Detection of Metabolic Alterations in Non-tumor Gastrointestinal

Jan 21, 2009 - London, Hammersmith Hospital, London W12 0NN, United Kingdom. Received September 18, 2008. In this study, we have used metabolic ...
0 downloads 0 Views 1MB Size
Download PDF
Detection of Metabolic Alterations in Non-tumor Gastrointestinal Tissue of the ApcMin/+ Mouse by 1H MAS NMR Spectroscopy Alexandra Backshall,† Denis Alferez,‡ Friederike Teichert,§ Ian D. Wilson,| Robert W. Wilkinson,| Robert A. Goodlad,‡,⊥ and Hector C. Keun*,† Department of Biomolecular Medicine, Division of Surgery, Oncology, Reproductive Biology and Anaesthetics, Faculty of Medicine, Imperial College London, Sir Alexander Fleming Building, South Kensington, London, United Kingdom, SW7 2AZ, Cancer Research U.K., Lincoln’s Inn Fields, London, United Kingdom, WC2A 3PX, Cancer Biomarkers and Prevention Group, Dept of Cancer Studies and Molecular Medicine, University of Leicester, United Kingdom, AstraZeneca Pharmaceuticals, Alderley Park, Macclesfield, Cheshire, United Kingdom, SK10 4TG, and Department of Histopathology, Division of Investigative Science, Imperial College London, Hammersmith Hospital, London W12 0NN, United Kingdom Received September 18, 2008

In this study, we have used metabolic profiling (metabolomics/metabonomics) via high resolution magic angle spinning (HRMAS) and solution state 1H NMR spectroscopy to characterize small bowel and colon tissue from the ApcMin/+ mouse model of early gastrointestinal (GI) tumorigenesis. Multivariate analysis indicated the presence of metabolic differences between the morphologically normal/non-tumor tissue from ∼10 week-old ApcMin/+ mice and their wild-type litter mates. The metabolic profile of isolated lamina propria and epithelial cells from the same groups could also be discriminated on the basis of genotype. Accounting for systematic variation in individual metabolite levels across different anatomical regions of the lower GI tract, the metabolic phenotype of ApcMin/+ lamina propria tissue was defined by significant increases in the phosphocholine/glycerophosphocholine ratio (PC/GPC, +21%) and decreases in GPC (-25%) and the gut-microbial cometabolite dimethylamine (DMA, -40%) relative to wild type. In the whole tissue, elevated lactate (+15%) and myo-inositol (+19%) levels were detected. As the metabolic changes occurred in non-tumor tissue from animals of very low tumor burden (0.9 confirming our assignments. NMR Data and Statistical Analysis. Data were imported and manipulated in Matlab (Mathworks) using in-house software written and compiled by Dr. T. M. D. Ebbels, Dr. H. C.Keun, Mr. J. T. Pearce, and Dr. O. Cloarec. Spectra were normalized25 (by the median fold change to the median spectrum) and ‘binned‘ in Matlab. Binned data (0.01 ppm per data-point) was exported to SIMCA for multivariate analysis. Partial Least Squares models were validated by examining the distribution of the cross-validated Q2 statistic by random permutation of the class membership26 (Supplemetary Figure 2, Supporting Information). For quantitation of individual metabolites identified as influential to discrimination between the genotypes signals were integrated with local linear baseline correction applied; see Table 2 for the metabolites and their resonance frequencies. The criteria for metabolite selection was correspondence between visual identification of average differences in resonance intensity and spectral position of integral regions with high model loading values. Unless otherwise specified statistical analyses were done using Welch’s t test, which assumes unequal sample variances. Initial visual inspection and PCA of the spectra indicated that for some of the tissue samples there were large lipid-based resonances at δ 0.85-1, 1.2-1.45, 2-2.3, 2.75-2.9, and 5.3-5.5 ppm that were causing separation in the models that interfered with and obscured the information that we wanted to acquire from the small metabolites in the tissue. These lipid-based resonances were attributed to fat deposits on the outer wall of the GI tract. Twelve of the whole tissue samples (n ) 72) were Journal of Proteome Research • Vol. 8, No. 3, 2009 1425

research articles

Backshall et al.

