SEARCH

SEARCH BY CITATION

Keywords:

  • metabolomic;
  • metabonomic;
  • metabolite profiling;
  • Crohn's disease;
  • ulcerative colitis

Abstract

  1. Top of page
  2. Abstract
  3. CURRENT UNDERSTANDING OF IBD
  4. METABOLOMIC ANALYSIS
  5. METABOLOMIC ANALYSIS IN IBD RESEARCH
  6. CONCLUSION
  7. Acknowledgements
  8. REFERENCES

Crohn's disease (CD) and ulcerative colitis (UC) are inflammatory bowel diseases (IBD) attributed to a dysregulated immune response towards intestinal microbiota. Although various susceptibility genes have been identified for CD and UC, the exact disease etiology is unclear and complicated by the influence of environmental factors. Metabolomic analysis enables high sample throughput measurements of multiple metabolites in biological samples. The use of metabolomic analysis in medical sciences has revealed metabolite perturbations associated with diseases. This article provides a summary of the current understanding of IBD, and describes potential applications and previous metabolomic analysis in IBD research to understand IBD pathogenesis and improve IBD therapy. Inflamm Bowel Dis 2011

Metabolomic analysis refers to the comprehensive study of the many small molecule metabolites present in biological samples, using technologies which enable the analysis of multiple metabolites with high sample throughput.1–7 The use of metabolomic analyses to study inflammatory bowel diseases (IBD) could potentially address key issues of IBD, which are unknown disease etiology and a requirement for improvement of therapy. This review provides a summary of the current understanding of IBD and explains the techniques of metabolomic analysis. The possible applications of metabolomic analysis to understand IBD disease etiology and improve treatment are discussed with supporting examples from other areas of medical research. The limited number of IBD studies that have utilized metabolomic analysis and their contributions to the current knowledge are also described.

CURRENT UNDERSTANDING OF IBD

  1. Top of page
  2. Abstract
  3. CURRENT UNDERSTANDING OF IBD
  4. METABOLOMIC ANALYSIS
  5. METABOLOMIC ANALYSIS IN IBD RESEARCH
  6. CONCLUSION
  7. Acknowledgements
  8. REFERENCES

Diagnosis and assessment

Crohn's disease (CD) and ulcerative colitis (UC) are the main types of IBD. Despite similarities in symptoms and disease location, both are distinct diseases with differences in clinical,8 pathological,8 and immunological features.9 Distinguishing the two can be difficult, such as in the case of colonic CD or indeterminate colitis. The autoantibodies ASCA (anti-Saccharomyces cerevisiae antibodies) and p-ANCA (perinuclear antineutrophil cytoplasmic antibodies), which are associated with CD and UC patients, respectively, have been proposed as diagnostic markers but their poor sensitivity and specificity are not ideal for clinical purposes.10 To date, a reliable test that can distinguish these two diseases does not exist. Biochemical markers commonly used for monitoring disease activity and response to treatment, such as the erythrocyte sedimentation rate and C-reactive protein, do not always correlate with disease activity and are not specific to IBD inflammation.11, 12 Thus, new biomarkers are still sought after.

Disease Etiology

The higher incidence and prevalence of IBD in developed Western countries compared with Asian countries13 suggests the role of environmental factors in IBD etiology. Furthermore, IBD incidence rates are increasing in Asian countries with industrialization and Westernization of lifestyles.14 Despite epidemiological differences, a specific food antigen causing IBD has not been discovered, although various types of foods have been suggested based on relationships between high incidence rates and food intakes.15–18 It is possible that dietary factors may influence IBD pathogenesis through the modulation of intestinal microbiota composition or immune responses.

