Ovine footrot: new insights into bacterial colonisation

Ovine footrot is characterised by interdigital dermatitis (ID) and by the separation of the skin and hoof horn (under-running footrot). Dichelobacter nodosus is the essential pathogen causing footrot; the role of other microorganisms in this disease remains unclear. The aims of this study were (i) to investigate the colonisation of D nodosus, Fusobacterium necrophorum and Treponema species in biopsies from the ovine interdigital skin of healthy, ID and footrot-affected feet and (ii) to characterise the virulence of D nodosus strains in those biopsies. Postslaughter biopsy samples (n=241) were collected and analysed by real-time PCR to determine prevalence and load of the different bacterial species. The highest prevalence and load of D nodosus were found on feet with ID. The vast majority of samples contained virulent D nodosus and some samples contained both virulent and benign D nodosus. Notably, the more pathogenic subspecies of F necrophorum was found in samples from UK sheep. Our findings provide further insights into the role bacterial colonisation may play in the early stage of ID and in the progression towards footrot.

Ovine footrot is a major cause of lameness affecting sheep welfare worldwide (Goddard and others 2006); it is characterised by two different clinical presentations, interdigital dermatitis (ID) and under-running footrot. ID is an initial inflammation of the interdigital skin where the superficial epidermal layers are inflamed, damaged and slough off irregularly and it may develop into under-running footrot, which is characterised by the separation of the hoof horn from the sensitive underlying tissue (Beveridge 1941, Egerton andothers 1969). In Australia, mild/ benign footrot is also used synonymously for ID and underrunning footrot is called virulent footrot (Raadsma and Dhungyel 2013).
Footrot is a complex disease with Dichelobacter nodosus, a Gram-negative anaerobic bacterium, as the essential pathogen causing under-running footrot (Egerton and others 1969, Kennan and others 2001, Han and others 2008. D nodosus load was found to be already increased in ID before the development of under-running footrot, thereby suggesting that D nodosus load drives the early stages of infection others 2014, 2015). Additionally, the occurrence of this disease is associated with different factors such as the virulence of D nodosus strains (Kennan and others 2010), farm management (Green and others 2007), environmental conditions others 2005, Muzafar andothers 2016) and initial damage in the epithelium of the interdigital skin (Beveridge 1941, Egerton andothers 1969).
Whole-genome sequencing demonstrated that D nodosus has a global conserved bimodal population, correlating with virulent and benign phenotypes (Kennan and others 2014). A large number of virulent D nodosus strains were identified in Australia (Kennan and others 2014), while in Scandinavian countries, such as Sweden, mainly benign strains have been found (Frosth and others 2015). In UK flocks, virulent D nodosus was more prevalent than benign in swabs from sheep with ID and footrot (Moore and others 2005). Virulent and benign D nodosus strains differ in their ability to degrade the extracellular matrix of the host due to enzymatic activity of extracellular proteases (Riffkin and others 1995). The acidic protease AprV2 is responsible for the overall elastase activity of virulent strains and was shown to be essential for the development of footrot, while the acidic protease AprB2 is associated with a benign phenotype (Kennan and others 2010). Importantly, presence of virulent D nodosus strains does not always correlate with severity of clinical presentations since virulent D nodosus has also been identified in sheep without any clinical sign and in ID cases others 2005, Stäuble andothers 2014).
In addition, Fusobacterium necrophorum, Treponema species and a range of other bacterial genera have been identified in the ovine interdigital skin , Bennett and others 2009, Calvo-Bado and others 2011, Frosth and others 2015. The role of F necrophorum in this disease still needs to be fully understood, with two hypothesis currently discussed: (i) F necrophorum is important to establish ID before D nodosus infection and hence initiates the disease (Egerton and others 1969) or (ii) F necrophorum is involved in the persistence and severity of footrot, once the under-running lesion has developed, playing a role as an opportunistic, secondary pathogen others 2014, 2015). F necrophorum is divided into subspecies necrophorum and funduliforme, the first is described to be more pathogenic (Tan and others 1996).
Veterinary Record (2016) doi: 10.1136/vr.103610 Treponema species are usually free-living spirochetes, but they have been associated with contagious ovine digital dermatitis (CODD) (Sullivan and others 2015) and bovine digital dermatitis (BDD) (Gomez and others 2012). BDD and CODD have polytreponemal aetiology with different Treponema species involved in their pathogenesis (Sayers andothers 2009, Sullivan andothers 2015). Initial identification of spirochetes in ovine footrot lesions was reported by Beveridge (1941). Recent studies identified Treponema species in a sheep with ID lesions from a flock with footrot history (Calvo-Bado and others 2011) and were found in both flocks and feet, with and without footrot (Frosth and others 2015). Hence, it suggests that further investigation to elucidate their role and whether different species of Treponema can be identified in ovine footrot is warranted.
Taken together, current data suggest that footrot is a polymicrobial disease and D nodosus and other microorganisms might have a synergistic relationship. However, the role of bacterial diversity and load and how that differ between healthy, ID and footrot feet remains unclear. In this context, the aims of this study were (i) to investigate the colonisation of D nodosus, F necrophorum, Treponema species and eubacteria, and (ii) to characterise the virulence of D nodosus strains in a cross section of healthy, ID and footrot abattoir biopsies from the ovine interdigital skin.

