Mastitis, or inflammation of the mammary gland, represents one of the most economically important diseases of production associated with small ruminant farming systems worldwide (Conington et al., Reference Conington, Cao, Stott and Bunger2008; EFSA, 2013). The most common aetiological agents isolated from cases of intramammary infection (IMI) in sheep flocks are the bacteria Staphylococcus aureus, Mannheimia haemolytica, Streptococcus uberis and several species of coagulate-negative staphylococci (Zadoks et al., Reference Zadoks, Tassi, Martin, Holopainen, McCallum, Gibbons and Ballingall2014; Smith et al., Reference Smith, Willis, Blakeley, Lovatt, Purdy and Green2015). In this study we focus on peracute and acute clinical mastitis which are painful conditions that can result in the death of the animal without immediate treatment (Arsenault et al., Reference Arsenault, Dubreuil, Higgins and Belanger2008). Animals that recover often have damaged mammary tissue and such animals are usually culled. Treatment is dependent on the use of antimicrobials often with limited effect when presented with acute or chronic disease (Mork et al., Reference Mørk, Waage, Tollersrud, Kvitle and Sviland2007). The overuse of antimicrobials is considered a danger to human health due to increasing levels of antimicrobial resistance (AMR), so efforts to reduce their use across different livestock production systems are required. Only limited data are available on the incidence of clinical mastitis across Scottish sheep flocks, the range of bacterial species involved and the prevalence of AMR. In this study we hypothesise that we can fill these knowledge gaps through a combination of questionnaire, farmer-based sampling and laboratory-based analyses.
Material and methods
Farm questionnaire
A questionnaire was supplied to each farm to generate a profile of flocks providing samples from cases of clinical mastitis. Participating farms were contacted directly through e-mail and 23 farms agreed to participate in the project. Farms were selected for initial contact to provide a geographical spread which reflects sheep density from across mainland Scotland. The distribution of flocks is shown in Fig. 1 and details of the farm questionnaire may be found in online Supplementary Table S1.
Figure 1. Distribution across Scotland of sheep flocks that supplied samples.
Sampling kits and sampling information
Three sampling kits were supplied to each participating farm. Each kit included sampling instructions (detailed in online Supplementary materials and methods), disposable gloves, alcohol wipes, a 30 ml sample collection tube and a sealable, padded return envelope. Samples were requested to be stored at 4°C and returned for laboratory processing as soon as possible after collection. The time from sample collection to laboratory processing was between 48 and 72 h.
Bacterial culture
Mammary secretions (10 μl) were plated onto 5% sheep blood agar (SBA) and MacConkey agar No. 3 (Acumedia, Neogen, UK) plates which allowed for categorisation into Gram-positive or Gram-negative isolates. Plates were incubated aerobically at 37°C and examined for bacterial growth after 24 and 48 h. Samples were considered contaminated if more than two species according to colony morphology were identified on each plate. Up to five individual colonies from each plate were sub-cultured in 5% SBA and ancillary tests were carried out for further bacterial classification including oxidase (Oxoid) and catalase tests. Phenotypic differentiation of S. aureus from other staphylococci was performed using the latex slide agglutination Staphytect Plus (Oxoid Limited, Basingstoke, UK) to limit the possibility that different species which are morphologically identical may be present in the same sample.
Bacterial species identification, species specific PCR
DNA was extracted from all bacterial isolates as detailed by Sousa et al. (Reference Sousa, Silva, Igrejas, Silva, Sargo, Alegria, Benito, Gómez, Lozano, Gómez-Sanz, Torres, Caniça and Poeta2014, Reference Sousa, Silva, Manageiro, Ramos, Coelhoi, Goncalves, Caniça, Torres, Igrejas and Poeta2017). PCR amplification of the nuc gene was used to confirm S. aureus species identity (Brakstad et al., Reference Brakstad, Aasbakk and Maeland1992) and a multiplex PCR was used to enable specific identification of Mannheimia isolates (Alexander et al., Reference Alexander, Cook, Yanke, Booker, Morley, Read, Gow and McAllister2008). For other bacterial colonies, species identification was carried out by amplification and sequence analysis of the 16s rRNA, rpoB and prokaryote elongation factor tuf loci as described by Korczak et al. (Reference Korczak, Christensen, Emler, Frey and Kuhnert2004) and Zadoks et al. (Reference Zadoks, Tassi, Martin, Holopainen, McCallum, Gibbons and Ballingall2014). Sequence data was used to determine species identity, using > 94% similarity for species identification or >5% difference between best and next-best match (Mellmann et al., Reference Mellmann, Becker, von Eiff, Keckevoet, Schumann and Harmsen2006; Omaleki et al., Reference Omaleki, Barber, Allen and Browning2010).
