Mechanisms of Bacterial Resistance to Antimicrobial Agents
Usha Stiefel
Louis B. Rice
The vast majority of antimicrobial agents employed in clinical settings are either natural products or chemical derivatives of natural products. The producers of these agents in nature are generally the microbes themselves. The purpose of this antibiotic production by microbes has traditionally been attributed to gaining a selective advantage in a mixed microbial environment. More recently, recognition that the natural production of antibiotics occurs at generally low levels and that exposure to subtherapeutic concentrations of some antibiotics alters bacterial transcription profiles has led to the concept of antibiotics as “signaling molecules” in nature. Since the production of antibiotics has been occurring in the microbial environment for (presumably) eons, it stands to reason that mechanisms to avoid their lethal action have been developed as well, either by species that produce the antibiotics or those that must share limited space and resources with those that do. In many instances, therefore, our discovery and growing use of antibiotics has led not to the development of resistance genes in bacteria but merely to the natural selection of intrinsically resistant species or the efficient scavenging of preexisting resistance genes by normally susceptible human pathogens. The emergence of Lactobacillus species during therapy with vancomycin and of Stenotrophomonas maltophilia during therapy with imipenem are examples of selection of intrinsically resistant species. Other phenotypes of resistance reflect more of the ease with which susceptible bacteria can mutate either structural or regulatory genes intrinsic to their species in a manner that results in decreased antibiotic susceptibility. Examples of this type of resistance include extended-spectrum cephalosporin resistance in Enterobacter species, fluoroquinolone resistance in many different species of bacteria, or the emerging resistance to linezolid in enterococci and staphylococci. Resistance to some antibiotics, in some species, is not readily achievable by mutation, and thus, must be acquired from other sources. This so-called acquired resistance accurately characterizes many different resistance phenotypes, including ampicillin resistance in Escherichia coli, penicillin resistance in staphylococci, and vancomycin resistance in enterococci. Finally, when antimicrobial agents are developed specifically to avoid the lethal action of acquired resistance genes, mutations within the acquired genes can lead to resistance to the newer agents. The emergence of resistance to extendedspectrum cephalosporins in Klebsiella pneumoniae and E. coli can represent this sort of amplified resistance.
Antimicrobial agents are effective because they target metabolic pathways or enzymes that are specific to bacteria and not to the host. A variety of mechanisms have been shown to result in bacterial resistance. Among these mechanisms are alterations in the antibiotic target such that binding or inhibition of function is decreased to the point of clinical irrelevance, decreased permeability that results in the inability of the agent to reach its target at a critical concentration, efflux of the agent from the cell, and destruction or modification of the antibiotic.
The expression of resistance and virulence by bacteria is often linked but sometimes in unpredictable ways. Selection of rifampin or streptomycin-resistant mutants in the laboratory is often associated with a decrease in the virulence of the strains when tested in animal models (1).
It is presumed that the point mutations in the targets (RNA polymerase in the case of rifampin, the ribosome in the case of streptomycin) lead to subtle but not fatal decreases in function in these resistant strains, conferring a competitive survival disadvantage relative to wild-type strains. Interestingly, continued passage in animals in the absence of antibiotic selective pressure does not always result in reversion to the susceptible genotype. Instead, compensatory mutations frequently occur that mitigate the deleterious effects of the primary mutation, restoring virulence while maintaining resistance (1). Acquired resistance and virulence determinants may also coalesce in environments that favor them, such as the modern hospital. Reports suggest that the worldwide rise of ampicillin- and vancomycinresistant Enterococcus faecium is because of the emergence and spread of genetically related strains enriched for high levels of ampicillin resistance as well as a variety of putative virulence determinants (2).
It is presumed that the point mutations in the targets (RNA polymerase in the case of rifampin, the ribosome in the case of streptomycin) lead to subtle but not fatal decreases in function in these resistant strains, conferring a competitive survival disadvantage relative to wild-type strains. Interestingly, continued passage in animals in the absence of antibiotic selective pressure does not always result in reversion to the susceptible genotype. Instead, compensatory mutations frequently occur that mitigate the deleterious effects of the primary mutation, restoring virulence while maintaining resistance (1). Acquired resistance and virulence determinants may also coalesce in environments that favor them, such as the modern hospital. Reports suggest that the worldwide rise of ampicillin- and vancomycinresistant Enterococcus faecium is because of the emergence and spread of genetically related strains enriched for high levels of ampicillin resistance as well as a variety of putative virulence determinants (2).