excluded from the analysis for this reason, and additionally, where necessary, these lipid-rich regions of the spectra were excluded from the PLS-DA analyses as they were not related to the small metabolites of interest in the GI tissue. Other resonances that were excluded from the analysis of the whole tissue were those of ethanol at δ 1.13-1.19 and 3.63-3.7; these signals were considered to be a result of residual ethanol from the tools used for the dissection, and so they were excluded in this instance. Immunohistochemistry. Staining for infiltrating macrophages into the areas on the small intestine and colon were conducted using an F4/80 antimouse antibody (Insight Biotech, Wembley, Middlesex, U.K.). Sequential incubation with primary F4/80 antibody was followed with a biotin-conjugated rabbit antirat polyclonal secondary antibody (DakoCytomation Ltd.), and a streptavidin-peroxidase reagent that was developed in DAB. After counterstaining with hematoxylin, the slides were assessed in a quantitative manner. All F4/80 positive cells within 25 high-power fields (×40) were scored and assigned by their staining.27

Results Whole tissue, lamina propria, and the aqueous fraction of epithelial cell extracts from colon, proximal (SB1), middle (SB2), and distal small bowel (SB3) from ApcMin/+ and C57BL/6J (wild type) mouse gastrointestinal (GI) tract were subjected to 1H NMR spectroscopic analysis. Figure 1A shows the aliphatic region of a representative 1H HRMAS NMR CPMG spectrum of the wild type mouse colon, with assignments. Analysis was limited to non-tumor GI tissue due to low numbers of polyps (5 polyps total for the 6 ApcMin/+ mice). There were no significant differences between the weights of the animals or the organs except for the small bowel (see Table 1). Of specific note is that there was no significant difference between the spleen weights of the two genotypes. The spleen can be used as a marker of tumor burden, and is usually enlarged in the Apcmin/+ mouse with a high tumor/polyp count.28 Here we observed no consistent enlargement, which agrees with the low tumor numbers observed in the Apcmin/+ mice in this study. NMR-Detectable Metabolic Differences between GI Regions Are Observed Irrespective of Genotype and Analytical Method. Figure 1B shows the scores plot for a PLS-DA model discriminating the whole tissue samples according to location along the GI tract; the tissue clusters according to GI region irrespective of genotype. The PLS-DA loadings indicated that the most influential separation was between the colon tissue and the small bowel, due to resonances at δ 3.61, 3.92, 3.53, 1.90, 3.02 and 3.25 (higher in colon) and 3.47, 3.48, 3.82 (higher in small bowel). Separation was also observed in the PLS-DA model scores for the epithelial cell extracts, the loadings indicated that the most influential resonance causing separation was that of phosphocholine at 3.22 ppm (not shown). Identification of Metabolic Differences between Morphologically Normal ApcMin/+ and Wild Type Mouse Gastrointestinal Tissue. PLS-DA model loadings (Figure 2A) and visual analysis of the CPMG spectra of the whole tissue highlighted resonances that may be influential in separating the different genotypes: δ 0.89 (lipid), 1.27 (lipid), 3.02 (creatine), 3.2 (choline), 3.22 (phosphocholine), 3.23 (glycerophosphocholine), 3.25 (taurine), 3.42 (taurine), 3.92 (creatine), and 5.3 (lipid). Statistical analysis of the integrals of these resonances indicated that there were significant differences between the levels of some of these metabolites between the ApcMin/+ and wild type. 1426

Journal of Proteome Research • Vol. 8, No. 3, 2009

Figure 1. 1H NMR analysis of the ApcMin/+mouse GI tract. (A) Aliphatic region of 600 MHz 1H HRMAS CPMG NMR spectrum of morphologically normal colon tissue from a C57BL/6J (wild type) mouse. (B) PLS-DA scores plot showing latent variables 1 and 2 of the binned data from the 600 MHz 1H HRMAS CPMG NMR spectra of morphologically normal tissue from the ApcMin/+ and wild type mice. Colored according to GI region: proximal small bowel (black), middle small bowel (red), distal small bowel (blue), colon (green). The PLS analysis was considered to be valid by permutation analysis with a Q2 (cum) value of 0.186 (for a 2 latent variable (LV) model). Different sections of ApcMin/+ mouse GI immunostained for F4/80, indicating typical morphology of the different GI regions are also shown.

