Metabolome of purulent materials of liver abscesses from crossbred cattle and Holstein steers fed finishing diets with or without in-feed tylosin

Abstract

Liver abscesses in feedlot cattle are a polymicrobial infection with Fusobacterium necrophorum and Trueperella pyogenes as the primary and secondary etiologic agents, respectively. Cattle with liver abscesses do not exhibit clinical signs and the abscesses are detected only at slaughter. The objective was to conduct metabolomics analysis of purulent materials of liver abscesses to identify biochemicals. Liver abscesses from crossbred cattle (n = 24) and Holstein steers (n = 24), each fed high-grain finishing diet with tylosin (n = 12) or no tylosin (n = 12), were included in the study. Abscess purulent materials were analyzed by ultrahigh-performance liquid chromatography-tandem mass spectroscopy. A total of 759 biochemicals were identified and were broadly categorized into carbohydrates, energy metabolism pathways intermediates, peptides, amino acids and their metabolites, lipids and their metabolites, nucleotides, vitamins and cofactors, xenobiotics, and partially characterized molecules. The top 50 biochemicals identified included amino acids, lipids, nucleotides, xenobiotics, peptides, and carbohydrates and their metabolites. Among the 15 amino acid metabolites in the top 50 biochemicals, four were tryptophan metabolites, indoleacrylate, indolepropionate, tryptamine, and anthranilate. The 3-phenylpropionate, a product of phenylalanine metabolism, was the predominant metabolite in purulent materials. Between the four treatment groups, a two-way ANOVA analysis identified biochemicals that exhibited significant main effects for cattle type and in-feed tylosin use and their interactions. A total of 59 and 85 biochemicals were different (P < 0.05) between the cattle type (crossbred vs. Holstein steers) and in-feed tylosin use (tylosin vs. no tylosin), respectively. Succinate, an intermediate of lactate fermentation by some bacterial species, was one of the top 30 biochemicals that differentiated the four treatment groups. A number of lysophospholipids, indicative of bacterial and host cell membrane lyses, were identified in the purulent materials. In conclusion, to our knowledge this is the first report on the metabolome of liver abscess purulent materials and several biochemicals identified were related to metabolic activities of the bacterial community, particularly F. necrophorum and T. pyogenes. Biochemicals unique to liver abscesses that appear in the blood may serve as biomarkers and be of diagnostic value to detect liver abscesses of cattle before slaughter.

A metabolomics analysis of purulent materials from 48 liver abscesses of cattle identified a total of 759 biochemicals. Biochemicals unique to liver abscesses that appear in the blood may have the potential to be used as biomarkers to identify cattle with liver abscesses before slaughter.

Introduction

Liver abscesses occur in feedlot cattle because of the feeding and management programs that include energy-dense, high-grain, and low-roughage diet (Nagaraja and Chengappa, 1998). It is generally accepted that the ruminal wall, damaged by exposure to chronic acidity from grain fermentation, becomes susceptible to invasion and colonization by opportunistic ruminal bacteria, leading to rumenitis, and subsequently the organisms enter portal circulation to reach the liver to cause abscesses (Jensen et al., 1954; Nagaraja and Chengappa, 1998). Liver abscesses are a polymicrobial infection and Fusobacterium necrophorum, a ruminal bacterium, is the primary etiologic agent. In addition, Trueperella pyogenes and Salmonella enterica are frequently isolated. A number of other species of bacteria, such as Escherichia coli, Klebsiella pneumoniae, Lactococcus garviae, and Enterococcus faecalis, have also been frequently isolated (Scanlan and Hathcock, 1983; Nagaraja and Chengappa, 1998; Amachawadi et al., 2017). The incidence of liver abscesses is highly variable, which generally ranges from 10% to 30%, and is influenced by a number of factors (Amachawadi and Nagaraja, 2016). One factor contributing to the variability is the cattle type, with incidence consistently higher in calf-fed Holstein steers compared to crossbred beef cattle (Reinhardt and Hubbert, 2015; Amachawadi and Nagaraja, 2016). Tylosin, a macrolide included in the feed, is the most commonly used antibiotic for the control of liver abscesses (Nagaraja and Chengappa, 1998).

Cattle with abscessed livers do not show any clinical signs and are detected only at necropsy or at the time of slaughter. Liver function tests, including blood cell counts, serum metabolites, and serological test targeting for F. necrophorum-specific antibodies have not proven to be of diagnostic value (Tan et al., 1994a; Macdonald et al., 2017; Herrick et al., 2020). Studies have attempted ultrasound scanning of the liver for the detection of abscesses with limited success (Lechtenberg and Nagaraja, 1991; Liberg and Jonsson, 1993). Ultrasonography may not identify abscesses if they are located on the visceral side, in deeper regions of the liver tissue, or in a lobe covered by the lung tissue (Dore et al., 2007; Braun, 2009).

