Antibiotic Resistance in Staphylococcus aureus and Coagulase Negative Staphylococci Isolated from Goats with Subclinical Mastitis

Antimicrobial resistance patterns and gene coding for methicillin resistance (mecA) were determined in 25 S. aureus and 75 Coagulase Negative Staphylococci (CNS) strains isolates from half-udder milk samples collected from goats with subclinical mastitis. Fourteen (56.0%) S. aureus and thirty-one (41.3%) CNS isolates were resistant to one or more antimicrobial agents. S. aureus showed the highest resistance rate against kanamycin (28.0%), oxytetracycline (16.0%), and ampicillin (12.0%). The CNS tested were more frequently resistant to ampicillin (36.0%) and kanamycin (6.7%). Multiple antimicrobial resistance was observed in eight isolates, and one Staphylococcus epidermidis was found to be resistant to six antibiotics. The mecA gene was not found in any of the tested isolates. Single resistance against β-lactamics or aminoglicosides is the most common trait observed while multiresistance is less frequent.


Introduction
Raw goat's milk can be a potential source of antibioticresistant pathogens of animal, human, and environmental origin. The microorganisms which contaminate raw milk may originate from the farm environment or from the goats and include the etiological agents responsible for clinical and Subclinical Mastitis (SCM).
In dairy goats with SCM, Coagulase Negative Staphylococci (CNS) make up 44.7% to 95.9% of the isolated pathogens from milk samples, and S. aureus, which is usually considered to have the greater pathogenicity, accounts from 4.1% to 18.0% of SCM agents [1]. The average prevalence of SCM in dairy goat's farms is between 20.0%-35.0% and results in significant economic losses due to reduction in milk production and poor milk quality [1].
The intramammary administration of antibiotics used on farms has increased, as it was proved to be effective for treating SCM in dry small ruminants [2,3]. The efficacy of intramammary antibiotic treatment could be compromised by staphylococci that produce biofilms in the udder.
The widespread use of antibiotics on dairy farms, could lead to the selection and to the emergence of antibioticresistant bacterial strains [4].
Most published scientific research papers had for object the antimicrobial resistance of intramammary infection pathogens isolated from raw bovine milk. The detection in raw milk of multiple resistant strains, and especially methicillin resistant S. aureus (MRSA) and CNS (MRCNS) strains, is regarded as an issue of great concern for their potential spread through the dairy food chain. In small ruminants, the spread of MRSA strains is controversial. MRSA strains are characterized by the presence of the mecA gene encoding low-affinity penicillin binding protein (PBP2 ), which mediates resistance to all classes of β-lactam antibiotics [5]. Few studies have reported that S. aureus strains, isolated from ewes milk affected by SCM, did not carry the mecA gene [6], or that, the few strains that were 2 Veterinary Medicine International resistant to methicillin (oxacillin), probably were of human origin [7]. Instead, other studies showed how MRCNS strains can be isolated from goats milk with SCM [8].
Antibiotic resistance pattern for staphylococci isolated from SCM refers mainly to cattle, and little is known about dairy goats [8][9][10][11][12]. Studies on antimicrobials susceptibility of these pathogens [13] have been mainly conducted using the agar disc diffusion method of Bauer et al. The broth microdilution and the agar dilution methods [14] instead allow the evaluation of the minimum inhibitory concentrations (MIC).
The evaluation of the antimicrobial susceptibility of Staphylococcus spp. isolated from goats with SCM is of great interest for clinical purposes in order to decide which antibiotics should be administered, as well as, for monitoring the spread of multiple resistant strains on farms. The current study seeks to support the sparse literature on antibiotic resistance of S. aureus and CNS isolated from goats with SCM. MICs and in vitro susceptibilities to ten antibiotics used in the medical and veterinary fields were determined against 25 S. aureus and 75 CNS strains. On these strains, the frequencies of single and multiple antibiotic-resistance were also evaluated.

Dairy Farms.
The milk samples were collected from eight goats flocks in the island of Sardinia (Italy) where mainly Sarda and Sarda-Maltese breeds were reared. The rearing system was extensive and the animals were handmilked. The goats, fed on bushes and grass, occasionally were supplemented with concentrates. The mean flock size was 187.5 ± 25.2 (mean ± SD) with a range from 160 to 234. Goats were not treated with systemic or intramammary antimicrobial agents during the lactation previous to the enrollment in this study nor during the dry period.

