Comparison of the Metabolomic Profiles of Irritable Bowel Syndrome Patients With Ulcerative Colitis Patients and Healthy Controls

New Insights Into Pathophysiology and Potential Biomarkers

Ammar Hassanzadeh Keshteli; Karen L. Madsen; Rupasri Mandal; Guy E. Boeckxstaens; Premysl Bercik; Giada De Palma; David E. Reed; David Wishart; Stephen Vanner; Levinus A. Dieleman


Aliment Pharmacol Ther. 2019;49(6):723-732. 

In This Article


In this study we demonstrate that the urinary metabolome differs significantly between IBS patients and healthy controls and also between UC patients in clinical remission and IBS patients. We also identified some metabolites in urine that correlated with severity of IBS symptoms. Together, these findings demonstrate proof-of-principle that urinary metabolic profiles have the potential to be useful in the diagnosis of IBS as well as furthering an understanding of disease pathogenesis.

Many factors including abnormalities in gastrointestinal motility, visceral sensation, brain-gut interaction, psychosocial distress, gut immune activation, intestinal permeability and intestinal and colonic microbiome have been suggested to play a role in the pathogenesis of IBS.[1] However, currently there are no clear IBS-related diagnostic biomarkers making the diagnosis rely solely on the identification of related symptoms and the exclusion of other organic conditions.[1] In addition, many IBS patients may present with IBD-like symptoms, which may require further invasive and expensive investigations including colonoscopy.[8] Therefore, the development of non-invasive biomarkers may provide more cost-effective tools to screen for IBS in suspected cases.

Metabolomics has increasingly been applied for identification of mechanistic, diagnostic and/or prognostic metabolites in patients with digestive diseases including inflammatory bowel disease.[9] Not surprisingly, these studies have identified significant abnormalities in lipid and amino acid metabolism, energy-related metabolism and membrane metabolites in patients with IBD. However, to date very limited work has been performed on metabolomic profiling of IBS patients.[12]

In the present study, we have demonstrated for the first time that urinary metabolites can discriminate between IBS patients and UC patients in clinical remission. Previously, 1H-NMR metabolite profiling of faecal extracts was shown to separate IBS from UC patients with decreased choline in IBS being an important discriminator.[23] Some,[24] but not all,[25] studies have found that IBS-D subjects could also be differentiated from UC patients with active disease using faecal volatile organic metabolites. Here we demonstrate that IBS patients have higher urinary levels of amino acids such as histidine, lysine and glutamine, but lower levels of proline and glutamic acid in comparison to UC patients. Earlier studies comparing IBD patients with healthy controls or IBS patients also demonstrated altered levels of urinary and/or plasma amino acids.[26–28] In the current study, we also found that UC patients had higher lactic acid and ethylmalonic acid than IBS patients. Although increased lactate in urine due to the presence of inflammation in colonic mucosa was previously shown in animal models of IBD,[29] the UC patients in this study were all in clinical remission based on partial Mayo scores. However, this does not preclude the possibility of colonic sub-inflammation in UC patients influencing epithelial metabolism and resulting in increased lactate levels. Methylmalonic acid is a malonic acid derivative and is a vital intermediate in the metabolism of fat and protein. Considering the high value of AUC (0.99) for the separation between IBS and UC patients, after being confirmed in future studies, new laboratory assays could be developed.

Metabolomic profiling has also been used to identify biomarkers capable of discriminating between IBS and healthy controls. However, previous studies mostly focused on faecal,[5,11] breath[10] or plasma[30] samples. In the present study, using metabolomic profiling of urine samples, we found that IBS patients had lower PC ae C38:6 and dopamine but higher hexanoylcarnitine (C6 (C4:1-DC)) compared with healthy controls. PC ae C38:6 is a glycerophospholipid involved in several immune signaling pathways and has been associated with fruits and vegetables intake.[31] C6 (C4:1-DC) is an acylcarnitine with pro-inflammatory properties and reflects fatty acid oxidation.[32] It has also been reported that sleep disturbances can increase acylcarnitine levels.[32,33] Therefore, it can be speculated that higher prevalence of sleep disorders in IBS patients[34] and reduced dietary intake of fruits and vegetables in this cohort[35] may have contributed to the observed alterations in C6 (C4:1-DC) and PC ae C38:6 respectively. It should also be noted that the observed negative correlation between C6 (C4:1-DC) and Bacteriodales may indicate a role for microbial changes in IBS patients that should be explored in future studies.

Interestingly, we also found that IBS patients had lower urinary dopamine than healthy controls. Dopamine is a neurotransmitter belonging to the catecholamine family that can modulate interactions between the brain and the enteric nervous system.[36] Decreased urinary dopamine levels have been related to anxiety and/or depression.[37,38] It is well-established that IBS patients have a higher prevalence of anxiety and depression[39] and our results suggest that dopamine may be involved in this process. It should also be noted that about one-fourth of the IBS patients in this study were on anti-depressant medications which may alter serotonin or dopamine metabolism.[40] In future studies, with larger sample sizes, the analyses should be adjusted for use of anti-depressant medication as an important confounding variable.

