Antimicrobial resistance patterns and biofilm formation of coagulase-negative Staphylococcus species isolated from subclinical mastitis cow milk samples submitted to the Onderstepoort Milk Laboratory

Background Increased prevalence of antimicrobial resistance, treatment failure, and financial losses have been reported in dairy cows with coagulase-negative Staphylococcus (CoNS) clinical mastitis, however, studies on CoNS infections are limited in South Africa. Therefore, the objectives of this study were to investigate the antimicrobial resistance patterns and biofilm formation in CoNS isolated from cow milk samples submitted to the Onderstepoort Milk Laboratory. Results A total of 142 confirmed CoNS isolates were used for this study. Biofilm formation was identified in 18% of CoNS tested. Staphylococcus chromogenes (11%) had the highest proportion of biofilm formation followed by S. haemolyticus (4.0%), S. epidermidis, S. hominis, S. xylosus, and S. simulans with 1% respectively. Ninety percent (90%) of CoNS were resistant to at least one antimicrobial (AMR) and 51% were multidrug-resistant (MDR). Resistance among CoNS was the highest to ampicillin (90%) and penicillin (89%), few isolates resistant to cefoxitin and vancomycin, 9% respectively. Similarly, MDR-S. haemolyticus (44%), MDR-S. epidermidis (65%), and MDR-S. chromogenes (52%) were mainly resistant to penicillins. The most common resistance patterns observed were resistance to penicillin-ampicillin (16%) and penicillin-ampicillin-erythromycin (10%). Only 42% of biofilm positive CoNS were MDR. Conclusion The majority of CoNS in this study were resistance to penicillins. In addition, most isolates were β-lactam resistant and MDR. Biofilm formation among the CoNS in this study was uncommon and there was no significant difference in the proportion of MDR-CoNS based on the ability to form a biofilm.

The increasing prevalence of resistance among CoNS could be due to the injudicious use of antimicrobials [21], the presence of penicillin binding protein 2a (PBP2a) [16,22,23], and mecA mediated oxacillin resistance [24][25][26][27]. In addition, the high prevalence of antimicrobial resistance among CoNS could be due to their ability to form a biofilm which facilitates persistent infections [9,28,29] and decreases susceptibility to commonly used antimicrobials [30]. The ability of Staphylococcus species to form biofilm has been linked to the presence of biofilm-forming genes such as icaA and bap gene [20,30,31], which have been isolated in CoNS species such as S. epidermidis, S. haemolyticus, S. xylosus, S. chromogenes and S. simulans [31]. To our knowledge, there are no studies that have reported an association between biofilm formation and high prevalence of MDR in CoNS isolated from subclinical mastitis cases in dairy cattle. In addition, no studies have been published on the antimicrobial resistance patterns of CoNS from dairy cattle in South Africa.
Thus, the aim of this study was to investigate the antimicrobial resistance patterns and biofilm formation of CoNS isolated from cow milk samples submitted to the Onderstepoort milk laboratory, Department of Production Animal Studies, South Africa. We hypothesize that CoNS with biofilm formation isolated from subclinical mastitis cases have an increased prevalence of resistance to commonly used antimicrobials. In addition, it also possible that these isolates are β-lactam and multidrugresistant (MDR).

Discussion
Coagulase negative Staphylococcus species (CoNS) have been reported as a cause of mastitis in dairy cattle [13] with prognosis in affected patients dependent on antimicrobial resistance profile of the isolate, the presence of virulence factors, and biofilm formation [29]. In this study, we investigated antimicrobial resistance patterns and biofilm formation of CoNS isolated from cow milk samples submitted to the Onderstepoort milk laboratory.

Biofilm formation of CoNS
We observed a low (18%) proportion of biofilm-forming CoNS compared to other studies [20,31,32] For example, Tremblay et al. [31] in Canada reported 96.7% proportion of biofilm formation in CoNS from dairy cattle with mastitis. Similarly, 85.1% of CoNS from subclinical and clinical mastitis cases of dairy cattle in Argentina formed biofilm   [20]. Simojoki et al. [32] in Finland also reported a high (31.3%) proportion of biofilm-forming CoNS from clinical mastitis cases in dairy cattle. The low proportion of biofilm formation of CoNS in this study compared to the above-mentioned studies may be attributed to the difference the growth media used [33], since, Tremblay et al. [31] and Simojoki et al. [32] used brain heart infusion for biofilm assay while in this study TSB was used. Nonetheless, the low proportion of biofilm formation in this study suggests that biofilm formation is not common in CoNS subclinical mastitis. Therefore, the role played by biofilm formation in the prognosis of CoNS subclinical mastitis in this study is limited. However, more studies need to be done to further explore the molecular epidemiology of biofilm formation CoNS from subclinical and clinical mastitis cases in dairy cattle, South Africa.

