Next Article in Journal
The Role of Comparative Psychology in the Training of Veterinarians
Next Article in Special Issue
Chronic Inflammatory Enteropathy and Low-Grade Intestinal T-Cell Lymphoma Are Associated with Altered Microbial Tryptophan Catabolism in Cats
Previous Article in Journal
Transcriptome-Based Identification of the Muscle Tissue-Specific Expression Gene CKM and Its Regulation of Proliferation, Apoptosis and Differentiation in Chicken Primary Myoblasts
Previous Article in Special Issue
Pathophysiology, Diagnosis, and Management of Canine Intestinal Lymphangiectasia: A Comparative Review
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
Animals
Volume 13
Issue 14
10.3390/ani13142313
animals-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles thumb_up
...
Endorse textsms
...
Comment
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Untargeted Analysis of Serum Metabolomes in Dogs with Exocrine Pancreatic Insufficiency

by
Patrick C. Barko
1,2,*,
Stanley I. Rubin
3,
Kelly S. Swanson
4,
Maureen A. McMichael
5,
Marcella D. Ridgway
1 and
David A. Williams
1,2
1
Department of Veterinary Clinical Medicine, University of Illinois at Urbana-Champaign, Urbana, IL 61802, USA
2
Department of Pathobiology, University of Illinois at Urbana-Champaign, Urbana, IL 61802, USA
3
VCA Animal Hospitals, Los Angeles, CA 90064, USA
4
Department of Animal Sciences and Division of Nutritional Sciences, College of Agricultural, Consumer and Environmental Sciences, University of Illinois at Urbana-Champaign, Urbana, IL 61801, USA
5
Department of Clinical Sciences, Auburn University, Auburn, AL 36849, USA
*
Author to whom correspondence should be addressed.
Animals 2023, 13(14), 2313; https://doi.org/10.3390/ani13142313
Submission received: 18 April 2023 / Revised: 3 July 2023 / Accepted: 4 July 2023 / Published: 14 July 2023
(This article belongs to the Special Issue Frontiers in Canine and Feline Gastrointestinal Disease)
Browse Figures
Review Reports Versions Notes

Abstract

:

Simple Summary

Exocrine pancreatic insufficiency (EPI) is a digestive disorder in dogs resulting from insufficient secretion of digestive enzymes from the exocrine pancreas. EPI is treated with oral pancreatic enzyme replacement therapy (PERT), but the persistence of clinical signs, especially diarrhea, is common after treatment. We sought to develop new insights into EPI using untargeted metabolomics analysis, a method that can measure hundreds of biochemicals (metabolites) in a sample. We analyzed 759 serum metabolites and identified 114 that varied significantly between dogs with EPI and healthy controls. Differences in fatty acid and amino acid metabolites were consistent with a state of malnourishment, and decreased vitamin B and C metabolites were suggestive of micronutrient deficiencies in dogs with EPI. Disturbances in gut microbial metabolites indicated the altered composition of the intestinal microbiome in dogs with EPI. Increased kynurenine, a tryptophan metabolite, in dogs with EPI may be associated with intestinal inflammation. As an exploratory study, causation cannot be determined from these results, but our findings have generated new data that can be used to inform future investigations of gastrointestinal and metabolic disturbances underlying the persistence of clinical signs in dogs with EPI treated with PERT.

Abstract

Exocrine pancreatic insufficiency (EPI) is a malabsorptive syndrome resulting from insufficient secretion of pancreatic digestive enzymes. EPI is treated with pancreatic enzyme replacement therapy (PERT), but the persistence of clinical signs, especially diarrhea, is common after treatment. We used untargeted metabolomics of serum to identify metabolic disturbances associated with EPI and generate novel hypotheses related to its pathophysiology. Fasted serum samples were collected from dogs with EPI (n = 20) and healthy controls (n = 10), all receiving PERT. Serum metabolomes were generated using UPLC-MS/MS, and differences in relative metabolite abundances were compared between the groups. Of the 759 serum metabolites detected, 114 varied significantly (p < 0.05, q < 0.2) between dogs with EPI and healthy controls. Differences in amino acids (arginate, homoarginine, 2-oxoarginine, N-acetyl-cadaverine, and α-ketoglutaramate) and lipids (free fatty acids and docosahexaenoylcarnitine) were consistent with increased proteolysis and lipolysis, indicating a persistent catabolic state in dogs with EPI. Relative abundances of gut microbial metabolites (phenyllactate, 4-hydroxyphenylacetate, phenylacetyl-amino acids, catechol sulfates, and o-cresol-sulfate) were altered in dogs with EPI, consistent with disruptions in gut microbial communities. Increased kynurenine is consistent with the presence of intestinal inflammation in dogs with EPI. Whether these metabolic disturbances participate in the pathophysiology of EPI or contribute to the persistence of clinical signs after treatment is unknown, but they are targets for future investigations.

1. Introduction

Canine exocrine pancreatic insufficiency (EPI) is a malabsorptive syndrome caused by insufficient secretion of pancreatic digestive enzymes [1,2]. Characteristic clinical signs of EPI include diarrhea, weight loss, and polyphagia. Clinical signs result from failure to digest dietary macromolecules (i.e., fats, carbohydrates, and proteins) resulting in malabsorption, negative energy balance, and altered composition of the gut microbiome (small intestinal bacterial overgrowth) [2]. Due to malabsorption, EPI is also associated with deficiencies in specific micronutrients, such as cobalamin and lipid-soluble vitamins [3,4,5]. The most common cause of canine EPI is end-stage pancreatic acinar atrophy (PAA), and clinical signs of EPI develop following an approximate 90% reduction in pancreatic acinar mass [6,7,8,9]. Serum concentrations of canine trypsin-like immunoreactivity (cTLI) less than 2.5 µg/L are diagnostic for EPI [10,11]. Dogs with EPI require life-long pancreatic enzyme replacement therapy (PERT), typically via the addition of enzyme extracts to each meal, and most require supplementation with cobalamin. The long-term prognosis for EPI is good, but responses to enzyme replacement therapy are variable, and a substantial proportion (~40%) of dogs with EPI have persistent clinical signs after the initiation of PERT, most commonly diarrhea and weight loss [12,13,14]. Thus, many dogs require adjunctive therapies, including therapeutic diets, antibiotics, probiotics, or some combination thereof, to control their clinical signs.
Little is known about the pathogenesis of PAA or its progression to EPI. Immunohistochemical and serologic studies have identified lymphocytic pancreatitis and circulating antibodies against pancreatic acini in dogs with subclinical PAA; however, in end-stage PAA associated with EPI, there is the replacement of acini with adipose tissue with minimal inflammation or fibrosis [9,15,16]. Over-representation of certain breeds, especially German Shepherds, among dogs with EPI has prompted investigations to identify a genetic basis for the disorder. The heritability of EPI is complex and does not follow a Mendelian pattern [17,18]. Recent genomics studies have identified a common risk allele in the dog leukocyte antigen complex in German Shepherd dogs and Pembroke Welsh Corgis with EPI [19,20]. However, the risk allele was not present in all dogs with EPI and was identified in some healthy dogs. Collectively, these findings suggest that there is an immune-mediated component to the pathogenesis of PAA and that one or more genetic or environmental risk factors may contribute to its pathogenesis.
Given the incomplete understanding of its etiopathogenesis, there is a need to generate new insights into the pathogenesis of EPI and its pathophysiologic consequences. There is also a need to identify pathophysiologic disturbances present in dogs with EPI after the initiation of PERT to develop novel therapeutic approaches for dogs with persistent signs of gastrointestinal dysfunction. Untargeted metabolomics analysis is an investigative tool that can provide new insights into poorly understood disease processes, especially those associated with nutritional and metabolic disturbances, such as EPI. The metabolome is the sum of all small molecules (<1500 kDa) present in an organism or biologic sample that are precursors, intermediates, and end-products of metabolic processes. Metabolites of endogenous and exogenous (diet, environment, and microbial) origin contribute to the serum metabolome of mammals [21]. The quantification of metabolites in a biologic sample can yield valuable insights into the metabolic phenotypes (i.e., metabotypes) of disease states by elucidating pathophysiologic mechanisms of disease and facilitating the discovery of novel biomarkers. The objectives of this investigation were to (1) detect differences in serum metabolite profiles between dogs with EPI and healthy controls and (2) to identify novel metabolic signatures relevant to the pathophysiology of EPI. While this was an untargeted, hypothesis-generating investigation, we hypothesized that metabolomics profiling would identify novel serum biomarkers plausibly associated with persistent gastrointestinal dysfunction in dogs with EPI following treatment with PERT.