ANTIMICROBIAL RESISTANCE TRANSFER
Although the primary concern of the healthcare epidemiologist is the prevention of the spread of bacterial strains among hospitalized patients, it is worthwhile to consider mechanisms by which resistance genes can spread among bacterial strains. A full discussion of the mechanisms of resistance transfer is beyond the scope of this chapter. Nevertheless, a few basic concepts should be understood.
Antimicrobial resistance determinants are commonly incorporated into extrachromosomal, independently replicating elements known as plasmids. Plasmids vary greatly in size (3 to >200 kb) and in the number of incorporated resistance determinants. In addition to genes responsible for replication and for antibiotic resistance, many plasmids also possess genes that stimulate their transfer between strains within a given genus, and occasionally, between strains of different (although usually closely related) genera. Large, transferable plasmids have been implicated in the spread of ceftazidime resistance among strains of Enterobacteriaceae, particularly in intensive and chronic care settings (3,4). Many of these plasmids also possess genes encoding resistance to a range of non-β-lactam antimicrobial agents, resulting in the elimination of several antibacterial options with a single transfer event (3,4). Transferable plasmids have also been identified in gram-positive genera, perhaps best characterized by the pheromone-responsive plasmids found in strains of Enterococcus faecalis (5). The widespread emergence of high-level gentamicin resistance in enterococci (see below), resulting from the production of a modifying enzyme most commonly encoded on plasmids, is a testament to the efficiency of plasmids in disseminating resistance determinants in this genus (6,7). Enterococci are also known to possess “broad host-range” plasmids. These plasmids transfer at a lower efficiency than do the pheromoneresponsive plasmids but have the advantage of being able to transfer to a wide variety of species. Evidence implicates broad host-range plasmids in the exchange of important resistance genes between enterococci and staphylococci, including β-lactamase production and high-level vancomycin resistance (8,9).
Plasmids need not encode their own transfer genes in order to spread between strains. Nonconjugative plasmids may be mobilized for transfer by conjugative plasmids. In addition, the presence of insertion sequences (ISs) (small regions of DNA capable of independent movement between replicons) has been shown to facilitate the co-integration of conjugative and nonconjugative plasmids, resulting in a larger, conjugative element (10). Appropriately sized plasmids may also be spread by transduction, resulting from the aberrant incorporation of plasmid rather than bacteriophage DNA into the phage head.
In addition to plasmids, antimicrobial resistance determinants are frequently incorporated into mobile elements known as transposons. Transposons may be rather simple elements whose mobility results from the presence of ISs flanking an antimicrobial resistance determinant (composite transposons), an arrangement in which mobility is due entirely to functions encoded by the ISs (11). Alternatively, transposons may be complex structures incorporating several genes. Tn21 is a Tn3-family transposon that has been found to contain a genetic locus (tnpI) that serves as a “hot spot” for the integration of a variety of antimicrobial resistance genes (12). Consequently, several Tn21-like transposons conferring resistance to a number of different antimicrobial agents, in varying combinations, have been described (13). These loci, referred to as integrons, appear to be important mechanisms for the dissemination of antimicrobial resistance genes in many gram-negative bacilli (14,15). Integrons may be critical vehicles of microbial genetic evolution and have only recently been employed by bacteria for purposes of stockpiling resistance determinants (16). Another Tn3-family transposon, Tn1546 (17), confers resistance to vancomycin and teicoplanin in enterococci and, more recently, in Staphylococcus aureus. It encodes nine genes involved in the regulation of transposition and the expression of glycopeptide resistance. More recently, a Tn21-based complex transposon carrying β-lactamase-mediated carbapenem resistance has been described in K. pneumoniae (18).
