|Home | About | Journals | Submit | Contact Us | Français|
Meropenem heteroresistance was investigated in six apparently meropenem-susceptible, Klebsiella pneumoniae carbapenemase (KPC)-producing K. pneumoniae (KPC-KP) clinical isolates, compared with that in carbapenemase-negative, meropenem-susceptible controls. In population analyses, the KPC-KP isolates grew at meropenem concentrations of 64 to 256 μg/ml. Heteroresistant colonies had significantly elevated expression of the blaKPC gene compared with the native populations but did not retain heteroresistance when subcultured in drug-free media. Time-kill assays indicated that meropenem alone was not bactericidal against KPC-KP but efficiently killed the control strains.
Since the beginning of the last decade, Klebsiella pneumoniae carbapenemase (KPC)-producing K. pneumoniae (KPC-KP) isolates have been increasingly detected in the United States and subsequently in several regions worldwide (3, 4, 13, 17, 21). KPC enzymes efficiently hydrolyze all β-lactam molecules (1, 22), conferring various levels of resistance to all β-lactam compounds, including carbapenems (13). However, KPC-producing K. pneumoniae may appear susceptible to carbapenems, mainly meropenem (2, 13), by reference CLSI agar dilution or broth microdilution methods as well as by automated systems (6, 15, 17). Characteristically, it has been reported that automated systems may identify as many as 87% of KPC-KP isolates to be susceptible to meropenem (13). The detection of the susceptibility level of KPC-KP isolates to carbapenems has been shown to be difficult due to the phenotypic heterogeneity that they commonly exhibit (3, 10, 13). For instance, in agar diffusion methods such as disk diffusion or Etest, the heterogeneous growth to carbapenems of KPC-KP results in the appearance of scattered colonies within the inhibition zones (9, 13).
These issues raise the need for cautious evaluation of susceptibility testing in KPC-KP isolates that are recovered in clinical laboratories. In our clinical laboratories, several KPC-KP isolates that appear susceptible by automated susceptibility assays or reference dilution assays contain heterogeneous subpopulations (D. Sofianou and K. Themeli-Digalaki, personal communications). It has been also shown that among Greek KPC-KP isolates, meropenem tends to exhibit lower MICs than imipenem or ertapenem (17, 20). In that respect, the aim of the present study was to characterize the heterogeneous mode of growth of apparently meropenem-susceptible KPC-KP clinical isolates by population analyses and bactericidal assays.
Six KPC-2-producing K. pneumoniae clinical isolates from our laboratory collection were randomly selected for the study among those that were meropenem susceptible by agar dilution (MIC ≤ 4 μg/ml) (5) but exhibited scattered heterogeneous colonies around carbapenem discs. The isolates were recovered from separate patients hospitalized in four hospitals located in different Greek regions. K. pneumoniae ATCC 13883, Pseudomonas aeruginosa ATCC 27853, and one meropenem-susceptible, non-carbapenemase-producing K. pneumoniae clinical isolate (KP 208) were used as controls. MICs of carbapenems, aminoglycosides, and other antimicrobials were determined by agar dilution and broth macrodilution (5). Pulsed-field gel electrophoresis (PFGE) of SpeI-digested genomic DNA of the KPC-KP strains was performed with a CHEF-DRIII system (Bio-Rad, Hemel Hempstead, United Kingdom) (17).
Population analyses utilizing meropenem were performed by spreading approximately 108 bacterial CFU on Mueller-Hinton agar plates with meropenem in serial dilutions for concentrations ranging from 0.25 to 256 μg/ml and incubating the plates for 48 h (16). The analysis was performed twice for all isolates, and the mean values of viable CFU were estimated and plotted on a semilogarithmic graph. The frequency of heteroresistant subpopulations at the highest drug concentration was calculated by dividing the number of colonies grown on antibiotic-containing plates by the colony counts from the same bacterial inoculum plated onto antibiotic-free plates. The stability of meropenem MICs for three distinct colonies grown at the highest drug concentration was determined by agar dilution after seven daily subcultures in antibiotic-free medium.
Killing curves utilizing meropenem, gentamicin, or meropenem plus gentamicin were performed as previously described (8). Briefly, tubes containing cation-supplemented Mueller-Hinton broth with antibiotics at concentrations equal to the MICs of the isolates were seeded with a log-phase bacterial inoculum of 5 × 106 CFU/ml (14) and incubated in a shaking water bath at 37°C. Samples (50 μl) of this broth culture were removed after 1, 3, 6, 9, 12, and 24 h for enumeration of viable bacteria by serial dilution and plating on antibiotic-free Mueller-Hinton agar plates. The analysis was performed three times for all isolates; the mean values of viable CFU were estimated and plotted on a semilogarithmic graph, and a 24-h time-kill curve was constructed for each isolate. Bactericidal activity (99.9% killing) was defined as a ≥3 log 10 CFU/ml reduction in viable cell counts with respect to the original inoculum (12, 14).
