To determine whether therapeutic treatment of pigs with chlortetracycline affects the susceptibility of their Campylobacter isolates for tetracycline, ciprofloxacin and erythromycin.
To determine whether therapeutic treatment of pigs with chlortetracycline affects the susceptibility of their Campylobacter isolates for tetracycline, ciprofloxacin and erythromycin.
Minimum inhibitory concentrations (MICs) and presence of a tetracycline resistance gene tet(O) were studied in Campylobacter collected before, during and after chlortetracycline treatment. Tetracycline MICs and the presence of tet(O) for additional Campylobacter coli isolates collected previously from seven farrowing farms were also determined. Isolates with ciprofloxacin MICs above the epidemiological cut-off value (ECOFF) were subtyped by flaA restriction fragment length polymorphism (RFLP). Tetracycline MICs of 221 Camp. coli isolates remained under the ECOFF at all sampling stages as well as the MICs for 63 isolates from the other farms. The ciprofloxacin MIC was above the ECOFF for 22% of the isolates, and one Camp. coli isolate had an erythromycin MIC above the ECOFF. None of the studied 300 Campylobacter isolates from nine herds carried tet(O). flaA-RFLP typing revealed the heterogeneity of Camp. coli isolates with high ciprofloxacin MICs.
Use of chlortetracycline did not increase the MIC values for the antimicrobials studied.
This study demonstrated that susceptibility of Camp. coli isolates is not affected by chlortetracycline therapy if tet(O) is not present in Camp. coli population.
Intestinal bacterial populations encounter selection pressure when oral antimicrobials are administered. Some commensal bacteria in animals, such as Campylobacter, are also major foodborne pathogens in many countries; some populations may be rendered resistant during the course of antimicrobial therapy of food production animals (Inglis et al. 2005; Juntunen et al. 2011). Campylobacter jejuni is the dominant human pathogen (EFSA/ECDC 2012b), but Campylobacter coli is more frequently resistant to ciprofloxacin, erythromycin and tetracycline (de Jong et al. 2009, 2012). If antimicrobial therapy is needed in severe campylobacteriosis, erythromycin and ciprofloxacin are commonly prescribed antimicrobial agents; tetracycline may be considered as an alternative, but frequently detected resistance is a limiting factor (Fitzgerald et al. 2008).
Extensive use of antimicrobials in food animals is a public health concern. Tetracycline resistance has been reported commonly in Camp. coli isolates from various animal sources (Qin et al. 2011; Schweitzer et al. 2011; Quintana-Hayashi and Thakur 2012; Scott et al. 2012). A high percentage of tetracycline-resistant Camp. coli has been reported even amongst pigs raised in antimicrobial-free production systems (Rollo et al. 2010; Tadesse et al. 2011; Quintana-Hayashi and Thakur 2012). Additionally, subtherapeutic chlortetracycline treatment of calves was shown to increase the recovery rate of erythromycin-resistant Campylobacter hyointestinalis (Inglis et al. 2005).
Bacteria may obtain resistance after target mutations or horizontal gene transfer (Canton and Morosini 2011). Tetracycline resistance in Campylobacter is mediated by the ribosomal protection protein, Tet(O), encoded either chromosomally or in transferable plasmids (Taylor and Chau 1996; Dasti et al. 2007). Other commensal bacteria, for example, Enterococcus spp., may also harbour the tet(O) gene (Fairchild et al. 2005). In addition, a nonspecific efflux pump mechanism (CmeACB) may decrease susceptibility to tetracycline (Gibreel et al. 2007).
Tetracycline resistance has been monitored in porcine Camp. coli, but longitudinal studies of antimicrobial resistance in pigs subsequent to chlortetracycline treatment have not been published. Therefore, we selected two finishing herds in which pigs were therapeutically treated with chlortetracycline, and we determined minimum inhibitory concentrations (MICs) for Camp. coli isolates before, during and after treatment. Prevalence of the tet(O) gene in nine herds was investigated. Furthermore, the presence of tet(O) in Enterococcus isolates was analysed from the same faecal samples, and Camp. coli isolates with high ciprofloxacin MICs were subtyped by flaA restriction fragment length polymorphism to analyse the heterogeneity of the isolates.
