The spread of CTX-M-type extended-spectrum β-lactamases

Authors


Corresponding author and reprint requests: G. M. Rossolini, Dipartimento di Biologia Molecolare, Sezione di Microbiologia, Università di Siena, I-53100 Siena, Italy
E-mail: rossolini@unisi.it

Abstract

CTX-M-type enzymes are a group of class A extended-spectrum β-lactamases (ESBLs) that are rapidly spreading among Enterobacteriaceae worldwide. More that 50 allotypes are known, clustered into six sub-lineages. The CTX-M-encoding genes have been captured from the chromosome of Kluyvera spp. on conjugative plasmids that mediate their dissemination among pathogenic enterobacteria. CTX-M-type ESBLs exhibit powerful activity against cefotaxime and ceftriaxone but generally not against ceftazidime, which has important implications for laboratory detection. However, several CTX-M variants with enhanced ceftazidimase activity have been detected. The rapid and massive spread of CTX-M-type ESBLs is rapidly changing the ESBL epidemiology and, in some geographical areas, these enzymes are now the most prevalent ESBLs in Enterobacteriaceae.

Introduction

Among the ‘new’β-lactamases emerging in Gram-negative pathogens, the molecular class A extended-spectrum β-lactamases (ESBLs), active against expanded-spectrum cephalosporins and monobactams (but not against cephamycins or carbapenems), are currently of great epidemiological and clinical interest [1,2]. The dissemination of these enzymes is a problem of global magnitude, with rates of ESBL production being particularly high in some enterobacterial species (e.g., Klebsiella pneumoniae and Escherichia coli) and in some areas (e.g., Europe and South America) [2].

The emergence of ESBLs followed the introduction into clinical practice of expanded-spectrum cephalosporins, which are acknowledged to be the most powerful selectors for these resistance determinants. The spread of ESBLs represents a major challenge to the activity of these drugs, which were a milestone in antimicrobial chemotherapy of infections caused by Enterobacteriaceae and other Gram-negative pathogens [2].

Enterobacteriaceae have adopted two major strategies for the production of ESBL activity: (i) expansion of substrate specificity of the broad-spectrum TEM- and SHV-type β-lactamases, which were already widespread as acquired β-lactamases in the clinical setting when expanded-spectrum cephalosporins were introducted, by substitution of single or multiple amino-acid residues at critical positions; and (ii) capture of new genes encoding enzymes with ESBL activity by horizontal transfer [3]. The latter strategy became more prominent at a somewhat later date (since the early 1990s), but has gained an increasingly important role in the evolution of ESBLs.

Several types of acquired ESBLs, other than TEM and SHV mutants, have been described in Enterobacteriaceae, including the CTX-M, VEB, GES/IBC, PER, TLA, BES and SFO enzymes [2]. Among these, the CTX-M-type ESBLs are by far the most successful in terms of spread and, in several settings, their impact is currently comparable to, or even greater than, that of TEM- and SHV-type ESBLs. This review aims to provide an overview of CTX-M-type ESBLs and their rapidly increasing epidemiological and clinical impact.

CTX-M-type β-lactamases: a rapidly growing group

CTX-M-type enzymes constitute a distinct lineage of molecular class A β-lactamases, and are a rapidly growing group. The first CTX-M-type enzyme of clinical origin, CTX-M-1, was described in enterobacterial strains isolated in Europe in the late 1980s [4]. Thereafter, new CTX-M variants have been described at an increasing pace and, presently, more than 50 allelic variants are known, clustered in six sub-lineages or groups. Each group is named after the first described member and, usually, includes minor allelic variants that differ from each other by single or a few amino-acid residues (Fig. 1).

Figure 1.