In general, transposons participate in the transfer of antimicrobial resistance determinants by virtue of their ability to move between bacterial chromosome and transferable plasmid. Exceptions to this rule are the conjugative transposons of gram-positive bacteria, which can transfer between strains without the necessity of a plasmid intermediate (19). These transposons possess their own genes responsible for transfer between microorganisms. In general, conjugative transposons encode resistance to tetracycline via the tetM gene, although some have been found to encode resistance to multiple antimicrobial agents (19). In addition to the transfer of the elements themselves, some investigators have found that the presence of conjugative transposons stimulates the transfer or deletion of unrelated chromosomal genes, raising the possibility that these elements could be involved in the transfer or deletion of a range of unrelated resistance determinants (8,20,21). A transposon in the Tn916 family has been described that encodes VanB-type vancomycin resistance in E. faecium (22). Conjugative transposons may also transfer determinants for antibacterial activity as well as antibiotic resistance. Several lactococcal and one enterococcal Tn916-like elements encoding determinants for production
of the antibacterial peptide nisin have been described (20). Many larger conjugative elements, especially those from gram-negative bacteria, have been generally categorized as integrating conjugative elements (23).
of the antibacterial peptide nisin have been described (20). Many larger conjugative elements, especially those from gram-negative bacteria, have been generally categorized as integrating conjugative elements (23).
Other mobile elements involved in the spread of antimicrobial resistance are the IS elements. These elements do not encode antimicrobial resistance themselves but may aid in the spread of resistance determinants via the formation of composite transposons or by serving as areas of homologous recombination between plasmid and chromosome. Insertion of IS elements may also result in the activation of poorly expressed genes via the presence of promoter sequences within the end of the mobile element (11). Evidence indicates that the expression of imipenem resistance in some strains of Bacteroides fragilis is due to the insertion of IS elements upstream of an unexpressed chromosomal gene encoding a carbapenemase (24). IS elements have also been implicated in plasmid:chromosome integration with subsequent transfer of chromosomal segments using the plasmid origin of transfer in E. faecalis (25).
Our ability to thwart the spread of resistance determinants between bacterial strains in the natural environment is poor. Factors affecting transfer between strains are poorly understood, but, in some cases, may involve exposure to antimicrobial agents. Transfer of conjugative transposons, for example, has been shown to be increased in vitro and in vivo after exposure of the donor strain to tetracycline (26,27). It is, therefore, reasonable to presume that environmental pressure from the overuse of antimicrobial agents plays some role in the spread of these determinants. In addition, the comingling of resistant strains of bacteria in the human gastrointestinal tract resulting from hospital and antibiotic exposure as well as from inattention to appropriate infection control techniques probably plays a role in the spread of resistant strains. In some cases, institution of infection control measures (such as barrier precautions for infected and colonized patients) has been shown to abort serious outbreaks of resistant microorganisms (28,29). In others, decreasing use of an antibiotic has been associated with a reduction in the prevalence of resistant strains in an institution (30). As such, judicious use of antimicrobial agents and proper attention to infection control recommendations are likely to be our best weapons to combat the spread of resistant bacteria for the foreseeable future.
 FIGURE 85-1 The two major peptide cross-links found in bacterial peptidoglycan. A: Cross-link between diamino pimelic acid and D-alanine, commonly found in peptidoglycan of gram-negative bacteria. B: Crosslink between Lysine and D-alanine, more commonly found in grampositive species. The pentaglycine linkage in this figure represents the cross-links found in S. aureus. (Adapted from Royet J, Dziarski R. Peptidoglycan recognition proteins: pleiotropic sensors and effectors of antimicrobial defences. Nat Rev Microbiol 2007;5(4):264-277). |
β-LACTAMS
Mechanism of Action
Targets of β-lactam antibiotics are a series of enzymes involved in the last step of peptidoglycan (cell wall) synthesis. This step involves a cross-linking reaction carried out by transpeptidases in which the terminal D-alanine of the pentapeptide stem of the peptidoglycan is cleaved. The energy resulting from this cleavage is used to form a peptide bond between the fourth residue of the pentapeptide (also D-alanine) and the cross-bridge, which is itself linked to the e-amino of diaminopimelic acid (in gram-negative microorganisms) or lysine (in gram-positive microorganisms) (Fig. 85-1). This cross-link is absolutely required for structural integrity of the bacterial cell wall. β-Lactam antibiotics, such as penicillin, are structural analogs of the pentapeptide terminal D-alanyl:D-alanine target covalently bound by the transpeptidases. The fact that these transpeptidases also bind penicillin (and other β-lactams) covalently has resulted in referral to them as penicillin-binding proteins (PBPs).
