Journal of Nutrition Animal Diets/Enrichment Products...

Home Help [Feedback] [For Subscribers] [Archive] [Search] [Contents]
QUICK SEARCH:   [advanced]


     

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Purchase Article
Right arrow View Shopping Cart
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Lee, W.-N. P.
Right arrow Articles by Go, V. L. W.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Lee, W.-N. P.
Right arrow Articles by Go, V. L. W.
© 2005 American Society for Nutrition J. Nutr. 135:3027S-3032S, December 2005

Supplement: International Conference on Diet, Nutrition, and Cancer

Nutrient-Gene Interaction: Tracer-Based Metabolomics1,2,3

Wai-Nang P. Lee and Vay Liang W. Go*,4

LABiomed Research Institute at Harbor-UCLA Medical Center, and * David Geffen School of Medicine at UCLA, University of California, Los Angeles, CA

4To whom correspondence should be addressed. E-mail: lee{at}labiomed.org.

ABSTRACT

Understanding nutrient-gene interaction requires tools for both the study of nutrigenomics and the characterization of phenotype. Metabolomics or metabolite profiling is a powerful tool for characterizing metabolic phenotype, and tracer-based metabolomics is a subset of metabolomics that focuses on metabolite distribution and flux determination using tracers. In this review, the characterizations of metabolic phenotype by metabolite profiling and by metabolic flux measurements are compared. The rationale and methodologies of tracer-based metabolomics are explained. Tracer-based metabolomics provides a relational database of metabolites linked by the relationship of shared metabolic pathways, common substrates, and cofactors. Such a collection of flux measurements provides precise and accurate information on the operation of the cellular metabolic network and its response to genetic and nutrient environment changes. Nutrient-gene interaction can be studied using the concept of constraint-based modeling, which states that the observed metabolic phenotype is a consequence of constraints from genetic factors and the nutrient environment. Thus, genetic inheritance (genomic constraints) confers a wide range of possible phenotypes whereas selection by metabolic (structural and pathway relationship) and environmental (physical environment and nutrient availability) constraints determines the final observed phenotype. The study of the contribution from nutrient and genetic factors to the survival advantage of cancer cells using flux measurements is a critical first step in our understanding of the relationship between nutrient intake and cancer risk.

KEY WORDS: • glucose intermediary metabolism • ribose synthesis • isotopomer analysis • constraint-based modeling

The current understanding of the malignant transformation of cells heavily emphasizes genetic regulation of the cell cycle, proliferation, apoptosis, angiogenesis, and metastasis. Uncontrolled proliferation and metastasis of cancer cells has been shown to be associated with the loss of regulation due to the presence of cancer promoters or the absence of tumor suppressor genes (1). However, this deterministic paradigm of the etiology of cancer is not sufficient to explain the great variation in the expression of human genes in response to mutation and environmental changes, nutrition, lifestyle, and age resulting in the variability of the phenotypes of cancer. Consequently, the relationship between nutrient intake and cancer risk is poorly understood.

To fully understand the links between nutrition and cancer, one must be able to answer 2 key scientific questions: What are the innate genetic responses to changes in nutrient environment? How do genetic mutations affect a cell’s metabolic response to changes in nutrient environment? The field of nutrigenomics was developed to address the former question by studying the activation of genes by nutrient components. The study of genotype-phenotype correlation provides answers to the second question (2).

The phenotype of an organism is generally defined by a set of physical (size and shape, body composition, fur color, etc.) or chemical (plasma glucose, lipids, creatinine, etc.) characteristics. Because the distribution of metabolites gives a more complete description of a phenotype, metabolite profiling is considered to be one of the basic approaches to phenotyping. Many methods have been available for detecting and quantifying metabolic intermediates, but the development of methodologies for characterizing phenotype lags behind those of genomics and proteomics because of the lack of high-throughput technology. Recent advances in the field of nuclear magnetic resonance spectroscopy (NMR) and mass spectrometry have greatly expanded our ability to characterize and quantify metabolites. The quantitation of metabolites by such high-throughput and multiplex technologies has become another "omics" modeled after genomics and proteomics. Just as genomics and proteomics are collections of genes and proteins that are expressed, metabolomics is a set of metabolite measurements corresponding to a certain cellular state (3,4). However, Hellerstein (5) noted that the measurement of metabolites is still inadequate for predicting phenotype, and measurement of metabolic fluxes provides an integrated picture and a better approximation to the metabolic phenotype. In this review, we compare the characterization of metabolic phenotype by metabolite profiling and by metabolic flux measurements. The rationale and methodologies of tracer-based metabolomics are explained. Examples of phenotypic changes in response to nutrient environment are provided. The use of flux measurement in metabolic phenotyping may help us better understand nutrient-gene interaction and the role of diet in the prevention of cancer.