Metabolomics is the study of metabolites in biofluids, cells, tissues, or organs, particularly of small molecule substrates, intermediates, and products of metabolism. Identification of metabolites, which are essentially “signaling” molecules, helps to understand the functional aspect of the biochemical process. In recent years, metabolomics has been used in the diagnosis and monitoring of a variety of disease progressions, such as infectious and noninfectious diseases and cancer in animals and humans (Wishart, 2019; Bhosle et al., 2022; Fraga-Corral., 2022). Since the first report on a metabolomics analysis of the ruminal fluid of cattle (Saleem et al., 2013), a number of studies on bovine biofluids (blood, urine, uterine fluid, milk, and semen) or tissues, including liver metabolome to relate to production traits, and to diagnose metabolic and infectious diseases, such as acidosis, ketosis, metritis, laminitis, and respiratory diseases, have been reported (Consolo et al., 2019; Foroutan et al., 2020; Polizel et al., 2022).

To our knowledge, there has been no study on untargeted metabolomics analysis of purulent material from liver abscesses in feedlot cattle. In addition to providing additional insights into the infection and the pathogenic process of the liver, the information obtained from this study may identify unique biomarkers that could aid in the diagnosis of liver abscesses in live cattle. Our objective was to conduct metabolomic analysis of purulent materials of liver abscesses of crossbred cattle and Holstein steers fed diets with or without tylosin to identify biochemicals.

Materials and Methods

Study design and treatment groups

Liver abscesses included in the study were collected from feedlot cattle at USDA-inspected slaughter houses, located in Arizona, California, Colorado, and Kansas, where humane slaughter practices were followed, according to the USDA guidelines. The study design, which included cattle type (crossbred and Holstein steers) and in-feed tylosin use (tylosin or no tylosin), has been described in our previous publications (Amachawadi et al., 2017, 2021). Briefly, the study was a 2 × 2 factorial design with liver abscesses were collected from crossbred beef cattle and Holstein steers, each fed high-grain finishing diet supplemented with or without tylosin (8 to 10 g per ton of feed; Tylan, Elanco Animal Health, Greenfield, IN). Abscess purulent material collected aseptically from intact abscesses were stored at −80 °C. Twelve liver abscess samples for each treatment group, which included three samples randomly selected from each of four feedlots, were shipped on dry ice to Metabolon Inc. (Morrisville, NC) for metabolomics analysis on Metabolon’s Global Metabolomics Platform.

Sample preparation

Samples were prepared using the automated MicroLab STAR system (Hamilton Co., Reno, NV). Several recovery standards were added prior to the extraction process for quality control. To remove proteins, dissociate small molecules bound to protein or trapped in the precipitated protein matrix, and to recover chemically diverse metabolites, proteins were precipitated with methanol under vigorous shaking for 2 min (GenoGrinder 2000, Glen Mills, Clifton, NJ), followed by centrifugation. The resulting extract was divided into five fractions: two fractions for analysis by two separate reverse-phase ultrahigh-performance liquid chromatography-tandem mass spectroscopy (RP)/UPLC-MS/MS methods with positive ion mode electrospray ionization (ESI), third fraction for the analysis by RP/UPLC-MS/MS with negative ion mode ESI, fourth fraction for the analysis by hydrophilic interaction liquid chromatography (HILIC) UPLC-MS/MS with negative ion mode ESI, and the fifth fraction was reserved for backup. Samples were placed briefly on a TurboVap (Zymark Corp., Hopkinton, MA) to remove the organic solvent. The sample extracts were stored overnight under nitrogen gas before preparation for analysis.

Ultrahigh-performance liquid chromatography-tandem mass spectroscopy

All methods utilized a Waters ACQUITY UPLC and a Thermo Scientific Q-Exactive high resolution/accurate mass spectrometer interfaced with a heated electrospray ionization (HESI-II) source and Orbitrap mass analyzer operated at 35,000 mass resolution. The sample extract was dried, then reconstituted in solvents compatible to each of the four methods. Each reconstitution solvent contained a series of standards at fixed concentrations to ensure injection and chromatographic consistency. One aliquot was analyzed using acidic positive ion conditions, chromatographically optimized for more hydrophilic compounds. In this method, the extract was gradient eluted from a C18 column (Waters UPLC BEH C18-2.1 × 100 mm, 1.7 µm) using water and methanol, containing 0.05% perfluoropentanoic acid and 0.1% formic acid (FA). Another aliquot was analyzed using acidic positive ion conditions, but it was chromatographically optimized for more hydrophobic compounds. In this method, the extract was gradient eluted from the same C18 column using methanol, acetonitrile, water, 0.05% pentafluoropropionic anhydride and 0.01% FA and was operated at an overall higher organic content. Another aliquot was analyzed using basic negative ion optimized conditions using a separate dedicated C18 column. The basic extracts were gradient eluted from the column using methanol and water, but with 6.5 mM ammonium bicarbonate at pH 8. The fourth aliquot was analyzed via negative ionization following elution from a HILIC column (Waters UPLC BEH Amide 2.1 × 150 mm, 1.7 µm) using a gradient consisting of water and acetonitrile with 10 mM ammonium formate, pH 10.8. The MS analysis alternated between MS and data-dependent MSⁿ scans using dynamic exclusion. The scan range varied between methods but covered 70 to 1,000 m/z. Raw data files were archived and extracted as described below.