The role of intestinal microbiota in IBD pathogenesis is indicated by the requirement of microbial exposure for intestinal inflammation to develop in animal models of IBD,19 and differences in the intestinal microbiota composition of CD and UC patients compared with healthy controls.20 An imbalance in bacteria composition (dysbiosis) may be causing a pathological response through the loss of beneficial immunomodulating bacteria or through an increase of opportunistic bacteria that are normally not pathogenic in a healthy host. A reduction in the phylum Firmicutes was generally observed in CD,20 with Faecalibacterium prausnitzii occurring at lower levels in patients with postoperative recurrence of CD.21 However, the role of dysbiosis in IBD etiology should be interpreted with caution, as dysbiosis could also be a result of chronic intestinal inflammation favoring the growth of certain bacteria. Studies on mice models showed that intestinal microbiota can be altered by pathogen-induced, chemically induced, or genetically predisposed intestinal inflammation.22

The increased risk of CD or UC in first-degree relatives of affected individuals indicate a genetic cause in the disease.23 Heritability is stronger in CD than UC, as disease concordance rates among monozygotic twins is higher for CD.24 Genome-wide association studies have identified over 30 susceptibility gene loci for CD and UC.25, 26 Approximately half of these susceptibility loci are shared by both diseases, indicating the same mechanistic pathways.26 Although several susceptibility gene loci have been identified for CD and UC, these loci only account for 10% of the overall variance in disease risk,25, 26 indicating that there are still unidentified susceptibility genes and/or other nongenetic factors involved in disease pathogenesis.

The immune-related functions of the various IBD susceptibility genes27 and the development of intestinal inflammation in animal models with defective gastrointestinal immune response28, 29 indicate that IBD may be caused by a dysregulated gastrointestinal immune response towards intestinal microbiota. Possible mechanisms by which defects in the gastrointestinal immune response give rise to IBD include increased penetration of enteric antigens,30 an excessive31–33 or insufficient34 immune response, or defective T-cell tolerance towards intestinal microbiota.35, 36

Overall, IBD appears to be a complex, multifactorial disease with unknown etiology and a requirement for improvement of therapy. Metabolomic analysis may help address these key issues. The following section explains the technique of metabolomic analysis and presents a brief summary of its methodologies. Potential applications of metabolomic analysis to understand IBD etiology and improve therapy are discussed later, with supporting examples.

METABOLOMIC ANALYSIS

  1. Top of page
  2. Abstract
  3. CURRENT UNDERSTANDING OF IBD
  4. METABOLOMIC ANALYSIS
  5. METABOLOMIC ANALYSIS IN IBD RESEARCH
  6. CONCLUSION
  7. Acknowledgements
  8. REFERENCES

Metabolomic analysis refers to the comprehensive study of the many small molecule metabolites present in biological samples.1–7 Typical features of metabolomic analysis are the analysis of multiple metabolites with high sample throughput. Fiehn6 refers to metabolomic analysis as “a comprehensive and quantitative analysis of the metabolome,” with the metabolome defined as the full suite of metabolites synthesized by a biological system. However, it is currently impossible to measure the complete metabolome due to the dynamic range of metabolite concentrations, chemical diversity of metabolites, and limitations of analytical techniques. A more appropriate definition of metabolomic analysis proposed by Villas-Boas et al37 is the “characterisation of metabolic phenotypes under specific sets of conditions.” This definition is similar to “metabonomic analysis,” defined by Nicholson et al7 as “the quantitative measurement of the time-related multi-parametric metabolic response of living systems to pathophysiological stimuli or genetic modification.” Nowadays “metabolomic” and “metabonomic” are generally considered to be synonymous, as both definitions share common features which are the measurement of large numbers of metabolites and high sample throughput. This review considers both “metabolomic” and “metabonomic” to have the same meaning which encompasses all the definitions mentioned, but the term “metabolomic” will be used throughout.