Metabolic phenotyping

The metabolic phenotype of an organism is usually characterized by the concentration of metabolic intermediates or by the changes in concentrations over time in the organism. The measurement of metabolites for phenotyping is known as metabolite profiling. Metabolite profiling evolved as an adjunct of genomics and proteomics. The belief that there may be a one-to-one correspondence between a certain gene mutation and changes in steady state distribution of metabolites encouraged the development of metabolic profiling as a tool for the narrowing the potential choices of candidate genes in determining potential gene mutation (6). Metabolomics (metabolite profiling), as defined by Harrigan et al. (7), "involves the acquisition of metabolome data sets of sufficient spectral and/or chromatrographic richness and resolution for multivariate statistical analysis and for metabolite identification and quantitation." As the definition implies, the data sets are method or technology dependent regardless of specific knowledge of the individual metabolites. As a result, such an approach is best characterized as finger printing, similar to as bar coding for genotyping. When spectral or chromatographic peaks can be resolved into individual metabolites, the collection of such a data set is defined as metabolomics. The data sets generated by the newer technologies of NMR and mass spectrometry are usually very large, and reducing the number of variables to a manageable few components by multivariate statistical analyses such as principal component analysis is usually required (7).

The utility of metabolomics depends on insights into the clinical correlation with a cluster of variables. Based on the normal ranges of these variables, categories of normal, high, or low values can be grouped. Thus, such metabolomic approaches, in a more narrow sense, can also be seen as pigeonholing. As shown in Figure 1, plasma glucose and triglyceride concentrations are used to segregate 9 possible clinical states. The insight into such clinical classification depends on the a priori knowledge of normal ranges of glucose and triglycerides, which cannot be derived from the measurement of metabolites alone. Metabolite profiling has been widely used in clinical medicine for laboratory diagnosis of diseases. Typically, clinical diseases are evaluated according to the presence of abnormal laboratory values. Most commonly, a single metabolite measurement is used for diagnosis and therapeutic endpoint in patient management. Such application of metabolite profiling is best illustrated by the use of plasma acylcarnitine levels for the detection of inborn errors of metabolism of fatty acid oxidation enzymes and the use of lipid profile in the clinical classification of dyslipidemia (8,9).



View larger version (9K):
[in this window]
[in a new window]
  		FIGURE 1 Metabolic phenotype as characterized by metabolite concentrations. Glucose and triglyceride concentrations are used to define 9 potential phenotypes [indicated as (H, H; H, N; H, L; etc.)]

 
When metabolite profiling is performed after a perturbation in a living system, the individual time series of metabolite concentration provides a measurement of substrate flux, which provides an integrated and dynamic picture of the metabolic system. Thus, measurement of fluxes in multiple metabolites gives a better approximation of the metabolic phenotype. Figure 2 illustrates the use of flux measurement for metabolic phenotyping. As in Figure , 1a system of 2 variables (glucose production and glucose utilization) is presented. The phenotype (flux) space is more accurately characterized by a multidimensional space of many substrate fluxes, which is often incompletely characterized because of the limitation of available methodologies. Each point on the graph represents a functional state (phenotype) characterized by a certain glucose production and utilization rate. Each point also has a corresponding steady state glucose concentration. Each line graph depicts a quantitative relationship between glucose production and utilization and bisects the plane into areas of potential high and low steady state glucose concentrations. The application of flux measurements in the characterization of metabolic phenotype has an additional advantage over metabolite profiling in that all known information of the metabolic network, including enzyme expression, proteomic, and enzyme kinetics data (as shown in Fig. 3), is embodied in the flux measurements, thus further constraining or defining the metabolic phenotype. In Figure 2 such constraints are shown as squares delimiting the range of possible phenotypes resulting from a specific genotype. Such representation is also known as constraint-based modeling (10). When glucose production exceeds utilization, as in insulin resistance, the biological system has a hyperglycemic phenotype as represented by the space within the shaded square. When utilization exceeds production, as in the fetal state, the phenotype is characterized by hypoglycemia represented by the space within the checkered square. In the population with the genotype that permits glucose production and utilization to vary with the box that straddles the euglycemic line, the phenotype space consists of blood glucose values that vary within a narrow range.



