|SEND TO PRINTER|
Feature Articles : Feb 15, 2006 ( )
Metabolomics Plays Crucial Discovery Role
Hastening the Early Stages of Drug R&D Could Save Time and Money
Metabolomics is emerging as a significant player in the drug development arena. Metabolomic analysis provides a biochemical snapshot of the small molecules produced during cellular metabolism, such as glucose, cholesterol, ATP, and lipid signaling molecules. The exact number of human metabolites has yet to be established, but some estimate it to be between 2,00010,000.
Since the metabolome directly reflects physiological status, it can biochemically monitor disease states and assess drug actions. Knowing early-on how drugs impact biochemistry would be a significant advantage. Impressed by its potential, the NIH has earmarked more than $70 million over the next five years for R&D projects. Interest by big pharma also is growing.
Many companies are conducting studies to evaluate metabolic profiling in their projects, uncover novel biomarkers, improve instrumentation, and enhance bioinformatics for data analysis and sharing, as evidenced by a number of presentations at Cambridge Healthtechs recent conference on metabolic profiling in Orlando.
Combining technologies, such as mass spectrometry (MS) and liquid or gas chromatography (LC or GC), can provide a powerful means for resolving a wide ranging number of metabolites.
Icoria (www.icoria.com) is using a combined LC/MS and GC/MS metabolomics approach for biomarker identification. We are focused on the discovery of novel, multiparameter biomarkers for early disease diagnosis using a multiplatform approach, states Alvin Berger, Ph.D., head of biochemical profiling. We are applying this strategy to various diseases, including liver disease, diabetes, obesity, and cancer. We recently demonstrated how these technologies could be applied to find biomarkers associated with clofibrate safety and efficacy.
Clofibrate has two actions, explains Dr. Berger. It lowers lipid levels but also can cause hepatotoxicity in rats, particularly at higher concentrations and with longer durations of exposure. It has rarely been studied using metabolomics but has been extensively studied using microarrays and transcriptomic approaches.
The aim of our studies was to determine if these two actions could be distinguished by using metabolomic approaches. First we used a broad lipophilic GC/MS strategy that we believed would identify mostly biomarkers of clofibrate efficacy, specifically compounds causing increases in mitochondrial and peroxisomal fat oxidation. We also undertook a broad hydrophilic LC/MS and LC/MSMS, or tandem mass spectrometry, metabolomic approach that we felt would favor the identification of biomarkers associated with potentially toxic effects in the liver, including oxidative stress, enlarged liver weight, and hyperplasia.
Dr. Berger indicates that Icorias studies demonstrate how the use of metabolomics, especially early in drug discovery, can help distinguish viable from nonviable drug candidates.
HTS and Metabolomics
Incorporating metabolomics in the early stages of drug discovery may save time and money, according to George G. Harrigan, Ph.D., co-founder, director, and treasurer of the Metabolomics Society. The attrition-adjusted costs in early discovery are enormous, Dr. Harrigan says. Introducing metabolomics early in the exploratory process would minimize costs by maximizing effective compound/target selection.
Dr. Harrigan describes how high-throughput screening linked to metabolomics can help. Many high-throughput screening campaigns centered on lipid enzymes, for example desaturases and prostaglandin synthases, often exploit LC/MS to confirm leads identified in more conventional screens.
Such orthogonal assays directly measure enzyme substrates and products. We demonstrated how this can be expanded to incorporate additional pathway-related metabolites when applied to biological systems. Enzyme product/substrates ratios can be regarded as mechanistic biomarkers of drug efficacy, while more comprehensive pathway profiling can further support efficacy claims or identify the potential for side effects.
In an example of such profiling, Pfizer (www.pfizer.com) collaborated with Lipomics Technologies (www.lipomics.com) to generate a pathway analysis of atherosclerosis-susceptible knock-out mice versus a nonsusceptible wild-type strain. We fed both sets of mice a diet designed to increase cholesterol and then characterized their lipids in plasma, adipose tissue, and liver at two-week intervals over 16 weeks, explains Dr. Harrigan, using gas chromatography coupled with flame ionization detection we measured 500 identified lipids in each sample.
This information allows us to identify dynamic changes in numerous pathways. Measuring cholesterol is pretty trivial, but our metabolomics approach revealed key metabolic pathways that were differentially regulated in an atherosclerosis-susceptible mouse. The more you can understand activity in models used, the more you can use that information to validate targets early on in the process and assess the impact of screening leads on them.
Characterizing Animal Phenotypes
How can scientists using animal models for drug-response studies distinguish between processes associated with disease mechanisms versus compensatory physiological adaptation? Metabolomics can help, says Matej Oresic, Ph.D., chief research scientist, VTT Technical Research Centre of Finland (www.vtt.fi), who leads research in the domain of quantitative biology and bioinformatics.
We characterize phenotypes of animal models in preclinical and basic studies, combining metabolomics with other levels, such as basic phenotyping and transcriptional profiling. The use of metabolomics is particularly helpful in translating complex phenotypes observed in preclinical studies into the clinical domain, as well as for studying distributed multitissue effects of interventions, says Dr. Oresic.
