Resistance of Gram-negative Bacilli to Antimicrobials
Bacterial pathogens exist in extremely large numbers, and their growth rates are generally rapid. This results in relentless evolution toward drug resistance under the selective pressure applied by the use of antibiotics in medicine and agriculture. Antimicrobial resistance has severely impacted the effectiveness of our current armamentarium of antibiotics, and the evolution of resistance will continue for any new agents that are introduced into clinical use. An understanding of drug resistance is important to prolong the effectiveness of currently used antibiotics and to inform the development of new agents. This chapter discusses antibiotic resistance in Gram-negative pathogens, beginning with the intrinsic resistance engendered by their unique outer membrane combined with active efflux and extending to the broad range of mechanisms including upregulation of efflux, alterations of cell envelope, mutation of antibacterial target genes, antibiotic-modifying enzymes, and target-protection mechanisms that are found in this diverse group of organisms. Resistance is often multifactorial, and the cumulative effect of multiple mechanisms is highlighted. Examples of these themes are provided for a range of important antibiotic classes, and efforts to address current resistance mechanisms are examined.
4.1 The Expanding Problem of Multidrug-Resistant (MDR ) Gram-negative Bacilli
Much has transpired in the realm of antibiotic resistance in the 10 years since the first edition of this text. This chapter began in 2007 with the line, “At the beginning of the twenty first century, we now find ourselves experiencing a taste of what life was like prior to the advent of the antibiotic age…,” and this reality is continuing to sink in. So much so, in fact, that antibiotic resistance is now routinely broached in the popular media and has the attention of government agencies and philanthropic groups and to some extent may be prompting a return to antibiotic discovery within the pharmaceutical industry. In 2009, the Infectious Diseases Society of America (IDSA) released the updated call to action for a coordinated effort to bring antibiotic development to the forefront, specifically regarding the “ESKAPE ” pathogens, Enterococcus faecium, Staphylococcus aureus, Klebsiella pneumoniae, Acinetobacter baumannii, Pseudomonas aeruginosa, and Enterobacter spp. . These pathogens cause the majority of US hospital infections, and resistance is a major issue. The Gram-negative bacilli are well represented in this group and pose a very significant emerging problem, particularly in the case of pan-antibiotic-resistant A. baumannii, multidrug-resistant (MDR) P. aeruginosa, and carbapenem-resistant Enterobacteriaceae (CRE). More recently the IDSA has begun “the 10x20 Initiative” (http://www.idsociety.org/10x20/). In February of 2017, the World Health Organization established its priority list for drug-resistant pathogens, and in the “critical” category are carbapenem-resistant A. baumannii and P. aeruginosa and carbapenem-resistant, extended-spectrum β-lactamase (ESBL )-producing Enterobacteriaceae (http://www.who.int/mediacentre/news/releases/2017/bacteria-antibiotics-needed/en/). The notion of tackling antimicrobial resistance was also addressed by the economist Jim O’Neill, who articulated the human costs and economic and security threats associated with a failure to act (https://amr-review.org/sites/default/files/160518_Final%20paper_with%20cover.pdf). Discussions about how to incentivize antibiotic discovery have followed and the establishment of research funding through agencies such as the Biomedical Advanced Research and Development Authority (BARDA , https://www.phe.gov/about/BARDA/Pages/default.aspx) and Combating Antibiotic-Resistant Bacteria Biopharmaceutical Accelerator (CARB-X , http://www.carb-x.org); Wellcome Trust and Pew Charitable Trust have also come online to promote antibiotic discovery. These activities, although encouraging, still underscore the challenge upon us. Therefore, an understanding of resistance in Gram-negative pathogens is informative and will be discussed in the following sections.
This chapter addresses intrinsic resistance and mutationally or horizontally acquired resistance mechanisms. Intrinsic (or innate) resistance varies widely among different pathogens and is determined by the general makeup of a cell, where the overall complement of genes and their expression levels establish a baseline susceptibility to an antibacterial. Here we focus on two broad elements important for intrinsic resistance, the impermeability of the Gram-negative cell envelope, which impedes a compound’s entry into the cell to exert its effect, and energy-dependent active efflux which extrudes a compound back out of the cell before it can engage its target. We begin there, since (i) these can be important hurdles to overcome in efforts to discover new antibiotics for Gram-negative pathogens and (ii) they can facilitate/exacerbate the emergence of mutationally or horizontally acquired resistance. The organism-specific genetic blueprint for intrinsic resistance provides the background within which mutations can be selected that further decrease susceptibility to antibacterial compounds. As well, the horizontal acquisition of new genetic material is an important route of acquired resistance. The progression to resistance is often multifactorial, and several acquired mechanisms can accumulate over time to cause clinically significant resistance and multidrug resistance (MDR). In that regard, the meaning of “resistance” is context-specific. In clinical antimicrobial susceptibility testing, resistance is based on a specific minimal inhibitory concentration (MIC ) of an antibiotic tested under standardized conditions, and clinical resistance occurs if the MIC of the antibiotic is above an established clinical resistance “breakpoint” . Here, we use the term more generally to convey the idea that the mechanisms discussed will alter (increase) the level of resistance (or decrease susceptibility), but not all resistance mechanisms will cause the specifically defined clinical resistance (shift over the breakpoint). The first edition of this chapter pertained mainly to antibiotics that had been in clinical use for some time (e.g., fluoroquinolones). These sections are updated here, but two additional aspects are now included. In 2007, tigecycline was just entering the clinic, and we update on what has happened in the approximately 12 years it has been in widespread clinical use (Sect. 4.2.4). Second, polymyxins were reintroduced into the clinic as a last line of defense against MDR Gram-negative pathogens. In a relatively short time, resistance has emerged and has begun to erode the clinical utility of these compounds, and this is discussed in Sect. 4.2.6.
4.2 Resistance in Gram-negative Bacilli
4.2.1 The Gram-negative Cell Envelope: Efflux and Outer Membrane Impermeability
126.96.36.199 Active Efflux
Example RND efflux pumps in Gram-negative pathogens and range of antibiotics accommodated by each pump
AG, CM, FQ, TC (MC), TG
BL, CM, EM, FQ, TC (MC), TG
CM, FQ, TM
BL, CM, FQ, ML, NO, RF
AG, FU, NO
BL, CM, EM, FQ, TG
AG, BL, CM, ML, NO, TC, TG, TM,CM, CP, FQ, TC
CM, EM, FQ, TC
AG, ML, TC, TG
AG, ML, TC, TG
S. enterica serovar
BL, CM, EM, FQ, NO, RF, TC
AG, BL, FQ
EM, FQ, TC (MC), TG
188.8.131.52 Mechanism of Efflux by RND Family Pumps
184.108.40.206 The Gram-negative Outer Membrane (OM) Permeability Barrier and Its Interrelationship with Efflux
The requirement for nutrient uptake across the OM bilayer is generally met by water-filled protein β-barrel channels that span the bilayer, known as porins (see Fig. 4.1). Porins allow for the passage of small hydrophilic molecules across the OM, essentially establishing the overall OM as a molecular sieve. These channels are also thought to allow influx of certain hydrophilic antibiotic molecules that are small enough to traverse the porin channels . Gram-negatives such as E. coli have several relatively nonspecific large porins such as OmpC and OmpF with molecular weight cutoffs of approximately 600 Da . In contrast P. aeruginosa harbors a number of more specialized or restrictive (smaller) porins to allow influx of nutrients. This organism also has the general porin OprF, but this exists only occasionally in the conformation that allows the channel to be open . This highlights that, along with variability in the lipid bilayer characteristics among different Gram-negative bacteria, the number and characteristics of the OM porin channels can also vary, causing large differences in the effectiveness of the OM permeability barrier. Reflecting this, the OM of P. aeruginosa was estimated to be more than tenfold less permeable than that of E. coli . The best studied examples of antibiotic permeation via porins center on various β-lactams. For example, carbapenems can enter E. coli via OmpC porins, and some carbapenems such as imipenem enter P. aeruginosa via OprD. In the latter case, the natural function of OprD is transport of basic amino acids, which bear some structural resemblance to certain carbapenems. Imipenem was also recently shown to enter P. aeruginosa via OpdP , suggesting that some antibiotics may access cells via multiple porins. Hydrophobic and/or larger compounds not able to enter by porins can enter the bacterium by diffusion across the membrane bilayer, although these processes are likely slower and are not well understood. As well, there are a limited number of specialized energy-dependent active transporters in the OM that import scarce nutrients such as iron (as siderophore or protein-bound complexes)  or cobalamin .
Overall, the OM of Gram-negative bacteria is highly evolved to provide a strong protective permeability barrier, but it still allows influx of important nutrients, either by passive diffusion in the case of porins or active transport in other cases. For most antibiotics though, the OM slows their influx considerably, and the combination of reduced antibiotic influx due to the OM permeability barrier and active efflux together typically determines the levels of susceptibility [34, 35]. RND family efflux pumps do not always exhibit a high velocity of antibiotic efflux, so beyond whether the pump recognizes a specific antibiotic, its effectiveness depends to a large extent on how slowly a given substrate antibiotic is entering the cell. If the influx rate is too fast, the pump may not “keep up,” and even if the compound is a substrate, the pump may not confer meaningful resistance. In contrast, if the membrane barrier is slowing influx, efflux pumps can then become a very significant resistance factor. This was elegantly shown for oxacillin compared to ampicillin in E. coli. These antibiotics were shown to be very similar as substrates of the AcrB efflux pump, but oxacillin traversed porins more slowly. Deletion of the acrB gene had a much larger impact on susceptibility to oxacillin than to ampicillin [36, 37]. Furthermore, disruption of either efflux or the OM permeability barrier in P. aeruginosa strongly increased antibiotic susceptibility, with an even greater increase in susceptibility when both were disrupted simultaneously, showing the interplay between these two factors . More recently it was reported that expression of an OM iron-siderophore transporter that had been engineered to create very large porin-like channels in the bacterial OM strongly increased susceptibility to a range of antibacterial compounds, further showing the important role of the OM permeability barrier and its interrelationship with active efflux in several Gram-negative pathogens [39, 585]. Correspondingly, mutations that impact either the OM permeability barrier or efflux in Gram-negative bacteria can decrease susceptibility to antibiotics, and this is discussed in the next section.
220.127.116.11 The Role of Efflux and the OM Permeability Barrier (Cell Permeability) in Decreasing Antibiotic Susceptibility
The combination of the OM permeability barrier and efflux is important for dictating the spectrum of Gram-negative pathogens that a clinically used or novel antibiotic under development will be sufficiently active. In those cases where useful antibacterial activity does occur (e.g., with currently used antibiotics), mutations that increase the expression of efflux pumps, alter substrate recognition, or impact the OM permeability barrier (decreased compound influx) erode this activity over time and limit a compound’s therapeutic longevity. Since many RND efflux pumps have broad substrate ranges, selection of pump upregulation with one compound will usually affect susceptibility to multiple antibiotics, contributing to multidrug resistance. These factors are particularly problematic since they contribute to the therapeutic demise of current antibiotics while also being the major impediment to the discovery of novel replacement antibiotics. As an example illustrating both of these points, tigecycline, a glycyl derivative of minocycline that evades the classic tetracycline-specific resistance mechanisms of ribosomal protection and efflux by tetracycline-specific single-component efflux pumps (e.g., TetA) (discussed in detail in Sect. 4.2.4), was still subject to intrinsic RND-mediated efflux in P. aeruginosa , and therefore its spectrum does not include this pathogen. Although tigecycline achieves useful antibacterial activity against other Gram-negatives and has been successfully implemented clinically, mutations leading to RND pump upregulation can erode this activity in organisms such as Proteus mirabilis , K. pneumoniae , E. coli , and A. baumannii .
The impact of cell impermeability on the discovery of new anti-Gram-negative antibiotics is difficult to overstate, as the vast majority of compounds are subject to some level of efflux and/or limited influx. In one study, the majority of antimicrobial compounds identified from direct antibacterial screening in E. coli were AcrAB-TolC pump substrates . In a broader discussion of overall screening efforts conducted at AstraZeneca, the inability of compounds to accumulate in Gram-negative bacteria was cited as a significant impediment to novel antibiotic discovery using corporate compound collections . Additional examples of novel compounds that inhibit specific bacterial targets but are subject to efflux include the peptide deformylase inhibitor LBM415, which is subject to AcrAB-TolC-mediated efflux in H. influenza , and CHIR-090 an LpxC inhibitor that has potent intrinsic activity against P. aeruginosa but selects in vitro for mutations that upregulate expression of several RND efflux pumps . Similarly, standard antibiotics with good intrinsic activity against Gram-negative pathogens, such as fluoroquinolones, many β-lactams, and aminoglycosides, also select for mutations causing increased pump expression which can decrease susceptibility substantially .
The selection of mutations leading to pump upregulation underscores the idea that although RND pump expression is generally subject to intricate regulation, pump expression is not typically induced by antibiotics of clinical importance. An exception is the strong induction of the MexXY efflux pump of P. aeruginosa by compounds that perturb protein synthesis. Even in that case, however, induction occurs in response to a range of structurally and mechanistically unrelated protein synthesis inhibitors including aminoglycosides and tetracyclines , novel ribosome inhibitors such as argyrin B  or to mutations that impair ribosome function [50, 51]. Therefore, MexXY expression is responsive to ribosome impairment [52, 53] rather than to the specific antibacterial compounds, and in some cases such as argyrin B, the inducing compound may not be a pump substrate. Novobiocin was also shown to directly bind to the NalD repressor of the MexAB-OprM efflux pump and induce pump expression .
In general, pump upregulation leading to decreased susceptibility to antibiotics generally occurs by selection of stable mutations in the pump gene promoter region or more often in regulatory genes. The circuits controlling efflux pump expression can be highly complex, and overall this topic is beyond the scope of this section, but in general, efflux pump regulation often involves repressor proteins that bind operators upstream of pump genes to reduce pump expression unless relieved by either an intracellular signal or interaction with other modulatory elements (e.g., MexR and NalD, which control expression of MexAB-OprM in P. aeruginosa [55, 56, 57]), positive activators that bind and induce pump gene expression in response to intracellular signals (e.g., MexT which activates expression of MexEF-OprN in P. aeruginosa [58, 59]), or in some cases two-component histidine kinase sensor response regulator pairs (e.g., BaeRS and CpxRA [60, 61, 62, 63]). In most Gram-negative bacteria, there is usually a housekeeping pump (e.g., MexAB-OprM of P. aeruginosa or AcrAB-TolC of E. coli) that is expressed constitutively, with additional pumps that are not appreciably expressed, at least under laboratory conditions (e.g., MexCD-OprJ of P. aeruginosa and AcrEF-TolC of E. coli). Expression of pumps such as MexAB-OprM and AcrAB-TolC can be increased by mutations in genes encoding their cognate repressors (e.g., mexR  and nalD  or other regulators (nalC ) for MexAB-OprM or acrR  for AcrAB-TolC). Mutations in genes controlling typically silent pumps, such as MexCD-OprJ and MexEF-OprN, can turn on expression to generally high levels with corresponding increases in resistance to their substrate antibiotics [59, 68]. Upregulated pumps are routinely found among clinical isolates . Since pump upregulation can result from simple loss-of-function mutations in repressor genes, these mutants can be selected at high frequencies under antibiotic exposure levels within which the efflux pump can accommodate. Furthermore, since RND or other pumps can extrude common biocides, such as chlorhexidine, pump upregulation is likely selected in the environment by biocide-containing cleaning solutions .
The complex regulatory circuits controlling expression of some efflux pumps can also control the expression of OM porins through which some antibiotics cross the OM. For example, the highly complex MAR (multiple antibiotic resistance (reviewed in [3, 7, 70]) regulatory circuit controls AcrAB-TolC expression and porin expression in E. coli. Therefore mutants having both reduced antibiotic influx and increased efflux via AcrAB-TolC can emerge. Similarly, P. aeruginosa mutants that overexpress MexEF-OprN are also downregulated for expression of the porin OprD, the main entry route of carbapenems into the cell . Reduced susceptibility to carbapenems in nfxC mutants is thought to be mediated mainly by reduced influx rather than efflux. Mutations affecting porin expression or function, including mutations within porin genes, have been described in several bacteria and in particular are associated with carbapenem resistance in Enterobacteriaceae and P. aeruginosa clinical isolates [71, 72, 73, 74]. Mutations in genes encoding the two-component regulator ParRS in P. aeruginosa were shown to cause inducible or constitutive resistance to four classes of antibiotic (polymyxins, aminoglycosides, fluoroquinolones, and β-lactams) via a combination of increased efflux (MexXY/OprM), porin downregulation, and aminoarabinose modification of lipopolysaccharide (LPS ) . The latter affects the ability of polymyxin (cationic peptides) to interact with LPS which is required for their entry into cells (discussed in Sect. 4.2.6). These examples serve to illustrate the extensive ability of many Gram-negative bacteria to survive exposure to toxic molecules by preventing their accumulation in the cell. The ability of mechanisms such as efflux and porin loss to enhance survival under exposure to antibiotics also supports the emergence of other resistance mechanisms such as specific target mutations, ultimately facilitating the emergence of very high levels of resistance.
18.104.22.168 Efforts to Address Compound Accumulation in Gram-negative Bacteria
Lack of sufficient compound accumulation in cells is arguably the single most important specific factor hindering the discovery and development of novel antibiotics for Gram-negative pathogens. The search for novel antibacterials with new mechanisms of action is predicated on avoidance of cross-resistance with existing mechanisms selected by currently used antibiotics. The presence of RND efflux pumps (and selection of pump over-expressors clinically), as a general broad resistance mechanism, often serves to defeat this strategy in cases where the novel compound is a pump substrate and has an intracellular target. It is reasonable to speculate that the efflux-/OM-mediated permeability barrier has coevolved with many intracellular targets that are essential for growth or viability in order to exclude most molecules with the characteristics required to optimally bind and inhibit intracellular essential targets. Therefore, strong interest has developed within the antibiotic discovery field in understanding the Gram-negative OM permeability barrier and efflux with a view toward two general goals: interfering with cell impermeability as a way of potentiating antibiotic activity by increasing the cellular accumulation of a partner antibiotic (combination therapy) and the understanding of the design of inhibitors that are less impacted by efflux and can penetrate cells effectively to reach their target. See the Innovative Medicines Initiative (IMI) translocation effort (https://www.imi.europa.eu/content/translocation) for more details.
Potentiation of the Cellular Activity of Antibiotics
One strategy to potentiate partner antibiotics that has garnered extensive interest over the years is the design of efflux pump inhibitors (EPIs ). In theory, a potent EPI could improve the spectrum and potency of a range of currently used antibiotics whose usefulness is compromised by efflux. EPIs could also serve to enhance and extend the clinical usefulness of novel agents that are or may become affected by efflux. Several EPIs have been described over the last two decades.