Table 2 gives a summary for selected metabolites (for whole tissue, lamina propria, and epithelial cell extracts), with those changes indicated that were considered to be significant by Welch’s t test. Figure 2B shows the scores plot for a PLS-DA model of the ApcMin/+ and wild type lamina propria tissue. The model indicated that the metabolic profile varied between the different genetic backgrounds. Visual inspection and PLS loadings suggested that this separation was due to resonances at δ 2.7, 3.23, 3.25, and 3.68 ppm. The chemical shifts 3.23 and 3.68 correspond to glycerophosphocholine (GPC). In lamina propria tissue the GPC levels showed a more consistent difference along the whole GI tract (Figure 3A), with only the SB2 tissue not showing a decrease in GPC; overall, for all GI regions there is a decrease in GPC in the ApcMin/+ compared to the wild type (Figure 3B). In addition, other metabolites that appeared to be having an influence on the separation were: dimethylamine (DMA) (Figure 3C and D), myo-inositol (which appeared to be most significant for the whole tissue, Figure 3F), and phosphoethanolamine. The difference in the ratios of PC:GPC for

research articles

Metabolic Profiling of the ApcMin/+ Mouse Using NMR

extracts is shown in Figure 2C. The scores show a separation between samples from the Apcmin/+ and C57BL/6J mice. The loadings indicated that the separation was due to the signals at 3.78, 3.77, 5.53, 1.34, 1.32, 2.53, 1.33 ppm. O-PLS analysis (not shown) also highlighted resonances at 1.32-1.34 (lactate), 2.5-2.55, 3.775, and 4.11 (lactate). Lactate signals were lower in the samples from the ApcMin/+ epithelial cells compared to those from the C57BL/6J mice. Although there were trends in the differences (summarized in Table 2), there were no significant differences between the genotypes. Immunohistochemical Analysis. Sections of colon tissue were stained with F4/80 in order to assess the degree of macrophage infiltration. Mean(S.E.M) counts for ApcMin/+ and wild type were 34.1((0.63) and 33.8((0.63) respectively. No significant difference was observed between the macrophage counts in the lamina propria of the wild type and the ApcMin/+ mice (p > 0.7, Student’s t test).

Discussion

Figure 2. PLS-DA models discriminating according to genotype: ApcMin/+(black), wild type (red). PLS-DA scores plots showing latent variables 1 and 2 of the binned data from the 600 MHz 1H CPMG NMR spectra of (A) morphologically normal whole tissue (HRMAS), (B) morphologically normal lamina propria layer tissue (HRMAS) and (C) aqueous fraction of CHCl3/MeOH extracts of epithelial cells. All data was UV scaled. The PLS-DA analyses were assessed for validity by permutation analysis (see Supplementary Figure 2, Supporting Information) with the following cumulative Q2 values: (A) 0.319 (3 LV model), (B) 0.315 (2 LV model), (C) 0.452 (3 LV model).

both genotypes was also analyzed. The PC:GPC ratio was higher in the ApcMin/+ tissue compared to the wild type for the lamina propria tissue and the whole tissue (Figure 3E) with a significant difference (p < 0.05) for the lamina propria tissue from the colon. Epithelial Cell Extracts. The scores plot for a PLS-DA model for the aqueous fraction of the epithelial cell CHCl3/MeOH

In this analysis we have used 1H NMR spectroscopy to investigate the metabolic background of the ApcMin/+ mouse, which is used extensively as a model of gastrointestinal carcinogenesis. This was an exploratory study to see if there were any NMR detectable differences between the ApcMin/+ and C57BL/6J genetic backgrounds. Differences observed in spectra between the two genotypes identified glycerophosphocholine, myo-inositol, DMA, and lactate as potential discriminators between the ApcMin/+ mice and their wild-type littermates. Correspondence of Metabolic Differences Across Sample Types. Different patterns of GPC concentrations were observed in the whole tissue, cell extracts, and lamina propria. Experiments conducted to define the stability of the tissue under HRMAS conditions (data not shown), indicated a tendency for the levels of choline-containing compounds to increase over time, as observed previously by Swanson and co-workers,29 so it is possible that discrepancies between sample types originate from the separation procedure, prior to freezing. However, we still observed differences in the ApcMin/+ mouse compared to the wild type, and for the duration of the HRMAS experiments the tissue samples remained stable. Additionally, the epithelial cells have been through the most “processing”svia the separation and then extraction. It is conceivable that the general difference between the epithelial cell extracts and the tissue could be due to export of metabolites from cells to interstitial fluid. Lactate was seen to be decreased in the ApcMin/+ epithelial cell extracts compared to the wild type. It has been observed that tumor cells increase export of lactate via specific monocarboxylate transporters.30 If this were the case with these cells (i.e., the beginnings of a tumorigenic phenotype) then this could be a reason for observing differing lactate levels between whole tissue and extracted cells. In the intact tissue samples the interstitial fluid is retained. In cell extracts, and to a certain extent the lamina propria, it will have been lost. In this case the measurements from the whole tissue are probably most representative of in vivo conditions followed by the lamina propria, and then the epithelial cells. Due to the separation and extraction processes it is likely that there will be some metabolic inconsistencies. Biochemical Differences Along the GI Tract. Differences in tissue biochemistry along the GI tract have been observed previously in Long Evans rats18 and humans.20 Wang and coworkers attributed the variation to the different functions of Journal of Proteome Research • Vol. 8, No. 3, 2009 1427

