Mouse Model - American Chemical Society

ARTICLE pubs.acs.org/jpr

Metabolic Phenotyping of the Crohn’s Disease-like IBD Etiopathology in the TNFΔARE/WT Mouse Model Pia Baur,† Franc-ois-Pierre Martin,‡,* Lisa Gruber,† Nabil Bosco,‡ Viral Brahmbhatt,‡ Sebastiano Collino,‡ Philippe Guy,‡ Ivan Montoliu,‡ Jan Rozman,§ Martin Klingenspor,§ Isabelle Tavazzi,‡ Anita Thorimbert,‡ Serge Rezzi,‡ Sunil Kochhar,‡ Jalil Benyacoub,‡ George Kollias,|| and Dirk Haller*,† †

)

Chair for Biofunctionality, ZIEL
Research Center for Nutrition and Food Science, CDD - Center for Diet and Disease, Technische Universit€at M€unchen, Gregor-Mendel-Strasse 2, 85350 Freising-Weihenstephan, Germany ‡ Nestle Research Center, P.O. Box 44, Vers-chez-les-Blanc, CH-1000 Lausanne 26, Switzerland § Chair for Molecular Nutritional Medicine, Else-Kr€oner Fresenius Center, Technische Universit€at M€unchen, Gregor-Mendel-Strasse 2, 85350 Freising-Weihenstephan, Germany Institute of Immunology, Alexander Fleming Biomedical Sciences Research Center, 34 Al. Fleming Street, Vari, Greece

bS Supporting Information ABSTRACT: The underlying biochemical consequences of inflammatory bowel disease (IBD) on the systemic and gastrointestinal metabolism have not yet been fully elucidated but could help to better understand the disease pathogenesis and to identify tissue-specific markers associated with the different disease stages. Here, we applied a metabonomic approach to monitor metabolic events associated with the gradual development of Crohn's disease (CD)-like ileitis in the TNFΔARE/WT mouse model. Metabolic profiles of different intestinal compartments from the age of 4 up to 24 weeks were generated by combining proton nuclear magnetic resonance (1H NMR) spectroscopy and liquid chromatography
mass spectrometry (LC
MS). From 8 weeks onward, mice developed CD similar to the immune and tissuerelated phenotype of human CD with ileal involvement, including ileal histological abnormalities, reduced fat mass and body weight, as well as hallmarks of malabsorption with higher energy wasting. The metabonomic approach highlighted shifts in the intestinal lipid metabolism concomitant to the histological onset of inflammation. Moreover, the advanced disease status was characterized by a significantly altered metabolism of cholesterol, triglycerides, phospholipids, plasmalogens, and sphingomyelins in the inflamed tissue (ileum) and the adjacent intestinal parts (proximal colon). These results describe different biological processes associated with the disease onset, including modifications of the general cell membrane composition, alteration of energy homeostasis, and finally the generation of inflammatory lipid mediators. Taken together, this provides novel insights into IBD-related alterations of specific lipiddependant processes during inflammatory states. KEYWORDS: metabonomics, inflammatory bowel disease (IBD), Crohn's disease (CD), tumor necrosis factor (TNF), nuclear magnetic resonance (NMR) spectroscopy, LC
MS, chemometrics, lipid metabolism

’ INTRODUCTION The clinically defined and idiopathic forms of inflammatory bowel disease (IBD), encompassing ulcerative colitis (UC) and Crohn's disease (CD), are spontaneously relapsing and immunologically mediated chronic disorders of the gastrointestinal tract.1 The clinical appearance of human IBD is highly heterogeneous across populations.2 CD can generally affect any region of the intestine in a segmental or discontinuous manner, thereby predominantly involving the ileum and colon. The inflammatory process is frequently transmural and associated with intestinal granulomas, strictures, and fistulas. On the other hand, UC-associated pathologies generally start at the rectum and then spread across the entire colon in an uninterrupted pattern, with inflammation being typically restricted to the r 2011 American Chemical Society

mucosa.2 Both manifestations are mediated by common and distinct mechanisms influenced by multiple environmental factors and specific genetic predispositions.2,3 For example, commensal microbiota and bacterial factors,4 loss of barrier function and mucosal homeostasis,5 and metabolic stress6 are well-accepted triggers modulating IBD pathogenesis. The high number of susceptibility genes, which have been identified during the last decades, illustrate the sheer complexity of the underlying pathology.3 Several animal models of human IBD were developed based on the identified gene loci to dissect the molecular processes contributing to the etiology of IBD.7 Received: August 18, 2011 Published: October 26, 2011 5523

dx.doi.org/10.1021/pr2007973 | J. Proteome Res. 2011, 10, 5523–5535

Journal of Proteome Research Most of them are valuable tools for studying many important disease aspects that are difficult to address in humans, such as the pathological mechanisms in early phases and during the evolution of the disease.8 Three main approaches are currently used for mouse models to trigger gut-associated inflammation including chemical induction (DSS, TNBS), immune cell adoptive transfer (naive T-cells), and genetic modification (IL-10 deficiency, TGFβ deficiency, etc.).9 The majority of these models are characterized by the incidence of colitis and, unlike human CD, do not involve the small intestine. Therefore, murine models of chronic ileal inflammation have recently been established, which spontaneously develop human-like CD.10 The tumor necrosis factor (TNF) plays a key role during the pathogenesis of IBD11 and treatment of IBD patients with anti-TNF molecules (e.g., infliximab) are clinically effective in the therapy of UC and CD.12 Kontoyiannis and co-workers13 genetically impaired the regulation of TNF in C57BL/6 mice by deletion of a repeated octanucleotide AU-rich motif in the 30 -untranslated region of the TNF encoding gene, referred as TNFΔARE. This genetic variation resulted in enhanced mRNA stability of TNF and TNFΔARE mice develop a severe CD8+ T cell-dependent ileitis closely resembling the immune and tissue-related phenotype of human CD with ileal involvement and arthritis.13 Mice, heterozygous for the genetic modification (TNFΔARE/WT), are immunologically well characterized and are used as a murine model for ileitis.14 There is increasing awareness, that system biology approaches, including transcriptomics, proteomics, and metabonomics may provide new insights into disease etiology by deciphering processes associated to specific inflammatory phases (acute, chronic, remission, etc.) and individual susceptibility.15 It is well established that metabolite concentrations and their timedependent changes in tissues and organs represent real endpoints of physiological regulatory processes.16 Since metabonomic approaches, including high-resolution proton nuclear magnetic resonance spectroscopy (1H NMR) and liquid chromatography coupled to mass spectrometry (LC
MS), provide a rapid analysis of a wide range of metabolites, they can be used to describe and understand physiological interactions during complex biological processes or disease states.16,17 For example, it is possible to identify different topographical regions of the intestine, characteristic for their structure and function, through specific metabolic profiles,18 as well as pathological changes during IBD.19 In fact, few studies have successfully applied metabonomics as a diagnostic tool to distinguish between UC and CD in human urine20 and fecal extracts.21,22 However, most of the studies describe metabolic phenotypes at an advanced disease state when symptoms can already be diagnosed by endoscopy. Recently, Martin et al.23 monitored metabolic alterations in plasma of IL10
/
mice before and during the development of the disease, which might help to define early IBD biomarkers. Unfortunately, the majority of studies including ours, have focused on UC in different animal models by exploring plasma,23,24 urine,24
27 and biopsies.28,29 Only very few investigations have identified CD-specific metabolic alterations, making the TNFΔARE/WT mouse model an excellent research target in the field of IBD metabonomics. In the present study, we monitored the effects and consequences of inflammatory processes associated with the gradual evolution of CD-like ileitis on the local metabolism taking advantage of TNFΔARE/WT mice. To obtain results

ARTICLE

comparable with previously published data on intestinal tissue samples from different animal models and human IBD patients (e.g., refs 28
32), nontargeted 1H NMR spectroscopy and two different targeted LC
MS-based metabolic profiling techniques were specifically chosen and used within this study. These three measurements were further combined with a histological analysis and measurements of body weight and composition as well as nutritional energy utilization to characterize site specific and systemic metabolic phenotypes at different time points (1, 4, 8, 12, 16, and 24 weeks) during the development of CDlike ileitis.