The first EPI described in detail as an inhibitor of multiple RND family pumps, MC207,110 (phe-arg-β-napthylamide, PAβN) was originally identified by screening for compounds that potentiated the activity of pump substrate fluoroquinolone antibiotics in P. aeruginosa . Inhibitors such as MC 207,110 are pump substrates and are thought to act through competitive binding and interference with substrate antibiotic recognition . MC207,110 is a lipophilic amine and as such falls into a chemical property space known for target promiscuity and associated challenges in achieving an acceptable safety profile . Perhaps reflecting this, MC207, 110 was also shown to have some bacterial membrane-disrupting activity [76, 79]. In contrast to MC207,110, the EPI D13-9001 is more specific to the MexAB-OprM efflux pump within P. aeruginosa . Compounds like D13-9001 would therefore need to be partnered with an antibiotic effluxed primarily by MexAB-OprM. The mechanism of pump inhibition involves binding of D13-9001 into a hydrophobic “trap” extending off the substrate translocation channel within MexB, ultimately compromising the pump’s functional rotation [81, 82]. This mechanism is more likely than that of a substrate competition mechanisms, such as that of MC207,110, to block efflux of multiple antibiotics and to enable better inhibitor potency. Furthermore, D13-9001 is zwitterionic which places it in a potentially less promiscuous chemical property space. Newer pyranopyridine EPI molecules (e.g., MBX2319 and analogs [83, 84, 85] that take advantage of the hydrophobic trap mechanism have achieved substantial increases in potency. Recently described EPIs (NSC 60339 and analogs) were shown to have a very novel mechanism of inhibition of AcrAB-TolC, by binding the membrane fusion protein (AcrA), inducing structural changes, and possibly interfering with assembly of the functional pump . Advances in the structural understanding of pump assembly and function, as well as the binding of several EPI molecules and the diversity of potential pump inhibitory mechanisms, may increase the ability to design new EPI molecules in the future.
Selection of a suitable partner antibiotic is potentially a complex issue, especially with clinically used antibiotics. This is because non-efflux-based resistance mechanisms (e.g., target or modifying enzyme-based mechanisms) affecting many standard antibiotics may have become widespread, and those mutants may be resistant even if efflux is fully inhibited. Secondly, many antibiotics are effluxed by several pumps within a given Gram-negative pathogen or across different bacteria, thereby requiring a broad spectrum of EPI activity to cover multiple pumps. To date no Gram-negative EPI has reached clinical use. MC207,110-based analogs were lipophilic cations, and ultimately unfavorable toxicity profiles could not be overcome . It remains to be seen if this approach will be successful with other novel inhibitors. Finally, the intriguing finding that a significant percentage of P. aeruginosa clinical isolates recovered from cystic fibrosis patients have mutationally lost MexAB-OprM function and become susceptible to ticarcillin has prompted suggestions that this antibiotic may find use in treating this subpopulation .
An alternative approach to efflux inhibition may be disruption of the bacterial membrane, thereby improving the ability of a partner antibiotic to gain access to the cell. It is well established that mutations affecting the synthesis or assembly of the Gram-negative OM cause hypersusceptibility to a range of antibiotics, especially those that are more hydrophobic in nature or are large molecular weight (i.e., too large to pass through porins and whose exclusion from cells is mediated mainly by the permeability barrier of the OM bilayer). Targets important for OM biosynthesis/assembly that may affect the OM permeability barrier if inhibited include the lipid A biosynthetic genes lpxA, lpxB, and lpxC [88, 89] and the LPS transport/assembly genes lptD and lptE in E. coli [90, 91, 92] and P. aeruginosa [93, 94]. Chemical inhibitors of targets such as these could induce disruption of the OM permeability barrier and strongly potentiate the activity of many antibiotics used in combination, although precisely where this may occur at a clinically useful level remains to be determined. Alternatively upon target inhibition, a corresponding progressive disruption of the permeability barrier may generate a cycle of increased uptake of the inhibitor itself, improving cellular potency, although again this remains to be shown for specific examples. Since LPS synthesis and assembly per se are essential in many Gram-negative bacteria, enzymes such as LpxC (first committed step in lipid A biosynthesis), LpxA/LpxD, and LptD have generated interest as targets for the design of novel inhibitors. For example, extensive medicinal chemistry efforts have resulted in potent small molecule inhibitors of LpxC with antibacterial activity [95, 96, 97, 98, 99, 100, 101, 102, 103, 104, 105, 106, 107, 108], at least one of which reached early phase clinical evaluation . P. aeruginosa-specific peptidomimetics targeting LptD (POL7001, POL7080) have very potent antipseudomonal activity [108, 110], suggesting the potential of this approach. Intriguingly, there are a small number of Gram-negative bacteria that can survive in the absence of lipid A biosynthesis, including the important pathogens Neisseria meningitides [111, 586] and some A. baumannii strains [112, 113, 114]. Loss of lipid A in the latter has been shown to result from mutations in lipid A biosynthetic genes lpxA, lpxC, or lpxD , and lpxC and lptD can be genetically deleted in some A. baumannii . Therefore, inhibitors of these targets would not be expected to have good, or any, antibacterial activity against such strains. However, reflecting the potential of inhibitors of these targets to potentiate antibiotics, strains harboring these genetic mutations or wild-type strains exposed to LpxC inhibitors became susceptible to several antibiotics [113, 115, 116]. Importantly, an intact OM permeability barrier is also important for survival of pathogens (virulence) during infection, by providing protection from serum complement and other host immune factors.
Molecules such as LPS are also strong activators of toll-like receptors (e.g., TLR4). Consistent with this, an LpxC inhibitor with no appreciable in vitro antibacterial activity against A. baumannii was highly protective in a mouse infection model . This was attributed to increased opsonophagocytic killing and lower levels of released LPS, leading to reduced inflammation. This example illustrates the additional potential of inhibitors of targets such as LpxC, used alone or in combination with other antibiotics, resulting from effects on the OM. A. baumannii represents an extreme example to show this, since some A. baumannii can tolerate a large or total loss of lipid A synthesis, providing the widest possible window to see this effect. It remains to be understood how broadly this might translate across different targets/inhibitors and different Gram-negative pathogens in clinically relevant scenarios. As with any novel antibacterial, in particular ones that must still reach an intracellular target such as LpxC, a variety of resistance mechanism are likely to be able to impact their cellular activity. In vitro, mechanisms such as target mutations, target overexpression, partial bypass by mutations in fabG, and upregulated efflux can all reduce the activity of the LpxC inhibitor CHIR-090 in P. aeruginosa , and mutations in fabZ reduce susceptibility of E. coli to LpxC inhibitors . In contrast to targets like LpxC which are intracellular, the OM itself is also a target, and compounds that could interact with and disrupt the OM might potentiate antibiotics without having to reach an intracellular target. Classic examples of this are cationic molecules, such as polymyxins, which interact with LPS via phosphates attached to lipid A and disrupt the OM permeability barrier causing sensitization to antibiotics [119, 120, 121]. Such an approach to potentiate antibiotics against Gram-negative pathogens (compound SPR741, an analog of polymyxin B nonapeptide) is currently being pursued .
Understanding Compound Penetration to Improve Access to Targets
While factors limiting a compound’s OM and inner membrane (IM) permeability are fairly well understood, structure activity relationships for efflux remain incomplete. Strategies to improve compound access to intracellular targets must consider the specific compartment in the cell where the target resides. A target might reside anywhere from the cell surface to the periplasmic space between the inner and OM through to the cytosol of the cell. Different issues come into play for each of these. One way to reduce the complexity of cell penetration is to pursue targets located near or on the cell surface, circumventing the need for significant cell penetration. As mentioned above, polymyxin-based antibiotic potentiators such as SPR741 directly target the LPS on the cell surface. Similarly, the cationic peptidomimetic POL7001 targets LptD, an essential protein localized in the OM  at or near the cell surface. The number of OM protein targets that are essential for growth is low (i.e., LptD and BamA), and these targets exist as components of complex machinery which could be more difficult to disrupt or inhibit by small molecules, but this remains to be fully understood. To date, only larger peptidomimetic inhibitors of these targets have been described [110, 123], providing some possible insights into the types of molecules expected to be active in this context. On a related note, eliminating the need for cell penetration is a salient feature of the monoclonal antibody approach, which functions specifically by exploiting surface-exposed antigens. An example is a recently described bispecific antibody (MEDI13902) targeting PcrV (type III secretion) and the exopolysaccharide Psl [124, 125] which is currently undergoing clinical trials for prevention of nosocomial pneumonia. Progressing further into the cell, some targets reside in the periplasmic space between the IM and OM. The number of essential targets here is also limited but includes the clinically validated penicillin-binding proteins (PBPs ) that are inhibited by the β-lactam class of antibiotics. Inhibitors of periplasmic targets need only to cross the OM, which lends itself to utilization of water-filled porins for compound access, with corresponding optimization for this route of entry. Entry through porins relies to a large extent on a size small enough to traverse the porin (porin cutoff is approximately 600 Da ), compound polarity (hydrophilicity), and appropriate charge distribution. The mechanisms by which translocation of compounds occurs via porins have been the subject of extensive investigation [126, 127, 128], and assays to evaluate porin translocation for use in drug design are being pursued [129, 130]. A recent study suggested that porin traversal may be optimal for small polar compounds with charged groups and a dipole moment having a component aligned perpendicular to its main axis . Since β-lactams have periplasmic targets, they may represent the class of antibiotic that can best be optimized specifically for permeation through porins. It is likely that this contributes in some cases to reducing the impact of efflux since rapid influx can overwhelm the capacity of RND family efflux pumps even if the compound is a pump substrate [36, 37]. The majority of novel antibacterial targets however are located in the cytosol, and understanding compound penetration (and evasion of efflux) becomes more complex in that case than for compounds such as β-lactams. This stems from the necessity to traverse the two distinctly different membranes. The OM severely limits influx of larger or more hydrophobic compounds, which must diffuse across the asymmetrical OM bilayer which has low fluidity and presents a high barrier for lipophilic compounds, since they are excluded from entry through porins . Smaller more hydrophilic molecules can traverse the OM through porins, but extensive optimization for polarity to maximize this can hinder entry across the symmetrical phospholipid inner membrane bilayer, which favors diffusion of hydrophobic molecules. Antibiotics directed at cytosolic targets may therefore need some element of amphiphilicity to cross both membranes, which also may increase recognition by RND efflux pumps . A much better understanding of the chemical property space required for the design of cell active inhibitors of intracellular targets, and the representation of this chemical property space typically found in corporate screening libraries, is likely required to overcome the ongoing inability to deliver novel antibacterials in this area. Initial efforts toward this understanding were described by O’Shea and Moser , who examined the properties of a wide range of antibacterial compounds and, consistent with our understanding of the cell envelope, correlated properties such as molecular weight and polarity with Gram-negative antibacterial activity. A more recent look at data from a wide range of screening efforts at AstraZeneca also suggested that polarity and small size correlated with reduced efflux and cellular activity and increasing compound hydrophobicity could drive biochemical target inhibition but possibly at the expense of cellular activity . However, increased polarity itself was not sufficient to ensure antibacterial activity , again suggesting a fine balance of properties is likely necessary. This is consistent with the notion that the cell envelope and efflux likely coevolved with intracellular targets to exclude molecules with the properties to strongly bind and inhibit essential targets. These properties may also be compound scaffold specific. Modulation of physicochemical properties (pKa and logD) improved the antibacterial activity of novel bacterial type II topoisomerase inhibitors . A very recent study directly measured accumulation of compounds in E. coli using mass spectrometry, and computational analysis indicated that rigid, amphiphilic compounds with low globularity and containing an amine moiety accumulated better. These rules were applied to convert a compound that was active only against Gram-positive bacteria into one with E. coli activity . This suggests it may be possible to derive some general rules to engineer compounds with Gram-negative accumulation and cellular activity, but more research will be necessary to validate this concept. Furthermore, compound accumulation must be achieved together with low toxicity in order for resulting compounds to be therapeutically useful. Efforts are also underway to explore new methods complementary to mass spectrometry to measure cellular accumulation of compounds, which could further assist in defining rules for cell penetration [135, 136]. Finally, the Trojan-horse concept has also been applied to antibacterial design. In this scenario, an antibiotic is linked to a compound that is actively transported into cells, thereby exploiting the active uptake mechanism to drive intracellular accumulation of the chimeric antibiotic molecule. Among others, a recent example of this is compound cefiderocol (S-649266), a catechol cephalosporin that is proposed to utilize energy-dependent siderophore-iron uptake systems in Gram-negative bacteria for improved cellular access .
4.2.2 β-Lactams and β-Lactamase Inhibitors
β-lactams are bactericidal and target penicillin-binding proteins (PBPs ) . Gram-negative bacteria possess multiple PBPs which are important for cross-linking of the peptidoglycan that makes up the rigid bacterial cell wall . β-lactams resemble segments of the growing peptidoglycan (e.g., D-Ala-D-Ala) and, after the formation of a low-affinity complex, covalently bind (acylate) the PBP at its active-site serine residue. Bacteria possess multiple PBPs, broadly classified into high-molecular-weight (HMW) and low-molecular-weight (LMW) categories . In general, the LMW PBPs are monofunctional D-Ala-D-Ala carboxypeptidases, whereas the HMW PBPs are either bifunctional (class A, transpeptidase and transglycosylase) or monofunctional (class B, transpeptidase) . Not all PBPs are essential for growth, but certain ones such as PBP3 (main target of aztreonam) are essential [141, 142]. Furthermore, many β-lactams can acylate the active-site serine of several PBPs which can contribute meaningfully to increased antibacterial activity. Inhibition of PBPs like PBP3 or multiple PBPs like PBP3 with PBP1a/PBP1b can ultimately lead to cell lysis [140, 142]. Recent studies have now elucidated that inhibition of PBPs also triggers a lethal malfunctioning of the cell wall synthetic machinery .
The high potency of many β-lactams against Gram-negative pathogens relies to a large extent on two factors, compound access and the nature of the targets themselves. PBPs are located in the Gram-negative periplasmic space. This means that β-lactams only need to cross the OM to access their targets. Therefore β-lactams can be polar molecules able to traverse water-filled porins, which have the fortuitous benefit of improving their safety. The second factor is that these relatively “exposed” PBPs are particularly good targets in conjunction with this class of inhibitor, since the compound-target interaction is covalent (essentially irreversible), and PBP inhibition causes severe impairment of a fundamental cellular process, leading ultimately to lethality. However, the PBP targets also form the basis of the overarching issue with resistance to β-lactams, which is the expression of β-lactamase enzymes. These enzymes are expressed in the periplasm and appear to have evolved from PBPs to attack and efficiently hydrolyze the β-lactam, mediating resistance . Non-β-lactamase mechanisms that affect susceptibility to β-lactams in Gram-negative pathogens include efflux , loss of uptake porins, and amino acid substitutions in the target PBPs. These topics are addressed in the following sections.
Classification of clinically relevant β-lactamases
Representative enzyme in clinical isolates
Extended-spectrum β-lactamases (ESBLs )
Penicillins, cephalosporins, monobactams
TEM, SHV, CTX-M, PER, VEB
TEM-3, SHV-2, PER-1, VEB-1, CTX-M-15, CTX-M-9, CTX-M-14, CTX-M-3
Penicillins, narrow-spectrum cephalosporins
Carbapenems, cephalosporin, cephamycins
KPC, IBC, IMI, NMC, SME, GES, SFC,
KPC-1, KPC-2, KPC-3, SME-1
Carbapenems, penicillins, cephalosporins, cephamycins
IMP, VIM, NDM, SPM, GIM, SIM, AIM, DIM, FIM, POM
Cephamycins, cephalosporins, narrow-spectrum monobactams, and penicillins
CMY, FOX, ACC, LAT, ACT, MOX, DHA, MIR,
CMY-1, CMY-2 ACT-1, DHA-1, DHA-2, CMY-13, CMY-4
Extended-spectrum β-lactamases (ESBLs )
OXA-10, OXA-13, OXA15, OXA-18, OXA-45
OXA-48, OXA-23 OXA-40, OXA-51, OXA-58
Extended-Spectrum β-Lactamases (ESBLs )
Extended-spectrum β-lactamases (ESBLs) confer resistance to nearly all β-lactam antibiotics except carbapenems and cephamycins. ESBLs were first identified in the mid-1980s in K. pneumoniae and Serratia marcescens . The occurrence of ESBLs in clinical isolates has been constantly increasing in the past two decades. A recent Centers for Disease Control and Prevention (CDC) report estimated nearly 26,000 healthcare-associated Enterobacteriaceae infections are caused by ESBL-producing Enterobacteriaceae (19% of isolates) causing 1700 deaths each year (https://www.cdc.gov/drugresistance/threat-report-2013/index.html). The fast spread of ESBL-producing strains is due to the presence of these genes on mobile genetic elements, such as plasmids, usually carrying other antibiotic resistance genes [145, 152, 153]. Early ESBLs evolved from the TEM and SHV enzymes to be able to hydrolyze oxyimino-cephalosporins, and these are molecular class A, functional group 2be. Subsequently, the ESBL category expanded to include enzymes such as the CTX-M family, mainly present in E. coli and K. pneumoniae; the PER family identified in Pseudomonas, Acinetobacter, and Salmonella species; and the VEB family reported in A. baumannii. These β-lactamases are not genetically related to TEM or SHV β-lactamases but have similar hydrolytic profiles and are part of the functional group 2be [153, 154, 155]. The most recent ESBLs are the OXA family, originally reported in P. aeruginosa, isolated in Turkey and France. The OXA family, in contrast to the other ESBLs, belongs to molecular class D and functional group 2de .
ESBLs are prevalent in the clinic and present serious problems in hospital-acquired infections, leading to increased mortality worldwide. ESBL prevalence varies across different geographic regions. In a recent report on Enterobacteriaceae isolates collected in 63 US hospitals from 2012 to 2014, 13.7% of these isolates had an ESBL profile. Different trends were observed among different species of Enterobacteriaceae; in E. coli the ESBLs occurrence increased from 12.7% (2012) to 15.1% (2014), whereas in K. pneumoniae, rates decreased from 18.9% (2012) to 15.5% (2014). Also, a statistically significant variation was observed across different regions in the United States. In the South Atlantic Region, the ESBL rates decreased from 20.8% (2012) to 9.2% (2014). Conversely in the Pacific region, the ESBL rates increased from 11.4% (2012) to 16.9% (2014). The predominant ESBLs identified in this study were CTX-M-15 (59% of ESBLs) followed by SHV (19% of ESBLs), both mainly in K. pneumoniae isolates, and CTX-M-14 (18% of ESBLs) . In Europe the ESBL rates vary considerably by country. The prevalence of ESBLs in E. coli in the 2014 European surveillance varies from 3.3% in Iceland to 40.4% in Bulgaria. Even more alarming is the prevalence of ESBLs in K. pneumoniae with rates over 70% in Greece, Bulgaria, and Romania. On a positive note, the rates of ESBLs did not increase from 2009 to 2014, attributed to the increased use of carbapenems . Also across Europe, the most prevalent ESBL types identified in clinical isolates were the CTX-M family β-lactamases, but the specific type varies considerably among countries, with CTX-M-9 and CTX-M-14 enzymes dominant in Spain and CTX-M-3 and CTX-M-15 dominant elsewhere .