Identification of S. aureus sequence type
Multi-locus sequence typing (MLST) was performed on isolates confirmed to be S. aureus to identify individual sequence types (ST). For each isolate, seven housekeeping genes: arcC, aroE, glpF, gmk, pta, tpi and yqiL were amplified and sequenced as described (Enright et al., Reference Enright, Day, Davies, Peacock and Spratt2000; Smyth et al., Reference Smyth, Feil, Meaney, Hartigan, Tollersrud, Fitzgerald, Enright and Smyth2009). Sequence data were analysed using the SeqMan Pro programme in the Lasergene package. Sequences were aligned and compared with reference sequences from the S. aureus MLST Database (https://pubmlst.org/saureus/, Jolley et al., Reference Jolley, Bray and Maiden2018) to assign allele number and sequence type. Novel alleles, not included in the MLST database were submitted to the S. aureus MLST Database to be assigned an allele number and a new sequence type.
Antimicrobial susceptibility testing
Staphylococcus and Mannheimia isolates were tested for susceptibility to 18 antimicrobial agents from the following groups: penicillins, cephalosporins, macrolides, tetracyclines, aminocyclitols, fluoroquinolones, phenicols, sulphonamides and folic acid, beta-lactamase inhibitors. Each bacterial isolate was tested using the Kirby–Bauer disk diffusion method and minimum inhibitory concentration (MIC), according to the European Committee on Antimicrobial Susceptibility Testing (EUCAST, 2018) and Clinical and Laboratory Standards Institute (CLSI, 2018a, 2018b). Further details of antimicrobial susceptibility testing are provided in online Supplementary materials and methods.
Antimicrobial resistance gene identification
Staphylococcus isolates resistant or intermediate-resistant to oxytetracycline were screened by PCR for the presence of four tetracycline-resistance determinants, tet(K), tet(L), tet(M) and tet(O) using primers and amplification conditions as detailed in Sousa et al. (Reference Sousa, Silva, Igrejas, Silva, Sargo, Alegria, Benito, Gómez, Lozano, Gómez-Sanz, Torres, Caniça and Poeta2014).
Results
Flock recruitment and sample collection
Of the 27 farms contacted, 23 farms responded. Of these, 16 forwarded 33 samples of mammary secretions from clinical cases of mastitis. The majority of samples showed alterations consistent with clinical mastitis, such as the presence of blood, clots, flakes or watery milk.
Bacteriology
One or more bacterial species grew following culture of each of the 33 samples for 48 h on SBA or MacConkey agar plates. Three samples provided two bacterial types while the remaining 30 had one, indicating that samples were not compromised by environmental contamination characterised by growth of a larger number of different bacterial types. In total 35 of the 36 isolates were archived as glycerol stocks. The farm code of origin for each sample, corresponding Moredun Research Institute (MRI) isolate number and bacterial species identification are shown in online Supplementary Table S2. One isolate (MRI-SI-026) failed to expand from the initial plate and was not analysed further. Three additional isolates, MRI-T1-020, 024 and 026 obtained from milk samples from the MRI sheep flock in an earlier study provided a total of 38 isolates for species identification, and antimicrobial susceptibility testing.
Bacterial species identification
Species-specific amplification of the nuc gene indicated that 23 of the 38 bacterial isolates were S. aureus (60.5%). Sequence of the 16S rRNA, and subsequent confirmation by multiplex PCR of the Mannheimia species, identified nine as Mannheimia haemolytica (23.7%). Sequence of the 16S rRNA, rpoB and tuf genes identified three isolates (7.9%) as coagulase-negative staphylococci (CNS) while the remaining three isolates were classified as a Bacillus spp, Streptococcus spp and Histophilus somni from sequence analysis of the 16S rRNA and rpoB locus (7.9%: Table 1 and online Supplementary Table S2).
Table 1. Breakdown of bacterial species identified
S. aureus MLST
Four previously reported (8, 130, 133 and 1640) and two novel S. aureus multi-locus sequence types (ST) were identified. The novel ST designated 5215 and 5224. ST 5215 were identified in isolate MRI-SI-017 and included a novel aroE allele (aroE: 722), while ST 5224, identified in MRI-T1-026 included a novel yqiL allele (yqiL: 687). The sequence of both novel alleles may be found in the PubMLST Database (https://pubmlst.org/). The distribution of ST across the 23 S. aureus isolates is shown in Table 2. ST 133 was the most abundant accounting for 14 of the 23 isolates (60.8%) followed by ST 130 which was found in five isolates (21.7%). ST 8, 1640, 5215 and 5224 each occurred once (4.3%). Six farms submitted multiple samples from which more than one S. aureus isolate was identified. All three S. aureus isolates from farms 15 and 6 correspond to ST 133 while both isolates from farm 7 were sequence type 130. In contrast, each of the three isolates from farm 14 were different ST (8, 133, 5215) while the three isolates from farm 17 provided two ST (133, 5224) and the two isolates from farm 12 provided two ST (130, 133).
