Multidrug-resistant coagulase-negative Staphylococci in food animals

Authors


Correspondence

Yifan Zhang, 3009 Science Hall, Department of Nutrition and Food Science, Wayne State University, 5045 Cass Avenue, Detroit, MI 48202, USA. E-mail: yifanzhang@wayne.edu

Abstract

Aims

To study the antimicrobial resistance of coagulase-negative staphylococci (CoNS) in animals.

Methods and Results

In the present study, a total of 87 CoNS recovered from food animals were characterized by antimicrobial susceptibility testing, resistance gene identification and conjugation. Of the seven species studied, Staphylococcus lentus, Staphylococcus sciuri, Staphylococcus xylosus and Staphylococcus haemolyticus accounted for over 96% of the isolates. In addition to β-lactam resistance (100%), high percentages of CoNS were resistant to tetracycline (67·8%), erythromycin (36·7%), clindamycin (27·5%) and quinopristin/dalfopristin (14·9%). Importantly, 47 (54%) isolates were resistant to at least three antimicrobial classes, including six CoNS resistant to six antimicrobial classes. The common genes for the above-mentioned resistance phenotypes were mec(A), tet(M), erm(A) and vga(A)LC, which were identified from 68·7%, 61%, 56·2% and 69·2% of the isolates, respectively. tet(M) was conjugatively transferable from 10 tetracycline-resistant CoNS to a Enterococcus strain, underlining the potential of antimicrobial resistance transfer from Staphylococcus to the commensal bacteria in human.

Conclusions

Multidrug resistance and resistance to non-β-lactam antimicrobials are common in CoNS in animals.

Significance and Impact of the Study

The data improve our understanding on the extent to which CoNS contribute to the dissemination of antimicrobial resistance in the food production environment.

Introduction

Although Staphylococcus aureus (S. aureus) has the highest pathogenic potential among the many species of staphylococci, coagulase-negative staphylococci (CoNS) have become an increasing concern in human and veterinary medicine (Rice 2006; Fessler et al. 2010). CoNS can be associated with a variety of animal and human diseases, such as suppurative disease, arthritis and urinary tract infection in animals (Lee 2003) and skin and soft tissue infections and bacteremia in humans (Martins and Cunha Mde 2007). Animals are common reservoir of Staphylococcus and can carry CoNS on their skin, noses, upper alimentary and urogenital tract and intestinal tract and thus may transmit the bacteria to human by animal handling (Werckenthin et al. 2001). CoNS carried by food animals can also contaminate the animal products and enter the food chain, thus posing potential threat to food consumers and handlers.

More importantly, methicillin-resistant CoNS (MRCoNS) have been isolated worldwide from food animals (Kawano et al. 1996; Fessler et al. 2010; Haenni et al. 2011), including pigs, cows, calves and chicken and therefore may compromise the treatment if animal or human infection occurs. Interestingly, resistance to non-β-lactam antimicrobials, such as erythromycin, tetracycline, clindamycin, ciprofloxacin and sulfamethoxazole/trimethoprim, is common in MRCoNS (Simeoni et al. 2008; Huber et al. 2011). A recent study from Switzerland revealed 48·2% of MRCoNS from livestock, and high percentages of CoNS were also resistant to non-β-lactam antimicrobials, suggesting that multidrug-resistant CoNS may become an emerging problem for veterinary medicine and public health (Huber et al. 2011). However, most studies on CoNS in animals have been performed in Europe. There are limited data on CoNS in US food animals as well as resistance of CoNS to non-β-lactam antimicrobials.

In addition to mec(A) that is carried by a mobile genetic element, staphylococcal cassette chromosome mec (SCCmec), and confers broad-spectrum β-lactam resistance, various antimicrobial resistance genes have been identified in CoNS from food production environment. erm(A), erm(B), erm(C), vga(B), tet(K) and tet(M) have been found in CoNS resistant to the macrolide–lincosamide–streptogramin B (MLSb) antimicrobials and tetracyclines (Aarestrup et al. 2000; Simjee et al. 2007; Graham et al. 2009), the antimicrobials commonly used in animal production in the United States. As many of these genes are carried on extra-chromosomal elements, including conjugative plasmids and transposons (Aarestrup et al. 2000; Schmitz et al. 2001; Khan et al. 2002; Simjee et al. 2007), CoNS strains may become an important reservoir of antimicrobial resistance genes that have potential to transfer to other commensal bacteria or even pathogenic ones.

