Transfer of vanA from an Enterococcus faecium isolate of chicken origin to a CC17 E. faecium isolate in the intestine of cephalosporin-treated mice

Sir, Vancomycin-resistant Enterococcus faecium isolates can be detected in both human patients and animals but seem in general to separate into different clonal groups. E. faecium strains acquired during hospitalization and associated with nosocomial outbreaks and infections belong to clonal complex 17 (CC17), whereas poultry isolates group into CC9. The vanA gene encoding vancomycin is located on transposon Tn1546. Even though several studies have shown variations in Tn1546, similar Tn1546 transposons have been detected in E. faecium isolates of both human and animal origin.1 This indicates a horizontal transfer of Tn1546 between E. faecium belonging to different clonal groups. The aim of the present study was to investigate whether transfer of the vanA gene was possible from an isolate of animal origin to a CC17 E. faecium isolate of human clinical origin in the intestine of cephalosporin-treated mice. An E. faecium isolate of chicken origin harbouring the vanA gene was used as the donor strain.2 The isolate had multilocus sequence type 245 (ST245) and belongs to CC9. The recipient was a human E. faecium isolate from a bloodstream infection and belongs to CC17. The isolate was resistant to ampicillin and susceptible to vancomycin. Six female NMRI mice weighing 25–30 g were caged in pairs. From day 0 to day 4 the mice were treated subcutaneously twice a day with cefuroxime (3.0 mg/day). On day 1 the mice were fed orally using a steel feeding tube with a suspension of the donor strain (1.4 × 109 cfu in 200 µL of 0.9% saline). The mice were fed the recipient strain on day 2 (3.75 × 109 cfu in 200 µL of saline). The number of donors, recipients and transconjugants in faecal samples was measured 24 h after the first treatment of cefuroxime and on days 2, 3, 4, 5, 6, 7, 8, 9 and 11. The samples were collected from each cage; each sample contained faeces from both mice in the particular cage. The faecal samples were homogenized in saline and plated on Enterococcosel agar plates (BD Diagnostic Systems, Sparks, MD, USA) supplemented with 32 mg/L vancomycin and 32 mg/L erythromycin, 32 mg/L vancomycin and 16 mg/L ampicillin or 32 mg/L erythromycin and 16 mg/L ampicillin. The experimental protocol was approved by the Ethics Committee for Animal Experimentation (animal inspectorate approval no. 2007/561-1430). None of the mice was colonized with vancomycin-resistant E. faecium prior to inoculation of the donor strain. There was a large variation in the colonization of the mice among the three cages. This could be explained by differences in the colonization resistance among the mice. Transfer of the vanA gene from the donor strain of animal origin to the CC17 recipient occurred in the intestine of the mice in two of the three cages (Figure 1). Antimicrobial susceptibility testing showed that all in vivo transconjugants were resistant to vancomycin. PCR confirmed the presence of vanA in all transconjugants. Transconjugants were recovered between day 3 and day 8 with a maximum of 250 cfu/g of faeces in both cages (Figure 1b). Faecal excretion of the strains by the mice. (a) Donor strain (open symbols). (b) Recipient strain (open symbols) and transconjugants (filled symbols). Each curve shows the results for one cage; each cage is represented by one symbol. The conjugation occurred in the two cages where the colonization of the mice was the highest; the mice in the last cage were poorly colonized. This indicates that in order for transfer to occur high numbers of both donor and recipient must be present; this is in accordance with previous in vivo conjugation studies.2 Higher transfer rates have been obtained in gnotobiotic animals.3,4 Gnotobiotic animals have no colonization resistance whereas mice treated with antimicrobial agents still have some colonization resistance, even though it is decreased compared with untreated animals. We previously demonstrated transfer of the vanA gene cluster from an E. faecium isolate of animal origin to an E. faecium isolate of human origin in the intestine of humans.2 However, this study was performed without any antimicrobial treatment and the recipient strain BM4105-RF was ST172, which is not part of the hospital-acquired CC17. A high level of gastrointestinal colonization by enterococci is a frequent predecessor of serious enterococcal infection and this is promoted by the use of systemic or local antimicrobial agents. In this study the mice were treated subcutaneously with the second-generation cephalosporin cefuroxime. Other studies have shown that intestinal colonization of vancomycin-resistant enterococci is promoted by exposure to broad-spectrum antimicrobial agents that have poor anti-enterococcal activity.5 Cefuroxime has poor anti-enterococcal activity and is the cephalosporin of choice in Denmark. A large increase in the consumption of second-generation cephalosporins for systemic use in hospitals has been recorded in Denmark over the past 10 years.6 In conclusion, the present study showed that vanA of animal origin can be transferred to a CC17 recipient obtained from a patient with a bloodstream infection. To our knowledge, this is the first in vivo study showing a zoonotic transfer of the vanA gene to a CC17 E. faecium isolate. This study was supported by the Danish Ministry of the Interior and Health as part of the Danish Integrated Antimicrobial Resistance and Research Programme (DANMAP) and the European Union Sixth Framework Programme ‘Approaches to control multiresistant enterococci (ACE): studies on molecular ecology, horizontal gene transfer, fitness and prevention’ under contract LSHE-CT-2007-037410. None to declare. Karin Sixhøj Pedersen is thanked for technical assistance. Ulrich Stab Jensen is thanked for advice regarding the chosen cephalosporin for the study.