Metabonomics is primarily concerned with the observation of small molecules within a biological sample and emerged because of the necessity for novel approaches to genetic and biomedical research. From the conception of molecular biology techniques during the early 90s, in particular after the decoding of the human genome, it was generally believed that the key to understanding disease processes was through characterisation of genetic variations and gene expression. However, the complexities underlying biological mechanisms and their interactions with the environment required a shift in research focus. This led to the development of various array based techniques, allowing researchers to investigate large numbers of genetic variations within a sample. Unfortunately, this approach produced results with limited therapeutic usefulness and few clinical applications. The utility of many current genetic techniques may, therefore, have reached their natural limit, leading to the development of novel approaches in this area, one of which is Metabonomics.

Metabonomics has been defined by Nicholson1 as "the quantitative measurement of the dynamic multiparametric metabolic response of living systems to pathophysiological stimuli or genetic modification". In essence, Metabonomics quantitatively measures changes within metabolic processes during an alteration of the ‘normal’ state due to a disease or other intervention. This field sits closer to the phenotype than conventional studies, as the collection of all metabolites in a biological cell is effectively the end point of the biochemical processes of that cell. Metabonomics is a ‘top-down’ systems biology approach that utilises global profiles describing the biological status of an individual. By studying the changes within the biological ‘system’ it is possible to efficiently model complex, multi-stage processes such as disease progression or drug metabolism, and correlate them to biomarkers.

In complex organisms, levels of biomolecular organization (genes, proteins, metabolites) and their control are inter-dependent and affected by environmental events. Changes at any level are presented to the world as biological ‘endpoints’; metabolic changes are an example of such endpoints and can be used to extract information of diagnostic or prognostic value. Many disease processes such as cancer and heart disease can be studied using Metabonomics. This research has identified metabolic profiles of cancer cells that have proved to be a useful tool for understanding tumour development and progression2, whereas biomarkers have been identified for obesity as well as heart disease and vascular disease3.

Metabonomics [...] emerged because of the necessity for novel approaches to genetic and biomedical research

Another core result of Metabonomics has been the characterization of various animal models for human disease achieved through metabolic profiling. Metabonomics is used to generate strains of phenotypes from either experimentally altered genotypes (transgenic models) or those derived by selective breeding. This allows the researcher to understand the biochemistry of their experimental model. Furthermore, the technology allows for the discrimination between phenotypically similar strains, which has allowed a deeper understanding of the underlying genetic mechanisms.

Metabonomics in Toxicology

Perhaps the most widely used application of Metabonomics in research is concerned with testing the adverse effects of chemicals on living organisms. This is especially important when testing a new pharmaceutical drug’s toxicity. Understanding the mechanisms of toxicity is a challenging endeavour. In the past the organ specific toxic effects of a drug have eluded detection in clinical trials and have only become apparent after the drug’s introduction into the market. Examples include Vioxx (rofecoxib) for pain and osteoarthritis and the weight-loss drug Acomplia (rimonabant), both of which had to be withdrawn due to safety concerns. The underlying issue is that while new drugs are tested in various in vitro, cell and animal models, the resulting data is not always applicable to humans. Furthermore clinical trials do not always reproduce real-world administration of medication and while clinical trials are typically very large, the sample size may often still be insufficient to detect rare side effects.

Metabonomics is important in this context because it establishes the ability to detect biomarkers of drug metabolism. Initially, known toxic effects can be investigated using a traditional metabolic approach. With this data, it is then possible to predict a toxic event in a novel drug before the occurrence of clinical events (biomarkers of early effect), to evaluate the severity of the poisoning (biomarkers of effect), and also to monitor exposed patients (biomarkers of exposure). Critically this allows for toxicity to be detected far earlier in the development process, leading to more efficient selection of drug candidates for further clinical trialling. Additionally, this generates significant financial benefits to the pharmaceutical company.