 The six known sub-lineages (or groups) of CTX-M-type β-lactamases, shown in a tree diagram that reflects similarity at the amino-acid sequence level. Filled triangles at the end of each branch indicate the presence of minor allelic variants within the corresponding group. The tree was constructed with the treeview program (http://taxonomy.zoology.gla.ac.uk/rod/treeview.html), based on a multiple sequence alignment of the publicly available CTX-M sequences (http://www.lahey.org/Studies/#). The degree of amino-acid sequence divergence among different groups ranges from 9.3% to 32% (or 25%, not considering the most divergent CTX-M-45 enzyme), being ≤3.5% within each group. CTX-M-14 is identical to CTX-M-18, and only the former is listed. Toho-1 corresponds to CTX-M-44, Toho-2 to CTX-M-45, UOE-1 to CTX-M-15, and UOE-2 and Toho-3 to CTX-M-14.

Capture and spread of CTX-M-type β-lactamase genes in the clinical setting

Unlike for most other acquired ESBLs, the original source of genes encoding CTX-M-type β-lactamases is known. The sources of CTX-M determinants are chromosomal genes resident in members of the genus Kluyvera, which includes a number of environmental species with little or no pathogenic activity against humans. In particular, precursors of genes encoding enzymes of the CTX-M-1 and CTX-M-2 groups have been detected in strains of Kluyvera ascorbata [4,5], while precursors of genes encoding enzymes of the CTX-M-8 and CTX-M-9 group occur in strains of Kluyvera georgiana [4,6]. The original sources of genes encoding members of the CTX-M-25 and CTX-M-45 subgroups remain to be identified, but most likely comprise other members of the genus Kluyvera. Altogether, the heterogeneity of CTX-M-type enzymes spreading in the clinical setting among enterobacterial pathogens probably reflects the occurrence of multiple independent events of capture of blaCTX-M genes from somewhat different sources within the Kluyvera genus.

Multiple genetic mechanisms have apparently been involved in the capture and dissemination of CTX-M determinants. Genes encoding enzymes of several CTX-M groups (CTX-M-1, CTX-M-2, CTX-M-9 and CTX-M-25) have been found downstream of an ISEcp1 insertion sequence (Fig. 2), a transposable element that can co-mobilise flanking DNA fragments via a one-ended transposition mechanism. This was experimentally shown to be able to mediate the capture of blaCTX-M genes from the chromosomes of Kluyvera spp. and to facilitate their inter-replicon mobility inside E. coli hosts [7,8]. Genes encoding enzymes of some CTX-M groups (CTX-M-2 and CTX-M-9) have also been found within the so-called CR1 region associated with some class 1 integrons (Fig. 2). ISEcp1 and the recombination system associated with CR1 regions have probably played a major role in the capture of blaCTX-M genes by conjugative plasmids, and in their further inter-replicon dissemination, while the association of CR1 regions with Tn402-like backbones could provide a further mechanism of mobility for resistance genes. Recently, a blaCTX-M-10 gene (CTX-M-9 group) was detected within an original genetic context, associated with phage-related sequences (Fig. 2), suggesting that additional types of mobile element might be involved in the mobilisation of blaCTX-M genes.

Figure 2.

 Examples of different genetic contexts of blaCTX-M genes. (a) blaCTX-M-15 located downstream of ISEcp1. A similar arrangement has also been observed with several other blaCTX-M allotypes of the CTX-M-1, CTX-M-2, CTX-M-9 and CXT-M-25 groups [4,33]. (b) blaCTX-M-2 located downstream of ISCR1, in association with a class 1 integron structure [4,34]. A similar arrangement has also been observed with several other blaCTX-M allotypes of the CTX-M-2 and CXT-M-9 groups [34]. (c) blaCTX-M-10 in an unusual genetic context, associated with phage-related genes and insertion sequences [35].

The horizontal spread of blaCTX-M genes among strains of the same or different enterobacterial species is largely promoted by plasmids, which often are self-conjugative and carry additional resistance determinants [4].

Functional properties of CTX-M-type β-lactamases

The CTX-M-type β-lactamases are natural ESBLs that exhibit a striking substrate preference for cefotaxime (and ceftriaxone) over ceftazidime (Table 1). Their strong cefotaximase activity is related to the unique geometry of the β-lactam-binding site, which allows efficient recognition of penicillins, narrow-spectrum cephalosporins and cefotaxime, but not of the bulkier ceftazidime molecule [9,10].