Mechanisms of β-Lactam Resistance
Target Resistance The binding affinity of β-lactams for their targets, the PBPs, varies with the β-lactam and the PBP. Enterococci, for example, are intrinsically resistant to the cephalosporins because these β-lactams do not bind one enterococcal PBP with high affinity (31). Within the genus Enterococcus, E. faecium tend to be more resistant to penicillins because many strains express a low-affinity PBP (PBP5) that carries out cell-wall synthesis at penicillin concentrations that inhibit the other PBPs (32).
Many cases of PBP-mediated β-lactam resistance result from the intrinsic characteristics of the PBPs of a given strain. PBP-mediated resistance may also be acquired. Resistance resulting from mutation can be readily demonstrated in the laboratory (33). Resistance to oxacillin in clinical S. aureus strains has been attributed to point mutations in PBP genes (33). In species that are naturally transformable (that can absorb naked DNA from the environment), the formation of mosaic PBP genes is common. Cloning and sequencing of Streptococcus pneumoniae or Neisseria gonorrhoeae genes encoding abnormal, low-affinity PBPs responsible for penicillin resistance has shown significant sections of these genes to be of foreign origin. In S. pneumoniae, the origin appears to have been from oral streptococci (34); in N. gonorrhoeae, from oral commensal neisserial species (35,36). The evolution of mosaic genes most likely occurred via DNA transformation followed by homologous recombination across areas of PBP sequence homology between the native and foreign DNA. Entire low-affinity PBPs can also be acquired by normally susceptible bacteria. Methicillin-resistant S. aureus (MRSA) has most commonly acquired low-affinity PBP2a, encoded by the mecA gene. The mec region is located within a larger mobile element (designated SCCmec) that varies in size depending on how much extra DNA it contains (37). Healthcare-associated strains (which are resistant to several unrelated classes of antimicrobial agents) contain a larger SCCmec, reflecting the insertion of additional DNA, some of which encodes additional antimicrobial resistance. In contrast, the recently described MRSA arising in the community (which is generally susceptible to a range of other antimicrobial agents) contains a relatively small SCCmec that encodes only resistance to methicillin (38). The mec region may have been acquired from coagulasenegative staphylococcal species (39).
The expression of resistance encoded by mosaic or acquired PBPs is often dependent on very specific conditions. Several staphylococcal genes, called fem (factors essential for methicillin resistance) or aux (auxiliary) factors, have been identified—the inactivation of which results in reversion to susceptible phenotype despite the expression of PBP2a (40). In most cases, these fem genes encode enzymes responsible for the synthesis of peptidoglycan precursors. Similarly, the expression PBP-mediated resistance in S. aureus, E. faecalis, and E. faecium is dependent upon the presence of specific glycosyltransferases with which the low-affinity transpeptidases can work (41, 42 and 43).
Enterococci are intrinsically resistant to some β-lactams, especially the cephalosporins, at high levels. Resistance is related to the low affinity of these compounds for the enterococcal PBP5 (32,44). Strains resistant to even higher levels of the penicillins, in the absence of production of β-lactamase, have been described with increasing frequency (45,46). These strains include several species, but E. faecium is most commonly reported from clinical laboratories. Most of these high-level resistant strains have one or more point mutations in pbp5 that are thought to lower the affinity for penicillin and other β-lactams (47). Accumulation of point mutations has been associated with lowered affinity and elevated minimum inhibitory concentrations (MICs) in the laboratory, supporting the notion that these point mutations contribute to higher levels of resistance (48). Enterococcal strains expressing high levels of resistance to β-lactams through low-affinity PBPs are also more resistant to β-lactam-aminoglycoside synergism, even in the absence of high levels of aminoglycoside resistance (49). Single-agent β-lactam therapy is precluded for such strains, leaving the glycopeptides as the antibiotic class of choice. The continued spread of glycopeptide resistance in penicillin-resistant enterococci (see below) is a problem at many large centers (46).