’ MATERIAL AND METHODS Animal Trial

The animal studies were conducted in accordance with national German law for animal protection. In total 48 heterozygous TNFΔARE/WT mice on C57BL/6 background as well as an equal number of corresponding wild-type (WT) C57BL/6 mice (Charles River) were conventionally raised at constant room temperature (22 ( 2 °C), air humidity (55 ( 5%), and a light/ dark cycle of 12/12 h. Water and conventional chow diet (ssniff R/ M-H, Soest, Germany) were given ad libitum. Groups of 8 animals per genotype (4 male and 4 female) were sacrificed at the age of 1, 4, 8, 12, 16, and 24 weeks. Sample Collection. One week before sacrifice, mice were separated into single cages and food uptake and water consumption was monitored during one week. Fecal samples were collected from individual mice every day, dried at 60 °C for 3 days, and stored airtight at RT until being assessed. Animals were sacrificed by cervical dislocation. For histological characterization, sections of the distal jejunum (dJ) and ileum (dI) and proximal (pC) and distal colon (dC) as well as samples of the subcutaneous, epididymal and mesenteric fat tissue were fixed in 4% neutral buffered formalin. For 1H NMR and LC
MS analysis, two times a 1 cm sample of the different gut sections (dJ, dI, pC, dC) was excised and flushed with 1 mL of phosphate buffered saline using a sterile syringe and then preserved at
80 °C. Body Composition Assessment. For body composition assessment, 4 additional WT and TNFΔARE/WT mice were sacrificed at the time points 12, 16, and 24 weeks, fixed in a stretched position and stored at
80 °C until analysis. Computer tomographic (CT) measurements were performed using the CTScanner LaTheta LCT-100A (Aloka, Japan and Zinsser Analytic, Deutschland). Gross Energy Content. Fecal pellets were manually ground and pressed into pellets and the gross energy content was determined in a 6300 Calorimeter (Parr Instrument Company, Moline, IL). Histological Scoring and Adipocyte Size Measurements. In formalin, fixed tissue samples were embedded in paraffin, cut in 5 μm sections followed by hematoxylin and eosin staining. Histological scoring was performed by blindly assessing the degree of lamina propria mononuclear cell infiltration, crypt hyperplasia, goblet cell depletion and architectural distortion in the different gut sections, resulting in a score from 0 (not inflamed) to 12 (inflamed), as previously described.33 The mean adipocyte size (μm2) was determined in the stained fat sections at a magnification of 200 fold and the size of adipocytes was calculated with the software AxioVision 4.6 (Carl Zeiss AG, Jena, Germany). 5524

dx.doi.org/10.1021/pr2007973 |J. Proteome Res. 2011, 10, 5523–5535

Journal of Proteome Research Cytokine Quantification. Cytokines, including mouse interferon gamma (IFNγ), interleukin 1 beta (IL-1β), interleukin 6 (IL-6), interleukin 10 (IL-10), interleukin 12 p70 (IL-12 p70), keratinocyte derived chemokine (KC) and tumor necrosis factor (TNF), were measured using a mouse pro-inflammatory multiplex kit (Meso Scale Discoveries, Gaithersburg, MD). Assay was carried out according to the manufacturer's manual. Metabolite Profiling

Sample Preparation and 1H NMR Spectroscopic Analysis. Around 5
10 mg of freeze-dried and ground tissue were used to extract hydrophilic and lipophilic metabolites applying an adapted Folch procedure34 as below. Samples were extracted three times with 0.5 mL of a chloroform
methanol solution (2:1, v/v). Combined extracts were washed first with 0.5 mL of water and second with 0.5 mL of water
methanol (1:1, v/v). The upper hydrophilic phases were collected each time and combined together. Lipophilic and hydrophilic fractions were afterward evaporated to dryness under nitrogen flow and freeze-dried, respectively. Hydrophilic fractions were dissolved in 60 μL of a deuteriated phosphate buffer (pH 7.4) containing 1 mM of 3-trimethylsilyl-1-[2,2,3,3,-2H4]-propionate (TSP) as a standard reference (δ = 0.0) and transferred into 1.7 mm NMR tubes. The lipophilic phases were reconstituted in 60 μL of a deuteriated chloroform
methanol solution (2:1, v/v) and transferred into 1.7 mm NMR tubes using octamethylcyclotetrasiloxane (OMS) as standard reference (δ = 0.0). 1H NMR spectra were registered at ambient temperature (300 K) on a Bruker Avance III spectrometer (Bruker Biospin, Rheinstetten, Germany) at 600 MHz using a standard 1D pulse sequence with solvent suppression and a relaxation delay D1 of 4.0 s. For each sample, 128 transients were collected into 65 K data points using a spectral width of 12 kHz and an acquisition time of 2.7 s, resulting in a total recycling time of 6.7 s. (For representative spectra of hydrophilic and lipophilic fraction see Supporting Information.) Acquired 1H NMR spectra were processed using the Topspin software package (version 2.1; Bruker Biospin, Rheinstetten, Germany). Prior to Fourier transformation, FIDs were multiplied by an exponential weighting function corresponding to a line broadening of 0.3 Hz for the hydrophilic and 1 Hz for the lipophilic fractions of tissue extracts. In addition, all spectra were manually phase- and baseline corrected. The chemical shifts were referenced to their respective standards (TSP or OMS) at δ = 0.0. The peak assignment to specific metabolites was achieved using an internal library of compounds and the literature35,36 and confirmed by standard two-dimensional NMR spectroscopy (JRES, TOCSY, HSQC) on selected samples. For statistical analysis all NMR spectra were converted into 22 K data points over the range of δ 0.2
10.0 and imported into the MATLAB software (version 7.0; The MathWorks Inc., Natick, MA) excluding the water residue (water δ = 4.50
5.19) and solvent signals (methanol δ = 3.24
3.27 and δ = 4.25
4.60; chloroform δ = 7.35
7.45). The spectra were normalized to a constant total sum of all intensities within the specified range and auto scaled. Sample Preparation and Biocrates Life Sciences AbsoluteIDQ Kit Analysis. The Biocrates Life Sciences AbsoluteIDQ kit, originally designed for plasma samples, was adapted as previously published37 to be used for intestinal tissue. Therefore samples were prepared as follows: dI tissue (20
30 mg wet weight) of the time points 4, 8, 12, 16, and 24 weeks were homogenized in 300 μL of EDTA (ethylenediaminetetraacetic

ARTICLE

acid; 0.292 mg/mL PBS) and 10 μL of BHT-buffer (butylated hydroxytoluene; 79.2 mg/mL PBS) using the FastPrep 24 system (MP Biomedicals LLC; Illkirch, France). Afterward, well plate preparation and sample application and extraction were carried out according to the manufacturer’s instructions. A final volume of 10 μL of tissue homogenate was loaded onto the provided 96-well plate. Liquid chromatography was realized on a Dionex Ultimate 3000 ultra high pressure liquid chromatography (UHPLC) system (Dionex AG, Olten, Switzerland) coupled to a 3200 Q TRAP mass spectrometer (AB Sciex; Foster City, CA) fitted with a TurboV ion source operating in electrospray ionization (ESI) mode. Sample extracts (20 μL) were injected two times (in positive and negative ESI modes) via direct infusion using a gradient flow rate of 0
2.4 min: 30 μL/min, 2.4
2.8 min: 200 μL/min, 2.9
3 min: 30 μL/min. MS source parameters were set at: desolvation temperature (TEM): 200 °C, high voltage:
4500 V (ESI
), 5500 V (ESI +), curtain (CUR) and nebulizer (GS1 and GS2) gases: nitrogen; 20, 40, and 50 psi; respectively, nitrogen collision gas pressure: 5 mTorr. MS/MS acquisition was realized in scheduled reaction monitoring (SRM) mode with optimized declustering potential values for the 163 metabolites screened in the assay. Raw data files (Analyst software, version 1.5.1; AB Sciex, Foster City, CA) were imported into the provided analysis software MetIQ to calculate metabolite concentrations. List of all detectable metabolites is available from Biocrates Life Sciences, Austria (http://biocrates.com). Sample Preparation and Inflammation Markers Quantification by UPLC
ESI
MS/MS using Isotope Dilution Technique. On the basis of previously published work,38 a method to measure a panel of 63 inflammatory markers was developed in house. DI tissue samples (20
30 mg wet weight) of the time points 4, 8, 12, and 16 weeks were homogenized in 300 μL of EDTA (ethylenediaminetetraacetic acid; 0.292 mg/ mL PBS) and 10 μL of BHT-buffer (butylated hydroxytoluene; 79.2 mg/mL PBS) using the FastPrep 24 system. These time points were chosen based on the remaining available biological material. For each sample a total of 50 μL of the tissue homogenates was mixed with 5 μL of the internal standard solution (0.1 ng/μL, for preparation see Supporting Information). The mixture was acidified by adding 15 μL of citric acid (1N). To precipitate the proteins, a volume of 550 μL of methanol/ ethanol (1:1, v/v) was added and samples were mixed during 15 min at 4 °C before being centrifuged (3500 rpm, 10 min, 4 °C). The organic phase was evaporated to dryness under constant nitrogen flow and the residues were solubilized with 80 μL water, followed by the addition of 20 μL of acetonitrile, before being centrifuged at 3500 rpm for 1 min at 4 °C. Analyses were carried out by liquid chromatography coupled to tandem mass spectrometry (LC
MS/MS). LC was realized on a Dionex Ultimate 3000 ultra pressure liquid chromatography (UPLC) system (Dionex AG, Olten, Switzerland). MS detection was realized on a 5500 Q TRAP mass spectrometer (AB Sciex; Foster City, CA) operating in ESI mode. Gradient chromatographic separation was performed on an Acquity BEH C18 column (2.1  150 mm, 1.7 μm; Waters, Milford, MA). The injection volume was 5 μL and the column was maintained at 50 °C. The mobile phase consisted of water containing 1% acetic acid (eluent A) and acetonitrile (eluent B) at a constant flow rate set at 450 μL/min. Gradient elution started from 20% B with a linear increase to 50% B at 6 min, from 50% to 95% B at 13 min, hold for 3 min at 95% B, before going back to 20% B at 16.1 min and reequilibration of the column for 5525