The present study was aimed to understand CoNS of animal origin as a reservoir of multidrug resistance. We examined the antimicrobial susceptibility phenotypes and genotypes of CoNS from agricultural animals and estimated the potential of tetracycline resistance gene transfer by conjugation.

Materials and methods

Bacterial strains

A total of 87 CoNS was selected from a collection of isolates that were originally recovered from oropharangeal cavity and rectum/cloacal of a variety of agricultural animals. All isolates were potentially resistant to β-lactam antimicrobials as the isolation protocol was designed to recover β-lactam-resistant Staphylococcus. Staphylococcal species identification have been performed in the previous study (Zhang et al. 2009), in addition to SCCmec typing and PFGE analysis, by Gram staining, catalase and tube coagulase tests, and API-Staph biochemical tests. The isolates consisted of Staph. lentus (33), Staph. sciuri (30), Staph. xylosus (12), Staph. haemolyticus (9), and one each of Staph. capitis, Staph. epidermidis, and Staph. hominis. CoNS-carrying animals included cattle (n = 27), sheep (n = 25), goats (n = 13), pigs (n = 7), chicken (n = 5), turkey (n = 5), duck (n = 3), geese (n = 1) and horses (n = 1). Bacterial strains were cultured on tryptic soy agar (TSA; Difco, Detroit, MI, USA) for further analysis.

Antimicrobial susceptibility testing of CoNS

Antimicrobial susceptibility of the 87 CoNS was determined using the Sensititre antimicrobial susceptibility system (Trek Diagnostic Systems, Westlake, OH, USA) and interpreted according to the Clinical and Laboratory Standards Institute (CLSI) guidelines (CLSI 2010). Briefly, 50 μl of 5 × 105 colony forming unit (CFU) ml−1 inoculum in cation-adjusted Mueller–Hinton broth (CAMHB; Difco, Detroit) was added to the Sensititre plates under aseptic conditions. Plates were covered with plate sealers and incubated at 35°C for 24 h. Ampicillin, cefoxitin (6 μg ml−1), chloramphenicol, ciprofloxacin, clindamycin, erythromycin, gentamicin, linezoid, levofloxacin, moxifloxacin, nitrofurantoin, oxacillin, penicillin, quinupristin–dalfopristin (Q/D), rifampin, streptomycin, tetracycline, tigecycline, sulfamethoxazole/trimethoprim and vancomycin were tested. Staphylococcus aureus ATCC 29213 was used as the quality control micro-organism.

Molecular determination of antimicrobial resistance

Genomic DNA template used for PCR was extracted by a boiling method as previously described (Zhang et al. 2005). mec(A) PCR (Perazzi et al. 2006) was performed in all 87 CoNS, whereas 10 genes conferring resistance to the MLSb group and tetracycline were tested only on isolates demonstrating the resistance phenotypes. The genes tested were erm(A), erm(B), erm(C), vat-1, vat-2, vat-3, vga(A), vga(B), tet(K) and tet(M) (Sutcliffe et al. 1996; Soltani et al. 2000; Volokhov et al. 2003; Weigel et al. 2007). In addition, tetracycline-resistant isolates were also examined for the presence of Tn916, a conjugative transposon (Soge et al. 2008).

Conjugation experiments

Tetracycline-resistant CoNS were used as donor strains in conjugation experiments to study tetracycline resistance gene transfer. Enterococcus faecalis JH2-2 (rifr fusr) was used as the recipient strain. Conjugation was performed by the filter-mating method (Agerso et al. 2006) with modifications. Briefly, overnight cultures of the donor strains grown in brain heart infusion (BHI) broth (Difco, Sparks, MD, USA) containing tetracycline (16 μg ml−1) and recipient grown in BHI containing fusidic acid (50 μg ml−1) and rifampicin (50 μg ml−1) were mixed (ratio, 1 : 1). The mixture was then placed on a 0·45-μm pore-size filter and incubated on BHI agar (Difco, Sparks) at 37°C overnight. The filter was washed and vortex-mixed in BHI broth. The mating mixture was spread onto BHI agar containing a combination of tetracycline (16 μg ml−1), fusidic acid (50 μg ml−1) and rifampicin (50 μg ml−1). Up to three potential transconjugants were purified on BHI agar containing appropriate antimicrobials, and resistance gene transfer by conjugation was confirmed by PCR.