Sampling.
A clinical examination of half-udders was conducted in order to exclude animals with signs of clinical mastitis [15]. A total of 3,000 half-udder milk samples (1,500 goats) were collected in a single sampling time in each of the eight flocks during the early lactation period (from January to April of 2006). The first few streams of foremilk were discarded, and duplicate half-udder milk samples were aseptically collected into sterile tubes after cleaning and disinfection of each teat end. One sample (10 mL) was used for bacteriological analysis and the other one (50 mL) was added with bronopol (2-bromo-2-nitropropane-1,3diol) and analysed for Somatic Cell Count (SCC). The milk samples were stored at +4 • C, and bacteriological and SCC examinations were carried out within 6 and 72 hours after sampling.

Bacteriological Analysis and Isolation Procedures.
Each half-udders milk sample was mixed by inversion and 10 μL were then inoculated onto 5.0% Sheep Blood Agar (SBA) plates. The SBA plates were then incubated aerobically at +37 • C and examined after 48 and 72 hours. A significant bacterial count was considered when a growth of ≥500 identical cfu/mL was detected. Bacterial strains were isolated on Brain Heart Infusion agar (BHI, Oxoid, Basingstoke, UK), and identified using routinary microbiological procedures such as colony morphology, microscopic characteristics and Gram staining, hemolysis pattern on SBA, catalase and oxidase reactions. The strains were identified using API ID32 STAPH system (bioMérieux, Lyon, France) and the strips were read by the mini API instrument and associated software V 1.5.2 (bioMérieux, Lyon, France). The strains were then frozen at −80 • C in BHI broth (Oxoid, Basingstoke, UK) with 15.0% glycerol.

Definition of Subclinical Mastitis.
For the purpose of this work, a mammary gland was considered as affected by subclinical mastitis when having no clinical signs or abnormal milk, in which ≥500 cfu/mL were isolated and a somatic cell count was ≥300,000 cells/mL [18].

Antimicrobial Agents and Minimal Inhibitory Concentration (MIC).
On each strain the MICs of ten antibiotics used in human and veterinary medicines were determined. The antibiotics tested were ampicillin (AMP), cephalothin (KF), cefoperazone (CFP), ceftriaxone (CRO), cloxacillin (OB), kanamycin (K), novobiocin (NV), ofloxacin (OFX), oxytetracycline (OT), and vancomycin (VA). The MICs were determined by the broth microdilution method [14,22] using cation-adjusted Mueller-Hinton broth (CAMHB, Oxoid, Basingstoke, UK). Each antimicrobial agent, in powder form (Sigma-Aldrich-Fluka, MI, Italy), was weighed and dissolved in an appropriate solvent [14,22], thus obtaining a stock solution (2,560 μg/mL). Stock solutions were stored at −80 • C until used. From each stock solution, 12 serial twofold working dilutions in deionized water (only for AMP, the diluents were phosphate buffer, pH 6.0, 0.1 mol/L) was prepared according to CLSI standard protocols, and the antimicrobial agent final concentrations in each microplate ranged between 0.06 and 128 μg/mL. Each strain stored at −80 • C until testing were subcultured twice on BHI agar (Oxoid, Basingstoke, UK) before inoculum preparation. Veterinary Medicine International 3 Two or more identical colonies were picked from BHI plates after overnight incubation and suspended in saline solution (0.85% w/v) to match a McFarland 0.5 turbidity standard, using an inoculum reader (portable photometer Densimat, bioMérieux, Lyon, France). Each suspension was further diluted 1 : 100 in CAMHB in order to achieve the adequate inoculum concentration ( 10 6 cfu/mL). Fifty microliters of the final suspension were inoculated into the wells of microtiter plates, which also contained 50 μL of the antimicrobial agent, so that the final inoculum density on test plates contained 5 x 10 5 cfu/mL in each well. Reference strains, Staphylococcus aureus ATCC 29213 and Enterococcus faecalis ATCC 29212, were used for quality control. Each microplate was then incubated at 35 ± 2 • C under aerobic environment for 20 hours. The susceptibility of each strain to the antimicrobial agents was then defined by comparing the results to those of the breakpoint values [19][20][21]. The MICs range and mode, MIC 50 and MIC 90 of each antimicrobial agent were also determined.

