It's not easy being the new kid on the block, but metabonomics is quickly finding its way into the world of Big Pharma. The term, coined by a research group at the Imperial College of Science (London), is defined as "the systemic profiling of metabolites and metabolic pathways in whole organisms through the study of biofluids and tissues."
Some of its potential applications include: preclinical evaluation of drug safety studies, patient stratification for drug treatment, and information on disease processes. Metabonomics also has several advantages over other measurements (genomics, proteomics), such as noninvasive samples (urine, plasma) and identification of individual molecules in complex samples using methods that allow for quantitative interpretation.
However, the major advantage of metabonomics is that the human metabolome contains only about 2,500 unique molecules. "What we're talking about are the small, non-proteinaceous, monomeric molecules made by humans, endogenously synthesized, like glucose and ATP," explains John Ryals, Ph.D., president and CEO, Metabolon (www.metabolon. com).
In order to analyze and identify these molecules, the company has developed a high throughput platform combining high-end LC/MS, GC/MS with proprietary software. Once identified, the molecules are related back to biochemical pathways.
"One of the benefits of this technology is we know the context with how these molecules are synthesized and degraded. So when the amount of a compound increases or decreases, we can understand that change as it relates to biochemical processes."
Dr. Ryals says that many companies are looking for "off-targeted" effects, or unanticipated side effects of drugs prior to clinical trials. "There are a lot of drugs that have many side effects that aren't related at all to their site of action. We can test panels of drugs and find out whether they are off- or on-target."
He adds that companies are also using metabonomics for preclinical work with animals and in early-stage clinical trials. "Many diseases are probably more like syndromes rather than singular molecular problems. We can sub-set patient populations based on how they show up biochemically."
This capability was recently well illustrated through a collaboration with Massachusetts General Hospital, where Metabolon used its platform to identify four unique groups of ALS patients. Sera samples of controls and patients with motor neuron disease were analyzed for total biochemical constituents.
The surprising result was that the analysis was able to differentiate patients with lower motor neuron disease from those with ALS. Further research has identified key biomarkers generic to motor neuron disorders in general, and specific biomarkers unique for ALS and other diseases.
What makes Metabolon's approach unique, says Dr. Ryals, is the ability to identify chemical structure and quantitate molecules, then cast that information against biochemistry. "Other companies are just fingerprinting.' We don't think that is the best way of doing it because you lose a lot of the value of having this type of information."
Building A Better Machine
The Metabolic Profiler combines nuclear magnetic resonance (NMR) and time-of-flight mass spectrometry (TOF-MS) with Bruker Biospin's (www.bruker-biospin.com) analysis software.
Introduced about a year ago, Bruker says the platform is applicable to toxicity and efficacy studies in preclinical and clinical development, as well as to clinical research in disease screening and patient stratification. Another application is the discovery of new small molecule diagnostic markers.
Werner Maas, Ph.D., vp of R&D, explains that NMR is nicely positioned to play a role in this growing field because of its throughput capabilities and the ease in which one can compare NMR spectra and apply statistical tools.
"Even though NMR is not the most sensitive technique, if you look at the overall spectrum of a large number of compounds, it's easy to detect the differences and to predict some type of metabolic event to see changes as a function of toxicological inputs or impulse. Also, NMR has a nice set of parameters to measure those characteristics."
The Profiler's software is in continuous development, states Dr. Maas, as the demands of the field change. "Our clients were demanding specific features in the software that traditionally weren't used in NMR, and this slowly led us into the direction of metabolomics/metabonomics."
The software module has a graphical user interface with integrated control of data acquisition for NMR and MS, a complete statistical tool kit with modeling capability, and integrated data analysis and evaluation on all data set types.
"We think by combining the NMR and MS data, we can gain quite a bit more than by the individual techniques. That's where our efforts are now," explains Dr. Maas. Another area the company is working on is a large-scale MS-NMR database for all endogenous metabolites found in the body.
Electrochemical Arrays Enhance Metabolite Detection
ESA (www.esainc.com) has addressed two of the challenges in metabolomicslarge data sets that make it difficult to find relevant data on marker compounds, and the ability to detect subtle changes to identify such molecules.
