Re-evaluation of the etiology and pathogenesis of liver abscesses in feedlot cattle: potential involvement of Fusobacterium varium and bacteriophages as an intervention tool

Abstract

Liver abscesses in cattle are the number one cause of liver condemnations at slaughter and have a significant economic impact on the beef cattle industry. Fusobacterium necrophorum, a common ruminal bacterium, has been recognized as the primary causative agent of liver abscess formation in feedlot cattle. Liver abscess formation in beef cattle has been historically attributed to the “ruminitis- liver abscess complex”, which hypothesizes that high grain feeding leads to ruminal acidosis and subsequent ruminal epithelial damage, which allows the translocation of F. necrophorum, and other liver abscess-causing pathogens, into the liver via portal vein circulation. Recent studies utilizing genomic sequencing-based technologies have revealed that liver abscesses are highly diverse, polymicrobial infections, and have also raised the question of whether the hindgut, specifically the colon, can be a source of liver abscess-causing pathogens. In light of these recent findings, my main research objective was to reevaluate our understanding of the etiology and pathogenesis of liver abscesses in feedlot cattle, with a focus on the potential involvement of Fusobacterium varium in liver abscess development. Additionally, we investigated the potential use of bacteriophages as an antimicrobial alternative for liver abscess prevention. Fusobacterium varium has recently been reported as the most predominant Fusobacterium species in the bovine rumen, as determined by 16S microbiome analysis. Therefore, our first experiment was to develop and validate a real-time quantitative PCR assay for the detection and quantification of F. necrophorum subsp. necrophorum, F. necrophorum subsp. funduliforme, and F. varium in ruminal fluid samples collected from feedlot cattle. The qPCR primers were designed to target the hgdA gene for species-level differentiation and the lktA gene for differentiation at the subspecies level. The qPCR assay was conducted on 345 ruminal fluid samples collected from feedlot cattle immediately after slaughter, of which 181 were from cattle with apparently healthy, non-abscessed livers, and 164 were from cattle with liver abscesses. The ruminal fluid samples were strained through cheesecloth to remove large particles and an aliquot of strained ruminal fluid was collected for qPCR analysis. Additionally, strained ruminal fluid samples were enriched using a selective media containing lactate or lysine as the energy source and supplemented with josamycin, vancomycin, and norfloxacin antibiotics. The qPCR assay quantified F. necrophorum subsp. necrophorum in only 22% and 29% of samples from cattle with healthy and abscessed livers, respectively, with the samples from abscessed livers significantly more likely to contain subsp. necrophorum (P = 0.004). Interestingly, F. necrophorum subsp. funduliforme and F. varium were found to be prevalent in nearly all samples (98 to 100%), and no significant difference in prevalence was found between cattle with healthy livers and cattle with liver abscesses. Mean concentrations of F. varium and both subsp. of F. necrophorum were not significantly different between animal status, with mean concentrations of 10³ to 10⁴ CFU/ml. Our second study was intended as a follow up to further investigate the potential for F. varium to be involved in liver abscesses since it was so highly prevalent in ruminal fluid samples. We obtained matched liver abscess, ruminal and colonic epithelial tissue, and ruminal and colonic content samples from 96 steers at the time of slaughter. Samples were homogenized and cultured onto blood agar and selective enrichment media for the presence of F. necrophorum and F. varium, and the previously described qPCR assay was used for detection and quantification of F. necrophorum and F. varium in the samples before and after enrichment. Three liver abscess samples (3.1%) yielded F. varium by culture-based methods, the first report of isolation of F. varium from bovine liver abscesses. Ruminal and colonic epithelial tissue samples had an F. varium isolation rate of 59.4% and 31.3%, respectively. The qPCR assay detected F. varium in 10 (10.4%) of samples, with only one of the samples quantifiable prior to enrichment. The total qPCR-based prevalence in ruminal and colonic epithelial tissues was 77.1% and 44.8%, respectively, while ruminal and colonic contents had a total qPCR-based prevalence of 86.5% and 70.1%. These findings showed support for the theory that the colon may be involved in liver abscess etiology. As a follow-up to this initial study, we investigated the effect of various feeding and management strategies on the prevalence of F. varium