dx.doi.org/10.1021/pr2007973 |J. Proteome Res. 2011, 10, 5523–5535

Journal of Proteome Research additional 11 min. Analytes were monitored in the scheduled reaction monitoring (SRM) mode provided within the Analyst software (version 1.5.1; AB Sciex, Foster City, CA). All mass transitions and MS source parameters of the internal standard solution are given in Supporting Information. The SRM detection window time was set at 120 s with a target scan time of 0.5 s. Nitrogen was used as curtain and desolvation gas at the respective pressure of CUR: 20, GS1: 70, GS2: 20 (arbitrary unit). Block source temperature was maintained at 600 °C, with the respective voltages: ISV:
4000 V, EP:
10 V, CXP:
5 V. A 15-points calibration curve was realized prior to sample analysis by measuring different dilutions of the standard solution (0
10 ng). Data processing was realized using Analyst software (version 1.5.1; AB Sciex, Foster City, CA). Peak area ratio of each analyte versus its corresponding internal standard or surrogate marker was calculated. The obtained concentration values were then normalized from the tissue amount used. Multivariate Statistical Data Analysis (MVDA)

All MVDA was carried out with the Simca-P+ software  Sweden) and the MATLAB (version 12.0; Umetrics AB, Umea, software package (version 7.0; The Mathworks Inc., Natwick, MA). Principal Component Analysis. First unsupervised Principal Component Analysis (PCA)39 was performed on NMR and LC
MS data to reduce the dimensionality of the initial data sets, to visualize the dominant global metabolic variances, and to detect outliers. In the corresponding PCA plots, data were represented by means of principal component scores, where each point represents an individual metabolic profile (data not shown). Variables, which were influential in the PCs, can be investigated by the PC loadings. Multivariate Curve Resolution
Alternating Least Squares (MCR-ALS). Multivariate Curve Resolution
Alternating Least Squares (MCR
ALS), as previously published by Martin et al.,19 provides a bilinear decomposition of the initial data matrix X in a set of factors, which can be defined as a combination of contributions C (pure concentration profiles) and S (pure spectral profiles). This decomposition can be expressed as: X ¼ CST þ E where X corresponds to the data from the different regions of the gut tract (1H NMR, MS data). Concentration profiles C reflect the changes in the contribution of the different regions to each factor and S is the matrix containing the spectral features that codefine each factor. While data can be structured in different ways before the algorithm application, we opted for a column-wise (CW) augmentation scheme onto a set of 4 matrices corresponding to the 4 sections of the gut tract. According to this ordering, the MCR model is then described as per equation below. 2 3 2 3 2 3 X1 E1 C1 6 7 7¼6 6l 7 7ST þ 6 6l 7 7 l XCW ¼ 6 4 5 4 5 4 5 XI CI EI The application of the ALS algorithm requires the initialization of one set of profiles. Several methods are available with this purpose to initialize C or S. In this case, contribution profiles were initialized by using evolving factor analysis (EFA). The fitting of the model was performed applying non-negativity constraints in both modes, and assuming the nontrilinearity of

ARTICLE

the data set. As MCR is an extremely flexible exploratory method, it is recommendable to have some estimative value on the number of real factors present in the data. Accordingly, PCA using random subset cross-validation was used in parallel to EFA to achieve a suitable estimation of the number of factors. In these conditions, the number of relevant factors for each MCR
ALS model was determined to be 4 or 5 as obtained from the interpretation on EFA results of the CW augmented matrix. Non-negativity constraints were applied by non-negative least-squares to both the contribution and the spectral profiles. Spectral profiles were normalized to have equal length. Stopping criterion of the algorithm was set to a difference of 0.1% between the standard deviations of the residuals of two consecutive iterations. Partial Least Squares (PLS) and Orthogonal PLS (OPLS) Discriminant Analyses (PLS-DA and O-PLS-DA). The supervised methods Partial Least Squares (PLS) and Orthogonal PLS (O-PLS) discriminant analyses (PLS-DA and O-PLSDA) were further applied to maximize the separation between inflamed and noninflamed samples.39 All models were generated using one predictive and one orthogonal component to discriminate between two groups of animals. The model robustness was assessed using the standard 7-fold cross validation method (repeatedly leaving out a seventh of the samples and predicting them back into the model). Also R2X and R2Y values were computed to show how much of the variation in the data set X (NMR or MS data) and in the data set Y (group information) was explained by the model. The validity of the model against overfitting was monitored by computing the cross-validation parameter Q2, which represents the predictability of the models and relates to the statistical significance. Negative or very low values of the Q 2 indicated that no statistically significant differences were observed. For the NMR data, differences between samples in the scores plot were extracted by using the variable coefficients according to a previously published method.39 Variables correlating with the group separation in the MS data were identified by using the S-plot. It visualizes the variable importance (VIP) score, representing the impact of a single metabolite to the group discrimination of the model.40 In addition, Mann
Whitney U tests were performed on representative NMR signals and MS variables to assess significant differences of peak integrals and metabolite concentrations, respectively.

’ RESULTS Ileal Inflammation was Associated with Alteration in Body Fat Composition

Heterozygous TNF ΔARE/WT mice develop two distinct pathologies called chronic inflammatory arthritis and CD-like IBD.13 However, the kinetic progression of ileitis has not precisely been addressed in this model. Hence, histological scoring of distal ileum (dI) sections from WT and TNFΔARE/WT animals aged 1, 4, 8, 12, 16, and 24 weeks revealed significant alterations in the tissue morphology of TNFΔARE/WT mice from 8 weeks onward reaching a plateau at 12 weeks (Figure 1A). The observed scores were associated with structural alterations and leukocyte infiltration into the mucosa, submucosa and muscularis region. In addition, significant body weight differences were observed between WT and TNFΔARE/WT mice at the age of 12 weeks, which were maintained until the age of 24 weeks (Figure 1B). In relation to the decreased body weight gain in 5526

dx.doi.org/10.1021/pr2007973 |J. Proteome Res. 2011, 10, 5523–5535

Journal of Proteome Research

ARTICLE

Figure 1. Histological scoring, body weight, nutritional energy utilization and adipose tissue morphology of WT and TNFΔARE/WT mice. (A) Histological scores [score: 0
12] of the distal ileum of WT and TNFΔARE/WT mice at 1, 4, 8, 12, 16, and 24 weeks of age. (B) Body weight, (C) daily food uptake and (D) energy content in feces over time. (E) Body composition (visceral and subcutaneous fat mass and lean mass) assessed by computerized tomography (CT) at 12, 16, and 24 weeks of age. (F) Adipocyte size of the subcutaneous, epididymal and mesenteric fat tissue. All scores are presented as mean values ( SEM from 8 animals per group (exception: CT-measurements, n = 4). Significant differences were assessed by Mann
Whitney U test and marked as follows: *p < 0.05., **p < 0.01, ***p < 0.001.