Another growing family of ESBLs is the OXA-type enzymes that confer resistance to ampicillin and cephalothin, are characterized by their high hydrolytic activity against oxacillin and cloxacillin, and are very poorly inhibited by clavulanic acid. The OXA-type enzyme genes differ genetically from all other ESBLs . To date, over 500 different OXA-type variants have been reported, but not all are ESBLs. The OXA-type enzymes with activity against oxyimino-cephalosporins are OXA-10 and its variants (OXA-11, OXA-14, OXA-16, and OXA-17), OXA-13 and its variants (OXA-19 and OXA-32), and some other OXA enzymes (OXA-15, OXA-18, and OXA-45). These enzymes have been identified mainly in P. aeruginosa isolates [155, 160].
Even though ESBL incidence rates have not been increasing in the past few years, they are still very high in some parts of the world and are a major health concern. Further, ESBLs are often present on mobile genetic elements with other antibiotic-resistant determinants, including those for aminoglycosides and fluoroquinolones. The use of carbapenems to treat infections caused by ESBL -producing pathogens is increasing the emergence of carbapenem-resistant strains, starting the debate on how to better treat those pathogens. Using a β-lactamase inhibitor (see Sect. 22.214.171.124) with a β-lactam is in principle a targeted and effective approach. A detailed analysis on the benefit of β-lactam/β-lactamase inhibitor combinations for the treatment of ESBL-producing pathogens can be found in a recent review by Viale et al. .
Class C β-Lactamase (AmpC )
AmpC β-lactamases belong to class C and functional group 1. They confer resistance to cephamycins, such as cefoxitin and cefotetan, and cephalosporins, including oxyimino-cephalosporins such as ceftazidime, cefotaxime, and ceftriaxone. They are also able to hydrolyze to a lesser extent penicillins and aztreonam . The majority of AmpC β-lactamases are not or are only weakly inhibited by inhibitors of class A enzymes such as clavulanic acid, sulbactam, and tazobactam. Some AmpC variants have been reported to be inhibited by tazobactam or sulbactam [162, 163]. Several AmpC β-lactamases are chromosomally encoded enzymes, found in Acinetobacter spp., C. freundii, Enterobacter spp., E. coli, Hafnia alvei, Morganella morganii, P. aeruginosa, and Yersinia enterocolitica. AmpC basal expression is generally low but can be induced to high levels in some bacteria (e.g., M. morganii and P. aeruginosa) upon exposure to some β-lactams. The regulation of AmpC expression varies among different organisms. In bacteria where AmpC is inducible, the ampC gene is accompanied by ampR, encoding a member of the LysR transcriptional regulator family. Disruption of peptidoglycan synthesis by β-lactams causes accumulation of peptidoglycan fragments that dislodge oligopeptides of UDP-N-acetylmuramic acid normally bound to AmpR, causing a conformational change where AmpR then positively activates transcription of ampC. The activity of AmpR is controlled indirectly by the activities of AmpD, a N-acetyl-muramyl-L-alanine amidase, and an inner membrane permease, AmpG, which are both involved in recycling of peptidoglycan intermediates. Therefore, although ampC expression is inducible in these cases, constitutive upregulation can occur via mutations in the genes encoding these regulatory factors. In clinical isolates, the most common cause of AmpC hyperexpression is mutation in ampD. Mutations in ampR causing AmpC hyperexpression have been reported but are not as common. Mutations in ampG only result in constitutive low-level expression and are the least common . In P. aeruginosa PAO1, AmpC expression is very tightly regulated by the presence of three AmpD genes with different affinities for their substrates. These AmpD genes are also reported to be involved in P. aeruginosa PAO1 virulence . Other organisms like E. coli and Shigella lack AmpR, and regulation occurs via a weak promoter and a strong attenuator. In these cases, AmpC expression is not inducible by β-lactams, and hyperexpression of AmpC leading to resistance results from mutation in the ampC promoter or attenuator [165, 166, 167]. AmpC β-lactamases can also be expressed from plasmids. The first plasmid-encoded AmpC variant, CMY-1, was identified in 1989 from K. pneumoniae isolated from a wound infection in South Korea. The high degree of resistance to cefoxitin was due to the high-level constitutive expression of CMY-1 [162, 168]. Since the identification of CMY-1, several families of plasmid-encoded AmpC variants have been reported in clinical isolates, especially in K. pneumoniae and E. coli.
Based on the source of the ampC gene, several plasmid-encoded AmpC families have been reported: the two CMY families (CMY-1 and CMY-2), the FOX family, the ACC family, the LAT family, the MIR family, the ACT family, the MOX family, and the DHA family [162, 169, 170]. Plasmids encoding ACT-1, DHA-1, DHA-2, and CMY-13 typically contain an ampR gene, and as such expression of these β-lactamases is inducible, whereas the other plasmid-encoded AmpC variants lack ampR and are not inducible. The high level of expression for the non-inducible plasmid-encoded AmpC variants is mainly due to strong promoters and high-gene copy number. As with several other plasmid-borne antibiotic resistance genes, plasmids harboring AmpC β-lactamase genes often carry resistance determinants for fluoroquinolones, sulfonamides, tetracyclines, aminoglycosides, chloramphenicol, trimethoprim, and other β-lactamases (ESBLs and metallo-β-lactamases) [162, 168, 171, 172]. Clinical isolates of K. pneumoniae and Salmonella enterica carrying plasmid-encoded AmpC have been reported to be resistant to cephalosporins and cephamycins as well as carbapenems. Detailed analysis of these strains showed that the resistance to cephalosporins and cephamycins was due to the plasmid-mediated CMY-4 β-lactamase for Salmonella enterica and DHA-1 for K. pneumoniae, whereas resistance to carbapenems also involved the lack of outer membrane porin proteins [173, 174, 175, 176, 177]. Combination of plasmid-encoded AmpC with deletions of porin genes (and possibly increased efflux) results in high-level resistance to most if not all β-lactams and leaves clinicians with few treatment options.
Carbapenems are considered the most effective β-lactams for the treatment of serious infections caused by Gram-negative bacteria and present a broad spectrum of antibacterial activity. Furthermore, carbapenems are relatively stable to most ESBLs and class C enzymes and are deemed to be safer to use than any other last-resort antibiotic. Therefore, the increasing number of reports of β-lactamases able to hydrolyze carbapenems over the last few years is of major concern. Carbapenemases are a heterogeneous group of β-lactamases, including members from classes A, B, and D [178, 179]. Most carbapenemases are able to hydrolyze a very broad spectrum of β-lactams. Several class A enzymes, functional group 2f, with carbapenem-inactivating activity, have been reported over the years. The first, SME-1, was reported in 1990 in a Serratia marcescens isolate from the United Kingdom . Subsequently, GES-1 in K. pneumoniae, SFC-1 in Serratia fonticola, and IBC-1/IMI-1/NMC-A in Enterobacter cloacae have been reported . One of the most recent and widespread class A carbapenemases is the K. pneumoniae carbapenemases (KPCs) . KPC-1 was the first carbapenemase identified for this family of enzymes, reported in 1996 in North Carolina [182, 183]. KPCs, in contrast to the other class A carbapenemases that are chromosomally encoded, are usually on mobile genetic elements and since their discovery have spread to many other organisms including most species of Enterobacteriaceae (including Enterobacter spp., Serratia marcescens, and Salmonella spp.), P. aeruginosa, and several other genera. A recent study that analyzed 147 cases of infections due to carbapenem-resistant K. pneumoniae from 2013 to 2014, in one hospital in Northern Italy, showed that the major resistance determinant was KPC-3 (83.8%). The death rate was an alarming 24.0% in 2013 and 37.5% in 2014 . This is of great concern especially in Southern European countries where CREs expressing KPC are spreading rapidly.
Several enzymes of the class D family (OXA type) are able to degrade carbapenems . The first identified enzyme in this class able to hydrolyze imipenem was OXA-23. It was isolated in the United Kingdom in 1985 from an A. baumannii isolate and was originally characterized as ESBL . Since this initial report, several other OXA enzymes have been described in Enterobacteriaceae, such as OXA-23-like, OXA-40-like, OXA-51-like, OXA-58-like, and OXA-48-like enzymes [185, 187]. The most widespread enzymes in this class are the OXA-48-like enzymes. OXA-48 was first isolated in a patient with UTI in Turkey in 2001 from a strain of K. pneumoniae and is now widely spread across Enterobacteriaceae and other Gram-negative species . In the recent past, OXA-48-like carbapenemases have been responsible for several outbreaks among carbapenem-resistant K. pneumoniae in Spain  and Greece , and currently no broadly active inhibitors of class D enzymes are on the market, in part due to high structural diversity within this class of enzymes .
Metallo-β-lactamases (MBLs) belong to Ambler class B, functional group 3, and all can inactivate carbapenems . MBL hydrolysis of β-lactams is mediated by zinc and inhibited by metal chelators such as ethylenediaminetetraacetic acid (EDTA) but not clavulanic acid or other clinically used β-lactamase inhibitors. MBLs can hydrolyze to varying degrees members of all β-lactam classes except monobactams. Since the early 1990s, when the first MBL, IMP-1, was detected, the number of transmissible genes encoding MBLs has been constantly growing. The major MBLs currently present in the clinic are IMP-like, VIM-like, and NDM-like enzymes. Other MBLs such us the SPM, GIM, SIM, AIM, DIM, FIM, and POM have been reported but are not widely distributed . More than 30 derivatives of IMP-like enzymes have been reported and are commonly found in Japan, China, and Australia causing sporadic outbreaks . In contrast, the VIM-like carbapenemases have been reported from hospitals worldwide. The first VIM-like carbapenemase, VIM-1, was identified in Italy from a P. aeruginosa isolate in 1997 . Currently, over 30 VIM-like carbapenemase have been reported around the world with VIM-2 being the most widespread MBL in P. aeruginosa . VIM-like enzymes are often harbored in gene cassettes and are also associated with integrons . The prevalence of VIM-like enzymes among MBL-producing Enterobacteriaceae isolates in Europe is very high, especially in countries like Italy and Greece. A multicenter European survey showed the presence of VIM-1-like enzymes in 98.9% of Enterobacteriaceae isolates producing MBLs .
The New Delhi metallo-β-lactamase (NDM), first reported in 2009, is the latest carbapenemase described that is threatening the usefulness of β-lactams globally. So far, 13 different variants of NDM have been reported; in several cases, the mutation in the blandm gene seems to predict rates of β-lactam hydrolysis . NDM producers have been isolated across the globe but predominately in the Asian continent and mainly in India. In the United Kingdom and the Middle East, outbreaks of NDM-producing strains have been reported in the recent past. Several reports have shown that NDM-1-producing pathogens are resistant to many other antibiotics, thus limiting options for treating these infections to a small number of agents such as polymyxins, fosfomycin, and aminoglycosides which are rarely used due to efficacy and/or safety concerns . As well, many NDM-1-producing strains also possess ribomethylase, leading to aminoglycoside resistance (see Sect. 126.96.36.199). In 2010 in Canada, a NDM-1-producing K. pneumoniae outbreak was reported, which included a patient with no prior history of traveling to Asian countries that are high risk for these infections. In vitro susceptibility tests showed that the strain was resistant to the majority of available antibiotics except colistin and tigecycline . Several studies summarized in a recent review by Zmarlicka et al. have suggested however that in vitro resistance to several β-lactams by NDM-1-producing organisms does not always translate to clinical outcomes, suggesting that some carbapenem/β-lactamase inhibitor combinations may still work in the clinic . More data and in-depth analysis would be needed to fully understand this.
188.8.131.52 Non-β-Lactamase-Mediated Resistance
Since β-lactam antibiotics cross the Gram-negative OM via porins, mutations causing loss of porins or affecting their structure or expression level can reduce susceptibility by reducing influx. Well-characterized examples of this are loss of OmpK35 and/or OpmK36 in K. pneumoniae [196, 197] and loss of OprD in P. aeruginosa [198, 199]. Loss or downregulation of these may have a significant impact on susceptibility to carbapenems and other classes. Mutations leading to upregulation of RND family efflux pumps can also reduce susceptibility to β-lactams, and in some cases, efflux upregulation occurs concomitant with porin downregulation (see Sect. 4.2.2). These mechanisms cause modest shifts in susceptibility generally but become significant in isolates where β-lactamases are also expressed [174, 200]. Alterations in PBPs (target mutations) had not garnered a lot of attention in Gram-negative pathogens, although it stands to reason that the widespread use of β-lactams/β-lactamase inhibitors is applying selective pressure for the emergence of altered patterns of PBP expression and/or mutations in PBPs of these organisms. Consistent with this, altered expression patterns of PBPs have been reported in pan-β-lactam-resistant clinical isolates of P. aeruginosa, although no changes in the amino acid sequences were found . Amino acid substitutions in PBP2 have been found in E. coli clinical isolates that affect susceptibility to carbapenems . Recently amino acid insertions in PBP3 were identified in clinical E. coli isolates that affect susceptibility to a range of β-lactams including the monobactam aztreonam [202, 203]. The mechanisms of porin loss, efflux, and PBP changes in isolation only shift β-lactam susceptibility modestly, but cumulatively they can have a large impact, especially when combined with the expression of β-lactamases. That PBP3 insertions modestly shift susceptibility to aztreonam is concerning since monobactams are the only class of β-lactams that are intrinsically stable to NDM metallo-β-lactamases, for which no inhibitors are currently available, and PBP3 insertions have been reported in NDM-1 expressing E. coli isolates in certain geographic areas . These strains often express various serine β-lactamases as well. β-lactamase inhibitors such as avibactam can address these β-lactamases, but underlying mutations altering PBPs combined with porin loss and efflux are likely to erode the effectiveness not only of β-lactams but also currently available β-lactam/β-lactamase inhibitor combinations.
184.108.40.206 β-Lactamase Inhibitors
β-Lactamase inhibitors on the market or in clinical development
Marketed (also generic)
Marketed (also generic)
Piperacillin or ceftolozane
Phase III, Pfizer
Phase II, Pfizer
Phase III, Merck
Nacubactam: RG6080, OP0505
Phase I, Roche
New drug application (NDA) (Carbavance)
Phase I, Entasis 
β-Lactam (tazobactam analog)
Cefepime or piperacillin
Phase I, Allecra
Phase I, Wockhardt
The first of these to be identified and brought to the clinic was clavulanic acid, a natural product isolated from Streptomyces clavuligerus , followed by the semisynthetic penicillanic acid sulfone class of inhibitors (sulbactam and tazobactam) [206, 207]. These BLIs all possess the basic core structure of a β-lactam which allows for recognition and binding to β-lactamase. However, key structural differences from β-lactams eliminate most or all intrinsic antibacterial activity against many bacteria and render them mechanism-based “suicide inhibitors” of sensitive β-lactamases . The mechanism of β-lactamase inactivation by these inhibitors is complex, but in general, the active-site serine of the β-lactamase attacks the carbonyl group in the β-lactam ring of clavulanic acid leading to acylation of the β-lactamase. This is then followed by a series of secondary reactions in the enzyme active site that irreversibly inactivate the enzyme [209, 210, 211]. A main difference between the BLI molecules and β-lactams that facilitates this mechanism is that BLIs possess good leaving groups at the C-1 position of their five-membered rings. This allows for secondary ring opening and subsequent β-lactamase enzyme modification. Important factors for BLI efficacy include high acylation and low deacylation rates, which localize them for a longer time period in the enzyme active site and a low number of hydrolytic events (inhibitor molecules hydrolyzed per unit time) necessary for complete enzyme inactivation (termed turnover number or tn). Differences exist in these factors between clavulanic acid, sulbactam, and tazobactam, and these are affected by differences in the active sites among β-lactamases. Clavulanic acid and tazobactam cover most class A β-lactamases including ESBLs . Sulbactam also covers these but is less potent against some enzymes. Tazobactam and sulbactam are better inhibitors of class C carbapenemases than clavulanic acid and notably differ from clavulanic acid in that they do not induce expression of AmpC in bacteria where this enzyme is inducible . However, none of these can cover strains producing metallo-β-lactamases such as NDM-1. Clavulanic acid is partnered in the clinic with amoxicillin or ticarcillin, sulbactam with ampicillin, and tazobactam with either piperacillin or ceftolozane.
Avibactam, the first non-β-lactam BLI approved for clinical use, is a broad inhibitor of class A (including KPCs), class C, and some class D β-lactamases. Avibactam is a member of the diazabicyclooctane (DBO) chemical class , and as such it has a different mechanism of inhibition from previous BLIs. This mechanism is not yet fully understood, but it appears that avibactam functions as a slowly reversible covalent inhibitor with release of intact avibactam for most class A and C β-lactamases . The enzymes appear to be slowly acylated and slowly deacylated, with no or only low-level hydrolysis of the inhibitor molecule. An exception to this was inhibition of KPC-2 which was rapidly acylated but slowly deacylated with hydrolysis of avibactam, so differences do exist. The release of intact avibactam in most cases however is thought to allow for recycling of the inhibitor by β-lactamases in the cell, leading to better inhibitory efficiency than BLIs like clavulanic acid or tazobactam that are hydrolyzed. Avibactam is a substantially more effective inhibitor of key β-lactamases like TEM-1, KPC-2, and AmpC from P. aeruginosa than clavulanic acid, sulbactam, or tazobactam but has limited coverage of class D enzymes, although it does cover OXA-48. Like previous BLIs, avibactam does not cover metallo-β-lactamases like NDM-1. Avibactam was introduced into the clinic very recently (2015), partnered with ceftazidime, and is in clinical trials for combination use with ceftaroline or aztreonam. The latter partnering with the monobactam aztreonam is meant to capitalize on the idea that monobactams are inherently stable to metallo-β-lactamases, and the combination should therefore cover strains expressing metallo-β-lactamases and/or serine β-lactamases. Although ceftazidime/avibactam has only been in clinical use a short while, resistance to this combination due to porin loss and upregulation of β-lactamase expression has been reported . It has recently been suggested that ceftazidime/avibactam could be used in combination with aztreonam for coverage of metallo-β-lactamase/serine-β-lactamase-producing clinical isolates , but appropriate dosing would need to be established.