Materials and methods Collection of tissue biopsies
This study included 241 ovine interdigital postslaughter biopsies collected at an abattoir using a convenience sampling approach due to variable availability of the clinical conditions at slaughter. The entire sample set included 79 healthy, 39 mild ID (slight lesion with ≤5 per cent of the interdigital skin space affected), 26 moderate/severe ID (interdigital skin lesions with ≥5 per cent of the interdigital space affected) and 97 footrot samples (Table 1). Of these, 40 animals had all four feet sampled, total of 160 samples. During two visits to the abattoir (November 1, 2013, and November 4, 2013), it was not possible to follow the same animal on the processing line; therefore, 78 feet biopsies were collected without the information if they belonged to the same sheep (Table 1). Since the animals were sampled in the processing line of the abattoir, no information regarding sheep breed or other characteristics was available for this study.
At the abattoir, all feet disease status was scored by two different scorers with one of the scorers present during all visits in order to standardise the sampling and scoring method; for details, see Table 1. The scoring system was adapted from Parsonson and others (1967), allowing classification into healthy, ID or footrot feet according to established scoring criteria: absence of interdigital skin lesion=healthy; slight interdigital skin lesion (≤5 per cent affected)=mild ID; moderate-to-severe ID lesion (>5 per cent affected); presence of under-running lesion=footrot.
Tissues were collected as described previously (Davenport and others 2014) and placed into RNALater (Sigma-Aldrich, Saint Louis, USA) at 4°C before long-term storage at −20°C.

DNA extraction and real-time PCR assays
For enzymatic digestion, each tissue was cut into small pieces and incubated with 180 μl of tissue lysis buffer (ATL) and 20 μl of proteinase K (20 mg/ml) (Qiagen, Hilden, Germany) at 56°C for three hours. DNA was isolated using the QIAamp cador kit according to manufacturer's recommendations and eluted in 50 μl AVE buffer (Qiagen). The final DNA concentration was determined using NanoDrop (ND-1000 (Thermo Fisher Scientific, Waltham, USA). Bacterial load was quantified using real-time PCR based on 16S rRNA gene for eubacteria (Strub and others 2007) and D nodosus (Frosth and others 2012) and the intergenic spacer region 2 containing a tRNAIle gene for Treponema species (Frosth and others 2015). Real-time PCR for F n subspecies necrophorum and F n subspecies funduliforme targeted the gyrB gene (Frosth and others 2015). Differentiation between virulent and benign D nodosus was performed based on the presence of aprV2 (virulent) and aprB2 (benign) genes (Frosth and others 2015). D nodosus and eubacteria assays were performed using PCR Lightcycler 480 (Roche Applied Science, Penzberg, Germany). Virulent and benign D nodosus, F necrophorum and Treponema species assays were carried out in an Applied Biosystems 7500 Fast Real-Time PCR System (Thermo Fisher Scientific).

Statistical analysis
Fisher's exact test was performed for bacterial prevalence and one-way analysis of variance followed by Dunn's multiple comparisons test for bacterial load using GraphPad Prism (V.6.0, La Jolla, USA). CIs of the prevalence data were calculated using GraphPad Software (http://graphpad.com/quickcalcs/confInterval2/). A P value ≤0.05 was considered significant.