Results

Antimicrobial susceptibility profiles of CoNS

All 87 CoNS showed resistance to oxacillin, whereas only 47 were able to grow at 6 μg ml−1 of cefoxitin. Resistance to the other two β-lactam antimicrobials were also high, 79·3% to ampicillin and 91·9% to penicillin. Fifty-nine isolates were resistant to tetracycline (67·8%), followed by 32 to erythromycin (36·8%), 24 to clindamycin (27·6%) and 13 to quinupristin–dalfopristin (14·9%). Resistance to other antimicrobials, such as chloromphenicol, ciprofloxacin, gentamicin and sulfamethoxazole–trimethoprim, were also identified, though at low prevalence. All 87 CoNS were resistant to at least one β-lactam antimicrobial, among which 16 were resistant to β-lactam only and 24 were resistant to β-lactam and one other antimicrobial. The remaining 47 isolates (54%) were resistant to at least three antimicrobial classes and defined as multidrug resistance, including 22 of 33 Staph. lentus, 15 of 30 Staph. sciuri, one of 12 Staph. xylosus and nine of nine Staph. haemolyticus. Twenty-one isolates demonstrated resistance to three classes, followed by 14–4 classes, six each to five and six classes. The common resistance profiles were β-lactam only (16), β-lactam and tetracycline (16), β-lactam, erythromycin and tetracycline (13), as well as β-lactam, clindamycin, erythromycin and tetracycline (7) (Fig. 1).

Figure 1.

Resistance of CoNS to various numbers of antimicrobial classes in this study. image No. of CoNS. Antimicrobial abbreviations: β-lactam, ampicilln/cefoxitin/oxacillin/penicillin; CHL, chloromphenicol; CIP, ciprofloxacin; CLI, clindamycin; ERY, erythromycin; GEN, gentamicin; MXF, moxifloxacin; NIT, nitrofurantoin; RIF, rifampin; SXT, sulfamethoxazole/trimethoprim; SYN, quinupristin–dalfopristin; TET, tetracycline.

All nine Staph. haemolyticus were β-lactam resistant, whereas prevalence of resistance to ampicillin, penicillin and oxacillin in Staph. lentus, Staph. sciuri and Staph. xylosus ranged from 72·7% to 100% (Table 1). Only one of nine Staph. haemolyticus was erythromycin resistant, in comparison with all but one Staph. xylosus being resistant. While high prevalence of resistance was observed in Staph. xylosus to β-lactam, erythromycin and tetracycline, all 12 Staph. xylosus were susceptible to other antimicrobials tested. A total of 13 CoNS, including six each of Staph. haemolyticus and Staph. lentus and 1 Staph. sciuri, were resistant to quinopristin–dalfopristin (Q/D). They were recovered from one cattle, two chicken, seven sheep and three turkey samples.

Table 1. Prevalence of common antimicrobial resistance in Staphylococcus lentus, Staph. sciuri, Staph. xylosus and Staph. haemolyticus
Classes of antimicrobialsAntimicrobials*No. of antimicrobial-resistant CoNS (%)
Staph. haemolyticus (N = 9)Staph. lentus (N = 33)Staph. sciuri (N = 30)Staph. xylosus (N = 12)
  1. *Resistance breakpoints for individual antimicrobial agents (μg ml−1): Ampicillin, ≥0·5; Chloromphenicol, ≥32; Ciprofloxacin, ≥4; Clindamycin, ≥4; Erythromycin, ≥8; Gentamicin, ≥16; Oxacillin, ≥0·5; Penicillin, ≥0·25; Quinupristin–dalfopristin, ≥4; Sulfamethoxazole/trimethoprim, ≥4/76; Tetracycline, ≥16.

