Received date: September 18, 2012; Accepted date: September 18, 2012; Published date: September 25, 2012
Citation: Yang Z, Marotta F (2012) Pharmacometabolomics in Drug Discovery & Development: Applications and Challenges. Metabolomics 2:e122. doi:10.4172/2153-0769.1000e122
Copyright: © 2012 Yang Z, et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Visit for more related articles at Metabolomics:Open Access
Metabolomics, a relatively new Omics platform, investigate small molecular metabolites by revealing any specific biomarker(s) in human disease or certain metabolism pattern changes after genetic and environmental intervention. It provides a direct read out of the metabolic state of an individual which cannot be obtained from transcriptomics and proteomics . These small molecules are the final products of cellular regulatory processes and may reflect the in vivo metabolism phenotype controlled by both genotype and xenobiotics including: environmental factors, diet and drugs. Metabolomics have been successfully applied in various research and clinical areas including: early diagnosis of diabetes , cancers , cardiovascular diseases , neonatology , phytotherapy  and even human nutrition research .
Recently, the concept of pharmacometabolomics is mentioned more frequently as an emerging discipline to study the effect of drugs on the whole pattern of small endogenous molecules and in applying the profiles of metabolomics for drug development. For the latter part, metabolomics is majorly used to differentiate patients into responder or non-responder groups in an effort to decrease large inter-individual variation in clinical trials [8,9]. It is a novel approach that combines metabolite profile and chemo-metrics to model and predict drug targets, efficacy, pharmacokinetics and toxicity on both individual and population basis . It attracts many scientists’ attention because of its intrinsic advantages and promising potentials in drug discovery and development . Considering personalized drug treatment is the desired goal for current drug development, pharmacometabolomics provide an effective and inexpensive strategy to evaluate drug efficacy and toxicology, which may make personalized medicine realistic both from scientific and financial perspectives [10,11]. Furthermore, the FDA also realized that metabolomics coupling with other “Omics” approaches could be a valuable tool in evaluating general toxicology and could eventually replace the use of animals after addressing certain challenges .
Drug discovery and development traditionally relied on various genotype cell/animal models for target identification and validation. With in-silico modeling (i.e., QSAR or ligand-receptor models) in rational drug design, compounds are synthesized and screened for potency and toxicity. Later, time-consuming ADMET (absorption, distribution, metabolism, elimination and toxicity) studies are performed among promising compounds to further screen and identify lead compounds. After optimizing the dosage form and dose regimen of the API, lengthy and costly clinical trials are conducted to provide final evaluation. After several years of development, pharmacometabolomics now provide seamless incorporation to each individual step in drug discovery and development as shown below.
The main application of metabolomics lies in the discovery of biomarkers, which could function as a diagnostic tool for cancer prevention, early detection, and monitoring the progression of diseases [13,14]. With high throughput/high resolution mass spectrometry and NMR techniques as well as various statistical models, many biomarkers in human disease were successfully identified to differentiate patients and health control groups, or to predict death . It is very critical for the pharmaceutical industry to find biomarkers for disease and to indicate how people respond to a drug. Biomarkers can be used in predicting therapy response and as surrogate end points for clinical trials . In one study, five metabolites were detected differently between the non-invasive tumors and the invasive tumors in gastric cancer patients, suggesting the selected tissue metabolites could probably be applied for clinical diagnosis of gastric cancer . Many other applications of metabolomics in biomarker discovery have been succinctly summarized in several reviews [18,19].
Since the biochemical bases of many diseases come from the activity of particular enzymes, the pharmacometabolomics profile not only diagnose disease, but also reveal new drug targets and explore new treatment strategies. The alterations of cellular metabolic stages describe combination changes of genome, transcriptome and proteome. Therefore, pharmacometabolomics is a complementary tool for drug target identification and validation. In a serum metabolomic analysis, Chen et al. identified stearoyl-CoA desaturase 1 (SCD1) and its related lipid species which may serve as potential targets for treatment of inflammatory diseases . 37 genetically determined metabolite traits were reported with strong association for various diseases . The identification of associated metabolic traits may generate many new hypotheses for biomedical and pharmaceutical research. Among these genes, SLC16A9 (MCT9) was demonstrated as a carnitine efflux transporter responsible for carnitine efflux from absorptive epithelia into the blood . Wei conducted a target-based metabolomics study to characterize metabolic response of Huh7.5 cells to genomic perturbation of HIF-1. The results identify a new therapeutic target by confirming HIF-1’s regulatory role in tumor energy metabolism .
