Of Pseudomonas, porins, pumps and carbapenems

Abstract

The carbapenems, imipenem and meropenem, are anomalous β-lactams in their antipseudomonal behaviour, as well as in their β-lactamase stability. In the case of imipenem, MICs for Pseudomonas aeruginosa isolates are unrelated to those of other β-lactams, whereas a strong relationship exists between the MICs of different penicillins and cephalosporins for most Pseudomonas isolates.1–4 This observation is partly because imipenem MICs are independent of derepression of the chromosomal AmpC β-lactamase, whereas this factor strongly co-determines the MICs of penicillins and cephalosporins.5 More critically, the MICs of imipenem are unaffected by the broad-spectrum intrinsic resistance expressed by many P. aeruginosa isolates, whereas this mechanism strongly co-determines the MICs of penicillins, cephalosporins and several unrelated drug classes, including quinolones, tetracyclines and chloramphenicol.1–4 Intrinsic resistance is expressed to a variable degree by all P. aeruginosa isolates, which are less susceptible than Enterobacteriaceae to most antibiotics. This behaviour was long thought to depend on impermeability,1,2 but is now found to involve an interplay of impermeability with multi-drug efflux, mediated principally by MexA-MexB-OprM.3,4,6,7 The MexB protein is a broad-spectrum pump, located in the cytoplasmic membrane; the OprM protein is a pore-forming protein that provides an exit portal through the outer membrane; and the MexA protein physically links these components. Upregulation of MexA-MexB-OprM arises at high frequency as a result of the nalB mutation at the mexR locus8–10 and raises the MICs of penicillins, cephalosporins, quinolones, tetracycline and chloramphenicol (Table), but not those of imipenem. nalB mutants may be selected in vitro, or during therapy with fluoroquinolones, penicillins and cephalosporins,11,12 and were the predominant phenotype selected when carbenicillin and ticarcillin were the mainstays of antipseudomonal treatment.13 The broad range of resistance conferred by upregulation is intriguing, and it is postulated that MexA-MexB-OprM has a natural role in removing amphipathic substances, which otherwise invade and disorganize the cytoplasmic membrane.14 Such a role would favour the evolution of a very broad substrate specificity. The lack of activity against imipenem may be because this carbapenem lacks any lipophilic phenyl or heterocyclic side chain(s). The advantage of evading intrinsic resistance is tempered by the fact that imipenem readily selects resistant mutants of P. aeruginosa, which are found to lack OprD protein (initially called D2 porin).15 The primary role of this protein is in the passive uptake of basic amino acids across the outer membrane, but it forms pores that are also permeable to carbapenems, though not to other β-lactams.16 Its loss raises the imipenem MICs from 1–2 mg/L, as for typical P. aeruginosa, to 8–32 mg/L, thus conferring clinical resistance. The MICs of non-carbapenems are unaffected. The resistance demands continued expression of the chromosomal AmpC β-lactamase, whether inducible or derepressed,17 but this aspect is peripheral in the present context. Carmeli et al.18 noted that selection of resistant P. aeruginosa during imipenem therapy was substantially more frequent than selection of ceftazidime-, piperacillinor ciprofloxacin-resistant mutants, and Calandra et al.19 found a 17% rate of resistance emerging during imipenem treatment of P. aeruginosa infections. The interactions of meropenem with these resistance mechanisms differ from those of imipenem, with implications that are provoking debate. The MICs of meropenem are 0.12–0.5 mg/L for typical strains of P. aeruginosa, and either upregulation of MexA-MexB-OprM or loss of OprD raises them to 2–4 mg/L,5,20 thereby reducing susceptibility without conferring resistance relative to either BSAC or NCCLS breakpoints. It is inferred that meropenem, like imipenem, can use the OprD pathway to enter the pseudomonal cell but that, unlike imipenem, it is recognized and ejected by MexB-mediated efflux, presumably because of its 2′ heterocyclic side chain. These differences between imipenem and meropenem allow arguments for either compound as ‘the less likely to cause resistance in P. aeruginosa’. The case for imipenem21 argues that, although it selects OprD– mutants, these have a narrow-spectrum insensitivity and remain fully susceptible to other drugs, whereas meropenem resistance co-depends on upregulation of MexA-MexB-OprM, a mechanism that compromises fluoroquinolones as well as other β-lactams. The case for meropenem17,22 is that substantive resistance is much harder to achieve than to imipenem, since two mutations (loss of OprD and upregulation of MexA-MexB-OprM) are needed rather than one. The likelihood that a cell will simultaneously undergo both mutations is c. <10–14, whereas imipenem-resistant mutants, lacking OprD, emerge at c. 10–7. Although formal analyses of the risk of meropenem