It is possible to provide unique real-time phenotypic information on the tissue state of the patient

In 2005 the Consortium on Metabonomic Toxicology (COMET) was established to develop expert models for the identification of toxicity based metabolic analysis, led by the Department of Surgery and Cancer at Imperial College with participation from several pharmaceutical companies. So far more than 200 drugs have been screened for toxicity in this project. Among these studies, acetaminophen (also known as N-acetyl-p-aminophenol, APAP) is frequently used as a model drug for liver injury (hepatotoxicity). By understanding the metabolic pathways of these compounds, it is possible to stipulate possible mechanisms of toxicity.

Clinical Applications

The health state of the body is directly linked to the metabolic activity of the individual; therefore, changes within the homeostasis and chemical equilibrium are reflected in the metabolic profile of that individual. It would be of great clinical benefit to measure the metabolic activity of an individual during a surgical treatment so that prognostic or diagnostic information can be increased. By extending some already well-understood principles of magnetic resonance imaging (MRI) to the Metabonomic field, it is possible to provide unique real-time phenotypic information on the tissue state of the patient. For example chemical fingerprints of tissue samples taken during surgery could be used to immediately analyse the type of tissue or disease state, greatly enriching the information available to the surgeon.

Scientists and clinicians have long since realised that not all patients respond equally to various treatments. Pharmacogenetics, which studies the genetic differences of metabolic pathways that affect an individual’s response to drug intake in terms of therapeutic effects and toxicity, has been established to address this problem. Information predicting drug response has led to the development of individualised treatment plans and patient stratification into responder groups. However, genetics is not the only factor that contributes to differences between patients’ response to drug treatment. Each human should be regarded as a complex ecological being. In addition normal metabolic states must be recognised as a complex continuum encapsulating both genetically and enzymatically controlled metabolism. Using this paradigm for future treatment strategies will pave the way towards truly personalised healthcare.

Integration of the Omics

For a comprehensive view of the biology of an organism it is necessary to identify and characterise all levels of biomolecular organisation, and correlate the various components in a network of overall interaction (what is now being called the ‘interactome’). The intuitive way to proceed would be to directly integrate the various ‘-omics’ into a global framework. However this poses various problems and there are many hidden stumbling blocks, mainly due to the complexities of cellular interactions.

Metabonomics provides complementary information to genomics and proteomics in the understanding of biological systems.

Various combinations of ‘-omic’ technologies have been attempted including integration of genomics, transcriptomics, proteomics, localizomics and phenomics. Metabonomics has also been a part of this integration, and combined studies with genetic data have become commonplace. It is possible to use measurements of metabolic concentrations and traditional genomic mapping to pinpoint the precise genetic contribution to phenotype. This approach has been named metabolomic QTL (mQTL) mapping and was first introduced in plant and rodent models4, and recently in humans. Metabolomic data is also routinely correlated with gut microbial data, especially in the study of obesity and of nutrition on the human genome and health. These extensions and combinations of technologies confirm the important role of metabonomics in the systems biology field.

While the use of metabolic profiling in toxicology and pathology is still in its infancy, this technology provides an additional layer of detail to biological or pathological change. In the future, development of ways to increase output coupled with better understanding of cellular metabolism will allow for the quick and precise description of biological status. Metabolic profiling can also complement other emerging technologies such as proteomics and transcriptomics, allowing for a complete systems view of biology. In the future it is expected that profiling will be able to be used in the clinical setting, as well as for the investigation of any change in the normal metabolic state of an organism.

One of the biggest remaining hurdles is the challenge of structural identification. This challenge arises due to the difficulty of distinguishing between similar compounds and the extensive training necessary to be able to perform the analyses. Initiatives such as a community-wide common database cataloguing various metabolites, along with their spectral representations and their associations, similar to what is currently available for genes, would allow for quicker identification. Furthermore the development of automated computational tools of identification and modelling would also expedite biomarker discovery.

Like many emerging technologies, metabonomics is associated with very high expectations, but has yet to reach maturity and be considered a mainstream research area. However, metabonomics has been readily adopted by industry, while the academic community is just starting to catch up. This is due to the fact that it is able to close the biological circle of gene to protein to metabolite; furthermore it allows the opportunity to work from the phenotype end-point backwards, which may prove easier than trying to untangle the distal effect of thousands of nucleotide changes and the downstream modifications. Further research and resources are needed for this exciting field to reach its potential.