Skip to main content

Table 1 Key microbial metabolomic studies of the past decade

From: Microbial metabolism of dietary components to bioactive metabolites: opportunities for new therapeutic interventions

Studies identifying major gut microbial metabolites
Aim of study Population Results Reference
Identify metabolites modulated by gut microbiota in various tissue and fluids Conventional versus germ-free C3H/HeJ mice •Higher bile acid levels in gut of germ-free mice.
•Higher phosphocholine and glycine in liver of germ-free mice.
•Hippurate and 5-aminovalerate reduced in germ-free animals.
•Higher levels of betaine, choline and myo-inositol in kidney of germ-free mice.
[4]
Determine effect of antibiotic treatment on metabolome Normal versus vancomycin-treated female NMRI mice •Vancomycin reduced urinary levels of hippurate, phenylacetylglycine, taurine, TMA and TMAO, and increased urinary levels of α-ketoisovalerate, n-butyrate, creatinine, guanidoacetic acid and glycine.
•Vancomycin reduced fecal levels of uracil, amino acids, SCFAs and urinary phenylacetylglycine and hippurate.
[5]
Identify metabolites derived from the gut microbiota Conventional versus germ-free Swiss Webster mice •Metabolites highly enriched or only present in conventional mice include indole derivatives (such as indoxyl sulfate and IPA), phenyl/benzoate derivatives. (hippurate, p-cresol), and conjugated fatty acids. [6]
Identify serum metabolites derived from gut microbiota Conventional versus germ-free Swiss Webster mice •Increased serum metabolites related to energy metabolism (pyruvic acid, citric acid, fumaric acid, malic acid) in conventional compared to germ-free mice. [7]
Determine effect of antibiotics on metabolite production Normal versus penicillin- and streptomycin-treated Han–Wistar rats •Antibiotics reduced urinary excretions of hippurate, phenylpropionic acid, phenylacetylglycine and indoxyl-sulfate, and elevated urinary excretions of taurine, glycine, citrate, 2-oxoglutarate and fumarate.
•Antibiotics reduced fecal SCFA.
[8]
Identify fecal and urinary metabolites derived from the gut microbiota Normal versus imipenem/cilastatin Wistar rats •Antibiotic treatment altered 202 urinary and 223 fecal metabolites.
•Major classes reduced by antibiotics include SCFAs, phenyl/benzoates (for example, p-hydroxyphenylacetate, m-hydroxyphenylacetate, hydroxycinnamic acid, phynylvalerate, p-aminobenzoate and hippurate), and indole-containing substances (indoxyl sulfate, indole-acetate, indole-carboxylate and indole-acetaldehyde), and urobilin. Antibiotic treatment increased tryptophan and tryptamine in feces.
[9]
Compare metabolomes of human versus humanized and conventionally raised mice. Germ-free versus humanized versus conventional Swiss Webstermice •Metabolome of humanized mice was more similar to metabolome of human donors than to metabolome of conventional mice, with more differences in feces than urine.
•Humanized mice had higher fecal levels of tryptamine and indoxyl glucuronide, and lower levels of trisaccharide, creatine and creatinine than conventional mice.
[10]
Studies examining microbial metabolites enriched or depleted in disease states
Disease Aim of study Population Results Reference
Metabolic disorders and CVD Identify metabolites associated with fatty liver disease Disease-susceptible (129S6) versus disease-resistant (BALBc) mice •Increased urinary dimethylamine, TMA, TMAO, formate and hippurate in 129S6 mice on HFD.
•Decreased plasma phosphatidylcholine seen in 129S6 mice likely due to microbial conversion to TMA.
[11]
Identify metabolites associated with obesity Fecal transplantation from ob/ob, ob/+, +/+ C57BL/J mice to germ-free mice •Recipients of fecal transplant reciprocate phenotype of donor. Cecal levels of acetate and butyrate increased in obese mice. [12]
Identify urinary metabolites associated with obesity Lean versus obese Zucker rat •Obese mice have higher urinary creatinine, TMAO, hippurate and acetate. [13]
Identify metabolites associated with obesity Healthy versus obese insulin-resistant male humans •Increased microbiota-derived hippurate acid, trigonelline, 2-hydroxyisobutyrate and xanthine was seen in the obese microbiota [14]
Identify metabolites that predict CVD Human subjects with CVD •Three metabolites of dietary phosphatidylcholine (choline, TMAO, betaine) predict risk for CVD.
•Studies in mice confirmed critical role for dietary choline and gut flora in TMAO production and CVD.
[15]
Identify metabolites elevated in mice highly susceptible to diet-induced obesity. C57J versus C57N mice •C57N more susceptible to diet-induced obesity than C57J.
•In cecum, C57N have decreased taurine-conjugated bile acids, bile acid sulfates, enterolactone and enterodiol; altered arachidonate metabolites and increased urobilins.
•In liver, C57N have increased taurine-conjugated bile acids, fatty acids and urobilins.
[16]
Identify urinary metabolite associated with human adiposity Human subjects from INTERMAP study (n = 2324) •Urinary metabolites associated with increased BMI included N-acetyl neuraminate, TMA, PCS, succinate, citrate ethanolamine. [17]
Determine effects of bariatric surgery on metabolome Severely obese human subjects undergoing bariatric surgery •Bariatric surgery reversed most metabolites associated with obesity such as increased aromatic and branched-chain amino acids, pyruvate, citrate, formate, methanol and isopropanol. [18]