Detection of mecA
Gene. Genomic DNA used as target for polymerase chain reaction (PCR) assay was extracted using the following procedure: strains grown in BHI broth at +37 • C overnight were centrifuged (10 minutes at 3,000 × g) and resuspended in 500 μL of Tris-EDTA buffer saline (Tris base 10 mM + EDTA 1 mM). The suspension was added with 10 μL of a 1.5 mg/mL lysostaphin solution (Sigma-Aldirich, MI, Italy) and incubated for 1 hour at +37 • C. Then, 5 μL of a 20 mg/mL proteinase K solution (Eurobio, Sarreguemines, France) were added and the incubation was continued at +50 • C for 60 minutes. An equal volume of phenol/chloroform/isoamyl alcohol (25 : 24 : 1) was added and mixed by inversion. After centrifuging (15 minutes at 10,000 × g), the upper layer was collected and 500 μL of chloroformisoamyl alcohol (24 : 1) solution were added. The mixture was centrifuged again (15 minutes at 10,000 × g) and the upper aqueous phase was transferred into a new tube. A volume of 800-1,000 μL of refrigerated absolute ethanol was added and gently mixed until DNA precipitation. DNA was resuspended in 100 μL of sterile deionized water. The DNA concentration was estimated spectrophotometrically. The mecA gene coding for methicillin resistance was detected by PCR as previously described [23]. The primers used for the detection of the mecA gene were AAAATCGATGGTAAAG-GTTGGC (forward) and AGTTCTGCAGTACCGGATTTGC (reverse). S. aureus HT 2004 0874 reference strain was used as positive control [24].  2%). The mean SCC of bacteriological positive samples was 6.3 Log 10 cells/mL, while in the negative ones it was 5.7 Log 10 cells/mL. The mean SCC was greater in milk samples positive for S. aureus (6.8 Log 10 cells/mL) when compared to those positive for CNS (6.4 Log 10 cells/mL).

Antimicrobials Susceptibility.
The MICs of antibiotics and the susceptibility of S. aureus and CNS strains isolated from goats with SCM are shown Tables 1-2. Fourteen (56.0%) of 25 S. aureus and thirty-one (41.3%) out of 75 CNS strains were resistant to one or more antimicrobials. The susceptibility of S. aureus was 92.0% or greater for seven out of ten antibiotics tested but was lower for kanamycin (60.0%), oxytetracycline (84.0%), and ampicillin (88.0%). The susceptibility of CNS was between 94.0%-100.0% for eight antimicrobials, but was somewhat lower for ampicillin (64.0%) and kanamycin (78.7%). Staphylococcus spp. isolates showed a poor susceptibility to AMP. The MIC 90 (1-4 μg/mL) of this antibiotic was higher than the breakpoint value for susceptibility. Among CNS resistant to AMP, the most prevalent species were S. caprae (37.0%) and S. chromogenes (22.2%), while only 12.0% of the S. aureus strains were resistant. On the other hand, almost all staphylococci (98.0%) were susceptible to OB and only 2 out of 75 (2.7%) of the CNS were resistant. Cephalosporins showed high activity against Staphylococcus spp. isolates. The percentages of susceptible staphylococci were 98.0%, 99.0% and 96.0% for KF, CFP, and CRO, respectively. The mecA gene was not found in any of the tested isolates. For K a low susceptibility of the isolates (74.0%) was recorded. The    MIC 90 of K against the CNS and S. aureus strains (32-≥128 μg/mL) was higher than the reference breakpoint for antimicrobial susceptibility. A different susceptibility to OT was observed in CNS isolates (94.7%) when compared to that of S. aureus (84.0%). The MIC 90 of OFX for both S. aureus and CNS was comparable with the reference breakpoint for antimicrobial susceptibility (≤1 μg/mL). All the tested staphylococci were susceptible to VA (100.0%), and 98.0% of these to NV. The NV resistant strains belonged to the S. chonii (n = 1) and S. xylosus (n = 1) species. The isolates which were resistant to two or more antimicrobial agents are shown in Table 3. It is remarkable that one S. epidermidis strain was resistant to six different antibiotics (AMP, CRO, K, OB, OFX, and OT).