Results
Prevalence and load of D nodosus, F necrophorum and Treponema species in tissues from the ovine interdigital skin The prevalence of D nodosus, F necrophorum, Treponema species and eubacteria was investigated in the ovine interdigital skin biopsies. All samples were positive for eubacteria (100 per cent of prevalence). Both total D nodosus and virulent D nodosus were significantly more prevalent in mild ID (P<0.05 and <0.01, respectively), moderate/severe ID (P<0.001 and <0.0001, respectively) and footrot (P<0.05 and <0.01, respectively) in comparison with healthy feet, with highest prevalence in moderate/ severe ID samples. Moreover, total D nodosus and virulent D nodosus were significantly more prevalent in moderateto-severe ID than in footrot samples (Fig 1a; see online supplementary appendix 1). In contrast, benign D nodosus was only detected in 7 per cent (17/241) of the samples (Fig 1b). Mixed populations of benign and virulent D nodosus strains were found in a small number of samples across all clinical conditions (Fig 1b).  subspecies. F necrophorum was significantly more prevalent in footrot than in healthy feet (P<0.05) (Fig 1c). Presence of both D nodosus and F necrophorum in the same tissue sample or virulent D nodosus and F necrophorum in the same tissue was significantly higher in footrot compared with healthy feet (P<0.01 and <0.01, respectively). Treponema species prevalence was very low (8 per cent, 20/241) and similar across all clinical conditions (Fig 1d) (see online supplementary appendix 1). Similar proportion of eubacterial DNA was detected at around 0.06 per cent±0.020 (mean±se of the mean) of total DNA for all samples, with 0.07 per cent±0.041 for healthy samples, 0.028 per cent±0.007 mild ID, 0.027±0.011 for moderate/severe ID and 0.073±0.039 for footrot samples. D nodosus load was significantly higher in moderate/severe ID and footrot in comparison to healthy feet (P<0.0001 for both) (Fig 2a). Virulent D nodosus load was significantly increased in mild ID, moderate/severe ID and footrot compared with a healthy feet (P=0.001, P<0.0001 and P<0.0001, respectively), with highest load in moderate/severe ID (Fig 2b). F necrophorum load was significantly increased in footrot but not in ID samples (P=0.022) (Fig 2c). The highest Treponema species load was found in footrot followed by healthy feet (Fig 2d).
In summary, while eubacterial loads were similar in all feet, both prevalence and load of total and virulent D nodosus were highest in moderate-to-severe ID, while F necrophorum were highest in footrot samples.