Determine effect of prebiotics in maternal diet on offspring adiposity Female Sprague Dawley rats fed high-fat/sucrose diet with and without 10% oligofructose •Addition of 10% oligofructose to diet normalizes body weight in diet-induced obese dams and inhibited adiposity in offspring.
•Microbiota composition of offspring similar to dams.
•Diet-induced obese dams have increased SCFAs, glycine, betaine, 2- and 3-hydroxybutyrate, cytidine, o-phosphocholine, formate, acetone and reduced levels or carnitine, methanol, amino acid, lactate and O-phosphorylcholine
•Subsequent addition of 10% oligofructose reduced O-phosphorylcholine, acetone, cytidine and 3-hydroxybutyrate, and increased propionate, urea, myo-inositol, isobutyrate, alanine, methionine, ornithine and proline.
[19]
Inflammatory bowel disorders Identify metabolites associated with Crohn’s disease Human twin pairs •In feces, twins with Crohn’s disease have increased fecal levels of hydroxyphenylacetylglycine, tyrosine, tryptophan, glycocholate, fatty acids and phenylalanine metabolites [20]
Identify metabolites associated with IBS Human subjects with IBS versus healthy •In feces, individuals with IBS have increased bile acid and decreased branched-chain fatty acids. Trends of increased taurine and cadaverine in ulcerative colitis.
•No change detected in SCFAs and amino acids.
[21]
Identify metabolites specific to Crohn's disease, ulcerative colitis or pouchitis Diseased versus healthy human subjects •Medium-chain fatty acids and some protein fermentation metabolites decreased in Crohn’s disease, ulcerative colitis and pouchitis.
•Hexanoate inversely correlated with Crohn’s disease.
•Styrene positively correlated with ulcerative colitis.
[22]
Develop simplified metabolomics approach to discriminate ulcerative colitis from Crohn’s disease Human •A single analytical platform based on reverse phase UHPLG-Orbitrap HRMS provided sufficient coverage to discriminate between ulcerative colitis and Crohn’s disease in fecal samples. [30]
CKD Determine effect of resistant starch on the gut metabolome in CKD Sprague Dawley rat with adenine-induced CKD fed high-fiber versus no additional fiber •High-fiber-resistant starch diet improved kidney function and ameliorated CKD.
•High-fiber-resistant starch decreased urinary indoxyl sulfate and p-cresol.
[23]
C. difficile infections To determine effect of antibiotics on C. difficile infection C57BL/6 mice infected with C. difficile given antibiotics versus no antibiotics •Antibiotics decrease secondary bile acids, glucose, free fatty acids and dipeptides while primary bile acids and sugars increase.
•Concluded that C. difficile exploits metabolites such as taurocholate or carbon sources for germination and growth.
[24]
To determine how bile acids impact C. difficile dynamics C57BL/6 mice infected with C. difficile given antibiotics versus no antibiotics •Susceptibility to C. difficile occurred only after antibiotic treatment, and was accompanied by a loss of secondary bile acids.
•Physiological concentrations of secondary bile acids inhibited C. difficile spore germination and growth.
[25]
To analyze fecal metabolome in C. difficile infection Human subjects with C. difficile versus healthy given antibiotics •In feces, subjects with C. difficile have decreased fecal cholesterol and increased fecal coprostanol.
63 microbes associated with increased coprostanol levels identified.
[26]
Neurological or behavior disorders To identify a pattern of metabolic perturbance in ASD Children with ASD versus healthy controls •82 metabolites were altered between ASD and controls.
•In ASD children, levels of amino acids (glycine, serine, threonine, alanine, histidine, taurine) and antioxidants such as carnosine were lower.
[27]
Determine if microbiota play a role in development of ASD Maternal immune activation model of ASD •Maternal immune activation treatment altered 8 % of all serum metabolites detected, with EPS most increased.
•Administration of B. fragilis normalized behavior and EPS levels.
[28]
Determine effects of antibiotics on cognition C57BL/6N mice given antibiotics versus no antibiotics •Antibiotic treatment impaired novel object recognition, but not spatial learning and memory.
•Antibiotic treatment reduced colon levels of SCFAs, TMA, adenine and uracil.
•Antibiotic treatment increased plasma levels of corticosterone and phospholipids, and reduced plasma levels of lysophospholipid and p-cresyl sulfate, TMAO, deoxycholate and chenodeoxycholate.
•Antibiotic treatment altered brain-derived neurotrophic factor, NMDA receptor subunit 2B, serotonin transporter and neuropeptide Y system.
[29]
Studies using metabolites as predictive biomarkers of physiological response to intervention
Aim of study Population Results Reference
To create a computational platform that predicts response to dietary intervention Obese human subjects •The CASINO (community and systems-level interactive optimization) toolbox was able to predict and quantitatively describe altered fecal and serum SFCA and amino acid levels in response to diet intervention. [31]
To develop a machine-learning algorithm that predicts postprandial glycemic response Healthy human subjects •High interpersonal variability in postprandial glycemic response
•Microbial metabolites key variables in algorithm that accurately predicts personalized responses to real life meals
[32]
  1. ASD autism spectrum disorder, BMI body mass index, CKD chronic kidney disease, CVD cardiovascular disease, EPS 4-ethylphenylsulfate, HFD high-fat diet, IBS irritable bowel syndrome, IPA indole-3-propionate, PCS para-cresyl sulfate, SCFAs short-chain fatty acids, TMA trimethylamine, TMAO trimethylamine N-oxide