The Metabolome: A Window on Cell Physiology & Portal to Understanding Complex Biological Systems, Diseases and Therapies

Web Sites

Duke University Center for Pharmacometabolomics
Home of the Interdisciplinary Research Consortium for Metabolomics.

Inborn Errors of Metabolism in Infancy and Early Childhood: An Update
Recent review of metabolic disorders published in American Family Physician.

Metabolomics
Research journal published by the Metabolomics Society.

Metabolomics Day in the Life: Chris Beecher
Metabolomics Day in the Life: Kyu Y. Rhee
Researcher profiles posted at the Web site of Agilent Technologies, Inc., a maker of bio-analytical instruments commonly used in metabolomics research.

Metabolomics Society
Scientific organization, composed of over 500 members in 20 countries, dedicated to promoting the growth, use, and understanding of metabolomics in the life sciences.

Metabolon, Inc.
Diagnostic products and services company that utilizes a patented biochemical profiling platform to provide global analysis of complex biological samples for the discovery of markers and pathways associated with drug action and disease.

METLIN Metabolite Database
A repository for mass spectral metabolite data, developed as a collaborative effort between the Siuzdak and Abagyan groups and Center for Mass Spectrometry at The Scripps Research Institute.

National Newborn Screening and Genetics Resource Center
Provides information and resources in the area of newborn screening and genetics to benefit health professionals, the public health community, consumers and government officials.

What is Mass Spectrometry?
Primer on technical and historical aspects of mass spectrometry, a critical tool for metabolomics.

Journal Articles

Steven Gross

Hao G, Gross SS. 2006. Electrospray tandem mass spectrometry analysis of S- and N-nitrosopeptides: facile loss of NO and radical-induced fragmentation. J. Am. Soc. Mass Spectrom. 17: 1725-1730.

Hao G, Wang D, Gu J, et al. 2009. Neutral loss of isocyanic acid in peptide CID spectra: a novel diagnostic marker for mass spectrometric identification of protein citrullination. J. Am. Soc. Mass Spectrom. 20: 723-727.

Nuriel T, Deeb RS, Hajjar DP, et al. 2008. Protein 3-nitrotyrosine in complex biological samples: quantification by high-pressure liquid chromatography/electrochemical detection and emergence of proteomic approaches for unbiased identification of modification sites. Methods Enzymol. 441: 1-17.

Upmacis RK, Crabtree MJ, Deeb RS, et al. 2007. Profound biopterin oxidation and protein tyrosine nitration in tissues of ApoE-null mice on an atherogenic diet: contribution of inducible nitric oxide synthase. Am. J. Physiol. Heart Circ. Physiol. 293: H2878-H2887. Full Text.

Gary Siuzdak

Benton HP, Wong DM, Trauger SA, et al. 2008. XCMS2: processing tandem mass spectrometry data for metabolite identification and structural characterization. Anal. Chem. 80: 6382-6389. Full Text

Crews B, Wikoff WR, Patti GJ, et al. 2009. Variability analysis of human plasma and cerebral spinal fluid reveals statistical significance of changes in mass spectrometry-based metabolomics data. Anal. Chem. 81: 8538-8544.

Patti GJ, Woo HK, Yanes O, et al. 2010. Detection of carbohydrates and steroids by cation-enhanced nanostructure-initiator mass spectrometry (NIMS) for biofluid analysis and tissue imaging. Anal. Chem. 82: 121-128.

Want EJ, Nordstrom A, Morita H, et al. 2007. From exogenous to endogenous: the inevitable imprint of mass spectrometry in metabolomics. J. Proteome Res. 6: 459-468.

Woo HK, Go EP, Hoang L, et al. 2009. Phosphonium labeling for increasing metabolomic coverage of neutral lipids using electrospray ionization mass spectrometry. Rapid Commun Mass Spectrom. 23: 1849-1855.

Yanes O, Woo HK, Northen TR, et al. 2009. Nanostructure initiator mass spectrometry: tissue imaging and direct biofluid analysis. Anal. Chem. 81: 2969-2975. Full Text

Rima Kaddurah-Daouk

Kaddurah-Daouk R, Kristal BS, Weinshilboum RM. 2008. Metabolomics: a global biochemical approach to drug response and disease. Annu. Rev. Pharmacol. Toxicol. 48: 653-683. Full Text

Members MSIB, Sansone SA, Fan T, et al. 2007. The metabolomics standards initiative. Nat. Biotechnol. 25: 846-848.