Data extraction and compound identification

Metabolon’s hardware and software were used to extract raw data, identify peaks, and assure quality control. The systems are built on a web-service platform utilizing Microsoft’s NET technologies, which run on high-performance application servers and fiber-channel storage arrays in clusters to provide active failover and load-balancing. Compounds were identified by comparison to library entries of purified standards or recurrent unknown entities. Metabolon maintains a library based on authenticated standards that contain the retention time/retention index (RI), mass-to-charge ratio (m/z), and chromatographic data (including MS/MS spectral data) on all molecules present in the library. Furthermore, biochemical identifications are based on three criteria: RI within a narrow RI window of the proposed identification, accurate mass match to the library ± 10 ppm, and the MS/MS forward and reverse scores between the experimental data and authentic standards. The MS/MS scores are based on a comparison of the ions present in the experimental spectrum to the ions present in the library spectrum. While there may be similarities between these molecules based on one of these factors, the use of all three data points can be utilized to distinguish and differentiate biochemicals. More than 3,300 commercially available purified standard compounds have been acquired and registered into Laboratory Information Management System for analysis on all platforms for determination of their analytical characteristics. Additional mass spectral entries have been created for structurally unnamed biochemicals, which have been identified by virtue of their recurrent nature (both chromatographic and mass spectral).

Metabolite quantification and data normalization

Peaks were quantified using area-under-the-curve. For analysis spanning multiple days, a data normalization step was performed to correct variation resulting from instrument interday tuning differences. Essentially, each compound was corrected in run-day blocks by registering the medians to equal one (1.00) and normalizing each data point proportionately (termed the “block correction”). For studies that did not require more than 1 d of analysis, no normalization was needed, other than for purposes of data visualization.

Statistical analysis

The metabolite composition of the purulent material from 48 liver abscesses consisting of a total of 759 compounds, quantified using area-under-the-curve, were log-transformed before analysis. The metabolites data were analyzed using a two-way (cattle type and in-feed tylosin use and their interaction) and a four-way (four treatment groups) ANOVA contrasts and a principal component analysis on natural log-transformed data in ArrayStudio (OmicSoft ArraySuite 10.0, Qiagen, N.V.). The short-chain fatty acids (SCFA) data were analyzed using STATA MP (v. 16.1; College Station, TX). A two-way ANOVA was used to compare mean differences between cattle type, tylosin supplementation, and their interaction. A random forest analysis (RFA) was used to identify the top 30 biochemicals that separate the four treatment groups (Brieman, 2001). A P-value of ≤ 0.05 was considered significant and a P-value between 0.05 and 0.1 was considered as a trend toward significance.

Results and Discussion

The liver is a pivotal organ of metabolism that plays a central role in energy metabolism, protein synthesis, and immunity. Physiologically, the liver is a demanding organ and its physiological state directly correlates to variation in metabolic output of animals (Johnson et al., 1990, Muller, 1998). Abscess formation in the hepatic parenchyma occurs when an acute inflammatory response fails to eliminate the pyogenic organism. Initially, a septic abscess consists of a collection of neutrophils that may not be encapsulated or may be encircled by a thin vascularized connective tissue wall. If the septic abscess is unresolved, the thin connective tissue wall matures into a fibrous capsule, which is thick and largely impermeable and the purulent material is walled off from the normal tissue (Ackerman, 2007). The enzymes and inflammatory mediators from tissue macrophages (Kupffer cells) and neutrophils liquefy the hepatic parenchymal tissue and neutrophils to form the purulent material. The nature of the purulent material in a liver abscess varies from liquid to caseous to dry and inspissated material. The purulent material of liver abscesses contains fluid, dead microbial and host cells, bacterial degradations of cytolytic products of bacterial cells, hepatic parenchymal cells, Kupffer cells, and blood cells.

In this study, purulent material of 48 liver abscesses of crossbred and Holstein steers fed high-grain diet with or without in-feed tylosin were subjected to untargeted metabolomics analysis. The metabolomics provide a snapshot of broadest range of biochemicals contained in the purulent material (Wishart, 2019). A total of 759 biochemicals were identified (Supplementary Table 1) and were broadly categorized into major biochemical classes or metabolic pathways, which included carbohydrates (n = 17), peptides (n = 43), amino acids and their metabolites (n = 193), energy pathway ([Krebs cycle and oxidative phosphorylation] intermediates; n = 11), lipids and their metabolites (n = 331), nucleotides (n = 54), vitamins and cofactors (n = 26), xenobiotics (n = 52), and partially characterized molecules (n = 1; Table 1). The metabolites in the lipids and peptides and amino acids classes were the most predominant biochemical classes contained in the purulent material. The types of lipids detected included intermediates and products of lipolysis (diacylglycerides, monoacylglycerides, glycerol, long-chain saturated or unsaturated fatty acids, medium-chain and short-chain fatty acids and their derivatives), variety of phospholipids and their intermediates (lysophopholipids), ceramides, plasmologens, and sphingomyelins. The nitrogenous compounds included peptides, mainly dipeptides and a number of amino acids and their metabolites, including intermediates of urea cycle. The amino acids and their metabolites included primarily aromatic amino acids (tryptophan, phenylalanine, and tyrosine), lysine, and branched-chain amino acids (valine, leucine, and isoleucine). The carbohydrate class of molecules included intermediates of glycolysis, pentose pathway, gluconeogenesis, metabolites of fructose, mannose, galactose, pentoses, and amino sugars. The energy metabolism class included metabolites of Krebs’s cycle and oxidative phosphorylation. The nucleotide class included metabolites of purines and pyrimidines. The list of biochemicals also included vitamins and cofactors and their metabolites, particularly nicotinamides. Additionally, a number of biochemicals categorized as xenobiotics, which included benzoate metabolites, plant components, bacterial and fungal components, drugs (e.g., salicylate), and other chemicals were also detected.