View larger version (16K):
[in this window]
[in a new window]
  		FIGURE 2 Metabolic phenotype as characterized by metabolite flux measurements. Glucose production and utilization are functionally related creating an infinite phenotype space. The actual possible phenotypes arising from genetic constraints are indicated by areas enclosed by the squares.

 


View larger version (30K):
[in this window]
[in a new window]
  		FIGURE 3 A subset of the glucose metabolic network. The substrate fluxes through these metabolic pathways are constrained by shared intermediates and enzyme cofactors. Thus, changes in the flux of one pathway is modulated by other pathways of the network. Three subsystems are identified: A, glycolytic/gluconeogenic subsystem; B, the tricarboxylic acid cycle; and C, the lipid synthesis subsystem. PC, pyruvate carboxylase; PDH, pyruvate dehydrogenase; PEP, phosphoenolpyruvate; PEPCK, phosphoenolpyruvate carboxykinase; PK, pyruvate kinase.

 
Nutrient-gene interaction can also be understood in terms of changes in the flux solution space. If carbohydrate intake is fixed such that total production is above 35 µmol · kg–1 · min–1, the observed metabolic phenotype will be predominantly hyperglycemic. The converse (hypoglycemic phenotype) is observed when intake is restricted such that total production is below 25 µmol · kg–1 · min–1. The white space bounded by these 2 limits represents the metabolic phenotype with normal blood glucose values.

The above illustrations draw on examples of phenotyping using 2 flux parameters. In reality, the phenotype space is multidimensional, constructed from flux measurement of many metabolites. Because living organisms all share a common characteristic—change over time—flux measurements of pathways of a metabolic network provide a more accurate characterization of a phenotype. Variations in gene expression and nutrient environment changes can be incorporated into the model as genotypic and metabolic constraints. Thus, nutrient-gene interaction can be systematically investigated using flux measurements of the metabolic system.

Tracer-based metabolomics

Tracer-based metabolomics is a subset of metabolomics that focuses on metabolite distribution and flux determination using tracers. The idea of using tracers to study metabolic pathways and flux analysis is not new. It has its origin in the creative works of Katz and Rognstad (11), who demonstrated specific positional labeling of carbon or hydrogen in a metabolite from labeled precursors. With the advent of NMR and mass spectrometry, stable isotope tracers have essentially replaced the original radioactive tracer. When a 13C-labeled substrate is introduced into a biological system, 13C is incorporated either through exchange or by direct synthesis into a wide range of metabolites collectively known as the metabolome. The incorporation of a labeled carbon molecule into a metabolic product generates a mass signature (a difference in molecular weight from the naturally existing compound) that can be detected by spectrometry. The use of stable isotope tracers has a further advantage over radioactive tracers in that a molecule can be labeled in several positions (mass isotopomer) to trace different metabolic pathways, and the incorporation of 13C or deuterium into metabolites can be easily assayed using mass spectrometry (12,13). For example, consider [1,2-13C2]glucose in the study of pentose cycle metabolism. [1,2-13C2]Glucose is a mass isotopomer with two 13C atoms in the carbon 1 and 2 positions. If glucose is metabolized through glycolysis, products such as glycerol, pyruvate, and lactate containing two 13C atoms (m2) are formed. However, if glucose undergoes oxidation via glucose 6 phosphate dehydrogenase pathways, one of the 13C atoms will be lost as carbon dioxide and products such as ribose containing a single 13C atom (m1) will be formed. When these pathways are linked to the transketolase-transaldolase pathways, newly synthesized pentose molecules containing odd and even numbers of 13C atoms (m1, m2, m3, and m4) are generated. These molecules can be clearly quantitated by GC/MS as shown in Figure 4.



