Previous Article | Next Article ![]()
Antimicrobial Agents and Chemotherapy, February 2005, p. 565-570, Vol. 49, No. 2
0066-4804/05/$08.00+0 doi:10.1128/AAC.49.2.565-570.2005
Copyright © 2005, American Society for Microbiology. All Rights Reserved.
School of Pharmacy University of Southern California,1 Department of Biological Sciences, California State University, Los Angeles,2 Mpex Pharmaceuticals, San Diego,2 Department of Pharmacy, Huntington Hospital, Pasadena, California2
Received 30 July 2004/ Returned for modification 3 October 2004/ Accepted 7 October 2004
|
|
|---|
8-fold MIC decrease when tested with EPI) was compared among subgroups with different resistance profiles. The EPO phenotype was more prevalent among levofloxacin-resistant than levofloxacin-sensitive strains (61%, 48/79 versus 9%, 6/64). EPO was present in 60% of fluoroquinolone-resistant strains without cross-resistance, while it was present at variable frequencies among strains with cross-resistance to other agents: piperacillin-tazobactam (86%), ceftazidime (76%), cefepime (65%), imipenem (56%), gentamicin (55%), tobramycin (48%), and amikacin (27%). The magnitude of MIC decrease with an EPI paralleled the frequency of which the EPO phenotype was observed in different subgroups. EPI reduced the levofloxacin MIC by as much as 16-fold in eight strains for which MICs were 128 µg/ml. Efflux-mediated resistance appears to contribute significantly to fluoroquinolone resistance and MDR in P. aeruginosa. Our data support the fact that increased fluoroquinolone usage can negatively impact susceptibility of P. aeruginosa to multiple classes of antipseudomonal agents. |
|
|---|
Efflux pumps present in P. aeruginosa serve physiologic functions such as the removal of intracellular toxic substances or metabolites as well as excretion of signaling molecules into the environment to facilitate cell-to-cell communication (9, 17, 28, 30, 31). Six EPs have been identified thus far: MexAB-OprM, MexCD-OprJ, MexEF-OprN, MexXY-OprM, MexJK-OprM, and MexVW-OprM (5, 20, 31). The first three have been well characterized. Each pump consists of a tripartite system: cytoplasmic membrane exporter protein in the resistance-nodulation-division family, an outer membrane protein, and a membrane fusion protein linking the exporter protein and the outer membrane protein. All of the pumps can expel a variety of compounds ranging from detergents to structurally unrelated antimicrobial agents from the cytoplasm or periplasmic space (9, 31). While each pump has a preferential set of antimicrobial agent substrates, the fluoroquinolones are universal substrates for all known Mex pumps (24, 31). Thus, FQ exposure has the potential to select for mutants with the multidrug-resistant (MDR) phenotype via efflux pump overexpression.
Large-scale surveillance studies have reported an increasing rate of FQ resistance among clinical isolates of P. aeruginosa; however, the role of efflux pump overexpression contributing to resistance is difficult to infer based on susceptibility pattern and has not been systematically examined (27). Several reports have documented the selection of the MDR phenotype by quinolones in vitro and in vivo. An early report described ciprofloxacin (CIP)-selected mutants with resistance to ß-lactams, aminoglycosides, and other quinolones with several alterations in outer membrane proteins including an increase in a 54-kDa protein (18). Another report described an isolate showing the nfxC phenotype with high-level quinolone and imipenem (IPM) resistance, which was obtained from the urine of a patient treated with norfloxacin. Attributable mechanisms of resistance were alterations in DNA gyrase, loss of D2 porin, and an increase in a 50-kDa protein (33). Several others have also reported the MDR phenotype with FQ resistance and an increase of an approximately 50-kDa outer membrane protein (7, 15, 23, 32).
A series of dipeptide amide compounds have been identified as efflux pump inhibitors (EPIs) with the ability to broadly inhibit several known multidrug efflux pumps in P. aeruginosa, MexAB-OprM, MexCD-Opr, and MexEF-OprN, and possibly other efflux pumps (10, 19, 21, 22, 34-36), by competing with antibiotic substrates for binding to the pumps. Devoid of intrinsic antibacterial activity, these inhibitors have been demonstrated to (i) potentiate the activity of FQs and other antibiotics, (ii) reverse acquired FQ resistance attributable to not only efflux mutations but also to target site mutations, and (iii) markedly decrease the frequency with which highly FQ-resistant strains could be selected in vitro (22, 34). These compounds have been used to screen for efflux pump overexpression in clinical isolates of P. aeruginosa (4, 22). For this study, we employed the EPI compound, MC-04,124, to characterize the efflux pump-overexpressed phenotype in clinical isolates of P. aeruginosa (34, 40). MC-04,124 has been shown to be as potent but less toxic than earlier derivatives (MC-207,110 and MC-02,595) in this series of compounds (22, 34, 40).