Table 1.

Metabolites of major biochemical classes and metabolic pathways that were detected in the purulent material of liver abscesses of crossbred cattle and Holstein steers fed high-grain finishing diets with or without tylosin

Major biochemical class and metabolic pathways	Subcategory of biochemical class and metabolic pathways	No. of biochemicals
Amino acids	Leucine, isoleucine, and valine metabolism	35
	Methionine, cysteine, and taurine metabolism	23
	Histidine metabolism	21
	Urea cycle, arginine, and proline metabolism	19
	Tryptophan metabolism	17
	Lysine metabolism	15
	Glutamate metabolism	14
	Polyamine metabolism	13
	Tyrosine metabolism	11
	Glycine, serine, and threonine metabolism	10
	Phenylalanine metabolism	9
	Alanine and aspartate metabolism	6
	Creatine metabolism	3
	Guanidino and acetamido metabolism	3
	Glutathione metabolism	3
Carbohydrate	Pentose metabolism	9
	Glycolysis, gluconeogenesis, pyruvate metabolism	6
	Amino sugar metabolism	6
	Fructose, mannose, and galactose metabolism	4
	Pentose phosphate pathway	2
Cofactors and vitamins	Nicotinate and nicotinamide metabolism	9
	Ascorbate and aldarate metabolism	5
	Riboflavin metabolism	3
	Vitamin B6 metabolism	3
	Tocopherol metabolism	2
	Pantothenate and CoA metabolism	1
	Hemoglobin and porphyrin metabolism	1
	Thiamin metabolism	1
	Vitamin A metabolism	1
Energy	Tricarboxylic acid cycle	10
	Oxidative phosphorylation	1
Lipids	Monohydroxy fatty acid	27
	Sphingomyelins	27
	Lysophospholipid	25
	Diacylglycerol	20
	Dicarboxylate fatty acids	19
	Endocannabinoid	18
	Long-chain polyunsaturated fatty acids	15
	Phosphatidylcholine	13
	Secondary bile acid metabolism	13
	Plasmalogen	11
	Ceramides	11
	Medium-chain fatty acids	10
	Phosphoatylethanolamine	10
	Fatty acid metabolism (acyl glycine)	8
	Phospholipid metabolism	8
	Hexosylceramides	8
	Fatty acid metabolism (acyl carnitine, long-chain saturated)	7
	Sterol	7
	Long-chain saturated fatty acids	6
	Dihydroxy fatty acid	6
	Sphingolipid synthesis	6
	Primary bile acid metabolism	6
	Fatty acid metabolism (acyl choline)	5
	Lactosylceramides	5
	Dihydrosphingomyelins	5
	Long-chain monosaturated fatty acids	4
	Fatty acid metabolism (acyl carnitine, monounsaturated)	4
	Eicosanoid	4
	Lysoplasmalogen	4
	Monoacylglycerol	4
	Branched-chain fatty acids	3
	Glycerolipid metabolism	3
	Dihydroceramides	3
	Sphingosines	3
	Fatty acid amides	2
	Amino fatty acids	2
	Fatty acid metabolism (acyl carnitine, hydroxy)	2
	Carnitine metabolism	2
	Mevalonate metabolism	2
	Fatty acid synthesis	1
	Short-chain fatty acids	1
	Fatty acid metabolism (acyl carnitine, short-chain)	1
	Fatty acid metabolism (acyl carnitine, polyunsaturated)	1
	Ketone bodies	1
	Trihydroxy fatty acid	1
	Inositol metabolism	1
	Phosphoatylserine	1
	Phosphoatylglycerol	1
	Phosphoatylinositol	1
Nucleotides	Pyrimidine metabolism, uracil containing	12
	Purine metabolism, xanthine/inosine containing	9
	Pyrimidine metabolism, cytidine containing	6
	Purine metabolism, guanine containing	5
	Purine metabolism, adenine containing	4
	Pyrimidine metabolism, orotate containing	4
	Pyrimidine metabolism, thymine containing	4
	Purine and pyrimidine metabolism	1
Partially characterized molecules	Partially characterized molecules	1
Peptides	Dipeptide	20
	Gamma-glutamyl amino acid	15
	Acetylated peptides	5
	Dipeptide derivative	2
	Modified peptides	1
Xenobiotics	Food component/plant	24
	Benzoate metabolism	12
	Chemicals	10
	Bacteria/fungal	3
	Drugs—topical agents	2
	Drug—other	1