Faecal samples were collected from ear-tagged finishing pigs (age 3–5 months) suffering from respiratory tract infections in two Finnish herds (herd 1: 420 finishers; herd 2: 1575 finishers) between November 2011 and February 2012. All pigs from herd 1 were medicated, and 33 pigs were sampled at three time points. All of herd 2, except for the pigs from two pens, received chlortetracycline treatment; 20 treated and 20 untreated pigs were sampled at three time points. Local veterinarians prescribed oral chlortetracycline (Aurofac vet 100 mg g−1; Scanvet Eläinlääkkeet Oy, Parola, Finland), which was mixed in feed (35 mg of chlortetracycline per kg of body weight) for a 10-day period. The first samples were collected on the day before the start of the antimicrobial medication. The second set of samples was taken on the 6th or 7th day of chlortetracycline therapy, and the third set was completed 22–24 days after cessation of the medication (33 and 35 days after the first sampling). Samples were transported as previously described (Juntunen et al. 2010). The following antimicrobial agents were used to treat other infections of individual pigs where indicated: procaine benzylpenicillin (Ethacilin vet 300 000 IU ml−1; Intervet International B.V., Boxmeer, the Netherlands) in both herds and injectable oxytetracycline (Terramycin vet 100 mg ml−1; Pfizer Animal Oy Health, Helsinki, Finland) and lincomycin (Lincocin vet 100 mg ml−1; Pfizer Oy Animal Health) in herd 1 only. The pigs included in this study had not previously been medicated in the finishing herds. Untreated pigs and sows from seven farrowing farms were previously sampled between 2007 and 2009 (Juntunen et al. 2012). These pre-existing isolates were used in the present study to obtain more isolates for the investigation of tet(O) gene prevalence.
Faecal samples were cultured on modified charcoal cefoperazone deoxycholate agar (mCCDA) (CM0739; Oxoid Ltd., Basingstoke, Hampshire, UK) that included the selective supplement (SR155; Oxoid Ltd.). Isolates were grown microaerobically (5%O2, 10%CO2 and 85%N2) at 37°C. Two presumptive isolates per pig from each sampling time point were cultured on Nutrient agar (CM0003; Oxoid Ltd.) containing 5% blood, where Camp. coli and Camp. jejuni were differentiated as previously described (Juntunen et al. 2010). Campylobacter isolates were stored in Mueller–Hinton broth (CM0405; Oxoid Ltd.) with glycerol at −70°C.
Enterococcus spp. were isolated from 17 and 13 faecal samples collected from herd 2 before and during chlortetracycline treatment, respectively. Faecal swabs were spread on Slanetz–Bartley agar (01-178; Scharlau Chemie S.A., Barcelona, Spain) containing 2 μg ml−1 tetracycline and incubated at 37°C for 48 h. Typical Enterococcus isolates growing on selective media were subcultured, and DNA was extracted as previously reported (Juntunen et al. 2010), and Enterocccus isolates were confirmed by genus-specific primers (Deasy et al. 2000).
The antimicrobial susceptibility of Campylobacter isolates to ciprofloxacin, erythromycin and tetracycline was determined by the broth microdilution method (CLSI 2008). Briefly, isolates were grown in cation-adjusted Mueller–Hinton broth supplemented with 5% lysed horse blood (bioTRADING Benelux B.V., Mijdrecht, the Netherlands). Antimicrobial agents were purchased from Sigma-Aldrich Chemie GmbH (Steinheim, Germany). Negative controls (without inoculum) and positive controls (with inoculum but without antimicrobial agent) were included, as was Camp. jejuni ATCC 33560 as a control strain. Tetracycline MICs for 63 Camp. coli isolates from the samples collected between 2007 and 2009 were determined by broth microdilution method (described above) or by the VetMIC method (National Veterinary Institute, Uppsala, Sweden, http://www.sva.se/en/service-and-products). Antimicrobial resistance was classified as microbiological resistance according to the epidemiological cut-off values (ECOFFs) published by the European Committee on Antimicrobial Susceptibility Testing (http://www.eucast.org). The ECOFF values are usually lower than clinical breakpoints, and therefore, a shift of MIC values towards resistance can be detected earlier than using clinical breakpoints, which refer to the outcome of treatment (clinical resistance). To define isolates as wild type or nonwild type, ECOFF values of ≤1, ≤8 and ≤2 μg ml−1 were used for ciprofloxacin, erythromycin and tetracycline, respectively. The presence of the tet(O) gene in Campylobacter and Enterococcus isolates was studied by PCR amplification of a 505-bp fragment (Fairchild et al. 2005). Total DNA was extracted as previously described (Juntunen et al. 2010), and the Camp. jejuni strain 81–176 was used as a positive control.