β-lactamAmpicillin9 (100)24 (72·7)24 (80)11 (91·6)
Oxacillin9 (100)33 (100)30 (100)12 (100)
Penicillin9 (100)28 (84·8)29 (96·6)12 (100)
MacrolideErythromycin1 (11·1)14 (42·4)15 (50)11 (91·6)
TetracyclineTetracycline5 (55·5)24 (72·7)20 (66·6)10 (83·3)
AmphenicolChloromphenicol0 (0)4 (12·1)1 (3·3)0 (0)
QuinoloneCiprofloxacin5 (55·5)0 (0)1 (3·3)0 (0)
LincosamideClindamycin7 (77·7)8 (24·2)8 (26·6)0 (0)
AminoglycosideGentamicin5 (55·5)1 (3)1 (3·3)0 (0)
StreptograminQuinupristin–dalfopristin6 (66·6)6 (3)1 (3·3)0 (0)
AntifolatesSulfamethoxazole/trimethoprim0 (0)1 (3)4 (13·3)0 (0)

There was no clear distinction with regard to the prevalence of β-lactam, tetracycline and macrolide resistance among animal groups. Multidrug-resistant CoNS were recovered from all animal species except for a goose and a horse. Eighty per cent each of CoNS from chicken and turkey were multidrug resistant, followed by 66·7% from duck, 61·5% from goats, 60% from sheep, 57·1% from pigs and 37% from cattle (Fig. 2).

Figure 2.

Antimicrobial resistance phenotypes and resistance genes identified in CoNS. *black, resistance phenotypes; grey, susceptible phenotypes. Antimicrobial abbreviations: β-lactam (AMP, ampicillin; FOX6, cefoxitin (6 μg ml−1); OXA, oxacillin; PEN, penicillin); CHL, chloromphenicol; CIP, ciprofloxacin; DAP, daptomycin; GEN, gentamicin; LEVO, levofloxacin; LZD, linezolid; MLS, macrolide–lincosamide–streptogramin (ERY, erythromycin; CLI, clindamycin; SYN, quinupristin/dalfopristin); MXF, moxifloxacin; NIT, nitrofurantoin; RIF, rifampin; STR, streptomycin; SXT, sulfamethoxazole/trimethoprim; TET, tetracycline; TGC, tigecycline; VAN, vancomycin. †underlines indicate multidrug-resistant strains. ‡Met-r, methicillin resistance genes; Tet-r, tetracycline resistance genes; MLSb-r, macrolide–lincosamide–streptogramin B resistance genes and Streptogramin-r, streptogramin resistance genes.

Distribution of antimicrobial resistance genes and Tn916

mec(A) was detected in 60 of 87 CoNS (67·8%). erm(A), erm(B), erm(C) and vga(A)LC, were detected in 31 of 47 isolates that were resistant to clindamycin, erythromycin and/or Q/D. The 16 remaining isolates, including 13 resistant to erythromycin, did not carry any of the genes tested and were comprised of eight Staph. scuiri, five Staph. lentus, two Staph. haemolyticus and one Staph. epidermidis. Moreover, three isolates from cattle, turkey and chicken (IDs 65, 79 and 86) which were resistant to all MLS antimicrobials carried at least three resistance genes tested. Of 32 erythromycin-resistant isolates, erm(A) predominated and was detected in 17 (53·1%) isolates, followed by erm(C) (25%) and erm(B) (21·9%). With regard to the 24 clindamycin-resistant isolates, erm(B) and erm(A) were detected in 12 (50%) and 7 (29·2%) isolates, respectively. Q/D resistance genes, vga(A)LC and erm(B), were both recovered from 9 of 13 (69·2%) Q/D-resistant CoNS (Fig. 2). tet(M) was carried by 36 of 59 tetracycline-resistant CoNS, among which 31 also had Tn916. tet(K) was detected in 27 isolates. Fourteen isolates contained both tet(K) and tet(M).

Tetracycline resistance was conjugatively transferable in 10 of 59 (16·9%) tetracycline-resistant CoNS, in which five were Staph. scuiri, four were Staph. lentus and one was Staph. haemolyticus. All transconjugants were positive for tet(M) and Tn916. Except for one Staph. scuiri isolate that was resistant to two antimicrobial classes (tetracycline and β-lactam), the remaining nine isolates carrying conjugatively transferable tetracycline resistance were all multidrug resistant.

