TECHNOLOGIES
DRUG DISCOVERY
TODAY
Metabolomic biomarkers: their role in
the critical path§
Laura K. Schnackenberg, Richard D. Beger*
Division of Systems Toxicology, National Center for Toxicological Research, Food and Drug Administration, Jefferson, AR 72079-9502, United States
Global metabolic profiling is being applied to identify
biomarkers of health. Some small molecules are exquisitely
sensitive indicators of health status. Metabolic
profiling analyses are being used to determine biomarkers
of drug safety and effectiveness as well as disease
diagnosis and prognosis including organ transplant
rejection. To understand the mechanism(s) of drug
toxicity and disease, a systems biology approach that
considers the information generated from metabolic
profiling, genetics, transcriptomics and proteomics
research paradigms is necessary. This will allow for a
better understanding of the mechanism(s) of drug
interactions and disease while possibly identifying susceptible
populations, an important goal in the move
toward personalized medicine.
Section Editor:
Janet Woodcock ­ Food and Drug Administration, Rockville,
MD, USA
Introduction
Global metabolic profiling has long been used to determine
biomarkers to aid in assessment of the pathophysiological
health status of patients. The emergence of genomics and
proteomics technologies has generated plausible mechanisms
that correlate with the diagnostic and prognostic metabolic
biomarkers in the health to disease continuum [1,2].
Global metabolic profiling has been referred to as either
metabolomics [3] or metabonomics [4], though both identify
metabolic alterations under varying conditions. In this
review, metabolomics and metabonomics will be referred
collectively to as global metabolic profiling or simply, metabolic
profiling. Together, genetics, transcriptomics, proteomics
and global metabolic profiling comprise the basis of the
systems biology approach.
Systems biology has been described as the computational
integration of data generated by a suite of `omic' platforms to
understand function across different levels of biomolecular
organization [4­6]. These new research paradigms are likely
to lead to new opportunities for personalized health because
they add more detail to our current knowledge of the health
and disease continuum. A systems biology approach provides
a better understanding of the mechanisms and progression of
disease as well as the ability to identify early and sensitive
biomarkers of drug efficacy and toxicity. The ability to link
changes in the metabolite profile to altered genes and proteins
will help to elucidate the source of metabolite biomarkers
in many cases. As part of a systems biology approach,
global metabolic profiling will also help the medical community
understand the complex interaction between metabolites,
genes and proteins to overall health status.
Metabolomic analytical platforms
Metabolic profiling studies have been performed with a wide
range of biofluids and tissues [7]. The large concentration
range, over eight orders of magnitude and chemical diversity
of metabolites found in cells, tissues and biofluids, requires
multiple analytical methods to detect as many metabolites as
possible. In general, the most common metabolic profiling
Drug Discovery Today: Technologies Vol. 4, No. 1 2007
Editors-in-Chief
Kelvin Lam ­ Harvard Stem Cell Institute, Harvard University, USA
Henk Timmerman ­ Vrije Universiteit, The Netherlands
Critical path
§
The views presented in this article do not necessarily reflect those of the U.S. Food
and Drug Administration.
*Corresponding author: R.D. Beger (Richard.Beger@fda.hhs.gov)
1740-6749/$ .Published by Elsevier Ltd. DOI: 10.1016/j.ddtec.2007.10.012 13
technologies consist of high-resolution NMR and hyphenated
mass spectrometric methods with an initial chromatographic
separation step (e.g. LC­MS). The advantages and
disadvantages of NMR and MS platforms have been thoroughly
reviewed [8]. Initially, NMR was used primarily for the
investigation of metabolite changes as it is highly reproducible
and quantitative in nature and it detects each proton in
the same manner. Mass spectrometry is a technology that is
more frequently being applied in metabolic profiling studies
due in part to its high sensitivity. The combination of NMR
and MS platforms in metabolic profiling research will permit
a more broad coverage of the metabolome [9].
Metabolomics in drug discovery
One of the major areas of research in pharmaceutical drug
discovery is directed toward the identification of biomarkers
of drug toxicity that can be used in preclinical and clinical
studies of drugs [10­14]. There is a strong need to develop
new biomarkers that can accurately predict toxicity in the
preclinical development of new chemical entities (NCEs)
early in the drug development process. Metabolic profiling
has the potential to impact pharmaceutical drug discovery by
lowering both the cost and time associated with the development
and marketing of a NCE. The current estimate of the
cost for developing a NCE is approximately one billion dollars
[15]. Additionally, there is a high attrition rate of NCEs with
only one in five actually making it to market. Of those NCEs
that fail, approximately half of them fail in phase III clinical
testing. Failed NCEs increase drug discovery costs and therefore,
ultimately increase cost to the consumers. Time is an
additional consideration with it taking an average of eight to
ten years for a NCE to make it through the developmental and
FDA approval processes. Therefore, a technique such as metabolic
profiling that has the potential to identify toxicity early
in the drug discovery process will save time and money for
pharmaceutical companies. For example, metabonomics
methods were used in a preclinical study at Merck on a
compound known to cause hepatotoxicity in several species
[16]. Multivariate statistics of the NMR spectra of urine
showed the dosed group separated from the control group
with the depletion of tricarboxylic acid cycle intermediates
and the appearance of medium-chain carboxylic acid. In vitro
experiments with this compound showed that it causes
defective metabolism of fatty acids. This is a case where
metabonomics was able to provide mechanistic insight to
the hepatotoxicity of a drug.
