Fluoroquinolone-Resistant Escherichia coli and Proteus mirabilis in Poultry of Middle Moravia , Czech Republic

Koláfi M., J . BardoÀ, P. Sauer , M. Kesselová, L. âekanová, I . Vágnerová, D. Koukalová, P. Hejnar : Fluoroquinolone-Resistant Escherichia coli and Proteus mirabilis in Poultry of Middle Moravia, Czech Republic. Acta Vet. Brno 2005, 74: 249-253. The aim of the study was to detect antibiotic resistance of Escherichia coli and Proteus mirabilis isolates from farm-reared poultry. During the period of June 2003 to June 2004, samples of cloacal swabs obtained from 5 poultry farms located in the central part of Moravia, Czech Republic were examined using aerobic cultivation and afterwards E. coli and P. mirabilis isolates were identified. Minimal inhibitory concentrations of antibiotics tested were determined by a microdilution method. Pulsed-field gel electrophoresis (PFGE) was performed using SmaI restriction endonuclease. A total of 300 samples of cloacal swabs from healthy layer hens were cultivated and 239 E. coli and 127 P. mirabilis strains were isolated. In the case of E. coli, 7 isolates (3%) from 4 different farms were identified manifesting resistance to ofloxacin and ciprofloxacin. Out of 24 P. mirabilis isolates resistant to the tested fluoroquinolones, 20 strains were isolated from one of the farms and PFGE analysis of DNA proved that 19 isolates were probably identical and represented one clonal type. The study confirmed the occurrence of multiresistant bacterial isolates with resistance to fluoroquinolones in poultry in the Czech Republic. Their clonal spread in farm-reared poultry can be suggested, too. Poultry, faecal bacteria, ofloxacin, ciprofloxacin, resistance Nowadays, increasing bacterial resistance to antibiotic agents including fluoroquinolones poses a serious problem (Jones et al. 1997; Neu 1992). This adverse trend is documented both in human and veterinary medicine and apparently its solution is to be seen in the synergism of preventive measures (Capriol i et al. 2000; Bogaard and Stobberingh 2000). The increasing bacterial resistance to fluoroquinolones both in human and animal populations is evidently connected with their overuse (Aguiar et al. 1992; Garau et al. 1999; Pena et al. 1995). In veterinary medicine, fluoroquinolones are used in the treatment of both farm and pet animals. According to the data provided by the Institute for the State Control of Veterinary Biologicals and Medicaments, the overall usage of quinolones in Czech veterinary medicine in 2003 represented 1532.9 kg (Hera 2005). Many fluoroquinolones used in veterinary medicine belong to “antibiotic agents with indication limits”, a group of antibiotics that should be applied in serious animal infections only, based on clinical experience, diagnostic confirmation of the causal microorganism and resistance to “common” antimicrobial agents. The aim of the study was to detect resistance to antimicrobial agents in Escherichia coli and Proteus mirabilis isolates from poultry bred in the central part of Moravia (Czech Republic), a region with extensive agricultural production. In fluoroquinolone-resistant ACTA VET. BRNO 2005, 74: 249–253 Address for correspondence: Doc. MUDr. M. Koláfi, Ph.D. Institute of Microbiology, Faculty of Medicine, Palack ̆ University Hnûvotínská 3 775 15 Olomouc, Czech Republic Tel: +420 585 632 407 Fax: +420 585 632 417 E-mail: kolar@fnol.cz http://www.vfu.cz/acta-vet/actavet.htm isolates, the analysis of genomic DNA was performed and the degree of their similarity determined. Materials and Methods Collect ion of isolates During the period of June 2003 to June 2004, samples of cloacal swabs from 5 poultry farms were examined. The farms were located in the central part of Moravia, Czech Republic. The parental breeds produced the final crossbreed of Brown Hissex. In each poultry farm, a total of 60 samples of cloacal swabs were taken from 60 healthy layer hens, i.e. 1 sample per hen. Each clinical sample was cultivated under aerobic conditions on conventional selective media (Endo agar, Xylose-Lysine-Desoxycholate agar). E. coli and P. mirabilis isolates were determined by standard biochemical procedures using Enterotest 24 (Pliva Lachema CZ). Antibiot ic suscept ibi l i ty tes t ing Minimal inhibitory concentrations (MICs) of antibiotics tested for E. coli and P. mirabilis isolates were determined by a microdilution method according to the National Committee for Clinical Laboratory Standards guidelines (NCCLS 2000). Concentrations of tested isolates in Mueller-Hinton broth (Oxoid UK) were prepared to obtain 0.5 McFarland turbidity. MICs were read after 18 h of incubation at 37 °C. The MIC was interpreted as the lowest concentration of the antibiotic that visibly inhibited bacterial growth. As MIC breakpoints, the following values based on the National Institute of Public Health’s National Reference Laboratory for Antibiotics recommendations (Urbá‰ková 1998) were used: 1 mg/l for ciprofloxacin, 2 mg/l for ofloxacin and tetracycline, 4 mg/l for ampicillin, cefazolin, cefuroxime, cefotaxime, ceftazidime, chloramphenicol, gentamicin, colistin and meropenem, 8 mg/l for amikacin, oxolinic acid, trimethoprim and ampicillin/sulbactam, and 32 mg/l for trimethoprim-sulfamethoxazole and nitrofurantoin. MIC90 and MIC50 include 90% and 50% isolates with the given or lower MIC value of the respective antibiotics in the appropriate bacterial species. Reference strains E. coli ATCC 25922, E. coli ATCC 35218 and Pseudomonas aeruginosa ATCC 27853 were used for protocol quality control. Molecular analysis of f luoroquinolone-resis tant isolates Pulsed-field gel electrophoresis (PFGE) was performed using SmaI restriction endonuclease. Genomic DNA isolation was carried out using protocols published by PantÛãek et al. (1997). Restriction cleavage of the blocks 1 × 1 × 5 mm was done in restriction solution [8 μl of restriction buffer for SmaI (Sigma-Aldrich), 65 μl of deionised water and 10 U of restriction enzyme SmaI]. The blocks were incubated at 25 °C overnight. PFGE was prepared in 1.2% agarose gel (Amresco) in 1xTBE buffer. PFGE was carried out at BioRad CHEFDR II for 28 h at a pulse time of 0.1 to 30 s, 5.0 V/cm, with angle 120°. The gel was stained in ethidium bromide solution (Sigma-Aldrich) (1 μg/ml). The data about administration of antibiotics in observed poultry farms are not available.