β-Lactamase-Mediated Resistance A more important (than target resistance) and frequent mechanism of bacterial resistance to β-lactam antibiotics, especially in gram-negative bacteria, is the production of β-lactamases— enzymes that hydrolyze the β-lactam ring (Fig. 85-2). The reactive β-lactam ring is required for the formation of a covalent bond between the antibiotic and its PBP target. Destruction of this ring results in the loss of antimicrobial activity. The β-lactamases form a broad family of enzymes and, along with the PBPs, are classified as serine D, D-peptidases (50). The homologies between many β-lactamases and PBP have led to the suggestion that β-lactamases have evolved from PBPs.
Two classification schemes for the β-lactamases are widely used. The first is based on primary structure and has been proposed by Ambler et al. (51,52) (Table 85-1). The other scheme (Bush-Jacoby-Medeiros classification) relies on the substrate specificity of the enzymes (53) (Table 85-2). There are many more classes and subclasses in this scheme, since single-point mutations in the gene encoding an enzyme may result in substantial changes in substrate specificity. The Ambler scheme is more frequently employed, likely because of its comparative simplicity.
Staphylococcal β-lactamase production became widespread within a few years of the clinical introduction of penicillin (54,55). By the mid-1940s, β-lactamase-producing S. aureus strains were prevalent within hospitals, necessitating the introduction of vancomycin and semisynthetic penicillins such as methicillin, nafcillin, and oxacillin. In contrast to observations of class A enzymes in gramnegative bacilli, staphylococcal β-lactamase has not evolved to a broader spectrum over the decades. The importance of β-lactamase production in gram-positive bacteria remains essentially restricted to staphylococci.
The epidemiology of β-lactamase-mediated resistance in gram-negative bacilli is far more complex than in gram-positive bacteria. Hundreds of different β-lactamases have been described in gram-negative bacteria over the past 3 decades. The most problematic and prevalent of these enzymes are those that confer resistance to expanded spectrum cephalosporins. Earlier versions of these extended-spectrum β-lactamases (ESBLs) were progeny of narrower spectrum enzymes that fall, like the staphylococcal β-lactamase, into Ambler class A. The most common enzymes of this class among clinical isolates are related to the widely prevalent TEM-1 and SHV-1 enzymes (53). TEM-1 is widely prevalent as the cause of ampicillin resistance in E. coli, Haemophilus influenzae, and in some cases N. gonorrhoeae, whereas SHV-1 is the chromosomal β-lactamase found in most K. pneumoniae strains. TEM-1 and SHV-1 are broad-spectrum β-lactamases that hydrolyze
the penicillins (ampicillin, mezlocillin, and piperacillin) with greater efficiency than the cephalosporins (56). ESBLs result from the accumulation of point mutations within the TEM-1 or SHV-1 enzymes that serve to “open up” the active site of the enzyme, allowing binding of the bulky extendedspectrum cephalosporins. These point mutations are often found in association with cellular characteristics that serve to enhance the phenotypic expression of resistance, such as the location downstream of strong promoters (leading to increased β-lactamase quantity) and reductions in the expression of outer membrane proteins (OMPs; porins that serve as conduits for the entry of antibiotics into the periplasmic space). Genes encoding TEM- and SHV-related ESBLs are most commonly found on transferable plasmids with resistance determinants to numerous other antimicrobial classes. Strains elaborating ESBLs, most commonly Klebsiella, have been responsible for several outbreaks of infection and colonization in Europe and the United States. Outbreaks have been ascribed to clonal dissemination, plasmid dissemination, or both (30,57,58).
the penicillins (ampicillin, mezlocillin, and piperacillin) with greater efficiency than the cephalosporins (56). ESBLs result from the accumulation of point mutations within the TEM-1 or SHV-1 enzymes that serve to “open up” the active site of the enzyme, allowing binding of the bulky extendedspectrum cephalosporins. These point mutations are often found in association with cellular characteristics that serve to enhance the phenotypic expression of resistance, such as the location downstream of strong promoters (leading to increased β-lactamase quantity) and reductions in the expression of outer membrane proteins (OMPs; porins that serve as conduits for the entry of antibiotics into the periplasmic space). Genes encoding TEM- and SHV-related ESBLs are most commonly found on transferable plasmids with resistance determinants to numerous other antimicrobial classes. Strains elaborating ESBLs, most commonly Klebsiella, have been responsible for several outbreaks of infection and colonization in Europe and the United States. Outbreaks have been ascribed to clonal dissemination, plasmid dissemination, or both (30,57,58).
