Curr Pharmacol Rep DOI 10.1007/s40495-017-0083-4

PHARMACOMETABOLOMICS AND TOXICOMETABOLMICS (C CHEN, SECTION EDITOR)

Metabolomics: Bridging Chemistry and Biology in Drug Discovery and Development Yuwei Lu 1 & Chi Chen 1

# Springer International Publishing AG 2017

Abstract Purpose of Review Metabolism, as a ubiquitous biological process, plays a decisive role in the wellbeing of cells, organs, and whole body. Since the endpoint of pharmacological intervention is to either recover or sustain the wellbeing of targeted biological systems, examining metabolic events reveals useful information on diseases and drug treatment. Based on its omics nature and capacity to examine both drug and endogenous metabolism, metabolomics is an ideal systems biology tool to examine chemical and metabolic events in drug discovery and development. Recent Findings Ongoing and potential applications of metabolomics in drug discovery and development include biomarker identification, bioactive compound identification, drug metabolism, metabolic profiling, and mechanistic investigation. Corresponding to four stages of drug discovery and development, why and how metabolomics is used in each stage of drug discovery and development are explained and discussed in four respective sections of this review. Summary Since metabolites, the direct targets of metabolomic analysis, are the end products of gene-, protein-, microbe-, and signal molecule-mediated biological processes, metabolomics possesses many functions that are irreplaceable by other system biology tools. Even though its application is still in the germination phase, metabolomics is a promising platform to

This article is part of the Topical Collection on Pharmacometabolomics and Toxicometabolomics * Chi Chen [email protected] 1

Department of Food Science and Nutrition, University of Minnesota, St. Paul, MN 55108, USA

bridge chemistry and biology in drug discovery and development. Keywords Metabolomics . Metabolite . Drug discovery . Drug development . Drug metabolism . Biomarker

Abbreviations NCE New chemical entity IND Investigational new drug APAP Acetaminophen MDA Multivariate data analysis PCA Principal components analysis

Introduction Owing to rapid progresses in analytical and informatics technology, metabolomics has become an indispensable component of systems biology in the past decade [1, 2]. Aiming to both qualitatively and quantitatively characterize endogenous and exogenous metabolites in a biological system, metabolomics has been adopted in diverse scientific fields, including the ones related to drug discovery and development, i.e., disease biomarker [3], xenobiotic metabolism [4], toxicology [5], and nutrition [6]. Based on experimental design and aims, metabolomics research could be categorized as either exploratory or hypothesis-driven investigation. Exploratory metabolomic investigation is capable of identifying novel and unexpected metabolic changes through nonbiased examinations and comparisons of samples from different treatments and health status, while hypothesis-driven metabolomic investigation, when combined with targeted biochemical analysis and specific experimental models, could reveal underlying

Curr Pharmacol Rep

mechanisms of metabolic changes by analyzing related upstream and downstream metabolic and regulatory pathways. Since metabolites in different biological samples reflect different metabolic events in the body and cells, metabolomics examines diverse biological samples, including cell extract, tissue, feces, blood, urine, bile, cerebrospinal fluid, and other biofluids. Prior to sample analysis, appropriate sample preparation is generally required to make the samples compatible with the instrumentation, such as filtration, extraction, protein precipitation, and chemical derivatization. Liquid chromatography mass spectrometry (LC-MS), gas chromatography mass spectrometry (GC-MS), and nuclear magnetic resonance spectroscopy (NMR) are commonly used as the analytical platforms for metabolomic analysis of biological samples [7–9]. Among them, LC-MS is the most widely used platform in reported metabolomics studies owing to its sensitivity, resolution, and its compatibility with the liquid matrix of diverse biological samples. Compared to LC-MS, GC-MS is advantageous for analyzing gaseous or volatile compounds, while NMR has the capability to examine intact samples (such as tissues) without destructive sample preparation. To deconvolute, visualize, and interpret large volumes of data generated by metabolomic analysis, chemometrics and bioinformatics tools have been utilized or developed [10–12]. Multivariate data analysis (MDA), which includes both unsupervised and supervised methods, is widely used to process data from untargeted metabolomic analysis. For example, unsupervised principal components analysis (PCA) could provide a non-discriminate overview on the relationships among examined samples by showing the clustering or grouping patterns and by identifying outlier samples, while supervised MDA, including partial least squares (PLS), orthogonal partial least squares (OPLS), and partial least squares discriminant analysis (PLS-DA), are effective for identifying the metabolites contributing to sample separation in the MDA models, such as vehicle versus drug or normal versus disease [13]. In mass spectrometry (MS)-based metabolomic analysis, chemical identities of the compounds or metabolites of interest can be determined by accurate mass measurement, elemental composition analysis, MS/MS fragmentation, and searches of metabolite and spectral databases, such as Kyoto Encyclopedia of Genes and Genomes (KEGG, http://www.genome.jp/kegg/), Human Metabolome Database (http://www.hmdb.ca/), Lipid Maps (http://www. lipidmaps.org/), METLIN database (http://metlin.scripps.edu/), BioCyc (http://biocyc.org/), Spectral Database for Organic Compounds (http://sdbs.riodb.aist.go.jp). However, compared to the identification processes in genomics, transcriptomics, and proteomics, how to conduct reliable and efficient metabolite annotation still remains as a major challenge in metabolomics studies despite the progresses in instrumentation and bioinformatics [14, 15]. Discovery and development of a new drug, prior to its approval by the Food and Drug Administration (FDA) in the USA, can be roughly divided into four consecutive stages,