Table 1.   Catalytic efficiencies of CTX-M-type enzymes, representative of major groups, for some β-lactam substrates
EnzymesGroupkcat/KM (/s/M)
PENLOTCTXCAZFEPATM
  1. PEN, penicillin G; LOT, cephalothin; CTX, cefotaxime; CAZ, ceftazidime; FEP, cefepime; ATM, aztreonam; NH, no detectable hydrolysis; ND, no data available.

  2. Data are from [4] and [13]

CTX-M-3CTX-M-1>5 × 1073 × 1073 × 106NH1 × 1031 × 106
CTX-M-8CTX-M-81 × 1072 × 1071 × 106<4 × 1031 × 1052 × 104
CTX-M-9CTX-M-91 × 1072 × 1074 × 1063 × 103ND5 × 104
CTX-M-44CTX-M-23 × 1061 × 1072 × 1061 × 1037 × 104ND

The peculiar substrate specificity of CTX-M-type enzymes has important implications for laboratory detection. Strains of Enterobacteriaceae with CTX-M enzymes normally appear to be resistant or exhibit reduced susceptibility to cefotaxime (also to ceftriaxone and cefpodoxime) and are readily detected as potential ESBL producers using the CLSI breakpoints indicating suspicion of ESBL production [11], whereas they often appear to be susceptible to ceftazidime, with MIC values and inhibition zone sizes, respectively, lower or larger than the CLSI breakpoints indicating suspicion of ESBL production. Therefore, to ensure that no isolate with CTX-M enzymes is missed, cefotaxime (or ceftriaxone or cefpodoxime) must always be included, in addition to ceftazidime, in screening for the presence of ESBL producers in the clinical microbiology laboratory.

Modulation of the substrate spectrum of CTX-M-type enzymes by point mutations is possible, similar to what happens with the TEM- and SHV-type enzymes [3]. In particular, a number of CTX-M mutants with increased ceftazidimase activity have been described. These have probably been selected by the massive use of ceftazidime in clinical practice [4,12,13]. The mutations in these variants occur in two of the structural elements that delimit the β-lactam-binding site, namely the terminal part of the B3 β-strand and the Ω-loop (Fig. 3). The Asp240→Gly substitution in the terminal part of the B3 β-strand is responsible for increased flexibility of the β-strand, rendering the active site more accessible to the bulkier ceftazidime molecule [9], while the substitutions in the Ω-loop (at position 167) apparently modify the mode of interaction of β-lactams with the binding site [14]. Overall, these mutations result in a modest increase in the catalytic efficiency of the enzyme against ceftazidime [4,12,13]. Nevertheless, this is sufficient to significantly increase the ceftazidime MICs for the strains producing the mutant enzymes (Table 2). Unlike the Asp240→Gly substitution, which does not significantly affect the activity of the enzyme against other β-lactam substrates, the substitutions in the Ω-loop are associated with a significant decrease in the catalytic efficiency of the enzyme for other substrates [4,12,13], which is also reflected in the corresponding MIC values (Table 2). This could account, at least in part, for the fact that the Asp240→Gly substitution has apparently been selected more frequently than mutations in the Ω-loop.

Figure 3.

 Modulation of the substrate spectrum of CTX-M-type enzymes by point mutations. The structure of a CTX-M-type enzyme (CTX-M-9; PDB entry 1YLJ) is shown here as a ribbon diagram generated with the deepview program (http://www.expasy.org/spdbv/). As with other serine β-lactamases, the overall molecular fold consists of an α-helical domain (on the left side) and a mixed α/β-domain (on the right side). The β-lactam-binding site, indicated by an arrow, is located in a cleft between the two domains. Positions 167 and 240, where amino-acid substitutions enhancing ceftazidimase activity occur, are located either in the Ω-loop (in green, at the bottom of the binding site) or in the terminal part of the B3 β-strand (β3, in red, on the right side of the binding site), and are shown by red dots. The active site serine residue (Ser*) in the active site, located at the end of H2 α-helix (α2, in yellow), is also shown. The Asp240→Gly substitution should increase the flexibility of the B3 β-strand, rendering the active site more accessible to the bulkier ceftazidime molecule [9]. The Pro167→Ser substitution in the Ω-loop is thought to modify the mode of interaction of β-lactams with the binding site, allowing better recognition of ceftazidime but impairing recognition of some other substrates [14].