The metabolites were quantified based on the area-under-the-curve and the top 100 biochemicals based on raw identity (Supplementary Table 2) were identified. The top 50 biochemicals (Figure 1) included the following biochemical classes: amino acids (n = 15), lipids (n = 11), nucleotides (n = 11), xenobiotics (n = 7), peptides (n = 3), and carbohydrates (n = 3). The 3-phenylpropionate, a product of Strickland reaction of phenylalanine, was the most dominant metabolite in the purulent material of liver abscesses. Among the 15 amino acid metabolites, four belonged to tryptophan metabolism (indoleacrylate, indolepropionate, tryptamine, anthranilate). Although there is no report on metabolomics analysis of liver abscesses in cattle, bovine liver metabolome, based on liver tissue, plasma, or serum metabolomics analyses have been published in relation to transition in dairy cows (Scharen et al., 2021a, 2021b), to feed efficiency (Fonseca et al., 2019) and prenatal or preweaning nutrition strategies in dairy cows or bulls (Leal et al., 2021; Polizel et al., 2022).

Figure 1.

Top 50 metabolites in purulent materials of liver abscesses of crossbred cattle and Holstein steers fed high-grain diet with or without tylosin.

A number of culture-based, and, recently, a few culture-independent studies have concluded that liver abscesses are a polymicrobial infection (Scanlan and Hathcock, 1983; Nagaraja and Chengappa, 1998; Amachawadi and Nagaraja, 2016; Weinroth et al., 2017; Amachawadi et al., 2021; Pinnell et al., 2022). There is a general agreement that F. necrophorum, a ruminal bacterium, is the primary etiologic agent and T. pyogenes, a likely inhabitant of the ruminal wall, is the secondary etiologic agent. Besides liver abscesses, T. pyogenes is also frequently associated with F. necrophorum in foot rot of cattle (Nagaraja et al., 2005) and metritis in dairy cows (Bicalho et al., 2012). The frequent association between the two organisms is attributed to a nutritional interaction and pathogenic synergy (Tadepalli et al., 2009). Trueperella pyogenes is a saccharolytic and lactic acid-producing organism and the possible substrate in the liver is likely to be glycogen (Yassin et al., 2011). Lactic acid is a major energy source for F. necrohorum, which does not or ferments sugars weakly. In the rumen, the biochemical niche for F. necrohorum is to ferment lactic acid to SCFA, mainly butyric, and degrade proteins and peptides and deaminate amino acids, particularly lysine and tryptophan (Tan et al., 1994b; Amachawadi et al., 2016).

In 2015, S. enterica, particularly a novel serotype, called Lubbock, was identified in liver abscesses of calf-fed Holstein steers (Amachawadi and Nagaraja, 2015). In a subsequent study, S. enterica has been detected not only in Holstein steers, but also in crossbred cattle (Amachawadi et al., 2017). The contribution of S. enterica to the abscess formation in cattle has not been determined. However, there are a few reports of S. enterica serotypes as the primary causative agent of liver abscesses in humans (Chaudhary et al., 2003; Qu et al., 2013). Besides F. necrophorum, T. pyogenes, and S. enterica, other bacterial species more frequently isolated than others include E. coli, K. pneumoniae, E. faecalis, and L. garviae (Amachawadi et al., 2017). In a few studies that have used 16S rRNA amplicon sequence-based analyses have indicated that phylum Proteobacteria (Gram-negative facultative or aerobic bacteria) and Bacteroidetes (Gram-negative anaerobes) are the second and third most dominant phyla in liver abscesses of cattle (Weinroth et al., 2017; Amachawadi et al., 2021; Pinnell et al., 2022). However, only rarely has Fusobacterium has not been reported in liver abscesses of cattle (Herrick et al., 2022).

In humans, liver abscesses are not uncommon, but mostly sporadic, and are caused by a multitude of etiologic acids, including protozoa (Entamoeba). In a metabolomic profiling of purulent material from ameobic (caused by Entamoeba histolytica) and pyogenic liver abscesses in humans, acetate, formate, and succinate were identified as markers for amebic liver abscesses (Bharti et al., 2012). In a study comparing drainage-resistant (thick purulent material) to ­drainage-sensitive (thin purulent material) liver abscesses caused by K. pneumoniae, five metabolites in the serum, glucose, lactate, 3-hydroxybutyrate, glutamine, and alanine, were able discriminate between the two types (Chang et al., 2018).