2. Materials and Methods

2.1. Patient Population and Sample Collection

Adult (≥1 year of age) dogs with previously diagnosed EPI were recruited from an online patient registry (epi4dogs.com/epi-registry; registry data accessed 5 February 2016). Dogs diagnosed with EPI were identified from the registry data, and their owners and primary care veterinarians were contacted to collect patient histories and medical records. Dogs were eligible for inclusion if a review of their medical records identified subnormal serum cTLI concentrations (cTLI < 2.5 µg/L; reference interval: 5.7–45.2 µg/L) from two or more consecutive measurements. Additionally, dogs with EPI had to have been receiving oral PERT with a commercially available enzyme extract of porcine origin for at least 30 days prior to enrollment. The diets of participating dogs were not standardized, and they were fed their current diet under the supervision of their owners and primary veterinarians. Information regarding the diets and PERT of dogs with EPI included in this study are presented in Supplementary File S1. Dogs with EPI were excluded if they had been diagnosed with any other gastrointestinal, metabolic, endocrine, systemic, or neoplastic diseases by their primary care veterinarian. After an overnight fast, samples of whole blood were collected by the dogs’ primary care veterinarians via jugular or lateral saphenous venipuncture using sterile 20 g hypodermic needles and syringes. Whole blood samples were collected into empty, sterile tubes, allowed to clot at room temperature for 30 min, and centrifuged on site at 2500–3500 rpm for 10 min. The resulting serum was aliquoted into empty, sterile polypropylene tubes and immediately frozen on site in clinical freezers (approximately −20 °C). To standardize sample collection, all required materials (hypodermic needles, syringes, transfer pipets, sample tubes) and detailed instructions were provided to participating veterinarians by the investigators. All samples were mailed overnight on dry ice to the investigators, where they were stored at −80 °C following acquisition. Samples were collected and received by the investigators between September, 2016 and February, 2017. Aliquots of serum were submitted for measurement of serum cTLI, cobalamin, and folate using solid-phase chemiluminescent assays run on the IMMULITE 2000 PXi immunoassay system (Siemens Inc., Munich, Germany) at the Texas A&M Gastrointestinal Laboratory (College Station, TX, USA). Dogs with serum cTLI concentrations <2.5 µg/L were assigned to the EPI group, and those with serum cTLI concentrations ≥2.5 µg/L were excluded.
Healthy, adult (≥1 year of age), client-owned dogs were recruited as a control group. Owners completed a survey to gather information about their dogs’ medical history, while their medical records were reviewed by the investigators. To screen for confounding diseases, each healthy dog received a physical exam, and the following diagnostic screening tests were performed: complete blood count, serum biochemistry panel, urinalysis, serum total T4, serum cTLI, serum cobalamin, serum folate, and fecal flotation. Complete blood count, serum chemistries, urinalysis, total T4, and fecal flotation assays were performed at the University of Illinois Veterinary Diagnostic Laboratory (Urbana, IL, USA). Dogs were excluded from this group if they had any historical clinical signs of gastrointestinal disease (vomiting, diarrhea, weight loss, anorexia, polyphagia) within 6 months of recruitment; serum cTLI concentrations <5.7 µg/L; clinical signs or laboratory abnormalities consistent with any systemic, metabolic, gastrointestinal, or pancreatic diseases; or received antibiotics, corticosteroids, or immunomodulatory drugs within 6 months of recruitment. To account for the potential impact of pancreatic enzyme replacement therapy on serum metabolite profiles, a powdered pancreatic enzyme extract (Pancreatin 6X; American Laboratories Inc., Omaha, NE, USA) was administered orally (1 tsp/cup of food) to the healthy dogs for 14 days prior to sample collection. Following the supplementation period, serum samples were collected after an overnight fast as described above and immediately frozen at −80 °C.

2.2. Untargeted Serum Metabolomics

Serum metabolite profiles were generated via ultrahigh performance liquid chromatography-tandem mass spectroscopy (UPLC-MS/MS) by a commercial laboratory (Metabolon Inc., Morrisville, NC, USA) between June and September, 2017. Serum samples (100 µL) were deproteinated using methanol (500 µL) precipitation with vigorous shaking for 2 min followed by centrifugation. The methanol contained four recovery standards (DL-2-fluorophenylglycine, tridecanoic acid, d6-cholesterol, and 4-chlorophenylalanine) used to assess extraction efficiency. The resulting extracts were dried and reconstituted in solvents compatible to each of the four UPLC-MS/MS methods. Two aliquots were reconstituted in 50 μL of 6.5 mM ammonium bicarbonate in water (pH 8) for negative ion analysis methods, and two were reconstituted using 50 μL 0.1% formic acid in water (pH 3.5) for the positive ion analysis methods. Reconstitution solvents contained internal standards (d7-glucose, d3-leucine, d8-phenylalanine, d5-tryptophan, d5-hippuric acid, Br-phenylalanine, d5-indole acetic acid, amitriptyline, d9-progesterone) to assess instrument performance and to aid in chromatographic alignment. Analytic controls consisting of pooled samples, generated from small volumes of each experimental sample, were analyzed simultaneously with experimental samples to serve as a technical replicate throughout the dataset. Methanol-extracted water samples served as process blanks. Experimental samples were randomized across the platform run with quality control samples spaced evenly among the injections.
For UPLC-MS/MS, all methods utilized a Waters ACQUITY ultraperformance liquid chromatographer and a Thermo Scientific Q Exactive high resolution/accurate mass spectrometer interfaced with a heated electrospray ionization source and Orbitrap mass analyzer operated at 35,000 mass resolution. One aliquot was analyzed using acidic positive ion conditions optimized for 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 (PFPA) and 0.1% formic acid (FA). Another aliquot was analyzed using acidic positive ion conditions chromatographically optimized for hydrophobic compounds. In this method, the extract was gradient eluted from the same C18 column using methanol, acetonitrile, water, 0.05% PFPA, and 0.01% FA. Another aliquot was analyzed using basic negative ion optimized conditions using a separate dedicated C18 column (Waters UPLC BEH C18-2.1 × 100 mm, 1.7 μm). The basic extracts were gradient-eluted from the column using methanol and water with 6.5 mM ammonium bicarbonate at pH 8. The fourth aliquot was analyzed via negative ionization following elution from an 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 covered 70–1000 m/z. Data extraction, compound identification, and data processing were performed using a proprietary software platform by Metabolon Inc. Compounds were identified by comparison to library entries of purified and authenticated standards. Metabolites were quantified by measuring the area-under-the-curve of the chromatographic peak.

2.3. Statistical Analysis

Categorical variables from the demographics data were compared using the Fisher’s exact or chi-squared test. The normality of numerical demographic and clinical variables were assessed using Shapiro–Wilk tests and by examining histograms. For data that were normally distributed, t-tests were used to compare means, and for those with non-normal distributions, nonparametric Wilcoxon rank-sum tests were used. Results were considered statistically significant when p < 0.05. Relative abundances of serum metabolites were normalized using median-scaling, missing values were imputed with the sample set minimum, and the data were log-transformed. Unsupervised analysis consisted of principal component analysis (PCA) and hierarchical clustering of Euclidian distances. Welch’s two-sample t-test was used to detect serum metabolites that varied between the EPI and healthy control groups. Metabolite set enrichment analysis (MSEA) was used to identify metabolic subpathways enriched in metabolites that varied between groups. Metabolites were ranked by the product of the −log(p-value) from the t-tests and the log2-fold-change in the metabolites between groups. This resulted in a ranked vector where the top of the vector contained significantly variable metabolites with relative abundances that were increased, and the bottom of the vector contained significantly variable metabolites that were decreased in the EPI group compared with healthy controls. Next, MSEA was implemented by using the “fgsea” R package (version 1.22.0) to calculate normalized enrichment scores (NES) for each metabolic subpathway [22,23]. Subpathways with a NES > 0 were significantly upregulated, and those with a NES < 0 were significantly downregulated in the EPI group compared with healthy controls.
The Benjamini–Hochberg false discovery rate (FDR) procedure was used to calculate FDR-corrected p-values (q-values) for multiple comparisons in the metabolomics data [24,25]. We sought to maximize the discovery of features that vary in association with EPI while controlling for false discovery within reasonable bounds. Metabolomics results were considered statistically significant when p < 0.05 and q < 0.2. All statistical analyses were performed in the R language for statistical computing (version 4.2.1) using the RStudio integrated development environment running on macOS (Monterey 12.5).