in liver abscesses or liver tissues and their corresponding ruminal and colonic epithelial tissues. Matched sets of samples were collected from two feedlot studies and one experimental study and subjected to culture-based and qPCR based methods for F. varium detection and quantification. In the first feedlot study of 190 cattle intended to evaluate the effect of tylosin on F. varium prevalence, no difference was observed between the tylosin treatment group compared to the non-tylosin fed group in ruminal epithelial tissues, liver abscesses, or healthy liver tissues. In the experimental study, cattle were assigned to either the control treatment (n = 10) or induced ruminal acidosis treatment (n =10), since ruminal acidosis is considered a necessary precursor to liver abscess formation. None of these cattle developed liver abscesses, and no difference in prevalence was observed in liver tissues, ruminal and colonic epithelial tissues, or ruminal and colonic contents between cattle with ruminal acidosis compared to control cattle. The second feedlot study compared samples from 159 cattle subjected to a 2x2 factorial design for the evaluation of two dietary starch levels (CON or HOT) and two management strategies (REG or ERR). Cattle fed a “hot” diet, intended to induce acidosis, and subjected to an erratic feeding schedule had a significantly higher F. varium prevalence in colonic epithelial tissues than cattle fed a control diet and subjected to an erratic feeding schedule (P < 0.01). No differences in ruminal or liver abscess F. varium were observed, nor was there any difference in F. varium concentrations. We also performed whole genome sequencing and analyses on a subset of 14 F. varium strains isolated from liver abscesses (n = 5), ruminal epithelial tissue (n = 5), and colonic epithelial tissue (n = 4). Using average nucleotide identity (ANI) comparison, very high levels of genetic similarity were found between strains recovered from the same animal, though no accepted cut-off value for establishing clonality has been set. Previous studies have reported resistance to tylosin in F. varium strains recovered from the bovine rumen, however, we did not find any macrolide resistance genes in our strains. These findings of high genetic similarity further emphasized that the hindgut may be a source of bacterial translocation into the liver. A whole genome sequencing study was then conducted on F. necrophorum subsp. necrophorum strains recovered from the two previously described feedlot studies, for the purpose of genetically comparing strains recovered from liver abscesses to strains recovered from ruminal and colonic tissues from the same animal, and to compare genomic differences between strains based upon their tissue of origin. A total of 62 F. necrophorum subsp. necrophorum strains were sequenced and used for bioinformatic analyses. Using core-genome SNP analyses along with ANI analysis, we determined clonal relationships between matched strains recovered from the colon and liver abscess from 2 animals, the ruminal and liver abscess strains from 3 animals, and the matched ruminal, colonic, and liver abscess strains from one animal. When the strains were subjected to AMR screening using StarAMR, it was found that two liver abscess strains carried the ermB gene, which encodes resistance to macrolide antibiotics such as tylosin. In feed tylosin administration is the most common liver abscess prevention method, and this was the second report of macrolide resistance in F. necrophorum isolated from bovine liver abscesses. Additionally, relatively high levels of aminoglycoside (17.7%) and tetracycline (24.2%) resistance was observed. Our final study aimed to investigate the use of bacteriophages as an antimicrobial alternative for liver abscess prevention. Characterization, including host ranging, titer determination, morphological examination, pH stability, and whole genomic sequences was performed on 5 novel bacteriophages previously isolated from untreated city sewage. All 5 phages were found to be highly specific to F. necrophorum subsp. necrophorum, and were generally stable in pH values ranging from 4-9 for up to 24 hours. Electron microscopy revealed four of the five phages were tailed, while one phage was unable to be visualized. A proof of concept study was conducted to evaluate the efficacy of prophylactic administration of a cocktail of all 5 bacteriophages in preventing liver abscesses caused by F. necrophorum in a mouse model. Significant reduction in both morbidity and mortality was observed (P < 0.05) in mice treated with the bacteriophage cocktail compared to the negative control group. These findings warrant further investigation into the potential application of bacteriophages as an alternative to tylosin for the prevention of liver abscesses in feedlot cattle.