 FIGURE 85-2 Reactions of serine-type peptidases (which include many β-lactamases) with the fourmembered β-lactam ring of β-lactam antibiotics. Rapid breakdown is characteristic of lactamases, whereas slow breakdown or complete inertness is characteristic of PBPs. (Modified from Ghuysen JM. Serine β-lactamases and PBPs. Annu Rev Microbiol 1991;45:37-67). |
TABLE 85-1 Molecular Classification of β-Lactamases | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| ||||||||||||
Mutations to extend the spectrum of TEM-1 or SHV-1 and allow hydrolysis of extended-spectrum cephalosporins commonly yield increased susceptibility to inhibition by β-lactamase inhibitors. In the clinical setting, however, the production of multiple enzymes and/or overproduction of individual enzymes often confer in vitro resistance to β-lactam/β-lactamase inhibitor combinations in ESBL producers. The relative scarcity of ESBL producers has made controlled studies of the efficacy of different therapies impractical, but carbapenems have been most effective in animal studies of infections with ESBL producers as well as in case reports and small series. Most of the clinical experience has been with imipenem (57,59).
In the past decade, resistance to extended-spectrum cephalosporins in Enterobacteriaceae has been increasingly attributed to the expression of β-lactamases of the CTX-M family. They are naturally resistant to cephalosporins. They fall into Ambler class A, are generally more active against ceftriaxone and cefepime than
ceftazidime, are susceptible to inhibition by β-lactamase inhibitors, are plasmid-mediated, and unlike the ESBL TEM and SHV variants, are very commonly found in E. coli as well as K. pneumoniae. In fact, the growing worldwide problem of increasing cephalosporin resistance in E. coli is almost exclusively attributed to the spread of CTX-M-type enzymes (60). The expression of these enzymes is frequently associated with resistance to fluoroquinolones, creating significant problems for empirical therapeutic regimens for community-acquired E. coli infection (60).
ceftazidime, are susceptible to inhibition by β-lactamase inhibitors, are plasmid-mediated, and unlike the ESBL TEM and SHV variants, are very commonly found in E. coli as well as K. pneumoniae. In fact, the growing worldwide problem of increasing cephalosporin resistance in E. coli is almost exclusively attributed to the spread of CTX-M-type enzymes (60). The expression of these enzymes is frequently associated with resistance to fluoroquinolones, creating significant problems for empirical therapeutic regimens for community-acquired E. coli infection (60).
TABLE 85-2 Bush-Jacoby-Medeiros Functional Classification Scheme for β-Lactamases | ||||||||||||||||||||||||||||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| ||||||||||||||||||||||||||||||||||||||||||||
Resistance to extended-spectrum cephalosporins may also be conferred by the expression of regulatory mutants of Ambler’s class C β-lactamases. These enzymes are broadly active cephalosporinases (which also hydrolyze penicillins) and are resistant to clinically achievable concentrations of β-lactam/β-lactamase inhibitor combinations (53). They are encoded by the ampC gene—a chromosomal gene widely disseminated among Enterobacteriaceae and Pseudomonas aeruginosa. In some species, such as E. coli, ampC is poorly expressed and not under regulatory control due to the absence of the ampR gene. The product of the ampR gene interacts with different cell wall breakdown products in a manner that results in AmpR becoming either a suppressor or an activator of ampC transcription (61, 62 and 63). Under normal circumstances, cells with inducible AmpC β-lactamases employ AmpD (a cellular amidase encoded by ampD) to reduce intracellular quantities of cellular breakdown product anhydro-muramyl-tripeptide, which results in an excess of uridine diphosphate (UDP)-muramyl-pentapeptide. UDP-muramyl-pentapeptide interaction with AmpR maintains AmpR as a repressor of ampC transcription. When exposed to certain antibiotics that favor the production of anhydro-muramyl-tripeptide (such as cefoxitin, clavulanic acid, and imipenem), the ability of AmpD to convert this substrate is overwhelmed, and interaction between anhydro-muramyl-tripeptide and AmpR converts AmpR into an activator of ampC transcription (induction). ampR is present and ampC is under regulatory control in Enterobacter species, Serratia marcescens, Citrobacter freundii, and P. aeruginosa, among others (63, 64 and 65). Imipenem is an efficient inducer of ampC expression, but it is a poor substrate for the ampC β-lactamase. It therefore remains active even in the presence of induced β-lactamase (as long as a concomitant mutation that decreases the entry of imipenem into the periplasmic space is not present—see below). Newer cephalosporins such as ceftazidime, ceftriaxone, and others are efficiently hydrolyzed by the AmpC but are poor inducers, and therefore, appear active in vitro against bacteria expressing inducible AmpC.