target identification, lead discovery, preclinical testing, and clinical trial. Among the four stages, target identification and lead discovery belong to drug discovery, while preclinical testing and clinical trial belong to drug development. Despite great progresses in biomedical sciences, the endeavor to discover and develop new drugs is still a prolonged and complex process, and is marked with low success rate and high cost [16]. In order to enhance the efficiency of this process, diverse chemical, biological, and informatics tools, including combinatorial chemistry, high-throughput screening, genomics, transcriptomics, and proteomics, have been adopted or developed. Since metabolomics, as an omics tool focusing on detecting small-molecule chemicals, matches with the chemical nature of many essential components of drug discovery and development, including drugs, leads, and metabolites, metabolomics is expected to be a very useful tool for drug discovery and development. This review aims to discuss ongoing and potential applications of metabolomics in each stage of drug discovery and development and also to provide insights on how metabolomics could function as a bridge between chemistry and biology in drug discovery and development (Fig. 1). Application of Metabolomics in Target Identification Target identification is a vital step in drug discovery as it largely determines the mechanism of action for a desired new drug [17]. Among numerous established drug targets, metabolites, which are the focus of metabolomics, are actually rarely set as direct targets of drug discovery, except for few cases, such as uric acid for hyperuricemia treatment [18]. In contrast, proteins and genetic products are more commonly designated as the targets of drug discovery based on their functions in pathogenesis as well as their Bdruggability^ for being the entities that can bind to or be regulated by smallmolecule drugs [19]. In fact, receptors (e.g., G-proteincoupled receptors), enzymes (e.g., protein kinases), ion channels, and transporters are four major drug targets [20]. In accordance with the protein or genetic natures of these drug targets, proteomics, transcriptomics, and genomics platforms have been widely used in target identification and validation research [21]. However, the lack of metabolites as the targets of drug discovery does not prevent the utilization of metabolomics in target identification. The values of metabolomics in the mechanistic investigations of protein and genetic targets of drug discovery are well justified by the following considerations. Firstly, modifying or correcting aberrant metabolism is a common goal of many drug discovery efforts for preventing or treating diseases that are caused by or related to metabolic disorders. As a ubiquitous biological process in body and cells, metabolism plays a central role in the pathogenesis and phenotype development of numerous diseases. The most prominent health issue associated with metabolism in modern

Curr Pharmacol Rep Fig. 1 A summary of applications of metabolomics in four stages of drug discovery and development. Some applications, such as identification of bioactive compounds and patient stratification, are specially associated with only one stage of drug discovery and development, while others, such as biomarker identification and drug metabolism, occur in multiple stages of drug discovery and development

society is the prevalence of metabolic syndrome in general population. This array of metabolic disorders, including abdominal obesity, elevated blood pressure, elevated fasting plasma glucose, high serum triglycerides, and low highdensity cholesterol (HDL) levels, contributes greatly to the increased incidences of cardiovascular diseases, diabetes, and other fatal morbidities [22]. In addition, rejuvenated research on cancer metabolism in recent years has revealed and highlighted the importance of metabolism in carcinogenesis and tumor metastasis [23]. The importance of metabolism in other chronic diseases, such as Alzheimer’s disease [24] and arthritis [25], has also been highlighted in recent studies. Secondly, the executers and regulators of deranged metabolism in many vicious human diseases are ideal targets of pharmacological interventions. Since bidirectional interaction commonly exists between the changes in gene and protein expression and the change in metabolites, defining early and specific metabolic changes induced by the disruption of biological system facilitates and rationalizes the identification of targets that are functionally important in pathogenesis. Thirdly, due to the complexity and unpredictability of metabolic system and its regulatory mechanisms, traditional targeted metabolite analysis has limited capacity to define and characterize metabolic changes if there is no robust hypothesis available. In contract, metabolomics, especially untargeted metabolomics, is capable of identifying novel or unexpected metabolic changes in a complex biological system or matrix without established hypothesis. Searching for robust and reliable metabolite markers that could be used for identifying pharmacological targets of human metabolic diseases is an active area in metabolomics research. Both targeted and untargeted metabolomics approaches have been utilized for this purpose (Fig. 2) [26]. For example, diverse perturbations of amino acid, lipid, carbohydrate metabolism as well as bacterial metabolism have been identified in diabetic patients, pre-diabetic subjects, or

animal models [27–30]. These metabolites could serve as the markers of different diabetic status or the efficacy indicators of therapeutic/preventive interventions and could also guide the target identification of anti-diabetic drugs. Similarly, metabolomics-based approaches have been adopted in other human diseases and morbidities, including obesity [31], cancer [32], cardiovascular [33], hepatobiliary [34], and neurological [35] diseases. Recently, research efforts on gut microbiota started to provide new insights on the targets of pharmacological interventions and may lead to the development of microbiota-derived therapies [36]. Within these efforts, metabolomic investigations of microbial metabolites and their correlations with pathogenesis have facilitated the identification of novel therapeutic targets. A recent metabolomicsbased study identified the intestinal microbiome as a potential target for treating that graft-versus-host disease (GVHD), as metabolomic analysis identified the decrease of butyric acid as a prominent change in intestinal epithelial cells after allogeneic bone marrow transplant, and further revealed that the restoration of butyric acid through the local administrations of exogenous butyrate or high butyrate–producing Clostridia alleviated the severity of GVHD [37]. Moreover, microbiotamediated choline metabolism has been identified as a pathogenic factor in cardiovascular diseases through a metabolomics-based study since trimethylamine N-oxide (TMAO), a metabolite of trimethylamine (TMA) from bacteria-mediated metabolism of dietary phosphatidylcholine, was shown to promote atherosclerosis [38•]. A subsequent study showed that resveratrol, a polyphenolic antioxidant, was able to reduce the TMAO-induced atherosclerosis largely through remodeling gut microbiota, suggesting that the gut microbiome is a valid target for therapeutic interventions of TMAO-induced atherosclerosis [39, 40]. In the field of infectious diseases, metabolomics showed the promise to be a useful tool for developing novel antimicrobials to overcome drug resistance, which was exemplified by the identification of