Nowadays, increasing bacterial resistance to antibiotic agents including fluoroquinolones poses a serious problem (Jones et al. 1997;Neu 1992).This adverse trend is documented both in human and veterinary medicine and apparently its solution is to be seen in the synergism of preventive measures (Caprioli et al. 2000;Bo g aar d and Stobberingh 2000).The increasing bacterial resistance to fluoroquinolones both in human and animal populations is evidently connected with their overuse (Aguiar et al. 1992;Garau et al. 1999;Pen a et al. 1995).
In veterinary medicine, fluoroquinolones are used in the treatment of both farm and pet animals.According to the data provided by the Institute for the State Control of Veterinary Biologicals and Medicaments, the overall usage of quinolones in Czech veterinary medicine in 2003 represented 1532.9 kg (Hera 2005).Many fluoroquinolones used in veterinary medicine belong to "antibiotic agents with indication limits", a group of antibiotics that should be applied in serious animal infections only, based on clinical experience, diagnostic confirmation of the causal microorganism and resistance to "common" antimicrobial agents.
The aim of the study was to detect resistance to antimicrobial agents in Escherichia coli and Proteus mirabilis isolates from poultry bred in the central part of Moravia (Czech Republic), a region with extensive agricultural production.In fluoroquinolone-resistant isolates, the analysis of genomic DNA was performed and the degree of their similarity determined.