The success of avibactam has inspired efforts to identify next-generation DBO β-lactamase inhibitors. These include relebactam (MK7655), directed at some class A, including KPCs, and class C β-lactamases (currently in Phase III trials, in combination with imipenem ). Relebactam possesses a narrower spectrum than avibactam since it does not include class D β-lactamases such as OXA-48. Nacubactam (RG6080, OP0595), directed at class A and C β-lactamases, is currently in Phase I trials and is intended for combination with meropenem (Table 4.3). Some DBOs also possess intrinsic antibacterial activity and this warrants some discussion. They inhibit PBP2 in some Gram-negative pathogens, similar to the β-lactam mecillinam . PBP2 inhibition can be synergistic with inhibition of other PBPs (i.e., with other β-lactams), as has been reported for nacubactam [218, 219], and this has been referred to as an “enhancer effect” independent of BLI activity. Compounds that inhibit PBP2, such as mecillinam (amdinocillin) or nacubactam, select for mutants with reduced susceptibility in vitro at high frequency [219, 220]. There is a multiplicity of mutations that engender tolerance of PBP2 inhibition, generally related to the stringent or envelope stress responses , as well as stringent response-independent mechanisms . Such mutations do not affect inhibition of PBP2 specifically; therefore the enhancer effect of DBOs is retained. Some intrinsically active DBOs may be considered as potent stand-alone antibacterials against Gram-negative pathogens, including P. aeruginosa, and this potential may be somewhat overlooked . Interestingly, the gene encoding PBP2 was reported as nonessential in P. aeruginosa based on genetic deletion , suggesting that chemical inhibition of PBP2 is distinct from genetic deletion, possibly due to induction of a futile cell wall pathway cycle as described in , but this is not fully understood. At least one DBO with potent activity against P. aeruginosa demonstrated a high frequency of in vitro resistance selection . It is not known if the intrinsic antibacterial activity of some PBP2-specific DBOs and their potential in vitro resistance profile will affect the clinical outcome when used in combination with a β-lactam. It has been proposed that mutants selected in vitro may not be fit enough to survive in the host , and it is possible that PBP2 inhibition would still provide efficacy against these mutants in vivo. Studies with potent antibacterial DBO molecules in relevant animal models to examine resistance potential in that context are needed to resolve this ongoing discussion.
Another DBO has recently been described (ETX2514 , Entasis) with broader inhibitory activity than other DBOs since it includes class D enzymes. Like nacubactam, ETX2514 has intrinsic antibacterial activity against some Gram-negative pathogens. ETX2514 is being paired with sulbactam (which has intrinsic antibacterial activity against A. baumannii)  for treatment of A. baumannii infections.
Another non-β-lactam BLI class is the boronic acid chemical scaffold. Boronic acid compounds were originally shown to inhibit serine proteases, and this observation was then extended to the serine β-lactamases. These inhibitors form a covalent reversible adduct between the boronate moiety and the catalytic serine of the β-lactamase. The most advanced of these is the cyclic boronate compound RPX7009 (Vaborbactam, The Medicines Company) which was the first of this class for which in vivo efficacy was demonstrated [218, 219]. Vaborbactam is active against class A carbapenemases (including KPCs), as well as other class A and class C β-lactamases [220, 221], but does not inhibit metallo-β-lactamases like NDM-1. A new drug application (NDA) has been filed for the combination of vaborbactam with meropenem for treatment of complicated urinary tract infections (cUTIs). Additional details on the inhibitors described above can be found in recent reviews [139, 210, 222].
The category of β-lactamases that has proven most challenging for the design of broad-spectrum inhibitors is the class B metallo-β-lactamases. To date, no inhibitor of these enzymes has reached the market. An alternative strategy to the design of a metallo-β-lactamase inhibitor is to exploit the fact that the monobactam aztreonam is intrinsically stable to metallo-β-lactamases and can therefore be partnered with avibactam; this combination has the potential to cover strains expressing both metallo-β-lactamase and serine β-lactamases. This combination is currently undergoing Phase III clinical trials. More recently, an innovative approach was undertaken to design novel next-generation monobactams that are not significantly impacted by most serine β-lactamases while retaining their intrinsic stability to the metallo-β-lactamases. One of these, LYS228 , demonstrated excellent potency against MDR Enterobacteriaceae, including CRE [224, 225], and has entered Phase II clinical trials (Novartis). Significant effort has also been devoted to the discovery of therapeutically useful inhibitors of the class B metalloenzymes. This has lagged to some extent since class B enzymes have a different mechanism than serine β-lactamases, and it appears that the design of inhibitors capable of covering multiple clinically important class B enzymes is technically challenging. Inhibitors of class B enzymes would also need to be highly specific and avoid human metalloenzymes to avoid toxicity issues. While the prevalence of class B enzymes remained relatively low in the past and their contribution to worldwide carbapenem resistance was initially considered minimal, this viewpoint has changed in recent years with increased spread of class B enzymes like NDM-1 and their linkage with other resistance determinants. Recent efforts in the search for class B inhibitors include the discovery of the natural product aspergillomarasmine, which is active against NDM-1 and VIM-2 . Novel bisthiazolidine (BTZ) inhibitors of class B enzymes have also recently been described . Of most interest are reports from The Medicines Company on a new series of cyclic boronate compounds derived from RPX7009 (Table 4.3) with broad-spectrum carbapenemase activity including metallo-β-lactamases, in preclinical development .
220.127.116.11 Resistance to β-Lactamase Inhibitors
The implementation of β-lactamase inhibitors extended the clinical usefulness of several β-lactam antibiotics for decades; however the emergence of variant β-lactamases and other mechanisms has eroded their usefulness. One factor that impacts the effectiveness of BLIs is the expression level of β-lactamases and/or the number of β-lactamases being expressed in a given isolate. Even if a BLI is potent inhibitor of serine β-lactamases, this can be overwhelmed by high-level expression of one or multiple β-lactamase enzymes . Overexpression can result from the β-lactamase gene residing on multicopy plasmids or via mutations in the promoter region causing high-level expression . In particular, high-level expression of TEM-1 was an early mechanism identified that reduced susceptibility to amoxicillin-clavulanate [231, 232]. This can also be related to the induction of β-lactamase by certain BLIs. Clavulanic acid induces the expression of AmpC β-lactamase in bacteria where this enzyme is inducible. Since clavulanic acid is not a good inhibitor of class C enzymes, this induction can be antagonistic toward the partner antibiotic . This is an issue in particular for the case of chromosomal inducible ampC (P. aeruginosa) or plasmid-borne inducible ampC such as DHA-1 in K. pneumoniae. In fact, antagonism by clavulanic acid is used as a diagnostic for the presence of inducible ampC [234, 235]. Sulbactam and tazobactam do not have this induction effect and as such can be better options (e.g., tazobactam paired with piperacillin against P. aeruginosa). Active efflux of BLIs [20, 236] or changes in influx, possibly due to porin loss, may also serve to reduce their concentration relative to the β-lactamases in the cells, decreasing their effectiveness, although influx of compounds such as avibactam is not currently well understood . Defects in porins OmpK35 and/or OpmK36 were, however, associated with decreased effectiveness of imipenem/relebactam and meropenem/RPX7009 in clinical isolates isolated from hospitals in New York [219, 238]. Efflux was implicated as an important mediator of resistance to ceftazidime/avibactam in P. aeruginosa, but this remains to be further explored . Porin mutations combined with upregulated expression of plasmid-borne KPC-3 have been associated with clinical resistance to ceftazidime/avibactam in K. pneumoniae , and porin mutations were associated with ceftaroline/avibactam resistance in E. cloacae mutants selected in vitro .
Resistance to BLIs, mainly clavulanic acid, also resulted from the emergence of new β-lactamases that are resistant to the BLI. Very soon after the introduction of clavulanic acid into clinical use, such variants began to emerge, with the first reported being variants of the class A TEM enzyme that were resistant to clavulanic acid, found in E. coli clinical isolates [241, 242]. These TEM variants were altered at their Arg244 residues to either Cys or Ser . This position had been shown earlier to be important for clavulanic acid inhibitory function, and so this clinical outcome might have been expected. These variants were initially designated inhibitor-resistant TEM (IRT-1 and IRT-2), but since then, 37 clavulanic acid-resistant TEM variants have been identified (cataloged at http://www.lahey.org/Studies/temtable.asp, functional group br), and the convention now is that these all have TEM numerical designations. These variants are found mainly in E. coli isolates but also occur in Klebsiella , Proteus , Shigella , and Citrobacter . Inhibitor resistance can also be combined with amino acid substitutions conferring β-lactamase activity against oxyimino-β-lactams (ESBL ), and these are referred to as complex mutant TEMs (CMT). Currently 11 of these variants have been described (http://www.lahey.org/Studies/temtable.asp, functional group ber). As well, seven inhibitor-resistant variants of the class A enzyme SHV have also been described (http://www.lahey.org/Studies/), with the most recent being SHV-107 found in a K. pneumoniae clinical isolate . It should be noted that inhibitor-resistant β-lactamases generally refer to clavulanic acid, and these can also be resistant to sulbactam, but generally they remain susceptible to tazobactam [249, 250, 251]. Therefore these enzymes mainly affect amoxicillin/clavulanate, ticarcillin/clavulanate, or ampicillin/clavulanate but not piperacillin/tazobactam. However, in 2010, the emerging class A ESBL KPC-2 carbapenemase  was shown to also be resistant to clavulanic acid, sulbactam, and tazobactam, raising serious concerns .
Although ceftazidime/avibactam is active against KPC producers, providing effective treatments in the near term, the emergence of β-lactamase variants resistant to avibactam has already been reported . In vitro selection studies demonstrated that variants of KPC-3  or certain inhibitor-resistant SHV enzymes (particularly the S130G variant)  were less susceptible to ceftazidime/avibactam. In the former study, the most frequently isolated variant of KPC-3 was Asp179Tyr, and the authors speculate this particular change may increase ceftazidime specificity rather than mediating resistance to avibactam per se. Interestingly, many of the alterations also appeared to impair the ability of the β-lactamase to hydrolyze carbapenems (reversal of ESBL ), thereby increasing susceptibility of the bacteria to those agents. Consistent with these in vitro studies, plasmid-borne variants of KPC-3 have been described with reduced susceptibility to avibactam in K. pneumoniae clinical isolates . These emerged within 10–19 days of ceftazidime/avibactam exposure. The KPC-3 variants found had either a D179Y/T243 M double substitution, D179Y single substitution, or V240G single substitution. Interestingly, these mutations also seemed to decrease the KPC-3 carbapenemase activity enough in some isolates to render them susceptible, and it has been suggested that agents like meropenem could be used to ameliorate to some extent the impact of such mutations. A large-scale analysis of the binding pockets of class A serine β-lactamases indicated that most would be susceptible to avibactam but some outliers were identified. In particular, PER-4 was shown to be highly resistant to avibactam , indicating the preexistence of class A β-lactamase variants in the clinic that are resistant to avibactam.
Since there is significant amino acid sequence homology between the GyrA and ParC and GyrB and ParE proteins, individual quinolone molecules can inhibit the activities of both enzymes, and most quinolones will inhibit both targets to varying degrees. Either topoisomerase can constitute the primary or secondary target of quinolones in different bacteria. Given their broad-spectrum and excellent tissue penetration, fluoroquinolones are well suited to empiric therapy and became one of the most broadly used classes of antibiotic. Quinolones are also used fairly extensively in agriculture . Given their widespread use, it is not surprising that resistance has emerged at a significant rate around the world. Resistance to fluoroquinolones in Gram-negative pathogens is mediated by several mechanisms, the most common being chromosomal mutations that alter the quinolone binding sites of the GyrA/B and ParC/E proteins. Additional mechanisms include chromosomal mutations that upregulate RND-mediated efflux or decrease compound penetration. Plasmid-based mechanisms also occur, including efflux, target protection, and compound modification. Each of these is discussed below, and the epidemiology of fluoroquinolone resistance is discussed in Chap. 10.
18.104.22.168 Target Mutations Conferring Quinolone Resistance
Summary of the impact of different resistance mechanisms on ciprofloxacin susceptibility of E. coli and P. aeruginosa
Organism and resistance mechanism
Fold change in ciprofloxacin MIC
gyrA + parC
gyrA + parC
gyrA + parC + efflux upregulation
22.214.171.124 Efflux and Reduced Compound Influx
The role of Gram-negative RND family efflux pumps in intrinsic and mutationally acquired resistance to antibiotics is covered in detail in Sect. 4.2.1 and in recent reviews . The potent broad-spectrum activity of fluoroquinolones against even intrinsically resistant Gram-negative bacteria such as P. aeruginosa indicates that RND efflux does not mediate enough intrinsic resistance to some fluoroquinolones to limit their spectrum. This may relate to some extent to the hydrophilicity  as well as to the overall target potency and cidality of fluoroquinolones. However, fluoroquinolones are substrates of a wide range of RND family pumps (Table 4.1) including the AcrAB-TolC pump of E. coli and Salmonella spp., the AcrEF pumps of E. coli and Salmonella enterica, the CmeABC pump of Campylobacter jejuni, and MexAB-OprM, MexXY-OprM, MexCD-OprJ, and MexEF-OprN of P. aeruginosa . Mutations in regulatory genes causing pump overexpression and decreased susceptibility can be readily selected by in vitro exposure to quinolones. However, in most cases overexpression of efflux pumps alone, in the absence of other mechanisms, affords only modest reductions in fluoroquinolone susceptibility (Table 4.4). Efflux pump overexpressing mutants are routinely found among clinical isolates . Upregulation of the AcrAB-TolC pump in fluoroquinolone-resistant E. coli clinical isolates contributed to high-level fluoroquinolone resistance along with QRDR mutations [67, 277]. Similarly, AcrAB-TolC upregulation played a role in fluoroquinolone resistance in K. pneumoniae and K. oxytoca clinical isolates . A P. aeruginosa clinical isolate with a mutation in gyrB and upregulated for MexAB-OprM emerged during ciprofloxacin monotherapy . Since RND pumps have broad substrate ranges, selection of pump overexpression during previous treatment with various antibiotics will result in selection of pump upregulation which will affect fluoroquinolones and vice versa. RND family efflux pumps function in concert with the OM permeability barrier, and therefore any reduction in a compound’s ability to cross the OM will have a corresponding enhancing effect on efflux-mediated resistance. Fluoroquinolones cross the outer membrane either through water-filled porin channels or by diffusion through lipid domains in the outer membrane depending on the hydrophobicity of the quinolone . Reduced porin levels have also been associated with fluoroquinolone resistance in E. coli [277, 281] and S. enterica  clinical isolates. Reduced porin levels often occur concomitantly with upregulation of efflux pumps  potentially linking reduced influx and increased efflux via a single mutation .
126.96.36.199 Plasmid-Mediated Quinolone Resistance (PMQR ): Topoisomerase Protection, Quinolone Modification, and Efflux
Three different plasmid-borne quinolone resistance mechanisms have been described. These are topoisomerase protection (Qnr), quinolone modification, and efflux. Each is discussed below. The plasmid-borne quinolone resistance determinant qnr (now termed qnrA1) was first identified from a quinolone-resistant K. pneumoniae clinical isolate in 1998 . Other qnr determinants were subsequently identified including qnrB  and qnrS , and over the years, the number has expanded to where there are currently seven families of Qnr proteins, identified from a range of organisms: QnrA, QnrB, QnrC, QnrD, QnrS, and QnrVC (cataloged at http://www.lahey.org/qnrStudies). The Qnr proteins are typified by having tandem repeats of a pentapeptide consensus sequence and as such are referred to as pentapeptide repeat proteins . The mechanism of Qnr proteins is referred to as topoisomerase protection and involves binding of the Qnr protein to gyrase and topoisomerase subunits and the holoenzymes [288, 289, 290]. Binding is not dependent on DNA or ATP and likely occurs prior to establishment of the ternary complex, reducing quinolone interaction with the topoisomerases. More recent structural information indicated that Qnr can assume a rodlike structure resembling B-form DNA, suggesting it may compete with quinolones by binding in the gyrase QRDR or DNA-gate region . Although Qnr proteins can bind to a number of subunits in vitro, they appear to mediate resistance to quinolones or other agents that bind the QRDR region of GyrA, but not to agents that target the ATPase function (e.g., GyrB) . Qnr proteins, specifically, only cause a marginal shift in quinolone susceptibility similar to that of single-target mutations  (Table 4.4). There are wide dissemination of qnr plasmids in Enterobacteriaceae clinical isolates around the world [293, 294, 295, 296] and significant diversity of plasmids that carry these genes (reviewed in ). In contrast they seem to be rare among non-fermenters such as P. aeruginosa and A. baumannii. Interestingly, qnr genes certainly predate the introduction of the synthetic quinolones into clinical use and are also found on the chromosomes of several bacteria . It has been suggested that mobilization from the genome to small transmissible plasmids may have originated in Proteeae . A final concern is that qnr genes typically reside on a range of plasmids that also encode other resistance markers, in particular extended-spectrum β-lactamases, such as SHVs and CTX-Ms, and AmpC-like enzymes such as DHA-1 (reviewed in ).
The second identified plasmid-borne resistance determinant was a bifunctional variant of the aminoglycoside-modifying enzyme aac(6′)-1b (see Sect. 188.8.131.52 on aminoglycoside-modifying enzymes) . This variant, designated aac(6′)-1b-cr, differs from aac(6′)-1b by encoding two amino acid substitutions, Trp102Arg and Asp179Tyr. These differences allow the enzyme to bind and acetylate fluoroquinolones that have an amino nitrogen on the piperazinyl ring (e.g., ciprofloxacin and norfloxacin), thereby reducing their activity. Fluoroquinolones that have modifications on the piperazinyl structure (e.g., levofloxacin or moxifloxacin) are not affected. Importantly, this variant enzyme retains its aminoglycoside-modifying activity, thus creating a single protein that can affect two different classes of antibiotic. Like qnr, the aac(6′)-1b-cr gene is usually found in a cassette as part of an integron in multiresistance plasmids that encode β-lactamases or qnr, is disseminated worldwide, and can also be found on the chromosome of some bacteria (summarized in ).
The most recently identified class of plasmid-borne resistance determinants are genes encoding fluoroquinolone efflux pumps. The first of these was oqxAB, identified in E. coli isolates of agricultural origin and which encodes an RND family efflux pump with a broad substrate range [299, 300, 301]. This was later found in a range of human, animal, and environmental isolates [302, 303, 304, 305]. Intriguingly, the oqxAB pump genes are found on the chromosome in K. pneumoniae, including drug-resistant human clinical isolates , and this appears to be the likely reservoir/origin of the plasmid-borne version . Typical of chromosomally encoded RND family pumps, upregulation of OqxAB expression in K. pneumoniae requires mutations in the oqxR regulatory gene ; however expression from plasmid-borne oqxAB genes is constitutive, and therefore, this was the first report of a constitutively expressed, mobile plasmid-borne efflux pump . The second was qepA, identified in 2008 an E. coli clinical isolate and which encodes a member of the major facilitator efflux pump superfamily . Another variant qepA2 has also been described . Of concern, qepA genes often reside on mobile elements with genes encoding ribosomal methyltransferases which mediate resistance to aminoglycosides , again linking fluoroquinolone resistance with resistance to other antibiotic classes.