Table 2.   Susceptibility to various β-lactams of Escherichia coli strains producing CTX-M derivatives with enhanced ceftazidimase activity, in comparison with that of E. coli strains producing the parent enzymes
E. coli strainsEnzymesMICs (mg/L)
CTXCAZFEPATM
  1. aThe β-lactam susceptibility of E. coli DH10B, as a representative host not producing β-lactamase activity, is shown for comparison.

  2. ND, no data available.

  3. CTX, cefotaxime; CAZ, ceftazidime; FEP, cefepime; ATM, aztreonam.

  4. Data are from [4] (for the CTX-M-3/CTX-M-15, CXT-M-9/CTX-M-16 and CTX-M-14/CTX-M-19 pairs), from [15] (for the CTX-M-1/CTX-M-32 pair) and from [16] (for CTX-M-54).

DH10BaNone≤0.06≤0.06≤0.060.06
DH10BCTX-M-3>25632128128
DH10BCTX-M-15
(CTX-M-3 Asp240→Gly)
>2562566464
J53CTX-M-54
(CTX-M-3 Pro167→Gln)
812812
TG1CTX-M-1>12864848
TG1CTX-M-32
(CTX-M-1 Asp240→Gly)
>128>25664>256
DH5αCTX-M-9161ND4
DH5αCTX-M-16
(CTX-M-9 Asp240→Gly)
168ND8
JM109CTX-M-186421664
JM109CTX-M-19
 (CTX-M-18 Pro167→Ser)
412844

The 1960s epidemiology of CTX-M-type β-lactamases

Strains producing CTX-M-type ESBLs were first reported, sporadically, in the late 1980s in Japan, Europe and Argentina [4]. In the early 1990s, a massive spread of CTX-M-producing strains occurred in Argentina and neighbouring countries. This regional CTX-M epidemic involved several enterobacterial species (including Salmonella enterica, Proteus mirabilis, E. coli, Shigella sonnei, Morganella morganii, Citrobacter freundii, Serratia marcescens and Enterobacter aerogenes) and, mostly, allotypes of the CTX-M-2 group [4]. During the past 15 years, CTX-M-type ESBLs have undergone a rapid and global spread. Enterobacterial strains producing these enzymes have now been reported almost everywhere (Fig. 4) and, in some settings, CTX-M-type enzymes outnumber the classic TEM- and SHV-type ESBLs [15,16]. This massive worldwide dissemination, which could be referred to as the ‘CTX-M pandemic’, is one of the most striking examples of rapid and global dissemination of plasmid-mediated resistance determinants among bacterial pathogens, and has been compared to the dissemination of the broad-spectrum TEM-type β-lactamases observed since the 1960s. The reason(s) for such an explosive dissemination of CTX-M-type ESBLs in Enterobacteriaceae remain(s) to be clarified. Carriage on plasmids that are highly efficient at conjugal transfer, and/or a lower fitness cost imposed by these enzymes and cognate genetic elements upon the bacterial hosts (as compared with other types of ESBLs), could be included among the possible explanations for the remarkable success of CTX-M-type enzymes, as compared with other types of ESBLs. However, these issues remain to be clarified and would constitute an interesting subject for future investigations.

Figure 4.

 World map showing locations where clinical isolates of Enterobacteriaceae producing CTX-M-type extended-spectrum β-lactamases have been reported. Data are according to Reference [4] and the subsequent literature available on the PubMed database (http://www.pubmed.com).