Methodologies

There are extensive reviews on the methodologies and analytical technologies for metabolomic analysis.38–43 Overall, metabolomic analysis requires analytical technologies that are capable of detecting and quantitating the large number of metabolites in a biological sample. Sample preparation is ideally minimal to enable the quantitation of as many metabolites as possible. However, no single analytical technique can measure all the metabolites that are present in a sample due to the diverse chemical properties of metabolites and the limitations of each analytical technique. Commonly used analytical platforms are mass spectrometry (MS)38 and nuclear magnetic resonance spectroscopy (NMR).39

MS measures the molecular mass of chemical compounds and their fragmentation products. The technique involves the ionization and fragmentation of compounds into smaller molecules which are then quantitated by a detector. The ionization and fragmentation of compounds are generally influenced by their chemical structure; thus, unique spectrums of mass fragments are generated for most compounds and can be used for compound identification. Gas chromatography (GC) and liquid chromatography (LC) are commonly used interfaces for MS in metabolomic analyses to enable physical separation of metabolite prior to MS analysis to enhance the detection of individual analytes.38, 40 GC-MS has the advantage of metabolite identification through matching of mass spectra to mass spectral libraries, as the use of electron impact ionization of a standard voltage produces reproducible mass spectra. In contrast, the use of mass spectral libraries in LCMS is limited by the poor reproducibility of mass spectra between different laboratories and instrument manufacturers due to different ionization techniques and ion suppression by sample matrices. However, LC-MS offers faster sample throughput and versatility, as a wider range of metabolites can be analyzed. GC-MS is limited to the analysis of volatile compounds, where nonvolatile metabolites require chemical derivatization to increase their volatility.

NMR measures the magnetic resonance of nuclei in molecules. The nuclei resonance frequency is influenced by the number and nature of surrounding nuclei; thus, the frequency and pattern of resonance provides structural information. Quantitation and identification of metabolites can be performed without prior separation of compounds in the sample, as specific metabolites give rise to unique resonances. However, not all metabolites can be individually identified if their signals overlap, which can be a problem with complex samples consisting of many metabolite species. Signals from metabolites present at lower concentrations may be overshadowed by signals from more abundant metabolites.44

The advantages and limitations of MS versus NMR in metabolomic analysis are summarized in Table 1. Quantitation and reproducibility of NMR are generally higher than MS.44 NMR analysis is nondestructive to the samples, unlike MS, thus NMR samples can be reused for other analysis. However, NMR is biased towards the detection of high abundance metabolites, whereas MS has a higher sensitivity for metabolites that can be easily ionized. Coupling of MS to chromatographic instruments enables MS to detect a wider range of metabolites through separation and quantitation of individual metabolites, whereas NMR is restricted by specific resonances and cannot always resolve overlapping signals easily.

Table 1. Summary of Advantages and Limitations of Mass Spectrometry Versus NMR in Metabolomic Analysis
 Mass spectrometryNMR
Quantitation & reproducibilityLowerHigher
SensitivityHigherLower
Detection rangeWiderNarrower
• couple with chromatographic separation (e.g. LC, GC)• bias towards high abundance metabolites
• restricted to specific resonances
• overlapping signals not easily resolved
SamplesCannot be reusedMay be nondestructive

Data Analysis

Metabolomic data consists of measurements of the concentrations of multiple metabolites, which collectively represent the metabolite profile of the sample or organism. The large volumes of raw data generated from measuring multiple chemical signals in several biological replicates of samples need to be transformed into an appropriate data form that can be statistically analyzed for relevant information. Data preprocessing involves extracting the relevant chemical signals from the raw data45–48 and normalizing for factors that could bias statistical analysis of the data, such as variation arising from analytical processing or biological characteristics of the samples (e.g., urine concentration).49–52

Multivariate statistical techniques are required to determine which metabolites differ between sample classes, and the use of such techniques in metabolomic analysis has been extensively reviewed elsewhere.53–55 Examples of commonly used multivariate discriminant tests in metabolomic studies are partial-least squares discriminant analysis (PLS-DA) and principal components analysis (PCA). PLS-DA is considered a supervised classification method, as the samples are designated into their classes for comparison. In contrast, PCA may be considered an unsupervised classification method, by analyzing the variation in the data without a priori designation of samples into their classes. Thus, PCA is useful for identifying unexpected variations and trends in the data.