Paige LA, Mitchell MW, Krishnan KR, et al. 2007. A preliminary metabolomic analysis of older adults with and without depression. Int. J. Geriatr. Psychiatry 22: 418-423.

Patkar AA, Rozen S, Mannelli P, et al. 2009. Alterations in tryptophan and purine metabolism in cocaine addiction: a metabolomic study. Psychopharmacology (Berl) 206: 479-489.

Quinones MP,Kaddurah-Daouk R. 2009. Metabolomics tools for identifying biomarkers for neuropsychiatric diseases. Neurobiol. Dis. 35: 165-176.

Yao JK, Dougherty GG, Jr., Reddy RD, et al. 2009. Altered interactions of tryptophan metabolites in first-episode neuroleptic-naive patients with schizophrenia. Mol. Psychiatry [Advance online publication 28 April 2009.] Full Text

Joshua Rabinowitz

Bennett BD, Yuan J, Kimball EH, et al. 2008. Absolute quantitation of intracellular metabolite concentrations by an isotope ratio-based approach. Nat. Protoc. 3: 1299-1311. Full Text

Lu W, Bennett BD, Rabinowitz JD. 2008. Analytical strategies for LC-MS-based targeted metabolomics. J. Chromatogr. B. Analyt. Technol. Biomed. Life Sci. 871: 236-242. Full Text

Lu W, Kimball E, Rabinowitz JD. 2006. A high-performance liquid chromatography-tandem mass spectrometry method for quantitation of nitrogen-containing intracellular metabolites. J. Am. Soc. Mass Spectrom. 17: 37-50.

Munger J, Bajad SU, Coller HA, et al. 2006. Dynamics of the cellular metabolome during human cytomegalovirus infection. PLoS Pathog. 2: e132. Full Text

Munger J, Bennett BD, Parikh A, et al. 2008. Systems-level metabolic flux profiling identifies fatty acid synthesis as a target for antiviral therapy. Nat. Biotechnol. 26: 1179-1186.

Yuan J, Bennett BD, Rabinowitz JD. 2008. Kinetic flux profiling for quantitation of cellular metabolic fluxes. Nat. Protoc. 3: 1328-1340. Full Text

Kyu Rhee

Blumenthal A, Isovski F, Rhee KY. 2009. Tuberculosis and host metabolism: ancient associations, fresh insights. Transl Res. 154: 7-14.

Bryk R, Gold B, Venugopal A, et al. 2008. Selective killing of nonreplicating mycobacteria. Cell Host Microbe 3: 137-45. Full Text

Rhee KY, Erdjument-Bromage H, Tempst P, et al. 2005. S-nitroso proteome of Mycobacterium tuberculosis: Enzymes of intermediary metabolism and antioxidant defense. Proc. Natl. Acad. Sci. USA 102: 467-472. Full Text

Chris Beecher

Glassbrook N, Beecher C, Ryals J. 2000. Metabolic profiling on the right path. Nat. Biotechnol. 18: 1142-1143.

Rozen S, Cudkowicz ME, Bogdanov M, et al. 2005. Metabolomic analysis and signatures in motor neuron disease. Metabolomics 1: 101-108. Full Text

Sreekumar A, Poisson LM, Rajendiran TM, et al. 2009. Metabolomic profiles delineate potential role for sarcosine in prostate cancer progression. Nature 457: 910-914. Full Text

Sponsorship

This event is part of the Dr. Paul Janssen Memorial Series at the New York Academy of Sciences.

Presented by

Supported by an educational grant from Talecris Biotherapeutics, Center for Science and Education.

This program is also supported by an educational grant from Celgene Corporation.