Subtyping by flagellin A gene restriction fragment length polymorphism (flaA-RFLP) was performed for 49 Camp. coli isolates with ciprofloxacin MICs above the ECOFF (herd 1: n = 18; herd 2: n = 31). The flaA fragments (1700 bp) were amplified with the consensus primers previously described (Wassenaar and Newell 2000). DNA was extracted, and the flaA fragment was amplified as previously published (Juntunen et al. 2010) using a PCR protocol comprising an initial denaturation step at 95°C for 15 min, 25 cycles of denaturation at 95°C for 30 s, annealing at 49°C for 90 s, and extension at 72°C for 2 min followed by a final extension at 72°C for 7 min. PCR products were digested with the DdeI restriction enzyme for 1 h at 37°C according to the manufacturer's instructions (New England Biolabs Inc., Ipswich, MA, USA). Restriction fragments were separated by gel electrophoresis, and the RFLP patterns were analysed with bionumerics v. 5.10 software (Applied Maths NV, Sint-Martens-Latem, Belgium) using the Dice similarity coefficient with 1% position tolerance and 0·5% optimization. The dendrogram was calculated by the unweighted pair group method using arithmetic averages.
Percentages of isolates above the ECOFFs were compared using pasw Statistics, Release 18.0.2 (SPSS Inc., Chicago, IL, USA). A P-value < 0·05 by Fisher's exact test was considered statistically significant.
A total of 236 Camp. coli were isolated from the two herds at three samplings. MIC values were determined for 221 Camp. coli isolates from herds 1 and 2 because some isolates did not grow after defreezing or in the positive controls of MIC testing. Similar distribution of tetracycline MICs was discovered amongst Camp. coli isolates collected before, during and after chlortetracycline treatment (Table 1). The tetracycline MICs ranged from ≤0·125 to 1 μg ml−1 in isolates obtained before treatment and from ≤0·125 to 0·5 μg ml−1 in those isolates collected during and after treatment. The tetracycline MIC50 values amongst treated and untreated pigs were ≤0·125 μg ml−1 at all sampling time points, and the MIC90 values were from 0·25 to 0·5 μg ml−1. Tetracycline MICs of the additional 63 Camp. coli isolates collected in the earlier study from seven farrowing farms ranged from ≤0·125 to 2 μg ml−1 (Table 2).
|Antimicrobial agent||Range of dilutions tested (µg ml−1)||n||No. of isolates carrying tet (O)||No. of isolates with MICs (μg ml−1) of||MIC50||MIC90|
The presence of the tet(O) gene was analysed in a total of 236 Camp. coli isolates and one Camp. jejuni isolate from herds 1 and 2; its prevalence was also assessed in 63 additional Camp. coli isolates from the other seven farms. According to the PCR analysis, none of the 300 Campylobacter isolates contained tet(O). One of the studied 30 Enterococcus spp. isolates carried the tet(O) gene; however, this tet(O)-positive isolate was present prior to chlortetracycline treatment.
Campylobacter coli isolates with ciprofloxacin MICs above the ECOFF were isolated before, during and after chlortetracycline treatment (Table 1). Altogether, 30·0 and 22·2% of the isolates had a ciprofloxacin MIC above the ECOFF during and after the treatment, respectively. Corresponding figures for untreated pigs were 16·0% (during treatment) and 11·1% (after treatment). However, the difference was not statistically significant (during: P = 0·34; after: P = 0·33). For ciprofloxacin, the MIC50 value was 0·125 and the MIC90 values ranged from 4 to ≥16 μg ml−1.
The erythromycin MICs fell below the ECOFF except for one Camp. coli isolate collected from a pig post-treatment (Table 1). MIC for ciprofloxacin fell below the ECOFF for this isolate. The MIC50 and MIC90 values for erythromycin were 1 and from 2 to 4 μg ml−1, respectively.
The MIC values of Camp. coli pairs isolated from one pig during the same sampling were compared (Table 3). More than 90% of the Camp. coli pairs had erythromycin and tetracycline MICs within ± one dilution step, and all MICs for these antimicrobial were within ± three dilution steps. The difference between two isolates was highest for ciprofloxacin MICs, for which ≥7 dilution step difference was noticed between 7 of 88 pairs. Ciprofloxacin, erythromycin and tetracycline MICs were identical for 75·0, 67·0 and 71·6% of the Camp. coli pairs, respectively.