View larger version (16K):
[in this window]
[in a new window]
  		FIGURE 4 Mass spectrum of ribose aldonitrile acetate showing incorporation of 13C from [1,2 13C2]-glucose via oxidative and nonoxidative pathways of the pentose cycle. The molecular ion is at m/z 256. The incorporation of 13C creates incremental masses.

 
The accumulation of 13C among the many metabolic intermediates is a measure of flux distribution from the labeled precursor to the various products. The distribution of 13C within the molecule provides an estimate of contribution to the synthesis through various pathways. The tracer distribution within individual compounds and among compounds is a metabolic phenotypic feature that depends on the functional state of the cell and its response to nutrient environment changes. When there is a large data set of mass isotopomers of metabolites from labeling experiments with 13C- or 2H-labeled substrates, the data set of isotopomer enrichment relative to that of the control can be organized and tabulated in an array format (SIDMAP array®) with pathway annotations (14). Flux measurements of pathways of a metabolic network accurately characterize a phenotype from a systems biology point of view integrating contributions from gene expression and nutrient environment changes. The "array" is a summary of substrate fluxes within the cellular metabolic network and is very useful for phenotype characterization in the study of nutrient-gene interaction.

[1,2-13C2]Glucose and other 13C-labeled glucose compounds are not the only tracers used for investigations in tracer-based metabolomics. Metabolic intermediates such as [2-13C]glycerol, [U-13C3]glycerol, and [U-13C3]lactate were used to probe the glycolytic-gluconeogenic pathways (15–17). [2-13C]Acetate, [1,2-13C2]acetate, and uniformly 13C-labeled fatty acids were used to investigate fatty acid synthesis and oxidation (18–20). The field of tracer-based metabolomics is still in its infancy. Future development will see the use of labeled substrates of sterols, amino acids, and essential fatty acids for probing other metabolic systems. It is anticipated that progress in tracer-based metabolomics will merge with progress in the other branches of bioinformatics and systems biology.

Influence of genes and nutrients on phenotype

In the past 10 y, tracer-based metabolomics has been applied to investigate phenotypic changes in cancer cells in culture. It is possible for the first time to characterize the contribution of oxidative and nonoxidative pathways to ribose and deoxyribose synthesis in addition to a number of other pathways connected with glucose metabolism (21). Genetic factors confer to cells a specific set of metabolic functions that correspond to a metabolic phenotype under a given nutrient environment. Thus, lung adenocarcinoma (H441) and myeloid leukemic (K562) cells have high transketolase activity for DNA and RNA ribose synthesis as compared with the relative preponderance of the oxidative pentose cycle activity in pancreatic adenocarcinoma (MIA) cells or inflammatory breast cancer cells (22–24). The patterns of substrate redistribution of these rapidly proliferating cells are also different from that of normal human fibroblasts, in which glycolysis is more prominent than pentose synthesis (25).

Genetic mutations leading to a proliferative phenotype are associated with the requirement for increased nucleic ribose and deoxyribose synthesis and glucose utilization (26,27). Such genetic changes often do not directly involve genetic changes altering enzymatic properties of metabolic enzymes. Many of the metabolic functions of the pretransformed cells may be conserved. For example, when NIH3T3 cells are transformed with the introduction of mutated k-ras oncogenes, these cells are unable to increase de novo lipogenesis and remain dependent on fatty acids in the medium (28). Metabolic changes are often the consequence of the activated pentose cycle affecting other pathways of the glucose metabolic network. The influence of hormones and growth factors provides the glucose metabolic pathways changes that permit certain cancer cells to compete even more effectively with surrounding normal tissues (29). Thus, transforming growth factor-ß promotes invasive transformation of H441 and similarly transformation of leukemic cells by isofenphos (30). The concept of competition among the various metabolic phenotypes of 3 cell lines is illustrated by Figure 5A. Under their respective genomic constraints, MIA cells have a higher glucose utilization rate and transketolase flux compared with those of fibroblasts cell lines. Thus, for the same glucose supply, MIA cells compete for glucose more effectively and proliferate better than fibroblasts of a normal subject and a patient with thiamine response megaloblastic anemia. However, such genotypically endowed properties for proliferation and survival can be dramatically altered by nutrient environment changes. In Figure 5B, when fibroblasts from patients with thiamine response megaloblastic anemia are cultured with thiamine supplementation, the high-thiamine environment permits normal function of the transketolase enzyme and glucose utilization and restores normal proliferation of these cells (31). Such an observation suggests that thiamine supplementation may promote cancer development in cancers that have a high requirement for transketolase activity and agrees with the epidemiologic finding of an association between thiamine supplements in the diet and the high cancer rate in Western countries (32). When the transketolase enzyme is inhibited by oxythiamine, MIA cells in culture enter into cell cycle arrest and apoptosis, and cell proliferation is reduced (Fig. 5B). The induction of cell apoptosis in MIA cells by glycogen phosphorylase inhibitor (33) and 2-deoxyglucose are further examples of metabolic constraints on cancer survival and growth (34). Therefore, it can be said that nutrient environment changes select cells with the actual observed metabolic phenotypes, which survive best under such environments.