Curr Pharmacol Rep Fig. 2 Approaches and applications of metabolomicsbased biomarker identification. Metabolomic analysis can function as an efficient discovery tool to reveal the metabolic changes caused by pathophysiological conditions or chemical treatments. The most significant changes observed in metabolomic analysis can be assigned as the metabolite biomarkers after confirmation and characterization. These biomarkers can be used for investigating the mechanisms of diseases, toxicity, and efficacy as well as the targets of drug discovery

glutamate racemase, an upstream enzyme in the early stage of peptidoglycan biosynthesis, as the primary target of β-chloroD-alanine, an anti-tuberculosis agent [41]. It is highly possible that the expansion of metabolomics platform into the studies of uncommon medical conditions and rare diseases will lead to the discovery of novel therapeutic targets for drug discovery. Application of Metabolomics in Lead Discovery Following target identification, a specific bioassay or multiple bioassays are required to examine the effects of testing compounds on expected targets of drug therapy, which can be binding affinity, inhibition, activation, or cell death. A novel chemical structures identified through this process can be designated as a new chemical entity (NCE). The screening efforts based on these bioassays, if successful, could lead to the discovery of promising leads for further development [42]. Two sources of testing compounds are organic synthesis and natural products. For purified synthetic compounds originated from rational drug design or combinatorial chemistry, metabolomics offers little value for their screening since their chemical identities are known. However, for undefined chemical mixtures, such as bioactive natural product extracts, metabolomics and its analytical approach offer a valuable platform for identifying bioactive components within the mixtures [43]. Metabolomics has been used for discovering novel bioactive compounds in plant, marine sponge, and microbial extracts, such as the ligands of adenosine A1 receptor through the correlations between metabolomic profiles and the results from adenosine A1 receptor assay [44–46]. In practice, in order to utilize this approach to identify bioactive chemicals, multiple

plant or microbial extracts originated from different locations, genetic backgrounds, treatments, or extraction methods are required to establish the correlations between bioactivities and specific compounds in chemical extracts (Fig. 3) [47]. Without exhaustive fractionation and purification, NMR and MS-based metabolomics is capable of defining chemical composition of multiple extracts and identifying major differences among them through chemometric analysis, such as PCA modeling. In addition, with correlation analysis, such as PLS or PLS-DA analysis, the data from NMR or MS analysis (chemical signals or metabolites as X variables) can be processed together with the data from bioassays (enzyme activity, gene function, or many other bioactivities as Y responses) to establish the correlation between specific chemicals in extracts and different responses observed in bioassay (Fig. 3) [48]. Therefore, instead of obtaining pure compound prior to bioassays, the interested bioactive compounds can be identified through the correlations between metabolomics and bioassays. Afterwards, purification and structural analysis can be followed to determine their structures, and a new round of bioassay can be conducted to validate the bioactivities of these purified novel compounds. Based on this approach, a combination of metabolomics, genomics, and bioinformatics has fostered the lead discovery from natural products [49, 50]. For example, the pairings of secondary metabolites of 178 actinomycete strains with their biosynthetic gene clusters, together with functional assays, led to the identification of tambromycin, a new chlorinated natural product that has antiproliferative activity against cancerous human B and T cell lines [51••]. Similar approaches have also been adopted for discovering new antibiotics to handle the prevalent drug resistance issues in medical practice [52].

Curr Pharmacol Rep

Fig. 3 Applications of metabolomics in screening bioactive compounds in undefined chemical mixtures. Different chemical mixtures, such as plant or microbial extracts, can be compared and examined by both metabolomic analysis and specific bioassays. The bioassays distinguish effective chemical mixtures from ineffective ones while the metabolomics analysis identifies the chemical components that differ between effective and ineffective chemical mixtures. The correlations between individual compounds in chemical mixture and specific bioactivity can lead to the identification of bioactive compounds

Application of Metabolomics in Preclinical Testing After a bioactive NCE is selected from lead discovery process for further development, preclinical testing, which uses animal models and in vitro systems, has to be conducted to examine its safety and efficacy before it can be tested in clinical trials. Since poor pharmacokinetic profile, unexpected toxicity, and insufficient efficacy are major causes behind the failures of many lead compounds [53], conducting stringent and comprehensive preclinical testing to address these risk factors and to provide mechanistic insights on safety and efficacy is essential for predicting the outcome of subsequent clinical trial and reducing its failure rate. Metabolomics is a valuable tool for achieving these goals of preclinical testing due to two important characters of preclinical testing. Firstly, metabolism, including both drug metabolism and endogenous metabolism, is an essential element for the pharmacokinetics, toxicity, and efficacy of a NCE, sometimes a determining factor. Therefore, information obtained from metabolomic analysis on the levels and turnover of NCE metabolites and endobiotics could offer direct insights on the pharmacokinetic profile of a NCE as well as its toxicity and efficacy. Compared to traditional metabolite analysis, the strength of metabolomics in studying NCE-related metabolic events mainly comes from its analytical capacity to detect subtle changes in complex biological matrix. Secondly, in many cases, poor pharmacokinetic profile and undesirable toxicity of NCEs could not be

