The metabolome is the full set of metabolites—the chemical fingerprints left behind by cellular processes—in an organism. The pharmaceutical industry is particularly interested in the potential of metabolomics research, because metabolites can serve as biomarkers for disease states and drug efficacy, and they can reveal toxicity during the drug discovery process. Metabolomics offers instrument manufacturers a tempting market: despite recent setbacks, the pharmaceutical industry can still be a significant source of revenue. In addition, instrument manufacturers do not need to make radical upgrades of their existing products in order to market them for metabolomics applications. Also, metabolomics research requires a wide array of instruments: gas chromatography (GC) and liquid chromatography (LC) systems are commonly used for separation, but the main analytical technologies for this research are mass spectrometry (MS) and nuclear magnetic resonance (NMR). Because metabolomics research generates significant amounts of data, software also plays an important role in these applications.

The size of the metabolome is smaller than the genome or proteome. One estimate places the number of genes in the human genome at 25,000 and the number of proteins in the proteome at one million; a “first draft” of the metabolome released in January by the University of Alberta indicated that there may be about 2,500 metabolites in the human metabolome. Other researchers have argued that this figure is low, as new metabolites are added with the introduction of foods and other environmental factors. The small-molecule metabolites that make up an organism’s metabolome are of a wide variety of molecular weights and concentrations and vary in polarity, thus presenting a number of analytical difficulties.

MS is perhaps the most common analytical technique used in metabolomics research. It can detect very small quantities of compounds and can be used to determine elemental composition. Prior to being analyzed with MS, metabolites need to be extracted from the tissues, cells or biofluids where they naturally occur. Solid-phase extraction (SPE) and liquid-liquid extraction are typically used for these purposes. Robert Bonsall, research administrator and senior scientist in the department of Plant Pathology at Washington State University, explained that current SPE methods are too limiting, which can prevent unknowns from being detected during later analysis. Following extraction, the metabolites are separated with GC, LC or capillary electrophoresis.

One drawback of MS is that the technique cannot provide quantitative information about certain compounds, but NMR can. John Shockcor, director of Metabolic Profiling Business Development at Waters, explained that MS makes up for these weaknesses with its high sensitivity: “Basically with NMR, although it’s quantitative, you get a subset of compounds and on a really good day you’ll see 30 things. On a really bad day with my mass spectrometer, I’ll see 3,000.” Dr. Shockcor added that quantitation using MS can be given in relative terms using isotope standards, but this method becomes expensive and impractical if it is used for a large number of compounds. Products that Waters sells for metabolomics applications include the Q-Tof Premier Q-TOF MS system, the Waters Metabonomics UPLC/MS/MS system, and Metabolynx, a software package that interprets MS data.

Dr. Shockcor explained that a particular difficulty with MS for metabolomics—and one that Waters is addressing—stems from the inexperience of many end-users with the technique: “The one thing I would like to see is people being more classically trained in all aspects of analytical chemistry. . . . But as an instrument manufacturer, I know that’s wishful thinking. So one of our major goals is . . . ensuring very good to excellent results without having particular expertise in that type of instrumentation.”

NMR provides quantitative data, yields structural information and, in some cases, requires less preanalytical preparation than MS methods. One concern with the technique, however, is its sensitivity, which is primarily based on the strength of the magnet it uses. Clemens Anklin, vice president of NMR Applications and Training at Bruker Biospin, explained that “500–600 MHz tends to be the optimum between sensitivity and cost of the equipment” for end-users using NMR for metabolomics applications. Dr. Anklin said that recent developments in NMR technology, such as combining the technique with LC, are answering these questions about sensitivity, even with NMR systems using midstrength magnets.

Dr. Anklin also acknowledged the need for some specialized training for metabolomics researchers using NMR: “Most of the time, these data are analyzed by statistical methods and you have to really try to eliminate all kinds of variations other than what’s really in the sample. So consistent sample preparation and consistent experimental conditions are really important.” For the future, Dr. Anklin said that improvements in NMR probe technology would be among the developments that would aid end-users in metabolomics applications.