Academy Friend

• Agilent Technologies
• Bristol-Myers Squibb R&D
• Umetrics

Over the past 20 years or so, the convergence of important scientific advances, such as DNA sequencing, widespread use of physical methods such as mass spectrometry and NMR, and large scale computational methods, has made it possible to develop the 'omics: genomics, proteomics, and now, metabolomics. These disciplines aim to understand the full complement of genes, proteins, or, in the case of metabolomics, small molecule metabolites found in a cell, tissue, organ, or organism. Until recently, genomics and proteomics have gotten all the attention, but there are many questions that these types of studies cannot answer. For example, they do not generally reveal the effects of protein post-translational modifications, nor can they elucidate the regulatory effects of small molecules on gene or protein functions within the living cell. For these questions, it is necessary to turn to metabolomics, which seeks to answer many of the same questions as traditional biochemistry, but on a much larger scale. As Steven Gross of Weill Cornell Medical College puts it, "Metabolomics represents the moment-to-moment phenotype of the cell," providing an important complement to our expanding knowledge of the genotype.

"Metabolomics represents the moment-to-moment phenotype of the cell."

Comprehensively identifying and quantitating all of the small molecules in a metabolome is not an easy task, however. Small molecule metabolites possess a much greater chemical diversity than nucleic acids or proteins, which are made up of a relatively small number of closely related building blocks. With time, researchers have developed a number of experimental metabolomics platforms with which to comprehensively study small molecules. Methods that are used to separate, identify, and quantitate small molecules include highly sophisticated applications of mass spectrometry, high performance liquid chromatography, capillary electrophoresis, nuclear magnetic resonance, and others. These methods take advantage of differences in molecule size, hydrophobicity, and charge characteristics.

Most researchers use multiple techniques that complement one another, thus avoiding biases inherent in each method and gaining a more comprehensive view of the molecules present. These methods must be very accurate, sensitive, and reproducible because researchers are looking both for compounds that are expected to be present, and compounds that were previously unknown. The large amount of data generated by metabolomics projects has also entailed the development of new software, statistical methods, and databases.

These endeavors are further complicated by the presence of small molecules that are not necessarily a consequence of normal metabolic processes. There is still some uncertainty about the exact number of small molecule metabolites normally found in human cells, but it is estimated that there are several thousand. Added to these are many more that may be related to the breakdown of xenobiotic compounds, for example, products of complex plant metabolites that are taken in by diet, or breakdown products of drugs and toxins.

On January 20, 2010, six veteran metabolomics investigators described their research, which has begun to bear fruit in the form of new insights into the metabolism of human cells in both health and disease. Steven Gross, of Weill Cornell Medical College, discussed his work developing metabolomic methods to detect inborn errors of metabolism in newborns. Gary Siuzdak of The Scripps Research Institute described work intended to uncover new therapeutic targets in the areas of chronic pain and multiple sclerosis. Rima Kaddurah-Daouk of Duke University discussed her work on establishing a metabolomics research infrastructure, as well as research results on CNS disorders such as depression and schizophrenia. Joshua Rabinowitz of Princeton University described new insights into the consequences of viral infection that have been gained from his work on metabolic fluxes in infected cells. Kyu Rhee of Weill Cornell Medical College has applied metabolomics methods to the discovery of new therapeutic targets for tuberculosis. And Chris Beecher of the University of Michigan presented his work applying metabolomics to the search for biomarkers and new therapeutic targets for prostate cancer.

The ultimate goal of these studies is to find places where interference with metabolic pathways might have a beneficial effect by stopping disease processes, as well as to generate clues that might lead to earlier detection or more accurate diagnoses. It seems that biology has come around full circle, returning to its early 20th century roots, where the emphasis was on understanding the underlying biochemistry of living processes. This time, however, instead of working on one pathway at time, researchers are striving to understand the interactions of hundreds of pathways and thousands of molecules, a very 21st century endeavor.

Speakers:
Steven Gross, Weill Cornell Medical College
Gary Siuzdak, The Scripps Research Institute
Rima Kaddurah-Daouk, Duke University Medical Center

Highlights

• Rapid metabolite profiling may one day be used to detect inborn errors of metabolism in newborns before significant developmental damage occurs.
• Increasing implementation of metabolomics has necessitated the development of new databases and software tools designed to catalog and analyze metabolomic information.
• Metabolomics can identify changes in metabolites in complicated diseases such as chronic pain, multiple sclerosis, depression, and schizophrenia.
• These changes may represent new leads for the development of therapeutics.
• Metabolomics may also shed light on differing responses to drugs between individuals.