research articles

Backshall et al.

Figure 3. Mean integrals for the regions of the spectra influential to separation according to genotype (A) GPC levels at 3.23 ppm for the lamina propria tissue in each GI region. (B) GPC levels in lamina propria tissue for all regions of the GI tract combined. (C) DMA levels in the lamina propria tissue in each GI region. (D) DMA levels in lamina propria tissue for all regions of the GI tract combined. (E) PC (3.22 ppm):GPC (3.23 ppm) ratio for both the lamina propria and the whole tissue, for all GI regions combined. (F) Myo-inositol levels in whole tissue for all regions of the GI tract combined. Error bars are standard error of the mean. §Combined values are the average across all regions where each value is normalized to the mean WT value for the matching bowel region. Asterisks indicate significance: *p e 0.05, **p e 0.01, ***p e 0.001 (Welch‘s t test).

the GI tract, resulting in different levels of energy metabolism, osmo-regulation, gut microbe activity, and oxidative protection.20 Our analysis indicates that there is an analogous difference in the biochemical composition of the tissue of the mouse intestine along the GI tract axis. This difference may be also attributed to the different digestive roles of the tissue along the GI tract as well as a difference in the gross morphology (see Figure 1B - histology panels). We specifically observed differences in the levels of GPC and PC along the axis for the epithelial cell extracts, with higher levels of PC at the proximal SB, and higher levels of GPC in the colon (data not shown). Wang and co-workers also observed differing GPC levels in the rats and humans, with higher levels of GPC in the jejunum (equivalent to SB2) compared to the ileum (equivalent to SB3). However, in the humans they observed decreased GPC in the colon compared to the small bowel, which is opposite to our observations in the mouse colon. These differences highlight two potential pitfalls when undertaking analysis of GI tissue. The first is the suggestion by Wang and co-workers that care that must be taken to maintain accuracy of localization; there may be an assumption that the tissue is of a uniform metabolic 1428

Journal of Proteome Research • Vol. 8, No. 3, 2009

profilesespecially in the mouse, which has a far less segregated GI tract than humans. The second is that there cannot be an assumption that the metabolic profiles of mice and men will be identical, as observed by Teichert et al., 2008, 13 who observed atypical phospholipid metabolism in TRAMP tumors compared to human prostate cancer. Metabolic Differences Associated with Genotype. Choline phospholipid metabolism has consistently been identified as altered in tumorigenic tissue and cells.31-33 In a study characterizing breast cancer cells and normal human mammary epithelial cells (HMECs), using 1H and 13C NMR it was observed that the breast cancer cells exhibited increased PC and total choline containing compounds, with significantly decreased GPC when compared with the normal HMECs.34 The increase in PC has also been observed in ovarian cell lines,35 and human colon cancer and adenoma tissue36 (with an associated increase in choline kinase activity).37 In the ApcMin/+ mouse altered levels of GPC were observed in the lamina propria tissue throughout the GI tract. Where it was observed to be significantly different, GPC levels were decreased in the ApcMin/+ mouse compared to wild type

