 in the same laboratories. All isolates were identified by the API test system (BioMérieux, Marcy-l'Etoile, France), and their identification was confirmed by additional laboratory tests (14).

Antibiotics. Teicoplanin and vancomycin were obtained from Hoechst Marion Roussel, Lainate, Italy, and Eli Lilly Italia, Sesto Fiorentino, Italy, respectively. Oxacillin was purchased from Sigma Chemical Co., St. Louis, Mo.

MIC tests. Teicoplanin, vancomycin, and oxacillin MICs were determined in accordance with the standard broth microdilution procedures recommended by the National Committee for Clinical Laboratory Standards (17), using Mueller-Hinton II broth (BBL Microbiology Systems, Cockeysville, Md.) as the test medium. Antibiotics were tested at final concentrations (prepared from serial twofold dilutions) ranging from 0.03 to 64 µg/ml. The inoculum was 5 × 10⁵ CFU/ml. The inoculated trays were incubated at 37°C for 24 h. S. aureus ATCC 29213 was used for quality control. We used the MIC breakpoints suggested by the National Committee for Clinical Laboratory Standards (17) for teicoplanin (susceptible, 8 µg/ml; intermediate, 16 µg/ml; resistant, 32 µg/ml), vancomycin (susceptible, 4 µg/ml; intermediate, 8 to 16 µg/ml; resistant, 32 µg/ml), and oxacillin (susceptible, 2 µg/ml; resistant, 4 µg/ml).

Population analysis. Population analysis profiles (PAPs) (29) were done by plotting colony counts against teicoplanin and vancomycin concentrations. Bacteria were grown overnight in tryptic soy (TS) broth (Difco Laboratories, Detroit, Mich.) at 37°C and then plated in duplicate at a series of dilutions on TS agar (Difco) containing antibiotic-free medium or twofold dilutions of the test antibiotic within the concentration range of 0.5 to 512 µg/ml. Plates were incubated at 37°C for 48 h, and the CFU were counted. The teicoplanin susceptibility profiles of S. haemolyticus isolates were also determined after overnight growth in TS broth supplemented with a subinhibitory teicoplanin concentration (one-fourth of the MIC). The lowest antibiotic concentration inhibiting 99.9% of growth (3-log₁₀ decrease in the number of CFU) was defined as the PAP MIC (28). The homogeneity or heterogeneity of the antibiotic susceptibility phenotype was established from the appearance of the curve, homogeneity usually being characterized by a steep slope which followed an almost horizontal course at the permissive drug concentrations and heterogeneity usually being characterized by one or more inflection points.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Antibiotic susceptibility. In vitro susceptibilities, in terms of the conventional MICs of teicoplanin, vancomycin, and oxacillin for the 18 bloodstream isolates studied, are reported in Table 1. Of the 12 S. haemolyticus strains, 4 were susceptible (MICs, 1 to 8 µg/ml) and 8 were resistant (MICs, 32 to 64 µg/ml) to teicoplanin while all were susceptible to vancomycin (MICs, 1 to 2 µg/ml). Of the six test strains of other Staphylococcus species (for all of which the teicoplanin MICs were 4 µg/ml), five were teicoplanin susceptible (MICs, 4 to 8 µg/ml) and one, an S. epidermidis isolate, was teicoplanin resistant (MIC, 32 µg/ml) while all were susceptible to vancomycin (MICs, 1 to 2 µg/ml). Oxacillin resistance was recorded in two of the four teicoplanin-susceptible and seven of the eight teicoplanin-resistant S. haemolyticus isolates, in two of the three S. aureus isolates, in one of the two S. epidermidis isolates, and in the S. simulans isolate.

PAPs of teicoplanin-susceptible S. haemolyticus isolates. The PAPs of the four strains for which the conventional teicoplanin MICs were within the range of susceptibility are reported in Fig. 1, and the frequencies of cells growing at high glycopeptide concentrations are indicated in Table 2 (teicoplanin) and Table 3 (vancomycin).

View larger version (24K): [in this window] [in a new window]   FIG. 1.   PAPs for glycopeptides of the four S. haemolyticus isolates for which the broth microdilution MICs of teicoplanin were in the range of susceptibility (8 µg/ml) (17). After overnight growth in TS broth, bacteria were plated on TS agar containing the indicated concentrations of teicoplanin () or vancomycin (). Population analysis for teicoplanin was also done after overnight growth in TS broth containing a subinhibitory concentration of teicoplanin ().

5

8

PAPs of teicoplanin-resistant S. haemolyticus isolates. The PAPs of the eight strains for which the conventional teicoplanin MICs were within the range of resistance are reported in Fig. 2, and the frequencies of cells growing at high glycopeptide concentrations are indicated in Table 2 (teicoplanin) and Table 3 (vancomycin).

View larger version (30K): [in this window] [in a new window]   FIG. 2.   PAPs of the eight S. haemolyticus isolates for which the broth microdilution MICs of teicoplanin were in the range of resistance (32 µg/ml) (17). After overnight growth in TS broth, bacteria were plated on TS agar containing the indicated concentrations of teicoplanin () or vancomycin (). Population analysis for teicoplanin was also done after overnight growth in TS broth containing a subinhibitory concentration of teicoplanin ().

