Effects of Daptomycin, Linezolid, and Vancomycin on Establishment of Intestinal Colonization with Vancomycin-Resistant Enterococci and Extended-Spectrum-β-Lactamase-Producing Klebsiella pneumoniae in Mice

Abstract

In mice, vancomycin and linezolid treatment disrupted the anaerobic intestinal microflora, based on denaturing gradient gel electrophoresis analysis, and promoted colonization by Klebsiella pneumoniae and vancomycin-resistant enterococci. However, the effects varied depending on dose and duration of treatment. Daptomycin treatment did not disrupt the anaerobic microflora or promote either pathogen.

In case-control studies, exposure to vancomycin has been associated with colonization and infection due to vancomycin-resistant enterococci (VRE) (8). Oral or parenteral vancomycin administration has also been shown to promote persistent overgrowth of VRE in the stool of colonized mice (3). It is not known whether other agents that are available for treatment of methicillin-resistant gram-positive pathogens (e.g., daptomycin, linezolid) also promote VRE colonization. Because linezolid and daptomycin possess inhibitory activity against most VRE strains (1, 5), we used a mouse model to test the hypothesis that these agents promote establishment of VRE colonization to a lesser degree than vancomycin. We also examined the effect of the antibiotics on establishment of colonization by extended-spectrum β-lactamase (ESBL)-producing Klebsiella pneumoniae, a pathogen with intrinsic resistance to each of the antibiotics.

The VRE strain used in this study (C68) is a clinical VanB isolate that we have used for previous mouse colonization studies (3). The MICs for C68 were 256 μg/ml with vancomycin, 0.25 μg/ml with daptomycin, and 1 μg/ml with linezolid. K. pneumoniae strain P62 produces an SHV ESBL and has been used in previous mouse studies (9).

Two doses of the antibiotics were studied, as follows: 1 or 12.5 mg/day for vancomycin, 0.6 or 7.5 mg/day for linezolid, and 0.12 or 1.5 mg/day for daptomycin. The lower dose of antibiotic was based on the daily dose recommended for human adults (in milligrams per kilogram of body weight), and the higher dose was the human equivalent dose calculated by the technique of Freireich et al. (6). The daptomycin, vancomycin, and linezolid doses were calculated based on 6, 33.3, and 20 mg/kg human doses, respectively. All antibiotics were administered subcutaneously once each day in 0.2 ml of phosphate-buffered saline, with the exception of the higher dose of linezolid, which was administered by orogastric gavage once each day because the formulations available were not conducive to subcutaneous dosing (i.e., the concentration of the intravenous formulation would have resulted in an inordinately high volume for subcutaneous administration).

The experimental protocol was approved by the Cleveland Veterans Affairs Medical Center's Animal Care Committee. Female CF1 mice (Harlan Sprague-Dawley, Indianapolis, Ind.) weighing 25 to 30 g were housed individually. Mice received treatment with vancomycin, linezolid, or daptomycin for 2 days before and 4 days after orogastric inoculation of 10⁴ or 10⁸ CFU of the pathogens (the day of orogastric inoculation was designated day 0). Separate experiments were conducted with VRE and K. pneumoniae. A subcutaneous clindamycin treatment (1.4 mg/day) group was included as a positive control for some experiments. The organisms were suspended in 0.5 ml of phosphate-buffered saline and administered using a stainless steel feeding tube (Perfektum; Popper & Sons, New Hyde Park, N.Y.). The densities of VRE and ESBL-producing K. pneumoniae in stool were measured on day 0 (prior to orogastric inoculation of the pathogens) and on days 1, 3, and 6 after orogastric inoculation as previously described (2, 5). The concentrations of antibiotics in stool were determined on day 4 of antibiotic treatment by an agar diffusion assay with Clostridium perfringens as the indicator strain (12). To confirm that daptomycin was systemically absorbed, the same assay was used to measure the concentration of this agent in the urine of five mice that received the higher treatment dose. For the experiments with the higher doses of antibiotics, stool samples were collected after 6 days of treatment for analysis of stool microflora using denaturing gradient gel electrophoresis (DGGE) of PCR-amplified bacterial rRNA genes as previously described (13). For the experiments with the lower doses of antibiotics, an additional set of experiments was conducted, in which vancomycin or daptomycin treatment was initiated 6 days prior to orogastric inoculation of the pathogens. All experiments were performed twice with three mice in each group (for a total of six mice per group), and the data were combined.