METABOLOMIC ANALYSIS IN IBD RESEARCH

  1. Top of page
  2. Abstract
  3. CURRENT UNDERSTANDING OF IBD
  4. METABOLOMIC ANALYSIS
  5. METABOLOMIC ANALYSIS IN IBD RESEARCH
  6. CONCLUSION
  7. Acknowledgements
  8. REFERENCES

The usefulness of metabolites as indicators of pathological conditions was recognized before the emergence of technological capabilities for metabolomic analysis. Classic examples are the elevation of glucose levels in diabetes and organic acidurias in inborn errors of metabolism.56 Metabolomic analysis now offers the advantage of a more comprehensive analysis of metabolites, as large numbers of metabolites in tissues and biofluids can be measured to produce metabolite profiles that are specific for the phenotypes. The possible applications of metabolomic analysis in IBD research with supporting examples from other research areas, and the contributions to current knowledge from IBD studies utilizing metabolomic analysis to date (Table 2) are discussed in the following sections.

Table 2. IBD Studies with Metabolomic Analyses
Disease or animal modelSampleMethodMetabolite differences between disease and control (sorted alphabetically)Studies
Upregulated in diseaseDownregulated in disease
  • *

    Confirmed with authentic standards.

UCColon mucosal tissue (intact), colonocyte extracts, urine, plasma lymphocytesNMRColon tissue: ascorbate, aspartate, glutamate, glutamine, glutathione, taurineColon tissue: betaine, glycerophosphocholine, lipid, myo-inositolBjerrum et al. (2010)79
Colonocytes: choline, glycerophosphocholine, myo-inositol
CD, UCUrineNMRCD: formate, glycine, glycolate, guanidoacetate, methylhistidineCD: 4-cresol sulfate, citrate, hippurateWilliams et al. (2009)60
UC: citrate, glycine, glycolate, guanidoacetate, methylhistidineUC: hippurate, trimethyllysine
CDFecal extractsFTMS(Z)/4/hydroxyphenyl-acetaldehyde-oxime, 3-(4-hydroxyphenyl)lactate, 4-hydroxyphenyl-acetylglycine, arachidonic acid, chenodeoxyglycocholate, dopaquinone, glycocholate, linoleic acid, octadecatrienoic acid, oleic acid, palmitic acid, phenylalanine, stearic acid, taurocholate, trihydroxy-6beta-cholanate, tryptophan, tyrosine2,3-dinor-8-iso-prostaglandin F2alpha, prostaglandin E2alpha, prostaglandin F1alpha, prostaglandin F2alphaJansson et al. (2009)84
CD, UCColon mucosal tissue extractsNMRCD: glucose, glycerophosphorylcholineCD: alanine, choline, formate, glutamine/glutamate, isoleucine/leucine/ valine, lactate, myoinositol, succinateBalasubramanian et al. (2009)85
UC: arginine, glucose, glycerophosphorylcholine, lysineUC: alanine, choline, formate, glutamine/glutamate, isoleucine/leucine/ valine, lactate, myoinositol, succinate
CD, UCFecal extractsNMRCD: alanine, glycerol, isoleucine, leucine, lysine, valineCD: acetate, butyrate, methylamine, trimethylamineMarchesi et al. (2007)61
UC: glutamate, lysineUC: methylamine, trimethylamine
CD, UCColon mucosal tissue (intact)NMRCD & UC (direction of differences unspecified): amino acids (nonspecific), choline, creatine, lipids (nonspecific), lysineBezabeh et al. (2001)78
IL10−/− miceUrineGC-MS5-aminovaleric acid,* fucose,* uracil,* xanthurenic acid* Lin et al. (2009, 2010)63, 64
IL10−/− micePlasmaNMRAlanine, arginine, choline in phospholipids, citrate, fumarate, HDL, isoleucine, lactate, LDL, phenylalanine, polyunsaturated lipids, pyruvateCreatinine, dimethylglycine, glucose, leucine, methionine, trimethylamine, tyrosine, VLDLMartin et al. (2009)86
IL10−/− miceUrineNMR3-indoxylsulfate, butyrate, dimethylamine, fucose, phenylacetylglycine, trimethylamine, trimethylamine-N-oxide, valine2-oxoglutarate, citrate, fumarate, N-isovaleroylglycine, succinateMurdoch et al. (2008)65
DSS-induced colitis, miceSerumLC-MS1-stearoyl-sn- glycero-3-phosphorylcholine*1-oleoyl-sn-glycero-3-phosphorylcholine,* 1-linoleoyl-sn-glycero-3-phosphorylcholine*Chen et al. (2008)62
Carrageenan-induced colitis, ratsColon mucosal tissueNMRCreatine, lipids (non-specific), phosphocholine Varma et al. (2007)87