TNFΔARE/WT mice, daily food intake and fecal energy loss were monitored. While food intake was significantly reduced in inflamed compared to noninflamed mice only at 24 weeks

(Figure 1C), bomb calorimetric analysis highlighted a significant increased energy loss in the feces of TNFΔARE/WT mice compared to WT mice from 8 weeks onward (Figure 1D). To gain 5527

dx.doi.org/10.1021/pr2007973 |J. Proteome Res. 2011, 10, 5523–5535

Journal of Proteome Research

ARTICLE

Figure 2. Topographical modeling of metabolic relationships in the gastrointestinal tract of WT and TNFΔARE/WT mice. (A) Sampling scheme of the different intestinal segments (distal jejunum, dJ; distal ileum, dI; proximal colon, pC; distal colon, dC) for WT and TNFΔARE/WT mice at 24 weeks of age. (B) Histological scores [score: 0
12] of different gut sections. (C) Cytokine expression level within the dI for interleukin-1 beta (IL-1β), interferon gamma (IFNγ), Interleukin-6 (IL-6), keratinocyte chemoattractant (KC) and tumor necrosis factor (TNF). Global MCR-ALS analysis of (D) WT and (E) TNFΔARE/WT 1H NMR spectra of the lipophilic extract. Median values of the pure concentration profiles of each factor for the four different functional gut compartments are displayed. (D) Integrals of the AUC for the signal of diacylglycerol (DAG; δ = 5.09) and polyunsaturated fatty acids (PUFA I δ = 2.74
2.79; PUFA II δ = 2.79
2.87). All scores are presented as mean values ( SEM from 8 animals per group. Significant differences were assessed by Mann
Whitney U test and marked as follows: *p < 0.05, **p < 0.01, ***p < 0.001.

further insight, body composition was assessed by using computerized tomography (CT) in mice aged 12, 16, and 24 weeks (Figure 1E), since changes in body weight were observed from 12 weeks onward. From this time point on, WT mice showed an increase in the percentage of their subcutaneous fat mass,

whereas the overall visceral fat mass remained constant. TNFΔARE/WT mice displayed a very different body composition pattern, showing no significant relative increase of fat mass over time, and therefore having a significantly lower fat mass than WT mice at all three measured time points. These findings were in 5528

dx.doi.org/10.1021/pr2007973 |J. Proteome Res. 2011, 10, 5523–5535

Journal of Proteome Research

ARTICLE

Table 1. Metabolic Differences in the Lipophilic Fraction of Different Intestinal Compartments (Distal Jejunum, Distal Ileum, Proximal Colon and Distal Colon) in TNFΔARE/WT Compared to WT Mice at the Age of 24 Weeks Detected by 1H NMR Spectroscopya

a NB, All scores are presented as mean values ( SD of normalized NMR data from 8 animals per group. Red color indicates a significant decreased, green color a significant increased metabolite level in TNFΔARE/WT compared to WT animals. Significant differences were assessed by Mann-Whitney U test and marked as follows: *p < 0.05, **p < 0.01, ***p < 0.001. KEY: DAG = diacylglycerides, FA = fatty acids, GPL = glycerophospholipids, PL = phospholipids, PUFA = polyunsaturated fatty acids, TG = triglycerides, UFA = unsaturated fatty acids.

agreement with a reduced fat cell size in the subcutaneous and epididymal fat tissues from 24 weeks old TNFΔARE/WT animals (Figure 1F). The mesenteric adipocytes were also of smaller proportions in different segments of the intestine (jejunum, ileum and colon). 1

H NMR Spectroscopy Highlighted Lipid Metabolism Alterations in the Inflamed Intestinal Parts

Histological scoring of different parts of the intestine (distal jejunum (dJ) and ileum (dI), proximal (pC) and distal colon (dC)) was assessed at the age of 24 weeks to determine the predominant site of the gut associated inflammation (Figure 2A). The most severe inflammatory changes were observed in the dI of the TNFΔARE/WT mice (Figure 2B). Investigation of the pC showed also a significant increase in the histology score in the TNFΔARE/WT samples, which remained lower than in the dI. The dJ and dC scores were similar to WT scores. In addition, cytokine analysis in the dI of 24 weeks old animals using a multiplex biological assay highlighted significantly elevated concentrations of pro-inflammatory cytokines in the TNFΔARE/WT animals, namely interferon gamma (IFNγ) and interleukin 6 (IL-6) (Figure 2C). The detection of increased levels of tumor necrosis factor (TNF) in the dI of TNFΔARE/WT mice validated the impact of the genetic modification. Metabonomic analysis of gut tissue extracts was conducted to provide additional insights into biochemical processes associated with structural and functional changes in the different intestinal compartments in relation to inflammation. Tissue extracts (hydrophilic and lipophilic fractions) of different intestinal parts (dJ, dI, pC and dC) obtained from 24 weeks old WT and

TNFΔARE/WT mice were therefore studied by high-resolution 1H NMR spectroscopy. Resulting metabolic profiles were then analyzed using a multivariate curve resolution (MCR
ALS) analysis as described by Martin et al.19 to model metabolic relationships along the gastrointestinal compartments (dJ, dI, pC, dC). This method was used to capture divergences across the gastrointestinal metabolic space in relation to its structure and function and to provide insights into metabolic deregulations involved in region-specific gastrointestinal disorders. Displaying the median values of the pure MCR
ALS concentration profiles of each MCR
ALS factor for the four different functional gut compartments highlighted region-specific pattern represented in the lipophilic profiles (Figure 2D, E; for factor contribution see Supporting Information). In the WT MCR
ALS model, the first factor captured a metabolic pattern being specific for the dJ, the second factor specific for the dI (Figure 2D). Factor 3 described a basal metabolic feature shared along all investigated gut segments. The fourth factor modeled a metabolic pattern specific to the pC and dC. Similar patterns were captured by applying MCR
ALS analysis to TNFΔARE/WT derived spectra of the organic tissue fraction, resulting as well in a dJ, a dI, a basal and a colon specific factor. Interestingly, the dI specific MCR
ALS factor highlighted a strong relationship between the ileal and colonic biochemical composition, not observable in the WT animals. By overlaying the dI specific spectrotypes from the WT (factor 2) and TNFΔARE/WT (factor 3) model, qualitative differences in the levels of diacylglycerol (DAG) and the composition of polyunsaturated fatty acids (PUFA) were highlighted. To gain semiquantitative information, peak areas in the original spectra 5529

dx.doi.org/10.1021/pr2007973 |J. Proteome Res. 2011, 10, 5523–5535

Journal of Proteome Research

ARTICLE

Table 2. Most Discriminating Metabolites in the Distal Ileum of TNFΔARE/WT Compared to WT Mice at the Age of 4, 8, 12, 16, and 24 Weeks Analyzed by UHPLC
ESI
MS/MSa

a NB, All scores are presented as mean values ( SD from 8 animals per group. Red color indicates a significant decreased, green color a significant increased metabolite level in TNFΔARE/WT compared to WT animals. Significant differences were assessed by Mann-Whitney U test and marked as follows: *p < 0.05, **p < 0.01, ***p < 0.001. KEY: C = acyl-carnitine, PCaa = cholin glycerophospholipid, PCae = cholin glycerophospholipid with an ether bond, SM = sphingomyelin.