Antimicrobial resistance of CoNS
A high (90%) proportion of CoNS in this study were resistant to at least one antimicrobial and no significant difference was observed in the proportion of resistance among CoNS species. Antimicrobial resistance in this study was higher than 21.4% reported in clinical mastitis cases of lactating cows in Sweden [34]. The reason for the high proportion of resistant CoNS isolates in this study is not clear. However, this could be due to selection pressure associated with injudicious use of antimicrobials for the treatment of clinical mastitis in dairy cattle from South Africa [21]. Moreover, antimicrobial drugs are easily available for farmers as over the counter medication [35]. Studies investigating antimicrobial use among farmers in the dairy industry in South Africa will be beneficial in understanding their role in antimicrobial stewardship. In addition, findings of this study suggest that interventions in the use of antimicrobial treatment among farmers in South Africa including restriction of antimicrobial use, limitation of over the counter antimicrobial, veterinary consultation, and the improvement of knowledge on antimicrobial resistance must be considered.

Penicillin resistance
A higher (63%) proportion of penicillins resistant CoNS was observed in this study compared to that reported in clinical mastitis of dairy cattle in Finland (32%) [32], Estonia (38.5%) [36], and Zimbabwe (8%) [5]. In contrast to other studies [37,38], we observed no difference in the proportion of penicillin resistance among CoNS species. The high proportion of penicillin resistance in this study could be due to the low affinity associated with penicillin-binding protein 2a (PBP2a) produced by Staphylococcus species [16,22,23]. In addition, this could also be due to overuse of these antimicrobials as they are readily available as over the counter antimicrobials for treatment of mastitis in dairy cattle in South Africa [39] mainly due to their narrow-spectrum activity [26,34,40].

Erythromycin resistance
The resistance to erythromycin among CoNS was higher (49%) in this study than reported in subclinical mastitis of dairy cattle in Argentina (29%) [17] and in Germany (22%) [41]. In contrast, a higher (73.2%) proportion of erythromycin resistant CoNS has been reported in a study done on subclinical mastitis cases of dairy cattle in Turkey [37]. The presence of erythromycin resistance in this study may be attributed to the presence of a ribosomal methylase based resistance, encoded by msrA and ermC [42]. Furthermore, Luthje et al. [41] suggest that alteration of the ribosomal methylase activity in Staphylococcus spp. could lead to horizontal gene transfer. Although erythromycin is one of the antimicrobial drugs used for control of CoNS isolates in bovine mastitis, it is not used for the treatment of CoNS in South Africa. Therefore, the high prevalence of resistance observed in this study needs further investigation.

Vancomycin resistance
Vancomycin resistance among CoNS in this study was uncommon (9%) compared to the 58.2% reported in subclinical bovine mastitis cases in Turkey [37]. In contrast, Bengtsson et al. [34] in Sweden reported no vancomycin resistance among CoNS isolated from mastitis cases in dairy cattle. Although vancomycin is not currently used for the treatment of clinical mastitis in South Africa, other peptides antimicrobial drugs such as bacitracin are used in combination intramammary applications. The presence of vancomycin resistance CoNS is of public health significance as vancomycin is the drug of choice for treatment of MRSA in human medicine [2]. Therefore, measures must be implemented including restriction on the use of peptides antimicrobial drugs in the treatment of mastitis in the dairy industry to curb the potential development of vancomycin-resistant CoNS.