Data analyses were performed with the use of Stata software (version 6.0; College Station, Texas). A one-way analysis of variance was performed to compare the groups, with P values adjusted for multiple comparisons using the Scheffe correction.

The effect of the higher antibiotic doses on the establishment of colonization with the pathogens is shown in Fig. 1. At baseline, none of the mice had detectable VRE or ceftazidime-resistant gram-negative bacilli in the stool (level of detection, ∼2 log₁₀ CFU/g). Vancomycin promoted overgrowth of VRE in comparison to saline controls (P < 0.001), whereas daptomycin did not (P = 0.64). Linezolid promoted overgrowth of VRE after the 10⁸ CFU inoculum (P = 0.03) but not the 10⁴ inoculum (P = 0.82). Vancomycin and linezolid (and clindamycin controls) each promoted overgrowth of ESBL-producing K. pneumoniae (P < 0.001), whereas daptomycin did not (P = 0.37). The mean concentrations of vancomycin and linezolid in stool on day 4 of treatment were 183 μg/g (range, 0 to 400 μg/g) and 810 μg/g (range, 600 to 1,400 μg/g), respectively. Daptomycin was not detectable in any of the stool samples tested (limit of detection, 6 μg/g). Subsequent experiments demonstrated that five of five mice treated with this dose of daptomycin had high concentrations of drug in their urine (>100 μg/ml) and therefore that the daptomycin was systemically absorbed. DGGE demonstrated that vancomycin and linezolid treatment caused marked disruption of the stool microflora (mean similarity indices of 15% and 11%, respectively, in comparison to saline controls), whereas daptomycin did not (mean similarity index of 68% in comparison to saline controls) (Fig. 2). Preliminary analyses in our laboratory demonstrate that most of the bands in the DGGE patterns of control mice represent anaerobic bacteria, with a predominance of bacteroides and clostridia (11).

FIG. 1.

Effect of higher doses of vancomycin, linezolid, and daptomycin on the establishment of colonization with VRE (A and B) or ESBL-producing Klebsiella pneumoniae (C) in mice. Clindamycin was included as a positive control. Mice received antibiotic treatment once daily from day 2 before infection to day 4 after infection, and 10⁴ or 10⁸ CFU of the pathogens were administered by orogastric gavage on day 0. If the pathogens were not detected in the stool, the lower limit of detection (∼2 log₁₀ CFU/g) was assigned. Filled square, vancomycin; filled circle, linezolid; open square, normal saline; open circle, daptomycin; open triangle, clindamycin. *, linezolid was administered orally; the other antibiotics were administered subcutaneously.

FIG. 2.

DGGE patterns of stool samples obtained from mice on day 4 of treatment with subcutaneous normal saline (lanes 1 to 3), subcutaneous daptomycin (lanes 4 to 6), oral linezolid (lanes 7 to 9), and subcutaneous vancomycin (lanes 10 to 12). Lane 13 shows a DGGE reference pattern.