Identifying the metabolic phenotype

Steven Gross and his colleagues at Weill Cornell Medical College began studying metabolomics as a result of their work on a particular type of protein modification, known as S-nitrosylation, that is an important component of the nitric oxide (NO) mediated signaling pathways within the cell. After cataloging the proteins that carry out this modification, describing their regulation, and identifying the sites that are modified on target proteins, they turned their attention to achieving a better understanding of how NO signaling actually alters the cell's metabolic state, in essence, its phenotype.

In order to obtain a more comprehensive understanding of this phenotype, they developed a metabolomics research platform. One of the critical features of this platform, and indeed a critical feature of all metabolomic platforms, is the ability to identify the metabolites that are present in the cell in an untargeted, unbiased way. This feature allows investigators to identify both previously unknown metabolites, and known metabolites whose roles in a given pathway were previously unsuspected. The objective of metabolomics, as an 'omic discipline, is to obtain as complete a description as possible of the compounds present, how they interact with each other, and how they influence the activities of genes and proteins.

As it turns out, this untargeted methodology is also highly applicable to the study of human disorders that arise from inborn errors of metabolism. Identification of such metabolic defects in newborns is an extremely important task, since deleterious consequences can occur quickly due to the build-up of unmetabolized intermediate compounds or due to a lack of critical building blocks, particularly in the rapidly developing brain. Methods to identify metabolic abnormalities in newborns must be fast, accurate, and use very small blood samples, since newborns contain on average only about 85 milliliters of blood. Hundreds of inborn errors of metabolism have been identified and more continue to be discovered. Because of the time and expense involved, only 8% of newborns are currently tested for the full complement of known defects, let alone tested for unknown metabolic disorders.

To develop new methods of testing for these disorders, Gross and his coworkers have been pursuing proof of principle studies in mice, utilizing very small samples: 1 microliter of blood. They are studying mice with a well known defect in purine catabolism, a mutation that inactivates the enzyme xanthine oxidoreductase. To test their ability to recognize this defect using their metabolomics methods, Gross's team compared blood samples from mice that were wild type, homozygous, or heterozygous for the mutation, and mice in which the metabolic defect was created biochemically with the inhibitory compound allopurinol.

Aqueous neutral phase chromatography and reverse phase chromatography were used to identify 3715 unique plasma features in wild-type and xanthine oxidoreductase (XOR-null) mutant mice.

They have identified around 3700 separate metabolic components in mouse plasma, some of which clearly differ between the wild-type, mutant, and allopurinol-treated mice. Some of the observed changes were expected based on previous knowledge of this pathway, but there were also a number of unanticipated changes which led them to uncover previously unrecognized defects in the mutant mice. In addition to identifying the direct effects of the mutation, they have uncovered a widening pool of metabolic perturbations that extends to many different pathways, resulting in previously unpredicted systemic consequences.

They have now begun to apply these methods to the analysis of blood samples from newborns diagnosed with inborn errors of metabolism. If their methods can be adapted to the needs of commercial testing laboratories, these methods are likely to make diagnoses faster and more accurate, and may allow many more newborns to be tested for these potentially debilitating disorders.

Mass spectrometry-based therapeutic metabolomics from tissues and biofluids

Gary Siuzdak and his coworkers at the Scripps Research Institute have been deeply involved in the development and application of metabolomics, including the establishment of the METLIN database, which compiles data on the identities and characteristics of metabolites as they are discovered, and XCMS, an open source software package that can be used to search tandem mass spectrometry data against data from known metabolites found in METLIN. A new version, XCMS(2), can also provide structural information on unknown metabolites by searching for similarities to known compounds. The METLIN database currently contains information on over 25,000 molecules. About 2500 of these represent endogenous metabolites, 4500 represent exogenous metabolites, 8000 are di- and tri-peptides, 8000 are various forms of lipids, and another 2000 are natural products.

Siuzdak is putting his highly developed metabolomic methods to use to investigate the complement of metabolites involved in the difficult-to-treat clinical syndromes of chronic pain and multiple sclerosis. He and his coworkers have developed not just solution-based methods but also a tissue-based form of mass spectrometry that can provide spatial information about the distribution of important metabolites in intact tissues. This imaging technique is called nanostructure-initiator mass spectrometry.