View larger version (22K):
[in this window]
[in a new window]
  		FIGURE 5 Phenotype changes according to genetic and metabolic constraints of pancreatic carcinoma (MIA) cells and fibroblasts of normal and patient with thiamine responsive megaloblastic anemia (TRMA): A, Metabolic phenotypes as a consequence of genetic constraints for these cell lines are shown in a phase plane of glucose utilization and transketolase flux; B, Metabolic constraints due to nutrient environment changes modify the metabolic phenotype of TRMA fibroblasts and MIA cells. Thus, several metabolic phenotypes are possible for the same genotype, and the final observed phenotype is the result of both genetic and metabolic constraints.

 
Nutrient environment and cancer risks

The mechanism that links dietary factors to increased cancer risks is not known. Although increased cancer risk due to other environmental factors such as smoking and sunlight exposure have been linked to mutagenic or carcinogenic effects of such factors, it is clear that the increased cancer risk from overnutrition is not due to increased intake of carcinogens or mutagens. The understanding of the relationship between metabolic phenotype changes and nutrient-gene interaction suggests the following relationship between nutrient environmental changes and cancer risks. The presence of a tumor-promoting gene or the loss of a tumor suppressor gene allows the cell with the mutation to proliferate indefinitely without becoming terminally differentiated. These cells continue to divide, and with successive division acquire new metabolic characteristics that confer a competitive advantage in the metabolic milieu.

Such a scenario is illustrated by our study in the colonic carcinoma (HT29) cells (Fig. 6) (35). Butyrate, a 4-carbon fatty acid, is a product of bacterial fermentation and is normally present in the colonic environment in high concentration. Colonic cells have evolved to possess enzymes for butyrate metabolism that allow these cells to thrive in such an environment. These cells are characterized by high butyrate utilization and relative independence from glucose utilization. As dietary intake of vegetable fibers decreases, the butyrate concentration in the colonic environment also decreases putting selection pressure on the normal colonic cells. In the presence of an oncogene (e.g., APC), colonic cells continue to proliferate instead of undergoing apoptosis. Finally, the cell type with an efficient glucose metabolic system is selected. The glucose metabolic phenotype permits the transformed colonic cells to compete effectively against normal tissues and become invasive. Interestingly, the transformed HT29 cells continue to possess butyrate metabolic enzymes. When they are exposed to a high-butyrate environment, butyrate becomes the preferred substrate and induces phenotypic changes in the HT29 cells. Because MIA cells are derived from a cell type that cannot utilize butyrate, the addition of butyrate in the culture medium does not result in any significant phenotypic changes (35).



View larger version (11K):
[in this window]
[in a new window]
  		FIGURE 6 The impact of nutrient gene interaction on metabolic phenotype of colon carcinoma (HT29) cells. The genotype of colonic cells confers a range of metabolic phenotypes with different capacities for butyrate and glucose utilization. Reduced butyrate level in the colonic environment selects for the high-glucose-utilization phenotype, which can revert to the high-butyrate-utilization phenotype when butyrate is reintroduced.

 
The above study supports the hypothesis that the evolution of cancer cells can be understood in terms of metabolic selection. Genetic mechanisms are the source of phenotypic variation whereas the final observed species is the result of metabolic selection.