The effect of the lower antibiotic doses on establishment of colonization with the pathogens is shown in Fig. 3. Linezolid (and clindamycin controls) promoted overgrowth of VRE in comparison to saline controls (P = 0.01 and P < 0.001, respectively), whereas daptomycin and vancomycin did not (P > 0.77). Linezolid (and clindamycin controls) promoted overgrowth of K. pneumoniae (P < 0.001), whereas daptomycin and vancomycin did not (P = 1.00). The mean concentration of linezolid in stool on day 4 of treatment was 100 μg/g; daptomycin and vancomycin were not detectable on day 4 (limit of detection, 4 μg/g). When the duration of antibiotic treatment prior to orogastric inoculation of pathogens was increased from 2 to 6 days, vancomycin treatment promoted overgrowth of K. pneumoniae (P < 0.001) and resulted in a trend toward increased density of VRE (P = 0.063), whereas daptomycin treatment did not (Fig. 3C and D). Vancomycin was not detectable in stool samples on days 3 or 5 of treatment, but was detectable in all samples tested on day 8, with a mean concentration of 102 μg/g (range, 54 to 200 μg/g).

FIG. 3.

Effect of lower doses of subcutaneous vancomycin, linezolid, and daptomycin on the establishment of colonization with VRE (A and C) or ESBL-producing Klebsiella pneumoniae (B and D) in mice. Subcutaneous clindamycin was included as a positive control. Mice received antibiotic treatment once daily from day 2 (A and B) or 6 (C and D) before infection to day 4 after infection, and 10⁴ CFU of the pathogens were administered by orogastric gavage on day 0. If the pathogens were not detected in the stool, the lower limit of detection (∼2 log₁₀ CFU/g) was assigned. Filled square, vancomycin; filled circle, linezolid; open square, normal saline; open circle, daptomycin; open triangle, clindamycin.

Our findings suggest the possibility that daptomycin treatment may not promote stool colonization with VRE or K. pneumoniae among patients. Because daptomycin was not detected in stool and did not cause significant changes in stool DGGE patterns, it is likely that it did not promote overgrowth of the pathogens because it did not disrupt competing anaerobic microflora (3-4, 9) rather than because of inhibitory activity. The absence of detectable daptomycin in the stool of mice could be due to low levels of biliary excretion or to inactivation of the drug in the gastrointestinal tract (14). Further research is needed to determine the concentration of daptomycin in the stool of humans and to evaluate the effect of this agent on the indigenous microflora. The administration of radioactively labeled daptomycin resulted in recovery of 5.7% of the radioactivity in the stool of human volunteers; however, this assay did not differentiate between active and inactive drug (data on file, Cubist Pharmaceuticals, Lexington, Mass.).

Linezolid and vancomycin caused significant disruption of stool DGGE patterns at higher doses (DGGE analysis was not performed at the lower doses). Both doses of these agents resulted in promotion of colonization by K. pneumoniae, although for the lower vancomycin dose, promotion occurred only when the duration of treatment prior to inoculation of K. pneumoniae was extended to 6 days. The higher dose of vancomycin promoted overgrowth of VRE; the higher dose of linezolid did not, unless the inoculum of VRE was increased to 10⁸ CFU. These data suggest that the inhibitory activity of linezolid against VRE strains may be sufficient to limit overgrowth of this pathogen under some circumstances. The concentrations of linezolid present in the stool of mice were much higher than those measured in the stool of healthy humans receiving 600 mg orally twice daily (i.e., a mean of 7.1 μg/g on day 4 of treatment) (10), suggesting that our findings could overestimate the impact of linezolid. However, the impact of linezolid on the indigenous microflora of mice was very similar to that observed in the study of healthy humans. Lode et al. (10) found that oral linezolid treatment resulted in suppression of anaerobes and enterococci, whereas the density of Klebsiella species increased significantly.

The concentrations of vancomycin in the stool of the mice receiving the lower dose were similar to those reported to be present in the stool of patients receiving intravenous vancomycin (i.e., 0 to 110 μg/ml) (2, 7). Currie and Lemos-Filho (2) found that vancomycin concentrations were low or undetectable in the stool of patients during the first 5 days of intravenous vancomycin treatment, whereas vancomycin was usually detectable in stool after 5 days of therapy (range, 3.3 to 94.8 μg/ml). These data in combination with our findings suggest that longer durations of intravenous vancomycin treatment may pose a significantly greater risk for promotion of pathogen overgrowth.