easily explained by existing knowledge on drug targets, diseases, and chemical-biological interactions. This uncertainty of preclinical testing could be overcome by the discovery power of metabolomics, especially by the exploratory nature of untargeted metabolomics. Biotransformation of NCE (or drug metabolism) and NCEinduced metabolic changes are two types of metabolic events related to the treatment of a NCE in a biological system. Metabolomic investigation of NCE-related metabolic effects can be defined as exploratory, such as the identification of new NCE metabolites and biomarkers, or hypothesis-driven, such as the role of enzymes and pathways in NCE-related metabolic effects. Drug metabolism could serve as a double-edged sword for safety and efficacy of a NCE since it could be either activated or deactivated by the biotransformation reactions catalyzed by drug-metabolizing enzymes in the body. The balance between activation and deactivation, in many cases, could determine the toxicity and efficacy of a NCE [54]. Therefore, studying the biotransformation of a NCE is vital for understanding and predicting its effects in vivo. Metabolomics-based approach in drug metabolism research has advantages over traditional analytical approaches like radiotracing and empirical metabolite identification, because metabolomics can circumvent some drawbacks and limitations of these approaches, such as synthesis of radiolabeled NCE and waste handling in radiotracing approach, and occurrence of unexpected biotransformation reactions in empirical metabolite identification [55]. In practice, a straightforward approach for identifying drug metabolites is to conduct a metabolomic comparison of samples from vehicle and NCE treatments, which could be urine, serum, feces, tissue extracts, and in vitro reaction mixtures. Since a NCE and its metabolites are only present in the samples of NCE treatment, analyzing metabolites that contribute to the separation of NCE and vehicle treatments in a metabolomic model could lead to the identification of NCE metabolites (Fig. 4). Using this approach, novel metabolites and metabolic pathways of therapeutic agents, toxicants, and dietary compounds, including aminoflavone [56], arecoline [57], and cocaine [58], have been identified and characterized. An alternative approach to facilitate metabolite identification and avoid interference from endobiotics is to use stable isotope (2H, 13C, 15N, or 18O)-labeled NCE in metabolomic analysis, in which equal amount of labeled and unlabeled xenobiotics are used in the treatments (Fig. 4). The efficacy of this approach has been demonstrated by the identification of novel metabolites of acetaminophen (APAP) [59], ethanol [60], and tempol [61]. Besides identifying novel NCE metabolites, metabolomics can also be adopted to investigate the roles of enzymes, such as cytochrome P450s, in drug metabolism when appropriate experiment models, such as mutation, knockout, or humanized animal models, are used [62]. For example, the role of CYP1A2 in metabolism of PhIP, a

Curr Pharmacol Rep

Fig. 4 The approaches of identifying NCE/IND metabolites through metabolomics. Unlike endobiotics (represented by open symbols), NCE/IND metabolites (represented by solid symbols) are not present in the metabolome prior to the treatment. Because of this feature, these exogenous metabolites are commonly presented as major contributors

to the separation of NCE/IND treatment samples from control samples in the metabolomics models. Therefore, metabolomic comparisons of vehicle vs NCE/IND treatments or unlabeled vs stable isotope-labeled NCE/IND (*) treatments facilitate the identification of NCE/IND metabolites

procarcinogen in human diet, was defined after metabolomic examination of urine samples from control and PhIP-treated wild-type, Cyp1a2-null, and CYP1A2-humanized mice. The results of this metabolomic study not only proved that CYP1A2 is a major PhIP-metabolizing enzyme but also revealed the existence of significant interspecies differences between human and mouse as well as the involvement of other cytochrome P450s in PhIP metabolism [63]. Similar to metabolomic investigation of drug metabolism, a straightforward comparison of samples from control and NCE-treated subjects can also be applied to investigate NCE-induced changes in endogenous metabolism. One major application of this approach is to screen and characterize NCE-elicited toxicities, resulting in observing chemicalinduced changes in lipid, amino acid, antioxidant, and carbohydrate metabolism [64–67]. For example, hepatotoxicity is a major form of toxicity that had led to discontinuation of many NCE developments and market withdrawal of approved drugs. Using APAP as the prototypic compound for druginduced liver injury, a metabolomic investigation of serum samples from APAP-sensitive and APAP-resistant mice identified the accumulation of long-chain acylcarnitines in serum as a prominent metabolic change in APAP overdose-induced hepatotoxicity. Because long-chain acylcarnitines are the substrates of mitochondrial fatty acid catabolism, this observation led to the identification of APAP-induced inhibition of fatty acid oxidation in the liver as well as the suppression of peroxisome proliferator-activated receptor alpha (PPARα) activation, which is a major regulator of fatty acid metabolism [68]. Another metabolomic examination of cocaine-induced liver injury also revealed progressive increase of serum acylcarnitine level during a 3-day cocaine treatment [69], suggesting acylcarnitine accumulation could be a common event in drug-induced hepatotoxicity. In fact, this metabolic event

has been since proposed as a diagnostic marker of chemicalinduced mitochondrial disorder [70]. Besides marker identification, determining underlying mechanisms of these metabolite markers is another application of metabolomic investigation (Fig. 2). In practice, when a hypothesis on the source and metabolic route of a marker is proposed, animals or cells could be treated with the stable isotope tracers of specific upstream and downstream endobiotics that are related to the identified metabolite markers. Afterwards, metabolomic analysis could trace the fates of stable isotope tracer and test the hypothesis [71]. This approach of combining metabolomics with stable isotope tracer has been effectively used to interrogate the chemical-induced changes in glucose, amino acid, and fatty acid metabolism, and is becoming a valuable tool in drug development [72]. Application of Metabolomics in Clinical Trial Metabolomics and metabolomics-based approaches can assist many aspects of a clinical trial on an investigational new drug (IND) (Fig. 1). Similar to its applications in target identification and preclinical testing, a direct application of metabolomics in human studies is to identify the metabolite markers of efficacy and toxicity as well as the metabolites of therapeutic agents (Fig. 4). For example, a novel metabolite of ethambutol, an anti-tuberculosis drug, and major metabolites of an herbal medicine were identified through a metabolomics analysis of patient urine and serum samples, respectively [73, 74]. Besides determining the metabolic fates of therapeutic agents, metabolomics-defined metabolic profile could also be used to predict pharmacokinetics and adverse effects of IND treatments, as demonstrated by the proof-of-concept human studies on tacrolimus [75], APAP [76], and busulfan [77]. In the case of busulfan study, the glycine, serine, and threonine