In general, the most widely used biofluid for toxicity
studies has been urine, which is easily obtained from a subject
and provides information about the whole system following a
toxic insult [17­19]. One advantage of using urine or plasma
in drug toxicology experiments is that the sample is collected
noninvasively so that it can be applied in clinical studies.
Another advantage is that multiple biofluid samples from a
single subject can be collected over a time course, which
allows the determination of a metabolic trajectory that
describes the toxic response and recovery period. The analyses
of metabolites in biofluids permit a toxicological evaluation
of the `health of many different organs' within the
same animal over time and may permit the simultaneous
evaluation of drug efficacy from the same biofluid sample
[19]. Further, since the same animal can be used over many
time points, the number of animals needed for a toxicological
study is greatly reduced. This makes the metabolic profiling
method much more cost effective than many other biomarker
detection methods. The potential to determine biomarkers
from easily obtained, noninvasive samples also makes
preclinical findings accessible in the clinical setting.
Another benefit of using a temporal study with the same
animal is that it is not necessary to know the pharmacodynamics
and pharmacokinetics of the drug before the biomarker
investigation. This removes the need for a preparatory
pharmacokinetic research, which saves time and financial
resources. This is especially important in the early ADME-Tox
(absorption, distribution, metabolism and excretion-toxicology)
stage of drug discovery. Temporal metabolic profiling
studies may permit a quick determination as to whether the
toxic insult to the animal causes temporary metabolite concentration
changes that return to normal after a period of
time or results in metabolite concentrations that stay in a
perturbed toxic state. This information can be evaluated to
determine whether and when the animal recovers from the
toxic distress [20].
Metabolomics in personalized health
NMR and GC­MS methods have long been applied to detect
inborn errors of metabolism following birth [21]. Thus, some
of the first attempts to determine biomarkers of disease by
global metabolic profiling was applied to the study of inborn
errors of metabolism [22,23]. Genetic alterations in DNA
sequence by nucleotide deletion, insertion or single nucleotide
polymorphism can alter the enzymatic activity of a
protein that is responsible for converting one metabolite to
another. If no other enzymes are able to interact with a
particular metabolite, the concentration of the metabolite
can build up. In an effort to return to a homeostatic state, the
metabolite is exported from the cell to biofluids like serum
and urine. In cases of inborn errors of metabolism, the
metabolite biomarkers are diagnostic for a particular inborn
disease and knowledge of genetics can be used to understand
the link between altered metabolic pathway and disease state.
Another success for metabolic profiling has been its ability
to diagnose renal, liver and heart organ transplant rejections
better than measurements of standard clinical chemistry
parameters [24­26]. Recent metabolic profiling evaluations
of kidney transplants have revealed biomarkers that include
altered levels of trimethylamine-N-oxide (TMAO), dimethyDrug
Discovery Today: Technologies | Critical path Vol. 4, No. 1 2007
14 www.drugdiscoverytoday.com
lamine, lactate, acetate and alanine. In many of these investigations,
TMAO was increased by a factor of 3­4 compared to
healthy controls. The increase in TMAO is believed to stabilize
proteins when there is an increased concentration of
protein denaturants such as urea and guanidine derivatives
following a toxic insult to the kidney [27].
Metabolic profiling has been applied to diagnose and predict
the outcome of diabetes [28], cirrhosis [29] and cancer
[30,31] in preclinical and clinical studies. Many metabolites
are species-independent and could form the basis of translational
biomarkers that are determined in preclinical studies
and applied during clinical studies. Metabolic profiling of
urine, plasma, serum and cerebral spinal fluid has been
applied effectively in a clinical research environment for
the assessment of a range of health issues. Urine samples
have been evaluated to investigate the efficacy of immunosuppressants
in renal transplant [32] and to detect inborn
metabolic diseases [22,23]. Plasma or serum samples have
been used to evaluate differences in fat metabolism in lean
and obese patients [28], to detect motor neuron diseases like
amyotrophic lateral sclerosis [33], to detect pancreatic cancer
[31], to detect coronary heart disease [34] and to detect
ovarian cancer [30]. Finally, cerebral spinal fluid has been
used for the detection of meningitis and ventriculitis [35].