Chronic pain is currently a widespread, often intractable clinical problem for which treatments are very limited. Its debilitating nature leads to a high cost for both individuals and society, in terms of lost productivity and quality of life. In a search for new therapeutic avenues, Siuzdak is using a mouse model of chronic pain created by transection of a nerve in the tibial region of the leg. Because chronic pain appears to be promoted by changes in the activity of nerves where they connect to the spinal column, they have examined the dorsal root ganglia of that region for changes in its metabolic profile. They have identified hundreds of metabolic changes, providing multiple new potential targets for therapeutic intervention.

Similarly, they have compared normal spinal column tissue to tissue with both active and dormant lesions of multiple sclerosis, a disease that is characterized by gradual destruction of the myelin sheath surrounding nerves. Here they found only about 20 to 30 metabolites that were different in the lesions when compared to healthy tissue. In this investigation, their tissue-based method, known as nanostructure initiated mass spectrometry, has provided important new insights into the disease processes that promote lesion development. These insights have provided new leads to potential drug targets for multiple sclerosis. Siuzdak and his group plan to continue to apply their methods to the understanding of the differences between healthy and diseased tissues, in an ongoing search for new clinical targets.

Metabolite profiling for personalized medicine

Another long time researcher in the field, Rima Kaddurah-Daouk of Duke University, spoke of her efforts to develop an infrastructure for the burgeoning metabolomics discipline. Kaddurah-Daouk was intimately involved in the creation of the Metabolomics Society, an organization intended to promote communication and collaboration among researchers working in this area. She has also been instrumentally involved in the creation of an NIH-funded metabolomics research network known as the Metabolomics Network for Drug Response Phenotype. This national research network, funded by the National Institute for General Medical Sciences (NIGMS) includes 15 linked research sites, mostly in the U.S., but also internationally represented at sites in Canada and the Netherlands. The consortium is headquartered at Duke University. By pooling their resources, the research groups involved hope to make important headway in major clinical areas, including depression, hyperlipidemia, antiplatelet therapy, and antihypertensives. They are developing core facilities that will serve as shared resources, including those for metabolomics instrumentation, bioinformatics, and database administration.

Kaddurah-Daouk explained some of the fundamental assumptions behind her work on the metabolic bases of disease susceptibility and differential response to therapeutic agents. She described how each person has a distinct metabolomic signature, or metabotype, that may explain differences in susceptibility to diseases and responses to medications. This metabolic signature is the manifestation of both genetic and environmental influences. Physicians now use a crude version of the metabotype in the form of the 5 to 10 metabolites that are routinely measured in serum, such as glucose, creatinine, and others, as a part of the diagnostic process. Kaddurah-Daouk envisions a greatly enhanced version of this testing that provides information on hundreds of metabolites, in a comprehensive report on health and disease status.

Each disease is likely to have a specific metabolomic signature.

Moreover, each disease is also likely to have a specific metabolomic signature, resulting from its effects on the metabolic processes of cells, tissues, and organs. A more comprehensive understanding of disease metabolomic signatures would lead to a better understanding of the mechanisms of action of current and emerging drug therapies, and might lead to the discovery of new therapies. Added to an understanding of the role of each person's metabotype, this knowledge could also help explain why some individuals experience side effects from certain medications while others do not.

The metabolomic signatures of diseases may also be used one day to classify diseases more intelligently. This is particularly important in areas such as psychiatric disease, where syndromes that appear very similar to outside observation may have widely varying underlying causes, often manifesting as highly variable responses to treatment. Depression is one such disorder, and Kaddurah-Daouk described considerable work from her laboratory and others on the metabolomic differences between individuals with this and other CNS disorders.

Kaddurah-Daouk has recently received NIH funding, under the federal economic stimulus act, to create a partnership between the Metabolomics Network for Drug Response Phenotype and two centers of excellence within the Pharmacogenomics Research Network, for the purpose of gaining deeper understanding of the mechanisms of variation in response to commonly used cardiovascular drugs, including antihypertensives and antiplatelet therapies.

In order for this vision of the uses of metabolomics to come to fruition, Kaddurah-Daouk stated that it will be necessary for the metabolomics community to develop critical resources, including the adoption of standards, a national metabolomics database and a metabolic pathway database that couples information about human metabolic pathways with information about the human genome. Metabolomics has the potential to substantially improve clinical assessments in health and disease, to provide information on individual variation in drug response that will lead to truly personalized medicine, and to enhance our knowledge of disease mechanisms leading to the development of more effective therapeutic interventions.