Curr Pharmacol Rep

metabolism pathway was associated with intravenous busulfan clearance in hematopoietic cell transplant recipients [77]. Another potential application of metabolomics in clinical trials is to evaluate the compliance of clinical study subjects on dosing and dietary pattern [78, 79], which are often key factors in clinical trials. Compared to preclinical testing, clinical trial has greater uncertainty, making it the most costly and time-consuming process in drug development. One factor contributing to this uncertainty is the genetic and epigenetic diversity of human subjects in clinical trials, which is in contrast to the uniformity of animal and in vitro models used in preclinical testing. Therefore, selecting the study subjects that are suitable with and also could benefit from intended therapeutic applications of IND is an essential element for the success of a clinical trial. Pharmacometabolomics has shown promise to be a useful tool for patient stratification [80, 81••]. For example, the metabolomic analysis of serum samples from healthy controls and arthritis patients was capable of distinguishing four major types of arthritis, rheumatoid arthritis, osteoarthritis, ankylosing spondylitis, and gout, based on their metabolic profiles [82]. Using this approach, specific metabolite marker or metabolic profile of a human disease could be used as a criterion to screen patients. Since the metabolic phenotypes defined by metabolomics could serve as the bridge between the genotypes of human patients and the responses to IND treatments, which include efficacy, adverse effect, and prognosis, one major impact of these metabolomics-based practices is the facilitation and development of personalized medicine in clinical trial as well as in clinical practice and post-market safety monitoring [83–85]. For example, metabolomics was applied to identify the treatment outcome biomarkers of escitalopram, a selective serotonin reuptake inhibitor for treating depression disorders. Through metabolomic comparison of plasma samples from the patients responsive and nonresponsive to escitalopram, the glycine level was found to be negatively associated with the treatment outcome. Subsequent pharmacogenomic investigation revealed that a single nucleotide polymorphism (SNP) in glycine dehydrogenase gene is associated with the treatment outcome of escitalopram [86]. In this case, metabolomic investigation sets up a foundation for adopting personalized usage of a therapeutic agent in clinic. Since all these metabolomics-based efforts affect the outcome of clinical studies, the adoption of metabolomics in clinical studies will likely improve the success rate of IND in drug development.

specifically associated with lead discovery and clinical trial, respectively, while biomarker identification and drug metabolism facilitate the mechanistic investigations in target identification, preclinical testing, and clinical trial (Fig. 1). Challenges in chemistry and biology exist in each stage of drug discovery and development. Failures to overcome these challenges are largely due to the lack of sufficient information and knowledge for making rational judgment. Development of omics platforms (genomics, epigenetics, transcriptomics, proteomics, metabolomics) in recent decades has provided powerful tools to help scientist gain more comprehensive view on these challenges without missing essential details. Compared to other omics platforms, metabolomics aims to measure the end products of all upstream events and activities that involve genes, proteins, enzymes, signal transduction, and regulations. This unique function of metabolomics renders its capacity to convey phenotypic information of a biological system. By applying this function in drug discovery and development, metabolomics could assist many research efforts in target identification, lead discovery, preclinical testing, and clinical trial (Fig. 1). Despite this great promise of metabolomics, the reality of utilizing metabolomics in drug discovery and development is far from ideal, and metabolomics remains to become a common tool in this process. Several reasons contribute to this status of metabolomics, including unfamiliarity of pharmaceutical research scientists with metabolomics technology, lack of efficient tools for data annotation and structural elucidation, and mediocre quality of many metabolomics research papers that only presented descriptive information on metabolic changes without conducting sufficient follow-up investigations to understand the mechanisms and significances. Like other omics platforms, the complete integration of metabolomics into drug discovery and development will take time and efforts from the scientists in multiple disciplines. Nevertheless, the capacity to bridge chemistry and biology will make metabolomics a valuable tool in future drug discovery and development.

Acknowledgments Dr. Chi Chen’s research is partially supported by the Agricultural Experiment Station project MIN-18-092 from the United States Department of Agriculture (USDA).

Compliance with Ethical Standards

Conclusion

Conflict of Interest The authors declare that they have no conflicts of interest.

Metabolomics has diverse applications in four stages of drug discovery and development. Among these applications, identification of bioactive compounds and patient stratification are

Human and Animal Rights and Informed Consent This article does not contain any studies with human or animal subjects performed by any of the authors.

Curr Pharmacol Rep

References

22. 23.

Papers of particular interest, published recently, have been highlighted as: • Of importance •• Of major importance 1. 2. 3.

4.

5. 6.

7.

8.

9.

10. 11. 12.

13. 14.

15.

16.

17.

18. 19. 20. 21.

Lindon JC, Holmes E, Nicholson JK. So what's the deal with metabonomics? Anal Chem. 2003;75:384A–91A. Weckwerth W, Fiehn O. Can we discover novel pathways using metabolomic analysis? Curr Opin Biotechnol. 2002;13:156–60. Mamas M, Dunn WB, Neyses L, Goodacre R. The role of metabolites and metabolomics in clinically applicable biomarkers of disease. Arch Toxicol. 2011;85:5–17. Johnson CH, Patterson AD, Idle JR, Gonzalez FJ. Xenobiotic metabolomics: major impact on the metabolome. Annu Rev Pharmacol Toxicol. 2012;52:37–56. Chen C, Kim S. LC-MS-based metabolomics of xenobioticinduced toxicities. Comput Struct Biotechnol J. 2013;4:e20130108. German JB, Roberts MA, Fay L, Watkins SM. Metabolomics and individual metabolic assessment: the next great challenge for nutrition. J Nutr. 2002;132:2486–7. Metz TO, Zhang Q, Page JS, Shen Y, Callister SJ, Jacobs JM, Smith RD. The future of liquid chromatography-mass spectrometry (LCMS) in metabolic profiling and metabolomic studies for biomarker discovery. Biomark Med. 2007;1:159–85. Styczynski MP, Moxley JF, Tong LV, Walther JL, Jensen KL, Stephanopoulos GN. Systematic identification of conserved metabolites in GC/MS data for metabolomics and biomarker discovery. Anal Chem. 2007;79:966–73. Dieterle F, Riefke B, Schlotterbeck G, Ross A, Senn H, Amberg A. NMR and MS methods for metabonomics. Methods Mol Biol. 2011;691:385–415. Katajamaa M, Oresic M. Data processing for mass spectrometrybased metabolomics. J Chromatogr A. 2007;1158:318–28. Trygg J, Holmes E, Lundstedt T. Chemometrics in metabonomics. J Proteome Res. 2007;6:469–79. Sugimoto M, Kawakami M, Robert M, Soga T, Tomita M. Bioinformatics tools for mass spectroscopy-based metabolomic data processing and analysis. Curr Bioinform. 2012;7:96–108. Want E, Masson P. Processing and analysis of GC/LC-MS-based metabolomics data. Methods Mol Biol. 2011;708:277–98. Iijima Y, Nakamura Y, Ogata Y, Tanaka K, Sakurai N, Suda K, Suzuki T, Suzuki H, Okazaki K, Kitayama M, Kanaya S, Aoki K, Shibata D. Metabolite annotations based on the integration of mass spectral information. Plant J. 2008;54:949–62. Kind T, Fiehn O. Metabolomic database annotations via query of elemental compositions: mass accuracy is insufficient even at less than 1 ppm. Bmc Bioinformatics. 2006;7:234. Kaitin KI, DiMasi JA. Pharmaceutical innovation in the twenty-first century: new drug approvals in the first decade, 2000-2009. Clin Pharmacol Ther. 2011;89:183–8. Schenone M, Dancik V, Wagner BK, Clemons PA. Target identification and mechanism of action in chemical biology and drug discovery. Nat Chem Biol. 2013;9:232–40. Pea F. Pharmacology of drugs for hyperuricemia. mechanisms, kinetics and interactions. Contrib Nephrol. 2005;147:35–46. Dixon SJ, Stockwell BR. Identifying druggable disease-modifying gene products. Curr Opin Chem Biol. 2009;13:549–55. Imming P, Sinning C, Meyer A. Drugs, their targets and the nature and number of drug targets. Nat Rev Drug Discov. 2006;5:821–34. Betts JC. Transcriptomics and proteomics: tools for the identification of novel drug targets and vaccine candidates for tuberculosis. IUBMB Life. 2002;53:239–42.

