Metagenomic analyses of the bovine foot skin microbiome; associations with development of Bovine Digital Dermatitis

Objectives: Previous studies have described the microbiome of the different lesion stages of Bovine Digital Dermatitis (BDD). The present study was designed to examine the healthy bovine foot-skin microbiome and determine differences in microbial populations associated with future development of BDD lesions.

Materials and Methods: Swabs were taken from the heel bulb region of the back-left foot of 259 Holstein-Friesian cows from three commercial UK dairy farms 3-4 weeks prior to calving. Feet were re-examined at one, four and 8-10 weeks post-calving to record any lesions that had developed and classify cows into four foot-health groups: those that remained healthy (HtHt), those that became infected (HtIn), those that were always infected (InIn) and those that were infected at enrolment but recovered (InHt).

16SrRNA gene amplicon sequencing was carried out for all samples using the Illumina® HiSeq 2500 platform, and taxonomic assignment of OTUs carried out using QIIME and the RDP classifier. Chao 1, Shannon and Simpson indices were calculated to describe alpha-diversity, and weighted and unweighted UniFrac distances were analysed using Principal Coordinates Analysis and used to describe beta-diversity for foot health groups and farms. Robust response screening (JMP Pro 12, SAS Institute Inc., USA), was performed to evaluate differences in relative abundance of OTU assignments at genus level between samples from the HtHt and HtIn groups.

A subset of five samples from the HtHt and HtIn groups were used for shotgun metagenomic analysis to provide higher taxonomic resolution and detect differences in the functional profiles of the microbiome. Prepared libraries were sequenced using the Illumina® HiSeq 4000 platform and results analysed using LefSe to determine taxa most likely to explain differences between the two groups. HUMAnN2 was used to functionally annotate read data to show biological pathway abundance and completeness.

Results: Differences in sample richness and evenness were identified between foot-health groups. Alpha-diversity metrics overall suggested that HtHt samples had significantly greater microbial diversity than InIn or InHt samples, and a tendency to greater microbial diversity than HtIn samples.

Focussing on comparing HtHt to HtIn groups (where samples originated from morphologically healthy feet), PCoA showed clustering between HtHt and HtIn groups, indicating that overall the two groups share the same species. However, response screening showed at genus level that Macrococcus spp. and Brachybacterium spp. were more abundant in samples from the HtHt group, whilst Succiniclasticum spp., Porphyromonas spp., Acholeplasma spp., Fastidiosipila spp., Peptoclostridium spp. and Prevotella spp. were more abundant in samples from the HtIn group. Treponema spp. were present in the top 20 most prevalent genera for HtIn samples, but not for HtHt samples.

Shotgun metagenomic analysis showed some differences in taxa present in HtHt compared to HtIn samples. HtHt samples had increased relative abundance of many gram positive bacteria from the phylum Actinobacteria which would be expected to be part of the healthy foot skin flora. HtIn samples showed biologically relevant increases in wall-less bacteria from the Tenericutes phyla, which has previously been found to be more abundant in BDD lesions. Taxonomic assignment of reads was low and therefore the significance of these findings is uncertain; however, there may be some agreement with the 16S rRNA gene sequencing analysis in finding increased Acholeplasma spp. in HtIn samples and increased Brachybacterium sp. in HtHt samples.

Eight functional pathways were identified as significantly more abundant in the HtIn group. One indicated production of 4-deoxy-L-threo-hex-4enopyranuronate, which is a uronic acid resulting from the degradation of many polymers. These include plant polymers such as pectin and gellan, but also important components of connective tissue such as heparin, heparin sulfate, hyaluronan and chondroitin sulfate.

Conclusion: This is the first study to associate taxonomical differences in the healthy foot-skin microbiome with future development of BDD lesions. Our data suggest that loss of diversity may begin before the appearance of morphological lesions, and that Macrococcus spp and Brachybacterium spp. may be protective whilst Succiniclasticum spp., Porphyromonas spp., Acholeplasma spp., Fastidiosipila spp., Peptoclostridium spp. and Prevotella spp. may initiate dysbiosis that leads to lesion development. Upregulation of a functional pathway associated with polymer degradation may indicate the beginning of a pathogenic process in HtIn samples detectable prior to development of visible BDD lesions.