Discussion

Most studies on antimicrobial-resistant Staphylococcus in agriculture have focused on Staph. aureus from food or food-producing animals, whereas less research effort has been put on CoNS, a group of staphylococci that are believed to be a larger reservoir of antimicrobial resistance genes (Lee 2003; O'Mahony et al. 2005; Juhasz-Kaszanyitzky et al. 2007). The recovery of 47 multidrug-resistant CoNS (54%) from food animals and the observation of resistance to macrolides, tetracyclines and streptogramin, in addition to β-lactam resistance, suggest food animals as important reservoir of antimicrobial-resistant CoNS.

The four Staphylococcal species, Staph. lentus, Staph. sciuri, Staph. xylosus and Staph. haemolyticus, identified in this study are commonly associated with farm animals (Aarestrup and Schwarz 2006). The high prevalence of resistance to tetracycline, erythromycin and clindamycin was not surprising as compared to previous studies on CoNS in poultry litter in the United States (Simjee et al. 2007) and MRCoNS from various sources, including animals, meat and contact persons, in Switzerland (Huber et al. 2011). When comparing the three studies, we found tetracycline resistance (67·8%) predominated in our study and so did erythromycin resistance (71%) in the Simjee study, whereas no marked difference was observed in the Swiss study in terms of the prevalence of resistance to erythromycin, tetracycline and clindamycin, ranging from 43·7 to 49% (Huber et al. 2011). This may be an indication of common use of tetracyclines and macrolides in animal production in the United States. Recovery of Q/D-resistant CoNS in both of the US studies, with prevalence of 14·9% and 13%, respectively, suggests a potential linkage between streptogramin usage on farms and Q/D resistance in this country, although comparison between the US data and data from Switzerland was not possible due to the exclusion of Q/D in the Swiss study. As Q/D can be used to treat methicillin-resistant Staphylococcal infection in humans (Hamel et al. 2008), our study suggests a need to monitor the Q/D resistance in CoNS of animal origin and to understand the potential of Q/D-resistant CoNS to cause human disease.

Because all CoNS included in this study were potentially resistant to β-lactam, it was not unexpected that the isolates were resistant to at least one β-lactam antimicrobial. The recovery of high percentages of multidrug-resistant CoNS from most animal categories suggests that animal commensal bacteria are important reservoir of antimicrobial resistance phenotypes. Because animal intestinal environment provides an optimal condition for antimicrobial resistance genes to transfer from commensals to pathogens (Marshall et al. 2009), gene dissemination across species or even genus borders is expected to be common in animal hosts, and the extent of antimicrobial resistance can thus be amplified substantially. Multidrug resistance phenotypes were commonly seen in the top three animals carrying CoNS and the prevalence was 37%, 60% and 61·5%, from cattle, sheep and goats, respectively, which again is an indication of widespread distribution of multidrug resistance in agriculture, although farm variation in agricultural practices and level of antimicrobial exposure of bacteria could not be excluded.

The observation of staphylococcal species variation in multidrug resistance is consistent with findings by Huber et al. (2011) that resistance of CoNS to erythromycin, tetracycline, clindamycin, ciprofloxacin, sulfamethoxazole/trimethoprim and gentamicin was more common in Staph. haemolyticus, Staph. epidermidis and Staph. sciuri than that in other species, with Staph. haemolyticus having the highest percentage of non-β-lactam resistance. Another study also found 4 multidrug-resistant Staph. haemolyticus of six total mec(A)-positive Staphylococci from clinical animals (van Duijkeren et al. 2004). Together with our results that all nine Staph. haemolyticus vs only one of 12 Staph. xylosus were multidrug resistant, it is reasonable to assume that species variation exists in Staphylococcus as to their antimicrobial resistance. Staphylococcus haemolyticus in animals may have stronger public health significance, considering that it is frequently involved in human clinical cases (van Duijkeren et al. 2004; Huber et al. 2011).