Metabolites in relation to cattle type and in-feed tylosin use

Between the four treatment groups, a two-way ANOVA analysis identified biochemicals, quantified by area-under-the-curve, which exhibited significant main effects for cattle type and in-feed tylosin use and their interactions. A total of 59 and 85 biochemicals were different (P < 0.05) between the cattle type (crossbred vs. Holstein steers) and in-feed tylosin use (tylosin vs. no tylosin), respectively. A four-way ANOVA contrasts identified the total number of biochemicals that exhibited significant differences (P ≤ 0.05) or tended to be different (P > 0.05 to ≤0.1) between the four treatment groups and the number of biochemicals that increased or decreased between the treatment groups (Table 2). For example, in crossbred cattle, 28 and 65 biochemicals were higher or lower, at P < 0.05 level, between the tylosin-fed and no-tylosin-fed groups, respectively. In our previous study that compared the bacterial community composition by culture method, liver abscesses from Holstein steers contained more diverse bacterial species than liver abscesses from crossbred cattle and one of the major differences was higher prevalence of T. pyogenes (a lactic acid producer) in crossbred cattle compared to Holstein steers (Amachawadi et al., 2017). The principal component analysis revealed no distinct grouping between the four treatment groups (Figure 2). The RFA identified 30 biochemicals, ranked by importance, that provide the best separation between the four treatment groups (Figure 3A). The predictive accuracy from analysis of all 48 samples was 48% (with 25% to be expected from random chance; Figure 3B). The 30 top-ranking biochemicals suggest differences in metabolites associated with amino acids (n = 14), lipids (n = 5), carbohydrates (n = 3), xenobiotics (n = 3), Krebs cycle (n = 2), cofactors and vitamins (n = 2), and nucleic acid (n = 1). A compound, 2,6-dihydroxybenzoic acid, categorized as a xenobiotic compound, was most different between the four treatment groups. The amino acid metabolites that were different between the four treatment groups included metabolites of aromatic amino acids, phenylalanine (N-acetylyphenylalanine and N-succinyl-phenylalanine) and tryptophan (kyneurenate and N-formylkynurenate). Creatine, a guanidino compound that plays a major role in the storage and transport of cellular energy, was the dominant nitrogenous metabolite different between the four treatment groups.

Table 2.

Statistical summary of metabolites in the purulent material in liver abscesses from crossbred cattle and Holstein steers fed high-grain diets with or without in-feed tylosin

ANOVA contrasts	Crossbred cattle Tylosin vs. no tylosin	Holstein steers Tylosin vs. no tylosin	No tylosin Crossbred vs. Holstein steers	Tylosin Crossbred vs. Holstein steers
Total biochemicals, P ≤ 0.05	93	57	111	19
Biochemicals	28; 65	32; 25	94; 17	12; 7
Total biochemicals, P ˃ 0.05 ≤ 0.1	66	40	49	24
Biochemicals	24; 42	30; 10	38; 11	11; 13

Arrows indicate biochemicals that were higher or lower in the treatment groups.

Figure 2.

Principal component analysis plot of the metabolites detected in purulent materials of liver abscesses from crossbred cattle or Holstein steers fed high-grain diets with or without tylosin.

Figure 3.

Random forest analysis indicating top 30 biochemicals (A) in purulent materials of liver abscesses separating crossbred cattle (CB) or Holstein steers (HS) fed high-grain diets with (Tyl) or without tylosin (no Tyl). The predictive accuracy from analysis of all samples was 48% with 25% to be expected from random chance (B).

A few of the metabolites of aromatic amino acids (tryptophan, phenylalanine, tyrosine) and branched-chain amino acids (valine, leucine, and isoleucine) were different between the treatment groups. Tryptophan metabolites, indoleacetate, N-acetyltryptophan, C-glycosyltrytophan, N-formylkynurenine, kynurenate, and anthranilate were significantly higher in crossbred cattle with no tylosin compared to Holstein steers with no tylosin and significantly decreased crossbred cattle with tylosin. Among the lipids, medium-chain fatty acids (C8, C10, and C12), particularly caprate (C10), were dominant fatty acids and were different between the four treatment groups.

Short-chain fatty acids

Because liver abscesses are primarily anaerobic infections and SCFA are major products of anaerobic fermentations, the types and concentrations of SCFA were determined in purulent materials (Table 3). Not surprisingly, acetic acid was the predominant SCFA contained in the purulent material, which is a product of fermentation of both anaerobic and facultative anaerobic bacteria (Gorbach et al., 1976). The presence of propionic acid, isobutyric acid, butyric acid, and isovaleric acid are strong indicators of anaerobic bacterial fermentation (Ladas et al., 1979). The concentrations of SCFA, except butyric acid, were not affected by cattle type or inclusion of tylosin in the diet. Butyric acid had a significant cattle-type effect with concentration higher (P < 0.05) in Holstein steers compared to crossbred cattle. Interestingly, concentration of butyric acid, which is the primary product of F. necrophorum from fermentation of lactic acid, was low. The presence of branched-chain fatty acids (isobutyric acid, 2-methylbutyric, and isovaleric) are products of deamination and decarboxylation of branched-chain amino acids (valine, leucine, and isoleucine, respectively) (Allison, 1979). In anaerobic infections, good correlations have been found in the recovery of Bacteroides or Fusobacterium from purulent material containing isobutyric, butyric, and succinic acids (Gorbach et al., 1976). Lactic acid was the second most dominant acid in the purulent material, which is likely a product of glycogen fermentation.

Table 3.