3. Results

3.1. Demographic and Clinical Data

Serum samples from 30 dogs were analyzed for this study, including 20 dogs with EPI and 10 healthy controls. There were no statistically significant differences in age, breed, or reproductive status between the EPI and healthy control groups (Table 1). The persistence of clinical signs associated with EPI were common in dogs with EPI following treatment with PERT. Diarrhea was reported in 40% (n = 8) and persistent weight loss in 15% (n = 3) of dogs in the EPI group. Other clinical signs that are considered atypical for EPI were also reported in the EPI group, including decreased appetite in 10% (n = 2) and vomiting in 30% (n = 6). Dogs with EPI had received a variety of adjunct therapies in addition to PERT. Antibiotics were administered in 35% (n = 7) of dogs with EPI, including metronidazole in 5% (n = 1) and tylosin in 30% (n = 6). Cobalamin and folate supplementation was administered to 70% (n = 14) and 40% (n = 8) of dogs with EPI, respectively. Other adjunctive medications administered to dogs with EPI included gastric-acid-suppressing drugs in 30% (n = 6) and probiotics in 70% (n = 14). As the presence of the active clinical signs of gastrointestinal disease and administration of gastrointestinal medication were an exclusion criterion for healthy dogs, no statistical comparisons were performed between the EPI and healthy control groups. Except for one dog in the EPI group that was fed a home-prepared diet, all dogs included in this study were fed diets that were formulated to meet the nutritional levels established by the AAFCO Dog Food Nutrient Profiles (https://www.aafco.org; accessed 7 June 2023) for maintenance or animal feeding tests using AAFCO procedures that substantiated that the diet provides complete and balanced nutrition for the maintenance of adult dogs. Four dogs in the EPI group were fed mixtures of different commercial diets that were formulated to meet the AAFCO profiles. With the exception of the dog fed a home-prepared diet and those fed mixtures of commercial diets, all dogs were fed dry kibble diets. Mean (± SD) dietary fat (as-fed basis) was significantly (p = 0.012) higher in dogs in the EPI group (15.4% ± 1.7) compared with the healthy controls (13.3% ± 2.1), but there were no other significant differences in dietary protein (p = 0.08) or crude fiber (p = 0.95) between the groups.
Serum biomarkers for exocrine pancreatic and gastrointestinal function were measured in the sera of dogs included in this study: cTLI, cobalamin, and folate (Table 1). Serum concentrations of cTLI (median, IQR) were significantly lower in dogs with EPI (1.0 µg/L, 1.0–1.1) compared with the healthy controls (18.8 µg/L, 15.7–29.1; p < 0.001). After excluding dogs that had received supplementation with cobalamin or folate, serum cobalamin (median, IQR) concentrations were significantly lower in dogs with EPI (231 ng/L, 207–313) compared with the healthy controls (411 ng/L, 297–593; p = 0.013), but there were no statistically significant differences in serum folate concentrations between the groups (p = 0.39).

3.2. Untargeted Metabolomics Analysis

The untargeted serum metabolome dataset contained 759 named biochemicals. Among metabolic superpathways, lipids were the most numerous (n = 355), followed by amino acids (n = 187), xenobiotics (n = 79), nucleotides (n = 47), peptides (n = 31), cofactors and vitamins (n = 29), carbohydrates (n = 21), and energy (n = 10) metabolites. PCA separated samples from dogs with EPI from healthy controls along the first principal component (Figure 1A), consistent with metabolome-wide differences in the abundance of serum biochemicals among groups. Similarly, hierarchical clustering of metabolite relative abundances based on Euclidian distances revealed the clustering of most samples from the healthy controls away from those with EPI (Figure 1B).
We sought to detect individual metabolites that varied in association with EPI by comparing relative abundances of serum metabolites between EPI and the healthy controls. After controlling for false discovery (q < 0.2), the relative abundances of 114 serum metabolites varied significantly between the groups, with 76 being increased and 38 being decreased in dogs with EPI compared with the healthy controls. A selection of significantly variable serum metabolites with a large effect size is presented in Table 2. The complete output is presented in Supplementary File S2, and the plots of all significantly variable metabolites are presented in Supplementary File S3.
The amino acid and lipid superpathways contained the greatest number of metabolites that varied significantly between the groups (Figure 2A). Metabolite set enrichment analysis was performed to identify metabolic subpathways that were enriched among significantly variable metabolites (the complete results are in Supplementary File S4). After controlling for multiple comparisons, the “Phenylalanine Metabolism” subpathway (NES = 1.42, p = 0.005, q = 0.17) was found to be increased, and the “Purine Metabolism, Adenine containing” subpathway (NES = −1.57, p = 0.002, q = 0.12) was decreased in dogs with EPI (Figure 2B).