184.108.40.206 Interplay of Resistance Mechanisms
Efflux and possibly lowered porin levels reduce susceptibility to fluoroquinolones in Gram-negative clinical isolates. Efflux upregulation in isolation may however cause only modest shifts in susceptibility. In cases of higher-level resistance, mutations in the QRDR regions are also found along with upregulation of efflux. In many Gram-negative pathogens, such as E. coli, target-based resistance progresses from single mutations (e.g., encoding alteration at Ser83 of GyrA), which cause only small shifts in susceptibility, to accumulation of multiple target mutations leading to high-level resistance. The accumulation of mutations depends on stepwise enrichment of mutants, in turn depending in part on the level of quinolone being within the mutant selection window, defined as being between the concentration required to block the growth of 99% of bacteria in culture (MIC99) and the MIC of the least susceptible next step mutant, (termed the mutant prevention concentration (MPC) . Since high-level resistance requires two or more target mutations, fluoroquinolone levels higher than the MPC can only select the simultaneous double mutants from a wild-type background at an extremely low frequency (approximately 10−12). The relatively rapid emergence of target-based high-level resistance to fluoroquinolones in the clinic in organisms such as P. aeruginosa and E. coli and a general association with efflux suggest that a key role for efflux may be in enhancing survival of first step target mutants which then rapidly accumulate additional mutations conferring stable high-level resistance. Factors such as suboptimal exposure to drug can contribute to enrichment of earlier stage mutants, and in cases where certain target mutations confer clinical resistance levels only in conjunction with efflux, the presence of pumps would be very important to this process. A hollow fiber model used to simulate human drug treatment with E. coli lends support to this notion in that mutants with a two- to eightfold shift in susceptibility due to upregulation of AcrAB-TolC emerged first, followed by emergence of target mutations (single or double) . The emergence of target mutations was also strongly delayed in a strain lacking AcrAB-TolC function. Efflux upregulation regressed after the emergence of target mutations suggesting that once target-based resistance was established, efflux upregulation may no longer be required and may revert back to wild-type expression.
In the case of P. aeruginosa, efflux has been shown to provide a significant contribution to establishing the intrinsic susceptibility to fluoroquinolones, which correspondingly enhances the ultimate levels of resistance caused by target QRDR mutations . Upregulation of various pumps including MexAB-OprM, MexCD-OprJ, or MexEF-OprN can also provide substantial shifts in fluoroquinolone susceptibility without target mutations [314, 315]. The combination of target mutations and efflux in P. aeruginosa can mediate very high-level resistance . Pump upregulation can occur at very high frequencies at lower compound levels since in many cases, this requires only loss-of-function mutations in regulatory genes like mexR (MexAB-OprM) or nfxB (MexCD-OprJ). Selection of resistance in P. aeruginosa in vitro at 4X MIC of levofloxacin occurred at 10−6–10−7, whereas the frequency at 4X MIC for an efflux-defective strain was <10−11, indicating that selection of resistance in the absence of efflux even at relatively modest multiples of the MIC could be rare . This suggests overall that efflux was even more of a factor in the emergence of target-based resistance in P. aeruginosa, consistent with the very rapid rise in fluoroquinolone resistance seen in P. aeruginosa in the United States after widespread fluoroquinolone use began, and the association of this with resistance to multiple antibiotics [316, 317]. The use of an efflux pump inhibitor to assess the prevalence of pump-mediated fluoroquinolone and multidrug resistance among P. aeruginosa clinical isolates also suggested a correlation between fluoroquinolone treatment and the co-emergence of target and pump-mediated multidrug resistance . More recently, associations were seen in clinical isolates between target mutations and efflux pump upregulation; however the expression levels of several pumps could not be correlated with higher-level resistance seen for some isolates that also harbored QRDR mutations, suggesting that other as yet unidentified factors may mediate higher-level resistance in some QRDR mutants . Similar interplay between plasmid-borne resistance mechanisms and other mechanisms is likely also occurring, both in terms of determining susceptibility and in facilitating the emergence of high-level resistance. Qnr proteins only cause a marginal shift in quinolone susceptibility  (Table 4.4), but this is additive with target-based or other mechanisms and will contribute to the emergence of higher-level clinically relevant resistance . Like qnr, the level of resistance conferred by aac(6′)-1b-cr alone was modest; however, more significant levels of resistance were observed when aac(6′)-1b-cr and qnrA were found together (Table 4.4). Furthermore, the presence of plasmid-borne aac(6′)-1b-cr in E. coli resulted in a greater recovery of resistant mutants during selection experiments with ciprofloxacin , essentially by widening the mutant selection window. This again highlights the interplay of determinants such as aac(6′)-1-cr and qnrA in the stepwise acquisition of clinically significant resistance . An additional factor that may have contributed to the emergence of high-level fluoroquinolone resistance is that DNA damage and interference with DNA replication caused by quinolones induce the SOS response, leading to upregulation of error-prone DNA polymerases. Evolution of quinolone resistance in E. coli in vitro and in an animal model of infection was curtailed in mutants lacking the SOS response .
220.127.116.11 New Strategies for Targeting Type II Topoisomerases
Delafloxacin (Melinta Therapeutics) [320, 321, 322] is a new structurally unique anionic fluoroquinolone that was recently approved by FDA for the treatment of acute bacterial skin and skin structure infections (ABSSSI ) and is in clinical trials for community-acquired pneumonia and complicated urinary tract infection. Delafloxacin is particularly potent against Gram-positive pathogens but is also active against several Gram-negative pathogens including H. influenzae, Enterobacteriaceae spp., and P. aeruginosa. Delafloxacin targets DNA gyrase and topoisomerase IV equally, which may reduce the emergence of resistance. Other efforts to discover novel agents that circumvent target-based or other fluoroquinolone resistance mechanisms in Gram-negative pathogens include the design of compounds referred to as novel bacterial type II topoisomerase inhibitors (NBTIs) that engage the GyrA/ParC targets via a mode of inhibition distinct from fluoroquinolones and that are not affected by QRDR mutations . These compounds have activity against Gram-negative pathogens including E. coli and P. aeruginosa. These NBTIs also seem to benefit from balanced inhibition of both gyrase and topoisomerase IV targets, thereby requiring at least two target mutations in E. coli in order to observe decreased susceptibility . Efforts to design novel inhibitors of the ATPase function of type II topoisomerases, in order to exploit DNA replication as a target but avoid QRDR-mediated resistance, have resulted in potent antibacterial compounds [324, 325]. Similarly, novel spiropyrimidinetrione agents with a mode of action distinct from fluoroquinolones that may involve targeting GyrB (AZD0914, now ETX0914, Entasis Therapeutics) are in clinical trials and may find utility in treating infections due to Gram-positive and/or fastidious Gram-negative pathogens, such as N. gonorrhoeae [326, 327, 328]. Another class of novel inhibitors of GyrB, which bind to the TOPRIM domain and are not affected by fluoroquinolone resistance mutations, has recently been described . A detailed discussion of non-quinolone inhibitors of topoisomerases is presented in Chap. 19.
Examples of tetracycline -specific resistance determinants found in selected Gram-negative pathogens
tet(A) tet(B) tet(G) tet(H), tet(L), tet(39), tet(Y)
tet(M) tet (O) tet(W)
tet(M) tet (S) tet(W)
tet(A-D), tet(G), tet(L), tet(39)
tet(A-E), tet(G) tet(J) tet(L), tet(Y),
Tetracycline-Specific MFS Family Efflux Pumps
There are many tetracycline-specific efflux pumps described for both Gram-positive and Gram-negative bacteria , and they function by actively extruding tetracycline from the cell and preventing accumulation to a level sufficient to fully inhibit the ribosome. Examples of pumps that are commonly found in Gram-negative pathogens are listed in Table 4.5, and an updated table of the distribution of these genes among Gram-negatives is maintained at https://faculty.washington.edu/marilynr/tetweb1.pdf and http://faculty.washington.edu/marilynr/tetweb2.pdf. The TetA efflux pump is perhaps the most broadly distributed among important Gram-negative pathogens, and indeed plasmid-encoded TetA was the first bacterial antibiotic efflux pump identified in 1980 [341, 342]. Generally, tetracycline-specific efflux pumps genes reside on mobile genetic elements and are thus horizontally acquired resistance mechanisms. The genes encoding TetA and TetB efflux pumps were later identified in natural oxytetracycline-producing Streptomyces as mechanisms protecting the producer organisms from the effects of the tetracyclines they were producing  suggesting this is the likely original source of this resistance mechanism. Genes such as tetA are commonly used as antibiotic selection markers for genetic engineering in bacteria, highlighting the effectiveness with which they can confer resistance to first-generation tetracyclines. The genes encoding tetracycline-specific pumps are usually accompanied by the tetR gene, which encodes a tetracycline-responsive repressor that controls expression of the TetA efflux pump [344, 345, 346, 347]. Unbound TetR functions as a repressor of tetA expression by binding to tandem operator sequences upstream of tetA as a homodimer and blocking expression . Upon binding tetracycline, TetR dissociates from the DNA, allowing transcription to occur.
The tetracycline-specific transporters, typified by TetA, are located in the bacterial inner (cytoplasmic) membrane, and those found in Gram-negative bacteria are about 46 kDa in size and have 12 transmembrane spanning regions . They belong to the major facilitator superfamily (MFS) and are tetracycline-H+ antiporters that operate through the exchange of a proton for the tetracycline molecule which drives transport of the tetracycline against a chemical concentration gradient, in this case from the cytoplasm across the inner membrane into the periplasmic space between the inner membrane and OM. Single-component pumps like TetA, located in the inner membrane, are generally thought to be more effective at extruding compounds from the cytoplasm than are pumps of the more broadly active RND family . The latter are generally thought to recognize compounds in the periplasm or when diffusing into the inner membrane. This is important, since in Gram-negative bacteria, the single-component pumps cannot extrude antibiotics completely out of the cell into the surrounding milieu but will deposit the compound into the periplasmic space between the inner and outer membranes. This can concentrate the tetracycline in the periplasm, from which it could diffuse back across the inner membrane into the cell in the absence of additional efflux across the OM. Consistent with this, higher levels of resistance mediated by pumps like TetA often require interplay with efflux across the OM by RND family pumps such as MexAB-OprM in P. aeruginosa or AcrAB-TolC in E. coli [6, 24]. These RND family pumps have a broad substrate specificity which includes tetracyclines, and the combined effect of specific single pump efflux from the cytosol and subsequent RND-mediated efflux from the periplasm to the outside of the bacterium can lead to high levels of resistance in some Gram-negative pathogens.
Although the TetA MFS family efflux pumps are currently widespread among clinical isolates, presumably driven by the extensive use of earlier-generation tetracyclines, there have been significant advancements in circumventing the impact of these pumps with each subsequent generation of tetracyclines. Understanding the emergence of clinical resistance to early generations of tetracyclines, along with improved understanding of tetracycline mechanism of action, was a key driving force for renewed interest in developing new tetracyclines that were not subject to this mechanism. Correspondingly, efforts leading to the identification of tigecycline were specifically directed toward achieving cellular activity against tetracycline-resistant bacteria [334, 349, 350, 351], including those expressing tetracycline efflux pumps. That this was achieved with tigecycline is shown by the potent antibacterial activity in broad susceptibility testing with resistant clinical isolates that supported clinical development as an “expanded-spectrum” antibiotic for treatment of multidrug-resistant Gram-negative infections (excluding P. aeruginosa) [349, 352, 353, 354, 355, 356, 357, 358, 359]. Specifically showing that tigecycline circumvents resistance mediated by these pumps, expression of Tet(A), Tet(B), or Tet(X) in a susceptible E. coli strain background conferred very high levels of resistance to tetracyclines (MIC ≥ 128 μg/mL) but had a much smaller or no impact on the third-generation tigecycline (or eravacycline (TP-434)), depending on the pump . Furthermore, no correlation was seen between the presence of tetracycline-specific efflux genes tet(A) to tet(E) and insusceptibility to tigecycline in strains of Enterobacteriaceae . It should be noted though that some variation in amino acid residues important for recognition of tetracyclines has been reported for TetA proteins expressed from tet(A) genes residing on different genetic elements and this can have a modest effect on how much susceptibility is shifted when the pump is expressed [360, 361]. Whether this portends the selection of mutations in tetracycline pump genes over time that increase recognition of tigecycline remains to be seen.
Efflux Mediated by RND Family Efflux Pumps
Tetracyclines are substrates of several RND family pumps  including AcrAB-TolC in E. coli and K. pneumoniae and the MexAB-OprM, MexXY-OprM, and MexCD-OprJ pumps in P. aeruginosa. RND efflux pumps are therefore important for determining the Gram-negative spectrum of these compounds. Second-generation compounds such as doxycycline and minocycline possess a broader antibacterial spectrum than tetracycline, most notably against Acinetobacter, Burkholderia, and Stenotrophomonas, but their activities against P. aeruginosa and most species of Enterobacteriaceae are still limited, in large part due to RND-mediated efflux . Tigecycline has an expanded spectrum, covering a range of Gram-negative pathogens including Enterobacteriaceae and non-fermenters such as Acinetobacter, Stenotrophomonas, and Burkholderia . Therefore basal-level RND-mediated efflux alone does not exclude these organisms from the spectrum of tigecycline. However, RND-mediated efflux is a factor excluding P. aeruginosa (MexAB-OprM, MexXY-OprM)  and Proteus mirabilis (AcrAB)  from the spectrum of tigecycline. Since tigecycline is a substrate of AcrAB-TolC present in many Gram-negative pathogens within the spectrum of tigecycline, RND-mediated efflux also posed a threat as a resistance mechanism, either via mutational upregulation of pump expression or indirectly by exacerbating other as yet unknown mechanisms. Supporting this, RND-mediated efflux has been implicated as a determinant of resistance in laboratory and clinical isolates of Morganella morganii , K. pneumoniae [42, 365, 366], E. coli , Enterobacter cloacae /E. aerogenes , and Salmonella enterica . As well, another RND family pump, OqxAB, may play a role in decreasing susceptibility to tigecycline in K. pneumoniae , and the AdeABC [61, 370, 371], AdeFGH [372, 373], and AdeIJK  efflux pumps have been associated with decreased susceptibility to tigecycline in A. baumannii.
18.104.22.168 Ribosomal Protection , Target Mutations, and Tetracycline-Modifying Enzymes
Resistance to some tetracyclines (first and second generation) can be caused by the action of ribosomal protection proteins (RPPs). Several of these proteins have been identified (e.g., Tet(B), Tet(O), Tet(M), Tet (S), Tet(Q), Tet(W))  (Table 4.5), and a comprehensive update on the distribution of these determinants in Gram-negative bacteria is maintained at http://faculty.washington.edu/marilynr/tetweb2.pdf. The best studied of these proteins are Tet(O) and Tet(M) . Plasmid-borne tet(O) was first identified in a Campylobacter jejuni clinical isolate  and later in Campylobacter coli. A similar gene, designated otr(A), was identified in the tetracycline-producing organism Streptomyces rimosus , suggesting that, like tetracycline-specific efflux, ribosomal protection likely originated as a mechanism to protect the tetracycline-producing organisms and has spread on mobile genetic elements. Genes encoding Tet(M) and Tet(Q) also occur on mobile elements . Ribosomal protection proteins are generally conserved GTPases that resemble the elongation factor EF-G (and to a lesser extent EF-Tu) [378, 379]. Early studies based on chemical probing and cryoelectron microscopy indicated that RPPs bind a similar site on the 50S ribosomal subunit as EF-G and subsequently cause the tetracycline to be released from the ribosome [375, 380, 381, 382]. It is thought that hydrolysis of GTP is not strictly necessary for causing the release of tetracycline from the ribosome but is required for RPP dissociation from the ribosome. Although the RPP-binding site is removed from the tetracycline-binding site on the 30S ribosomal mRNA, it was originally hypothesized that RPP binding caused an overall conformational shift in the ribosome sufficient to dislodge bound tetracycline  and stimulate binding of tRNA to the A-site which also reduced rebinding of tetracycline . More recent cryoelectron microscopy and modeling of Tet(O) and Tet(M) bound to the 70S ribosome suggest that bound RPPs may also intrude directly into the binding site of tetracyclines located around residue C1054 of the 16SrRNA [383, 384]. Ribosomal protection confers resistance primarily to tetracycline, doxycycline, and minocycline, but as is the case for tetracycline-specific efflux, ribosomal protection has been circumvented by the third-generation tetracyclines including tigecycline, omadacycline, and eravacycline [360, 385, 386]. Target-based (active-site) resistance to tetracyclines is relatively rare but does occur. Mutations in the rRNA target were initially reported in the Gram-positive Propionibacterium acnes (G1058C)  and later in Helicobacter pylori [388, 389]. Resistance to tetracycline in Neisseria gonorrhoeae can be mediated by a mutation in rpsJ, encoding Val57Met substitution in the 30S ribosomal protein S10 . This mechanism was later found in K. pneumoniae KPC-2-producing clinical isolates and associated with reduced susceptibility to tigecycline [391, 392].The first identified tetracycline-modifying enzyme, TetX, was encoded on transposons isolated from Bacteroides fragilis , and several more have been identified since then (see Table 4.5). These enzymes are monooxygenases that act by hydroxylating the tetracycline, interfering with the tetracycline magnesium-chelating properties which are needed for ribosome binding [393, 394]. The hydroxylated tetracycline is also less stable and can then decompose. These enzymes interact with the central core of the tetracycline molecule, explaining why they can act on all tetracyclines including third-generation compounds [395, 396]. Nonetheless they appear to be less effective in conferring resistance to third-generation tetracyclines . These enzymes have not emerged or spread as a major source of resistance yet, particularly to tigecycline, but should be monitored in the clinic.
22.214.171.124 Evasion of Tetracycline-Specific Resistance by Third-Generation Glycylcyclines (Tigecycline)
Third-generation tetracyclines were developed in direct response to the emergence and spread of resistance , and this effort led to the synthesis of a novel class of glycyl-substituted C9-aminotetracyclines that are referred to as glycylcyclines . One of these, GAR-936, now tigecycline, bears a t-butyl amine substitution and is very potent against a broader range of Gram-positive and Gram-negative bacteria than tetracycline. Moreover, it evades the two main categories of acquired resistance to tetracycline, tetracycline-specific efflux and ribosomal protection. This is partly because tigecycline binds the ribosome with a much higher affinity (10–100-fold higher) than does tetracycline, and this is reflected in more potent inhibition of translation as measured using in vitro translation assays [397, 398, 399]. The basis for the improved affinity of tigecycline is an additional stacking interaction between the 9-t-butylglycylamido portion of tigecycline (C-9 moiety) and the C1054 nucleobase of the 16S rRNA . There is also additional steric clash between tigecycline and the anticodon stem loop of the A-site tRNA compared to tetracycline, making it more effective in preventing tRNA entry into the A-site. This is likely important for the overall improved potency against a broader range of Gram-negatives than tetracycline (i.e., overcoming intrinsic resistance). The C-9 moiety also may enhance the target on-rate of tigecycline and sterically clash with important residues of the RPP TetM (within loop 3 of domain 4 of TetM) that interact with C1054, thus preventing TetM from dislodging tigecycline from the ribosome. Therefore the C-9 moiety itself is likely preventing TetM and other RPPs from conferring resistance rather than this being a function solely of higher tigecycline binding affinity . As mentioned above, tigecycline also appears to escape recognition by tetracycline-specific efflux pumps, shown using TetB containing vesicles . Tigecycline may also be less effective as an inducer of tetracycline efflux pump expression .
