Footrot, or infectious pododermatitis, is a hoof infection commonly found in sheep, goats, and cattle. Although footrot was first described more than 180 years ago, it is a complex disease still endemic in many countries (Zanolari et al., 2021). Ovine footrot is caused by an interaction of two anaerobic, Gram-negative bacteria: D. nodosus (formerly Bacteroides nodosus) and F. necrophorum (formerly Sphaerophorus necrophorus). However, Dichelobacter nodosus is the primary causative agent of footrot in sheep (Kennan et al., 2014). Initial colonization of the hoof by opportunistic bacteria, including the ruminant digestive tract commensal bacterium Fusobacterium necrophorum, is followed by infection with D. nodosus, and interaction between these two pathogens causes footrot in sheep. Ovine footrot is characterized by the separation of the keratinous hoof from the underlying tissue and causes severe lameness (Kennan et al., 2011). The annual costs of footrot were estimated at £24.4 million in UK (Nieuwhof and Bishop, 2005) and $18.4 M in Australia (Sackett et al., 2007; Smith et al., 2022), which corresponds to £1.32 and £0.15 per living ewe and lamb, respectively (Nieuwhof and Bishop, 2005). In Switzerland, annual costs for footrot were estimated at CHF33 million for the sheep population (Zingg et al., 2017). Affected sheep frequently experience pain, discomfort, and reduced mobility which affects their ability to access feed (Abbott and Lewis, 2005). Thus, it is not surprising that affected sheep can experience reduced growth rates and wool production. For instance, one study showed that lambs with footrot reached slaughter weight 31.9 days later than lambs without footrot (Zingg et al., 2017).

Currently, a variety of different footrot management and treatment approaches are utilized world-wide. These include foot trimming, foot baths/foot soaks with zinc sulfate and copper sulfate, injection of antibiotics (penicillin and streptomycin combinations), and topical medications or vaccination against D. nodosus. None of these interventions is perfect, and the best results are obtained when several methods are combined (Bennett and Hickford, 2011). Variation in management and treatment reflects variation in stocking rate (of importance with a contagious disease), the size of flocks, the cost of labor for labor-intensive management practices, and the cost and availability of remedies and acceptability of the various management and treatment regimes in different markets. There remains a need for additional minimally labor-intensive tools to reduce both losses and treatment costs of ovine footrot.

Since pathogen persistence in the environment depends on the host (Clifton et al., 2019), one possibility is to use genetic resistance to footrot as a prevention tool. Demonstration of genetic variance can provide a sense of the promise of such a strategy, and studies dissecting genetic variation involved in degrees of footrot resistance have been ongoing for the last 4 decades (Escayg et al., 1997). Moderate heritability has been estimated for ovine footrot, generally between 0.20 and 0.30 depending on breed and phenotypic scoring method (Emery et al., 1984; Raadsma et al., 1994; Nieuwhof et al., 2008a; Raadsma and Conington, 2011), demonstrating that footrot resistance is a heritable trait. Breed differences have been observed, including that Merino sheep are particularly susceptible to footrot, while others such as Romney are more resistant (Emery et al., 1984). These results suggest there is potential for development of genetic tools to improve footrot resistance, as simple phenotypic selection has led to long-term genetic improvement (Parker et al., 1983; Conington et al., 2008).

To enhance genetic gains, the identification of specific genes and molecular markers associated with footrot resistance is needed. Although a few genetic markers for natural resistance to footrot have been identified (Litchfield et al., 1993; Escayg et al., 1997; Niggeler et al., 2017), there is still a paucity of information about genetic variation in susceptibility to ovine footrot. The role of the major histocompatibility complex (MHC) in modulating immune responses, and subsequently disease susceptibility for both the Class I and Class II regions, has been investigated in relation to footrot resistance (Litchfield et al., 1993; Escayg et al., 1997; Hickford et al., 2004; Raadsma and Dhungyel, 2013). For genome wide association studies, only two studies (Mucha et al., 2015; Niggeler et al., 2017), have been reported and they were focused solely on European wool sheep breeds. Katahdins and other hair sheep show distinct genetic heritage (Spangler et al., 2017) which has manifested in demonstrated differences in disease resistance traits between hair and wool sheep (Vanimisetti et al., 2004). Some have suggested that hair breeds like Katahdin and Blackbelly might be more resistant to footrot than other sheep (Yazwinski et al., 1979; Zajac, 1995; Bishop and Morris, 2007; Azarpajouh, 2014; de Almeida, 2018), and there has not been an investigation of the genetics of ovine footrot susceptibility at the genome-wide level in these breeds. Therefore, the main aim of this study was to undertake a genome-wide association study (GWAS) to investigate Single Nucleotide Polymorphisms (SNPs) and identify genes linked to footrot susceptibility in North American hair and wool sheep.