VTT functional studies using animal models identify function-specific genes and metabolites involved in homeostatic as well as toxicological responses. The institute primarily employs LC or GC coupled to MS to analyze metabolite samples, followed by a bioinformatics platform for data mining and processing.
Dr. Oresic suggests that two of the most pressing challenges in the metabolomics arena are making sense of the mounds of data generated, as well as exchanging that information among scientists. Our institute is pursuing a strong informatics effort. Everyone is floating their own data, but what is lacking is an appropriate standardization of metabolomics data, so that we can better compare and validate each others findings. We also need a better understanding of the metabolic pathways and the mechanisms that are leading to changes in metabolite levels and fluxes. We need to have these issues solved before metabolomics is more broadly utilized and accepted in the domain of drug discovery.
VTT is helping with the solution. They are developing analytical methods for metabolic profiling, software tools for processing data, and establishing a bioinformatics infrastructure.
MS as a versatile metabolomics technology can serve as a stand-alone method for identifying compounds from complex mixtures or can create libraries identifying metabolic fragmentation patterns, useful for establishing a samples fingerprint.
Metabolon (www.metabolon. com) is focusing on the quantitative measurement and identification of the repertoire of biochemicals contained in biological samples. Mass spectrometry is a tremendous improvement over other methods, such as nuclear magnetic resonance, in that we can find hundreds of metabolites per sample and do so in the femtomolar range, points out Chris Beecher, Ph.D., vp biochemistry and technology. Also, you only need small amounts of sample. But we also have found that no single technology will provide the entire answer. We couple MS with LC and GC separations and merge all of this data using informatics to reveal novel metabolic fingerprints.
To prove its point, the company recently performed a study in collaboration with Massachusetts General Hospital to identify novel disease biomarkers for amyotrophic lateral sclerosis (ALS). It conducted a metabolomics analysis on blinded plasma samples from patients treated for motor neuron disease versus control patients.
We measured metabolites quantitatively and clearly saw that there were multiple populations within the ALS population, reports Dr. Beecher. We defined subsets as to whether patients were or were not taking the drug riluzole. We were able to identify physical subsets of disease and see the effect the drug had on patients. This drug was seriously altering patients blood chemistry.
This technology may provide answers as to whether specific drugs can revert patients to more normal conditions or if they cause significant side effects. This could identify which drugs are the best candidates to develop. Also, such information may tell us which pathways are impacted, and that would help identify mechanisms. The FDA has shown great interest in this area. They also feel it may provide insights into drug action that have never been possible before.
Metabolon is expanding its study to include hundreds of patients in a longitudinal investigation to follow disease progression and is initiating similar studies in Alzheimers and Parkinsons diseases.
So-called targeted approaches focus on preselected metabolites reflecting relevant metabolic pathways. This strategy has the advantage of honing in on specific metabolites that can be linked to a specific therapeutic area. In principle, this also could increase the analytical rigor and precision for identification and analysis.
Biocrates Life Sciences (www.biocrates. at) is pursuing targeted approaches utilizing tandem mass spectrometry (MS/MS) for metabolomic analysis.
MS/MS has a number of benefits, according to Armin Graber, Ph.D., CEO, including the ability to measure low nanomolar concentrations of metabolites in complex mixtures without needing chromatographic instruments. This allows high-throughput methods for accurate identification and quantification of arrays of metabolites in multiple pathways.
The company recently demonstrated the performance of its technology by identifying biomarkers characterizing a disease treated with an already approved therapeutic and a candidate drug class. It compiled comparative pharmacodynamic profiles for a PPAR-gamma agonist and the novel compound in a mouse model for diabetes mellitus type 2. PPAR (peroxisome proliferator-activated receptor) is an orphan nuclear receptor that plays a key role in regulating triglycerides, blood glucose homeostasis, and insulin resistance.
The results were so informative that the pharmaceutical company that asked us to perform this study launched a program for a new drug indication and will enter clinical development in 2006, says Dr. Graber. Furthermore, a joint research collaboration was initiated including several biomarker identification and validation projects.
The task of data interpretation, especially when applied across different models, can be both complex and daunting. Lipomics Technologies is developing proprietary analytical technologies and bioinformatics tools for accurate and quantitative profiling of lipid metabolites.
Lipid profiling has undergone a rapid expansion recently, because there is a growing awareness of the central role lipid metabolism plays in metabolic and inflammatory disorders, states Steven Watkins, Ph.D., president and CSO.
A major challenge of metabolomics today is the need to integrate data from different metabolomic platforms. We also believe that a logical and productive step forward is the use of more focused platforms that can produce quantitative data on known metabolites.
Our technologies target known lipid metabolites and define quantitative changes in panels of samples. For example, our TrueMass analysis produces lipomic profiles that measure hundreds of lipid metabolites from samples such as tissue, plasma, or serum. Because the resulting data are quantitative, TrueMass data can be integrated with databases. Additionally, Surveyor identifies statistically significant differences between the lipomic profiles of experimental groups and presents the data as an interactive heat map format.
According to Dr. Watkins, bioinformatics data linked to biological context must provide another critical aspect for drug discovery-translation. Results must be generalizable to other biological situations or populations. The great advantage of metabolites is that they are largely translatable from preclinical to clinical situations.
© 2012 Genetic Engineering & Biotechnology News, All Rights Reserved