were integrated for these two specific metabolite groups and statistical significance was confirmed by Mann
Whitney U test (Figure 2F). Within the PUFA specific signal (δ = 2.74
2.87) an increase for the fraction δ = 2.74
2.79 (PUFA I), containing most-notably the n-3 and n-6 fatty acids C20:3, C20:4, C20:5, and C22:6, was combined with a decrease of the segment δ = 2.79
2.87 (PUFA II), mainly representing the PUFAs C18:2 and C18:3 (peak assignment was confirmed by literature,35,36 spiking experiments and 2D-NMR spectroscopy). Supervised chemometric data analysis using O-PLS-DA was also employed on the hydrophilic and lipophilic fractions of the intestinal extracts to maximize group separation and to model more subtle metabolic changes. The metabonomic analysis of the dJ and dC showed in agreement with the histological scoring no metabolic differences in the hydrophilic and lipophilic extracts. Interestingly, the analysis of the lipophilic extracts from the inflamed tissues, namely dI and pC, showed significant metabolic variations between WT and TNFΔARE/WT mice, but not in the hydrophilic extracts. Afterward, influential metabolites were identified through the analysis of the corresponding coefficients plots. In addition, semiquantitative information was generated by 1 H NMR peak integration, and statistical significance was assessed by Mann
Whitney U test (Table 1). These results described a significant modulation of cholesterol, triglycerides, glycerophospholipids, saturated and unsaturated fatty acids as well as sphingomyelin and plasmalogen metabolism in TNFΔARE/WT mice when compared to WT controls. Altogether these data highlight the main inflammation location in dI and pC where most if not all significant metabolic changes were observed and mirror the histological findings. Inflammation-associated Metabolic Changes in the Distal Ileum Detected by Targeted LC
MS Metabonomics

Metabolic differences related to the development of inflammation in the dI were first monitored by applying a targeted LC
MS metabonomic method on dI tissues of WT and TNFΔARE/WT mice. O-PLS-DA was performed on quantitative

information for 163 metabolites, including amino acids, sugars, acyl-carnitines, sphingolipids, and glycerophospholipids, to maximize the separation of WT and TNFΔARE/WT mice at the age of 4, 8, 12, 16, and 24 weeks in a pairwise manner. Here, models for the 12, 16, and 24 weeks time point displayed high and positive R2X, R2Y and Q2 parameters (Table 2), whereas the most robust model was obtained at 12 weeks. The ten most influential metabolites, based on the computed VIP score, are listed in Table 2. Statistical significances were assessed by Mann
Whitney U test (all significantly regulated metabolites are listed in Supplementary table, Supporting Information). Some of the listed metabolites showed a high variability over time due to variations during the normal aging process as well as interanimal variability. The data show increased concentrations of sphingomyelins as well as decreased levels of acyl-carnitines, especially of medium chain acyl-carnitines (C:10, C:12, C:14OH, C:16-OH), in the dI of inflamed TNFΔARE/WT mice. Within the group of choline containing glycerophospholipids (PCaa/ae) a concentration rise was observed for acyl ether lipids (PCae's), whereas the level of PCs with ester bonds (PCaa's) decreased. For two groups of compounds (C12:1; PCaa C26:0) a representative histogram displaying the concentration development of the metabolites over time is shown in Figure 3A. It highlights in both cases significant differences between the TNFΔARE/WT and WT samples at the time points 12 and 16 weeks. In addition, a targeted LC
MS method was employed to investigate specific inflammatory lipid markers within the dI over time (4, 8, 12, and 16 weeks), including arachidonic acid, isomers of thromboxane, hydroxyeicosatetraenoic (HETE) acid, dihydroxyeicosatrienoic (DiHETrE) acid, hydroxyoctadecadienoic (HODE) acid and prostaglandins. The application of O-PLSDA revealed significant differences from 8 weeks onward between WT and TNFΔARE/WT samples. Table 3 lists the O-PLSDA model parameters as well as all inflammatory markers changing in the dI of TNFΔARE/WT compared to WT mice. 5530

dx.doi.org/10.1021/pr2007973 |J. Proteome Res. 2011, 10, 5523–5535

Journal of Proteome Research

ARTICLE

Figure 3. Metabolic differences during the inflammation development in the distal ileum of TNFΔARE/WT compared to WT mice. (A) Concentration levels (μmol/mg tissue) of 2 representative metabolites (C:12, PCaa C26:0) measured by targeted UHPLC
ESI
MS/MS metabonomic analysis at the time points 4, 8, 12, 16, and 24 weeks. (B) Concentration levels (μg/g tissue) of 2 representative inflammatory markers (arachidonic acid, thromboxane B2) measured by targeted UPLC
ESI
MS/MS method at the time points 4, 8, 12, and 16 weeks. All scores are presented as mean values ( SEM from 8 animals per group. Significant differences were assessed by Mann
Whitney U test and marked as follows: *p < 0.05., **p < 0.01, ***p < 0.001. KEY: C, acyl-carnitine; PCaa, cholin glycerophospholipid.

Significant differences were assessed applying the Mann
Whitney U test. Some of the listed metabolites showed a high variability over time due to variations during the normal aging process as well as interanimal variability. Metabolites generated from the precursor linoleic acid, namely 9- and 13- HODE, were decreased in the TNFΔARE/WT compared to WT mice at the time points 12 and 16 weeks. Also the levels of 9 HETE and DiHeTrE were lower at 12 weeks in the inflamed animals. In addition, TNFΔARE/WT mice showed significantly increased levels of the prostaglandins PGF1 alpha and PGF2 alpha as well as thromboxane B2 compared to WT mice. For the arachidonic acid (AA) and thromboxane B2 (TXB2) a representative histogram displaying the concentration of these two specific inflammatory markers over time is shown in Figure 3B. AA shows significant differences at 12 and 16 weeks with increased levels at 12 weeks and decreased amounts at 16 weeks in WT mice compared to TNFΔARE/WT mice. Concentration of TXB2 was elevated in the TNFΔARE/WT mice from 8 weeks onward.

’ DISCUSSION In the present study we demonstrate that the TNFΔARE/WT mouse model displays similarity with the immune and tissuerelated phenotype of human CD with ileal involvement.

Specifically, it represents ileal histological abnormalities, reduced total body fat mass and weight, as well as hallmarks of malabsorption with higher energy wasting. Therefore, we believe that the TNFΔARE/WT mouse is an appropriate model to study metabolic changes associated with the gradual development of CD-like ileitis. To do so, we carried out a holistic metabonomic characterization of systemic (not shown) and ileal outcomes to get metabolic signatures of disease development across age. Comprehensive analysis of the results revealed three main ongoing biological processes associated with disease onset. As depicted in Figure 4, these include modifications of the general cell membrane composition, alteration of energy supply machinery and finally the generation of lipid inflammatory mediators. All these aspects work together to modify histological features and body composition in TNFΔARE/WT mice. During CD large portions of the small intestinal mucosa are functionally impaired and patients subsequently suffer from malnutrition not exclusively due to lower food uptake. Hence, these nutritional deficiencies could primarily explain modification of body composition and weight loss.41 In the current study, TNFΔARE/WT mice had a lower body weight from 12 weeks onward when compared to WT mice. These observations were 5531

dx.doi.org/10.1021/pr2007973 |J. Proteome Res. 2011, 10, 5523–5535

Journal of Proteome Research

ARTICLE

Table 3. Concentration Levels (μg/g tissue) of Inflammatory Markers in the Distal Ileum of TNFΔARE/WT Compared to WT Mice at the Age of 4, 8, 12, and 16 Weeks Analyzed by UPLC
ESI
MS/MSa

NB, All scores are presented as mean values ( SD from 8 animals per group. Red color indicates a significant decreased, green color a significant increased metabolite level in TNFΔARE/WT compared to WT animals. Significant differences were assessed by Mann-Whitney U test and marked as follows: *p < 0.05, **p < 0.01, ***p < 0.001. KEY: DiHETrE = dihydroxyeicosatrienoic acid, HETE = hydroxyeicosatetraenoic acid, HODE = hydroxyoctadecadienoic, PG = prostaglandin. a

Figure 4. Schematic model of inflammation-driven alterations of lipid metabolism during ileitis.