erm(A) and erm(C) were commonly found in erythromycin-resistant Staphylococci in this study, which is in agreement with previous reports that they were the predominant erythromycin resistance genes in Staphylococcus isolated from various sources (Aarestrup et al. 2000; Khan et al. 2002; Simjee et al. 2007; Graham et al. 2009). Unlike erm(A) and erm(C), erm(B) is less common in Staphylococcus than in Enterococcus and Streptococcus. However, erm(B) was identified in seven erythromycin-resistant CoNS in our study, in contrast to its absence in previous studies (Aarestrup et al. 2000; Simjee et al. 2007; Graham et al. 2009), suggesting genetic diversity in erythromycin resistance genes in Staphylococci in different geographical locations. As erm(B) can be carried on plasmids, the chromosome and on transposons (Khan et al. 2002), future examination of the genetic background of erm(B) in these isolates would help assess the potential of its dissemination in Staphylococcus in animals. Our finding that 13 of 32 erythromycin-resistant isolates had no resistance genes identified was not surprising as compared to previous reports (Simjee et al. 2007), although it could also be due to the limited number of genes tested.

As for streptogramin resistance genes, the vga(A) gene detected was confirmed by DNA sequencing as vga(A)LC. This evolutionary variant of vga(A) encodes for an ABC transporter and has been reported in Staphylococcus resistant to lincosamides and streptogramin (Novotna and Janata 2006; Qin et al. 2011). Failure to observe clindamycin resistance in some of our vga(A)LC-positive isolates can be explained by the different phenotypic methods used to determine clindamycin resistance because we used a broth microdilution method whereas agar dilution or disc diffusion was applied in previous studies. In fact, the clindamycin resistance level of Staph. haemolytics carrying vga(A)LC reported by Novotna et al. was not very high (Novotna and Janata 2006), so borderline resistance phenotypes would not be unexpected when using a different method. In addition, source of isolation may also have indication on antimicrobial resistance phenotypes. All previously reported Staphylocccus carrying vga(A)LC was from clinical settings where constant antimicrobial selective pressure maintains the resistance phenotypes more effectively than what the antimicrobial selection does in agriculture.

The two mechanisms of tetracycline resistance reported in Staphylococcus, ribosomal protection and active efflux, were identified in this study as evidenced by the presence of tet(M), encoding for ribosomal protection, in more than 60% of CoNS, and tet(K), encoding for an active efflux, in 45·7% of the isolates. Tn916, a conjugative transposon, which is often associated with tet(M), was detected in most tet(M)-positive isolates. The fact that nearly 17% of CoNS could transfer tet(M) to Enterococcus faecalis, together with Tn916, suggests that tetracycline resistance from Staphylococcus can transfer to other Gram-positive bacteria that have potential to cause human diseases. However, considering the high prevalence of Tn916 in our isolates, this conjugation rate was low, which might be due to the variation in resistance transfer to E. faecalis from different Staphylococcus species, especially when lineage variation in tetracycline resistance transfer has been observed from Staph. aureus strains to Staph. aureus recipients (de Vries et al. 2009). As antimicrobial resistance genes other than tet(M) have been identified on Tn916 (Fletcher and Daneo-Moore 1992; Del Grosso et al. 2004; Lancaster et al. 2004), transfer of additional antimicrobial resistance genes should also be expected via this conjugative transposon from a multidrug resistance reservoir.

In conclusion, our data indicate that CoNS in agricultural animals are an important reservoir of multidrug resistance in addition to the resistance to β-lactam antimicrobials and underline the importance of surveillance of multidrug-resistant CoNS in the food production environment. As our CoNS strains were all resistant to at least one β-lactam antimicrobial, further research is needed as to whether methicillin resistance predisposes CoNS to become multidrug resistant as compared to general CoNS, including the potential linkage, if any, between β-lactam resistance and other resistance phenotypes. Species variation exists in the prevalence of multidrug resistance in Staphylococci. Certain Staphylococcal species, such as Staph. haemolyticus, may have stronger potential to become multidrug resistant and thus require closer research and public health attention.

Acknowledgements

We thank Jeffrey T. LeJeune from The Ohio State University for sharing the CoNS isolates. We also thank Vasjana Tomco and Farah Yousif for their excellent technical assistance.

Conflict of Interest

None to declare.

Ancillary