Each of these metabolic profiling studies was able to develop a
predictive relationship between the biofluid spectral patterns
and a health disease state in humans. Most diseases occur
later in life and have not only a genetic component but also
environmental contributions. The later the onset of a disease,
the more probably it is that contributions from the environment
played a larger and significant role. Once a disease can
be defined as a pattern of metabolites in tissue or biofluids,
global metabolic profiling can be used as a diagnostic tool.
Metabolomics and the critical path
In March 2006, FDA published the Critical Path Opportunities
Report and List as a follow up to FDA's initial Critical Path
Challenges and Opportunities Report that was released in
2004. The Opportunities Report and List presented 6 major topic
areas and 76 specific scientific opportunities. Metabolomics
can play a significant role in the major topics for developing
biomarkers, streamlining clinical trials and defining at-risk
populations. The ability of metabolomics to provide translational
safety biomarkers related to kidney, liver, heart and
vascular damage should allow it to play a major role in many
opportunities presented in Topic 1: better evaluation tools developing
new biomarkers and disease models. In Critical
Path Topic 2: streamlining clinical trials ­ the ability of
metabolomics to give noninvasive biomarkers of efficacy
and toxicity will facilitate the advancement of clinical trial
designs. By evaluating patient responses through pharmacometabonomic
phenotyping techniques, pharmaceutical
companies may be able to adjust which patients are used
in each stage of the trial through accurate prediction of
nonresponders and responders in terms of toxicity [36].
Finally, linking genetics, transcriptomics, proteomics, metabolic
profiling, nutrition and gut microflora in relation to
patient health to disease status is an essential component of
the FDA's critical path to personalized medicine. Metabolic
profiling has a potential to play a vital role in many facets of
the FDA's Critical Path.
Conclusions
Metabolic profiling is an essential component that along with
genetics, transcriptomics and proteomics data will permit a
detailed description of the interactions between metabolites,
proteins, transcripts and genes in the health and disease
continuum. In many errors of inborn metabolism, the relationship
between disease state, metabolic biomarker and
genetics is easily understood. However, in many diseases,
the relationship between health status, genetics and metabolic
state is highly complex and not easily determined. The
ability of metabolic profiling to provide noninvasive translational
biomarkers makes it an integral part of a systems
biology approach as well as in the move toward personalized
medicine and within the FDA Critical Path opportunities.
References
1 Fiehn, O. (2001) Combining genomics, metabolome analysis, and
biochemical modeling to understand metabolic networks. Comp. Funct.