Discussion and conclusion

Nutritional and environmental factors are important mechanisms for the development of common human malignancies in addition to tumor-inducing genes and growth-modulating signal transduction pathways. The mechanisms linking diet to cancer are poorly understood because of a lack of conceptual and experimental tools (36). We hypothesized that tumor cells, as a consequence of genetic alteration, possess a genotype with the potential for many phenotypes such as proliferation, differentiation, cell cycle arrest, and apoptosis (25). The diverse genetic mechanisms of tumor induction have in common their resultant influence on metabolism (altering the respective genetic constraints), thus altering the normal potential for differentiation, cell cycle arrest, and apoptosis. Depending on the availability of energy fuel and special substrates necessary for the cells to function at that particular state (metabolic constraints), the cells will manifest a particular metabolic phenotype. Therefore, understanding nutrient-gene interaction contributing to cancer risk requires tools for both the study of nutrigenomics and characterization of phenotype.

Metabolomics is a powerful tool for characterizing metabolic phenotypes. Tracer-based metabolomics provides a relational database of metabolites linked by the relationship of metabolic pathways, common substrates, and cofactors. Tracer-based metabolomics permits metabolic fluxes of multiple pathways to be simultaneously determined in a single experiment. Such a collection of flux measurements provides precise and accurate information on the operation of the cellular metabolic network and its response to genetic and nutrient environment changes in terms of a flux solution space. Nutrient-gene interaction can be studied using the concept of constraint-based modeling promoted by Covert and Palsson (10), which states that the observed metabolic phenotype (flux solution space) is a consequence of constraints from genetic factors and substrate environment. Genetic inheritance (genomic constraints) confers a set of possible phenotypes. Selection by metabolic (structural and pathway relationship) and environmental (physical environment and substrate availability) constraints determines the final observed phenotype. Examples of genetic and metabolic constraints on cancer cell phenotypes were previously reviewed (25). Dietary changes such as increased consumption of fat and inadequate consumption of vegetable have been recognized as important dietary factors in determining cancer risks of the gastrointestinal tract. However, the increased risks are not attributable to the presence of dietary mutagens or carcinogens. We propose that these dietary changes alter the metabolic milieu in the intestinal tract, thus permitting the selection for survival of the cancer cells. In conclusion, understanding the contribution of nutrient and genetic factors to the survival advantage of cancer cells using flux measurements is a critical first step in our study of the relationship between nutrient intake and cancer risk.

FOOTNOTES

1 Published in a supplement to The Journal of Nutrition. Presented as part of the International Research Conference on Food, Nutrition, and Cancer held in Washington, DC, July 14–15, 2005. This conference was organized by the American Institute for Cancer Research and the World Cancer Research Fund International and sponsored by (in alphabetical order) California Avocado Commission; California Walnut Commission; Campbell Soup Company; The Cranberry Institute; Danisco USA, Inc.; The Hormel Institute; National Fisheries Institute; The Solae Company; and United Soybean Board. Guest editors for this symposium were Vay Liang W. Go, Ritva R. Butrum, and Helen A. Norman. Guest Editor Disclosure: R. R. Butrum and H. Norman are employed by conference sponsor American Institute for Cancer Research; V.L.W. Go, no relationships to disclose. Back

2 Author Disclosure: No relationships to disclose. Back

3 This work was supported in part by PHS M01-RR00425 of the General Clinical Research Unit and P01-CA42710 of the UCLA Clinical Nutrition Research Unit Stable Isotope Core. Back

LITERATURE CITED

1. Greenwald P, Kelloff GJ, Boone CW, McDonald SS. Genetic and cellular changes in colorectal cancer: proposed targets of chemopreventive agents. Cancer Epidemiol Biomarkers Prev. 1995;4:691-702.[Abstract]

2. German B, Young VR.. Genetic and cellular changes in colorectal cancer: proposed targets of chemopreventive agents. Allison SP Go VLG eds. Genetic and cellular changes in colorectal cancer: proposed targets of chemopreventive agents. Nutrition and genomics in metabolic issues of clinical nutrition. :243-261 Karger Basel, Switzerland.

3. Harrigan GG Goodacre R eds. Genetic and cellular changes in colorectal cancer: proposed targets of chemopreventive agents. Metabolic profiling: its role in biomarker discovery and gene function analysis. Kluwer Academic Publishers Boston.