4. Discussion

The purpose of this investigation was to identify differences in serum metabolomes between healthy dogs and dogs with EPI. Unsupervised analysis with PCA and hierarchical clustering of Euclidian distances revealed the separation of samples from dogs with EPI and the healthy controls, consistent with metabolome-wide differences in serum metabolite profiles between the groups. The relative abundances of 143 serum metabolites varied in association with EPI (p < 0.05), and 114 of these were considered statistically significant after controlling for false discovery (q < 0.2). Pathway enrichment analysis identified two metabolic subpathways that were differentially enriched among significantly variable metabolites. The “Purine Metabolism, Adenine Containing” subpathway was downregulated in dogs with EPI. Within this subpathway, adenosine and adenosine 5′-monophosphate (AMP) were significantly lower in the serum of dogs with EPI. Previous studies have identified adenosine receptors in pancreatic acinar and ductal cells, and purines are implicated in the regulation of exocrine pancreatic secretions in rodents, humans, and dogs [26,27,28,29,30]. Additionally, adenosine has been found to exert an anti-inflammatory effect on intestinal epithelial cells, and it is plausible that the decreased availability of adenosine could contribute to clinical signs of gastrointestinal dysfunction that persist after PERT due to the dysregulation of intestinal mucosal inflammatory responses [31].
Several gut microbial metabolites were found to vary significantly in the sera of dogs with EPI compared with the healthy controls. The “Phenylalanine Metabolism” subpathway was significantly upregulated in association with EPI. This subpathway contains several gut microbial metabolites, including phenylpyruvate, 4-hydroxyphenylacetate, and phenyllactate, which were all significantly increased in the serum of dogs with EPI compared with the healthy controls. These biochemicals are microbial catabolites of phenylalanine that are generated by bacteria in the genera Lactobacillus and Bifidobacterium [32,33,34,35,36,37]. Previous studies of fecal microbiomes identified enrichment in genes associated with phenylalanine metabolism and increased abundances of Lactobacillus and Bifidobacterium in dogs with EPI [38]. The acetylated peptides phenylacetylglutamine and 4-hydroxyphenylacetylglutamine were significantly increased in dogs with EPI. A recent study integrating the deep metagenomic sequencing of gut microbiota with untargeted serum metabolomics in humans revealed strong, inverse correlations between serum concentrations of these acetylated peptides and the abundance of Faecalibacterium prausnitzii, an important commensal bacterium that maintains host health [39]. The sulfated benzoate derivates, 3-methoxycatechol sulfate and 4-hydroxycatechol sulfate, were decreased, and o-cresol-sulfate was increased in the sera of dogs with EPI. In a previous metagenomic study of human feces, the genes involved in benzoate catabolism via catechol were affiliated with diverse clades of bacteria including Bacteroidetes, Proteobacteria, Firmicutes, and Chloroflexi [40]. Notably, o-cresol sulfate is a well-characterized uremic toxin highlighting the interconnected nature of gut microbial and host metabolic systems in dogs. Consistent with our finding of the variation in numerous gut microbial metabolites in association with EPI, previous studies have identified microbiota dysbiosis in dogs with EPI characterized by small intestinal bacterial overgrowth, higher fecal abundance of E. coli, Lactobacillus, and Bifidobacterium and lower fecal abundance of Fusobacterium, C. hiranonis, and Faecalibacterium [38,41,42,43,44]. Interestingly, differences in fecal microbiomes in some dogs with EPI persist following treatment, and fecal microbiomes of dogs with EPI following the initiation of PERT are significantly different from healthy dogs and similar to those with EPI [38,44]. Consequently, it seems likely that these differences in the serum relative abundances of microbial metabolites in dogs with EPI are due to changes in gut microbiome composition, metabolic function, or both. As all dogs enrolled in this study were treated with PERT, our findings suggest that pancreatic enzyme supplementation may not be sufficient to restore eubiosis in dogs with EPI. These results also suggest that phenylalanine derivatives and acetylated peptides may be promising markers for enteric microbiota dysbiosis in dogs with EPI.
The presence of enteric microbiota dysbiosis in dogs with EPI is likely secondary to the malabsorption of dietary macromolecules due to the insufficient secretion of pancreatic digestive enzymes. However, it is possible that enteric microbiota participate in the pathogenesis or progression of PAA. In experiments in the 1970s, PAA-like lesions were induced in rats fed a liquid elemental diet [45,46]. The PAA-like lesions were rare in germ-free rats but invariably present in conventional rats. Germ-free rats purposefully contaminated with cecal contents from conventional mice also developed PAA-like lesions, but those contaminated with sterilized cecal contents did not. These findings suggest that PAA was induced by nutritional/metabolic disturbances, the impacts of which were at least partially conditioned by the presence of commensal enteric microbiota. Whether enteric microbiota participate in the pathogenesis of PAA or its progression to EPI in dogs is unknown.
Some of the metabolic disturbances we identified in dogs with EPI are similar to those observed during fasting or starvation, including greater relative abundances of urea cycle metabolites, purines, pyrimidines, and fatty acids [47]. Urea cycle amino acid derivatives (arginate, homoarginine, and 2-oxoarginine) were significantly lower, and N-acetyl-cadaverine (a lysine derivative) was greater in dogs with EPI, consistent with the increased rate of deamination reactions due to proteolysis. Greater relative abundances of free fatty acids, docosahexaenoylcarnitine (an acylcarnitine), and α-ketoglutaramate (a ketogenic amino acid) are consistent with the mobilization (lipolysis) of peripheral adipose stores and increased mitochondrial beta-oxidation reactions. Interestingly, some of the changes in microbial metabolites discussed above also indicate a state of malnourishment. In particular, increased plasma phenyllactate and urinary p-cresol have been identified in undernourished mice [35]. Collectively, these findings are consistent with the presence of a catabolic state in dogs with EPI, characterized by the mobilization of bodily stores of protein and adipose tissue, to maintain energy homeostasis.
Kynurenine, a derivative of tryptophan, was significantly greater in the serum of dogs with EPI (Supplementary Files S2 and S3). This metabolite is noteworthy owing to its association with intestinal mucosal inflammation in humans and rodent models of inflammatory bowel disease (IBD). Kynurenine is a mammalian catabolite of tryptophan that is generated in hepatic and peripheral tissues, primarily intestinal epithelial cells and leukocytes. In peripheral tissues, kynurenine is synthesized by indoleamine 2,3-dioxygenase-1 (IDO-1), the rate-limiting step in the peripheral kynurenine pathway [48]. The increased expression of IDO-1 and increased circulating kynurenine concentrations are associated with endoscopic inflammation and disease activity and predictive of outcomes in humans with IBD [49,50,51]. Thus, greater relative abundances of kynurenine are suggestive of the presence of intestinal mucosal inflammation in dogs with EPI, a factor that could contribute to the pathophysiology of gastrointestinal clinical signs, especially those atypical of EPI (e.g., vomiting, anorexia). Additional studies of kynurenine and other tryptophan catabolites in EPI are currently underway to confirm and further elucidate the importance of this finding.
We compared pancreatic and gastrointestinal biomarkers and identified significantly lower serum cobalamin in dogs with EPI compared with healthy controls. Cobalamin deficiency has been documented in up to 67% of dogs with EPI and is a negative prognostic indicator for therapeutic response [4,5]. The persistence of altered serum cobalamin in this population of dogs with EPI treated with PERT emphasizes the necessity to monitor serum cobalamin concentrations during treatment. Persistent cobalamin deficiency following PERT suggests that oral pancreatic enzyme supplementation alone is insufficient to restore cobalamin homeostasis. A previous study identified lower concentrations of lipid-soluble vitamins in this same population of dogs with EPI, consistent with persistent deficiency of micronutrient absorption in dogs with EPI treated with PERT [3]. We identified evidence of other micronutrient disturbances involving vitamin B6 and vitamin C, as evidenced by decreased pyridoxal and threonate, respectively. The causes for the altered metabolism of these water-soluble vitamins is unknown and should be investigated in future studies.
Our findings should be interpreted in light of several important limitations. First, this was an untargeted study intended to reveal novel metabolic markers and associations that are plausibly involved in the pathophysiology of EPI. Direct causal associations cannot be determined from any of these results. Many hundreds of metabolites were measured in a relatively small number of dogs, increasing the risk of discovering falsely significant features. We attempted to control for false discovery, but it is likely that some of the results we report as significant are false positives. Controlling the FDR at q < 0.2 means that 22/114 significant metabolites were likely to be false discoveries. We considered this to be an acceptable rate of false discovery given the objectives of the study. Future, hypothesis-driven investigations should confirm our findings using targeted analytic methods. As the serum samples from the dogs with EPI were collected in primary care veterinary clinics, preanalytical conditions could have impacted our results. We provided detailed instructions for primary care veterinarians to collect, process, store, and ship the samples. However, the samples were collected and processed by different personnel and in different laboratories. We cannot exclude the possibility that differences in the personnel, methods, and equipment (e.g., centrifuges) used to process the samples or a deviation from the sample collection protocol could have influenced our results. The sample sizes of the EPI and healthy control groups were different. We cannot exclude the possibility that the difference in group sizes could have influenced our results. Future investigations should assess serum metabolite concentrations using larger numbers of animals with similar sample sizes, and our results could inform such future investigations.
Regarding the patient population, we were unable to control environmental factors, including the dogs’ geographic locations or diets. This was a field study of animals with spontaneously occurring EPI, and some patients were receiving therapeutic diets as a component of therapy. Owing to this, we were not able to control the brands, nutrient compositions, quantities, or feeding frequencies of the diets fed to dogs enrolled in this study. Previous investigations have revealed the significant effects of diet on the serum metabolomes of dogs and other animals [52,53,54]. Though we did not standardize the diets of dogs in this study, we compared the macronutrient composition (protein, fat, and crude fiber) of their diets to understand whether there were significant differences between the EPI and healthy control groups. We identified a small, but statistically significant, difference in the dietary proportions of fat between the groups, with the diets of dogs in the EPI group containing a higher proportion of fat compared with the healthy controls. There were no significant differences in the dietary proportions of protein or crude fiber between the groups. However, because energy-basis nutrient profiles were not available for all diets, macronutrient data on an as-fed basis was analyzed which does not account for differences in moisture content or energy density between the diets. Data related to the quantities and frequencies of feedings were not collected, so their impacts cannot be assessed. Thus, we cannot exclude the possibility that differences in diet could have influenced our findings. Future investigations are needed to understand the influence of different diets on the serum metabolomes of dogs with EPI; these future studies should aim to include assessments of diets on an energy basis. There were small differences in the mean ages of the EPI and healthy controls groups, though the differences were not statistically significant. Serum metabolomes are known to vary with age in dogs and other mammals [55,56,57]. To understand whether differences in age were likely to have influenced our results, we attempted to detect correlations among relative abundances of serum metabolites and age. Of 114 serum metabolites that varied significantly between the EPI and healthy control groups, only 9 were significantly correlated with the age (Supplementary File S5). It is unlikely that the small differences in age between groups had a significant impact on our findings.
Treatments administered to dogs in the EPI group could have influenced our findings. The brands, strengths, and dosages of oral pancreatic enzyme extracts administered to dogs in the EPI group were not standardized. A study in piglets in which EPI was experimentally induced by pancreatic duct ligation, suggests that PERT can result in the normalization of enteric microbiota dysbiosis [58]. Thus, it is possible that some of our findings could have been influenced by inadequate PERT, resulting in some dogs with EPI having persistent digestive dysfunction and associated dysbiosis. While we attempted to control for the potential impact of PERT on serum metabolomes by administering pancreatic enzyme extracts to the healthy controls, it is likely that dogs with EPI respond differently to PERT compared with healthy dogs. It is also plausible that differences between the dogs with EPI and healthy controls could have been influenced by the duration of treatment. Dogs with EPI had been treated for months–years with PERT, whereas the healthy controls received enzyme extracts for only 14 days. We compared serum metabolite profiles in the healthy dogs before and after oral pancreatic enzyme administration, but there were no significant differences in metabolite abundances between the two groups (Supplementary File S6). Though we detected persistent clinical signs of gastrointestinal dysfunction in dogs with EPI, the clinical assessment of dogs with EPI was not among the objectives of this investigation. For this reason, we did not attempt to compare their serum metabolomes to dogs with a resolution of clinical signs. Lacking clinical assessments (e.g., weight, body condition, fecal quality scoring) before and after the initiation of PERT, we were not able to make accurate assessments of the clinical responses to PERT in dogs with EPI. The determination as to whether dogs had persistent clinical signs was based on subjective assessments made by the dogs’ owners and provided to the investigators during historical interviews. Owing to these factors, we concluded that our clinical data were insufficient to compare serum metabolomes between dogs with EPI that had resolved vs. persistent clinical signs. Future investigations should assess changes in the metabolomes of dogs following the initiation of PERT using more rigorous clinical assessments collected before and after the initiation of PERT. Based on historical interviews and reviews of the medical records, antibiotics, probiotics, and gastric-acid-suppressing drugs were administered to some dogs in the EPI group within the 30 days prior to sample collection. The exact brands, dosages, and administration frequencies of these treatments were not known for all dogs. These medications are known to impact enteric microbial communities in ways that could have affected our findings, especially in relation to the metabolites derived from enteric microbiota [59,60,61]. Targeted follow-up studies in dogs with EPI not exposed to antibiotics and probiotics are needed to confirm our findings.