We also found that IBS patients had lower urinary p-hydroxybenzoic acid and 2-methylsuccinic acid compared with healthy controls. P-hydroxybenzoic acid is a phenolic derivative of benzoic acid. Significant amounts of benzoic acid have been found in most berries. In addition, benzoic acid is a byproduct of phenylalanine metabolism in bacteria and is produced when gut microbes process polyphenols from plant sources.[41] Recently, we showed that improvement in IBS symptoms after following a low fermentable oligosaccharides, disaccharides and monosaccharides and polyols (FODMAP) diet was associated with a relative increase in urinary p-hydroxybenzoic acid.[14] We also showed that p-hydroxybenzoic acid level was correlated with Lactococcus levels.[14] Therefore, decreased level of this metabolite in our IBS patients may be due to dietary intake and gut microbial composition making it a potential target for future dietary and microbiota-targeted therapies in these patients. 2-methylsuccinic acid is carboxylic acid related to tricarboxylic acid cycle. Although we found a negative correlation between levels of this metabolite and Erysipelotrichales, its role in IBS requires to be explored in future studies.

Based on patients' self-reported stool consistency, IBS patients are classified into different subtypes. It has been suggested that multifactorial pathophysiology of disease may differ between IBS subtypes.[42] Differences in serum-derived melatonin/tryptophan[43] or serotonin[44] levels between IBS-D and IBS-C patients have been reported previously. In the present study, we aimed to investigate the underlying pathophysiological mechanisms of IBS subtypes through a metabolomic approach. Although we did not find a distinctive metabolic profile in IBS-M and IBS-D patients, we found that 3-methyladipic acid and suberic acid had significantly higher levels in IBS-M and IBS-D cases respectively. 3-Methyladipic acid and suberic acid are medium chain fatty acids which have dietary sources. Their role in the pathophysiology of IBS subtypes is not known yet and therefore requires further investigations.

Metabolomic profiles involve both host and microbial-produced metabolites and can be significantly impacted by dietary intake and alterations in gut microbial composition. Some studies have shown differences in gut microbial composition in UC and IBS patients,[45] and it is possible that the differences in urinary metabolic profiles we observed could be related to altered gut microbiota. Indeed, gut microbiota are involved in the production of a range of neurotransmitters such as dopamine[46] and it has been shown that metabolic activity of the gut microbiota may also influence the fatty acid composition of different host tissues.[47] In the present study, we also showed that levels of some metabolites in urine were correlated with abundance of bacteria in stool samples. Therefore, differences in gut microbial composition between IBS and UC patients, as well as differences between IBS subtypes[48] could have contributed to the altered metabolic profiles.

In the present study, we identified several urinary metabolites to have a positive correlation with IBS-SSS scores as an indicator of IBS severity. In particular, histamine levels were related to the severity of abdominal pain, while aspartic acid, methylmalonic acid, PC aa C38:4 and PC ae C36:2 were related to the severity of abdominal distention. Histamine produced by mast cells, at least in part can play an important role in the pathogenesis of IBS through increasing visceral hypersensitivity, abnormal intestinal barrier function, motility and secretion.[49] It has recently been reported that visceral hypersensitivity, symptoms and abdominal pain in IBS patients were reduced following administration of a histamine receptor H1 antagonist.[50] In addition, improvement in IBS severity following a low FODMAP diet was associated with reduction of urinary histamine.[14] These exciting findings suggest that histamine is a potential biomarker for assessment of IBS severity as well as a potential therapeutic target as part of a low FODMAP diet. In another study, IBS patients with high levels of short chain fatty acids as colonic bacteria-derived fermentation products presented with significantly worse gastrointestinal symptoms.[11] The identification of biomarkers which correlate with disease severity would be very useful as markers of efficacy of different therapeutic interventions in IBS patients.

Limitations of our study include the relatively small sample sizes and significant differences in age and gender between the three groups of participants. However, we addressed these limitations by using robust multivariate analysis while adjusting for age and gender. In addition, no assessment of lifestyle factors such as differences in specific dietary intake which may have affected the observed differences of urinary metabolites between the three groups of subjects is another weakness of our study. However, none of the patients were on a specific diet such as a gluten free, paleolithic or low FODMAP diet. It should also be noted that the urinary metabolome may also be affected by different medications (eg anti-depressant medications that may interfere with dopamine or serotonin metabolism). Therefore, some of the observed differences in the metabolome between the groups could be explained by differences in the medications or supplements that our participants were taking and all of these factors need to be taken into account in future studies. Recruiting few IBS-C and IBS-U patients was another limitation of the present study. In addition, definition of UC remission in our study was based on partial Mayo scoring, not endoscopic assessment. Considering the values of FCP, it can be speculated that not all UC patients in this study were in endoscopic remission and some of them had subclinical inflammation which could have affected their gut microbial composition and urinary metabolome. Furthermore, we could not validate findings from previous metabolomic studies. However, while only urine samples were available for metabolomic assays in the present study in all three groups of participants, findings from previous studies were driven by metabolomic measurements in other biological samples such as breath, stool or serum. Although previous studies that applied metabolomic assays on other biological samples identified some IBS-related metabolites, it has been suggested that using urine samples has some advantages that have resulted in the extensive use of urine in most metabolomic studies. It can easily be collected at multiple time points and in large quantities and needs less complex sample pre-treatment due to lower protein content and lower sample complexity.[51] Therefore, performing further studies focused on metabolic profiling using urinary samples are required.

In conclusion, in this exploratory study we identified specific novel metabolites in urine that could be used for discriminating IBS patients from healthy controls and UC patients in clinical remission. We also found that some urinary metabolites were associated with severity of abdominal pain and distention in IBS patients. Some of these metabolites were likely originating from microbial metabolism, thus suggesting the microbiome as a plausible therapeutic target. Further investigation into these metabolites in larger prospective and clinical trial studies will help to determine the prognostic value of these metabolic profiles and their potential role in pathogenesis and future novel treatments.