24.

25.

26.

27. 28. 29. 30.

31.

32. 33. 34. 35.

36. 37.

38.•

39.

40.

41.

Moller DE. Metabolic disease drug discovery—Bhitting the target^ is easier said than done. Cell Metab. 2012;15:19–24. Schulze A, Harris AL. How cancer metabolism is tuned for proliferation and vulnerable to disruption. Nature. 2012;491:364–73. Cai H, Cong WN, Ji S, Rothman S, Maudsley S, Martin B. Metabolic dysfunction in Alzheimer’s disease and related neurodegenerative disorders. Curr Alzheimer Res. 2012;9:5–17. da Cunha VR, Brenol CV, Brenol JC, Fuchs SC, Arlindo EM, Melo IM, Machado CA, de Castro Chaves H, Jr, Xavier RM. Metabolic syndrome prevalence is increased in rheumatoid arthritis patients and is associated with disease activity. Scand J Rheumatol. 2012;41:186–91. Griffiths WJ, Koal T, Wang Y, Kohl M, Enot DP, Deigner HP. Targeted metabolomics for biomarker discovery. Angew Chem Int Ed Engl. 2010;49:5426–45. Suhre K. Metabolic profiling in diabetes. J Endocrinol. 2014;221: R75–85. Friedrich N. Metabolomics in diabetes research. J Endocrinol. 2012;215:29–42. Lu J, Xie G, Jia W. Metabolomics in human type 2 diabetes research. Front Med. 2013;7:4–13. Lu Y, Yao D, Chen C. 2-Hydrazinoquinoline as a derivatization agent for LC-MS-based metabolomic investigation of diabetic ketoacidosis. Metabolites. 2013;3:993–1010. Zhang A, Sun H, Wang X. Power of metabolomics in biomarker discovery and mining mechanisms of obesity. Obes Rev. 2013;14: 344–9. Aboud OA, Weiss RH. New opportunities from the cancer metabolome. Clin Chem. 2013;59:138–46. Senn T, Hazen SL, Tang WH. Translating metabolomics to cardiovascular biomarkers. Prog Cardiovasc Dis. 2012;55:70–6. Beyoglu D, Idle JR. The metabolomic window into hepatobiliary disease. J Hepatol. 2013;59:842–58. Hassan-Smith G, Wallace GR, Douglas MR, Sinclair AJ. The role of metabolomics in neurological disease. J Neuroimmunol. 2012;248:48–52. Garber K. Drugging the gut microbiome. Nat Biotech. 2015;33: 228–31. Mathewson ND, Jenq R, Mathew AV, Koenigsknecht M, Hanash A, Toubai T, Oravecz-Wilson K, Wu S-R, Sun Y, Rossi C, Fujiwara H, Byun J, Shono Y, Lindemans C, Calafiore M, Schmidt TM, Honda K, Young VB, Pennathur S, van den Brink M, Reddy P. Gut microbiome-derived metabolites modulate intestinal epithelial cell damage and mitigate graft-versus-host disease. Nat Immunol. 2016;17:505–13. Wang Z, Klipfell E, Bennett BJ, Koeth R, Levison BS, DuGar B, Feldstein AE, Britt EB, Fu X, Chung Y-M, Wu Y, Schauer P, Smith JD, Allayee H, Tang WHW, DiDonato JA, Lusis AJ, Hazen SL. Gut flora metabolism of phosphatidylcholine promotes cardiovascular disease. Nature. 2011;472:57–63. Metabolomic analysis of plasma samples from CVD patients set up the foundation for discovering the microbial metabolism of choline as a source of CVD pathogenesis. It demonstrates the power of metabolomics in target identification. The finding has broad applications in therapy and prevention Koopen AM, Groen AK, Nieuwdorp M. Human microbiome as therapeutic intervention target to reduce cardiovascular disease risk. Curr Opin Lipidol. 2016;27:615–22. Chen ML, Yi L, Zhang Y, Zhou X, Ran L, Yang J, Zhu JD, Zhang QY, Mi MT. Resveratrol attenuates trimethylamine-N-oxide (TMAO)-induced atherosclerosis by regulating TMAO synthesis and bile acid metabolism via remodeling of the gut microbiota. MBio. 2016;7:e02210-15. Prosser GA, Rodenburg A, Khoury H, de Chiara C, Howell S, Snijders AP, de Carvalho LP. Glutamate racemase is the primary

Curr Pharmacol Rep

42.