Unfortunately, the oxyiminocephalosporins (e.g., ceftazidime, cefotaxime, and ceftriaxone) are very good selectors of mutants that express high levels of the ampC β-lactamase constitutively. Their ability to select constitutive mutants results from their status as weak inducers. Constitutive AmpC production commonly results from null mutations in ampD, with subsequent intracellular accumulation of anhydro-muramyl-tripeptide and constitutive activation of ampC expression (61). Thus, from among a population of microorganisms, the small number (1 in 106-7) of preexisting cells with mutations ampD are selected for growth by the presence of antibiotic with potent activity against strains in which ampC expression is repressed. Once constitutive expression occurs, the strains are essentially resistant to all β-lactams except for carbapenems and cefepime (64). Cefepime’s major advantage in this regard appears to be its status as a zwitterion, allowing it to achieve high periplasmic concentrations by rapid passage through the outer membrane. Caution should be exercised in using cefepime to treat deregulated ampC mutants of Enterobacter species, however, since reports of the emergence of cefepime resistance (associated with a reduction in an OMP) in these strains during therapy have been published (66).
In a study of Enterobacter bacteremia by Chow et al. (67), the major class of antibiotics associated with selection of resistance was the newer cephalosporins as opposed (especially) to the newer penicillins. Concomitant use of aminoglycosides did not prevent the emergence of this resistance.
In this study, resistance developed in 19% of all patients treated with newer cephalosporins. Therapeutic failure occurred in about half of those patients. For all patients infected with a multiply resistant strain, the mortality rate was significantly increased. Infection with a multiply resistant strain was closely associated with prior use of a new cephalosporin. Although cephalosporins have been most frequently associated with the emergence of AmpC mutants in the clinical setting, virtually any antibiotic active against repressed strains but inactive against overexpressing strains should be avoided when treating Enterobacter infections.
In this study, resistance developed in 19% of all patients treated with newer cephalosporins. Therapeutic failure occurred in about half of those patients. For all patients infected with a multiply resistant strain, the mortality rate was significantly increased. Infection with a multiply resistant strain was closely associated with prior use of a new cephalosporin. Although cephalosporins have been most frequently associated with the emergence of AmpC mutants in the clinical setting, virtually any antibiotic active against repressed strains but inactive against overexpressing strains should be avoided when treating Enterobacter infections.
Plasmid-encoded versions of AmpC enzymes have been observed in several species of Enterobacteriaceae, including E. coli and K. pneumoniae, among others (68). These strains express high levels of the AmpC enzyme constitutively and have resistance profiles identical to multiply β-lactam-resistant Enterobacter species and P. aeruginosa. The most prevalent of these enzymes is CMY-2, derived from the Citrobacter AmpC enzyme (69). The carbapenems are the only therapeutically reliable β-lactams against these strains. It is noteworthy, however, that one such enzyme, designated ACT-1, was identified in a porin-deficient strain of K. pneumoniae, where it conferred resistance to imipenem and was associated with failures of this antibiotic in clinical settings (70).