ADMET (absorption, disposition, metabolism, elimination and toxicity) could serve as a screen and optimization step in drug discovery and development. The drug candidates with unfavorable ADMET properties (i.e., very poor solubility, extremely low bioavailability, general toxicity) usually failed during the development phase. On the other side, the lead compounds may obtain optimized formulation or dosage design to improve its efficacy and decreased toxicity with good understanding of these biopharmaceutical properties. Drug metabolism is an important activation (for prodrug) or detoxifying route in vivo to access drug’s efficacy and toxicity. It determined drug candidate’s bioavailability, pharmacokinetics and pharmacodynamics . Coupling with various transporters, drugs including natural compounds may go through extensive metabolism after administration [24,25] and may cause large variations of pharmacokinetics and bioavailability in animal and clinical studies [26,27]. Pharmacometabolomics now provides a faster, simpler, less invasive approach to characterize biopharmaceutical properties and toxicity as well as revealing their mechanisms. In one study, NMR derived pharmacometabolomic analysis showed significant difference of cellular response at high and low doses of docetaxel in MCF7 breast cancer cells . The LC/ MS/MS based pharmacometabolomic approach applied to cisplatin’s cytoxotixicty in human lung cancer cell lines was also reported . These results indicated that the efficacy of pharmacometabolomics to evaluate cytotoxicity of various pharmaceutical ingredients and this has been widely accepted . Pharmacometabolomic phenotyping also reveals different responses to xenobiotics in animals. Clayton first demonstrated that a pre-dose metabolic profile of urine could predict the toxicity and metabolism of paracetomol in rats . Two classical experimental models demonstrated that pre-dose urinary metabolic profiles had strong association with post-dose pharmacological outcomes .
As mentioned above, increasing numbers of studies showed practical values of pharmacometabolomics in ADMET and predicted individual drug response and variation. Pharmacometabolomics have been widely applied in clinical studies . Trupp et al.  found that metabolite profiling can successfully separate patients into two response groups and could predict LDL-cholesterol response related to simvastatin treatment. Phapale et al.  evaluated the potential of a metabolic phenotype to predict individual variation in the pharmacokinetics of tacrolimus. Using four metabolites, one could successfully predict the plasma AUC (area under concentrationtime curve) of tacrolimus in individualized PK. The results indicated that integrative pharmacometabolimic approach combined with the metabolic profiling of pre-dose urine can serve as a useful tool for personalized drug therapy. Clayton showed that p-cresol sulfate reversely correlate to the level of acetaminophen sulfonation in human urine and indicated the metabolism pathway of acetaminophen . These studies demonstrate that monitoring small molecular metabolites in vivo is possible for the development of personalized health care.
Although pharmacometabolomics have been successfully applied in drug discovery and development, and are now widely used as complementary tools in understanding biological changes after a drug intervention or a gene knockout, it bears the same challenges as metabolomics faces right now . There are three major challenges ahead in pharmacometabolomics.
Firstly, accurate quantification of small molecular metabolites in cell, tissue, organ or whole organism is the cornerstone of metabolomics. For mass spectrometry based pharmacometabolomics, various analytical aspects including: sample preparation/stability, extraction recovery, carryover and matrix effects, choices of internal standards, and quality control should be systematically validated . For NMR based pharmacometabolomics, the spectrum is generally very complex and difficult to interpret due to function-group based multiple peaks per analyte . Therefore, the quality of the analytical data is very critical for both targeted and non-targeted approaches. We should always keep in mind that compromise and biases do exist when analyzing hundreds of metabolites in various matrix samples without efficient separation, adequate recovery and good reproducibility.
Secondly, selected cohort of samples and sample size should be large enough to confirm pharmacometabolomics results. One example is that there are some controversies about the biomarkers in prostate cancer . The level of sarcosine was reported to show significant elevation in biopsy-positive patients in a high impact journal, indicating its potential to serve as a biomarker for prostate cancer [38,39]. However, an independent confirmation in larger cohorts of patients showed negative results for the value of urinary sarcosine as a biomarker for prostate cancer . The underlying reasons may be compounding, but the selection of sample cohorts could significantly contribute to it.
Thirdly, one needs to pay attention to statistical strategies for avoiding false discoveries in metabolomics and related experiments. Many studies failed to take consideration of minimal statistical requirements to obtain true significance by multiple hypothesis testing, inappropriate choice of statistical modeling, and over fitting of experimental data . The classical p-values, p<0.05, are too optimistic when multiple tests are done simultaneously. There is also a challenge of statistically modeling datasets with large numbers of variables, but relatively small sample sizes .
Overall, this editorial is meant to highlight that pharmacometabolomic approach can complement traditional drug discovery, develop processes to identify new pharmacological targets, facilitate drug efficacy and toxicity studies, as well as, reveal new insights into pharmacodynamic and pharmacokinetics. With rapid development of mass spectrometry, NMR techniques and various available statistical methods, pharmacometabolomic will play a more important role through drug discovery and development.