Elucidating the Disease Mechanism

The metabolite profile of an organism is part of its phenotype, which is linked to genotype; genes code for proteins which perform enzymatic processing to produce metabolites representing the phenotype. Therefore, genetic mutations affecting the enzymatic function of the proteins will alter the metabolite profile of the organisms. These metabolite changes may be detected by metabolomic analysis and may provide insights into disease pathogenesis. For example, Dang et al57 used LC-MS to profile over 100 metabolites in extracts of glioblastoma cells with the isocitrate dehydrogenase 1 (IDH1) mutation R132H that is commonly found in human brain cancer. Wildtype IDH1 enzyme converts isocitrate into α-ketoglutarate. The investigators found that glioblastoma cells with the IDH1 mutation had higher levels of 2-hydroxyglutarate, leading to the discovery that the R132H mutation resulted in a new ability of the enzyme to convert α-ketoglutarate into 2-hydroxyglutarate. Excess levels of 2-hydroxyglutarate arising from the R132H mutation may be contributing to the formation and malignant progression of gliomas, as individuals that accumulate 2-hydroxyglutarate in their brains due to 2-hydroxyglutarate dehydrogenase deficiencies have an increased risk of brain tumors.57 This study by Dang et al demonstrates that comparisons of metabolite concentrations between mutant and wildtype states can reveal insights into disease mechanism. To date, the use of metabolomic analysis to study the metabolic effects of the various genetic mutations of IBD in human patients has not been reported.

Although the metabolite profile of an organism is part of its phenotype, which is linked to genotype, phenotype is also influenced by external factors such as nutrition, pathogens, and commensal bacteria. Xenobiotics arising from digestion of food and intestinal microbiota contribute to the host's metabolite profile. Therefore, studying the metabolite profile of an organism can also provide information on the influence of environmental factors. For example, Holmes et al,58 who performed NMR metabolomic analyses on urine samples from 4630 individuals from East Asian and Western populations. The urinary metabolite profiles could discriminate between populations according to their contrasting diets and blood pressure. Metabolites associated with the blood pressure of individuals were identified, such as alanine, formate, and hippuric acid, which reflect diet and intestinal microbial activity. Further information regarding the use of metabolomic analysis in epidemiological studies is provided by Bictash et al.59

The use of metabolomic analyses to study the influence of intestinal microbial metabolism on IBD has been demonstrated by comparisons of the metabolite profiles of IBD patients and healthy controls. Using NMR analysis, Williams et al60 profiled the urinary levels of microbial metabolites and found that urine samples from CD patients have higher levels of formate and lower levels of hippurate and 4-cresol sulfate compared to those of controls or UC patients, whereas no significant differences were observed for trimethylamine-oxide or dimethylamine levels. Based on perturbations in bacterial species reported for IBD, the investigators suggested that the reduced hippurate and 4-cresol sulfate levels may be related to a reduction in Clostridia spp., whereas higher formate levels may be related to an increase in Enterobacteriacea.60