126.96.36.199 Mechanisms of Tigecycline Resistance Emerging in the Clinic
Resistance to early-generation tetracyclines in the clinic emerged rapidly after their introduction in the late 1940s, and the epidemiology of tetracycline resistance is described in Chap. 10. As described above, third-generation tetracyclines, exemplified by tigecycline, are able to largely overcome established resistance by circumventing tetracycline efflux and/or ribosomal protection. Tigecycline entered the clinic in 2005 (often used as a last line of defense in treating MDR isolates), and since it was refractory to the main resistance mechanisms, it was not clear what mechanisms of resistance would emerge in clinical use, although it seemed likely that RND efflux would play a role. The Tigecycline Evaluation and Surveillance Trial (TEST ) is an ongoing global study to monitor in vitro susceptibility to tigecycline and other antibiotics in MDR isolates. The most recent report  examined isolates collected worldwide between 2004 and 2014 and found that rates of MDR E. coli ranged from 4% in North America to 18% in Latin America. Approximately 94% of MDR E. coli were resistant to minocycline, but only 0.2% were resistant to tigecycline, the lowest rate for all antibiotics tested. Tigecycline-resistant E. coli isolates did not appear in this study until 2008; however eight resistant isolates have been identified between 2009 and 2014 across a wide geographic range that included one isolate from North America. For K. pneumoniae, rates of MDR were approximately 12%. Among those, the rate of tigecycline resistance was higher than seen for E. coli, at approximately 6%, although this was still the lowest of all the antibiotics tested. A. baumannii had the highest frequency of MDR, at 44% with some geographic areas having >50% MDR. The lowest rate of resistance among the MDR isolates was reported for minocycline (13%). Tigecycline resistance breakpoints are not established for A. baumannii, but tigecycline had a lower MIC90 than minocycline. The most recent report  concluded that tigecycline has remained active against most MDR isolates (excluding P. aeruginosa which has high intrinsic resistance), although this varies by geographical region. The acquisition of tigecycline resistance in small numbers of E. coli isolates over the course of the collection of these strains was observed and should be further monitored.
So far, decreased susceptibility to tigecycline in the clinic has been attributed to upregulation of RND family efflux pumps. For example, resistance has been correlated with upregulation of AcrAB-TolC expression in clinical isolates of E. coli [43, 402], K. pneumoniae [42, 365, 403], and E. cloacae  and of AdeABC in A. baumannii [44, 370]. Interestingly the tetX gene encoding enzymatic modification was detected in a clinical isolate of A. baumannii from China . Decreased susceptibility in Salmonella enterica was attributed to the combined activity of a plasmid-borne tet(A) gene and mutation in ramR, presumably leading to RND efflux pump upregulation. There is not always a direct correlation between efflux and susceptibility however, and mechanisms of resistance to tigecycline may ultimately prove to be more complex as tigecycline is used longer in the clinic. For example, a recent study of tigecycline-resistant A. baumannii found involvement of AdeABC efflux but also uncovered a potential role for mutational disruption in the trm methyltransferase gene in resistance in clinical isolates . Finally, in vitro tigecycline selection studies using strains harboring the well-characterized tetracycline resistance genes tet(A), tet(K), tet(M), and tet(X) selected for mutations in these genes that increased the ability of the encoded proteins to act on tigecycline . Since these genes are widespread in clinical isolates, it will be of interest to see if this occurs in the clinic going forward.
188.8.131.52 Novel Agents and New Approaches: Circumvention of Tetracycline-Specific Resistance Mechanisms in Third-Generation Tetracyclines (Glycylcyclines )
Current efforts in the search for next-generation tetracyclines are largely being done by Tetraphase Pharmaceuticals, specifically centered on the fully synthetic fluorocyclines eravacycline, TP-271, and TP-6076. Eravacycline (currently in Phase III) is generally more potent overall and has a slightly better spectrum than tigecycline, but also does not cover P. aeruginosa. TP-271 has a much more limited spectrum and is directed at the bacterial pathogens responsible for community-acquired pneumonia . TP-6076 has potent activity against a range of pathogens including A. baumannii and carbapenem-resistant Enterobacteriaceae and has entered Phase I trials as of this writing. It was also selected for funding support from CARB-X (www.carb-x.org).
Aminoglycoside resistance in the clinic occurs via a number of different mechanisms: mutations altering the target (rRNA or ribosomal proteins), transport defects, efflux and, importantly, modifying enzymes that inactivate the drug [148, 409, 437]. Chromosomal mutations of the target are very rare in Gram-negative bacilli mainly due to the high number of copies of the 16S rRNA [409, 437]. Reports of target-site mutations in clinical isolates have been limited to Mycobacterium spp. [438, 439] and Borrelia burgdorferi , and so will not be addressed further. Each of the remaining mechanisms is discussed below, and the epidemiology of aminoglycoside resistance is discussed in Chap. 10.
184.108.40.206 Aminoglycoside-Modifying Enzymes
Major aminoglycoside-modifying enzymes present in clinical isolates and their resistance profile
Aminoglycoside acetyltransferases (AACs)
Tobramycin, amikacin, netilmicin, dibekacin, sisomicin, kanamycin, isepamicin
Tobramycin, gentamicin, netilmicin, dibekacin, sisomicin
Gentamicin, sisomicin, fortimicin
Kanamycin, amikacin and tobramycin, ciprofloxacin, and norfloxacin
Aminoglycoside phosphotransferases (APHs)
Kanamycin, neomycin, streptomycin, lividomycin, paromomycin, and ribostamycin
Kanamycin, neomycin, lividomycin, paromomycin, butirosin, and ribostamycin
Aminoglycoside nucleotidyltransferases (ANTs)
Tobramycin, gentamicin, dibekacin, sisomicin, kanamycin
Tobramycin, amikacin, dibekacin, kanamycin, isepamicin
The aminoglycoside acetyltransferases (AACs) are the major class among these modifying enzymes. These enzymes are acetyl CoA-dependent, and they acetylate various amino groups found on the aminoglycoside structure [442, 443, 454, 455]. The most common AACs in Gram-negative bacteria are AAC(6′)-I, AAC(3)-IIa, and AAC(3)-I. Recently, a broadening in activity spectra for some of these enzymes has also been observed. AAC(6′)-Ib-cr has acquired the ability to modify fluoroquinolones by acylation of the secondary amine of the piperazine ring of the antibiotic present on ciprofloxacin and norfloxacin but not levofloxacin [447, 448]. The aminoglycoside phosphotransferases (AHPs) and nucleotidyltransferases (ANTs) are both ATP-dependent enzymes. APHs phosphorylate the hydroxyl groups on the aminoglycoside similar to ATP-dependent kinases, sharing high similarity with serine-threonine eukaryotic kinases [442, 443]. ANTs utilize ATP as AMP donor that is added on the aminoglycoside hydroxyl groups. The major representatives of this class present in the clinic are ANT(2″) and ANT(4′) described in Table 4.6 . Bifunctional enzymes with a broader spectrum of activity have been reported. ANT(3″)-Ii/AAC(6′)-IId that confers resistance to streptomycin, spectinomycin, and gentamicin has been isolated from Serratia marcescens, a human enteropathogen [456, 457]. Among the aminoglycoside-modifying enzyme genes, aac(6′)-Ib was the most prevalent (37.5% of isolates were positive), in a study looking at 200 Gram-negative bacilli resistant to aminoglycosides . In another study from Spain of 330 aminoglycoside resistant Enterobacteriaceae isolates, the predominant resistance determinant was Aph(3″)-Ib (65.4% of isolates were positive) in accordance with the observed streptomycin resistance phenotype [409, 432].
220.127.116.11 Ribosomal Protection
Posttranscriptional methylation of 16S rRNA by aminoglycoside rRNA methyltransferases (RMTs) is an emerging resistance mechanism for this class of antibiotics. RMTs modify specific nucleotide residues (N7 position of nucleotide G1405 or N1 position of nucleotide A1408) of the 16S rRNA, thereby preventing aminoglycosides from binding to their target [408, 458]. This mechanism was originally identified and characterized in antibiotic-producing organisms as a self-protection mechanism  but has now been emerging in several important Gram-negative nosocomial pathogens . In 2003, aminoglycoside rRNA methyltransferases were reported in K. pneumoniae (encoded by the armA gene) and P. aeruginosa (encoded by the rmtA gene) both conferring high-level resistance to 4,6-disubstituted deoxystreptamines [461, 462]. After the first identification of these genes, a series of plasmid-encoded RMTs have been identified (encoded by rmtB1, rmtB2, rmtC, rmtD, rmtD2, rmtE, rmtF, rmtG, and rmtH) in several clinical isolates . In 2007 an aminoglycoside rRNA methyltransferase (encoded by the npmA gene) was reported from E. coli isolated in 2003 from the urine of an inpatient in a general hospital in Japan, which conferred resistance to 4,6- and 4,5-disubstituted 2-deoxystreptamines . Further information on the class of enzyme including their origin and the impact on the use of aminoglycoside can be found in a review published in 2016 by Doi et al. . Even though a low prevalence of this class of enzyme has been reported in clinical isolates, their ability to confer high-level pan aminoglycoside resistance in conjunction with their presence on mobile elements threatens the future use of aminoglycosides.
18.104.22.168 Decreased Permeability and Efflux
As described in Sect. 4.2.1, a major issue in Gram-negative bacilli is the inability of many drugs to penetrate the cell membrane. However, aminoglycosides are cationic and therefore can interact with negatively charged LPS to facilitate “self-promoted uptake” across the OM. This is followed by energy-dependent (electron transport-mediated) uptake across the inner membrane. Correspondingly, alterations in LPS or reductions in uptake across the inner membrane were proposed to play a role in reducing susceptibility. In the case of P. aeruginosa, this may involve aminoarabinose modification of the lipid A moiety of LPS, controlled by the PhoP-PhoQ two-component regulator pair. This system is a well-characterized determinant of resistance to the polymyxin class of antibiotics (see Sect. 4.2.6), but its involvement in aminoglycoside resistance is less well understood [464, 465]. Reduced expression of some oligopeptide transporters, such as OppA, may also reduce entry of aminoglycosides .
Efflux by RND family pumps has been shown to play a significant role in aminoglycoside extrusion and therefore resistance in several pathogens. In E. coli, the AcrAD pump has been shown to be able to efflux aminoglycosides [466, 467]. Other pumps involved in aminoglycoside efflux are AmrAB-OprA and BpeAB-OprB in Burkholderia pseudomallei [466, 468], AdeABC in A. baumannii , and MexXY-OprM of P. aeruginosa [469, 470]. The MexXY-OprM efflux pump is unique among the complement of P. aeruginosa pumps in its ability to extrude aminoglycosides. It is also induced by agents inhibiting protein synthesis, contributing to both impermeability and adaptive aminoglycoside resistance [470, 471]. The latter refers to the induction of reversible resistance by exposure to aminoglycosides, which is now known to result largely from induction of MexXY and possibly by a concomitant upregulation of anaerobic respiration genes which may compromise aminoglycoside uptake across the inner membrane . The regulation (inducibility) of MexXY expression involves the MexZ repressor and PA5471, a protein of unknown function, which sense disruptions of protein synthesis (translation) [52, 473]. More recent work showed the induction of MexXY by aminoglycosides also depends on the two-component system AmgRS . This appears to be related to the role of AmgRS as a cell envelope stress response regulator and, in the case of aminoglycosides, in responding to incorporation of misfolded proteins in the inner membrane that results from aminoglycoside action on the ribosome. Mutations in amgS can also cause constitutive activation of MexXY expression. Mutations in genes encoding another two-component system ParRS, involved in resistance to polymyxins, also cause upregulation of MexXY and aminoglycoside resistance [75, 475]. It should be noted that although aminoglycoside-modifying enzymes are generally the most important resistance mechanism in Gram-negative bacteria, this does not appear to be the case in P. aeruginosa isolates recovered from CF patients. Since these patients tend to be colonized by strains common in the natural environment, there is less chance for these enzymes to accumulate, and therefore only a small percentage of CF isolates harbor aminoglycoside-modifying enzymes . Therefore, efflux by MexXY, likely in conjunction with other mechanisms, is comparatively more important in this instance.
Growth in the biofilm mode is another barrier for the entry of aminoglycosides in bacteria, contributing to intrinsic and adaptive resistance to this class of antibiotics. Biofilms are defined as an intertwined community of bacteria adhering on a surface and surrounded by a self-produced matrix composed of extracellular DNA, proteins, and polysaccharides. Biofilms play a key role in chronic P. aeruginosa infections and have been associated with pulmonary infections in patients with CF where aminoglycosides and, in particular, tobramycin are routinely used [477, 478]. Therefore understanding the role of biofilms in relation to aminoglycoside resistance is of high importance. Subinhibitory concentrations of aminoglycosides, especially tobramycin, have been shown to induce biofilm formation in P. aeruginosa by the induction of the aminoglycoside response regulator (arr) gene. This gene is postulated to be involved in the regulation of cell surface adhesiveness and therefore contributes to biofilm-specific aminoglycoside resistance. Some studies have hypothesized the ability of the biofilm matrix to bind aminoglycosides and therefore drastically reduce their antibacterial activity [479, 480]. Several in vitro studies have shown co-dosing of an aminoglycoside with a cationic steroid antibiotic, CSA-13, or a cationic peptide, DJK-5, is effective in overcoming biofilm-mediated resistance [481, 482]. More study to understand the molecular mechanisms of biofilm induction by aminoglycosides as well as device strategies to inhibit biofilm formation are needed.
22.214.171.124 Novel Approaches and Treatment Strategies
Better understanding of aminoglycoside resistance, molecular mechanisms and dosing regimens to minimize toxicity, together with the increase of resistance in the clinic, has created interest in discovering new approaches or novel aminoglycosides aimed at overcoming resistance. One approach that has been pursued to extend the useful lifespan of aminoglycosides is the pursuit of aminoglycoside-modifying enzyme (AME) inhibitors that would prevent inactivation of aminoglycosides, similar to the successful combinations of β-lactamase inhibitors with β-lactams. Despite several three-dimensional structures of all representative members of the three classes of AMEs [483, 484, 485, 486, 487, 488, 489, 490, 491, 492] and better understanding of their molecular mechanism of action, screening efforts and structure-based drug designs have not yet led to viable clinical candidates. Labby and Garneau-Tsodikova have reviewed several efforts aimed at identifying suitable AME inhibitors in a recent review .
New delivery strategies for aminoglycosides have been adopted in the past few years that provide higher local concentration at the infection site with a lower total amount of drug delivered, which reduces systemic exposure and safety liabilities of this class of drug. The best example of new delivery method for inhaled aminoglycoside is tobramycin inhalation powder (TOBI® Podhaler®), delivered via the T-326 inhaler (Novartis). Long-term safety studies in patients with CF have shown that it is well tolerated, there was no evidence of serum tobramycin accumulation with successive cycles, and no unexpected adverse events were observed. Further, the new powder delivery method improved compliance due to shorter administration time, convenience, and ease of use [498, 499]. Bayer Healthcare, in collaboration with Nektar, is currently developing BAY41-6551, a drug-device combination of a specially formulated amikacin. BAY41-6551 has recently completed Phase III as an adjunctive treatment for intubated and mechanically ventilated patients with Gram-negative pneumonia and showed bactericidal activity against most isolates tested with amikacin MICs ≤ 256 μg/mL [500, 501, 502].
Aminoglycosides are a key class of antibiotic used by physicians to treat serious infections caused by MDR Gram-negative and Gram-positive pathogens. Continued characterization of aminoglycoside resistance mechanisms may enable the design of resistance determinant inhibitors and/or new aminoglycosides capable of circumventing these mechanisms. Additionally, efforts aimed at optimization of dosing regimens and discovery of new delivery strategies should help to maintain and extend the clinical utility of this important antibiotic class.
Polymyxins are an older class of cationic cyclic lipopeptide antibiotics that were introduced into clinical use in the 1950s (polymyxin B and colistin, also known as polymyxin E). However, the use of polymyxins declined sharply around the early 1970s due to concerns of toxicity [503, 504] and the availability of safer antibiotics. The mechanism by which polymyxins kill bacteria is not fully understood. One mechanistic step that is well established is an initial interaction of the cationic peptide with negative charges on the lipopolysaccharide (LPS ) that forms the outer leaflet of the Gram-negative OM [505, 506]. This interaction occurs mainly via negatively charged phosphate residues located on lipid A and is required for the “self-promoted uptake” of polymyxins into the bacteria. Binding of polymyxin is thought to displace divalent cations (Mg2+, Ca2+) that cross-link adjacent LPS molecules, and this can disrupt to some extent the permeability barrier of the OM, which also increases uptake of the polymyxin. This is unlikely to be responsible entirely for cell killing. As discussed in more detail in Sect. 126.96.36.199 below, derivatives of polymyxin (e.g., polymyxin B nonapeptide) exist with dramatically reduced antibacterial activity that retain the OM disruption activity. Polymyxins may ultimately kill via mechanisms that include lysis of the inner membrane  generation of toxic hydroxyl radicals  and inhibition of respiration via targets such as type II NADH-quinone oxidoreductases (NDH-2) . The interaction of polymyxins with the Gram-negative OM LPS has two main implications: first, that susceptibility to this class of compound can be decreased by restructuring LPS to reduce its negative charge (discussed below) and, second, that it limits the spectrum of polymyxins to some but not all Gram-negative pathogens. This includes the important Gram-negative ESKAPE pathogens E. coli, K. pneumoniae, P. aeruginosa, and A. baumannii. Polymyxins were reintroduced into the clinic in the early 2000s as a last-line therapeutic option to address the emergence of MDR and XDR in these pathogens. Not long after this reintroduction into clinical practice, emergence of colistin-resistant strains began to increase, prompting concern about their ongoing therapeutic usefulness, especially considering that no other options may exist in scenarios where colistin is being used. The main mechanism by which Gram-negative pathogens that are susceptible to polymyxins develop resistance is via alterations in their lipopolysaccharide that reduce its net negative charge, thereby reducing uptake of polymyxins. This can occur through selection of chromosomal mutations in genes involved in regulating LPS remodeling or by horizontal acquisition of plasmids harboring genes mediating this process. Some Gram-negative pathogens (e.g., Burkholderia cepacia, Proteus mirabilis, Serratia marcescens) are not susceptible to polymyxins (intrinsic resistance), because their LPS always possesses such modifications. These mechanisms are discussed in the sections below.