Cefoxitin and β-lactam resistance
The cefoxitin test is the preferred method for testing the CoNS for mecA mediated oxacillin resistance [43][44][45].
We observed a lower (9%) proportion of cefoxitin resistant CoNS compared to the 29.41% reported in clinical mastitis cases from dairy cattle in Tunisia [46] and 40% reported in subclinical mastitis from dairy cattle in Switzerland [47]. In addition, there was no significant difference in the proportion of cefoxitin resistant isolates within CoNS species. The presence of mecA mediated oxacillin resistance is suggestive of methicillin-resistant coagulase negative Staphylococcus species [1] and also encodes for penicillin-binding protein PBP2a. Together with the blaZ gene have been reported in β-lactam resistance among Staphylococcus species [9,48]. Antimicrobial resistance of β-lactams is attributed to the hydrolysis and alteration of the β-lactam ring in bacteria [9] and is a common resistance mechanism to penicillins [31]. In this study, we observed a higher proportion of β-lactam resistant CoNS compared to the 23% reported in subclinical mastitis cases of dairy cows in Finland [13]. In contrast, all (100%) CoNS isolates from clinical mastitis cases of dairy cattle in Argentina were β-lactam resistant [38]. The high presence of β-lactam resistant isolates and the potential presence of methicillin resistance among CoNS are likely to result in the poor clinical outcome as these isolates are likely to be resistant to other antimicrobial groups including tetracyclines, lincosamides, aminoglycosides, and macrolides [23,43,49,50].

Multidrug resistance and biofilm formation
We observed a high (51%) proportion of MDR-CoNS compared to 45% reported in clinical mastitis cases in India [27]. The high proportion of MDR-CoNS in this study could be attributed to the presence of mecA mediated oxacillin resistance and a high proportion of βlactams resistant CoNS isolates [24][25][26][27]. Nonetheless, the high occurrence of MDR-CoNS further emphasizes the need for judicious use of antimicrobial drugs in the dairy industry in South Africa. There was no significant difference in the presence of MDR among CoNS with biofilm formation compared to those without. To our knowledge, this is the first study to compare antimicrobial resistance patterns and biofilm formation in subclinical mastitis in dairy cattle. In contrast, a study done in clinical patients in human medicine reported a high prevalence of multidrug resistance in biofilm positive CoNS compared to biofilm negative CoNS [51]. In addition, multidrug resistance among S. aureus from human clinical isolates in Korea was more common in isolates with biofilm formation compared to those without [52]. In human medicine, Oliveira et al. [53] reported that bacteria grown as biofilm are up to 1000 times more resistant compared to planktonic. The results of this study suggest that there is no association between biofilm formation and multidrug resistance.
The study is not without limitation, the type of growth media used in the study to assay biofilm formation may have played a role in the low proportion of biofilm identified in this study as the chemical composition of growth media have been shown to influence the expression of biofilm-forming genes in bacteria [33]. In addition, vancomycin resistance was assessed using the disk diffusion method, however, the broth dilution antimicrobial test is the preferred method for analysis of vancomycin resistance [45]. The population of isolates used in this study came from samples submitted to one laboratory. Therefore, the results of this study should not be generalized to the entire dairy industry in South Africa.

Conclusion
Biofilm formation among the CoNS isolates in this study was uncommon and there was no significant different proportion of MDR-CoNS based on the ability to form a biofilm. The majority of CoNS in this study were resistance to penicillins. In addition, most isolates were βlactams resistant and MDR.
The presence of high antimicrobial resistance in this study is a clinical concern and urgent actions should be taken to address the situation. Farmers in South Africa need to be made aware of the high MDR among CoNS and the need for judicious use of antimicrobials in the treatment of CoNS subclinical mastitis. The role of antimicrobial use practises in the development of resistance in subclinical mastitis in the dairy industry should be investigated. The relationship between antimicrobial resistance and biofilm formation in CoNS biofilm formation in the dairy industry is not clear and this concept needs further investigation.

Data source
Coagulase negative Staphylococci isolated from composite milk samples of subclinical mastitis cases that were submitted to the Onderstepoort milk laboratory, Faculty of Veterinary Science, University of Pretoria in 2017 were used. The laboratory receives milk samples from dairy farms across South Africa for routine diagnosis of mastitis. In total, 142 pure CoNS isolates were included in this study.

Species identification using MALDI-TOFMS
All Staphylococcus isolates were subjected to matrixassisted laser desorption ionisation time-of-flight mass spectrometry (MALDI-TOF-MS) as previously described [55,56]. Single pure colonies were transferred onto MALDI plates (Sigma Aldrich, St. Louis, MO) in duplicate and covered with 1 ll of cyano-4-hydroxycinnamic acid in an organic solution (50% acetonitrile and 2.5% tri-fluoro-acetic acid). The preparation was crystallized by air drying at room temperature. Flex Control software (Bruker Daltonics) recorded spectra set for bacterial identification. MALDI Biolayer 3.0 software (Bruker Daltonics) with an integrated pattern-matching algorithm was used to compare generated peak lists against the reference library and a score was generated based on similarity.