There were 92 sheep with healthy scores of 1 (no footrot in any foot), 52 with intermediate scores of 2 (at least one foot with a score of 2), and 107 with scores of 3 for footrot disease (at least one foot with a score of 3). Out of 553,197 SNPs, the call rate screen removed no SNPs with call rates <95%. The minor allele frequency screen removed 37,043 SNPs, and the Hardy-Weinberg equilibrium test at p < 10^–25 removed an additional 519 SNPs, which left 515,635 SNPs after quality control for further analysis.

The genomic inflation factor (lambda) for the overall genome-wide association analysis was 1.01. Figure 1 shows a Manhattan plot of genome-wide association, and a quantile-quantile plot is given in Supplementary Figure S2. Detailed information on the top loci demonstrating genome-wide significant associations are highlighted in Table 1. Additional information on loci with genome-wide suggestive association are shown in Table 2.

Since footrot susceptibility is heritable (Parker et al., 1983; Raadsma et al., 1994; Nieuwhof et al., 2008b; Raadsma and Conington, 2011), a genetic selection approach may help to decrease lesion score and the number of animals affected through selective breeding if loci associated with footrot can be identified. The aim of the present study was to perform the first GWAS for footrot in North American wool sheep and also the first GWAS for hair sheep. Our analysis yielded an appropriate model fit with a genomic inflation factor of 1.01. We identified 3 genome-wide significant and 33 genome-wide suggestive loci associated with ovine footrot susceptibility on 13 different autosomes (Table 1; Table 2). Below, we explore the regions surrounding these regions, the potential involvement of nearby genes in ovine footrot, and compare our results to those identified in other studies to date.

Multiple genome-wide significant positional candidate genes were identified in the present study with functions in the immune system (Table 1). First, two genome-wide associate SNPs (rs421352693; p = 1.08 × 10⁻⁹ and rs411314769; p = 3.74 × 10⁻⁷; Table 1) were located in an intergenic region on OAR25. Although this locus did not contain a positional candidate gene, this region could affect susceptibility to footrot through the regulation of distant genes due to the presence of regulatory elements, non-coding RNAs, or other features (Scacheri and Scacheri, 2015; Georges et al., 2019). It is interesting that the most significant SNP (rs421352693) was located less than 10 Kb away from a seven base-pair highly conserved element from an analysis of 91 eutherian mammal genomes (Ensembl release 97; Zerbino et al., 2018). The specific function of this highly conserved element has not been fully elucidated, but it is located between ARID5B and RTKN2 (Jiang et al., 2014), both of which have been linked to roles in regulation of immune responses. ARID5B encodes a DNA-binding protein with roles in NK cell function (Cichocki et al., 2018), cancers of both B and T lymphocytes (Healy et al., 2010; Zeng et al., 2014; Leong et al., 2017), and autoimmune disease (Wang et al., 2012; Yang et al., 2013). The RTKN2 gene is expressed in lymphocytes (Collier et al., 2004), induces an NF-kB-dependent hold on apoptosis (Collier et al., 2009) that can change counts and function of available immune cells, and has been implicated in autoimmune disease (Myouzen et al., 2012). In the cases of both of these genes, roles in immune cells suggest possible immune mechanisms for differential control of footrot in sheep. There are ongoing efforts to identify and annotate regulatory elements in sheep and other ruminants through the Functional Annotation Animal Genomes (FAANG) consortium, among others (Andersson et al., 2015; Tuggle et al., 2016). The functional importance of this genome-wide association may become clearer once data from such annotation projects are complete.

The guanylate-binding protein 6 (GBP6) gene is located on OAR1 (Jiang et al., 2014) within a peak defined by two SNPs, one genome-wide significant (rs159679616; p = 3.91 × 10⁻⁷; Table 1) and one genome-wide suggestive (rs159679642; p = 1.02 × 10⁻⁶; Table 2). Guanylate-binding proteins (GBPs) are abundantly expressed cellular proteins with seven highly homologous members in sheep, termed GBP1 to GBP7, expressed in response to interferon-gamma (IFN-γ) and other pro-inflammatory cytokines (Olszewski et al., 2006; Kim et al., 2011; Praefcke, 2018). GBP6 stimulates phagocyte oxidase, antimicrobial peptides, and autophagy effectors in an immune response capable of killing multiple types of bacteria (Kim et al., 2011). Furthermore, it has been proposed that GBPs might promote lysis of vacuoles, thereby triggering detection of pathogen associated molecular patterns (PAMPs) and further immune responses, as well (Meunier et al., 2014; Meunier and Broz, 2015).