4. Fiehn O. Metabolomics–the link between genotypes and phenotypes. Plant Mol Biol. 2002;48:155-171.[Medline]

5. Hellerstein MK. In vivo measurement of fluxes through metabolic pathways: the missing link in functional genomics and pharmaceutical research. Annu Rev Nutr. 2003;23:379-402.[Medline]

6. Raamsdonk LM, Teusink B, Broadhurst D, Zhang N, Hayes A, Walsh MC, Berden JA, Brindle KM, Kell DB, et al. A functional genomics strategy that uses metabolome data to reveal the phenotype of silent mutations. Nat Biotechnol. 2001;19:45-50.[Medline]

7. Harrigan GG, Brackett DJ, Boros LG. Medicinal chemistry, metabolic profiling and drug target discovery: a role for metabolic profiling in reverse pharmacology and chemical genetics. Mini Rev Med Chem. 2005;5:13-20.[Medline]

8. Chace DH, DiPerna JC, Naylor EW. Laboratory integration and utilization of tandem mass spectrometry in neonatal screening: a model for clinical mass spectrometry in the next millennium. Acta Paediatr Suppl. 1999;88:45-47.[Medline]

9. Root M, Smith T. Prescribe by risk: the utility of a biomarker-based risk calculation in disease management to prevent heart disease. Dis Manag. 2005;8:106-113.[Medline]

10. Covert MW, Palsson BO. Constraints-based models: regulation of gene expression reduces the steady-state solution space. J Theor Biol. 2003;221:309-325.[Medline]

11. Katz J, Rognstad R.. The labeling of pentose phosphate from glucose-14C and estimation of the rates of transaldolase, transketolase, the contribution of the pentose cycle, and ribose phosphate synthesis. Biochemistry. 1967;6:2227-2247.[Medline]

12. Boros LG, Cascante M, Go VLW, Heber D, Hidvégi M, Lee WNP. Metabolic profiling of cell growth and death in cancer: applications in drug discovery. Drug Discovery Today. 2002;7:366-374.

13. Boros LG, Cascante M, Lee WN.. Stable Isotope-based dynamic metabolic profiling in disease and health. Harrigan GG Goodacre R eds. Stable Isotope-based dynamic metabolic profiling in disease and health. Metabolic profiling: Its role in biomarker discovery and gene function analysis. :141-169 Kluwer Academic Publishers Boston.

14. Boros LG, Serkova NJ, Cascante MS, Lee W-N P. Use of metabolic pathway flux information in targeted cancer drug design. Drug Discovery Today: Therapeutic Strategies. 2004;1:435-443.

15. Neese RA, Schwarz JM, Faix D, Turner S, Letscher A, Vu D, Hellerstein MK. Gluconeogenesis and intrahepatic triose phosphate flux in response to fasting or substrate loads. Application of the mass isotopomer distribution analysis technique with testing of assumptions and potential problems. J Biol Chem. 1995;270:14452-14466.[Abstract/Free Full Text]

16. Previs SF, Fernandez CA, Yang D, Soloviev MV, David F, Brunengraber H. Limitations of the mass isotopomer distribution analysis of glucose to study gluconeogenesis. Substrate cycling between glycerol and triose phosphates in liver. J Biol Chem. 1995;270:19806-19815.[Abstract/Free Full Text]

17. Mao CS, Bassilian S, Lim SK, Lee WN.. Underestimation of gluconeogenesis by the [U-(13)C(6)]glucose method: effect of lack of isotope equilibrium. Am J Physiol Endocrinol Metab. 2002;282:E376-E385.[Abstract/Free Full Text]

18. Hellerstein MK. Methods for measurement of fatty acid and cholesterol metabolism. Curr Opin Lipidol. 1995;6:172-181.[Medline]

19. Lee WP, Edmond J, Bassilian S, Morrow J. Mass isotopomer study of glutamine oxidation and synthesis in primary culture of astrocytes. Dev Neurosci. 1996;18:469-477.[Medline]

20. Lee WP, Lim S, Bassilian S, Bergner EA, Edmond J. Troglitazone stimulates peroxisomal b-oxidation in human hepatoma cells. J Biol Chem. 1998;273:20929-20934.[Abstract/Free Full Text]