preceded by an increased fecal energy loss and an increased fecal fat excretion of TNFΔARE/WT mice as assessed by Fourier transforminfrared (FT-IR) spectroscopy of feces (see Supplementary Figure 2C,D, Supporting Information). These findings suggest that TNFΔARE/WT mice parallel at least in part the malabsorption

phenotype observed in CD patients with ileal resection.42 Huybers et al.43 also reported calcium malabsorption and bone alteration in this animal model, providing therefore additional evidence of insufficient nutrient uptake in the intestine of TNFΔARE/WT mice. Besides this, inflamed mice showed a significantly reduced food 5532

dx.doi.org/10.1021/pr2007973 |J. Proteome Res. 2011, 10, 5523–5535

Journal of Proteome Research uptake at the age of 24 weeks. For CD patients a decreased caloric intake, correlating with disease severity, was reported and attributed to several factors, such as anorexia, abdominal pain, nausea, vomiting, and intestinal obstruction.44 Anorexia could have applied in a similar manner for the reduced food uptake of the inflamed animals in the current study. The heterozygous TNFΔARE/WT mouse model spontaneously developed a CD like IBD pathology, predominantly localized in the distal ileum and to a lower degree in the proximal colon, with mucosal and submucosal infiltration of inflammatory cells partially extending into the muscular layers of the bowel wall. Tissue pathology was associated with local expression of pro-inflammatory cytokines (i.e., TNF, IFNγ and IL-6) likely produced by infiltrated immune or epithelial cells. Many of the lipid metabolites detected by 1H NMR spectroscopy, such as cholesterol, plasmalogen, sphingomyelin, GPL, as well as PUFAs, are important cell membrane constituents. We argue that their accumulation could be subsequent to the described massive immune cell infiltration leading to an alteration of the overall lipid membrane composition in the same tissue region. In addition, it is known that inflammatory cells typically contain a high proportion of PUFAs and release them under inflammatory conditions especially arachidonic acid.45 We indeed detect elevated concentrations of PUFA in the inflamed intestinal tissues. Generation and maintenance of the lipid composition of different cellular compartments is an important issue for proper cell function. Quantitative changes of any membrane lipid or microdomain like lipid rafts during ileitis might have a strong impact on cellular fluidity, structure and function. Modifications of the cholesterol46 or PUFA47 content in lipid rafts of intestinal epithelial cells, known to stabilize the tight-junction assembly, can lead to an increased paracellular permeability, a well-known phenomenon in the pathophysiology of CD.5 This mechanism seems to be also regulated by the action of TNF48 and therefore can be exacerbated in TNFΔARE/WT mice. Studies assessing body composition in CD patients previously described a reduction of body weight due to a depletion of fat but not the fat-free mass.49 In the current study, WT mice constantly increased the proportions of their adipose tissue from 12 weeks onward, especially the subcutaneous fat fraction, whereas TNFΔARE/WT mice showed significant proportional reduction of subcutaneous and visceral fat depots over time. As previously suggested in CD patients, these observations may indicated a shift toward lipolysis and fatty acid oxidation.50 In addition, elevated TNF levels in the adipose tissue of the TNFΔARE/WT mice (see Supplementary Figure 2B, Supporting Information), whose induction of lipolytic actions is well reported, may promoted increased lipolysis and fat oxidation in this animal model.51 In contrast to the frequent finding of mesenteric fat hypertrophy in CD patients, called “creeping fat”,52 TNFΔARE/WT mice had smaller adipocytes when compared to WT cells in the mesenteric compartment, but also in the subcutaneous and epididymal fat. Detection of low levels of TGs and modified levels of DAG, the degradation product of TGs, in the intestine as well as reduced levels of TGs in the liver of TNFΔARE/WT mice (see Supplementary Figure 2A, Supporting Information) may suggest that lipid malabsorption, leading to insufficient precursor availability, and increased TNF-driven lipolysis synergistically interact at the epithelial interface. In addition, studies in CD patients showed elevated lipolysis and fatty acid oxidation, rather than the use of glucose as a metabolic fuel.50,51 As we did not observe

ARTICLE

alterations of other energy metabolites (amino acids, lactate, glucose as reported in other IBD models 21,29,30 ) we may hypothesize that under these conditions intestinal epithelial cells mainly rely on beta-oxidation for their energy needs. A decrease of acyl-carnitines (C:10, C:12) was detected by LC
MS from the age of 12 weeks onward, possibly indicating a higher use of these fatty acids for oxidation. As the decrease was mainly observed for medium chain fatty acids, we further speculate that these fatty acids are now predominantly used for beta-oxidation as the long chain fatty acids likely served as substrate for the PUFA and lipid mediator synthesis. PUFAs are not only involved in the maintenance of cellular integrity by altering the fluidity and functional compartmentalization of cell membranes, but are also released as regulators of cell signaling, either by themselves or after conversion to other lipid mediators like eicosanoids. Elevated concentrations of the n-6 PUFA arachidonic acid as well as some eicosanoids were found in the inflamed intestinal mucosa in the acute phase of human CD and are therefore thought to play a key role in the IBD pathogenesis.45 In the current study, an increase of long chain polyunsaturated n-3 and n-6 fatty acids (C20:3 n-6, C20:4 n-6, C20:5 n-3, C22:6 n-3) was combined with a decrease of essential precursors (C18:2 n-6, C18:3 n-3) in inflamed tissue regions of TNFΔARE/WT mice. Similar results were observed in earlier studies where chronic intestinal inflammation caused a similar shift in the mucosal fatty acid profile.31 It has been hypothesized that in active IBD the local PUFA biosynthesis is enhanced, coexisting with increased PUFA utilization related to the inflammatory process.31 Moreover, the hydrolysis of GPL is catalyzed by phospholipases liberating PUFAs, predominantly arachidonic acid, to further initiate the production of eicosanoids to initiate and/or resolve inflammation according to the nature and the timing of eicosanoids formed.32 In our study, a decrease of GPL/PCaa was observed in TNFΔARE/WT mice by 1 H NMR spectroscopy and LC
MS measurements respectively. The expression and activity of the responsible enzyme seems to be increased under inflammation and has been shown to be elevated after TNF stimulation.53 We may hypothesize that increased biosynthesis in combination with elevated release sustain the pool of available PUFAs to fuel the generation of eicosanoids. In studies with human biopsies, leukotriene B4 and prostaglandin E2 are often reported to be elevated in the inflamed tissues of IBD patients.54 Unfortunately, we were not able to detect these two metabolites in our samples due to limited tissue availability. However, we observed significant changes of arachidonic acid as well as pro-inflammatory eicosanoids, such as prostaglandin F1 alpha and F2 alpha, which showed reduced concentration levels in TNFΔARE/WT compared to WT mice at 12 weeks followed by a significant increase by 16 weeks. It has been shown that the level of these metabolites as well as of thromboxane B2 was also elevated during human IBD.54 During the release of PUFAs many other lipid mediators can be formed such as lysophosphatidylcholine (lyso PCa) and diacylglycerol (DAG). Diminished amounts of lyso PCa were observed under inflammation (see Supporting Information) in contrast to increased amounts of DAG. DAG is a key metabolic intermediate in glycerolipid synthesis that also serves as a second messenger in regulating classical and novel protein kinase C enzymes. Besides this, elevated levels of sphingomyelin, which can be hydrolyzed by sphingomyelinases to release ceramides acting as pro- or anti-inflammatory mediators were detectable in the intestine of TNFΔARE/WT mice. A reduction of the enzyme 5533

dx.doi.org/10.1021/pr2007973 |J. Proteome Res. 2011, 10, 5523–5535

Journal of Proteome Research expression and activity, responsible for the sphingomyelin digestion in the intestine, was present in colorectal cancer as well as in patients with chronic colitis.55 In conclusion, we demonstrated the potential of using different metabonomic tools to get additional insights into the molecular mechanisms underlying the development and progression of CD. We observed that the metabolic profile of the quiescent parts (dJ, dC) of the TNFΔARE/WT mice tended to be different from the WT profiles, becoming significantly modified in the inflamed parts (dI, pC), as previously reported in IBD patients.21,29,56 These changes were most prominently reflected in the intestinal lipid metabolism. These results provide novel insights into IBD-related alterations of specific metabolic processes in the tissue during the different inflammatory states and may help to further understand the link between inflammatory processes and metabolic alterations.

’ ASSOCIATED CONTENT

bS

Supporting Information Supplementary materials, figures, and tables. This material is available free of charge via the Internet at http://pubs.acs.org.

’ AUTHOR INFORMATION Corresponding Authors

*Dirk Haller, Ph.D., Professor, Biofunctionality, Technische Universit€at M€unchen, Gregor-Mendel-Str. 2, 85350 FreisingWeihenstephan, Germany. E-mail: [email protected]. Phone +49-8161-71-2026. Fax +49-8161-71-2824 *Francois-Pierre Martin, Ph.D., Nestle Research Center, BioAnalytical Sciences, P. O. Box 44, Vers-chez-les-Blanc, CH-1000 Lausanne 26, Switzerland. Email: Francois-Pierre.Martin@rdls. nestle.com. Phone +41-21-785-8771.