5. Conclusions

This investigation identified relative abundances of 114 serum metabolites that varied significantly between dogs with EPI and healthy controls after pancreatic enzyme replacement therapy. The differences in amino acids and lipids were consistent with increased proteolysis and lipolysis, indicating a persistent catabolic state in dogs with EPI. The relative abundances of gut microbial metabolites were altered in dogs with EPI, consistent with disruptions in gut microbial communities. Increased kynurenine suggests the presence of intestinal inflammation in dogs with EPI. Our findings have revealed evidence of persistent gastrointestinal dysfunction, enteric microbiota dysbiosis, and altered lipid, amino acid, nucleotide, and peptide metabolism in dogs with EPI receiving PERT. The extent to which these metabolic alterations participate in the pathogenesis or pathophysiology of EPI is unknown. These findings should be interpreted with caution and should be subject to future, hypothesis-driven research to confirm and further elucidate their pathophysiologic significance. Future studies should investigate the metabolic disturbances we have identified in larger populations of dogs with EPI while attempting to standardize diets and therapeutic strategies. Additionally, efforts to understand whether the persistence of gastrointestinal clinical signs and dysbiosis in dogs with EPI following the initiation of PERT is related to inadequate treatment, a concurrent chronic enteropathy, or an unidentified facet of its pathophysiology are needed.

Supplementary Materials

The following supporting information can be downloaded at: https://www.mdpi.com/article/10.3390/ani13142313/s1. File S1: Information about diet and pancreatic enzyme histories; File S2: Complete output of t-test comparing serum metabolites between dogs with EPI and healthy controls; File S3: Boxplots of all significantly variable serum metabolites; File S4: Complete output of metabolite set enrichment analysis; File S5: Correlations among serum metabolite abundances and age; File S6: Comparisons of serum metabolomes in healthy dogs before and after pancreatic enzyme supplementation.

Author Contributions

Conceptualization, P.C.B., S.I.R., D.A.W.; methodology, P.C.B., S.I.R., K.S.S., M.A.M., M.D.R., D.A.W.; formal analysis, P.C.B.; data curation, P.C.B.; writing—original draft preparation, P.C.B.; writing—review and editing, P.C.B., S.I.R., K.S.S., M.A.M., M.D.R., D.A.W.; visualization, P.C.B.; supervision, D.A.W.; project administration, D.A.W.; funding acquisition, P.C.B. All authors have read and agreed to the published version of the manuscript.

Funding

This investigation was funded by a grant program for veterinary medical residents sponsored by Nestle Purina PetCare, Inc. and a generous donation from Paula and John Gatens. Funding was awarded to one of the primary authors (PCB) and the principal investigator (DAW). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Institutional Review Board Statement

This investigation was approved by the University of Illinois Institutional Animal Care and Use Committee (IACUC #14284; approved 12 December 2014).

Informed Consent Statement

All samples from dogs in the EPI groups were collected by primary care veterinarians from client-owned dogs with spontaneously occurring EPI. Samples from dogs in the healthy control group were collected from client-owned dogs lacking a history or clinical signs of gastrointestinal dysfunction. All owners provided informed consent.

Data Availability Statement

All data and R code necessary to replicate this analysis are located in a GitHub repository (https://github.com/pcbarko/K9_EPI_Serum_Metabolome, accessed on 28 June 2023).

Acknowledgments

The authors thank Olesia Kennedy, Epi4Dogs Foundation Inc., and Paula and John Gatens for their invaluable support for this investigation. The authors would also like to acknowledge Diane Sloan for donating the pancreatin used in this investigation. Without their generous assistance, this investigation would not have been possible.

Conflicts of Interest

The authors declare no conflict of interest.