Table 2. Molecular characterisation of S. aureus sequence type
Antimicrobial susceptibility
Of the 26 staphyloccocci isolates tested for antimicrobial susceptibility, one CNS isolate (AMR-S1-018) was resistant to oxytetracycline and the presence of the tetK resistance gene was confirmed. Five S. aureus isolates were intermediate-resistant to oxytetracycline and one intermediate-resistant to kanamycin. No evidence for AMR was detected in the nine M haemolytica. isolates. All were susceptible with MIC ranging from 0.125–0.064 mg/l with the resistance breakpoint ≥ 1 mg/l. These data are summarised in online Supplementary Table S2.
Integration with questionnaire data
Eighteen flock profiles were reported, and most were upland sheep flocks. Questionnaire data identified a high level of diversity in flock profiles; however, they provide important background information on the scale required for future surveys. Samples were derived from 26 ewes with twin lambs, four with a single lamb, one triplet and two with no details. The spread of breeds and crossbreeds included Scottish Blackface, Lleyn, North Country Cheviots, Texel, Suffolk, Highlander, Aberfield and Beltex sires. The majority of mastitis cases appeared in younger sheep (Table 3). Most flocks lambed inside with six moving to grass post lambing. The incidence of observed clinical mastitis as reported by visual inspection by livestock keepers was low, ranging from 0.3 to 2.4% of ewes in each flock. The peak of clinical mastitis was 4–5 weeks post lambing, and treatments relied on systemic delivery of antimicrobials. Some use of intramammary tubes for antimicrobial delivery and anti-inflammatories was also reported. Penicillin and streptomycin were used by several producers while oxytetracycline and tulathromycin (Draxxin) were also used successfully.
Table 3. Age of ewes with clinical mastitis that provided samples
Discussion
A key aspect of this study was to demonstrate the effectiveness of the information provided to farmers to ensure that environmental bacteria did not contaminate samples received for laboratory culture. Environmental contamination as evidenced by more than two bacterial species following initial culture was not recorded in the 33 samples received. As such we were able to gain a considerable amount of data from a small survey of samples from clinical cases of mastitis from farms across Scotland. As selected farms had participated in previous studies unrelated to mastitis, they may not represent a fully unbiased group. This combination of farm questionnaire, farm-based sampling and laboratory-based analysis provided background information on the scale required for a comprehensive future study and identified areas for improvement in the sampling instructions and genetic analyses to incorporate into a systematic large-scale survey. Importantly this includes requesting information on whether antimicrobial treatment was provided prior to sampling which would impact on the bacterial populations identified. We would also include bacterial genome sequencing to identify and compare virulence gene repertoires between isolates presenting with different mastitis phenotypes.
In this study, 100% of the clinical samples were bacteria positive in contrast to similar surveys in dairy cows where 20–50% of samples from clinical cases may be culture negative (Schukken et al., Reference Schukken, Grommers, Vandegeer and Brand1989; Bradley et al., Reference Bradley, Leach, Breen, Green and Green2007; Keane et al., Reference Keane, Budd, Flynn and McCoy2013). This is likely to reflect differences between production systems in terms of mammary function, milk flow rates and infection levels for lactating dairy cows and sheep raised to produce wool and lambs for meat.
The bacteriological analysis identified S. aureus as the dominant pathogen in over 60% of the clinical cases of mastitis submitted from mostly upland sheep flocks. This is consistent with previous surveys of clinical mastitis cases in sheep across the UK and Northern Europe (Bergonier and Berthelot, Reference Bergonier and Berthelot2003; Mork et al., Reference Mørk, Waage, Tollersrud, Kvitle and Sviland2007). S. aureus infections are also common in dairy cows where they can cause clinical mastitis but are more often associated with subclinical chronic mastitis (Peton and Le Loir, Reference Peton and Le Loir2014). The lower prevalence of CNS and M. haemolytica identified in this study is also consistent with previous analyses. The failure to identify Streptococcus isolates is likely to reflect the clinical nature of the cases as Streptococcus has been reported to be associated more with subclinical chronic forms of mastitis (Bergonier et al., Reference Bergonier, De Cremoux, Rupp, Lagriffoul and Berthelot2003; Hariharan et al., Reference Hariharan, Donachie, Macaldowie and Keefe2004).