188.8.131.52 Mutationally Acquired Resistance Mediated by Reduction of Negative Charge Status of LPS
Gram-negative bacteria have a broad ability to remodel their OM. The best understood and most widespread mechanism that decreases susceptibility to polymyxins utilizes this capacity by modification of LPS to reduce its negative charge and, consequently, the initial binding step of cationic polymyxin to the cell surface. This is highly complex and varies among different strains but in general occurs in Enterobacteriaceae and P. aeruginosa via masking the negatively charged 4′-phosphates on lipid A, and to a lesser extent the 1 position phosphate or the 3-deoxy-d-manno-octulosonic acid (KDO), by addition of 4-amino-4-deoxy-l-arabinose (l-Ara4N) . Synthesis and transfer of l-Ara4N are mediated by the products of the arn locus (e.g., arnBCADTEFpmrE in P. aeruginosa, also known as pmrHFIJKLME). This is perhaps the most common mechanism of reducing the negative charge of LPS described in these organisms. Enterobacteriaceae can also add phosphoethanolamine (pEtN) (mainly to the 1 position but also to other locations such as 4′-lipid A, KDO, or core oligosaccharide) via transferases such as PmrC. The contribution of pEtN to resistance appears to be smaller than that of l-Ara4N, but both clearly play a role, and both decorations can occur together. Recently, P. aeruginosa has also been shown to be able to modify its LPS with pEtN when zinc is present, under the regulation of the ColRS two-component regulatory system . A. baumannii lacks an arn locus and therefore cannot carry out L-Ara4N modification but can undergo pEtN modification  or, as recently described, galactosamine modification . Regulatory control of these modifications is highly complex and is often mediated by interrelated networks of two-component regulatory systems (TCSs). The PmrAB regulator pair controls l-Ara4 and/or pEtN modification and is widespread in Gram-negative pathogens. Similarly the PhoPQ system, also present in several organisms such as S. enterica, E. coli, K. pneumoniae, and P. aeruginosa (but absent in A. baumannii), is important for control of LPS modification and can be interconnected with the PmrAB system. For example, these two regulatory systems are interconnected in S. enterica and E. coli via the PmrD protein . These systems can upregulate LPS modification (e.g., upregulate expression of the arn locus) in response to certain conditions such as magnesium limitation or exposure to polymyxin or other cationic peptides [513, 514, 515]. Although there is likely some adaptive change in susceptibility to polymyxins mediated by these systems upon drug exposure, or by exposure to cationic peptides in the host, resistance is generally mutationally acquired, via selection of mutations in the genes encoding these regulators, which leads to strong constitutive upregulation of LPS modification. Some mutations in these regulatory genes may also lead to stronger inducibility of the systems by the polymyxin . Regulatory mutations can be selected in vitro and have also been associated with resistance in the clinical setting, including mutations found in pmrA/pmrB in K. pneumoniae [517, 518, 519, 520], P. aeruginosa [516, 521, 522], and A. baumannii [523, 524, 525, 526, 527] and phoP-phoQ in K. pneumoniae [528, 529, 530] and P. aeruginosa [464, 522, 531]. A. baumannii lacks both phoPQ and an arn locus, so pmrB mutations are frequently found in this pathogen, and these mutants will have pEtN modification through activation of the pmrC transferase gene located in the pmrABC locus . However, mutations in pmrAB are not always found in colistin-resistant A. baumannii clinical isolates suggesting that mutations elsewhere on the chromosome can upregulate pmrABC .
The importance of the PmrAB and PhoPQ systems in controlling OM remodeling and resistance to polymyxins is well established; however a full understanding of these phenomena is still forthcoming. Recently in the case of K. pneumoniae, mutation of mgrB, which encodes a negative feedback regulator of the PhoPQ two-component system, was revealed as an important mediator of LPS modification and colistin resistance [533, 534]. Loss of MgrB function leads to constitutive activation of PhoPQ and LPS modification. This mechanism appears to be relatively widespread in K. pneumoniae clinical isolates [529, 535, 536] and can be mediated by insertion of genetic elements that may also carry other resistance genes such as β-lactamases . The fact that colistin resistance can arise from any loss-of-function mutation of mgrB likely explains its relatively high prevelance among colistin-resistant K. pneumoniae isolates. Several additional TCSs involved in polymyxin resistance have been characterized more recently including the CrrAB (colistin resistance regulon) in K. pneumoniae . Changes in CrrB result in upregulation of the PmrAB system via CrrR, which then upregulates arn genes and pmrC, leading to LPS modification and colistin resistance . However, not all K. pneumoniae harbor the crrAB genes. In P. aeruginosa, three additional TCSs also known to be involved in modulating susceptibility to polymyxins have been described: ParRS , ColRS , and CprRS . ParRS and CprRS participate in adaptive resistance to polymyxins by upregulating expression from the arn locus upon sensing polymyxins or other cationic peptides. ColRS and CprRS are required for high-level polymyxin resistance resulting from mutations in phoPQ . These interactions appear complex, and mutational analysis also suggested that additional factors beyond l-Ara4N modification of lipid A could be involved in resistance in P. aeruginosa, but this remains to be fully elucidated . Mutations in the parRS genes were subsequently shown to reduce susceptibility to multiple classes of antibiotic due to coordinately upregulating expression from the mexXY efflux pump genes and arn and downregulating expression of the oprD porin gene . For additional details on polymyxin resistance mechanisms, see Jeannot et al. . Another aspect of resistance relevant to colistin is the phenomenon of heteroresistance, which refers to the presence of a substantial stable resistant subpopulation in cultures of isolates that may score as susceptible to an antibiotic by standard susceptibility testing. Colistin heteroresistance has been described mainly in K. pneumoniae [528, 543, 544] and A. baumannii , and resistant subpopulations can harbor a range of resistance mutations . Heteroresistant isolates can be recovered from patients with no prior treatment with polymyxins, and it is expected that the use of colistin could rapidly enrich for the resistant subpopulation leading to clinical failures. Heteroresistance can be missed by standard susceptibility tests, suggesting that the rates of, and potential for, selecting colistin resistance in the clinic may be underestimated. For more information on heteroresistance, see Chap. 9 in this volume. Finally, as mentioned above, a number of Gram-negative bacteria are intrinsically resistant to polymyxins. These include Burkholderia cepacia complex, Proteus, Serratia, Providentia, and others. These organisms differ from susceptible strains in that their LPS is always constitutively modified with l-Ara4N or has other alterations affecting polymyxin binding.
184.108.40.206 Mutations Causing Loss or Reduction of LPS
Synthesis of LPS (in particular the lipid A portion) and assembly of the LPS-containing OM are essential for the growth and/or viability of most Gram-negative pathogens, but there are a few exceptions to this. Neisseria meningitidis, Moraxella catarrhalis [586, 587] and a subset of A. baumannii have been shown to tolerate loss of LPS biosynthesis. Indeed, this was uncovered in the case of A. baumannii during in vitro studies of colistin resistance, where mutations in genes involved in lipid A biosynthesis were directly selected from A. baumannii strain ATCC 19606 or other strains on polymyxin-containing medium [112, 546]. Point mutations were initially identified in the lpxA, lpxC, or lpxD genes  that encode enzymes catalyzing the first three steps of lipid A biosynthesis . A follow-up experiment identified mutants where lpxA or lpxC were inactivated by insertion sequence ISAb11 , and subsequently an engineered mutant deleted for lpxC was reported (described in ), confirming that lpxA and lpxC are dispensable in A. baumannii ATCC 19606, at least under laboratory growth conditions. These mutants lack lipid A, the target of the initial interaction with polymyxins, and as such are highly resistant to polymyxins but are also highly susceptible to a range of other antibiotics due to loss of the protective lipid A-containing OM . To date, it appears that this mechanism may be confined to mutations in genes encoding enzymes occurring early in the lipid A biosynthesis pathway since inactivation of steps occurring later in the lipid A biosynthetic pathway (e.g., LpxH  or LpxK ) causes toxic accumulation of lipid A synthetic pathway intermediates and is not tolerated. Furthermore, this mechanism does not apply across all A. baumannii, as only a subset appears to tolerate loss of lipid A biosynthesis (e.g., ATCC 19606). The reasons for this are not fully understood, but a recent study showed potentially compensatory transcriptomic changes in response to loss of lpxA , whereas others showed that expression of penicillin-binding protein (PBP)-1A rendered lipid A loss lethal in strains that could otherwise tolerate lipid A loss and that cells lacking both PBP 1A and lipid A had increased expression of lipoproteins on their surface that may compensate for lipid A loss . Although lipid A loss and colistin resistance can be readily selected in vitro, it stimulates debate about its relevance in the clinic, both in terms of colistin resistance and, as discussed above in Sect. 220.127.116.11, with respect to the evaluation of novel antibacterial targets within the lipid A biosynthetic pathway (e.g., LpxC). This stems from the notion of whether A. baumannii lacking lipid A (LPS ) can survive during infection and therefore could be selected during colistin treatment.
Since the Gram-negative OM provides protection from the host immune system, it is generally thought that loss of lipid A (OM) would render the cells unfit in the host environment. Supporting this, colistin-resistant isolates with mutations in lpxA, lpxC, or lpxD  were highly attenuated in C. elegans and mouse models of infection , and an LpxC inhibitor that lacked in vitro antibacterial activity against A. baumannii was efficacious in a mouse model of infection . Overall this is consistent with the very high detection of pmr mutations among colistin-resistant clinical isolates rather than loss of LPS and implies that total loss of lipid A may be more of an in vitro phenomenon. More recent analyses of colistin-resistant A. baumannii XDR clinical isolates identified mutations in pmrA, lpxC, and lpxD (and lpsB, involved in synthesis of the core region attached to lipid A and shown to be involved in intrinsic polymyxin resistance ) occurring together . Isolates selected for further study (AC12 and AC30) produced considerably less LPS than either the laboratory strain ATCC 19606 or polymyxin-susceptible clinical isolates . This suggests a possibility that mutations reducing, but not abolishing, lipid A biosynthesis may emerge over time in the clinic and that the combination of pEtN and/or galactosamine modification with reduced lipid A synthesis could conspire to decrease susceptibility. These isolates were generally drug resistant, suggesting that the reduction in lipid A may not be enough to severely compromise the OM permeability barrier, also allowing for their survival in the host. It is tempting to speculate that the mutation in lpsB may serve to further stabilize the reduced levels of modified lipid A core present in these cells and that other factors are likely involved in determining the overall susceptibility of A. baumannii to colistin [552, 554], including the level of lipid A acylation. A full understanding of this mechanism awaits further study. Whether this phenomenon can extend to other Gram-negative pathogens that strictly require LPS for growth or viability remains to be seen, but one recent study forcing the in vitro evolution of colistin resistance in P. aeruginosa using a morbidostat approach generated mutations in pmr genes and lpxC among others .
18.104.22.168 Plasmid-Mediated Modification of LPS
Mutations mediating resistance to colistin can occur fairly rapidly in some Gram-negative pathogens as described above, but there were initially no reports of horizontal transfer of mobile elements carrying genes mediating LPS modification and colistin resistance. This changed in 2015 with reporting of the plasmid-borne mcr-1 gene encoding a pEtN transferase in E. coli strains in China and its distribution in strains isolated from raw meat, animals (pigs), and humans . It was quickly established that: plasmid-borne mcr-1 was widespread in many regions of the world; occurred in isolates from food animals, meat and vegetables, the environment, and humans; was found mainly in E. coli but also occurred in other bacteria; and was detected in isolate collections dating back to the 1980s [557, 558]. The first identification of the mcr-1 gene in E. coli from a patient in the United States was reported in 2016 . Additional mcr-1.2 , mcr-2 , and mcr-3  variants have now also been identified. The specific impact of the mcr-1 gene in all four Gram-negative ESKAPE pathogens has very recently been reported. Mcr-1 mediates pEtN modification in E. coli, K. pneumoniae, A. baumannii, and P. aeruginosa; however it only seems to shift susceptibility in the first three . Given that colistin was reintroduced into clinical use primarily as a last-line therapy for treating MDR Gram-negative infections, including carbapenem-resistant Enterobacteriaceae (CRE), the identification of a mobile element conferring colistin resistance raised immediate concern as to its potential dissemination into strains such as CREs, and indeed this has already occurred. For example, the mcr-1.2 gene was originally found in a KPC carbapenemase-producing K. pneumoniae human clinical isolate of the clinically important ST512 lineage, isolated in Italy . Two multidrug-resistant K. pneumoniae human clinical isolates were shown to harbor both the blandm-5 metallo-β-lactamase and mcr-1 genes . E. coli isolates from food and human origin had both the blandm-9 and mcr-1 genes [565, 566], and an isolate from a human urinary tract infection in the United States had blandm-5 and mcr-1 . The first death known to result from such an infection in the United States occurred in Nevada in 2016 and was attributed to an untreatable K. pneumoniae harboring an NDM metallo-β-lactamase and plasmid-borne mcr-1 . Although still generally of lower incidence worldwide , the spread of these untreatable strains is inevitable. If colistin therapy will continue to be used, vigilance in detection and surveillance of both carbapenem and colistin resistance and corresponding implementation of effective infection control and stewardship procedures are very important . Furthermore, the need for new antibiotics to treat these infections has now become an extreme priority. Finally, the widespread distribution of mcr-1 in food animals and food products is entirely consistent with the use of large amounts of colistin in agriculture  and therefore the dissemination of these antibiotics generally into the environment, particularly localized around farms. MCR-1 and variants are related to a resistance protein from natural producers of polymyxin and to another pEtN transferase, LptA from Neisseria [571, 572]. The evolutionary history remains to be fully understood in this case, but it is difficult not to speculate that this process was enhanced by extensive agricultural and veterinary use of colistin. Now that it has occurred, such continued use of colistin will continue to facilitate the maintenance and spread of mcr-1-containing strains, and so significant benefit may be derived from finding creative ways to address this issue (for more information on agricultural use of antibiotics, see Chap. 10 in this volume). It is interesting to note that since colistin was out of favor for some time in human clinical usage, the dissemination of mcr-1-containing strains into reservoirs, such as the human gut, and diversity of mcr-1-containing genetic elements may be underestimated .
22.214.171.124 Novel Approaches and New Agents
A derivative of polymyxin B, polymyxin B nonapetide (PMBN), lacks this moiety and is less toxic and less potent as an antibacterial but retains the ability to interact with the bacterial OM and permeabilize cells. PMBN itself has been the subject of much interest and research over the years as a possible potentiating molecule for use in combination with other antibiotics, but the potential for unacceptable residual toxicity still exists. The number of positive charges on polymyxin has also been associated with toxicity. Northern Antibiotics/Spero has exploited this to design a PMBN derivative (SPR741) that contains an N-acetyl-threonine-d-serine side chain, thereby reducing the number of positive charges from five to three relative to PMBN . This molecule does not have significant antibacterial activity and is reported to be less toxic but retains antibiotic potentiation activity in E. coli, K. pneumoniae, and A. baumannii, although it does not potentiate in P. aeruginosa [574, 575, 576]. As of this writing, SPR741 has entered Phase I clinical trials. Cubist Pharmaceuticals (now Merck) have designed a polymyxin decapeptide derivative containing a halo-aryl moiety at its N-terminus (CB-182804) to pursue a reduction in toxicity . CB-182804 exhibited slightly lower antibacterial potency relative to polymyxin B but was efficacious in animal models of infection and is reported to have reduced toxicity. CB-182804 entered Phase I clinical trials but appears to be discontinued. Along the same lines, Pfizer reported a series of analogs replacing the N-terminal acyl chain with biaryl moieties and substituting the diamino-butyrate moiety at amino acid 3 with diamino-propionate. One of these, named 5 X, had slightly improved antibacterial activity and indications of reduced toxicity, but based on studies in dogs, the therapeutic index was not significantly better than polymyxin B . Researchers at Monash University are exploring novel polymyxin lipopeptides to define structure activity relationships for gaining activity against colistin-resistant isolates  and have presented data on other less toxic polymyxin derivatives in conjunction with the Medicines Company . Cantab has also reported on the piperazine derivative that showed reduced cytotoxicity and improved in vivo efficacy over polymyxin B in A. baumannii and P. aeruginosa lung infection models. Additional details and chemical structures for these and other novel polymyxins can be found in the review by Brown and Dawson . Finally, there is renewed interest in the octapeptin natural products which also interact with and traverse the Gram-negative OM but do so via a different mechanism and therefore may be active against polymyxin-resistant strains [582, 583, 584].
4.3 Concluding Remarks
The discovery of antibiotics, along with vaccines and improved concepts in hygiene, could be considered the greatest achievement in healthcare-related science in history. Unfortunately, decades of antibiotic use and perhaps misuse, in both medicine and agriculture, have enriched for resistant bacteria in the clinical setting, eroding the effectiveness of the antibiotics upon which we have relied and setting the stage for a potentially very different reality in medicine from what most of us had grown accustomed to. This is especially unfortunate since so much of medical practice, for example, surgery, has relied on antibiotics for success. As can be seen from the above discussions, antibacterial resistance is complex and multifactorial. However there are key mechanisms that affect susceptibility to certain classes, such as β-lactamases for β-lactams and AMEs for aminoglycosides, which may provide specific strategies for next-generation versions of these antibiotics that address those mechanisms. It is clear that no effort should be spared on these approaches for the near term and that new agents directed at previously unexploited novel targets should be aggressively pursued where they show promise. Hopefully the new awareness of the issue of antimicrobial resistance, and the various incentivizing efforts spawned from this, will be successful in moving us in the right direction to address this threat. Finally, even if new agents come along in the near term, it is imperative that complacency in antibiotic discovery never again sets in. New agents will likely be the last line of defense, and as such, when resistance to them emerges, the overall issue of untreatable infections would again be upon us.
Gram-negative pathogens have a unique additional asymmetric outer membrane (OM). This membrane establishes a significant permeability barrier (reduces influx) to toxic molecules including antibiotics. Mutations decreasing compound permeability can be selected under antibiotic exposure.
Gram-negative pathogens have unique RND family efflux pumps that extrude most antibiotics and other toxic molecules; these work together with the OM to reduce intracellular compound accumulation. Upregulation of efflux pump expression, or changes in compound specificity, can be selected under antibiotic exposure.
An understanding of the design of new compounds that accumulate sufficiently (overcome efflux) in Gram-negative bacteria is lacking.
Mechanisms that cause resistance to specific antibiotics include mutations that alter the antibiotic target, acquisition of proteins that bind and protect the target, or acquisition of enzymes that modify antibiotics.
Clinical resistance to specific classes of antibiotics often results from combinations of efflux and OM changes together with compound-specific mechanisms.
Multidrug resistance arises from various combinations of all of the above.