References

  1. Westermarck, E.; Wiberg, M. Exocrine Pancreatic Insufficiency in the Dog: Historical Background, Diagnosis, and Treatment. Top. Companion Anim. Med. 2012, 27, 96–103. [Google Scholar] [CrossRef]
  2. Batt, R.M. Exocrine Pancreatic Insufficiency. Vet. Clin. N. Am. Small Anim. Pract. 1993, 23, 595–608. [Google Scholar] [CrossRef]
  3. Barko, P.C.; Williams, D.A. Serum concentrations of lipid-soluble vitamins in dogs with exocrine pancreatic insufficiency treated with pancreatic enzymes. J. Vet. Intern. Med. 2018, 32, 1600–1608. [Google Scholar] [CrossRef]
  4. Soetart, N.; Rochel, D.; Drut, A.; Jaillardon, L. Serum cobalamin and folate as prognostic factors in canine exocrine pancreatic insufficiency: An observational cohort study of 299 dogs. Vet. J. 2019, 243, 15–20. [Google Scholar] [CrossRef]
  5. Chang, C.; Lidbury, J.A.; Suchodolski, J.S.; Steiner, J.M. Effect of oral or injectable supplementation with cobalamin in dogs with hypocobalaminemia caused by chronic enteropathy or exocrine pancreatic insufficiency. J. Vet. Intern. Med. 2022, 36, 1607–1621. [Google Scholar] [CrossRef]
  6. Wiberg, M.E.; Nurmi, A.-K.; Westermarck, E. Serum Trypsinlike Immunoreactivity Measurement for the Diagnosis of Subclinical Exocrine Pancreatic Insufficiency. J. Vet. Intern. Med. 1999, 13, 426–432. [Google Scholar] [CrossRef]
  7. Wiberg, M.E.; Westermarck, E. Subclinical exocrine pancreatic insufficiency in dogs. J. Am. Vet. Med. Assoc. 2002, 220, 1183–1187. [Google Scholar] [CrossRef]
  8. Wiberg, M. Pancreatic acinar atrophy in German shepherd dogs and rough-coated Collies. Etiopathogenesis, diagnosis and treatment. A review. Vet. Q. 2004, 26, 61–75. [Google Scholar] [CrossRef] [Green Version]
  9. Westermarck, E.; Batt, R.M.; Vaillant, C.; Wiberg, M. Sequential study of pancreatic structure and function during development of pancreatic acinar atrophy in a German shepherd dog. Am. J. Vet. Res. 1993, 54, 1088–1094. [Google Scholar]
  10. Williams, D.A.; Batt, R.M. Sensitivity and specificity of radioimmunoassay of serum trypsin-like immunoreactivity for the diagnosis of canine exocrine pancreatic insufficiency. J. Am. Vet. Med. Assoc. 1988, 192, 195–201. [Google Scholar]
  11. Williams, D.A.; Batt, R.M. Diagnosis of canine exocrine pancreatic insufficiency by the assay of serum trypsin-like immunoreactivity. J. Small Anim. Pract. 1983, 24, 583–588. [Google Scholar] [CrossRef]
  12. Batchelor, D.J.; Noble, P.-J.M.; Taylor, R.H.; Cripps, P.J.; German, A.J. Prognostic Factors in Canine Exocrine Pancreatic Insufficiency: Prolonged Survival is Likely if Clinical Remission is Achieved. J. Vet. Intern. Med. 2007, 21, 54–60. [Google Scholar] [CrossRef]
  13. German, A.J. Exocrine Pancreatic Insufficiency in the Dog: Breed Associations, Nutritional Considerations, and Long-term Outcome. Top. Companion Anim. Med. 2012, 27, 104–108. [Google Scholar] [CrossRef]
  14. Wiberg, M.; Lautala, H.; Westermarck, E. Response to long-term enzyme replacement treatment in dogs with exocrine pancreatic insufficiency—PubMed. J. Am. Vet. Med. Assoc. 1998, 213, 86–90. [Google Scholar]
  15. Wiberg, M.E.; Saari, S.A.; Westermarck, E.; Meri, S. Cellular and humoral immune responses in atrophic lymphocytic pancreatitis in German shepherd dogs and rough-coated collies. Vet. Immunol. Immunopathol. 2000, 76, 103–115. [Google Scholar] [CrossRef]
  16. Wiberg, M.E.; Saari, S.A.M.; Westermarck, E. Exocrine Pancreatic Atrophy in German Shepherd Dogs and Rough-coated Collies: An End Result of Lymphocytic Pancreatitis. Vet. Pathol. 1999, 36, 530–541. [Google Scholar] [CrossRef]
  17. Westermarck, E.; Saari, S.; Wiberg, M. Heritability of Exocrine Pancreatic Insufficiency in German Shepherd Dogs. J. Vet. Intern. Med. 2010, 24, 450–452. [Google Scholar] [CrossRef]
  18. Tsai, K.L.; Noorai, R.E.; Starr-Moss, A.N.; Quignon, P.; Rinz, C.J.; Ostrander, E.A.; Steiner, J.M.; Murphy, K.E.; Clark, L.A. Genome-wide association studies for multiple diseases of the German Shepherd Dog. Mamm. Genome 2011, 23, 203–211. [Google Scholar] [CrossRef] [Green Version]
  19. Evans, J.M.; Tsai, K.L.; Starr-Moss, A.N.; Steiner, J.M.; Clark, L.A. Association ofDLA-DQB1alleles with exocrine pancreatic insufficiency in Pembroke Welsh Corgis. Anim. Genet. 2015, 46, 462–465. [Google Scholar] [CrossRef]
  20. Tsai, K.L.; Starr-Moss, A.N.; Venkataraman, G.M.; Robinson, C.; Kennedy, L.J.; Steiner, J.M.; Clark, L.A. Alleles of the major histocompatibility complex play a role in the pathogenesis of pancreatic acinar atrophy in dogs. Immunogenetics 2013, 65, 501–509. [Google Scholar] [CrossRef] [Green Version]
  21. Wishart, D.S. Metabolomics for Investigating Physiological and Pathophysiological Processes. Physiol. Rev. 2019, 99, 1819–1875. [Google Scholar] [CrossRef]
  22. Sergushichev, A. Algorithm for cumulative calculation of gene set enrichment statistic. Sci. Technol. J. Inf. Technol. Mech. Opt. 2016, 16, 956. [Google Scholar] [CrossRef]
  23. Korotkevich, G.; Sukhov, V.; Budin, N.; Shpak, B.; Artyomov, M.N.; Sergushichev, A. Fast gene set enrichment analysis. bioRxiv 2021, 060012. [Google Scholar] [CrossRef] [Green Version]
  24. Benjamini, Y.; Hochberg, Y. Controlling the False Discovery Rate: A Practical and Powerful Approach to Multiple Testing. J. Royal Statistical Soc. Ser. B Methodol. 1995, 57, 289–300. [Google Scholar] [CrossRef]
  25. Storey, J.D.; Bass, A.J.; Dabney, A.; Robinson, D. Qvalue: Q-Value Estimation for False Discovery Rate Control. 2020. Available online: http://github.com/jdstorey/qvalue (accessed on 28 June 2023).
  26. Hayashi, M.; Inagaki, A.; Novak, I.; Matsuda, H. The adenosine A2B receptor is involved in anion secretion in human pancreatic duct Capan-1 epithelial cells. Pflugers Arch. 2016, 468, 1171–1181. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  27. Iwatsuki, K. Subtypes of adenosine receptors on pancreatic exocrine secretion in anaesthetized dogs. Fundam. Clin. Pharmacol. 2000, 14, 203–208. [Google Scholar] [CrossRef] [PubMed]
  28. Novak, I.; Hede, S.E.; Hansen, M.R. Adenosine receptors in rat and human pancreatic ducts stimulate chloride transport. Pflugers Arch. 2008, 456, 437–447. [Google Scholar] [CrossRef]
  29. Burnstock, G.; Novak, I. Purinergic signalling in the pancreas in health and disease. J. Endocrinol. 2012, 213, 123–141. [Google Scholar] [CrossRef] [Green Version]
  30. Hayashi, M. Expression of Adenosine Receptors in Rodent Pancreas. Int. J. Mol. Sci. 2019, 20, 5329. [Google Scholar] [CrossRef] [Green Version]
  31. Ye, J.H.; Rajendran, V.M. Adenosine: An immune modulator of inflammatory bowel diseases. World J. Gastroenterol. 2009, 15, 4491–4498. [Google Scholar] [CrossRef]
  32. Beloborodova, N.; Bairamov, I.; Olenin, A.; Shubina, V.; Teplova, V.; Fedotcheva, N. Effect of phenolic acids of microbial origin on production of reactive oxygen species in mitochondria and neutrophils. J. Biomed. Sci. 2012, 19, 89. [Google Scholar] [CrossRef] [Green Version]
  33. Li, X.; Jiang, B.; Pan, B.; Mu, W.; Zhang, T. Purification and Partial Characterization of Lactobacillus Species SK007 Lactate Dehydrogenase (LDH) Catalyzing Phenylpyruvic Acid (PPA) Conversion into Phenyllactic Acid (PLA). J. Agric. Food Chem. 2008, 56, 2392–2399. [Google Scholar] [CrossRef] [PubMed]
  34. Jia, J.; Mu, W.; Zhang, T.; Jiang, B. Bioconversion of Phenylpyruvate to Phenyllactate: Gene Cloning, Expression, and Enzymatic Characterization of d- and l1-Lactate Dehydrogenases from Lactobacillus plantarum SK002. Appl. Biochem. Biotechnol. 2010, 162, 242–251. [Google Scholar] [CrossRef] [PubMed]