Speakers:
Joshua Rabinowitz, Princeton University
Kyu Rhee, Weill Cornell Medical College and New York Presbyterian Hospital
Chris Beecher, University of Michigan Medical School

Highlights

• Some viruses hijack their host cell's metabolic pathways, forcing them to produce large amounts of the components needed to make new viral particles.
• Metabolic flux labeling adds the dimension of time to metabolomics.
• Metabolomics can identify new pathways in bacteria that may serve as targets for future antimicrobial drugs.
• These new targets will be especially important for combating tuberculosis and other infections where resistance is rapidly eroding therapeutic options.
• Metabolomics can be used to identify new biomarkers and new drug targets in oncology.

Viral hjacking of host cell metabolism

Viral infections are an interesting case in point when it comes to our understanding of the metabolomic signatures of diseases. Although quite a bit is known about the effects of viral infection on gene and protein expression, very little is known about the effects of viral infection on host cell metabolic processes. Joshua Rabinowitz of Princeton University described his work in this area, which has overturned some previously held assumptions.

The mainstream view of viruses has been that they take advantage of host genes and proteins, and in some cases take a few lipid molecules with them when they leave by budding off. Although viruses need nucleotides and amino acids to replicate, as well as sources of energy, it was thought that they could get these materials simply by depleting existing host cell pools of metabolites. Rabinowitz and his colleagues investigated the idea that viruses might actually require very high levels of some metabolites, leading to a need to actively hijack the host cell's metabolic pathways.

In order to address this question, they are investigating the metabolic changes occurring in cells infected by several different viral species, including human cytomegalovirus (CMV), influenza A, and herpes simplex. In some cases, they have identified large changes in host cell metabolism as a result of these infections. In addition to shedding new light on the processes of viral infection, knowledge of these changes could lead to new and possibly superior therapeutic approaches to combating viral infections. In particular, therapeutic approaches based on metabolic changes might be able to circumvent the development of resistance, since such approaches would not necessarily depend on a specific gene or protein product. Most forms of antiviral resistance occur as a result of the acquisition of a mutation that changes the structure of the targeted gene or protein.

CMV infection greatly increased intracellular pools of metabolites.

Rabinowitz found that human CMV infection dramatically changed the metabolome of host cells. Instead of depleting intracellular pools of metabolites as expected, CMV infection greatly increased these pools. This observation led to an important question: did these increases result because the virus had used up all of a particular metabolite, preventing downstream metabolic pathways from proceeding and leading to the build up of intermediate metabolites? Or was the virus actively directing host metabolic pathways to increase their output? In order to answer this question, he and his group devised a method to measure the flow of metabolic processes in infected cells, using 13C-labeled precursor compounds such as glucose and glutamine. The method, which is called kinetic flux profiling, added the parameter of time to their understanding of the infected cell's metabolome.

Using this method, they found that infection with human CMV markedly upregulated flux through much of the central carbon metabolism of the host cell, including such central processes as glycolysis. Metabolic flux became much faster with CMV infection, rather than being blocked up, suggesting that the virus had actively hijacked host metabolic pathways. Rabinowitz and his group have followed clues provided by identities of the elevated metabolites to determine which pathways have been hijacked and through what mechanism, leading to considerable new insight into the effects of CMV on infected cells.

Their work with herpes simplex provided similar results, which is not unexpected since CMV is also a member of the herpes virus family. However, cells infected with influenza A showed very different results. In these cells, very little new metabolism was turned on and longer infection times were instead associated with greatly reduced concentrations of many metabolites. A distinct few metabolites showed raised levels; interestingly, these belonged to pathways which included known protein targets of viral replication inhibitors. Rabinowitz suggested that measuring metabolic flux may become an important method for identifying new antiviral targets, possibly including targets that would be less susceptible to the development of antiviral resistance.