Collection of isolates
During the period of June 2003 to June 2004, samples of cloacal swabs from 5 poultry farms were examined.The farms were located in the central part of Moravia, Czech Republic.The parental breeds produced the final crossbreed of Brown Hissex.In each poultry farm, a total of 60 samples of cloacal swabs were taken from 60 healthy layer hens, i.e. 1 sample per hen.Each clinical sample was cultivated under aerobic conditions on conventional selective media (Endo agar, Xylose-Lysine-Desoxycholate agar).E. coli and P. mirabilis isolates were determined by standard biochemical procedures using Enterotest 24 (Pliva Lachema CZ).

Antibiotic susceptibility testing
Minimal inhibitory concentrations (MICs) of antibiotics tested for E. coli and P. mirabilis isolates were determined by a microdilution method according to the National Committee for Clinical Laboratory Standards guidelines (NCCLS 2000).Concentrations of tested isolates in Mueller-Hinton broth (Oxoid UK) were prepared to obtain 0.5 McFarland turbidity.MICs were read after 18 h of incubation at 37 °C.The MIC was interpreted as the lowest concentration of the antibiotic that visibly inhibited bacterial growth.As MIC breakpoints, the following values based on the National Institute of Public Health's National Reference Laboratory for Antibiotics recommendations (U r b á ‰ k o v á 1998) were used: 1 mg/l for ciprofloxacin, 2 mg/l for ofloxacin and tetracycline, 4 mg/l for ampicillin, cefazolin, cefuroxime, cefotaxime, ceftazidime, chloramphenicol, gentamicin, colistin and meropenem, 8 mg/l for amikacin, oxolinic acid, trimethoprim and ampicillin/sulbactam, and 32 mg/l for trimethoprim-sulfamethoxazole and nitrofurantoin.MIC 90 and MIC 50 include 90% and 50% isolates with the given or lower MIC value of the respective antibiotics in the appropriate bacterial species.
Reference strains E. coli ATCC 25922, E. coli ATCC 35218 and Pseudomonas aeruginosa ATCC 27853 were used for protocol quality control.

Molecular analysis of fluoroquinolone-resistant isolates
Pulsed-field gel electrophoresis (PFGE) was performed using SmaI restriction endonuclease.Genomic DNA isolation was carried out using protocols published by PantÛãek et al. (1997).
Restriction cleavage of the blocks 1 × 1 × 5 mm was done in restriction solution [8 µl of restriction buffer for SmaI (Sigma-Aldrich), 65 µl of deionised water and 10 U of restriction enzyme SmaI].The blocks were incubated at 25 °C overnight.
PFGE was prepared in 1.2% agarose gel (Amresco) in 1xTBE buffer.PFGE was carried out at BioRad CHEF-DR II for 28 h at a pulse time of 0.1 to 30 s, 5.0 V/cm, with angle 120°.The gel was stained in ethidium bromide solution (Sigma-Aldrich) (1 µg/ml).
The data about administration of antibiotics in observed poultry farms are not available.

