Fluoroquinolone Resistance Is a Poor Surrogate Marker for Type II Topoisomerase Mutations in Clinical Isolates ofStreptococcus pneumoniae

Fluoroquinolone antimicrobial agents have been used increasingly since the late 1980s. Recently, new quinolones were developed with enhanced activity against gram-positive species (15, 17), including Streptococcus pneumoniae. However, emerging resistance in S. pneumoniae is considered an increasing problem. S. pneumoniae resistance to older fluoroquinolones arises from active efflux (1, 2, 4, 14) and as a result of amino acid substitutions in the quinolone resistance-determining region (QRDR) domains of type II topoisomerase, DNA gyrase, and topoisomerase IV. Several specific substitutions have been associated with resistance to fluoroquinolone antimicrobial agents (5, 11, 12-14). Thus, we hypothesized that there is a direct correlation between the presence of these substitutions and reduced susceptibility that can be detected by routine testing in the clinical microbiology laboratory. If this is the case, then antimicrobial agent susceptibility testing of clinical isolates could serve as an effective surrogate marker for the surveillance of first-step mutations, allowing for early recognition of when emerging resistance may become a significant clinical problem.

The amino acid sequences of QRDR domains and the MICs of ciprofloxacin, ofloxacin, levofloxacin, moxifloxacin, sparfloxacin, norfloxacin, and trovafloxacin were determined for 44 defined clinical isolates ofS. pneumoniae from the collection of strains at Northwestern Memorial Hospital in Chicago, Ill., and for one archived laboratory strain (CP1000) (16). All agents except moxifloxacin and sparfloxacin were selected because they had been in clinical use. The latter two compounds were chosen because of their ability to avoid the development of resistance mutations (1, 14). The strains were chosen to represent a distribution of isolates; for approximately 40%, the ofloxacin MIC was <2 μg/ml, and for the remaining isolates, the ofloxacin MIC was ≥2 μg/ml. Ofloxacin was chosen as the reference agent since it seemed the least active compound and had been observed in our laboratory to be especially potent for selecting strains with resistance mutations (unpublished observation). Susceptibility testing of the isolates was done by the microdilution method using Mueller-Hinton broth supplemented with 3 to 5% lysed (laked) horse blood prepared in our laboratory (9). Susceptibility panels were inoculated at a density of 1 × 10⁵ to 5 × 10⁵ CFU/ml and incubated at 35 to 37°C for 24 h.

To analyze amino acid substitutions in ParC, ParE, GyrA, and GyrB of the selected strains, nucleotide sequences including QRDR domains of the respective proteins (amino acids 115 to 198 in GyrA, 361 to 511 in GyrB, 55 to 167 in ParC, and 392 to 529 in ParE) were determined and compared to the corresponding sequences from the reference strain, CP1000. A 253-bp fragment of gyrA (bp 342 to 595), a 453-bp fragment of gyrB (bp 1080 to 1533), a 337-bp fragment ofparC (bp 164 to 501), and a 413-bp fragment ofparE (bp 1175 to 1587) were amplified and then sequenced using an ABI PRISM dye terminator cycle sequence ready reaction kit (PE Biosystems, Foster City, Calif.) and an ABI PRISM 310 genetic analyzer according to the protocol of the manufacturer. Amino acid sequence alignment was done using MegAlign (DNASTAR, Inc., Madison, Wis.). All sequences and MICs were determined in duplicate.

The MIC results for the tested fluoroquinolones are shown in Table1. Also shown in Table 1 are amino acid substitutions in the QRDR domains relative to the sequence of the reference strain, CP1000. Of the 44 isolates, 23 bore at least one amino acid substitution; 17 of these 23 isolates had a single mutation (7 in ParC only, 9 in ParE only, and 1 in GyrB only) and 6 isolates had two or more substitutions. As is evident from Table 1, there was no correlation between resistance to the fluoroquinolone agents tested and the presence or type of amino acid substitutions, despite the fact that many changes were previously reported (8, 10-13). Interestingly, a prevalent amino acid substitution in ParE (Ile460 → Val) demonstrated the problem of using phenotypic resistance as a marker for genetic mutations. Although it was found in 14 strains for which the MIC of ofloxacin was ≥2 μg/ml, it did not seem to affect the levels of ciprofloxacin resistance, nor did it show a correlation with resistance levels for the other tested quinolones. In addition, in 10 strains for which the MIC of ofloxacin was 2 to 4 μg/ml, no Ile460 → Val substitution was detected. These results demonstrate the difficulty of relating phenotypic resistance to a specific amino acid change and how the selection of a screening antimicrobial agent can affect the potential detection of any mutation(s).

Only 7 of the 23 strains bearing amino acid substitutions had decreased susceptibility to ciprofloxacin (MICs of ≥2 μg/ml), and the MICs of ciprofloxacin for all the other strains were ≤1 μg/ml. As is evident from the data presented, even in the seven strains with reduced susceptibility to ciprofloxacin, which appeared to be the most affected agent (likely due to the fact that it is the most widely used fluoroquinolone), no correlation could be determined between the number, nature, and phenotypic significance of the substitutions and the specific level of in vitro susceptibility.

In summary, our data demonstrate little correlation between specific mutations in DNA gyrase or topoisomerase IV and phenotypic susceptibility to frequently used fluoroquinolones. This finding is disconcerting because there is no evidence that any QRDR mutations occurred naturally in S. pneumoniae strains that were archived before the advent of quinolone chemotherapy, such as strain CP1000. Furthermore, the presence of first-step mutations has been shown to facilitate the development of resistance to newer fluoroquinolones that otherwise resist the emergence of resistance in wild-type strains (14). Since different mutations likely arise from exposure to and determine resistance to different quinolones (3, 6, 7, 14, 17), these results indicate that routine testing cannot sufficiently detect emerging quinolone resistance in S. pneumoniae clinical isolates because some mutations do not exhibit phenotypic changes with many drugs. However, the accumulation of secondary mutations often produces pronounced phenotypic effects appearing only after the initial, first-step changes. Thus, in screening for resistance, it is not possible to monitor susceptibility data for one fluoroquinolone in order to predict emerging resistance to the entire class. In combination, these observations indicate that there is no simple screen for first-step mutations using routine clinical laboratory susceptibility patterns as an early warning system for emerging resistance to fluoroquinolones. Since surveillance is a key element for detecting and managing emerging antimicrobial agent resistance (18), caution must be exercised in the interpretation of phenotypic data suggesting no evidence for developing resistance to fluoroquinolones. Only appropriate use of this class will likely maintain its useful activity for a necessary prolonged period of time.

• Copyright © 2001 American Society for Microbiology