|Antimicrobial agent||No. (cumulative %) of isolates with difference of twofold dilution steps in MICs||No. of Camp. coli pairs|
|Ciprofloxacin||66 (75·0)||10 (86·4)||0 (86·4)||3 (89·8)||0 (89·8)||1 (90·9)||1 (92·0)||7 (100·0)||88|
|Erythromycin||59 (67·0)||21 (90·9)||7 (98·9)||1 (100·0)||88|
|Tetracycline||63 (71·6)||19 (93·2)||4 (97·7)||2 (100·0)||88|
A total of 13 flaA types were observed amongst 49 Camp. coli isolates by flaA-RFLP typing (Fig. 1). The isolates from herd 1 (n = 18) were clustered into six flaA types, and 31 isolates from herd 2 were distributed amongst eight flaA types. The isolates from two herds did not share same flaA types, except for type 9, in which one isolate originated from herd 1 and the other three from herd 2. The most predominant flaA type amongst the isolates from herd 2, type 12, was observed 17 times from 10 pigs both before and during treatment, but seven different flaA types (1, 3, 5, 8, 9, 11 and 13) were isolated during or after treatment. Type 2 was the most commonly detected flaA type from herd 1; it was present before, during and after treatment. Two other flaA types (4 and 6) were also observed before treatment in herd 1, but three other types (7, 9 and 10) were present only during or after treatment.
Ribosomal protection protein Tet(O) commonly confers tetracycline resistance in Campylobacter isolates (Taylor and Chau 1996; Dasti et al. 2007; Kurincic et al. 2012). The lack of the tet(O) gene in our Camp. coli isolates resulted in low tetracycline MIC values. A 10-day period of chlortetracycline treatment in pigs did not select for Camp. coli isolates with high tetracycline MICs from populations negative for tet(O) prior to treatment. All tetracycline MICs were ≤1 μg ml−1 and therefore below the ECOFF (≤2 μg ml−1). We additionally studied the prevalence of the tet(O) gene amongst Camp. coli isolates previously collected from seven Finnish pig farms. None of the 63 isolates carried tet(O), and the tetracycline MICs of those isolates also remained under the ECOFF. Combined with earlier monitoring studies (FINRES-Vet 2011; EFSA/ECDC 2012a), these results indicate that tetracycline MICs are low in Finnish porcine Camp. coli, and this is due to the absence of the tet(O) gene.
The presence of the tet(O) gene was commonly observed both in Enterococcus spp. and Camp. coli isolated from single pigs in a recent study (Frye et al. 2011). To discover whether commensal Enterococcus spp. harboured tet(O), we studied its presence in Enterococcus isolates from herd 2. Surprisingly, only one Enterococcus spp. isolate was tet(O) positive; moreover, this positive isolate was collected prior to chlortetracycline treatment. However, a Finnish monitoring programme has revealed a high percentage of tetracycline resistance amongst porcine Ent. faecium (33%) and Ent. faecalis (74%) (EFSA/ECDC 2012a). No further isolates were shown to carry tet(O) even following selection by tetracycline, but our limited number of Enterococcus isolates from a single herd may have contributed to this disparity; however, Enterococcus spp. may harbour other tetracycline resistance determinants such as tet(A), tet(K), tet(L), tet(M), tet(R), tet(S) or tet(W) (Fairchild et al. 2005; Frye et al. 2011). Therefore, tetracycline resistance in many Finnish Enterococcus isolates may be conferred by other less common tet genes, and tet(O) is probably not horizontally transferred to Camp. coli from Enterococcus present in the same pigs. In a study of commercial chicken flocks, transfer of tet(O) from commensal bacteria to Camp. jejuni was not observed either (Fairchild et al. 2005).
Efflux pumps contribute to tetracycline MICs, but this activity seems unable to lead to high tetracycline MICs in the absence of tet(O). According to a previous study, if MIC values were below the ECOFF, the efflux pump inhibitor PAßN (phenylalanine–arginine–β-naphthylamide) decreased tetracycline MICs only by one or two dilution steps (Kurincic et al. 2012).
In addition to its preventative or therapeutic uses, chlortetracycline may be administered simply to promote growth in pigs. It was estimated to be the most common in-feed antimicrobial used in the U.S. swine industry (Apley et al. 2012). However, tetracycline has not been used as a feed additive in Finland, and Finnish feed manufacturers ceased the use of other growth promoters during the 1990s (FINRES-Vet 2011). This history of limited antimicrobial use may contribute to the susceptibility of Finnish Camp. coli isolates compared with the tetracycline resistance observed in many other countries (Qin et al. 2011; EFSA/ECDC 2012a; Quintana-Hayashi and Thakur 2012).