Results
In total, 300 samples of cloacal swabs from poultry (parental breed of Brown Hissex) were cultivated and 239 E. coli and 127 P. mirabilis isolates were identifiety.Antibiotic resistance of the investigated species is given in Table 1.A higher rate of resistant strains in the case of E. coli was proved in tetracycline (48%) and ampicillin (29%).As for the P. mirabilis strains, higher resistance to trimethoprim (31%) and trimethoprim-sulfamethoxazole (28%) was documented, apart from their natural resistance to colistin, nitrofurantoin and tetracycline.
Among E. coli isolates, 7 (3%) resistant to ofloxacin and ciprofloxacin were isolated from 4 poultry farms, while 4 isolates were detected in farm IV.Out of the 24 P. mirabilis isolates resistant to the tested fluoroquinolones, 20 were isolated from farm IV (Table 2).
In E. coli, 5 other isolates were detected, resistant to oxolinic acid (MICs ≥ 16 mg/l), but with persisting sensitivity to ofloxacin and ciprofloxacin.However, MICs of these isolates reached higher values (1-2 mg/l in ofloxacin and 0.5 -1 mg/l in ciprofloxacin).
Fluoroquinolone-resistant isolates of E. coli were also resistant to tetracycline, trimethoprim-sulfamethoxazole and trimethoprim; 4 isolates out of 7 were resistant even to gentamicin.P. mirabilis isolates resistant to fluoroquinolones were also resistant to trimethoprim-sulfamethoxazole, trimethoprim and, in 67%, to gentamicin.
In total, 20 fluoroquinolone-resistant isolates of P. mirabilis isolated from the same farm were analyzed by PFGE.The results are shown in Fig. 1 (see Plate XIII).Of these, 19 isolates were identical and represented one clonal type.A unique isolate shows 60% similarity of a DNA profile in SmaI spectrum with other probably identical isolates.

Discussion
The relation between the application of antibiotics and the dissemination of bacterial resistance from animals to humans has been described by Hummel et al. (1996).Also, the study by van den Bo g aar d et al. (2001) confirmed that transmission of resistant clones and resistance plasmids of E. coli from poultry to humans commonly occurs.
E. coli is a major pathogen of worldwide importance in commercially produced poultry (Gross 1994).Although this bacterial species is commonly found in the intestinal tract of 251 hens, some serotypes can cause diseases.Antibiotic therapy is therefore an important tool for reducing mortality associated with E. coli infections (Dho-Moulin 1993;Gonzáles et al. 1990).On the other hand, the use of antibiotics can select multiresistant bacteria.
A study performed in Spain showed the resistance of avian E. coli isolates to ofloxacin and ciprofloxacin in 14% and 13%, respectively (Blanco et al. 1997).In the present study, the corresponding value reached 3%.In P. mirabilis isolates, the frequency of fluoroquinoloneresistant isolates was much higher.However, it must be stressed that 83% of the isolates were obtained from one farm, suggesting that these probably originated from one resistant clone.Thus the average resistance level cannot be determined.
Isolates of both species resistant to fluoroquinolones were multi-resistant, manifesting resistance to trimethoprim-sulfamethoxazole, trimethoprim and, in the case of E. coli isolates, even to tetracycline.Higher occurrence of gentamicin-resistant isolates was also notified.Similar results were reported by Bazile-Pham- Khac et al. (1996) who described simultaneous resistance to tetracycline, trimethoprim and tobramycin in fluoroquinolone-resistant E. coli isolates.It is to be noted that in E. coli five other isolates were identified, manifesting resistance to oxolinic acid.Ofloxacin and ciprofloxacin MIC ranges were 1 -2 mg/l and 0.5 -l mg/l, respectively, and, according to the breakpoints applied, these isolates were sensitive to fluoroquinolones.These are probably isolates with single mutations in gyrA, i.e. with decreased sensitivity (McDonald et al 2001).This finding should be the first sign of the negative phenomenon of increasing resistance to fluoroquinolones.Therefore, sensitivity testing of gram-negative rods isolated from poultry should involve not only fluoroquinolones, such as ofloxacin or ciprofloxacin, but oxolinic acid as well.
Based on the analysis of the genomic DNA in P. mirabilis isolates resistant to fluoroquinolones and isolated from one of the farms, it is possible to assume that it is most probably one identical isolate spread in the farm.
In conclusion, the results of our study support the occurrence of multiresistant bacterial isolates with resistance to fluoroquinolones in poultry in the Czech Republic and underline the importance of antibiotic policy implementation in veterinary medicine, including monitoring bacterial isolates with dangerous phenotypes of resistance.
Proteus mirabilis isolates 22 marker Sigma Pulse Marker 50-1000 kB in lanes 1 and 22 was used

Table 1 .
Resistance of Escherichia coli and Proteus mirabilis to antibiotics with MIC range, MIC 50 and MIC 90