Short-chain fatty acids and lactic acid concentrations (µM) in the purulent material of liver abscesses of crossbred cattle and Holstein steers fed high-grain diets with or without in-feed tylosin

Fermentation acids	Treatment groups				P-value
	Crossbred cattle with no tylosin (n = 12)	Crossbred cattle with tylosin (n = 12)	Holstein steers with no tylosin (n = 12)	Holstein steers with tylosin (n = 12)	Cattle type	Tylosin	Cattle type × Tylosin
Acetic acid	45.9	45.4	64.1	54.0	0.0651	0.4618	0.5027
Propionic acid	6.9	5.7	10.3	12.7	0.2664	0.8884	0.6898
Isobutyric acid	1.1	0.1	0.9	1.8	0.3252	0.9739	0.2771
Butyric acid	0.2	0.5	2.5	1.5	0.0418	0.6462	0.4248
2-Methylbutyric acid	0.5	0.2	0.5	1.4	0.3063	0.6912	0.3376
Isovaleric acid	1.7	0.3	1.5	1.1	0.5650	0.0894	0.2994
Valeric acid	0.005	0.009	0.022	0.1	0.2570	0.3656	0.3970
Caproic acid	0.012	0.022	0.014	0.027	0.5472	0.0548	0.7053
Lactic acid	14.9	33.9	27.9	21.4	0.9823	0.5443	0.2184

Lactate, lysine, and tryptophan concentrations and their metabolites

Because F. necrophorum is the primary etiologic agent and its biochemical role, at least in the rumen, is to ferment lactic acid and degrade amino acids, particularly lysine and tryptophan, their concentrations and their metabolites are of interest (Russell, 2006; Tadepalli et al., 2009). The metabolites related to lactate production and fermentation are shown in Figure 4A. The absolute concentration of lactate was higher in crossbred cattle with tylosin compared to no tylosin (Table 1). The increase in lactate in crossbred cattle with tylosin is likely because of higher prevalence of T. pyogenes, a lactic acid-producing organism in cattle with tylosin compared to no tylosin (Amachawadi et al., 2017). However, in Holstein steers, lactate was higher with no tylosin compared to tylosin-fed cattle. Among the metabolites of lactate fermentations, pyruvate, malate, fumarate, succinate, propionate, acetate, butyrate, and valerate were detected, but not oxaloacetate (Figure 4C). An intermediate of lactate fermentation product, succinate, was one of the top 30 biochemicals that differentiated the four treatment groups. The succinate had a cattle type and tylosin interaction, with concentration higher in crossbred cattle with no tylosin than the other three treatment groups (Figure 4B). Lysine (Figure 5A) and tryptophan (Figure 5C) levels were not affected by the treatment groups. The metabolites related to lysine and tryptophan fermentations are shown in Figures 6 and 7, respectively. The tryptophan metabolites, indoleacrylate, indolepropionate, tryptamines, and anthranilate were among the top 50 dominant ­biochemicals detected in the purulent material, and indolepropionate and tryptamine were among the top 10 metabolites.

Figure 4.

Lactate metabolic pathway (A), succinate concentration (B), and heat map (C) of lactate metabolites (fold change) in purulent materials of liver abscesses of crossbred cattle (CB) and Holstein steers (HS) fed high-grain diets with (Tyl) or without tylosin (no Tyl). Significant increases are indicated by red (P ≤ 0.05) and magenta colors (P ≥ 0.05 and ≤ 0.1) and significant reductions are indicated by dark green (P ≤ 0.05) and light green colors (P ≥ 0.05 and ≤ 0.1). Main effects significance are indicated by dark blue (P ≤ 0.05) and light blue (P ≥ 0.05 and ≤ 0.1).

Figure 5.

Lysine (A) and tryptophan (B) concentrations and their metabolic pathways (C and D) in purulent materials of liver abscesses of crossbred cattle (CB) and Holstein steers (HS) fed high-grain diets with (Tyl) or without tylosin (no Tyl).

Figure 6.

Heat map of lysine metabolites (fold change) in purulent materials of liver abscesses of crossbred cattle and Holstein steers fed high-grain diets with (Tyl) or without tylosin (no Tyl). Significant increases are indicated by red (P ≤ 0.05) and pink colors (P ≥ 0.05 and ≤ 0.1) and significant reductions are indicated by dark green (P ≤ 0.05) and light green colors (P ≥ 0.05 and ≤ 0.1). Main effects significance are indicated by dark blue (P ≤ 0.05) and light blue (P ≥ 0.05 and ≤ 0.1).

Figure 7.

Heat map of tryptophan metabolites (fold change) in purulent materials of liver abscesses of crossbred cattle (CB) and Holstein steers (HS) fed high-grain diets with (Tyl) or without tylosin (no Tyl). Significant increases are indicated by red color (P ≤ 0.05) and significant decreases are indicated by dark green (P ≤ 0.05) and light green (P ≥ 0.05 and ≤ 0.1). Main effects significance are indicated by dark blue P ≤ 0.05) and light blue (P ≥ 0.05 and ≤ 0.1).