Genom. 2, 155­168
2 German, J.B. et al. (2004) Metabolomics in the opening decade of the 21st
century: building the roads to individual health. J. Nutr. 134, 2729­2732
3 Fiehn, O. (2002) Metabolomics ­ the link between genotypes and
phenotypes. Plant Mol. Biol. 48, 155­171
4 Nicholson, J.K. et al. (1999) `Metabonomics': understanding the metabolic
responses of living systems to pathophysiological stimuli via multivariate
statistical analysis of biological NMR spectroscopic data. Xenobiotica 29,
1181­1189
5 Thomas, C.E. and Ganji, G. (2006) Integration of genomic and
metabonomic data in systems biology ­ are we `there' yet? Curr. Opin. Drug
Discov. Dev. 9, 92­100
6 Goodacre, R. et al. (2004) Metabolomics by numbers: acquiring and
understanding global metabolite data. Trends Biotechnol. 22, 245­252
7 Lindon, J.C. et al. (1999) NMR spectroscopy of biofluids. In Annual Reports
on NMR Spectroscopy, (Vol. 38) (Webb, G.A., ed.), pp. 2­88, Academic Press
8 Dunn, W.B. and Ellis, D.I. (2005) Metabolomics: current analytical
platforms and methodologies. Trends Anal. Chem. 24, 285­294
9 Lenz, E.M. and Wilson, I.D. (2007) Analytical strategies in metabonomics.
J. Proteome Res. 6, 443­458
10 Nicholson, J.K. et al. (2002) Metabonomics: a platform for studying drug
toxicity and gene function. Nat. Rev. Drug Discov. 1, 153­161
11 Robertson, D.G. (2005) Metabonomics in toxicology: a review. Toxicol. Sci.
85, 809­822
12 Portilla, D. et al. (2006) Metabolomic study of cisplatin-induced
nephrotoxicity. Kidney Int. 69, 2194­2204
13 Schnackenberg, L.K. et al. (2006) An integrated study of acute effects of
valproic acid in the liver using metabonomics, proteomics, and
transcriptomics platforms. OMICS 10, 1­14
14 Espandiari, P. et al. (2007) The utility of a rodent model in detecting
pediatric drug-induced nephrotoxicity. Toxicol. Sci. 99, 637­648
15 Harrigan, G.G. and Yates, L.A. (2006) High-throughput screening,
metabolomics and drug discovery. IDrugs 9, 188­192
Vol. 4, No. 1 2007 Drug Discovery Today: Technologies | Critical path
www.drugdiscoverytoday.com 15
16 Mortshire-Smith, R.J. et al. (2004) Use of metabonomics to identify
impaired fatty acid metabolism as the mechanism of a drug-induced
toxicity. Chem. Res. Toxicol. 17, 165­173
17 Lindon, J.C. et al. (2005) The Consortium for Metabonomic Technology
(COMET): aims, activities, and achievements. Pharmacogenomics 6, 691­699
18 Wilson, I.D. et al. (2005) HPLC-MS-based methods for the study of
metabonomics. J. Chromatogr. B: Analyt. Technol. Biomed. Life Sci. 817, 67­76
19 Bollard, M.E. et al. (2005) NMR-based metabonomic approaches for
evaluating physiological influences on biofluid composition. NMR
Biomed. 18, 143­162
20 Keun, H.C. et al. (2004) Geometric trajectory analysis of metabolic
responses to toxicity can define treatment specific profiles. Chem. Res.
Toxicol. 17, 578­587
21 Lehnert, W. and Hunkler, D. (1986) Possibilities of selective screening for
inborn errors of metabolism using high-resolution 1H-FT-NMR
spectrometry. Eur. J. Pediatr. 145, 260­266
22 Constantinou, M.A. et al. (2005) 1H NMR-based metabonomics for the
diagnosis of inborn errors of metabolism in urine. Anal. Chim. Acta 511,
303­312
23 Kuhara, T. (2005) Gas chromatographic­mass spectrometric urinary
metabolome analysis to study mutations of inborn errors of metabolism.
Mass Spectrom. Rev. 24, 814­827
24 Serkova, N. et al. (2005) 1H-NMR-based metabolic signatures of mild and
severe ischemia/reperfusion injury in rat kidney transplants. Kidney Int.
67, 1142­1151
25 Silva, M.A. et al. (2006) Hepatic artery thrombosis following orthotopic
liver transplantation: a 10-year experience from a single centre in the
United Kingdom. Liver Transpl. 12, 146­151
26 Wishart, D.S. (2005) Metabolomics: the principles and potential
applications to transplantation. Am. J. Transplant. 5, 2814­2820
27 Bell, J.D. et al. (1991) Nuclear magnetic resonance studies of blood plasma
and urine from subjects with chronic renal failure: identification of
trimethylamine-N-oxide. Biochim. Biophys. Acta 1096, 101­107
28 Griffin, J.L. and Nicholls, A.W. (2006) Metabolomics as a functional
genomic tool for understanding lipid dysfunction in diabetes, obesity and
related disorders. Pharmacogenomics 7, 1095­1107
29 Yang, J. et al. (2004) Diagnosis of liver cancer using HPLC-based
metabonomics avoiding false-positive result from hepatitis and
hepatocirrhosis diseases. J. Chromatogr. B 813, 59­65
30 Odunsi, K. et al. (2005) Detection of epithelial ovarian cancer using 1HNMR-based
metabonomics. Int. J. Cancer 113, 782­788
31 Beger, R.D. et al. (2006) Metabonomic models of human pancreatic cancer
using 1D proton NMR spectra of lipids in plasma. Metabolomics 2, 125­134
32 Holmes, E. et al. (1990) Proton NMR analysis of plasma from renal failure
patients: evaluation of sample preparation and spectral-editing methods.
J. Pharm. Biomed. Anal. 8, 955­958
33 Rozen, S. et al. (2005) Metabolomic analysis and signatures in motor
neuron disease. Metabolomics 1, 101­108
34 Brindle, J.T. et al. (2002) Rapid and noninvasive diagnosis of the presence
and severity of coronary heart disease using 1H-NMR-based
metabonomics. Nat. Med. 8, 1439­1445
35 Coen, M. et al. (2005) Proton nuclear magnetic resonance-based
metabonomics for rapid detection of meningitis and ventriculitis. Clin.
Infect. Dis. 41, 1582­1590
36 Clayton, T.A. et al. (2006) Pharmaco-metabonomic phenotyping and
personalized drug treatment. Nature 440, 1073­1077
Drug Discovery Today: Technologies | Critical path Vol. 4, No. 1 2007
16 www.drugdiscoverytoday.com