research articles

Metabolic Profiling of the ApcMin/+ Mouse Using NMR littermates. This decrease in GPC may be indicative of an intermediate metabolic phenotype between normal and transformed cells. One of the genes that modifies the ApcMin/+ phenotype is Mom1, which codes for a secretory phospholipase (PL)A2, an enzyme in the phosphatidylcholine metabolic pathway.38,39Further, Affymetrix microarray data for the ApcMin/+ mouse GI tissue (http://compbio.dfci.harvard.edu/cancer/ data/APC/index.shtml) has indicated that mRNA levels for phospholipase A2 group IIA (platelets/synoviral fluid) are increased over 2-fold in the ApcMin/+ mouse compared to the wild type. PLA2 is part of the glycerophospholipid synthesis and degradation pathway, leading to the tentative suggestion that regulation of a modifier such as PLA2 could be affecting the GPC metabolism in the ApcMin/+ mouse. Other pathway-related metabolic differences were also observed. Levels of trimethylamine (TMA), the precursor to DMA have been linked to choline availability and gut microbial activity.40,41 The differences that we have assigned to DMA that were observed in the lamina propria tissue suggest that there may be a difference in the choline availability and metabolism between the ApcMin/+ and C57BL/6J genetic backgrounds, independent of phospholipid synthesis or degradation. Myoinositol, increased in the ApcMin/+ whole tissue compared to the wild type, is an osmolyte but also a precursor for phosphatidylinositol lipids, so alterations in the levels of this metabolite could reflect more general perturbation to lipid metabolism with the ApcMin/+ genotype Immune System Involvement. The metabolic alteration localized to the lamina propria could also be explained by a change in its nature due to macrophage infiltration into the lamina propria. As part of their genetic analysis Chen et al.,42 observed expression of COX-2 in macrophages, but only weak staining on the epithelial cells. It has been observed that upregulation of COX-2 may occur in macrophages in histologically normal lamina propria tissue prior to adenoma formation,43 potentially providing an environment that is conducive to epithelial cell growth.44 We stained sections of colon tissue with F4/80 to see if macrophage infiltration could be a contributing factor to the differences observed, but saw no significant differences in the macrophage numbers between the wild type and the ApcMin/+ tissue. A Physiological Model for Early Phenotypic Change in the ApcMin/+ Mouse. Analysis of whole tissue, lamina propria tissue, and epithelial cell extracts indicated that there are differing metabolic profiles between the non-tumor tissue of the ApcMin/+ mouse compared to the wild type. The differences appear to be subtle and variable, depending on tissue type, and location along the GI tract axis. This type of effect localization has been observed previously in ApcMin/+ mice in relation to induction of tumorigenesis, where increased tumor multiplicity caused by TGFR overexpression was limited to the jejunum.45 It has also been observed that genes most highly up-regulated in polyps were also up regulated in the surrounding polyp free mucosa in ApcMin/+ mice,42 suggesting that a region of tissue that gives rise to polyps is altered compared to nonpolyp regions (or wild-type tissue). Our results suggest that the reduced function of APC is producing an altered metabolic phenotype. It is generally considered that loss of both alleles of Apc is sufficient for adenoma formation.46 The morphologically normal GI tissue we have analyzed contains cells with heterozygous mutation, and we are therefore potentially observing a phenotype alteration due to this heterozygosity, prior to LOH. In their review Humphries and Wright47

(and references therein) propose that the crypts of mice with Apc+/– heterozygosity are able to expand at a faster rate than normal, creating a field of Apc+/– crypts, with more stem cells that are at risk from further APC mutation and LOH, resulting in adenoma (consistent with the “field cancerization” hypothesis16). The subtle metabolic differences that we have reported may be a result of us observing a mixture of normal and fissioning Apc+/– crypts. As a preliminary metabolic analysis of the ApcMin/+ mouse phenotype, this work indicates some subtle differences that define an intermediate state between the wild type profile and the expected tumor profile detected by NMR. The work has also complemented studies indicating that care must be taken when comparing tissue from different regions of the mouse gut. Further investigation of different Min phenotypes at varying ages will give a clearer indication of the relationship between these metabolic signatures and the genetic background in the Min mouse with the eventual aim of identifying biomarkers for detection and monitoring intervention that have the potential to be translated to the clinical management and prevention of colorectal cancer.