Resistance to β-lactam/β-lactamase inhibitor combinations can result from several different mechanisms, all of which involve the production of β-lactamase. As noted above, expression of an AmpC enzyme confers resistance to both cephalosporins and β-lactam/β-lactamase inhibitor combinations. Resistance to inhibitor combinations alone can be conferred by increased production of a normally susceptible enzyme (i.e., TEM-1), permeability defects, or a combination of both mechanisms (71). Specific inhibitorresistant enzymes can also result from mutation of TEM-1 or SHV-1, similar to extending the cephalosporin spectrum of these β-lactamases (72). Resistance to both extendedspectrum cephalosporins and β-lactam/β-lactamase inhibitor combinations is quite common in the clinical setting. This phenotype can be conferred by production of AmpC enzymes, by the increased production of an ESBL, or by the expression of more than one enzyme (one an ESBL, the other a more common enzyme such as SHV-1) (73,74).
Carbapenem hydrolyzing enzymes are increasingly identified. S. maltophilia is an intrinsically carbapenemresistant species that can emerge as an important pathogen in clinical settings (75). It owes its resistance to the synthesis of an inducible, zinc-dependent carbapenemase encoded on the chromosome. Cation (usually zinc)-dependent β-lactamases (generally classified as IMP or VIM enzymes) capable of hydrolyzing carbapenems have been described in several species (76). Among anaerobic bacteria, a French study showed that approximately 1% to 2% of examined B. fragilis isolates carried a carbapenemase gene, although the gene was expressed in only about half of these (77). In Acinetobacter baumannii, carbapenem resistance has been associated with the expression of class D enzymes (OXA type) (78). Finally, class A carbapenemases —previously described as chromosomally encoded enzymes occurring in scattered isolates of Enterobacter and Serratia (79)—have now spread among gram-negative bacteria via plasmids. The most common variants are the K. pneumoniae carbapenemases (KPCs). In 2001, a novel class A carbapenemase, encoded on a 50-kb nonconjugative plasmid, was described in a clinical K. pneumoniae isolate displaying high-level imipenem resistance (16 µg/mL) and termed at that time KPC-1 (80). Kinetic studies revealed that the purified enzyme hydrolyzed not only carbapenems, but penicillins, cephalosporins, and—in stark contrast to the class B metallo-carbapenemases— aztreonam as well. Concomitant losses of porin genes (ompK35 and ompK37) were also felt to play a small role in carbapenem resistance, as MICs for carbapenems were reduced in E. coli transformants with blaKPC-1 as compared to the parent strain (although still above the susceptible range). Shortly after this report, an outbreak of KPC-producing K. pneumoniae was described among ICU patients in a New York medical center (81); over the succeeding years, KPC β-lactamases have disseminated not only among most continents on the earth but also among numerous other Enterobacteriaceae and to other families of microorganisms, such as P. aeruginosa (82). The combination of carbapenem resistance, an inhibitor-resistant phenotype (83), resistance to monobactams, and the (largely unexplained) success of Klebsiella species expressing these enzymes at, disseminating geographically render these microorganisms particularly difficult to treat.
β-Lactamase Expression Combined with Reduced Access
The ultimate outcome of an interaction between a β-lactamase molecule and a β-lactam antibiotic will depend not only on the intrinsic activity of the β-lactamase but also on the quantity of the two molecules present at the time of interaction. Weak β-lactamases in sufficient concentration will be able to successfully defend against β-lactam attack, whereas even highly active β-lactamases can be overwhelmed by a sufficient quantity of β-lactam antibiotic. The ability of gramnegative bacilli to restrict β-lactam access to the periplasmic space and to concentrate β-lactamases within that space offers a powerful advantage for tilting the balance of power in favor of the β-lactamase. In P. aeruginosa, a single OMP— OMP D2—is required for transport of imipenem into the periplasmic space (84,85) (Table 85-3). Strains that decrease the expression of OMP D2 are resistant to imipenem only in the presence of the expression of the AmpC β-lactamase, even though it is an inefficient hydrolyzer of imipenem (84). This same combination of mechanisms has been shown to lead to carbapenem resistance in Enterobacter species and Proteus rettgeri (67,86,87). Clinical isolates of carbapenem-resistant K. pneumoniae expressing a plasmid-mediated AmpC β-lactamase combined with the loss of expression of two nonspecific porins have also been reported (70).
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