43.

44.

45.

46.

47.

48. 49.

50.

51.••

52. 53.

54. 55. 56.

57.

58.

target of beta-chloro-d-alanine in mycobacterium tuberculosis. Antimicrob Agents Chemother. 2016;60:6091–9. Bleicher KH, Bohm HJ, Muller K, Alanine AI. Hit and lead generation: beyond high-throughput screening. Nat Rev Drug Discov. 2003;2:369–78. Rochfort S. Metabolomics reviewed: a new Bomics^ platform technology for systems biology and implications for natural products research. J Nat Prod. 2005;68:1813–20. Yuliana ND, Khatib A, Verpoorte R, Choi YH. Comprehensive extraction method integrated with NMR metabolomics: a new bioactivity screening method for plants, adenosine A1 receptor binding compounds in Orthosiphon stamineus Benth. Anal Chem. 2011;83: 6902–6. Ali K, Iqbal M, Yuliana ND, Lee Y-J, Park S, Han S, Lee J-W, Lee H-S, Verpoorte R, Choi YH. Identification of bioactive metabolites against adenosine A1 receptor using NMR-based metabolomics. Metabolomics. 2013;9:778–85. Hou Y, Braun DR, Michel CR, Klassen JL, Adnani N, Wyche TP, Bugni TS. Microbial strain prioritization using metabolomics tools for the discovery of natural products. Anal Chem. 2012;84:4277– 83. Yuliana ND, Khatib A, Choi YH, Verpoorte R. Metabolomics for bioactivity assessment of natural products. Phytother Res. 2011;25: 157–69. Wold S, Sjostrom M, Eriksson L. PLS-regression: a basic tool of chemometrics. Chemom Intell Lab Syst. 2001;58:109–30. Doroghazi JR, Albright JC, Goering AW, Ju KS, Haines RR, Tchalukov KA, Labeda DP, Kelleher NL, Metcalf WW. A roadmap for natural product discovery based on large-scale genomics and metabolomics. Nat Chem Biol. 2014;10:963–8. Johnston CW, Skinnider MA, Wyatt MA, Li X, Ranieri MR, Yang L, Zechel DL, Ma B, Magarvey NA. An automated genomes-tonatural products platform (GNP) for the discovery of modular natural products. Nat Commun. 2015;6:8421. Goering AW, McClure RA, Doroghazi JR, Albright JC, Haverland NA, Zhang Y, Ju KS, Thomson RJ, Metcalf WW, Kelleher NL. Metabologenomics: correlation of microbial gene clusters with metabolites drives discovery of a nonribosomal peptide with an unusual amino acid monomer. ACS central science. 2016;2:99–108. The combination of metabolomics, genomics, and functional assay in this study presents a highly effective platform for identifying bioactive compounds from complex natural product mixtures. The progress in this metabologenomics platform will likely revive the efforts to explore natural products as the major sources of new drugs Wencewicz, T. A., New antibiotics from nature's chemical inventory. Bioorganic & medicinal chemistry 2016. Schuster D, Laggner C, Langer T. Why drugs fail—a study on side effects in new chemical entities. Curr Pharm Des. 2005;11:3545– 59. Ioannides, C., Enzyme systems that metabolise drugs and other xenobiotics. Wiley: New York, 2002; p viii, 578 p. Chen C, Gonzalez FJ, Idle JR. LC-MS-based metabolomics in drug metabolism. Drug Metab Rev. 2007;39:581–97. Chen C, Meng L, Ma X, Krausz KW, Pommier Y, Idle JR, Gonzalez FJ. Urinary metabolite profiling reveals CYP1A2mediated metabolism of NSC686288 (aminoflavone). J Pharmacol Exp Ther. 2006;318:1330–42. Giri S, Idle JR, Chen C, Zabriskie TM, Krausz KW, Gonzalez FJ. A metabolomic approach to the metabolism of the areca nut alkaloids arecoline and arecaidine in the mouse. Chem Res Toxicol. 2006;19: 818–27. Yao D, Shi X, Wang L, Gosnell BA, Chen C. Characterization of differential cocaine metabolism in mouse and rat through metabolomics-guided metabolite profiling. Drug Metab Dispos. 2013;41:79–88.

59.