In recent surveillance studies, high rates of CTX-M enzymes among ESBL-producing E. coli and K. pneumoniae isolates have been reported from South America, Asia and Europe, while a lower impact has been observed in Canada (Table 3). In some of those settings, rates of CTX-M-production as high as 89.7% in E. coli and 58.5% in K. pneumoniae have been reported (Table 3).

Table 3.   Rates of isolates with CTX-M enzymes, among extended-spectrum β-lactamase (ESBL)-producing Escherichia coli and Klebsiella pneumoniae isolates, as reported in recent surveillance studies
CountryYear(s)E. coli (%)K. pneumoniae (%)References
  1. aKlebsiella spp.

Russia1997–199835.935.0[17]
Canada2000 6.313.7a[36]
Argentina200080.050.0[37]
Spain200252.312.5[19]
Italy200354.812.3[20]
Taiwan200389.758.5[16]
South Korea200330.432.7[38]

In Europe, where the TEM- and SHV-type ESBLs were first described and have played a major role as ESBL determinants [3], the CTX-M-type ESBLs have recently achieved a remarkable diffusion in several countries. The spread of CTX-M-producing strains was first reported in eastern Europe [4,17], but has subsequently involved also western and southern European countries. One of the most striking examples of rapid dissemination of these ESBLs has been reported in the UK, where CTX-M enzymes were first reported in 2000 and have subsequently undergone an explosive spread involving E. coli and also K. pneumoniae, with a predominance of group 1 enzymes (mostly CTX-M-15) [18]. In the London area, E. coli resistance to expanded-spectrum cephalosporins increased from 1.8% in 2001 to 7.5% in 2004, largely as a consequence of the diffusion of CTX-M-type ESBLs [18], raising considerable concern, since E. coli is one of the most common species isolated from clinical samples. A remarkable diffusion of CTX-M-type ESBLs has also been observed in the most recent nationwide surveys of ESBL production carried out in Spain [19] and Italy [20]. In Spain, rates of production of CTX-M enzymes were found to be 52.3% and 12.5% among ESBL-producing isolates of E. coli and K. pneumoniae, respectively, with a predominance of group 9 (CTX-M-9 and CTX-M-14) and group 1 (CTX-M-10) [19]. In Italy, the rates of CTX-M production were found to be 54.8% and 12.3% among ESBL-producing isolates of E. coli and K. pneumoniae, respectively, with an absolute predominance of group 1 enzymes (mostly CTX-M-1 and CTX-M-15 and, less frequently, CTX-M-32) [20,21]. High-rates of CTX-M enzymes have also been reported in Greek hospitals [15], and the presence of CTX-M-type ESBLs of various groups has been well-documented in other European countries, including France, Austria and Sweden [22–24].

The progressive change in the epidemiology of ESBLs due to the spread of CTX-M-type enzymes has been clearly documented in two longitudinal studies, carried out at a regional level, in northern Italy and in Austria [23,25]. In those areas, the prevalence of CTX-M enzymes among ESBL-producing E. coli was found to be low or nil at the beginning of the surveillance period (12.5% in Italy in 1999; 0% in Austria in 1998), but a steady increase in prevalence was observed during the following years, concluding with rates that were as high as 38.2% of all ESBLs in Italy in 2003 and 85% in Austria in 2004, respectively. Notably, the increase of CTX-M enzymes in E. coli as compared with other ESBL types did not reflect a relative, but rather an absolute, increase in the former isolates.

Although originally confined to hospitals, ESBL-producing strains are now emerging also in the community [26]. This is an alarming phenomenon that could have major implications for antimicrobial chemotherapy. Most ESBL-producing isolates that cause community-acquired infections have been E. coli, with CTX-M-type β-lactamases [26]. In a recent study, producers of CTX-M enzymes were found to be involved in eight of ten cases of bacteraemia caused by community-acquired ESBL-producing enterobacteria [27].