’ ABBREVIATIONS: CD, Crohn's disease; IBD, inflammatory bowel disease; MS, mass spectrometry; MCR-ALS, Multivariate Curve Resolution
Alternating Least Squares; NMR, nuclear magnetic resonance; O-PLS-DA, orthogonal projection to latent structure discriminant analysis; PCA, principal component analysis; PLS-DA, projection to latent structure discriminant analysis; TNF, tumor necrosis factor; UC, ulcerative colitis. ’ REFERENCES (1) Kaser, A.; Zeissig, S.; Blumberg, R. S. Inflammatory bowel disease. Annu. Rev. Immunol. 2010, 28, 573–621. (2) Xavier, R. J.; Podolsky, D. K. Unravelling the pathogenesis of inflammatory bowel disease. Nature 2007, 448 (7152), 427–34. (3) Cho, J. H.; Brant, S. R. Recent insights into the genetics of inflammatory bowel disease. Gastroenterology 2011, 140 (6), 1704–12e2. (4) Clavel, T.; Haller, D. Molecular interactions between bacteria, the epithelium, and the mucosal immune system in the intestinal tract: implications for chronic inflammation. Curr. Issues Intest. Microbiol. 2007, 8 (2), 25–43. (5) McGuckin, M. A.; Eri, R.; Simms, L. A.; Florin, T. H.; RadfordSmith, G. Intestinal barrier dysfunction in inflammatory bowel diseases. Inflamm. Bowel Dis. 2009, 15 (1), 100–13. (6) Shkoda, A.; Ruiz, P. A.; Daniel, H.; Kim, S. C.; Rogler, G.; Sartor, R. B.; Haller, D. Interleukin-10 blocked endoplasmic reticulum stress in intestinal epithelial cells: impact on chronic inflammation. Gastroenterology 2007, 132 (1), 190–207.

ARTICLE

(7) Mizoguchi, A.; Mizoguchi, E. Inflammatory bowel disease, past, present and future: lessons from animal models. J. Gastroenterol. 2008, 43 (1), 1–17. (8) Wirtz, S.; Neurath, M. F. Mouse models of inflammatory bowel disease. Adv. Drug Delivery Rev. 2007, 59 (11), 1073–83. (9) Sartor, R. B. Review article: How relevant to human inflammatory bowel disease are current animal models of intestinal inflammation? Aliment Pharmacol. Ther. 1997, 11 (Suppl 3), 89–96; discussion 96
7. (10) Pizarro, T. T.; Arseneau, K. O.; Bamias, G.; Cominelli, F. Mouse models for the study of Crohn’s disease. Trends Mol. Med. 2003, 9 (5), 218–22. (11) Scaldaferri, F.; Correale, C.; Gasbarrini, A.; Danese, S. Mucosal biomarkers in inflammatory bowel disease: key pathogenic players or disease predictors? World J Gastroenterol. 2010, 16 (21), 2616–25. (12) Cosnes, J. Can we modulate the clinical course of inflammatory bowel diseases by our current treatment strategies? Dig Dis 2009, 27 (4), 516–21 (13) Kontoyiannis, D.; Pasparakis, M.; Pizarro, T. T.; Cominelli, F.; Kollias, G. Impaired on/off regulation of TNF biosynthesis in mice lacking TNF AU-rich elements: implications for joint and gut-associated immunopathologies. Immunity 1999, 10 (3), 387–98. (14) Apostolaki, M.; Manoloukos, M.; Roulis, M.; Wurbel, M. A.; Muller, W.; Papadakis, K. A.; Kontoyiannis, D. L.; Malissen, B.; Kollias, G. Role of beta7 integrin and the chemokine/chemokine receptor pair CCL25/CCR9 in modeled TNF-dependent Crohn’s disease. Gastroenterology 2008, 134 (7), 2025–35. (15) Kussmann, M.; Rezzi, S.; Daniel, H. Profiling techniques in nutrition and health research. Curr. Opin. Biotechnol. 2008, 19 (2), 83–99. (16) Nicholson, J. K.; Lindon, J. C.; Holmes, E. ’Metabonomics’: understanding the metabolic responses of living systems to pathophysiological stimuli via multivariate statistical analysis of biological NMR spectroscopic data. Xenobiotica 1999, 29 (11), 1181–9. (17) Bjerrum, J. T.; Nielsen, O. H.; Wang, Y. L.; Olsen, J. Technology insight: metabonomics in gastroenterology-basic principles and potential clinical applications. Nat. Clin. Pract. Gastroenterol. Hepatol. 2008, 5 (6), 332–43. (18) Martin, F. P.; Wang, Y.; Yap, I. K.; Sprenger, N.; Lindon, J. C.; Rezzi, S.; Kochhar, S.; Holmes, E.; Nicholson, J. K. Topographical variation in murine intestinal metabolic profiles in relation to microbiome speciation and functional ecological activity. J. Proteome Res. 2009, 8 (7), 3464–74. (19) Martin, F. P.; Montoliu, I.; Kochhar, S.; Rezzi, S. Chemometric strategy for modeling metabolic biological space along the gastrointestinal tract and assessing microbial influences. Anal. Chem. 2010, 82 (23), 9803–11. (20) Williams, H. R.; Cox, I. J.; Walker, D. G.; North, B. V.; Patel, V. M.; Marshall, S. E.; Jewell, D. P.; Ghosh, S.; Thomas, H. J.; Teare, J. P.; Jakobovits, S.; Zeki, S.; Welsh, K. I.; Taylor-Robinson, S. D.; Orchard, T. R. Characterization of inflammatory bowel disease with urinary metabolic profiling. Am. J. Gastroenterol. 2009, 104 (6), 1435–44. (21) Marchesi, J. R.; Holmes, E.; Khan, F.; Kochhar, S.; Scanlan, P.; Shanahan, F.; Wilson, I. D.; Wang, Y. Rapid and noninvasive metabonomic characterization of inflammatory bowel disease. J. Proteome Res. 2007, 6 (2), 546–51. (22) Bezabeh, T.; Somorjai, R. L.; Smith, I. C. MR metabolomics of fecal extracts: applications in the study of bowel diseases. Magn. Reson. Chem. 2009, 47 (Suppl 1), S54–61. (23) Martin, F. P.; Rezzi, S.; Philippe, D.; Tornier, L.; Messlik, A.; Holzlwimmer, G.; Baur, P.; Quintanilla-Fend, L.; Loh, G.; Blaut, M.; Blum, S.; Kochhar, S.; Haller, D. Metabolic assessment of gradual development of moderate experimental colitis in IL-10 deficient mice. J. Proteome Res. 2009, 8 (5), 2376–87. (24) Schicho, R.; Nazyrova, A.; Shaykhutdinov, R.; Duggan, G.; Vogel, H. J.; Storr, M. Quantitative metabolomic profiling of serum and urine in DSS-induced ulcerative colitis of mice by (1)H NMR spectroscopy. J. Proteome Res. 2010, 9 (12), 6265–73. (25) Murdoch, T. B.; Fu, H.; MacFarlane, S.; Sydora, B. C.; Fedorak, R. N.; Slupsky, C. M. Urinary metabolic profiles of inflammatory bowel 5534