Discussion
In this study, the authors provided further insights into the bacterial colonisation present in healthy, ID and footrot ovine feet. The authors found similar patterns regarding the prevalence and load of D nodosus and F necrophorum in postslaughter biopsies from the interdigital space as previous studies in UK sheep flocks using swabs and biopsies (Moore and others 2005, Calvo-Bado and others 2011, others 2014, 2015).
As expected, the highest D nodosus prevalence and load found in this study was in ID samples, hence supporting the hypothesis that D nodosus drives the development of the early stages of footrot others 2014, 2015). Interestingly, D nodosus was found in a large proportion of biopsy samples from healthy feet (58 per cent, 46/79), suggesting it might be present in the stratum corneum (horny layer) but not necessarily causing disease. It is also possible that these visibly healthy feet might have had subclinical footrot and may have developed  (Egerton and others 1969), sheep breed (Emery and others 1984) and presence of co-infecting bacteria such as F necrophorum others 1969, Roberts and. In this study, the majority of D nodosus present in the ovine interdigital skin biopsies were virulent strains. Similarly, high prevalence of virulent D nodosus in the UK sheep was identified previously using gelatinase gel protease assay (Moore and others 2005). Therefore, these studies demonstrate that virulent strains are currently circulating in UK flocks. In contrast, in Sweden where under-running footrot is not endemic, most of the D nodosus were found to be benign (Frosth and others 2015). The authors found a mixed population of benign and virulent D nodosus strains in the same feet; a potential synergistic role of benign and virulent strains still needs to be investigated.
F necrophorum prevalence and load were higher in footrot than in ID and healthy samples. These results, together with other published data that also found an increased presence of F necrophorum in footrot lesions (Beveridge 1941, Bennett and others 2009, Witcomb and others 2014, support the hypothesis that F necrophorum contributes to the pathogenesis of under-running footrot. F necrophorum was presumed to facilitate D nodosus invasion (Egerton and others 1969); in the present study, the authors found that the presence of both D nodosus and F necrophorum in the same tissue was significantly higher in footrot than in healthy feet; nevertheless, the exact nature and the role of the synergy between F necrophorum and D nodosus remains unclear.
Only a small number (9 per cent, 7/79) of healthy biopsy samples were positive for F necrophorum in this study. Similarly, Witcomb and others (2015) found low prevalence of F necrophorum in swabs (8 per cent, 1/12) and biopsies (8 per cent, 1/12) from healthy feet, but in an earlier study where swabs were repeatedly collected from 18 sheep during five weeks, F necrophorum was found in 62 per cent (140/225) of healthy feet (Witcomb and others 2014). This suggests that the prevalence of F necrophorum in healthy feet varies according to sampling structure and collection methods. F necrophorum is a commensal in the alimentary tract (Smith and Thornton 1997) and might be present in faeces contaminating ovine feet. Moreover, it was also detected on swabs taken from the oral cavity of sheep and suggested it might be transmitted from the mouth of sheep to the paddock (Bennett and others 2009); hence, the significance of F necrophorum in healthy ovine interdigital skin remains unclear; it may colonise healthy skin as a commensal microorganism without causing any skin disease, while in damaged skin, F necrophorum colonisation may initiate ID and, thus, predispose the invasion of D nodosus. Whether D nodosus essentially requires F necrophorum colonisation to facilitate its skin invasion remains unclear.
F necrophorum is divided into subspecies necrophorum and funduliforme, the first is described to be more pathogenic due to a higher lipopolysaccharide content and higher production of leukotoxin (Tan and others 1996). In this study, the majority of positive samples for F necrophorum was subspecies necrophorum. Previous studies investigating this bacterium in UK flocks did not differentiate the subspecies of F necrophorum others 2014, 2015). Therefore, despite the fact that this sample set is small, this is the first study suggesting that F necrophorum subspecies necrophorum may be the more prevalent subspecies circulating in UK flocks. Since F n subspecies necrophorum is described to be more virulent than F n subspecies funduliforme (Tan and others 1996) and considering the fact that this bacterium may exacerbate footrot lesions, there might be an association between the high prevalence of severe footrot lesions in the UK and the presence of F n subspecies necrophorum. In contrast, in Swedish flocks where most of the footrot lesions were associated with mild footrot, F n subspecies funduliforme was more prevalent than F n subspecies necrophorum (Frosth and others 2015).
Spirochaetes have also been identified in ID and/or footrot lesions (Beveridge 1941  were positive for Treponema species with similar prevalence in healthy, ID and footrot feet. Similarly, low detection of Treponema species in ovine biopsies from UK sheep was also reported by Calvo-Bado and others (2011); moreover, no significant association between Treponema species and footrot was reported by Frosth and others (2015). Hence, whether the low detection of Treponema species reflects its importance in the footrot pathogenesis remains an open question to be further elucidated. The authors amplified treponemal DNA using a genus-specific qPCR and not a species-specific assay, hence detecting free living as well as pathogenic Treponema species; therefore, more studies are warranted to characterise the Treponema species commonly present in ovine footrot. In contrast to early investigations reporting that an initial infection with D nodosus is often followed by an infection with Treponema species (Beveridge 1941, Thomas 1962, the authors only found 3 per cent of the biopsies (8/241) positive for both virulent D nodosus and Treponema species. A limitation of using abattoir samples is that it is impossible to investigate the progression of the disease and thus verify whether healthy or ID feet positive for D nodosus would develop footrot lesions. On the other hand, an advantage of using abattoir samples is the ability to collect biopsies from animals that are slaughtered for other purposes than this study and detect bacteria localised within tissues.

Conclusions
The results presented in this study, together with other published data, confirm that D nodosus is mainly associated with ID stage and F necrophorum with footrot stage, thereby supporting that D nodosus drives the early stages of footrot and F necrophorum plays a role in the pathogenesis of ovine footrot. Moreover, virulent D nodosus population is more prevalent than benign in UK flocks. Treponema species were detected in few samples; hence, further studies are warranted to provide more detailed information about the role Treponema species may have in ovine footrot. Additionally, this study reports novel results regarding the higher prevalence of F necrophorum subspecies necrophorum than subspecies funduliforme in sheep from the UK, and a mixed population of virulent and benign D nodosus present in the same skin biopsy.