We hypothesized that the widespread use of FQ agents could have a deleterious collateral effect on the susceptibility of P. aeruginosa to existing antipseudomonal agents through FQ-selected overexpression of multidrug efflux pumps. The specific objectives of our study were to use an EPI as a screening agent to evaluate (i) the prevalence of efflux pump-mediated resistance among clinical isolates of P. aeruginosa and (ii) the contribution of EP overexpression as the putative mechanism for the MDR phenotype in P. aeruginosa.
(This study was presented in part at the 43rd Interscience Conference on Antimicrobial Agents and Chemotherapy, Chicago, Ill., 14 to 17 September 2003.)
|
|
|---|
|
View this table: [in a new window] |
TABLE 1. Synergistic in vitro activity of MC-04,124 combined with LVX against laboratory strains of P. aeruginosa
|
Broth microdilution tests were performed according to the National Committee for Clinical Laboratory Standards (NCCLS) recommendations to determine strain susceptibility (26). Each strain was freshly thawed and subcultured prior to testing. The following agents were tested in twofold dilutions in Mueller-Hinton broth (Becton Dickinson, Franklin Lakes, N.J.): CIP, LVX, cefepime, ceftazidime, piperacillin-tazobactam, imipenem (IPM), gentamicin, tobramycin, and amikacin. Additionally, the isolates were tested against the FQsLVX, CIP, moxifloxacin (MOX), and gatifloxacin (GAT)alone at concentrations ranging from 0.125 to 32 µg/ml and in the presence of 20 µg of EPI/ml. For strains with a LVX MIC >32 µg/ml and a decrease in MIC when tested in combination with EPI, additional testing with LVX concentrations up to 256 µg/ml was performed. Isolates were tested in duplicate and results were read visually and spectrophotometrically after overnight incubation at 35°C.
Definitions. The MIC was the lowest concentration of the antimicrobial agent at which no growth was detected. Resistance was defined by MIC breakpoints according to NCCLS interpretive criteria for the respective agent. No MIC breakpoints for P. aeruginosa were provided by NCCLS for GAT and MOX, thus LVX breakpoints were used arbitrarily. An efflux pump-overexpressed phenotype was arbitrarily defined as any strain exhibiting at least an eightfold decrease in MIC to FQ when tested in the presence of an EPI (MC-04,124). This measure was based on the observation of a four- to eightfold potentiation of levofloxacin activity with this and other diamine EPI derivatives against the wild-type strain with constitutive expression of MexAB-OprM (4, 22).
Data analysis. LVX was used as the FQ marker for comparison between MICs of FQ alone and in the presence of an EPI. The prevalence of the efflux pump-overexpressed phenotype was determined for strains overall and also for subgroups based on FQ susceptibility and pattern of cross-resistance to other antipseudomonal agents.
Subgroupings. Isolates were subgrouped into the following categories. The levofloxacin-sensitive (LvxS) group included isolates obtained from adults and children, where isolates from children served as a FQ-naive control group. Levofloxacin-resistant (LvxR) group isolates were further divided into four groups arbitrarily based on the varied Mex pump substrate profile (1, 24, 29, 39): 1, LvxR-Only; 2, with ß-lactam resistance (Lvx-BL, defined as LvxR plus resistance to all ß-lactamsceftazidime, cefepime, and piperacillin-tazobactam, excluding IPM); 3, with imipenem resistance (Lvx-IPM); and 4, with other resistance such as an aminoglycoside agent or single ß-lactam agent (Lvx-Other).
The frequency of the EP-overexpressed phenotype was compared for FQ-resistant strains with and without cross-resistance to other agents. In addition, the magnitudes of activity potentiation (as measured by the magnitude of change in MICs alone versus in the presence of an EPI) were examined for each FQ agent tested and compared with each other.