Discussion
In the present study, the average prevalence of SCM was 15.6%, which is within the range (6.5%-67.0%) reported in previous studies carried out on dairy goat farms [1].
Staphylococci, which made up 88.5% of the isolated microorganisms in this study, are the most common pathogens associated with SCM in dairy goats [25]. AMP was less effective than the other β-lactam antimicrobial agents. The susceptibility of the isolates against AMP was within the range reported by other authors [9,26], even if Moroni et al. [27], have found a markedly greater prevalence of AMP resistant S. aureus (67.9%). These findings are consistent with AMP sensitivity to the penicillinases, frequently produced by Staphyloccus spp. and particularly by CNS strains. OB, a penicilinase-resistant penicillin (PRP), was very effective in vitro. Two CNS resistant to OB were simultaneously resistant to other β-lactamic antimicrobials, such as AMP (n = 1) and AMP-CRO (n = 1). The Cephalosporins showed a strong activity with regard to staphylococci. The MIC 90 values of CFP ranged between 0.25 and 8 μg/mL, that were comparable (1.87-3.75 μg/mL) to those found by Moroni et al [12]. Nevertheless, a low susceptibility to CFP was previously found in some CNS strains isolated from goats with SCM, particularly with  regard to S. chromogenes, S. warneri, S. simulans, and S. kloosii [11]. Other authors found that the susceptibility to KF of CNS isolated in goats with SCM was between 86.0% and 100.0% [10,28], that is, comparable with the results (98.7%) obtained in this study. The mecA gene was not found in any of the strains tested, in agreement with the results of a previous study carried out on S. aureus strains isolated from sheep with SCM [6]. The results of the present study confirm that methicillin-resistant staphylococci prevalence is still low in ruminants as observed in previous research [29]. The finding of some mecA-negative isolates which were phenotipically resistant to β-lactam antimicrobial agents could be related to a less common type of resistance due to either overproduction of β-lactamase or the presence of altered Penicillin Binding Protein (PBP) not related to 2a or 2 [30]. The susceptibility to OT was lower in S. aureus (84.0%) than in CNS (94.7%). In previous studies, a number of authors have observed a marked variability in the susceptibility of both these microorganisms to tetracycline, as it ranged between 10.0% and 100.0% [8,10,12,28]. The susceptibility of CNS against NV was 97.3% and this peculiarity is of interest in the taxonomy for bacterial typing since it is also well related with pathogenic activity [31]. All staphylococci were sensitive to VA, thus confirming the results of other authors [10,32]. The VA breakpoint value was recently reduced from ≤4 μg/mL to ≤2 μg/mL in testing the susceptibility of bacterial strains isolated from humans [20]. Some staphylococci isolated from milk samples taken from goats with SCM showed MIC values of 4 μg/mL. When these strains are transferred from animals to humans, they could increase the spreading of vancomycin-intermediate S. aureus (VISA) strains. Indeed, comparing the MIC values of VA that we found with the breakpoint actually used for human origin strains, a relevant percentage of S. aureus (16.0%) and CNS (13.3%) would be classified as intermediate.

Conclusion
This study confirms that staphylococci are the most common pathogens associated with SCM in dairy goats. As expected, the bacteriological positive milk samples had a SCC greater than the negative ones. Over 40.0% of the tested staphylococci were resistant to at least one antimicrobial agent. Single resistance against β-lactamics or aminoglicosides is the most common trait observed. Multiple antibiotic resistance was found in few of the tested Staphylococci strains, mainly in CNS. Although the methicillin-resistant staphylococci represent the most important pathogens responsible of humans severe hospital-acquired infections, the absence of mecA gene and the low prevalence of single and multiple antibiotic resistance suggest that SCM in goats does not play a significant role in the spreading of multiresistant staphylococci and it does not represent a great public health concern.