The company's Metabolomics System eliminates both problems by only detecting molecules that are redox active. Small molecules are measured by their selective oxidation across an array of proprietary, flow-through, electrochemical sensors. The result is a 3-D metabolic profile with known and unknown analytes that can be used to characterize changes in the metabolome.
"It focuses only on compounds involved in oxidation-reduction processes, like cellular metabolism, and typically these are exactly the kinds of molecules people are looking at from a metabolomics standpoint," explains Darwin Asa, Ph.D., marketing manager. "It also allows you to see a lot more of the low abundance marker compounds that you can't see with other techniques."
This system includes a 16-channel CoulArray Detector, auto sampler, interface hardware to MS or LC-MS, and data analysis software. Although the company has made single cell electrochemical detectors for many years, the array format grew from customer demand.
"We had customers who wanted to analyze six to ten different compounds in one sample at a time. The only way to do that was with the array format," says Dr. Asa. He explains that each array has multiple cells, each set to a different redox potential. The sample flows through these cells, and the molecules are separated by their redox potential. It is now capable of analyzing thousands of molecules simultaneously.
Although this technology cannot identify specific biomarkers, it can quickly (in a few seconds) identify changes from one sample to another. "These electrochemical detectors are really good at identifying key biomarkers that are changing in your metabolomic sample."
Data analysis is easier than with MS or NMR because data can be segmented into smaller pieces and the differences between samples can be quickly determined without having to identify everything in the sample.
New Software For Data Analysis
Most scientists in the metabolomics field agree that NMR remains a useful tool. However, "there hasn't been much in the way of software tools to make it faster, more convenient, and more accurate," explains Neil Taylor, CEO, Chenomx (www. chenomx.com).
This is where the company's Eclipse Version 3.0 comes into play. Originally developed two years ago, the software has been recently upgraded based on customer feedback.
Taylor says most NMR researchers use Principal Components Analysis (PCA), which evaluates NMR spectra and looks at the differences between the actual spectral peaks.
"We found that these peaks tend to move around with different conditions of the sample. So when you do a comparison between two samples, it doesn't show the peaks in the exact locations. We built complex software to predict where those peaks will move."
In addition, the software provides a method for quantitative analysis of these compounds. "We analyze the spectrum for the actual components, the concentration of the chemicals present. We produce a table of compounds and their concentrations and then the user can compare those concentrations rather than comparing a spectra," he states.
The company has analyzed about 200 compounds to date and plans to add new compounds to its database. The software will continue to be improved for faster analysis and to provide more knowledge from the data. Current Eclipse applications include biomarker discovery and toxicology and safety testing for drugs.
"We really don't see ourselves as just a software company because there's so much science in our product. We think the biggest future applications for our software will be diagnostics and personalized medicine," summarizes Taylor.
Pioneers of Metabonomics
Metabometrix (www. metabometrix.com) is also working to enhance NMR spectroscopic and data-processing methods. Collaborating with the academic group at Imperial College that pioneered the use of NMR spectroscopy of biofluids coupled with chemometrics, the company has developed a proprietary platform of metabonomics technologies.
"If one chooses the right tissue and the right analytical technique, metabonomics will be able to classify samples according to toxicity and say something about the biomarkers that distinguish the classes," explains John Lindon, Ph.D., professor of biological chemistry at Imperial College.
"Looking at biochemical changes, which is what metabonomics does, the technique is likely to be more sensitive and specific than many others. Metabolic changes are in the real world, unlike gene expression changes, which may not have any relevance to real biochemistry," he adds.
Dr. Lindon states that what makes Metabometrix' approach unique is that they are able to tie any metabolic analyses to real biological and biochemical interpretations. The firm's platform uses a mixture of commercially available software with methods and refinements developed in house.
The company has performed studies for many major pharmaceutical companies in preclinical and clinical studies. It has also identified proprietary clinical markers. "We have been working in a number of key disease areas and have been successful in generating valid statistical models for disease diagnosis."