Given the wide variety in the physical properties of metabolites, metabolomics researchers often use both MS and NMR, rather than choosing between one or the other. With each instrument, there is a strong demand for software that can analyze metabolomics data. Bruker Biospin’s metabolomics software, AMIX, is used for spectroscopic and statistical analysis of data from NMR, MS and LC systems. According to Alex Cherniavsky, Global Accounts Manager for Chenomx, a producer of software for NMR data analysis in metabolomics applications, some NMR users have a need for software that can quickly convert their data into formats that can be used in third-party data analysis programs. The Chenomx NMR Suite, which can work with Bruker, JEOL and Varian’s data formats, also includes a number of software tools for interpreting and managing NMR data.

MS manufacturers also provide software for data deconvolution, statistical analysis and visualization, such as Applied Biosystems’ MarkerView software. In addition, compound databases that can accelerate the compound identification process are used with both NMR and MS systems. For example, Agilent offers a desktop version of the Scripps Center for Mass Spectrometry’s METLIN database of metabolite spectra. Dr. Anklin said that Bruker Biospin collaborated with several European firms and universities to create a database of NMR spectra for common metabolites.

Asked about the strongest markets for MS systems in metabolomics, Dr. Shockcor responded, “Right now, I would say universities and pharma are still probably our major customer base . . . I think there’s probably a little bit more interest in the techniques in the universities but it takes them a long time to come up with funding, so right now, pharma is the bulk of the business.” Elaborating on the use of MS for metabolomics in the pharmaceutical industry, Dr. Shockcor explained that the technique was widely used, particularly “in discovery for early toxicity screening.” He added that MS-based metabolomics approaches were least often used in the preclinical segment of drug discovery.

One of the hottest areas within the pharmaceutical industry for metabolomics applications is finding what Dr. Shockcor called markers of efficacy: “If I can come up with two or three compounds that are indicative of a person having early onset of Alzheimer’s, then I can flag that person for treatment with my drug very early on. I could also use those markers and, if they’re reversible . . . I have a marker of efficacy. This is huge: if somebody comes up with something of that nature, it’s going to be worth billions of dollars. So there’s a lot of effort going there in universities and research hospitals and in pharma.” Dr. Shockcor said that sales of MS systems for metabolomics applications were split almost evenly among North America, Europe and Asia.

According to Dr. Anklin, most of the sales of Bruker Biospin’s NMR systems for metabolomics applications are going to the pharmaceutical industry, but he added that demand from academic research labs is increasing. Dr. Anklin explained that NMR has historically been used early in the drug discovery process, but that metabolomics studies of toxicology and drug metabolism have pushed the use of NMR into preclinical and clinical applications. Sales of Chenomx’s products also attest to the importance of the pharmaceutical industry for sales of analytical instruments used for metabolomics applications. Mr. Cherniavsky said that the company’s initial sales were to pharmaceutical companies and that “academic researchers performing biomarker discovery and other pharma-related applications are also becoming significant customers.” But he also said that there has been an increase over the last few years in academic customers doing nonpharma research with Chenomx software.

One example of metabolomics research in nonpharma contexts is Mr. Bonsall’s metabolomic studies of soil bacteria that produce natural antibiotics to protect plants from certain diseases. This research focuses on the chemical signaling that takes place between plants and bacteria in the soil area known as the rhizosphere. Mr. Bonsall uses HPLC for separation and Q-TOF MS, NMR and a photodiode array for analysis. Mr. Bonsall agreed that the tradeoffs between MS and NMR were between sensitivity and structural information, but said that he uses both techniques in his research.

Applied Biosystems, Thermo Fisher Scientific, Waters and Agilent are among the major companies that offer systems combining GC or LC separation with MS analysis. However, only the Bruker companies can offer a full complement of analytical technologies—MS and NMR—for metabolomics research, as well as a software package that can interpret data from MS, NMR and LC systems. Whether this will provide them with great success in the market remains to be seen.