Acknowledgment. A.B. is supported by an MRC PhD studentship. D.A. was funded by BBSRC-AstraZeneca-CRUK. We acknowledge the technical support provided by Nikki Mandir (Histopathology), Robert Rudling, and Barbara Cross (Cancer Research, Biological resources). Supporting Information Available: Supplementary Figures 1 and 2. This material is available free of charge via the Internet at http://pubs.acs.org. References (1) Ilyas, M.; Straub, J.; Tomlinson, I. P.; Bodmer, W. F. Genetic pathways in colorectal and other cancers. Eur. J. Cancer 1999, 35 (14), 1986–2002. (2) Fodde, R.; Smits, R.; Clevers, H. APC, signal transduction and genetic instability in colorectal cancer. Nat. Rev. Cancer 2001, 1 (1), 55–67. (3) Fodde, R.; Smits, R. Disease model: familial adenomatous polyposis. Trends Mol. Med. 2001, 7 (8), 369–73. (4) Moser, A. R.; Pitot, H. C.; Dove, W. F. A dominant mutation that predisposes to multiple intestinal neoplasia in the mouse. Science 1990, 247 (4940), 322–4. (5) Shoemaker, A. R.; Gould, K. A.; Luongo, C.; Moser, A. R.; Dove, W. F. Studies of neoplasia in the Min mouse. Biochim. Biophys. Acta 1997, 1332 (2), F25–48. (6) Aoki, K.; Taketo, M. M. Adenomatous polyposis coli (APC): a multifunctional tumor suppressor gene. J. Cell Sci. 2007, 120 (Pt 19), 3327–35. (7) Clarke, A. R. Wnt signalling in the mouse intestine. Oncogene 2006, 25 (57), 7512–21. (8) Beckonert, O.; Keun, H. C.; Ebbels, T. M. D.; Bundy, J.; Holmes, E.; Lindon, J. C.; Nicholson, J. K. Metabolic profiling, metabolomic and metabonomic procedures for NMR spectroscopy of urine, plasma, serum and tissue extracts. Nat. Protocols 2007, 2 (11), 2692–2703. (9) Fan, T. W.-M. Metabolite profiling by one- and two-dimensional NMR analysis of complex mixtures. Prog. Nucl. Magn. Reson. Spectrosc. 1996, 28 (2), 161–219. (10) Nicholson, J. K.; Foxall, P. J.; Spraul, M.; Farrant, R. D.; Lindon, J. C. 750 MHz 1H and 1H-13C NMR spectroscopy of human blood plasma. Anal. Chem. 1995, 67 (5), 793–811. (11) Ala-Korpela, M. 1H NMR spectroscopy of human blood plasma. Prog. Nucl. Magn. Reson. Spectrosc. 1995, 27 (5-6), 475–554. (12) Griffin, J. L.; Kauppinen, R. A. Tumour metabolomics in animal models of human cancer. J. Proteome Res. 2007, 6 (2), 498–505. (13) Teichert, F.; Verschoyle, R. D.; Greaves, P.; Edwards, R. E.; Teahan, O.; Jones, D. J.; Wilson, I. D.; Farmer, P. B.; Steward, W. P.; Gant, T. W.; Gescher, A. J.; Keun, H. C. Metabolic profiling of transgenic adenocarcinoma of mouse prostate (TRAMP) Tissue by (1)H-NMR analysis: evidence for unusual phospholipid metabolism. Prostate 2008, 68 (10), 1035–47.