Marchesi et al61 performed NMR metabolomic analysis on fecal extracts from CD, UC, and healthy individuals and found that fecal samples from CD and UC patients have decreased levels of butyrate, acetate, methylamine, and trimethylamine compared to those of healthy controls. The investigators attributed the depletion of butyrate and acetate to a reduction of Clostridium coccoides and C. leptum, as these bacteria groups were reported to be reduced in IBD patients compared to healthy controls and are mainly responsible for the production of short chain fatty acids.

These studies by Williams et al and Marchesi et al demonstrate that profiling the metabolite changes arising from dysbiosis may help understand the influence of intestinal microbial metabolism on the host in IBD.

Animal Models

The use of animal models in IBD research has helped to understand how defects in immunity may give rise to intestinal inflammation. To date, only a few studies have reported the use of metabolomic analysis on animal models of IBD (Table 2).

LC-MS analysis performed by Chen et al62 on the serum of mice with dextran sulfate sodium (DSS)-induced colitis revealed that DSS-treated mice have higher levels of stearoyl lysophosphatidiylcholine and lower levels of oleoyl lysophosphatidiylcholine. The perturbations in the levels of these lipids led to the discovery that these levels were caused by DSS inhibition of stearoyl-coA desaturase 1 (SCD1)-mediated olecic acid biogenesis in the liver, resulting in exacerbation of proinflammatory responses. The findings of the study implied that SCD1 and associated lipids could be potential drug targets for treating IBD.

Lin et al63, 64 and Murdoch et al65 performed GC-MS and NMR metabolomic analyses, respectively, on interleukin-10-deficient (IL10−/−) mice, which develop Crohn's-like intestinal inflammation unless raised in germ-free conditions.19, 28 Both investigators found increased urinary levels of fucose in IL10−/− mice, which was associated with the early stages of intestinal inflammation. Increased urinary fucose levels have been observed in liver diseases such as cirrhosis and liver cancer.66, 67 Fucosylation of membrane proteins is important in leukocyte trafficking68 and IL10 is involved in regulation of the migration of intestinal T-cells to the liver.69 Extraintestinal manifestations of CD involving the liver, such as primary sclerosing cholangitis, have been attributed to aberrant trafficking of immune cells from the intestine to the liver.70 Therefore, perhaps fucosylation of immune cell recognition and signaling proteins regulated by IL10 are involved in the development of CD-associated liver conditions.

Lin et al also discovered that urinary xanthurenic acid levels were elevated in IL10−/− mice during the early stage of intestinal inflammation, leading to the discovery that the levels of the xanthurenic acid precursors, kynurenine, and 3-hydroxykynurenine were elevated in plasma.63 Induction of immune tolerance in T cells is attributed to the action of kynurenine metabolites produced by indoleamine 2,3-dioxygenase (IDO)-mediated catabolism of tryptophan by dendritic cells. Lin et al63 suggested that kynurenine metabolites are involved in regulating IL10 expression in the induction of immune tolerance towards intestinal microbiota.

The findings from metabolomic analyses of animal models of IBD have provided new insights into potential IBD disease mechanisms. These techniques should be extended to other animal models of IBD to enhance knowledge of the disease process.

Diagnosis and Monitoring

Metabolite profiles produced from metabolomic analysis can be used to discriminate between disease and nondisease states, as demonstrated by Sreekumar et al,71 who used LC-MS to profile 1126 metabolites in 262 clinical samples of tissue, plasma, or urine that were related to prostate cancer. The metabolite profiles produced from these samples could discriminate between benign prostate, clinically localized prostate cancer, and metastatic disease. Sarcosine was identified as a potentially important metabolic intermediary of cancer cell invasion, as its levels were elevated during prostate cancer progression to metastasis, and it was able to induce an invasive phenotype in benign prostate epithelial cells.