Statistical analysis. Prevalence of efflux pump overexpression, cross-resistance, and magnitude of MIC decrease with EPI were compared by chi-square or Fisher's exact test, as deemed appropriate. The significance level was 0.05. Analyses were performed using GraphPad Prism version 4.00 for Windows (GraphPad Software, San Diego, Calif.).
|
|
|---|
|
View this table: [in a new window] |
TABLE 2. Prevalence of the efflux pump-overexpressed phenotypea
|
95%) showed cross-resistance to other FQ agents: ciprofloxacin (75/79, 95%), gatifloxacin (76/79, 96%), and moxifloxacin (74/79, 95%). In addition, a substantial portion of LvxR isolates with cross-resistance to structurally unrelated antipseudomonal agents exhibited the efflux pump-overexpressed phenotype (Fig. 1). LvxR strains showing cross-resistance towards a ß-lactam antipseudomonal agent have the highest prevalence of the EP-overexpressed phenotype compared to other cross-resistant subgroups. Specifically, LvxR strains which are cross-resistant to piperacillin-tazobactam have the highest prevalence of the efflux-overexpressed phenotype (85%), followed by ceftazidime (76%). The EP-overexpressed phenotype was present in about half of the strains with cross-resistance to cefepime (56%), imipenem (55%), gentamicin (54%), and tobramycin (48%).
![]() View larger version (38K): [in a new window] |
FIG. 1. Frequency of the efflux pump-overexpressed phenotype (EP) in LvxR strains showing cross-resistance to non-FQ antipseudomonal agents. n, number of LvxR isolates showing cross-resistance to the respective agent. CAZ, ceftazidime; FEP, cefepime; TZP, piperacillin-tazobactam; GEN, gentamicin; TOB, tobramycin; AMK, amikacin.
|
![]() View larger version (37K): [in a new window] |
FIG. 2. Effect of EPI on fluoroquinolone MICs. All strains were tested over the same range of concentrations (0.5 to 8 µg/ml) for all four FQ agents. n, number of strains showing at least an 8-fold decrease in MIC with an EPI; an additional 28 isolates showed a 8-fold decrease in the LVX MIC when tested at a broader concentration range (0.25 to 256 µg/ml). P was 0.0002 for CIP versus MOX, LVX, or GAT. P was >0.05 for comparison amongst MOX, LVX, and GAT.
|
|
View this table: [in a new window] |
TABLE 3. Levofloxacin MIC alone and in the presence of an EPI
|
|
|
|---|
Large-scale surveillance studies thus far have tracked the change in P. aeruginosa susceptibilities over time but have not evaluated the underlying mechanisms responsible for the increase in P. aeruginosa resistance (2, 8, 14, 27). To date, only one epidemiologic study has investigated the contribution of EP overexpression to FQ resistance among clinical isolates of P. aeruginosa obtained from 12 countries using EP-specific and broad-spectrum inhibitors (4). Pump profiles among studied strains varied widely by country of origin. Results specific to strains from the United States (n = 258) indicate that 50% of the clinical isolates overexpressed the MexAB-OprM pump, with one-third of the isolates (32%) expressing a combination of efflux pumps and/or target mutations. However, this study did not explore the role of EP overexpression in strains with the MDR phenotype.
In our study, we determined the prevalence of the efflux pump-overexpressed phenotype and compared FQ-susceptible and FQ-resistant clinical isolates of P. aeruginosa obtained from non-cystic fibrosis adult and pediatric patients. We included susceptible strains from patients (i.e., pediatrics) who were unlikely to have been preexposed to the FQ agents as controls. We screened for the presence of the EP-overexpressed phenotype using a broad-spectrum EPI (MC-04,124) based on its ability to potentiate levofloxacin activity (10, 22, 34, 40). The efflux pump-overexpressed phenotype was present in the majority (61%) of FQ-resistant strains compared to only 9% of FQ-susceptible strains.
In addition, we observed that strains expressing multidrug resistance were significantly associated with the efflux pump-overexpressed phenotype. The presence of the EP-overexpressed phenotype in all FQ-resistant strains with cross-resistance to non-imipenem ß-lactams is remarkable. By analyzing the individual ß-lactam agent, we found that the EP-overexpressed phenotype is most frequently encountered in the presence of coresistance to piperacillin-tazobactam (85%) followed by ceftazidime (76%).
