Molecular epidemiology of fosfomycin-resistant Escherichia coli from humans and animals

Abstract

The diminishing choice of effective antibiotics against resistant pathogens has forced clinicians to revive the use of old antibiotics. Hence, fosfomycin has been frequently suggested for alternative therapies given its track record of low resistance rates despite extensive use. However, there have been recent reports of plasmid-mediated fosfomycin resistance among animals and healthy humans in Asia. Accordingly, comparison of shared fosfomycin resistance mechanisms between animals and humans will shed light on the spread of resistance and guide future use of antimicrobials.

This study aimed to investigate plasmid-mediated fosfomycin resistance in E. coli isolates collected from patients and animals in Hong Kong. Non-duplicate E. coli isolates were cultured from 1711 urinary isolates and 167 blood clinical samples collected from multiple centres during 1996-2008 and from fecal samples of 210 cattle, 214 pigs, 460 chickens, 398 stray cats, 368 stray dogs and 456 wild rodents during 2008-2010.

A total of 2106 animal samples yielded 1693 E. coli isolates of which 831 were ESBL-producers. Fosfomycin-resistant isolates were more likely than fosfomycin-susceptible isolates to be ESBL-producers and multidrug resistant (≥3 antimicrobial classes). Of the 101 fosfomycin-resistant isolates, 97 (96.0%) were fosA3 positive and 94 (93.1%) were blaCTX-M positive. Of the 1878 clinical isolates, 18 were fosfomycin-resistant of which six were fosA3-positive and two were positive for another fosA variant (designated fosKP96). All but one fos-carrying clinical isolate was ESBL-producing. The majority of the fos-carrying E. coli strains belonged to diverse clones under two main clonal complexes CC58 accounting for 38 (36.2%) strains and CC10 for 32 (30.5%) strains.

PCR mapping showed that all fosA3-containing regions were followed by a downstream IS26. In all clinical isolates and 81 (83.5%) of animal isolates, an IS26 was also found upstream. In 14 (14.4%) of animal isolates, the fosA3 gene was preceded by an upstream blaCTX-M-14-containing transposon-like structure (ΔISEcp1-blaCTX-M-14-ΔIS903 or ISEcp1-IS10-blaCTX-M-14-ΔIS903). For the remaining two animal isolates, the upstream region could not be defined.

In a random subset of 18 animal isolates, fosA3 was carried on transferable plasmids with sizes of 50–200 kb and the following replicons: F2:A-:B- (n = 3), F16:A1:B- (n = 2), F24:A-B- (n = 1), IncHI2 (n = 3), IncN (n = 2), IncI1 (n = 1), B/O (n = 1) and untypeable (n = 5). Among six fosA3-carrying clinical isolates, the distributions were: F2:A-:B- (n = 2), IncN (n = 1), multi-replicon F-:A-:B1/IncN (n = 1) and untypeable (n = 2). Both fosKP96-carrying plasmids belonged to IncN. Restriction fragment length polymorphism analysis showed that the four F2:A-:B- plasmids carrying fosA3 and blaCTX-M-3 genes from a pig, dog and two patients shared an almost identical pattern. The complete plasmid sequences further demonstrated their homology.

This study demonstrated the emergence of fosA3-mediated fosfomycin resistance among E. coli isolates from multiple sources. Highly similar IncFII plasmids and IS26 transposon-like structures appear to be the main vehicles for dissemination. This study also highlighted occurrence of plasmids carrying fosKP96 that may have been overlooked by others as this variant could not be detected by established PCR markers.