However, the use of metabolite differences between disease and healthy states as clinical biomarkers for disease diagnosis requires extensive validation to ensure that the metabolite biomarkers can diagnose the disease accurately. For example, Brindle et al72 reported the identification of an NMR signature of coronary heart disease in human serum from 36 patients and 30 controls, where the signature could correctly classify 93% of the samples. However, two other larger studies with at least 95 diseased patients per study and also using NMR analysis of serum failed to find a predictor with an accuracy of more than 80%.73, 74 Possible factors affecting the accuracy of the predictor include data overfitting, gender effect, and medication effect.

Correct diagnosis of CD and UC would assist in disease management, as each disease has different treatment approaches. However, a metabolite profile specific for CD or UC may not exist, as both diseases are not single gene mutation disorders. The discovery of various disease susceptibility genes has raised the suggestion that IBD may consist of several disease subtypes sharing the same clinical features.75 Thus, different genetic variants of CD and UC may give rise to different disease mechanisms, resulting in different metabolite profiles. Discrimination of these metabolite profiles may be possible by combining metabolomic analysis with genome-wide association studies. For example, Illig et al76, 77 linked genetic variants of metabolism-related genes to metabolic traits represented by metabolite concentration ratios, where the gene variants were detected from genome-wide association studies.

Nevertheless, the potential use of metabolomic analysis to identify unique metabolite profiles of CD and UC was demonstrated on colonic tissue and urine samples of IBD patients and non-IBD controls. Bezabeh et al78 used an optimal region selection algorithm to classify NMR spectra of colon mucosal biopsies from 31 CD, 45 UC, and 25 controls. CD was distinguished from UC with an accuracy of 98.6%. IBD was distinguished from normal controls with an accuracy of 97.9%. 82% of noninflamed tissues from IBD patients were classified unambiguously, of which some were classified as inflamed, suggesting that these normal-looking tissue may actually be in a preclinical disease stages.

Williams et al60 showed that PLS-DA modeling of NMR spectra of urine samples from 68 CD, 60 UC, and 60 controls could distinguish the cohorts, with sensitivities of 75%–100%, specificities of 75%–100%, positive predictive values of 62%–100%, and negative predictive values of 71%–100% depending on the validation method. In contrast, PLS-DA modeling of NMR spectra of urine, colon mucosal biopsies, or blood lymphocytes from 61 UC (32 active, 29 inactive) and 25 controls by Bjerrum et al79 could only strongly differentiate active UC from control using the NMR spectra of intact colon mucosal tissue or extracts from colonocytes. No metabolic differences were found between the urine samples, unlike Williams et al, which may be due to differences in NMR data preprocessing and urine sampling conditions.80 Similar to Bezabeh et al, Bjerrum et al found that the colonic tissues of some inactive UC (20%) had similar profiles to active UC, suggesting that the NMR metabolomic analysis could detect early pathogenic changes.

The use of metabolomic analysis to identify biomarkers of inflammation for monitoring disease activity in IBD has been demonstrated in the IL10−/− mouse model by Lin et al.63 Lin et al compared GCMS urinary metabolite profiles of IL10−/− mice and wildtype housed in either conventional or specific pathogen-free (SPF) conditions. Metabolomic comparisons of both genotypes in either housing conditions discriminated inflammatory and noninflammatory metabolite effects of genetic modification, as inflammation in SPF IL10−/− mice is less severe than conventional IL10−/− mice. Fifteen metabolites were likely to be associated with intestinal inflammation, which included xanthurenic acid, fucose, and 5-aminovaleric acid.

Overall, these studies demonstrate that metabolomic analysis can identify specific metabolite profiles of IBD that can potentially be used for disease diagnosis and monitoring. However, the possible effects of various factors, such as genetic variation, dietary factors, medication, and sample collection have to be taken into account when designing studies to identify accurate metabolite biomarkers.