Mining the mycobacterial metabolome

Continuing on the theme of resistance, Kyu Rhee of Weill Cornell Medical College noted that resistance to therapeutic agents is an important issue in the fight against bacterial infections as well. Antibacterial resistance is so prevalent and widespread that researchers must continually look for new antibiotic targets. Current antibiotics take advantage of only a very small portion of metabolic space, since almost all target one of four processes: cell wall biosynthesis, protein biosynthesis, nucleic acid synthesis, or maintenance of cell membrane potential. Rhee and his colleagues hope to take advantage of a much larger metabolic space by discovering new antibacterial targets with the use of metabolomics.

As a specific target, Rhee and his laboratory have long been engaged in the pursuit of antibiotics targeting Mycobacterium tuberculosis, the causative organism of tuberculosis. Tuberculosis is the leading cause of death due to bacterial infection worldwide, and represents a growing public health crisis because of the development of resistant strains. In recent years, both multidrug resistant and extremely resistant strains have developed and spread, while at the same time, the last antibiotic that was developed for tuberculosis was approved in the U.S. over 40 years ago. Rhee pointed out that, despite these problems, tuberculosis is a potentially eradicable disease, because much like small pox, which has been eradicated, humans are its only reservoir.

The metabolism of M. tuberculosis is quite different from that of a human cell.

One of the challenges of treating tuberculosis is that M. tuberculosis lives most of its life inside a phagocytic vacuole within infected tissues, which can make it difficult for antibacterial compounds to reach and kill it. However, this lifestyle also presents opportunities because conditions inside the vacuole dictate that the organism make considerable adaptive changes in its metabolism in order to sustain life. These changes make its metabolism quite different from that of a human cell, presenting the possibility of developing agents that kill M. tuberculosis without causing harm to human host cells. This could be achieved by selectively targeting enzymes that are needed for M. tuberculosis to function but not for human metabolism.

In pursuing this possibility, Rhee and his group have produced a metabolomic library for M. tuberculosis that represents the convergence of three independently produced libraries. This was no small feat, since it required the growth of large volumes of M. tuberculosis cultures under very strict biohazard controls. They have identified about 1700 reproducible features of this library as well as specific metabolic pathways that appear to be important for M. tuberculosis' intravesicular lifestyle. They have begun to elucidate important aspects of the tricarboxylic acid cycle, a central component of energy metabolism that is not well understood in this organism. They have identified new enzymes and reassessed the functions of previously known ones. This untargeted exploration of metabolic pathways and protein function should lead to the discovery of important new targets for eradicating the global epidemic of tuberculosis.

Metabolite profiling for discovery of cancer biomarkers

Chris Beecher of the University of Michigan Medical School presented his work on one of the potentially most fruitful applications of metabolomics: the discovery of new biomarkers and new therapeutic targets in the war on cancer. He and his colleagues are comparing the metabolomic profiles of plasma, urine, and prostate tissue from men with early stage or advanced prostate cancer, and from benign tissues adjacent to the cancer, in an effort to identify differences and similarities that may lead to therapeutic insights.

A comparison of prostate tissue samples shows the differences between normal and cancerous tissues. PCA=low-grade tumor tissue, Benign=benign adjacent tissue, Mets=metastatic tissue.

What they have found is that "the chemistry of the cancerous tissue is overwhelmingly different" from normal tissues, Beecher said. Their studies have identified numerous metabolic pathways that are altered in the cancerous tissue. Because of the complexity of the data, Beecher and his group are using molecular concept mapping to assist in exploring the relationships between these pathways, and in determining their possible significance to the growth and development of prostate cancer.

Using these methods, they have identified sarcosine, a methylated derivative of the amino acid glycine, as a specific metabolite whose levels become highly elevated as prostate cancer progresses to metastasis. Several lines of evidence have now established that the pathway that produces sarcosine is closely associated with aggressive behavior and changes that might lead to metastasis in cultured prostate cancer cells. They found, for example, that the addition of exogenous sarcosine, or genetic knockdown of the enzyme that leads to sarcosine degradation, sarcosine dehydrogenase, causes benign prostate epithelial cells in culture to take on an invasive phenotype. They have identified both genetic and protein components of this pathway that appear to be involved in the process of becoming invasive. Importantly, sarcosine can be detected in urine, which would make it highly useful as a non-invasive biomarker for prostate cancer metastasis. It is may also be that this pathway will present a novel therapeutic target for drugs that inhibit growth or prevent metastasis of prostate cancer. These studies represent a highly promising beginning for the use of metabolomics as a tool in the fight against cancer.