In this study, two new S. aureus ST associated with a novel aroE allele (722) and a novel yqiL allele (687) were identified and assigned ST 5215, and ST 5224 respectively. While we identified a total of six ST in this study, they are dominated by ST 133 and ST 130 which correspond to 14 (61%) and 5 (22%) of the 23 isolates respectively. These dominant ST's are broadly distributed in the isolates collected from across Scotland. Both ST's which are included in clonal complex's CC133 and CC130 were also the most common identified in milk samples from Swiss sheep and goats showing signs of mastitis (Merz et al., Reference Merz, Stephen and Johler2016) suggesting that they are widely distributed in small ruminants across Europe (Porrero et al., Reference Porrero, Hasman, Vela, Fernandez-Garayzabal, Dominguez and Aarestrup2012; McMillan et al., Reference McMillan, Moore, McAuley, Fegan and Fox2016) and North Africa (Achek et al., Reference Achek, El-Adawy, Hotzel, Tomaso, Ehricht, Hamdi, Azzi and Monecke2020). ST's 133 and 130 appear better adapted to small ruminants as they are less common in mastitis samples from dairy cows or from human isolates (Shepheard et al., Reference Shepheard, Fleming, Connor, Corander, Feil, Fraser and Hanage2013; Naushad et al., Reference Naushad, Nobrega, Naqvi, Barkema and De Buck2020).
To our knowledge, this study is the first report of ST 1640 isolated from a case of clinical mastitis in sheep. ST 1640 had previously been identified in nasal samples from healthy sheep and deer in Scotland (Fountain et al., Reference Fountain, Blackett, Butler, Carchedi, Schilling, Meredith, Gibbon, Lloyd, Loeffler and Feil2021) and from nasal and faecal samples from horses in Spain (Mama et al., Reference Mama, Gomez, Ruiz-Ripa, Gomez-Sanz, Zarazaga and Torres2019) suggesting a broad host range. ST 8 and its associated clonal complex CC8 has previously been described in isolates from human infections (Jackson et al., Reference Jackson, Davis and Barrett2013) and based on genotyping and phylogenetic analysis, it is thought to have transferred and adapted to dairy cattle relatively recently (Boss et al., Reference Boss, Cosandey, Luini, Artursson, Bardiau, Breitenwieser, Hehenberger, Lam, Mansfeld, Michel, Mösslacher, Naskova, Nelson, Podpečan, Raemy, Ryan, Salat, Zangerl, Steiner and Graber2016). The identification of ST 8 in a sheep mastitis isolate indicates that the process of transfer and adaptation is continuing into small ruminants.
Three of the six farms that supplied multiple samples had multiple ST, with two of these farms having different ST in each of the three samples submitted. Two of the low frequency ST associated with this study, 8 and 5215 were identified on a single farm. The human origin of ST8 is described above while ST 5215 is a single locus variant of the common sheep ST 133. The relationships between multiple ST and individual farm management systems are questions for a larger study, however the presence of other ruminant livestock species on both farms may be linked with the multiple ST identified.
In this survey, AMR in the Gram-positive and Gram-negative bacteria isolated from clinical cases of mastitis was limited to a single CNS isolate demonstrating resistance to oxytetracycline and in which the tetK resistance gene was confirmed. No evidence of methicillin-resistant S. aureus (MRSA) was observed as all isolates were sensitive to antimicrobials oxacillin and cefoxitin which may be used as markers for MRSA (Fernandes et al., Reference Fernandes, Fernandes and Collignon2005). These data indicate that levels of AMR in bacterial samples from these clinical cases of mastitis is low which is in agreement with recent data from the UK (Silva et al., Reference Silva, Phythian, Currie, Tassi, Ballingall, Magro, McNeilly and Zadoks2020) and Norway (NORM-VET, 2017).
The survey described here collected data from 29 flocks from across Scotland. This confirmed the high degree of diversity within the sheep industry across Scotland, which is reflected in the variety of breeds, lambing systems, and numbers of mastitis cases. With such diversity only a much larger survey conducted over several lambing seasons would provide data on the scale required to draw significant conclusions. Importantly, we have demonstrated that samples provided by livestock producers are free from environmental contamination which will simplify sampling at the scales required for a future study.
Supplementary material
The supplementary material for this article can be found at https://doi.org/10.1017/S0022029924000517
Acknowledgements
The authors acknowledge the administrative support of Penny Middleton at LHS and all livestock producers who completed the questionnaire and supplied samples for analysis.