PCR detection and sequencing of blaKPC-2 genes had been performed previously. Quantitative real-time reverse transcriptase PCR (qRT-PCR) for the expression of blaKPC-2 was applied to the native populations and the colonies grown at the highest meropenem concentrations, collected directly for testing from the plates and also after 1-week subcultures in drug-free medium. The qRT-PCR was performed as described previously (11), using the 16S rRNA gene to standardize the expression level results (7). The measurements were done in triplicate using an MX3005P instrument (Stratagene, La Jolla, CA) with Brilliant SYBR green (Qiagen, Hilden, Germany). Gene expression in heteroresistant subpopulations was expressed as fold of increment relative to those of the respective native populations which were set as calibrators. Control reactions of untranscribed RNA were also included to check for contaminating DNA.
The meropenem Etest MICs of the KPC-KP isolates ranged from 1 to 4 μg/ml, but colonies grown within the zone of inhibition in Etest were observed at meropenem concentrations of 6 to 32 μg/ml. The KPC-KP isolates had imipenem, meropenem, and ertapenem agar dilution MICs of 4 to 16 μg/ml, 2 to 4 μg/ml, and 8 to >32 μg/ml, respectively. They exhibited resistance to most antimicrobial classes and were susceptible only to colistin, tigecycline, and gentamicin. PFGE analysis showed two distinct genotypes; type I included five isolates with three subtypes, while type II had a single isolate (Table (Table11).
Population analysis assays for meropenem showed that 2.3 × 10−5 to 1.5 × 10−8 of the initial inoculum grew at concentrations ranging from 64 to 256 μg/ml for all six KPC-KP isolates (16 to 128 times their MICs), while the control isolates grew up to four times their initial meropenem MICs (Table (Table1).1). The dilution meropenem MICs of the heteroresistant colonies after 1-week subcultures in antibiotic-free medium ranged from 4 to 8 μg/ml, being similar to or slightly higher than those of the parental isolates (Table (Table1).1). qRT-PCR showed a statistically significant increase in the expression of the blaKPC-2 gene between the native and the meropenem-heteroresistant subpopulations in all isolates (ranging from 0.475-fold to 1.07-fold, P < 0.01), while the expression of blaKPC-2 did not differ between native and heteroresistant colonies after seven subcultures in drug-free medium. It is therefore probable that the expression levels of blaKPC have contributed to the heteroresistant phenotype.
Time-killing studies utilizing the respective antimicrobials at 1× MIC showed that meropenem when tested alone did not demonstrate bactericidal activity against the KPC-KP study isolates. Although a moderate reduction in viable cell counts was detected between 1 and 6 h of incubation, a substantial bacterial regrowth was detected in all KPC-KP isolates (Fig. (Fig.11 B, panel I). Gentamicin was bactericidal against all KPC-KP isolates tested (Fig. (Fig.1B,1B, panel II), while the combination of meropenem and gentamicin killed more rapidly all KPC-KP isolates (Fig. (Fig.1B,1B, panel III). For gentamicin and meropenem with gentamicin, a bacterial regrowth was observed at 24 h in all KPC-KP isolates (Fig. (Fig.1B,1B, panels II and III). The control strains were rapidly killed by all antibiotics, alone or in combination, and no bacterial regrowth was observed.
Carbapenem heteroresistance has been characterized previously in several Gram-negative rods, such as Acinetobacter baumannii and P. aeruginosa (8, 15). KPC-producing Enterobacteriaceae commonly exhibit a heterogeneous mode of growth to carbapenems, rendering difficult their detection and the estimation of their susceptibility levels (18). Population analyses in our collection of KPC-KP revealed that a proportion of the initial bacterial inocula consisted of heteroresistant subpopulations that grew at considerably higher meropenem concentrations. However, after subcultures in antibiotic-free medium, the colonies grown at the highest meropenem concentration exhibited MICs similar to those of the native isolates. The findings of the present study have shown that the growth of KPC-KP at high meropenem concentrations might be attributed to a temporary overexpression of the blaKPC-2 gene, to the heavy bacterial inocula used in population analyses, or to other undefined mechanisms. Moreover, meropenem bactericidal assays revealed in all isolates a less pronounced killing that was followed by a substantial regrowth. This regrowth could be also due to a proportion of the meropenem-heteroresistant subpopulations that survived the meropenem exposure. The combination of meropenem and gentamicin at 1× MIC resulted in an effective bacterial killing of the isolates and this could have significant implications for the treatment of infections caused by apparently meropenem-susceptible KPC-KP strains.
The detection of KPC-KP isolates poses a number of difficulties, since it cannot be based on reference dilution assays. It is shown that a low proportion of the initial bacterial inocula may grow at considerably higher meropenem concentrations. The CLSI agar dilution using 104 CFU/spot or the broth microdilution using 5 × 105 CFU/ml may identify such strains as meropenem susceptible. In that respect, phenotypic assays easily applicable in the clinical laboratory, such as the boronic acid disc assay (19) or, for larger laboratories, molecular methods, are needed for the effective identification of KPC-KP isolates.
Overall, the present study suggests that apparently meropenem-susceptible KPC-KP isolates may contain resistant subpopulations, which could survive under treatment with meropenem alone, possibly leading to treatment failures.
Published ahead of print on 26 May 2010.