This study builds upon our previous analysis of transcriptomes at arrival that were derived from post-weaned beef cattle that ultimately developed BRD versus cattle that remained healthy (14). The present investigation was conducted with the intent to identify potential at-arrival biomarkers and pathways that indicate risk of BRD-associated mortality in post-weaned beef cattle. Our overarching goal with these studies is to identify gene expression profiles and biological pathways in blood samples at arrival that segregate with later BRD morbidity or mortality. By analyzing the transcriptomes of post-weaned cattle before they exhibit clinical signs of BRD, our approach will also improve understanding of the mechanistic basis of both susceptibility and resistance to BRD in this cohort. Previous research to determine early antemortem indications of BRD and risk of severity has yielded varied results (19) (20) (21) (22). At present, the diagnosis and classification of BRD is primarily assessed using clinical factors that have proven to be imprecise, effectively limiting BRD management (6) (10) (23). Nonetheless, cattle diagnosed with BRD in this investigation using these same clinical factors (including DART scoring, treatment records, and average daily weight gain) exhibited differences in gene product and molecular pathway expression at arrival that ultimately segregated with their BRD-associated mortality (14).
In DEAD cattle, the expression of genes related to type I interferon production/signaling and viral defense were increased. These viral defense genes included IFIT1/2/3, IRF4, HERC5/6, OAS2, MX2, and ISG15/20. Several investigations have similarly identified these genes as relevant in cattle with prolonged inflammation and ongoing viral infection (24) (25) (26) (27). These findings, when considered with the increased expression of viral defense genes at arrival in cattle that ultimately died of BRD, provide evidence that cattle in the DEAD cohort were combating a viral agent at arrival. Though cattle that died of BRD in this investigation did not show clinical evidence of BRD at arrival, viral BRD is often subclinical and may initially present as an upper airway disease. Subclinical viral respiratory infection at arrival would not only account for the observed viral defense pathways in the cattle that died of BRD, but would facilitate secondary infectious processes in the lung that contribute to the observed BRD mortality (25) (5) (28) (29). One challenge with our study is that differences in gene expression were characterized in peripheral blood. It has been suggested that the blood transcriptome represents an amalgamation of gene expression patterns and pathways in distinct physiological sites, such as airway epithelium, lymph nodes, and splenic tissue (30) (31). Necropsy demonstrated that disease was limited to the lung in cattle that died. Overt disease was not evident at arrival, but subclinical disease cannot be ruled out.
In this study, we identified increased expression of gene products that interact with toll-like receptor 4 (TLR4) and interleukin 6 (IL6) in both live and dead cohorts. While there was no difference in the expression of these specific genes between DEAD vs ALIVE cattle at arrival, we have previously described increased BGN, MARCO and POMC expression (known to be involved in TLR4-dependent pro-inflammatory pathways) in arrival blood transcriptomes from cattle that ultimately developed BRD when compared to cattle that remained clinically healthy (14). IL6 and several other type I interferon-associated genes have been reported to be differentially expressed within lymph node samples of virus-challenged cattle (25) (26). TLR4 possesses high avidity for lipopolysaccharide (LPS) and some viral structural proteins, and is capable of inducing type I interferons production and increased levels of IL6 (32) (33) (34) (35). Additionally, elevated levels of IL6 may reciprocatively induce type I interferon production, enhancing natural killer cell cytotoxic activity, M1 macrophage maturation, and interleukin 12 (IL12) production (36) (37) (38) (39). It is important to note that IL6 and TLR4 are not differentially expressed between DEAD and ALIVE groups but are predicted to be active based on associations with DEG products increased within each cohort. It is possible that TLR4 and IL6, relatively non-specific markers of inflammation, are initiated in both DEAD and ALIVE groups albeit through differing mechanisms. Several studies have demonstrated that TLR4 expression is increased in active respiratory disease and is responsible for proinflammatory cytokine production, in both viral and bacterial induced infections (40) (41) (42) (43). Furthermore, in ALIVE cattle, the increased expression of several proinflammatory genes was identified: CD300LG, COL1A1, CX3CR1, KIR2DL5A, LOC104968634 (NK2B), OGN, LOC782922 (PRXL2B), TARP. These gene products, largely involved in natural killer cell activation, leukocyte adhesion, prostaglandin synthesis, and initiation of the acquired immune system, possess known interactions or promotion of TLR4 and IL6 activity. In conjunction with TLR4 interactions, it is possible that the ALIVE cattle were actively combating extracellular antigens or etiological agents. The commonality between ALIVE and DEAD cattle is antigenic and immunogenic signaling without inflammatory mitigation. Notably, our research did not ascertain the order of TLR4 and IL6 association, therefore further research is necessary to define mechanistic characteristics and signaling order.
One limitation of this study is the lack of antemortem pathogen identification, particularly viral isolation at arrival. Modeling genes that were differentially expressed between live and dead cohorts identified increased antiviral pathways in the DEAD cohort. Accordingly, we utilized de novo alignment and BLAST toolkits to mine unmapped reads for viral sequences that would account for the gene expression changes and pathways identified in our study. Reads that fail to map to the host reference assembly have been previously used to identify pathogens within RNA-Seq datasets (44) (45) (46) (47) (48). The de novo assembly reads aligned predominantly to annotated ungulate sequences (Fig. 4, 5). This is an expected occurrence in which the unmapped reads representing gene products in the tested cattle were not identified with the Bos taurus ARS-UCD1.2 reference assembly. This is not an uncommon occurrence with reference assemblies from non-model organisms that reflects errors in the assembly’s structural annotation and has been otherwise reported in transcriptomic experiments using cattle (44) (48).
A notable finding in this investigation is the lack of alignments to pathogenic organisms associated with BRD. This finding was consistent in three instances 1) when assembled contigs of the de novo transcriptome were mapped to all sequences in the NCBI non-redundant nt database; 2) when unmapped reads were mapped against all known viral sequences; and 3) when unmapped reads were mapped against all known bovine viral pathogen sequences. One key limitation of this experiment is the poly-A tail dependent capture of reads for library preparation. It is possible that pathogen genes within each whole blood sample were never captured prior to sequencing. Therefore, while we were unable to identify pathogen genes in these instances, it does not rule out the possibility of these cattle harboring etiological agents at arrival. All biological replicates possessed reads that matched only to BVDV1 sequences from the NCBI nt database. However, these reads matched solely to ubiquitin C (UBC) mRNA within Bovidae assemblies and no alignment was detectible when realigned to the NC_001461.1 complete BVDV1 genome. This demonstrates that the unmapped reads failed to align to viruses related to BRD, but rather aligned to bovine genome sequences that have been incorporated into BVDV1. It has been shown that several BVDV1 sequences possess Bovidae genomic sequence contamination, specifically to UBC mRNA. This finding agrees with the alignment discovery reported by Usman and colleagues (44). Despite the absence of viral sequences, the identified DEGs and pathways provide evidence that the anti-viral mechanisms were activated at arrival in cattle within the DEAD cohort.