Another relevant issue in the epidemiology of ESBL-producing Enterobacteriaceae (especially E. coli) is their presence in the commensal microbiota of humans. Here again, the CTX-M-type ESBLs appear to play a consistent role. In a recent study from Spain, rates of faecal carriage of ESBL-producing E. coli during non-outbreak situations were found to be as high as 11.8% among inpatients and 5.5% among outpatients, while 42% and 69% of the ESBL-positive isolates, respectively, had CTX-M-type enzymes [28]. In another recent study, carried out in Israel, a 10.8% rate of faecal carriage of ESBL-positive isolates was found upon hospital admission, and most of the community-acquired isolates produced a CTX-M-type enzyme [27]. The risk of developing subsequent bacteraemia caused by an ESBL-producing strain was found to be significantly higher in colonised vs. non-colonised patients [27]. Finally, E. coli isolates with CTX-M enzymes have also been detected in samples from livestock [29,30] and companion animals [31], showing that animals might act as an important reservoir.

Clinical impact of CTX-M-type β-lactamases

Overall, production of CTX-M-type β-lactamases has the same clinical implications as production of other ESBLs, leaving carbapenems as the only reliable β-lactams for treatment of serious infections caused by producer pathogens [2]. Similar to what was observed with enterobacterial strains producing TEM- and SHV-type ESBLs, high-rates of co-resistance to potentially active drugs, e.g., fluoroquinolones and aminoglycosides, have also been reported for strains with CTX-M enzymes. In the most recent nationwide survey carried out in Italy, the gentamicin- and ciprofloxacin-resistant rates, among E. coli with CTX-M enzymes, were 62% and 19%, respectively [21].

Although carbapenems are stable to CTX-M-type enzymes, the emergence of resistance during carbapenem therapy has recently been reported in a K. pneumoniae strain producing the CTX-M-15 enzyme, following the loss of an outer-membrane porin [32]. Such mutations tend to be rapidly lost in the absence of a selective pressure; however, the possibility of selection of carbapenem-resistant mutants of ESBL-producing Enterobacteriaceae during carbapenem therapy is a matter of concern, due to the increasing prevalence of these enzymes and to the limited number of therapeutic alternatives.

Concluding remarks

The CTX-M-type ESBLs have recently undergone a rapid and global spread in Enterobacteriaceae. This CTX-M ‘epidemic’ is changing the epidemiology of ESBLs and, in several settings, the CTX-M-type ESBLs are now the most prevalent ESBLs encountered in E. coli and K. pneumoniae. The propensity of CTX-M-encoding genes to rapidly spread in E. coli is a matter of major concern, considering that this is one of the most common species isolated from clinical samples. Dissemination of the CTX-M-type ESBLs is not restricted to the nosocomial setting but also involves the community. This phenomenon is acting to modify the epidemiology of ESBLs, whereas those enzymes were, previously, mostly restricted to the nosocomial setting.

The CTX-M-type β-lactamases exhibit a remarkable allotypic diversity, and their genes occur in a notable variety of genetic contexts. Both these features, and especially the latter, could account for the rapid dissemination exhibited by these resistance determinants. Evolution of substrate specificity by point mutations is also possible, as with TEM- and SHV-type β-lactamases.

The spread of CTX-M-type ESBLs is causing rapid, important and unpredictable changes in the epidemiology of antibiotic resistance. It underlines the need for strict surveillance. In this regard, it should be emphasized that testing molecules such as cefotaxime, ceftriaxone and/or cefpodoxime is essential for laboratory detection of isolates producing CTX-M enzymes, and that testing ceftazidime as the sole representative of expanded-spectrum cephalosporins can lead to false-negative results, thereby underestimating the prevalence of ESBL producers.

Acknowledgements

The experimental work on ESBLs carried out in our laboratory was supported in part by research grants from the European Commission (LSHM-CT-2003-503335, COBRA Specific Targeted Research Project), the Italian Ministry of Research and University (MIUR, PRIN 2005), and Wyeth Pharmaceuticals. Owing to the limit on the number of references, it was not possible to cite all the original sources of the reviewed material. Consequently, recent review articles have mostly been cited for data that have already been referenced therein. The authors would like to acknowledge all the original scientific contributions relevant to the field that could not be cited owing to space limitations.

Ancillary