Investigation of bovine leukemia virus transmission pathways within dairy herds through analysis of genomic and epidemiological data

Bovine leukemia virus (BLV) is a highly prevalent retrovirus in US dairy cattle herds and has negatively affected dairy herds by reducing milk production, increasing culling rates, and contributing to economic losses. The objective of this study was to link pathogen genomic data with animal-level data to provide inference to BLV transmission pathways in dairy herds. Blood samples were collected from cattle of at least 8 mo of age during 5 different sampling periods in 2 different dairy herds. Following testing, we extracted, sequenced, and assembled the BLV proviral DNA using a prior validated and published protocol from positive samples. The assembled near-complete genomes were aligned and converted to phylogenetic trees, and single nucleotide variant (SNV) matrices were created. When compared with global BLV genomes published in GenBank, the BLV sequences from our 2 Minnesota study herds (herd A and F) showed additional clade diversity not shared with previously characterized global variants. The BLV genomes identified from the 2 study herds were organized in multiple clades, and herd A had more BLV genomic diversity than herd F. Of the 195 total BLV sequences from the herds, 51 were from cows at multiple time points, and from herds A and F, respectively 84.5% and 98.3%, of within-host pairs of cattle had ≤10 SNV differences between the pairs, and 82% and 98.4% of within-host pairs were in identical clades. These results indicated low variability of the BLV proviral DNA genome through time within individual infected cattle. Within the dam-daughter (DD) cattle pairs from herds A (n = 24) and F (n = 10), 62.5% and 60% had ≤10 SNV and 50% and 60% clustered in identical clades, respectively. In comparison, other between-cattle pairs in herds A and F showed 35.7% (n = 4,981) and 48.4% (n = 4,300) had ≤10 SNV and 11.2% and 50.6% in identical clades, respectively. These results are consistent with DD transmission in a subset of BLV-infected cattle. Pairs of cattle born at similar times in the herd had a similar SNV distribution to other cattle pairs within their herds, indicating minimal transmission early in life among similar age cohorts. Genomes from cattle with a high BLV proviral load (relative concentration of BLV proviral DNA to bovine DNA in copies/µL >2,000) were distributed across most clades, indicating the potential for local transmission for BLV from multiple clades rather than a single unusually transmissible lineage dominating the herd. In summary, this study shows the potential benefit of using near-complete genomes to identify high-resolution within-herd BLV differences to infer transmission.