Journal of Proteome Research • Vol. 8, No. 3, 2009 1429

research articles (14) Liotta, L. A.; Kohn, E. C. The microenvironment of the tumourhost interface. Nature 2001, 411 (6835), 375–9. (15) Gatenby, R. A.; Gillies, R. J. Why do cancers have high aerobic glycolysis. Nat. Rev. Cancer 2004, 4 (11), 891–9. (16) Braakhuis, B. J. M.; Tabor, M. P.; Kummer, J. A.; Leemans, C. R.; Brakenhoff, R. H. A Genetic Explanation of Slaughter‘s Concept of Field Cancerization: Evidence and Clinical Implications. Cancer Res. 2003, 63 (8), 1727–1730. (17) Nicholson, J. K.; Holmes, E.; Wilson, I. D. Gut microorganisms, mammalian metabolism and personalized health care. Nat. Rev. Microbiol. 2005, 3 (5), 431–8. (18) Wang, Y.; Tang, H.; Holmes, E.; Lindon, J. C.; Turini, M. E.; Sprenger, N.; Bergonzelli, G.; Fay, L. B.; Kochhar, S.; Nicholson, J. K. Biochemical characterization of rat intestine development using high-resolution magic-angle-spinning 1H NMR spectroscopy and multivariate data analysis. J. Proteome Res. 2005, 4 (4), 1324– 9. (19) Wang, Y.; Cloarec, O.; Tang, H.; Lindon, J. C.; Holmes, E.; Kochhar, S.; Nicholson, J. K. Magic Angle Spinning NMR and (1)H-(31)P Heteronuclear Statistical Total Correlation Spectroscopy of Intact Human Gut Biopsies. Anal. Chem. 2008, 80 (4), 1058–1066. (20) Wang, Y.; Holmes, E.; Comelli, E. M.; Fotopoulos, G.; Dorta, G.; Tang, H.; Rantalainen, M. J.; Lindon, J. C.; Corthesy-Theulaz, I. E.; Fay, L. B.; Kochhar, S.; Nicholson, J. K. Topographical variation in metabolic signatures of human gastrointestinal biopsies revealed by high-resolution magic-angle spinning 1H NMR spectroscopy. J. Proteome Res. 2007, 6 (10), 3944–51. (21) Tugnoli, V.; Mucci, A.; Schenetti, L.; Righi, V.; Calabrese, C.; Fabbri, A.; Di Febo, G.; Tosi, M. R. Ex vivo HR-MAS Magnetic Resonance Spectroscopy of human gastric adenocarcinomas: a comparison with healthy gastric mucosa. Oncol. Rep. 2006, 16 (3), 543–53. (22) Moreno, A.; Rey, M.; Montane, J. M.; Alonso, J.; Arus, C. 1H NMR spectroscopy of colon tumors and normal mucosal biopsies; elevated taurine levels and reduced polyethyleneglycol absorption in tumors may have diagnostic significance. NMR Biomed. 1993, 6 (2), 111–8. (23) Goodlad, R. A.; Ryan, A. J.; Wedge, S. R.; Pyrah, I. T.; Alferez, D.; Poulsom, R.; Smith, N. R.; Mandir, N.; Watkins, A. J.; Wilkinson, R. W. Inhibiting vascular endothelial growth factor receptor-2 signaling reduces tumor burden in the ApcMin/+ mouse model of early intestinal cancer. Carcinogenesis 2006, 27 (10), 2133–9. (24) Rudling, R.; Hassan, A. B.; Kitau, J.; Mandir, N.; Goodlad, R. A. A simple device to rapidly prepare whole mounts of murine intestine. Cell Prolif. 2006, 39 (5), 415–20. (25) Dieterle, F.; Ross, A.; Schlotterbeck, G.; Senn, H. Probabilistic quotient normalization as robust method to account for dilution of complex biological mixtures. Application in 1H NMR metabonomics. Anal. Chem. 2006, 78 (13), 4281–90. (26) Westerhuis, J. Assessment of PLSDA cross validation. Metabolomics 2008, 4 (1), 81–89. (27) Alferez, D.; Goodlad, R. A. To best measure cell proliferation in samples from the intestine. Cell Prolif. 2007, 40 (2), 231–40. (28) Orner, G. A.; Dashwood, W. M.; Blum, C. A.; Diaz, G. D.; Li, Q.; AlFageeh, M.; Tebbutt, N.; Heath, J. K.; Ernst, M.; Dashwood, R. H. Response of Apc(min) and A33 (delta N beta-cat) mutant mice to treatment with tea, sulindac, and 2-amino-1-methyl-6-phenylimidazo[4,5-b]pyridine (PhIP). Mutat. Res. 2002, 506-507, 121–7. (29) Swanson, M. G.; Zektzer, A. S.; Tabatabai, Z. L.; Simko, J.; Jarso, S.; Keshari, K. R.; Schmitt, L.; Carroll, P. R.; Shinohara, K.; Vigneron, D. B.; Kurhanewicz, J. Quantitative analysis of prostate metabolites using 1H HR-MAS spectroscopy. Magn. Reson. Med. 2006, 55 (6), 1257–64. (30) Halestrap, A. P.; Price, N. T. The proton-linked monocarboxylate transporter (MCT) family: structure, function and regulation. Biochem. J. 1999, 343 Pt 2, 281–99.