Chen C, Krausz KW, Idle JR, Gonzalez FJ. Identification of novel toxicity-associated metabolites by metabolomics and mass isotopomer analysis of acetaminophen metabolism in wild-type and Cyp2e1-null mice. J Biol Chem. 2008;283:4543–59. 60. Shi X, Yao D, Chen C. Identification of N-acetyltaurine as a novel metabolite of ethanol through metabolomics-guided biochemical analysis. J Biol Chem. 2012;287:6336–49. 61. Li F, Pang X, Krausz KW, Jiang C, Chen C, Cook JA, Krishna MC, Mitchell JB, Gonzalez FJ, Patterson AD. Stable isotope- and mass spectrometry-based metabolomics as tools in drug metabolism: a study expanding tempol pharmacology. J Proteome Res. 2013;12: 1369–76. 62. Jiang XL, Gonzalez FJ, Yu AM. Drug-metabolizing enzyme, transporter, and nuclear receptor genetically modified mouse models. Drug Metab Rev. 2011;43:27–40. 63. Chen C, Ma X, Malfatti MA, Krausz KW, Kimura S, Felton JS, Idle JR, Gonzalez FJ. A comprehensive investigation of 2-amino-1methyl-6-phenylimidazo[4,5-b]pyridine (PhIP) metabolism in the mouse using a multivariate data analysis approach. Chem Res Toxicol. 2007;20:531–42. 64. Boudonck KJ, Mitchell MW, Nemet L, Keresztes L, Nyska A, Shinar D, Rosenstock M. Discovery of metabolomics biomarkers for early detection of nephrotoxicity. Toxicol Pathol. 2009;37:280– 92. 65. Beger RD, Sun J, Schnackenberg LK. Metabolomics approaches for discovering biomarkers of drug-induced hepatotoxicity and nephrotoxicity. Toxicol Appl Pharmacol. 2010;243:154–66. 66. Clarke CJ, Haselden JN. Metabolic profiling as a tool for understanding mechanisms of toxicity. Toxicol Pathol. 2008;36:140–7. 67. Aranibar N, Bhaskaran V, Ott KH, Vassallo J, Nelson D, Lecureux L, Gong L, Stryker S, Lehman-McKeeman L. Modulation of ascorbic acid metabolism by cytochrome P450 induction revealed by metabonomics and transcriptional profiling. Magn Reson Chem. 2009;47(Suppl 1):S12–9. 68. Chen C, Krausz KW, Shah YM, Idle JR, Gonzalez FJ. Serum metabolomics reveals irreversible inhibition of fatty acid beta-oxidation through the suppression of PPARalpha activation as a contributing mechanism of acetaminophen-induced hepatotoxicity. Chem Res Toxicol. 2009;22:699–707. 69. Shi X, Yao D, Gosnell BA, Chen C. Lipidomic profiling reveals protective function of fatty acid oxidation in cocaine-induced hepatotoxicity. J Lipid Res. 2012;53:2318–30. 70. McGill MR, Li F, Sharpe MR, Williams CD, Curry SC, Ma X, Jaeschke H. Circulating acylcarnitines as biomarkers of mitochondrial dysfunction after acetaminophen overdose in mice and humans. Arch Toxicol. 2013;88:391–401. 71. Creek DJ, Chokkathukalam A, Jankevics A, Burgess KE, Breitling R, Barrett MP. Stable isotope-assisted metabolomics for networkwide metabolic pathway elucidation. Anal Chem. 2012;84:8442–7. 72. Fan TW, Lorkiewicz PK, Sellers K, Moseley HN, Higashi RM, Lane AN. Stable isotope-resolved metabolomics and applications for drug development. Pharmacol Ther. 2012;133:366–91. 73. Das MK, Arya R, Debnath S, Debnath R, Lodh A, Bishwal SC, Das A, Nanda RK. Global urine metabolomics in patients treated with first-line tuberculosis drugs and identification of a novel metabolite of ethambutol. Antimicrob Agents Chemother. 2016;60:2257–64. 74. Guo M, Zhao B, Liu H, Zhang L, Peng L, Qin L, Zhang Z, Li J, Cai C, Gao X. A metabolomic strategy to screen the prototype components and metabolites of Shuang-Huang-Lian injection in human serum by ultra performance liquid chromatography coupled with quadrupole time-of-flight mass spectrometry. Journal of Analytical Methods in Chemistry. 2014;2014:12. 75. Phapale PB, Kim SD, Lee HW, Lim M, Kale DD, Kim YL, Cho JH, Hwang D, Yoon YR. An integrative approach for identifying a metabolic phenotype predictive of individualized pharmacokinetics of tacrolimus. Clin Pharmacol Ther. 2010;87:426–36.

Curr Pharmacol Rep 76.

Clayton TA, Lindon JC, Cloarec O, Antti H, Charuel C, Hanton G, Provost JP, Le Net JL, Baker D, Walley RJ, Everett JR, Nicholson JK. Pharmaco-metabonomic phenotyping and personalized drug treatment. Nature. 2006;440:1073–7. 77. Navarro SL, Randolph TW, Shireman LM, Raftery D, McCune JS. Pharmacometabonomic prediction of busulfan clearance in hematopoietic cell transplant recipients. J Proteome Res. 2016;15:2802– 11. 78. Andersen MB, Rinnan A, Manach C, Poulsen SK, Pujos-Guillot E, Larsen TM, Astrup A, Dragsted LO. Untargeted metabolomics as a screening tool for estimating compliance to a dietary pattern. J Proteome Res. 2014;13:1405–18. 79. Wallner-Liebmann S, Gralka E, Tenori L, Konrad M, Hofmann P, Dieber-Rotheneder M, Turano P, Luchinat C, Zatloukal K. The impact of free or standardized lifestyle and urine sampling protocol on metabolome recognition accuracy. Genes Nutr. 2015;10:441. 80. Lindon JC, Nicholson JK. The emergent role of metabolic phenotyping in dynamic patient stratification. Expert Opin Drug Metab Toxicol. 2014;10:915–9. 81.•• Everett JR. From metabonomics to pharmacometabonomics: the role of metabolic profiling in personalized medicine. Front Pharmacol. 2016;7:297. This review provides a balanced

overview on the development of pharmacometabon(l)omics in the past 10 years and the promise of its applications in predicting the outcome of drug treatment 82. Jiang M, Chen T, Feng H, Zhang Y, Li L, Zhao A, Niu X, Liang F, Wang M, Zhan J, Lu C, He X, Xiao L, Jia W, Lu A. Serum metabolic signatures of four types of human arthritis. J Proteome Res. 2013;12:3769–79. 83. Everett JR. Pharmacometabonomics in humans: a new tool for personalized medicine. Pharmacogenomics. 2015;16:737–54. 84. Eckhart AD, Beebe K, Milburn M. Metabolomics as a key integrator for Bomic^ advancement of personalized medicine and future therapies. Clin Transl Sci. 2012;5:285–8. 85. Schnackenberg LK. Global metabolic profiling and its role in systems biology to advance personalized medicine in the twenty-first century. Expert Rev Mol Diagn. 2007;7:247–59. 86. Ji Y, Hebbring S, Zhu H, Jenkins GD, Biernacka J, Snyder K, Drews M, Fiehn O, Zeng Z, Schaid D, Mrazek DA, KaddurahDaouk R, Weinshilboum RM. Glycine and a glycine dehydrogenase (GLDC) SNP as citalopram/escitalopram response biomarkers in depression: pharmacometabolomics-informed pharmacogenomics. Clin Pharmacol Ther. 2011;89:97–104.