  35. Preidis, G.A.; Ajami, N.J.; Wong, M.C.; Bessard, B.C.; Conner, M.E.; Petrosino, J.F. Microbial-Derived Metabolites Reflect an Altered Intestinal Microbiota during Catch-Up Growth in Undernourished Neonatal Mice. J. Nutr. 2016, 146, 940–948. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  36. Nierop Groot, M.N.; de Bont, J.A.M. Conversion of Phenylalanine to Benzaldehyde Initiated by an Aminotransferase in Lactobacillus plantarum. Appl. Environ. Microbiol. 1998, 64, 3009–3013. [Google Scholar] [CrossRef] [Green Version]
  37. Visconti, A.; Le Roy, C.I.; Rosa, F.; Rossi, N.; Martin, T.C.; Mohney, R.P.; Li, W.; de Rinaldis, E.; Bell, J.T.; Venter, J.C.; et al. Interplay between the human gut microbiome and host metabolism. Nat. Commun. 2019, 10, 4505. [Google Scholar] [CrossRef] [Green Version]
  38. Isaiah, A.; Parambeth, J.C.; Steiner, J.M.; Lidbury, J.A.; Suchodolski, J.S. The fecal microbiome of dogs with exocrine pancreatic insufficiency. Anaerobe 2017, 45, 50–58. [Google Scholar] [CrossRef]
  39. Dekkers, K.F.; Sayols-Baixeras, S.; Baldanzi, G.; Nowak, C.; Hammar, U.; Nguyen, D.; Varotsis, G.; Brunkwall, L.; Nielsen, N.; Eklund, A.C.; et al. An online atlas of human plasma metabolite signatures of gut microbiome composition. Nat. Commun. 2022, 13, 5370. [Google Scholar] [CrossRef]
  40. Yadav, M.; Lomash, A.; Kapoor, S.; Pandey, R.; Chauhan, N.S. Mapping of the benzoate metabolism by human gut microbiome indicates food-derived metagenome evolution. Sci. Rep. 2021, 11, 5561. [Google Scholar] [CrossRef]
  41. Williams, D.A.; Batt, R.M.; McLean, L. Bacterial overgrowth in the duodenum of dogs with exocrine pancreatic insufficiency. J. Am. Vet. Med. Assoc. 1987, 191, 201–206. [Google Scholar]
  42. Simpson, K.; Morton, D.; Sørensen, S.; McLEAN, L.; Riley, J.; Batt, R. Biochemical changes in the jejunal mucosa of dogs with exocrine pancreatic insufficiency following pancreatic duct ligation. Res. Vet. Sci. 1989, 47, 338–345. [Google Scholar] [CrossRef] [PubMed]
  43. Simpson, K.W.; Batt, R.M.; Jones, D.; Morton, D.B. Effects of exocrine pancreatic insufficiency and replacement therapy on the bacterial flora of the duodenum in dogs. Am. J. Vet. Res. 1990, 51, 203–206. [Google Scholar] [PubMed]
  44. Blake, A.B.; Guard, B.C.; Honneffer, J.B.; Lidbury, J.A.; Steiner, J.M.; Suchodolski, J.S. Altered microbiota, fecal lactate, and fecal bile acids in dogs with gastrointestinal disease. PLoS ONE 2019, 14, e0224454. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  45. Geever, E.F.; Seifter, E.; Levenson, S.M. Pancreatic pathology, chemically defined liquid diets and bacterial flora in the rat. Br. J. Exp. Pathol. 1970, 51, 341–347. [Google Scholar] [PubMed]
  46. Levenson, S.M.M.; Kan, D.M.; Gruber, C.B.; Crowley, L.B.; Jaffe, E.R.M.; Nakao, K.M.; Geever, E.F.M.; Seifter, E. Hemolytic Anemia and Pancreatic Acinar Atrophy and Fibrosis Conditioned by “Elemental” Liquid Diets and the Ordinary Intenstinal Microflora. Ann. Surg. 1971, 174, 469–510. [Google Scholar] [CrossRef]
  47. Kondoh, H.; Teruya, T.; Yanagida, M. Metabolomics of human fasting: New insights about old questions. Open Biol. 2020, 10, 200176. [Google Scholar] [CrossRef]
  48. Haq, S.; Grondin, J.A.; Khan, W.I. Tryptophan-derived serotonin-kynurenine balance in immune activation and intestinal inflammation. FASEB J. 2021, 35, e21888. [Google Scholar] [CrossRef]
  49. Gupta, N.K.; Thaker, A.I.; Kanuri, N.; Riehl, T.E.; Rowley, C.W.; Stenson, W.F.; Ciorba, M.A. Serum Analysis of Tryptophan Catabolism Pathway: Correlation with Crohn’s Disease Activity. Inflamm. Bowel Dis. 2012, 18, 1214–1220. [Google Scholar] [CrossRef] [Green Version]
  50. Sofia, M.A.; Ciorba, M.A.; Meckel, K.; Lim, C.K.; Guillemin, G.J.; Weber, C.R.; Bissonnette, M.; Pekow, J.R. Tryptophan Metabolism through the Kynurenine Pathway is Associated with Endoscopic Inflammation in Ulcerative Colitis. Inflamm. Bowel Dis. 2018, 24, 1471–1480. [Google Scholar] [CrossRef]
  51. Nikolaus, S.; Schulte, B.; Al-Massad, N.; Thieme, F.; Schulte, D.M.; Bethge, J.; Rehman, A.; Tran, F.; Aden, K.; Häsler, R.; et al. Increased Tryptophan Metabolism Is Associated with Activity of Inflammatory Bowel Diseases. Gastroenterology 2017, 153, 1504–1516. [Google Scholar] [CrossRef] [Green Version]
  52. Deng, P.; Jones, J.C.; Swanson, K.S. Effects of dietary macronutrient composition on the fasted plasma metabolome of healthy adult cats. Metabolomics 2013, 10, 638–650. [Google Scholar] [CrossRef]
  53. Lyu, Y.; Liu, D.; Nguyen, P.; Peters, I.; Heilmann, R.M.; Fievez, V.; Hemeryck, L.Y.; Hesta, M. Differences in Metabolic Profiles of Healthy Dogs Fed a High-Fat vs. a High-Starch Diet. Front. Vet. Sci. 2022, 9, 801863. [Google Scholar] [CrossRef] [PubMed]
  54. Wu, G.D.; Compher, C.; Chen, E.Z.; Smith, S.A.; Shah, R.D.; Bittinger, K.; Chehoud, C.; Albenberg, L.G.; Nessel, L.; Gilroy, E.; et al. Comparative metabolomics in vegans and omnivores reveal constraints on diet-dependent gut microbiota metabolite production. Gut 2016, 65, 63–72. [Google Scholar] [CrossRef] [PubMed]
  55. Puurunen, J.; Ottka, C.; Salonen, M.; Niskanen, J.E.; Lohi, H. Age, breed, sex and diet influence serum metabolite profiles of 2000 pet dogs. R. Soc. Open Sci. 2022, 9, 211642. [Google Scholar] [CrossRef]
  56. Yue, T.; Tan, H.; Shi, Y.; Xu, M.; Luo, S.; Weng, J.; Xu, S. Serum Metabolomic Profiling in Aging Mice Using Liquid Chromatography—Mass Spectrometry. Biomolecules 2022, 12, 1594. [Google Scholar] [CrossRef]
  57. Yu, Z.; Zhai, G.; Singmann, P.; He, Y.; Xu, T.; Prehn, C.; Römisch-Margl, W.; Lattka, E.; Gieger, C.; Soranzo, N.; et al. Human serum metabolic profiles are age dependent. Aging Cell 2012, 11, 960–967. [Google Scholar] [CrossRef]
  58. Ritz, S.; Hahn, D.; Wami, H.T.; Tegelkamp, K.; Dobrindt, U.; Schnekenburger, J. Gut microbiome as a response marker for pancreatic enzyme replacement therapy in a porcine model of exocrine pancreas insufficiency. Microb. Cell Factories 2020, 19, 221. [Google Scholar] [CrossRef]
  59. Whittemore, J.C.; Price, J.M.; Moyers, T.; Suchodolski, J.S. Effects of Synbiotics on the Fecal Microbiome and Metabolomic Profiles of Healthy Research Dogs Administered Antibiotics: A Randomized, Controlled Trial. Front. Vet. Sci. 2021, 8, 665713. [Google Scholar] [CrossRef]
  60. Stavroulaki, E.M.; Suchodolski, J.S.; Xenoulis, P.G. Effects of antimicrobials on the gastrointestinal microbiota of dogs and cats. Vet. J. 2023, 291, 105929. [Google Scholar] [CrossRef]
  61. Garcia-Mazcorro, J.F.; Suchodolski, J.S.; Jones, K.R.; Clark-Price, S.C.; Dowd, S.E.; Minamoto, Y.; Markel, M.; Steiner, J.M.; Dossin, O. Effect of the proton pump inhibitor omeprazole on the gastrointestinal bacterial microbiota of healthy dogs. FEMS Microbiol Ecol. 2012, 80, 624–636. [Google Scholar] [CrossRef]
Animals 13 02313 g001 550
Figure 1. Unsupervised analysis of serum metabolite profiles. Samples are annotated by color according to their group: EPI (orange), healthy (blue). (A) Principal component analysis of relative metabolite abundances. (B) Statistical heatmap of relative metabolite abundances using Euclidian distances and hierarchical clustering.
Figure 1. Unsupervised analysis of serum metabolite profiles. Samples are annotated by color according to their group: EPI (orange), healthy (blue). (A) Principal component analysis of relative metabolite abundances. (B) Statistical heatmap of relative metabolite abundances using Euclidian distances and hierarchical clustering.
Animals 13 02313 g001
Animals 13 02313 g002 550