1430

Journal of Proteome Research • Vol. 8, No. 3, 2009

Backshall et al. (31) Ackerstaff, E.; Glunde, K.; Bhujwalla, Z. M. Choline phospholipid metabolism: a target in cancer cells. J. Cell Biochem. 2003, 90 (3), 525–33. (32) Glunde, K.; Jacobs, M. A.; Bhujwalla, Z. M. Choline metabolism in cancer: implications for diagnosis and therapy. Expert Rev. Mol. Diagn. 2006, 6 (6), 821–9. (33) Podo, F. Tumour phospholipid metabolism. NMR Biomed. 1999, 12 (7), 413–39. (34) Glunde, K.; Jie, C.; Bhujwalla, Z. M. Molecular causes of the aberrant choline phospholipid metabolism in breast cancer. Cancer Res. 2004, 64 (12), 4270–6. (35) Iorio, E.; Mezzanzanica, D.; Alberti, P.; Spadaro, F.; Ramoni, C.; D’Ascenzo, S.; Millimaggi, D.; Pavan, A.; Dolo, V.; Canevari, S.; Podo, F. Alterations of choline phospholipid metabolism in ovarian tumor progression. Cancer Res. 2005, 65 (20), 9369–76. (36) Chan, E. C. Y.; Koh, P. K.; Mal, M.; Cheah, P. Y.; Eu, K. W.; Backshall, A.; Cavill, R.; Nicholson, J. K.; Keun, H. C. Metabolic profiling of human colorectal cancer using high-resolution magic angle spinning nuclear magnetic resonance (HR-MAS NMR) spectroscopy and gas chromatography mass spectrometry (GC/MS). J. Proteome Res. 2008, 10.1021/pr8006232. (37) Nakagami, K.; Uchida, T.; Ohwada, S.; Koibuchi, Y.; Suda, Y.; Sekine, T.; Morishita, Y. Increased choline kinase activity and elevated phosphocholine levels in human colon cancer. Jpn. J. Cancer Res. 1999, 90 (4), 419–24. (38) Haines, J.; Johnson, V.; Pack, K.; Suraweera, N.; Slijepcevic, P.; Cabuy, E.; Coster, M.; Ilyas, M.; Wilding, J.; Sieber, O.; Bodmer, W.; Tomlinson, I.; Silver, A. Genetic basis of variation in adenoma multiplicity in ApcMin/+ Mom1S mice. Proc. Natl. Acad. Sci. U.S.A. 2005, 102 (8), 2868–73. (39) MacPhee, M.; Chepenik, K. P.; Liddell, R. A.; Nelson, K. K.; Siracusa, L. D.; Buchberg, A. M. The secretory phospholipase A2 gene is a candidate for the Mom1 locus, a major modifier of ApcMininduced intestinal neoplasia. Cell 1995, 81 (6), 957–66. (40) Zeisel, S. H.; daCosta, K. A.; Youssef, M.; Hensey, S. Conversion of Dietary Choline to Trimethylamine and Dimethylamine in Rats: Dose-Response Relationship. J. Nutr. 1989, 119 (5), 800–804. (41) al-Waiz, M.; Mikov, M.; Mitchell, S. C.; Smith, R. L. The exogenous origin of trimethylamine in the mouse. Metabolism 1992, 41 (2), 135–6. (42) Chen, L. C.; Hao, C. Y.; Chiu, Y. S.; Wong, P.; Melnick, J. S.; Brotman, M.; Moretto, J.; Mendes, F.; Smith, A. P.; Bennington, J. L.; Moore, D.; Lee, N. M. Alteration of gene expression in normalappearing colon mucosa of APC(min) mice and human cancer patients. Cancer Res. 2004, 64 (10), 3694–700. (43) Hull, M. A.; Booth, J. K.; Tisbury, A.; Scott, N.; Bonifer, C.; Markham, A. F.; Coletta, P. L. Cyclooxygenase 2 is up-regulated and localized to macrophages in the intestine of Min mice. Br. J. Cancer 1999, 79 (9-10), 1399–405. (44) Inaba, T.; Sano, H.; Kawahito, Y.; Hla, T.; Akita, K.; Toda, M.; Yamashina, I.; Inoue, M.; Nakada, H. Induction of cyclooxygenase-2 in monocyte/macrophage by mucins secreted from colon cancer cells. Proc. Natl. Acad. Sci. U.S.A. 2003, 100 (5), 2736–41. (45) Bilger, A.; Sullivan, R.; Prunuske, A. J.; Clipson, L.; Drinkwater, N.; Dove, W. F. Widespread hyperplasia induced by transgenic TGFR in ApcMin mice is associated with only regional effects on tumorigenesis. Carcinogenesis 2008, 29, 1825–1830. (46) Lamlum, H.; Papadopoulou, A.; Ilyas, M.; Rowan, A.; Gillet, C.; Hanby, A.; Talbot, I.; Bodmer, W.; Tomlinson, I. APC mutations are sufficient for the growth of early colorectal adenomas. Proc. Natl. Acad. Sci. U.S.A. 2000, 97 (5), 2225–8. (47) Humphries, A.; Wright, N. A. Colonic crypt organization and tumorigenesis. Nat. Rev. Cancer 2008, 8 (6), 415–24.

PR800793W