21. Lee WNP, Boros LG, Puigjaner J, Bassilian S, Lim S, Cascante M.. Investigation of the pentose cycle using [1,2-13C2]-glucose and mass isotopomer analysis: estimation of transketolase and transaldolase activities. Am J Physiol. 1998;274:E843-E851.[Medline]

22. Boros LG, Torday JS, Lim S, Bassilian S, Cascante M, Lee WN.. Transforming growth factor beta2 promotes glucose carbon incorporation into nucleic acid ribose through the nonoxidative pentose cycle in lung epithelial carcinoma cells. Cancer Res. 2000;60:1183-1185.[Abstract/Free Full Text]

23. Boren J, Marin S, Cascante M, Lim S, Bassilian S, Syed A, Lee W-NP.. STI571 decreases glucose derived nucleic acid synthesis but increases direct glucose oxidation and lactate production in K562 myeloid blast cells. J Biol Chem. 2001;276:37747-37753.[Abstract/Free Full Text]

24. Boros LG, Bassilian S, Lim S, Lee WNP. Genistein inhibits non-oxidative ribose synthesis in MIA pancreatic adenocarcinoma cells: a new mechanism of controlling tumor growth. Pancreas. 2001;22:1-7.[Medline]

25. Boros LG, Lee WNP, Go VLW. A metabolic hypothesis of cell growth and death in pancreatic cancer. Pancreas. 2002;24:26-33.[Medline]

26. Warburg O.. A metabolic hypothesis of cell growth and death in pancreatic cancer. The metabolism of tumors. Costable London.

27. Horecker BL. Pathways of carbohydrate metabolism and their physiological significance. Journal of Chemical Education. 1965;42:244-253.[Medline]

28. Vizan P, Boros LG, Figuerasy A, Capellay G, Mangues R, Bassilian S, Lim S, Lee WNP, Cascante M.. K-ras codon-specific mutations produce distinctive metabolic phenotypes in human fibroblasts. Cancer Res. ; (in press).

29. Martin AM, Weber BL. Genetic and hormonal risk factors in breast cancer. J Natl Cancer Inst. 2000;92:1126-1135.[Abstract/Free Full Text]

30. Boros LG, Williams RD.. Isofenphos induced metabolic changes in K562 myeloid blast cells. Leuk Res. 2001;25:883-890.[Medline]

31. Boros LG, Steinkamp MP, Fleming JC, Lee WNP, Neufeld EJ. Defective RNA ribose synthesis in thiamine-responsive megaloblastic anemia (TRMA): mechanism for the syndrome. Blood. 2003;102:3556-3561.[Abstract/Free Full Text]

32. Boros LG. Population thiamine status and varying cancer rates between western, Asian and African countries. Anticancer Res. 2000;20:2245-2248.[Medline]

33. Lee WNP, Cascante M, Go VLG, Guo PZ, Boros LG.. Glycogen phosphorylase inhibitor CP-320626 inhibits pentose cycle metabolism, glycolysis and fatty acid synthesis with a subsequent cycle arrest and apoptosis in pancreatic tumor cells. Br J Cancer. 2004;91:2094-2100.[Medline]

34. Spitz DR, Sim JE, Ridnour LA, Galoforo SS, Lee YJ. Glucose deprivation-induced oxidative stress in human tumor cells. A fundamental defect in metabolism?. Ann N Y Acad Sci. 2000;899:349-362.[Abstract/Free Full Text]

35. Boren J, Lee W-N P, Bassilian S, Lim S, Centelles JJ, Syed A, Boros LG, Cascante M.. The stable isotope-based dynamic metabolic profile (SIDMAP) of butyrate induced HT29 cell differentiation. J Biol Chem. 2003;278:28395-28402.[Abstract/Free Full Text]

36. Johnson IT. Cancers of the gut and Western ills. Science. 2005;307:1839.[Abstract]





This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Purchase Article
Right arrow View Shopping Cart
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Lee, W.-N. P.
Right arrow Articles by Go, V. L. W.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Lee, W.-N. P.
Right arrow Articles by Go, V. L. W.

Home Help [Feedback] [For Subscribers] [Archive] [Search] [Contents]