EPI appears to have the strongest effect on potentiating LVX activity in strains showing the greatest degree of resistance to LVX. Notably, MIC90s for the subgroup of strains resistant to all ß-lactam agents (excluding imipenem) decreased by 16-fold (128 to 8 µg/ml) in the presence of an EPI compared to fourfold (2 to 0.5 µg/ml) for levofloxacin-susceptible clinical strains and twofold (0.5 to 0.25 µg/ml) for the PAO1 wild-type strain. The effect of the EPI on FQ MICs appears to be similar for LVX, MOX, and GAT but lower for CIP when tested over the same MIC range. An accurate assessment and comparison of the potentiating effect of EPI on FQ agents other than LVX may be limited by the substantial number of strains with MICs beyond the upper limit of concentration range tested. However, it is possible that the differential effects observed may be a function of differential binding affinity of the EPI or the Mex systems expressed towards each FQ agent. Kohler et al. demonstrated that the quinolones differed in their ability to select for a particular efflux system depending on its C-7 and C-6 constituents (16). Ciprofloxacin demonstrated preferential selection for the MexCD-OprJ system (16). Levofloxacin, however, had an eightfold decrease in MIC (in the presence of an EPI) for all strains expressing the mexAB-oprM (22).
Interestingly, use of an EPI with broad activity against known Mex pumps (MC-04,124) did not potentiate LVX activity at all in 18% (14/79) of our FQ-resistant strains while the LVX MIC was decreased fourfold in 19% and twofold in 3% of the remaining FQ-resistant strains, suggesting that other mechanisms such as target mutations may contribute significantly to resistance in these strains. It is possible that the prevalence of the EP-overexpressed phenotype may be greater if testing to determine actual MICs was performed at the higher range of concentrations for other FQ agents besides LVX. In addition, our arbitrary measure of the efflux-overexpressed phenotype based on an eightfold MIC decrease may have underestimated the contribution of efflux pumps to FQ resistance.
Our demonstration of the widespread presence of the EP-overexpressed phenotype in FQ-resistant and MDR strains requires genotypic confirmation. Further studies are planned to explore other mechanisms that may have contributed to FQ resistance (i.e., target mutations) and MDR in our strains (i.e., target mutations, ß-lactamase production, and loss of outer membrane porins).
Efflux pump-mediated resistance is an underrecognized mechanism contributing to resistance in clinical isolates of P. aeruginosa. We have demonstrated that the efflux pump-overexpressed phenotype appears to be widespread among clinical strains which have developed resistance to the fluoroquinolone class of agents. Broader testing of P. aeruginosa isolates in large-scale surveillance studies is warranted to establish the true prevalence of the efflux-overexpressed phenotype. More importantly, we have provided support for the link between FQ resistance and multidrug resistance through phenotypic confirmation of EP overexpression. Strains which overexpress efflux pumps are frequently resistant to the fluoroquinolones and the ß-lactams simultaneously, leaving few available treatment options. Considering that the FQs appear to be universal substrates for the Mex pumps and that these pumps have broad substrate specificity for structurally unrelated compounds, increased exposure through widespread prescribing of the fluoroquinolones will undoubtedly promote the selection for strains overexpressing efflux pumps. The continued emergence and increasing prevalence of FQ-resistant and MDR P. aeruginosa is of grave concern and urgently calls for clinicians to spare the use of fluoroquinolones as first-line therapy in order to preserve the utility of the existing antipseudomonal armamentarium.
We are thankful for the assistance of Jasmine Razeghi, Megan Nguyen, Clark Inderlied, and Misa Nakayama and the Microbiology Laboratories at Huntington Hospital and Childrens Hospital Los Angeles in saving bacterial isolates.
We thank Kevin Ward and Janet Hindler at the Clinical Microbiology Laboratory at UCLA Medical Center for preparing the microtiter trays for susceptibility testing and Lyn Cote from Mpex Pharmaceuticals for performing checkerboard titration assays.
|
|
|---|
mexA-mexB-oprM mutants of Pseudomonas aeruginosa. Antimicrob. Agents Chemother. 42:1938-1943.
This article has been cited by other articles:
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Copyright © 2009 by the American Society for Microbiology. For an alternate route to Journals.ASM.org, visit: http://intl-journals.asm.org | More Info»