In a Canadian study of sheep, an association between tetracycline use and tetracycline resistance in Campylobacter was not detected, even though 78·9 and 39·4% of the Camp. coli and Camp. jejuni isolates were tetracycline resistant, respectively (Scott et al. 2012). Furthermore, in a Japanese study, use of tetracycline within the previous 6 months did not correlate with oxytetracycline resistance in porcine Camp. coli isolates (Asai et al. 2007). However, they reported high oxytetracycline resistance both at the tetracycline-free farms (87%; 101/116) and at the farms where tetracycline had been used within the previous 6 months (94%; 32/34). Additionally, studies conducted in chickens did not detect emergence of tetracycline-resistant Camp. jejuni after oxytetracycline or chlortetracycline treatment (Fairchild et al. 2005; Piddock et al. 2008). Even though tetracycline-resistant Campylobacter does not seem to emerge after oxy- or chlortetracycline treatment, a previous study of orally administered chlortetracycline detected a bioavailable form in pig faeces, and tetracycline-resistant coliforms were isolated throughout the study period until 5 weeks after treatment (Hansen et al. 2002).
Ciprofloxacin MICs were above the ECOFF for 22% of the Camp. coli isolates in this study, and high MICs were observed before, during and after treatment. Comparison of the MICs of Camp. coli pairs showed that isolates with both low and high ciprofloxacin MICs were observed in a single pig. Fluoroquinolones were not used in the herds, indicating that isolates with high ciprofloxacin MICs may have originated from the farrowing farms. It is also possible that once the isolates with high ciprofloxacin MICs had been transported to the herds, they may have persisted at the farm environment and colonized following batches of pigs. We have previously detected high ciprofloxacin MICs in Camp. coli at a farrowing farm in which fluoroquinolones were not used (Juntunen et al. 2010). The results of our present study together with previous findings (Luo et al. 2005; Nannapaneni et al. 2009) emphasize the persistence of ciprofloxacin-resistant Campylobacter even without antimicrobial selection pressure. Moreover, tetracycline resistance is commonly detected in Camp. coli even at antimicrobial-free pig farms (Gebreyes et al. 2005; Tadesse et al. 2011).
We did not detect Camp. coli isolates with high erythromycin MICs, suggesting that these finishing pigs were not exposed to tylosin; we have previously demonstrated that tylosin selects for high erythromycin MICs in Camp. coli at farrowing farms (Juntunen et al. 2010, 2011). Instead, subtherapeutic chlortetracycline treatment of calves was previously shown to select for erythromycin-resistant Camp. hyointestinalis (Inglis et al. 2005).
Comparison of the MIC values of two Camp. coli isolates originating from single pigs revealed that almost all Camp. coli pairs had erythromycin and tetracycline MICs within ± 2 dilution steps. As observed with the distribution of ciprofloxacin MICs, the presence of tet(O) would probably have resulted in higher range of tetracycline MICs within the same pigs.
Isolates with ciprofloxacin MICs above the ECOFF were subtyped by flaA-RFLP. Forty-nine isolates were distributed amongst 13 flaA types. Similar diversity was detected by an earlier study in which 66 porcine Camp. coli were divided into 18 flaA types (Oporto et al. 2007). Our isolates from two herds were clustered into separate flaA types, except for one isolate from herd 1 that was the same flaA type as three other isolates from herd 2. Our present genotyping results obtained by flaA-RFLP support our earlier observations made by pulsed-field gel electrophoresis (Juntunen et al. 2010, 2011) that porcine Camp. coli isolates with high ciprofloxacin MICs are genetically diverse, despite the presence of some predominant genotypes. Prior to treatment, we observed four flaA types in these two herds; only two of these were found during or after treatment, but nine additional flaA types were also distinguished. The dynamics of Camp. coli genotypes seem to be complex; the similarity of the MIC values for ciprofloxacin, erythromycin and tetracycline detected before, during and after treatment do not explain the change observed in the flaA types.
In conclusion, tetracycline MIC values of Camp. coli isolates were low before, during and after the administration of therapeutic in-feed chlortetracycline. The absence of the tetracycline resistance gene tet(O) in 300 Campylobacter isolates from nine farms explained the low tetracycline MIC values detected in Finnish porcine isolates. Furthermore, treatment did not increase MICs for ciprofloxacin or erythromycin, those antimicrobials used to treat human campylobacteriosis.
Authors acknowledge Rauha Mustonen for her laboratory work and Joana Revez for her contribution to flaA typing. Persons who assisted in samplings at the farms as well as Hanna Juntunen for her technical assistance are thanked. This study was financially supported by Walter Ehrström Foundation and performed at the Centre of Excellence on Microbial Food Safety Research, Academy of Finland (FCoE MiFoSa, grant no. 141140). P. Juntunen was funded by the Graduate School of the Veterinary Faculty of the University of Helsinki.
There is no conflict of interest to declare.