dx.doi.org/10.1021/pr2007973 |J. Proteome Res. 2011, 10, 5523–5535

Journal of Proteome Research disease in interleukin-10 gene-deficient mice. Anal. Chem. 2008, 80 (14), 5524–31. (26) Lin, H. M.; Barnett, M. P.; Roy, N. C.; Joyce, N. I.; Zhu, S.; Armstrong, K.; Helsby, N. A.; Ferguson, L. R.; Rowan, D. D. Metabolomic analysis identifies inflammatory and noninflammatory metabolic effects of genetic modification in a mouse model of Crohn’s disease. J. Proteome Res. 2010, 9 (4), 1965–75. (27) Otter, D.; Cao, M.; Lin, H. M.; Fraser, K.; Edmunds, S.; Lane, G.; Rowan, D. Identification of urinary biomarkers of colon inflammation in IL10
/
mice using Short-Column LCMS metabolomics. J. Biomed. Biotechnol. 2011, 974701. (28) Balasubramanian, K.; Kumar, S.; Singh, R. R.; Sharma, U.; Ahuja, V.; Makharia, G. K.; Jagannathan, N. R. Metabolism of the colonic mucosa in patients with inflammatory bowel diseases: an in vitro proton magnetic resonance spectroscopy study. Magn. Reson. Imaging 2009, 27 (1), 79–86. (29) Sharma, U.; Singh, R. R.; Ahuja, V.; Makharia, G. K.; Jagannathan, N. R. Similarity in the metabolic profile in macroscopically involved and un-involved colonic mucosa in patients with inflammatory bowel disease: an in vitro proton ((1)H) MR spectroscopy study. Magn. Reson. Imaging 2010, 28 (7), 1022–9. (30) Shiomi, Y.; Nishiumi, S.; Ooi, M.; Hatano, N.; Shinohara, M.; Yoshie, T.; Kondo, Y.; Furumatsu, K.; Shiomi, H.; Kutsumi, H.; Azuma, T.; Yoshida, M. GCMS-based metabolomic study in mice with colitis induced by dextran sulfate sodium. Inflamm. Bowel Dis. 2011, 17 (11), 2261–74. (31) Buhner, S.; Nagel, E.; Korber, J.; Vogelsang, H.; Linn, T.; Pichlmayr, R. Ileal and colonic fatty acid profiles in patients with active Crohn’s disease. Gut 1994, 35 (10), 1424–8. (32) Morita, H.; Nakanishi, K.; Dohi, T.; Yasugi, E.; Oshima, M. Phospholipid turnover in the inflamed intestinal mucosa: arachidonic acid-rich phosphatidyl/plasmenyl-ethanolamine in the mucosa in inflammatory bowel disease. J. Gastroenterol. 1999, 34 (1), 46–53. (33) Katakura, K.; Lee, J.; Rachmilewitz, D.; Li, G.; Eckmann, L.; Raz, E. Toll-like receptor 9-induced type I IFN protects mice from experimental colitis. J. Clin. Invest. 2005, 115 (3), 695–702. (34) Folch, J.; Lees, M.; Sloane Stanley, G. H. A simple method for the isolation and purification of total lipides from animal tissues. J. Biol. Chem. 1957, 226 (1), 497–509. (35) Siddiqui, N.; Sim, J.; Silwood, C. J.; Toms, H.; Iles, R. A.; Grootveld, M. Multicomponent analysis of encapsulated marine oil supplements using high-resolution 1H and 13C NMR techniques. J. Lipid Res. 2003, 44 (12), 2406–27. (36) Oostendorp, M.; Engelke, U. F.; Willemsen, M. A.; Wevers, R. A. Diagnosing inborn errors of lipid metabolism with proton nuclear magnetic resonance spectroscopy. Clin. Chem. 2006, 52 (7), 1395–405. (37) R€omisch-Margl, W.; Prehn, C.; Bogumil, R. R€ohring, C.; Suhre, K.; Adamski, J. Procedure for tissue sample preparation and metabolite extraction for high-throughput targeted metabolomics. Metabolomics 2011 (doi:10.1007/s11306-011-0293-4). (38) Deems, R.; Buczynski, M. W.; Bowers-Gentry, R.; Harkewicz, R.; Dennis, E. A. Detection and quantitation of eicosanoids via high performance liquid chromatography-electrospray ionization-mass spectrometry. Methods Enzymol. 2007, 432, 59–82. (39) Trygg, J. O2-PLS for qualitative and quantitative analysis in multivariate calibration. J. Chemom. 2002, 17, 53–64. (40) Wiklund, S.; Johansson, E.; Sjostrom, L.; Mellerowicz, E. J.; Edlund, U.; Shockcor, J. P.; Gottfries, J.; Moritz, T.; Trygg, J. Visualization of GC/TOF-MS-based metabolomics data for identification of biochemically interesting compounds using OPLS class models. Anal. Chem. 2008, 80 (1), 115–22. (41) Fisher, R. L. Wasting in chronic gastrointestinal diseases. J. Nutr. 1999, 129 (1S Suppl), 252S–255S. (42) Filipsson, S.; Hulten, L.; Lindstedt, G. Malabsorption of fat and vitamin B12 before and after intestinal resection for Crohn’s disease. Scand. J. Gastroenterol. 1978, 13 (5), 529–36. (43) Huybers, S.; Apostolaki, M.; van der Eerden, B. C.; Kollias, G.; Naber, T. H.; Bindels, R. J.; Hoenderop, J. G. Murine TNF(DeltaARE)

ARTICLE

Crohn’s disease model displays diminished expression of intestinal Ca2+ transporters. Inflamm. Bowel Dis. 2008, 14 (6), 803–11. (44) Bannerman, E.; Davidson, I.; Conway, C.; Culley, D.; Aldhous, M. C.; Ghosh, S. Altered subjective appetite parameters in Crohn’s disease patients. Clin. Nutr. 2001, 20 (5), 399–405. (45) Calder, P. C. Polyunsaturated fatty acids, inflammatory processes and inflammatory bowel diseases. Mol. Nutr. Food Res. 2008, 52 (8), 885–97. (46) Lambert, D.; O’Neill, C. A.; Padfield, P. J. Depletion of Caco-2 cell cholesterol disrupts barrier function by altering the detergent solubility and distribution of specific tight-junction proteins. Biochem. J. 2005, 387 (Pt 2), 553–60. (47) Innis, S. M.; Dai, C.; Wu, X.; Buchan, A.M.; Jacobson, K. Perinatal lipid nutrition alters early intestinal development and programs the response to experimental colitis in young adult rats. Am. J. Physiol. Gastrointest. Liver Physiol. 2010, 299 (6), G1376–85 . (48) Li, Q.; Zhang, Q.; Wang, M.; Zhao, S.; Ma, J.; Luo, N.; Li, N.; Li, Y.; Xu, G.; Li, J. Interferon-gamma and tumor necrosis factor-alpha disrupt epithelial barrier function by altering lipid composition in membrane microdomains of tight junction. Clin. Immunol. 2008, 126 (1), 67–80. (49) Rocha, R.; Santana, G. O.; Almeida, N.; Lyra, A. C. Analysis of fat and muscle mass in patients with inflammatory bowel disease during remission and active phase. Br. J. Nutr. 2009, 101 (5), 676–9. (50) Al-Jaouni, R.; Hebuterne, X.; Pouget, I.; Rampal, P. Energy metabolism and substrate oxidation in patients with Crohn’s disease. Nutrition 2000, 16 (3), 173–8. (51) Mingrone, G.; Greco, A. V.; Benedetti, G.; Capristo, E.; Semeraro, R.; Zoli, G.; Gasbarrini, G. Increased resting lipid oxidation in Crohn’s disease. Dig. Dis. Sci. 1996, 41 (1), 72–6. (52) Peyrin-Biroulet, L.; Chamaillard, M.; Gonzalez, F.; Beclin, E.; Decourcelle, C.; Antunes, L.; Gay, J.; Neut, C.; Colombel, J. F.; Desreumaux, P. Mesenteric fat in Crohn’s disease: a pathogenetic hallmark or an innocent bystander? Gut 2007, 56 (4), 577–83 (53) Lee, C. W.; Lin, C. C.; Lee, I. T.; Lee, H. C.; Yang, C. M. Activation and induction of cytosolic phospholipase a(2) by TNF-alpha mediated through Nox2, MAPKs, NF-kappa B, and p300 in human tracheal smooth muscle cells. J. Cell Physiol. 2010, 226 (8), 2103–2114. (54) Hawkey, C. J.; Karmeli, F.; Rachmilewitz, D. Imbalance of prostacyclin and thromboxane synthesis in Crohn’s disease. Gut 1983, 24 (10), 881–5. (55) Sjoqvist, U.; Hertervig, E.; Nilsson, A.; Duan, R. D.; Ost, A.; Tribukait, B.; Lofberg, R. Chronic colitis is associated with a reduction of mucosal alkaline sphingomyelinase activity. Inflamm. Bowel Dis. 2002, 8 (4), 258–63. (56) Bjerrum, J. T.; Nielsen, O. H.; Hao, F.; Tang, H.; Nicholson, J. K.; Wang, Y.; Olsen, J. Metabonomics in ulcerative colitis: diagnostics, biomarker identification, and insight into the pathophysiology. J. Proteome Res. 2010, 9 (2), 954–62.

5535

dx.doi.org/10.1021/pr2007973 |J. Proteome Res. 2011, 10, 5523–5535