2

3

5

6

6

7

Comparison of conventional MICs and PAP data. To evaluate possible correlations between resistant subpopulations and MICs in each of our S. haemolyticus strains, we compared the teicoplanin and vancomycin MICs obtained by the standard broth microdilution (BMD) method as recommended by the National Committee for Clinical Laboratory Standards (17) (BMD MICs) with the antibiotic concentrations inhibiting the growth of the majority of cells as derived from PAPs (PAP MICs) and correlated MIC data with the frequency of highly resistant subpopulations.

PAPs of strains of other staphylococcal species. None of the six bloodstream isolates of other Staphylococcus species exhibited heterogeneous teicoplanin susceptibility profiles. As regards vancomycin, three strains (AS95A, AS95B, and SS96) demonstrated homogeneous susceptibility while the three other strains showed curves with inflection points, but cells growing on 8 µg of vancomycin per ml were not found (AS98 and ES82) or were present at a frequency as low as 10⁸ (ER94) (Table 3). For all six strains, the PAP MICs of both teicoplanin and vancomycin were identical to or twice the BMD MICs (Tables 2 and 3).

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

Our results suggest that heterogeneous expression of teicoplanin resistance is prevalent among S. haemolyticus strains and may be associated with heterogeneous resistance to vancomycin. All test strains of this species for which the conventional teicoplanin MICs were within the range of susceptibility and the majority of those for which the conventional MICs were within the range of resistance showed heterogeneous teicoplanin phenotypes. On the other hand, while all isolates were susceptible to vancomycin on the basis of conventional MIC tests and breakpoints, population analysis indicated homogeneous susceptibility to the antibiotic in all four S. haemolyticus strains conventionally susceptible to teicoplanin but in only one of the eight strains conventionally resistant to teicoplanin. The seven other strains of the latter group showed heterogeneous vancomycin profiles: all contained subpopulations growing on 8 µg of vancomycin per ml, and two also contained subpopulations that grew on 16 µg/ml. Overall, our results are in agreement with those reported by Sieradzki et al. (26-28), who, however, found heterogeneous phenotypes for teicoplanin in clinical isolates not only of S. haemolyticus (28) but also of S. epidermidis (26-28) and S. hominis (28) and identified heterogeneous phenotypes for vancomycin also in isolates for which the conventional MICs of teicoplanin were within the range of susceptibility (27). It is worth noting, however, that heterogeneous phenotypes for teicoplanin, but not for vancomycin, have been identified in some "historical" isolates of S. epidermidis and S. haemolyticus collected between 1925 and 1964 (28).

Heterogeneous expression of glycopeptide resistance has recently been reported in clinical S. aureus isolates with reduced susceptibility to glycopeptides (1, 13, 25). In particular, heterogeneous resistance to vancomycin was defined by Hiramatsu et al. (13) as that observed in strains for which the conventional vancomycin MICs are 4 µg/ml but that generate subpopulations capable of growing in plates containing 8 µg of vancomycin per ml at a frequency of 10⁶. This definition applies to six of our seven S. haemolyticus isolates with heterogeneous vancomycin profiles, all of which contained subpopulations growing on 8 µg of vancomycin per ml at frequencies of 10⁵ to 10⁶; the seventh strain contained a subpopulation that grew on 8 µg of vancomycin per ml but at a frequency of only 10⁷.

It is noteworthy that the clinical staphylococci with heterogeneous glycopeptide resistance profiles reported thus far (1, 13, 15, 25-28, 34) were all methicillin resistant, whereas 3 (2 teicoplanin susceptible and 1 teicoplanin resistant on the basis of conventional MICs) of our 12 S. haemolyticus isolates were methicillin susceptible with homogeneous oxacillin susceptibility profiles (data not shown). The fact that these three methicillin-susceptible strains were fully comparable to the methicillin-resistant strains in their heterogeneous teicoplanin resistance, with one (HR97) even heterogeneously resistant to vancomycin, does not support the hypothesis, advanced for S. aureus strains (15, 16), of a possible relationship between heteroresistance to glycopeptides and heteroresistance to methicillin. On the other hand, the levels of resistance to glycopeptides and -lactams often denote a reverse relationship (7, 9).

Resistant subpopulations selected during glycopeptide therapy have been thought to be responsible for therapeutic failures in patients infected by S. aureus (13, 25), S. epidermidis (26), or S. haemolyticus (23, 31) strains initially demonstrating glycopeptide susceptibility by conventional MIC tests but showing heterogeneous resistance on population analysis. We have little information about antibiotic treatments and outcomes for our S. haemolyticus bloodstream isolates. However, in all of these strains, including those demonstrating conventional susceptibility to teicoplanin, growth in the presence of a subinhibitory teicoplanin concentration clearly selected glycopeptide-resistant subpopulations. A propensity for such a selection appears to be a species-related feature of S. haemolyticus strains. Further studies should address the issue of the stability of the selected subpopulations, as we often noted a reversion toward the original population distribution after growth in the absence of the selecting drug (data not shown). In conclusion, susceptibility to glycopeptides as determined by standard dilution methods is unlikely to represent a reliable basis for glycopeptide treatment of S. haemolyticus infections. But PAP MICs, in the absence of a careful evaluation of population profiles, may also correlate poorly with the presence of glycopeptide-resistant subpopulations of S. haemolyticus strains.