Bacterial phospholipids

Phospholipids are primary components of host cell and bacterial cell membranes and lipoprotein particles. Lysophospholipids are intermediate products generated by the action of bacterial lipases on phospholipids (Sohlenkamp et al., 2016; Tan et al., 2020). Therefore, presence of lysophospholipids is indicative bacterial cell and host cell membrane degradation (Zheng et al., 2017). The list of lysophospholipids identified in the purulent material and the relative fold increase in the four treatment groups are shown in a heat map (Figure 8). A few of the lysophospholipids, such as 1-oleoyl-glycero-3-phospocholine (18:1), 1-stearoyl-glycero-3-phosphate (18:0), 1-oleoyl-glycero-3-phosphoethanolamine (18:0), 1-linoleoyl-glycero-3-phosphocholine (18:2), 1-palmitoyl-glycero-3-phosphoethanolamine (16:0), 1-oleoyl-glycero-3-phosphoethanolamine (18:1), 1-palmitoyl-glycero-3-phosphoglyceril (16:0), and 1-stearoyl-glycero-3-phosphoglyceril (18:0), were higher in abscesses of tylosin-fed compared to no-tylosin-fed group, perhaps because of bacterial cell deaths in response to tylosin feeding.

Figure 8.

Heat map of phospholipid metabolites (fold change) in purulent materials of liver abscesses of crossbred cattle (CB) and Holstein steers (HS) fed high-grain diets with (Tyl) or without tylosin (no Tyl). Significant increases are indicated by red (P ≤ 0.05) and pink colors (P ≥ 0.05 and ≤ 0.1) and significant decreases are indicated by dark green (P ≤ 0.05) and light green colors (P ≥ 0.05 and ≤ 0.1). Main effects significance are indicated by dark blue (P ≤ 0.05) and light blue (P ≥ 0.05 and ≤ 0.1).

Cattle with liver abscesses, even those that have severely abscessed livers (one or more large abscesses, often with adhesions or hundreds of small- to medium-sized abscesses), do not show any clinical signs. Attempts to relate blood changes—complete blood counts, serum chemicals, or liver enzymes—have not proven to be consistent. Lechtenberg and Nagaraja (1991) assessed liver function tests in peripheral blood of Holstein steers with experimentally induced liver abscesses. Hepatic dysfunction was evidenced by decrease in serum albumin concentration and low sulfobromophthalein clearance, an indicator of hepatocellular function. Dairy cows diagnosed with liver abscesses by ultrasonography had decreased serum globulin concentration (Dore et al., 2007). MacDonald et al. (2017) evaluated blood and biliary changes associated liver abscesses in cattle at slaughter and reported decreased concentrations of thyroxine, albumin, cholesterol, and alkaline phosphatase enzyme and higher concentrations of cortisol and aspartate aminotransferase in association with liver abscesses. In a recent study (Herrick et al., 2020) that assessed changes in complete blood cell counts and serum chemistry in fed Holstein steers with or without liver abscesses, increased platelet counts and decreased hemoglobin concentrations and hematocrit values in cattle with severely abscessed livers were observed. Serum analysis indicated increased globulin and decreased sodium, albumin, alanine aminotransferase, and aspartate aminotransferase compared to cattle with no abscesses. A possible outcome of the metabolomics analysis of purulent material may be identification of biochemicals that may be of diagnostic value. Many of the biochemicals are likely to be contained in healthy liver tissue also. The biomolecules need to be unique to liver abscesses, derived from bacterial activity and not a molecule contained in the healthy liver tissue, and importantly appear in the peripheral blood to be of value in diagnosis before slaughter. It is possible that diagnosis of liver abscesses in live cattle could aid in management decisions to decrease the impact of liver abscesses on animal health and animal performance. In fact, a major application of an antemortem diagnosis may be in the evaluation of novel interventions.

In conclusion, the metabolomics analysis of purulent material identified a total of 759 biomolecules, which included carbohydrates, energy metabolism pathways, peptides, amino acids and their metabolites, lipids and their metabolites, nucleotides, vitamins and cofactors, xenobiotics, and partially characterized molecules. The metabolites in the lipids and peptides and amino acids classes were the most predominant biochemical classes. The top 50 biochemicals in the purulent materials included biochemical classes and their metabolites of amino acids, lipids, nucleotides, xenobiotics, peptides, and carbohydrates. The 3-phenylpropionate, a product of Strickland reaction of phenylalanine, was the most dominant metabolite. Among the 15 amino acid metabolites, four (indoleacrylate, indolepropinate, tryptamine, anthranilate) were tryptophan metabolites. Biomolecules unique to liver abscesses that appear in the blood may have the potential to be used as a biomarker in the diagnosis of liver abscesses in cattle before slaughter. It is possible that antemortem diagnosis of liver abscesses could aid in management decisions to decrease the impact of liver abscesses on animal health and animal performance. In fact, a major application of an antemortem diagnosis may be in the evaluation of novel interventions.

Glossary

Abbreviations

ESI

electrospray ionization

FA

formic acid

HILIC

hydrophilic interaction liquid chromatography

RFA

random forest analysis

RI

retention index

RP

reverse phase

SCFA

short-chain fatty acids

UPLC-MS/MS

ultrahigh-performance liquid chromatography-tandem mass spectroscopy

Conflict of Interest Statement

The authors declare no real or perceived conflicts of interest.