Enterococcus faecalis Clones in Poultry and in Humans with Urinary Tract Infections, Vietnam

Transmission routes and reservoirs need to be elucidated.

E nterococci are commensals of the human and animal gastrointestinal tract and opportunistic pathogens that cause urinary tract infections (UTIs), endocarditis, and sepsis (1). Nosocomial infections caused by enterococci have increased; these pathogens are now the third most common at hospitals after Escherichia coli and Staphylococcus aureus (2); and enterococci are frequently recorded as the cause of UTIs, wound infections, bacteremia, and endocarditis (3)(4)(5)(6).
The sources of enterococcal infections in humans are not clear, but animal reservoirs have been suggested (2,4,(7)(8)(9). A study comparing enterococcal isolates from 4 European countries and the United States demonstrated that E. faecalis isolated from pigs in Portugal had pulsedfi eld gel electrophoresis (PFGE) patterns identical to those of multidrug-resistant isolates at hospitals in Spain, Italy, and Portugal, all of which were shown by multilocus sequence typing (MLST) to belong to sequence type (ST) 6 (7). In Denmark, high-level gentamicin-resistant E. faecalis of ST16 with an identical PFGE pattern was isolated from pigs and from humans with endocarditis (9). Identical and closely related PFGE patterns were demonstrated by isolates from humans and from pork and chicken meat in the United States, all of which contained high-level gentamicin-resistant genes (4). Our objective was to characterize epidemiologically related E. faecalis isolated from humans with UTIs and from poultry living in the same households in Vietnam to evaluate the zoonotic potential of E. faecalis.

Recruitment of Patients, Urine Collection, and Bacterial Culture of Urine
Urine samples were collected during January 2008-January 2010 at the Military Medical University, Hospital 103, in Ha Dong, Hanoi. Patients with clinical symptoms of UTI (i.e., >1 of the following symptoms: frequent urination; painful urination; hematuria; cloudy urine; or pain in pelvic area, fl ank, or low back) were referred from nearby pharmacies and informed about the project. A midstream urine sample was collected at the hospital under supervision of a nurse. Only patients with uncomplicated UTIs were included; patients reporting underlying diseases, such as hematologic disorders, respiratory infections, diarrhea, diabetes, cancer, HIV/AIDS, liver cirrhosis, alcoholism, anatomic malformations of urinary tract, nephrolithiasis, or urolithiasis were excluded, as were patients with hospitalacquired UTIs. The urine was cultured immediately after collection. Thirty-one UTI patients met the study criteria of having E. faecalis CFU >10 3 /mL isolated from a urine sample in pure culture and were raising poultry in their households.
The urine samples were cultured on Flexicult agar plates (Statens Serum Institut, Copenhagen, Denmark), where E. faecalis grows as small green/blue-green colonies and E. faecium as small green colonies (10). Three colonies were isolated from each UTI patient. All 31 participants were interviewed when urine samples were collected. Personal information recorded included age, sex, and underlying diseases. The following clinical symptoms were recorded: frequent urination, painful urination, cloudy urine, blood in urine, pain in pelvic area, fl ank pain, pain in low back, and fever. In addition, information about duration of symptoms; previous UTIs; and medical treatment before arrival at the hospital, including type of antimicrobial drug used, was recorded.
Species identifi cation of all 31 presumptive E. faecalis isolates from urine and 83 isolates from poultry were confi rmed by species-specifi c PCR as described by Dutka-Malen et al. (11). Only isolates identifi ed as E. faecalis by PCR were further characterized.
All study participants were informed orally and in writing about the study and provided written consent. The ethics committee at Army Hospital 103 approved the study protocols.

Collection of Cloacal Swabs from Poultry
When a urine sample was positive for E. faecalis, the patient's household was visited within 1 week, and cloacal swabs were taken from 2-4 chickens in the household. Fecal samples were taken with a sterile cotton swab and immediately placed in Cary-Blair media (Oxoid, Basingstoke, Hampshire, UK) for transportation to the laboratory. Samples were then streaked on Slanetz and Bartley agar medium (Merck, Darmstadt, Germany) the same day and incubated for 24-48 h at 37°C. Subsequently, 2 individual colonies were randomly selected and subcultured on nonselective LB-agar, Lennox plates (Difco, Becton Dickinson, Sparks, MD, USA), which were incubated overnight at 37°C to obtain pure cultures. Colonies were then grown in brain-heart infusion broth (Oxoid) overnight at 37°C and stored for further characterization at -80°C in cryotubes containing 30% glycerol.

MLST and PFGE
To investigate whether isolates of E. faecalis from urine and poultry belonged to identical STs, we characterized isolates from urine and poultry by MLST. Urine isolates were characterized by sequencing of all 7 housekeeping genes used in the MLST scheme: gdh, gyd, pstS, gki, aroE, xpt, and yqil. To confi rm that the UTIs were caused by a single strain, 1 additional colony from 9 (29%) of 31 urine samples was characterized by sequencing the gki and yqil genes. Two isolates from each chicken were characterized by sequencing the gki and yqil genes. When sequences of both genes in 2 isolates corresponded to the sequence of the same genes in the urine isolate, which occurred in 11 cases, 1 of the 2 isolates from poultry was randomly selected and further characterized. When gene sequences in only 1 isolate from poultry were identical to the isolate from urine, the isolate was further characterized. Primers and PCR conditions are described on the E. faecalis MLST website (http://efaecalis.mlst.net/). Amplicons were sequenced in both directions by Macrogen (Seoul, South Korea). DNA sequences obtained were assembled using CLC Main Workbench 5.2 software (CLC bio, Aarhus, Denmark) and compared with published alleles, and an ST was assigned to each strain (http://efaecalis.mlst.net/). PFGE was performed as described (12) by using the restriction enzyme smaI (New England BioLabs, Ipswich, MA, USA).