- 47.Dean CR, Narayan S, Daigle DM, Dzink-Fox JL, Puyang X, Bracken KR, et al. Role of the AcrAB-TolC efflux pump in determining susceptibility of Haemophilus influenzae to the novel peptide deformylase inhibitor LBM415. Antimicrob Agents Chemother. 2005;49(8):3129–35.PubMedPubMedCentralCrossRefGoogle Scholar
- 49.Jones AK, Woods AL, Takeoka KT, Shen X, Wei JR, Caughlan RE, et al. Determinants of antibacterial spectrum and resistance potential of the elongation factor G inhibitor argyrin B in key gram-negative pathogens. Antimicrob Agents Chemother. 2017;61(4):e02400–16.PubMedPubMedCentralCrossRefGoogle Scholar
- 71.Poulou A, Voulgari E, Vrioni G, Koumaki V, Xidopoulos G, Chatzipantazi V, et al. Outbreak caused by an ertapenem-resistant, CTX-M-15-producing Klebsiella pneumoniae sequence type 101 clone carrying an OmpK36 porin variant. J Clin Microbiol. 2013;51(10):3176–82.PubMedPubMedCentralCrossRefGoogle Scholar
- 76.Lomovskaya O, Warren MS, Lee A, Galazzo J, Fronko R, Lee M, et al. Identification and characterization of inhibitors of multidrug resistance efflux pumps in Pseudomonas aeruginosa: novel agents for combination therapy. Antimicrob Agents Chemother. 2001;45(1):105–16.PubMedPubMedCentralCrossRefGoogle Scholar
- 80.Yoshida K, Nakayama K, Ohtsuka M, Kuru N, Yokomizo Y, Sakamoto A, et al. MexAB-OprM specific efflux pump inhibitors in Pseudomonas aeruginosa. Part 7: highly soluble and in vivo active quaternary ammonium analogue D13-9001, a potential preclinical candidate. Bioorg Med Chem. 2007;15(22):7087–97.PubMedCrossRefGoogle Scholar
- 87.Chalhoub H, Pletzer D, Weingart H, Braun Y, Tunney MM, Elborn JS, et al. Mechanisms of intrinsic resistance and acquired susceptibility of Pseudomonas aeruginosa isolated from cystic fibrosis patients to temocillin, a revived antibiotic. Sci Rep. 2017;7:40208.PubMedPubMedCentralCrossRefGoogle Scholar
- 89.Young K, Silver LL, Bramhill D, Cameron P, Eveland SS, Raetz CR, et al. The envA permeability/cell division gene of Escherichia coli encodes the second enzyme of lipid A biosynthesis. UDP-3-O-(R-3-hydroxymyristoyl)-N-acetylglucosamine deacetylase. J Biol Chem. 1995;270(51):30384–91.PubMedCrossRefGoogle Scholar
- 94.Shen X, Johnson AK, Jones AK, Barnes SW, Walker JR, Ranjitkar S, Woods AL, Six DA, Dean CR. Genetic characterization of the hypersusceptible Pseudomonas aeruginosa strain Z61: identification of a defect in LptE. Abstr C-105 Abstr 54th Intersci Conf Antimicrob Agents Chemother American Society for Microbiology, Washington, DC; 2014.Google Scholar
- 106.Titecat M, Liang X, Lee CJ, Charlet A, Hocquet D, Lambert T, et al. High susceptibility of MDR and XDR Gram-negative pathogens to biphenyl-diacetylene-based difluoromethyl-allo-threonyl-hydroxamate LpxC inhibitors. J Antimicrob Chemother. 2016;71(10):2874–82.PubMedPubMedCentralCrossRefGoogle Scholar
- 107.Kasar R, Linsell MS, Aggen JB, Lu Q, Wang D, Church T, et al. Hydroxamic acid derivatives and their use in the treatment of bacterial infections. Google Patents. 2012.Google Scholar
- 113.Garcia-Quintanilla M, Carretero-Ledesma M, Moreno-Martinez P, Martin-Pena R, Pachon J, McConnell MJ. Lipopolysaccharide loss produces partial colistin dependence and collateral sensitivity to azithromycin, rifampicin and vancomycin in Acinetobacter baumannii. Int J Antimicrob Agents. 2015;46(6):696–702.PubMedCrossRefPubMedCentralGoogle Scholar
- 122.Murray B, Pillar C, Pucci M, Shinabarger D. Mechanism of action of SPR741, a potentiator molecule for Gram-negative pathogens. Abstr Saturday-491. New Orleans: Microbe; 2017.Google Scholar
- 133.Reck F, Ehmann DE, Dougherty TJ, Newman JV, Hopkins S, Stone G, et al. Optimization of physicochemical properties and safety profile of novel bacterial topoisomerase type II inhibitors (NBTIs) with activity against Pseudomonas aeruginosa. Bioorg Med Chem. 2014;22(19):5392–409.PubMedCrossRefPubMedCentralGoogle Scholar
- 137.Ito A, Kohira N, Bouchillon SK, West J, Rittenhouse S, Sader HS, et al. In vitro antimicrobial activity of S-649266, a catechol-substituted siderophore cephalosporin, when tested against non-fermenting Gram-negative bacteria. J Antimicrob Chemother. 2016;71(3):670–7.PubMedCrossRefPubMedCentralGoogle Scholar
- 157.Castanheira M, Mendes RE, Jones RN, Sader HS. Changes in the frequencies of β-Lactamase genes among Enterobacteriaceae isolates in U.S. hospitals, 2012 to 2014: activity of ceftazidime-avibactam tested against β-Lactamase-producing isolates. Antimicrob Agents Chemother. 2016;60(8):4770–7.PubMedPubMedCentralCrossRefGoogle Scholar
- 158.European Centre for Disease Prevention and Control. Antimicrobial resistance surveillance in Europe 2015. Annual report of the European Antimicrobial Resistance Surveillance Network (EARS-Net). Stockholm: ECDC; 2017.Google Scholar
- 192.Lauretti L, Riccio ML, Mazzariol A, Cornaglia G, Amicosante G, Fontana R, et al. Cloning and characterization of blaVIM, a new integron-borne metallo-β-lactamase gene from a Pseudomonas aeruginosa clinical isolate. Antimicrob Agents Chemother. 1999;43(7):1584–90.PubMedPubMedCentralCrossRefGoogle Scholar
- 193.Giakkoupi P, Petrikkos G, Tzouvelekis LS, Tsonas S, Legakis NJ, Vatopoulos AC, et al. Spread of integron-associated VIM-type metallo-β-lactamase genes among imipenem-nonsusceptible Pseudomonas aeruginosa strains in Greek hospitals. J Clin Microbiol. 2003;41(2):822–5.PubMedPubMedCentralCrossRefGoogle Scholar
- 199.Moya B, Beceiro A, Cabot G, Juan C, Zamorano L, Alberti S, et al. Pan-β-lactam resistance development in Pseudomonas aeruginosa clinical strains: molecular mechanisms, penicillin-binding protein profiles, and binding affinities. Antimicrob Agents Chemother. 2012;56(9):4771–8.PubMedPubMedCentralCrossRefGoogle Scholar
- 200.Bradford PA, Urban C, Mariano N, Projan SJ, Rahal JJ, Bush K. Imipenem resistance in Klebsiella pneumoniae is associated with the combination of ACT-1, a plasmid-mediated AmpC β-lactamase, and the foss of an outer membrane protein. Antimicrob Agents Chemother. 1997;41(3):563–9.PubMedPubMedCentralCrossRefGoogle Scholar
- 203.Zhang Y, Kashikar A, Brown CA, Denys G, Bush K. Unusual Escherichia coli PBP 3 insertion sequence identified from a collection of carbapenem-resistant Enterobacteriaceae tested In vitro with a combination of ceftazidime-, ceftaroline-, or aztreonam-avibactam. Antimicrob Agents Chemother. 2017;61(8):e00389.PubMedPubMedCentralGoogle Scholar
- 217.Fass RJ, Gregory WW, D’Amato RF, Matsen JM, Wright DN, Young LS. In vitro activities of cefoperazone and sulbactam singly and in combination against cefoperazone-resistant members of the family Enterobacteriaceae and nonfermenters. Antimicrob Agents Chemother. 1990;34(11):2256–9.PubMedPubMedCentralCrossRefGoogle Scholar
- 224.Blais J, Lopez S, Ruzin A, Leeds JA, Dean CR, Simmons RL, Casarez A, Reck F. LYS228 is a novel monobactam with potent activity in vitro against ESBL-producing and Carbapenem-resistant Enterobacteriaceae. Abstracts of microbe 2017. New Orleans, USA; 2017; June 1–5.Google Scholar
- 225.Mendes RE, Rhomberg PR, Schaeffer B, Huband MD, Flamm RK. In vitro activity of LYS228 against Enterobacteriaceae, including molecularly characterized multidrug-resistant isolates. Abstracts of microbe 2017. New Orleans; 2017; June 1–5.Google Scholar
- 228.Hecker SJ, Reddy KR, Glinka TW, Totrov M, Lomovskaya O, Tsivkovski R, Rubio-Aparicio D, Clifton MC, Atkins K, Fairman JF, Griffith DC, Dudley MN. Discovery of a new series of broad-spectrum carbapenemase inhibitors (BCIs) with activity vs serine and metallo Β-lactamases. Abstracts of the 55th interscience conference on antimicrobial agents and chemotherapy. San Diego: 2015; F–270.Google Scholar
- 229.Jones-Dias D, Manageiro V, Ferreira E, Louro D. Antibiotic resistance surveillance program in Portugal p, Canica M. Diversity of extended-spectrum and plasmid-mediated AmpC β-lactamases in Enterobacteriaceae isolates from Portuguese health care facilities. J Microbiol. 2014;52(6):496–503.PubMedCrossRefPubMedCentralGoogle Scholar
- 233.Lister PD, Gardner VM, Sanders CC. Clavulanate induces expression of the Pseudomonas aeruginosa AmpC cephalosporinase at physiologically relevant concentrations and antagonizes the antibacterial activity of ticarcillin. Antimicrob Agents Chemother. 1999;43(4):882–9.PubMedPubMedCentralCrossRefGoogle Scholar
- 239.Winkler ML, Papp-Wallace KM, Hujer AM, Domitrovic TN, Hujer KM, Hurless KN, et al. Unexpected challenges in treating multidrug-resistant Gram-negative bacteria: resistance to ceftazidime-avibactam in archived isolates of Pseudomonas aeruginosa. Antimicrob Agents Chemother. 2015;59(2):1020–9.PubMedPubMedCentralCrossRefGoogle Scholar
- 251.Pagan-Rodriguez D, Zhou X, Simmons R, Bethel CR, Hujer AM, Helfand MS, et al. Tazobactam inactivation of SHV-1 and the inhibitor-resistant Ser130 -->Gly SHV-1 β-lactamase: insights into the mechanism of inhibition. J Biol Chem. 2004;279(19):19494–501.Google Scholar
- 256.Shields RK, Chen L, Cheng S, Chavda KD, Press EG, Snyder A, et al. Emergence of ceftazidime-avibactam resistance due to plasmid-borne blaKPC-3 mutations during treatment of carbapenem-resistant Klebsiella pneumoniae infections. Antimicrob Agents Chemother. 2017;61(3):e02097.PubMedPubMedCentralCrossRefGoogle Scholar
- 276.Kriengkauykiat J, Porter E, Lomovskaya O, Wong-Beringer A. Use of an efflux pump inhibitor to determine the prevalence of efflux pump-mediated fluoroquinolone resistance and multidrug resistance in Pseudomonas aeruginosa. Antimicrob Agents Chemother. 2005;49(2):565–70.PubMedPubMedCentralCrossRefGoogle Scholar
- 282.Miro E, Verges C, Garcia I, Mirelis B, Navarro F, Coll P, et al. Resistance to quinolones and β-lactams in Salmonella enterica due to mutations in topoisomerase-encoding genes, altered cell permeability and expression of an active efflux system. Enferm Infecc Microbiol Clin. 2004;22(4):204–11.PubMedCrossRefPubMedCentralGoogle Scholar
- 294.Briales A, Rodriguez-Martinez JM, Velasco C, de Alba PD, Rodriguez-Bano J, Martinez-Martinez L, et al. Prevalence of plasmid-mediated quinolone resistance determinants qnr and aac(6’)-Ib-cr in Escherichia coli and Klebsiella pneumoniae producing extended-spectrum β-lactamases in Spain. Int J Antimicrob Agents. 2012;39(5):431–4.PubMedCrossRefPubMedCentralGoogle Scholar
- 296.Jacoby GA, Strahilevitz J, Hooper DC. Plasmid-mediated quinolone resistance. Microbiol Spectr. 2014;2(5). https://doi.org/10.1128/microbiolspec.
- 306.Perez F, Rudin SD, Marshall SH, Coakley P, Chen L, Kreiswirth BN, et al. OqxAB, a quinolone and olaquindox efflux pump, is widely distributed among multidrug-resistant Klebsiella pneumoniae isolates of human origin. Antimicrob Agents Chemother. 2013;57(9):4602–3.PubMedPubMedCentralCrossRefGoogle Scholar
- 327.Basarab GS, Kern GH, McNulty J, Mueller JP, Lawrence K, Vishwanathan K, et al. Responding to the challenge of untreatable gonorrhea: ETX0914, a first-in-class agent with a distinct mechanism-of-action against bacterial Type II topoisomerases. Sci Rep. 2015;5:11827.PubMedPubMedCentralCrossRefGoogle Scholar
- 328.Foerster S, Golparian D, Jacobsson S, Hathaway LJ, Low N, Shafer WM, et al. Genetic resistance determinants, In vitro time-kill curve analysis and pharmacodynamic functions for the novel topoisomerase II inhibitor ETX0914 (AZD0914) in Neisseria gonorrhoeae. Front Microbiol. 2015;6:1377.PubMedPubMedCentralCrossRefGoogle Scholar
- 335.Solomkin J, Evans D, Slepavicius A, Lee P, Marsh A, Tsai L, et al. Assessing the efficacy and safety of Eravacycline vs Ertapenem in complicated intra-abdominal infections in the investigating Gram-Negative infections treated with Eravacycline (IGNITE 1) trial: a randomized clinical trial. JAMA Surg. 2017;152(3):224–32.PubMedCrossRefPubMedCentralGoogle Scholar
- 352.Bradford PA, Weaver-Sands DT, Petersen PJ. In vitro activity of tigecycline against isolates from patients enrolled in phase 3 clinical trials of treatment for complicated skin and skin-structure infections and complicated intra-abdominal infections. Clin Infect Dis. 2005;41(Suppl 5):S315–32.PubMedCrossRefPubMedCentralGoogle Scholar
- 355.Fritsche TR, Sader HS, Stilwell MG, Dowzicky MJ, Jones RN. Potency and spectrum of tigecycline tested against an international collection of bacterial pathogens associated with skin and soft tissue infections (2000–2004). Diagn Microbiol Infect Dis. 2005;52(3):195–201.PubMedCrossRefPubMedCentralGoogle Scholar
- 369.Hentschke M, Christner M, Sobottka I, Aepfelbacher M, Rohde H. Combined ramR mutation and presence of a Tn1721-associated tet(A) variant in a clinical isolate of Salmonella enterica serovar Hadar resistant to tigecycline. Antimicrob Agents Chemother. 2010;54(3):1319–22.PubMedCrossRefPubMedCentralGoogle Scholar
- 371.Bratu S, Landman D, Martin DA, Georgescu C, Quale J. Correlation of antimicrobial resistance with β-lactamases, the OmpA-like porin, and efflux pumps in clinical isolates of Acinetobacter baumannii endemic to New York City. Antimicrob Agents Chemother. 2008;52(9):2999–3005.PubMedPubMedCentralCrossRefGoogle Scholar
- 374.Rumbo C, Gato E, Lopez M, Ruiz de Alegria C, Fernandez-Cuenca F, Martinez-Martinez L, et al. Contribution of efflux pumps, porins, and β-lactamases to multidrug resistance in clinical isolates of Acinetobacter baumannii. Antimicrob Agents Chemother. 2013;57(11):5247–57.PubMedPubMedCentralCrossRefGoogle Scholar
- 390.Hu M, Nandi S, Davies C, Nicholas RA. High-level chromosomally mediated tetracycline resistance in Neisseria gonorrhoeae results from a point mutation in the rpsJ gene encoding ribosomal protein S10 in combination with the mtrR and penB resistance determinants. Antimicrob Agents Chemother. 2005;49(10):4327–34.PubMedPubMedCentralCrossRefGoogle Scholar
- 391.Villa L, Feudi C, Fortini D, Garcia-Fernandez A, Carattoli A. Genomics of KPC-producing Klebsiella pneumoniae sequence type 512 clone highlights the role of RamR and ribosomal S10 protein mutations in conferring tigecycline resistance. Antimicrob Agents Chemother. 2014;58(3):1707–12.PubMedPubMedCentralCrossRefGoogle Scholar
- 402.Sato T, Suzuki Y, Shiraishi T, Honda H, Shinagawa M, Yamamoto S, et al. Tigecycline nonsusceptibility occurs exclusively in fluoroquinolone-resistant Escherichia coli clinical isolates, Including the major multidrug-resistant lineages O25b:H4-ST131-H30R and O1-ST648. Antimicrob Agents Chemother. 2017;61(2):e01654.PubMedPubMedCentralGoogle Scholar
- 417.Anderson NH, Bowman J, Erwin A, Harwood E, Kline T, Mdluli K, Ng S, Pfister KB, Shawar R, Wagman A, Yabannavar A. Antibacterial agents. WO/2004/062601. July 29, 2004.Google Scholar
- 421.Clock SA, Tabibi S, Alba L, Kubin CJ, Whittier S, Saiman L. In vitro activity of doripenem alone and in multi-agent combinations against extensively drug-resistant Acinetobacter baumannii and Klebsiella pneumoniae. Diagn Microbiol Infect Dis. 2013;76(3):343–6.PubMedCrossRefPubMedCentralGoogle Scholar
- 429.Haidar G, Alkroud A, Cheng S, Churilla TM, Churilla BM, Shields RK, et al. Association between the presence of aminoglycoside-modifying enzymes and in vitro activity of gentamicin, tobramycin, amikacin, and plazomicin against Klebsiella pneumoniae carbapenemase- and extended-spectrum-β-lactamase-producing Enterobacter species. Antimicrob Agents Chemother. 2016;60(9):5208–14.PubMedPubMedCentralCrossRefGoogle Scholar
- 432.Miro E, Grunbaum F, Gomez L, Rivera A, Mirelis B, Coll P, et al. Characterization of aminoglycoside-modifying enzymes in Enterobacteriaceae clinical strains and characterization of the plasmids implicated in their diffusion. Microb Drug Resist. 2013;19(2):94–9.PubMedCrossRefPubMedCentralGoogle Scholar
- 433.Petrosillo N, Vranic-Ladavac M, Feudi C, Villa L, Fortini D, Barisic N, et al. Spread of Enterobacter cloacae carrying blaNDM-1, blaCTX-M-15, blaSHV-12 and plasmid-mediated quinolone resistance genes in a surgical intensive care unit in Croatia. J Glob Antimicrob Resist. 2016;4:44–8.PubMedCrossRefPubMedCentralGoogle Scholar
- 445.Rather PN, Munayyer H, Mann PA, Hare RS, Miller GH, Shaw KJ. Genetic analysis of bacterial acetyltransferases: identification of amino acids determining the specificities of the aminoglycoside 6’-N-acetyltransferase Ib and IIa proteins. J Bacteriol. 1992;174(10):3196–203.PubMedPubMedCentralCrossRefGoogle Scholar
- 446.Wohlleben W, Arnold W, Bissonnette L, Pelletier A, Tanguay A, Roy PH, et al. On the evolution of Tn21-like multiresistance transposons: sequence analysis of the gene (aacC1) for gentamicin acetyltransferase-3-I(AAC(3)-I), another member of the Tn21-based expression cassette. Mol Gen Genet. 1989;217(2-3):202–8.PubMedCrossRefPubMedCentralGoogle Scholar