Biofilm formation
The biofilm formation of CoNS isolates was investigated using the tissue culture plate method [57]. Isolates were cultured in BTA (blood tryptose agar) for 24 h at 37°C. A loopful of a colony was then inoculated into 5 mL of trypticase soy broth (TSB) for 24 h at 37°C. The inoculated broth was diluted using 1:100 to make a final volume of 2 ml (1.98 ml TSB: 0.02 ml inoculum). Individual wells of sterile 96 well flat bottom polystyrene tissue culture-treated plates (Sigma-Aldrich, Costar, USA) were filled with 200 μL of the diluted broth, positive control and negative control in triplicate. The plates were incubated for 24 h at 37°C. After incubation, the plates were read to obtain optical density (OD) before washing at a wavelength of 570 nm using a micro ELISA (enzymelinked immunosorbent assay) auto-reader (model 680, Biorad, UK). The contents of each well were then removed by gently tapping. The wells were washed with 200-300 ul of phosphate buffer saline (pH 7.4) three times while gently flicking the plates after each wash and left to dry for about 15 min.
Biofilm formed and adhered to the wells were fixed using 150 ul of (96%) methanol for 20 min, where after the contents were removed, and the plates were left to dry for 60 min. Each well was stained with 150 ul (0.2%) of crystal violet for 15 min, 150 ul of (96%) ethanol were then added into each well and covered for 30 min to elute the stain. The plates were read after washing at a wavelength of 570 nm using a micro ELISA auto-reader (model 680, Biorad, UK). This method was repeated 3 times and the OD (optical density) was averaged and subtracted from the cut off value to obtain the final OD for each isolate. A reference strain S. epidermidis ATCC 35984 was used as a control (Thermo Fischer).
The interpretation of the results was divided into the following categories; OD ≤ ODc (Optical density cut-off value) = no biofilm producer; ODc < OD ≤2XODc = weak biofilm producer; 2XODc < OD ≤4XODc = moderate biofilm producer; 4XODc < OD = strong biofilm producer [57]. For the purposes of analysis, weak, moderate, and strong biofilm were classified as biofilm positive.

Antibiotic susceptibility testing
Coagulase negative Staphylococcus isolates were subjected to antimicrobial susceptibility testing against a panel of 11 drugs using the disc diffusion method (Kirby-Bauer method) [45] on Mueller-Hinton agar according to Clinical and Laboratory Standards Institute guidelines. The antimicrobials investigated included 10 mcg ampicillin (AMP), 10 iu penicillin G (P), 30 μg oxytetracycline (OT), 15 μg erythromycin (E), 30 μg chloramphenicol (C), 10 μg streptomycin (S), 5 μg ciprofloxacin (CIP), 30 μg cefoxitin (FOX), 10 μg vancomycin (VAN), 10 mcg clindamycin (DA) and 5 mcg cloxacillin (OB) [45]. Based on the diameter of the zone of inhibition, isolates were classified as sensitive, intermediate or resistant [45]. For the purpose of analysis, the intermediate susceptibility was considered as resistant. Isolates that were resistant to at least one antimicrobial drug were defined as "resistant" while those resistant to three or more antimicrobial categories were defined as "multidrug resistant" [58]. β-lactams resistance was classified as resistant to at least penicillins, cephalosporins or carbapenems [9,59]. The interpretation of vancomycin was based on the criteria by Rezaeifar et al. [60].

Data analysis
The proportions and frequencies of all the variables together with their 95% of confidence intervals were calculated and presented in table format.

Acknowledgments
The author would like to thank the Onderstepoort milk laboratory for providing the isolates for this study and staff members from the Section Veterinary Public Health (VPH) for all the assistance. Lastly, the authors thanks Dr. Joseph Erhabor for assisting with biofilm formation identification.
Authors' contributions LP contributed to the design of the study, data analysis, interpretation, writing of the manuscript, and performed all the laboratory analysis. DN was the main supervisor and contributed to the design of the study, data analysis, the writing and final editing of the manuscript. IM assistance with study design, laboratory analysis and review of the final draft of the manuscript. All authors read and approved the final manuscript.

Availability of data and materials
All the data used to support the findings of this study is available in the manuscript. Raw datasets may also be requested from the corresponding author provided that all ethical requirements have been met.

Ethics approval
The study was approved by the University of Pretoria Animal Ethics Committee (v046-18) as well as the Research Ethics Committee (REC024-18).

Consent for publication
Not applicable.

Competing interests
The authors declare that they have no competing interests.