Virulence Genes
The presence and sequence of the following 6 virulence genes were used to further characterize the isolates from urine and poultry: asa1, CylA, efaA, Esp, gelE, and EF0591 (13). After detecting the virulence genes by PCR (13), we sequenced the genes in both directions using Macrogen. DNA sequences were compared, and possible nucleotide differences were calculated by using Smith-Waterman local alignment (EMBOSS) available online from the European Bioinformatics Institute: (www.ebi.ac.uk/).

Results
In 7 (23%) of 31 UTI cases, E. faecalis isolated from patient urine and poultry demonstrated identical STs and an indistinguishable (4 pairs) or closely related PFGE pattern (3 pairs, defi ned as showing <3 fragment difference) (Figure).
In addition, antimicrobial drug susceptibility patterns were similar, and only 1 variation was found in the virulence gene profi les (Tables 1, 2). Five of these 7 patients reportedly had a profession where they worked with poultry. A total of 22 patients who did not share a clone of E. faecalis found in poultry in their household reported working with poultry.  Figure).

Antimicrobial Drug Susceptibility Testing
When we compared isolates from urine and poultry from individual households, we detected similar MICs of each tested antimicrobial drug, showing a 1-dilution factor deviation ( Table 2). For several isolates, an MIC could not be established because the MIC fell outside the test intervals. We detected different MICs for 7 antimicrobial drugs when we compared strains 204U and 204P. All isolates were fully susceptible (lowest or second lowest MIC tested) to ampicillin, avilamycin, linezolid, penicillin, salinomycin, tigecycline, and vancomycin (results not shown in Table 2).

Virulence Genes
PCR for the 6 virulence genes showed that the isolates from urine and poultry from an individual household contained identical virulence genes that varied from 1 to 5 genes, except for 1 household in which the isolate from urine (90U) did not contain the asa1 gene (Table 1). When we compared the DNA sequences from the epidemiologically related urine and poultry strains, we found that all 23 sequenced gene pairs showed 100% similarity.

Discussion
We document isolation of the same clone of E. faecalis in urine and poultry from the same households in which patients had close contact with the poultry. The potential for zoonotic transmission of E. faecalis has been suggested, but to our knowledge, only epidemiologically unrelated isolates have been investigated (3,4,(7)(8)(9)14).
Most of the isolates in our study belonged to ST16, which has been isolated from animals and humans, including clinical and nonclinical isolates (14). ST93 was isolated from a patient with an ulcer in Poland and from an unknown source in the United States, and ST141 was isolated from chickens in Denmark and from a blood sample of a hospitalized person in Poland (http://efaecalis.mlst.net/). When we interpreted PFGE patterns for their relatedness using criteria suggested by Tenover et al. (15), we found 4 pairs of E. faecalis strains with indistinguishable band patterns that could be "considered to represent the same strain" (15). From 3 individual households, isolates from urine and poultry showed PFGE patterns with 1 or 2 band differences and thus can be considered closely related (15). These identical or closely related PFGE patterns, together with the supporting fi ndings by MLST and virulence gene profi ling, suggest that E. faecalis might be transmitted from poultry to humans, causing UTIs. However, the fi nding of similar isolates from humans and poultry also could result from sharing a common clone of E. faecalis. ST16 has been reported from various epidemiologically unrelated human and animal sources (14), which could indicate a common clone in humans and animals. Because no data about ST16 in the environment are available, an environmental reservoir cannot be ruled out. Because 27 of the 31 patients reported having contact with poultry through their work, contact with poultry outside the household environment cannot be excluded as the source of E. faecalis. Epidemiologic risk factor studies are needed to document actual transmission routes.
The variation found in resistance patterns might have resulted from exposure to different antimicrobial drugs, resulting in different selection pressure on E. faecalis in the human and poultry hosts. The 7 patients studied had UTI symptoms for an average of 514 days (range 5 days-10 years), which is unusually long for UTI (Table  1). Although self-medication is well established to be a common practice in Vietnam (16), only 2 of the 7 patients acknowledged use of antimicrobial drugs to treat their UTI symptoms before they participated in the study (data not shown). Over time, patients tend to forget what kind of medication they received. Furthermore, the questionnaire asked only whether antimicrobial drugs were used against UTI, not whether they were used to treat other diseases. In addition, poultry might have been exposed to antimicrobial drugs through growth promoters added in the feedstuff and during therapeutic or preventive treatments, but information about such use was not available.
In most Western countries, contact with poultry occurs mainly through handling and consumption of poultry meat. However, the risk for zoonotic transmission of E. faecalis from poultry meat remains to be investigated. Thus, similar studies and risk factor studies should be conducted in more countries to evaluate the effect on zoonotic transmission of differences in human habits of poultry consumption and contact with poultry. In addition, animals other than pigs and poultry should be investigated as sources of zoonotic E. faecalis transmission. Finally, we cannot exclude the possibility that E. faecalis pathotypes found in poultry might represent transmission from humans, e.g., in this study, from UTI patients. However, poultry as carriers of  ST16 has been documented (17), and it seems more likely that humans are exposed to poultry litter than that poultry are exposed to human feces. We did not investigate the route of E. faecalis transmission, but the route could be colonization of the human intestine and subsequently ascending the urethra as reported for E. coli (18). Further studies are required to explain routes of transmission. The emergence of enterococci as causes of human infections and their resistance to some of the crucial antimicrobial drugs used for human treatment emphasizes the need to elucidate transmission routes and reservoirs for the enterococci and their resistance genes (5,6,(19)(20)(21).