Figure 2. Metabolic pathway enrichment among serum metabolites. (A) Counts of significantly variable metabolites from each metabolic superpathway. Each bar represents the number of significantly variable metabolites in each superpathway. The red bars show the number of metabolites that were increased in dogs with EPI compared with healthy controls. The blue bars show the number of metabolites that were decreased in dogs with EPI compared with healthy controls. (B) Metabolite set enrichment analysis of metabolic subpathways (MSEA). MSEA was used to detect metabolic subpathways enriched among metabolites with high (positive NES) and low (negative NES) stability. The 20 subpathways with the highest absolute value of NES are shown. Bars are annotated using the FDR-adjusted p-value (q-value), such that grey bars represent pathways with q-value ≥ 0.2 and orange bars represent subpathways with q-value < 0.2.
Figure 2. Metabolic pathway enrichment among serum metabolites. (A) Counts of significantly variable metabolites from each metabolic superpathway. Each bar represents the number of significantly variable metabolites in each superpathway. The red bars show the number of metabolites that were increased in dogs with EPI compared with healthy controls. The blue bars show the number of metabolites that were decreased in dogs with EPI compared with healthy controls. (B) Metabolite set enrichment analysis of metabolic subpathways (MSEA). MSEA was used to detect metabolic subpathways enriched among metabolites with high (positive NES) and low (negative NES) stability. The 20 subpathways with the highest absolute value of NES are shown. Bars are annotated using the FDR-adjusted p-value (q-value), such that grey bars represent pathways with q-value ≥ 0.2 and orange bars represent subpathways with q-value < 0.2.
Animals 13 02313 g002
Table
Table 1. Patient population statistics. Categorical variables compared using Fisher’s exact or chi-squared tests. Variables with normal distributions are presented as means and standard deviations (SD) and are otherwise presented as medians and interquartile ranges (IQRs). Comparisons of serum cobalamin and folate omit dogs receiving supplementation. Laboratory reference intervals (RIs) are listed for cTLI, cobalamin, and folate.
Table 1. Patient population statistics. Categorical variables compared using Fisher’s exact or chi-squared tests. Variables with normal distributions are presented as means and standard deviations (SD) and are otherwise presented as medians and interquartile ranges (IQRs). Comparisons of serum cobalamin and folate omit dogs receiving supplementation. Laboratory reference intervals (RIs) are listed for cTLI, cobalamin, and folate.
HealthyEPIp-Value
Age (years) 0.36
Mean (SD)5.0 (± 1.74)4.3 (± 2.0)
Breed 0.62
Akita0 (0%)1 (5%)
Australian Shepherd0 (0%)2 (10%)
Border Collie0 (0%)1 (5%)
Cavalier King Charles Spaniels0 (0%)1 (5%)
German Shepherd4 (40%)9 (45%)
Labrador Retriever1 (10%)1 (5%)
Mixed Breed4 (40%)4 (20%)
Pit Bull Terrier1 (10%)0 (0%)
West Highland White Terrier0 (0%)1 (5%)
Sex 0.16
Spayed Female3 (30%)11 (55%)
Intact Male0 (0%)2 (10%)
Neutered Male7 (70%)7 (35%)
Serum cTLI (RI: 5.7–45.2 µg/L) <0.001
Median (IQR)18.8 (15.7–29.1)1.0 (1.0–1.1)
Serum cobalamin (RI: 251–908 ng/L) 0.01
Median (IQR)411 (297–593)231 (207–313)
Serum folate (RI: 7.7–24.4 ng/L) 0.39
Mean (SD)12.1 (±5.2)13.9 (±4.5)
Table
Table 2. Significantly variable serum metabolites. Significantly variable serum metabolites with large effect size (|log2-fold-change|> 0.6; corresponding > 1.5 fold difference in means). p-values derived from Welch’s t-test. q-value, false discovery rate-adjusted p-values; log2FC, log2 of the fold difference in means between dogs with EPI and healthy controls.
Table 2. Significantly variable serum metabolites. Significantly variable serum metabolites with large effect size (|log2-fold-change|> 0.6; corresponding > 1.5 fold difference in means). p-values derived from Welch’s t-test. q-value, false discovery rate-adjusted p-values; log2FC, log2 of the fold difference in means between dogs with EPI and healthy controls.
BiochemicalSubpathwayp-Valueq-Valuelog2FC
Amino Acids
Alpha-ketoglutaramateGlutamate Metabolism0.0010.0190.99
Cysteinylglycine DisulfideGlutathione Metabolism0.0010.023−1.31
Cysteine-glutathione Disulfide0.0030.043−0.93
4-guanidinobutanoateGuanidino and Acetamido Metabolism0.0020.0341.65
HomocarnosineHistidine Metabolism0.0010.0171.24
5-aminovalerateLysine Metabolism0.0020.0340.99
N-acetyl-cadaverine0.0010.0241.19
CystineMethionine, Cysteine, s-Adenosylmethionine, and Taurine Metabolism0.000<0.001−2.29
PhenyllactatePhenylalanine Metabolism<0.0010.0100.93
Phenylpyruvate<0.0010.0111.31
4-hydroxyphenylacetate<0.0010.0002.09
Phenol SulfateTyrosine Metabolism0.0030.042−1.93
4-hydroxyphenylpyruvate0.0010.0230.83
4-hydroxyphenylacetatoylcarnitine<0.0010.0013.61
Pro-hydroxy-proUrea cycle; Arginine and Proline Metabolism0.0010.023−1.57
Argininate0.0010.0140.66
2-oxoarginine<0.0010.0012.44
Carbohydrates
1,5-anhydroglucitolGlycolysis, Gluconeogenesis, and Pyruvate Metabolism<0.0010.006−0.73
Glycerate<0.0010.0011.03
RibosePentose Metabolism<0.001<0.0011.97
Cofactors and Vitamins
ThreonateAscorbate and Aldarate Metabolism<0.0010.002−0.98
Alpha-CEHC SulfateTocopherol Metabolism<0.0010.010−2.03
PyridoxalVitamin B6 Metabolism0.0020.028−0.79
Lipids
2-hydroxydecanoateFatty Acid, Monohydroxy0.0010.0231.48
1-arachidonoyl-GPA (20:4)Lysophospholipid<0.0010.0081.02
1-oleoyl-GPA (18:1)<0.001<0.0012.27
1-palmitoyl-GPA (16:0)<0.001<0.0012.38
Heptanoate (7:0)Medium Chain Fatty Acid0.0030.0460.80
CholinePhospholipid Metabolism<0.0010.0010.62
1-(1-enyl-palmitoyl)-2-oleoyl-GPE (P-16:0/18:1)Plasmalogen<0.0010.0060.81
1-(1-enyl-stearoyl)-2-oleoyl-GPE (P-18:0/18:1)<0.0010.0110.92
Sphingomyelin (d18:1/25:0, d19:0/24:1, d20:1/23:0, d19:1/24:0)Sphingolipid Metabolism0.0010.020−1.02
Sphingomyelin (d18:1/14:0, d16:1/16:0)0.0010.0250.65
Nucleotides
XanthinePurine Metabolism, (Hypo)Xanthine/Inosine-containing<0.0010.0011.53
Adenosine 5′-monophosphatePurine Metabolism, Adenine-containing<0.0010.001−2.89
Adenosine<0.0010.004−2.45
DihydroorotatePyrimidine Metabolism, Orotate-containing<0.0010.001−1.17
UracilPyrimidine Metabolism, Uracil-containing<0.0010.0081.33
Peptides
4-hydroxyphenylacetylglutamineAcetylated Peptides<0.0010.0011.93
Phenylacetylthreonine0.0020.0281.96
4-hydroxyphenylacetylglycine0.0000.0012.10
LeucylglutamineDipeptide0.0010.0241.90
Xenobiotics
Tartronate (hydroxymalonate)Bacterial/Fungal<0.0010.0080.94
Perfluorooctanesulfonic AcidChemical0.0020.034−2.11
S-(3-hydroxypropyl)mercapturic Acid0.0010.0260.76
Hydroquinone SulfateDrug–Topical Agents0.0020.028−1.86
QuinateFood Component/Plant<0.0010.0063.16
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Barko, P.C.; Rubin, S.I.; Swanson, K.S.; McMichael, M.A.; Ridgway, M.D.; Williams, D.A. Untargeted Analysis of Serum Metabolomes in Dogs with Exocrine Pancreatic Insufficiency. Animals 2023, 13, 2313. https://doi.org/10.3390/ani13142313

AMA Style

Barko PC, Rubin SI, Swanson KS, McMichael MA, Ridgway MD, Williams DA. Untargeted Analysis of Serum Metabolomes in Dogs with Exocrine Pancreatic Insufficiency. Animals. 2023; 13(14):2313. https://doi.org/10.3390/ani13142313

Chicago/Turabian Style

Barko, Patrick C., Stanley I. Rubin, Kelly S. Swanson, Maureen A. McMichael, Marcella D. Ridgway, and David A. Williams. 2023. "Untargeted Analysis of Serum Metabolomes in Dogs with Exocrine Pancreatic Insufficiency" Animals 13, no. 14: 2313. https://doi.org/10.3390/ani13142313

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop