Emergence of extended-spectrum-AmpC-expressing Escherichia coli isolates in Belgian hospitals

Sir, The overproduction of AmpC chromosomal cephalosporinases with broadened substrate activity has recently been reported in clinical isolates of Escherichia coli.1 These so-called extended-spectrum AmpC (ESAC) β-lactamases confer reduced susceptibility to all cephalosporins, including the fourth-generation agents cefepime and cefpirome. We aimed to investigate the possible occurrence of ESAC-expressing E. coli isolates in Belgium. Among the 6850 non-duplicate E. coli clinical isolates (one isolate/patient) that were recovered from two Belgian hospitals (Erasme Hospital, Brussels and UCL de Mont-Godinne Hospital, Yvoir) between 2004 and 2006, 83 were found to display a β-lactam resistance pattern consistent with AmpC overproduction on the basis of resistance to amoxicillin, amoxicillin/clavulanic acid, cefazolin and cefoxitin by the disc diffusion method using the CLSI interpretative guidelines.2 The presence of AmpC was further confirmed by analytical isoelectric focusing demonstrating a band at a pI of ∼8.5–9.0, disappearing in the presence of oxacillin. Fourteen out of the 83 AmpC hyper-producing isolates (Table 1) were resistant to ceftazidime and cefotaxime by current EUCAST breakpoints3 and intermediate to cefepime (MIC above 1 mg/L; 1.5–8 mg/L). Sequence analysis of their chromosomal ampC genes revealed that eight of the isolates harboured an S287N mutation, while two harboured an H296P mutation, one harboured a V298L mutation and one harboured an L293P mutation. All these mutations, except the L293P mutation that is reported here for the first time, have already been reported in E. coli and were shown to be involved in the ESAC phenotype.4,5 Phenotypic and genotypic characterization of AmpC-overproducing E. coli isolates recovered at two Belgian hospitals ESAC, extended-spectrum AmpC; CAZ, ceftazidime; CTX, cefotaxime; FEP, cefepime; FOX, cefoxitin; MEM, meropenem; NA, not analysed. aThe reduced susceptibility to cefepime could be ascribed to the presence of a blaOXA-30 β-lactamase detected in this isolate. bEC WT AmpC is an E. coli clinical isolate hyper-producing a wild-type AmpC presenting exactly the same sequence as ESAC (L293P) except for the absence of a mutation at amino acid L293. Phenotypic and genotypic characterization of AmpC-overproducing E. coli isolates recovered at two Belgian hospitals ESAC, extended-spectrum AmpC; CAZ, ceftazidime; CTX, cefotaxime; FEP, cefepime; FOX, cefoxitin; MEM, meropenem; NA, not analysed. aThe reduced susceptibility to cefepime could be ascribed to the presence of a blaOXA-30 β-lactamase detected in this isolate. bEC WT AmpC is an E. coli clinical isolate hyper-producing a wild-type AmpC presenting exactly the same sequence as ESAC (L293P) except for the absence of a mutation at amino acid L293. No mutation was found in the chromosomal AmpC of the two remaining isolates. Instead, ISEcp1-blaCMY-2 (plasmid-mediated ampC gene) was detected in those E. coli. In one of the isolates (MIC of cefepime=4 mg/L), the reduced susceptibility to cefepime could be ascribed to the presence of a blaOXA-30 β-lactamase gene, which was identified by PCR/sequencing and isoelectric focusing (pI = 7.3, not susceptible to clavulanate and oxacillin). A TEM-1-β-lactamase-encoding gene alone was evidenced in the last isolate (MIC of cefepime=1.5 mg/L). In 69 out of the 83 isolates with cefepime MICs ≤1 mg/L, no single mutation assigned to the ESAC phenotype could be detected in the chromosomal AmpC. Among the 12 isolates expressing ESAC β-lactamases, 11 strains (9 of which were recovered from urine or stool samples) belonged to phylogenetic group A, a group usually associated with the commensal non-virulent flora, and one isolate (from a stool sample) belonged to phylogroup D, a group usually associated with virulent and extra-intestinal strains.6 Ten out of 12 of the ESAC-expressing isolates were closely related by PFGE (>80%) and rep-PCR (DiversiLab™ System, bioMérieux, France) analysis (>95%) while the ESAC-expressing isolate EC03, belonging to phylogroup D, harboured a different rep-PCR and PFGE profile sharing <70% and 60%, respectively, with the previous isolates. Isolate EC12, belonging to phylogroup A, was not related to the other ESAC-producing isolates. The putative new ESAC-β-lactamase-encoding gene (accession number FJ439686) was cloned in E. coli TOP10 using pCR®-Blunt II-TOPO® plasmid as vector (Invitrogen®, Merelbeke, Belgium) (Table 1). AmpC from E. coli EC WT, expressing a wild-type AmpC with exactly the same sequence as L293P ESAC except for the absence of a mutation at amino acid L293, was used as a control. ESAC transformant L293P TOP10 was resistant to cefepime while transformant WT TOP10 remained susceptible to this cephalosporin. Since the AmpC sequences of L293P TOP10 and WT TOP10 only differed by the L293P amino acid substitution located inside the R2 loop region, already shown to be involved in the ESAC phenotype in Enterobacteriaceae,1 this strongly suggested that this mutation was responsible for the resistance to cefepime. This study represents the first report of ESAC-expressing E. coli outside of France. As reported by Mammeri et al.,5 our study confirms the occurrence of ESAC isolates among AmpC-overproducing E. coli isolates in Belgian hospitals (12 out of 83, i.e. 14.5%). In contrast to previous reports, two different molecular typing methods showed a consistent clonal relatedness between the vast majority of the ESAC isolates, most of which belong to phylogroup A. It is possible that the selection of specific commensal endemic clones of chromosomal AmpC hyper-producers might be the reflection of similar antibiotic policies that obviously favour the clinical usage of β-lactam/β-lactamase inhibitor associations and of expanded-spectrum cephalosporins. In any case, this observation underlines the need to further investigate the epidemiology of AmpC-producing E. coli on a wider scale at a national level, as well as to evaluate the clinical relevance of this mechanism of resistance. This work was funded by a grant from the Belgian Antibiotic Policy Coordination Committee and by a grant from the Belgian Fond National de la Recherche Scientifique Médicale (FRSM 3.4522.09) to UCL Mont-Godinne/ULB Erasme. None to declare.