The SNP association peak on OAR1 was also very near the trichohyalin (TCHH) gene (Table 1; Table 2). Trichohyalin crosslinks with keratin intermediate filaments to provide mechanical strength in hair follicles, as well as hooves (O’Keefe et al., 1993; Steinert et al., 2003). This could affect hoof formation in ways that predispose or protect sheep from ovine footrot. Thus, this most significant pair of SNPs result points to involvement of hoof structure as well as to stimulation of the immune response as potential mechanisms for susceptibility to ovine footrot. Other immune related QTLs have been reported in the literature from the same region where GBP6 and TCHH are located on OAR1. For instance, a facial eczema susceptibility QTL has been reported based on microsatellite marker genotyping on OAR1 (Phua et al., 2009). It is possible that this association could reflect contributions from both genes (GBP6 and TCHH) through one or more regulatory elements, but future work would be required to elucidate the specific underlying functional mutation(s) in this region.

A cluster of six SNPs spanning just over 100 Kb on OAR3 included the lowest p-value genome-wide suggestive association results (Table 2). Of these, the SNP with the lowest p-value (p = 6.6 × 10⁻⁷) was rs415053617 (Table 2), which was located less than 10 Kb from SLC38A1 (Oguchi et al., 2012). The SLC38A1 gene encodes a glutamine transporter expressed in hair cells (Oguchi et al., 2012) and extracellular exosomes (Willms et al., 2016) and is involved in oxidative stress responses (Ogura et al., 2011). Taken together, these data suggest SLC38A1 might inhibit D. nodosus and/or F. necrophorum by limiting glutamine nutrient availability, either before or during infection.

Two GWAS have been performed previously for ovine footrot susceptibility in sheep with wool. Niggeler et al. (2017) identified the only other genome-wide association with footrot (aside from the present work) on OAR2 near the multi-PDZ domain protein 1 (MPDZ) gene in Swiss White Alpine sheep. The genome-wide associated SNP was rs418747104, and additional genome-wide suggestive markers nearby included rs426927857 and rs406749947. However, this locus was not associated with footrot in our dataset (all p > 0.05). The other prior GWAS for ovine footrot in sheep with wool (Mucha et al., 2015) identified no genome-wide associations but did identify chromosome-wise significant associations on ovine chromosomes 4, 8, 14, 17, 18, 24, and 26. None of these loci were associated with footrot in this study (all p > 0.05). The non-overlapping associated genomic regions among different experiments may be caused by differences in pathogen characteristics on different continents or by breed differences, especially since this is the first study to include hair sheep.

In addition to GWAS, prior candidate gene studies have been conducted to identify associations with ovine footrot. Candidate gene studies have identified associations with footrot in DQA1, DQA2, DQB, and DRA in the Major Histocompatibility Complex (MHC) on ovine chromosome 20 (Escayg et al., 1997; Gelasakis et al., 2013). This is a complex region with paralogs derived from tandem repeats and even variable numbers of genes (Herrmann-Hoesing et al., 2008; Gao et al., 2010; Bickhart and Liu, 2014). Haplotypes in this region can be quite long, and some are ancient in origin (Raymond et al., 2005). The specific markers used in prior work are difficult to assess in our study because of the widely differing marker types. However, no markers in that region on OAR 20 achieved genome-wide significant or genome-wide suggestive support in this study (Table 1; Table 2).

This is the first genome-wide study examining footrot susceptibility using hair sheep and the first GWAS to identify multiple genome-wide associations with footrot. These results provide insight into mechanisms that may affect footrot susceptibility and resistance. In particular, both genome-wide significant and genome-wide suggestive associations illustrated themes of immune function, nutrient availability, and hoof formation and integrity. Thus, this study met its objective in improving understanding of host genetics of footrot susceptibility. In addition to biological insights, these results provide a foundation for future work developing predictive genetic marker tests. Such technology offers the possibility of disproportionate benefits to flock health through selection against the most susceptible animals before infection, disease, and transmission to other animals (Galvani and May 2005; Lloyd-Smith et al., 2005). However, more research is needed to identify the specific functional mutations in linkage disequilibrium with the markers in this study. In addition, the functional mutations will need to be validated and examined for potential correlated responses to selection, including production traits (White and Knowles, 2013) such as some already identified for immune loci in sheep (Cinar et al., 2016; Cinar et al., 2018). Further, additional animals can be used to examine potential for genomic selection to leverage these data into additional breeding applications that may not require identification of causal mutations. In these ways, sheep production can benefit from reliable, predictive genetic tests for which selection does not lead to deleterious effects on other traits.