Personalizing Medication

Azathiopurine and mercaptopurine are thiopurine immunosuppressants frequently used in the treatment of IBD. Although these thiopurines are effective in inducing and maintaining long-term remission for CD and UC, both drugs are not effective in one-third of patients and cause adverse effects in up to one-fifth of patients due to interindividual differences in thiopurine metabolism.81 Thiopurine metabolism is complex due to the involvement of multiple enzymes.81 Monitoring the relative levels of toxic and therapeutic thiopurine metabolites in the blood or measuring the activity of thiopurine S-methyltransferase (TPMT), a thiopurine metabolizing enzyme, are carried out to assess therapeutic response and myelotoxicity.82 However, conflicting results have been reported for therapeutic drug monitoring, and myelotoxicity can still occur in the presence of normal TPMT activity; thus, blood count monitoring remains standard practice for myelotoxicity.82

Assessment of the therapeutic response of IBD patients to thiopurines or other drugs could possibly be improved by using metabolomic analysis to identify metabolite profiles that can discriminate between responders and nonresponders. Genetic polymorphisms in thiopurine metabolizing enzymes that are associated with therapeutic response to thiopurine have been identified, such as for TPMT.81 Genetic factors in therapeutic response to thiopurine may be reflected in the individual's metabolite profile, which could be used to predict responders and nonresponders. Clayton et al83 showed that the urinary metabolite profiles of rats can predict their response towards drugs prior to dosing. Predose metabolite profiles which were generated by NMR analysis of urine discriminated between responders and nonresponders to galactosamine hydrochloride.83 Similarly, the predose urinary metabolite profiles of rats could predict the amount of paracetamol metabolites excreted after dosing, and the profiles could be associated with the severity of the drug-induced liver damage.83

Metabolite profiles for assessment of therapeutic response of IBD patients not only have to consist of drug metabolites, but may also include endogenous metabolites that are affected by the drug treatment. In addition, the use of urinary metabolite profiles to monitor drug response would be more convenient than the use of blood samples, as multiple urine samples can be obtained noninvasively.

CONCLUSION

  1. Top of page
  2. Abstract
  3. CURRENT UNDERSTANDING OF IBD
  4. METABOLOMIC ANALYSIS
  5. METABOLOMIC ANALYSIS IN IBD RESEARCH
  6. CONCLUSION
  7. Acknowledgements
  8. REFERENCES

Although to date there are only a few studies utilizing metabolomic analysis in IBD research, the results from these studies demonstrate how high-throughput analysis of multiple metabolites to identify metabolic perturbations in IBD may provide new insights into IBD pathogenesis and improve therapy. Metabolite differences reported for IBD patients and controls vary between studies, which may be attributed to differences in sample type, research approach (targeted or nontargeted analysis), and the lack of confirmation of metabolite identity with authentic standards. Further research is required to validate these metabolite differences and improve the accuracy of metabolite markers for IBD clinical applications. Finally, the study of metabolite differences in IBD patients should be extended to other analytical techniques besides NMR to enable the discovery of additional metabolite differences, as no single analytical platform can completely analyze the metabolome.

Acknowledgements

  1. Top of page
  2. Abstract
  3. CURRENT UNDERSTANDING OF IBD
  4. METABOLOMIC ANALYSIS
  5. METABOLOMIC ANALYSIS IN IBD RESEARCH
  6. CONCLUSION
  7. Acknowledgements
  8. REFERENCES

Nutrigenomics New Zealand is a collaboration between AgResearch Limited, the New Zealand Institute for Plant & Food Research Limited, and the University of Auckland.

REFERENCES

  1. Top of page
  2. Abstract
  3. CURRENT UNDERSTANDING OF IBD
  4. METABOLOMIC ANALYSIS
  5. METABOLOMIC ANALYSIS IN IBD RESEARCH
  6. CONCLUSION
  7. Acknowledgements
  8. REFERENCES