Metabolomics - an important emerging science

Metabolomics is generally accepted as the study of the repertoire of non-proteinaceous, endogenously synthesised small molecules present in an organism. The metabolome refers to the catalogue of those molecules in a specific organism, eg, the human metabolome. As with any emerging science there is some controversy regarding nomenclature, but the newly formed Metabolomics Society seems to be adhering to a convention consistent with the other '-omic' sciences in this regard:

• gene - genome - genomics
• protein - proteome - proteomics
• metabolite - metabolome - metabolomics

As shown in Figure 1, according to the 'central dogma' of molecular biology, DNA is transcribed into RNA, which is then translated into proteins. While the exact numbers are still not determined, there appear to be about 37,000 to 40,000 genes encoded in the human genome.

Genes are transcribed into RNA molecules and, following different types of RNA processing, there may be 125,000 to 250,000 unique transcripts represented in the transcriptome. These RNAs are translated into proteins and, following post-translational processing of the proteins, there could be as many as 1,000,000 unique proteins represented in the proteome. In a sense, one can view DNA as coding compressed information in a way analogous to compression of computer files.

These multitudes of proteins are organised into elegant and intricate signal transduction pathways that function to perceive inputs and trigger outputs. The inputs can be highly varied, from hormone or neurotransmitter signalling to changes in the physical environment. The pathways themselves can be extremely complex and involve many different proteins. It is not unusual for certain pathways to modify other pathways, so many of the proteins in the proteome are components in a complex web of events.

However, the ultimate outcome of these signalling pathways is that metabolic enzymes may be up- or down-regulated, and this influences the synthesis or degradation of the small molecules. In metabolomics, we measure the repertoire of small molecules in a sample (eg, cells, tissues, organs, organisms) to understand more clearly what has changed in a system.

The human metabolome

One of the most important advantages of metabolomics is that the human metabolome is relatively small. We can find evidence for only about 2500 unique molecules. There have been many confusing statements around this number, with some investigators speculating that there must be tens of thousands to millions of small molecules in the human metabolome. However, in the tree of life, humans are not very biochemically complex. They lack the ability to make secondary products and even lack certain aspects of primary metabolism such as the ability to synthesise essential amino acids and vitamins.

The NAPRALERT (NAtural PRoducts ALERT) database, probably the most complete database listing all known biologically synthesised small molecules, has in the order of 140,000 entries. Most of these molecules are the products of intermediate and secondary metabolism from plants, fungi and bacteria.

Humans have little capacity for intermediate and secondary metabolism and so have very few of these compounds.

Figure 2 shows a mass distribution of a portion of the human metabolome comprising approximately 1400 molecules. Each data point represents a window of 20 daltons. The peak of compounds around 180 daltons contains many of the sugars in the metabolome, the peak around 360 daltons contains many cortisones and prostaglandins, and the largest of the molecules represent the dolicols which anchor proteins to membranes. However, it is clear that most small molecules are below about 600 daltons in mass.

The measurement of small molecules can be very precise

A second powerful benefit of metabolomics is that the individual molecules in the complex mixtures can be identified with certainty. For example, by using a combination of chromatography followed by mass spectrometry, the individual molecules can be separated and identified based on their mass. By using appropriate standards a molecule can be identified with great precision.

This level of precision is even more significant when one considers that glucose is the same molecule when measured from a bacterium to a fly to a human. There is no need to employ arguments regarding homology or homeology as is necessary for genomic or proteomic analysis. Therefore, comparative studies have more precision and become less interpretive.

Metabolomics can be quantitative

One of the limitations in RNA profiling or protein profiling is that the methods are not quantitative. This lack of quantitation greatly limits applications of the various approaches. For example, contemporaneous controls must always be run and in many experiments much of the effort involves analysing controls. Metabolomics can be performed in ways that allow for quantitative interpretation. This opens up the ability to apply statistical methodologies and databasing approaches that have only limited value in other technologies.

Biochemistry is relevant for disease and drug action

Finally, metabolomics as a measure of biochemistry is a more direct measure for a disease state or the action of a drug. Disease states result ultimately from a change in the biochemistry of a system, and most drugs act at the level of biochemistry. Therefore, measuring the biochemical status is very relevant to the understanding of how disease is manifest, how drugs work and who is responding (or not responding) to them.

Using transcriptomics or proteomics as a measure for drug action requires that the change in biochemistry causes a change in gene or protein expression, and it is not evident that this often happens.

In summary, metabolomics has many benefits over current technologies like genomics, transcriptomics and proteomics including reduced complexity, greater precision, better quantitation and higher relevance.

Tissue specificity of the metabolome

Another characteristic of the metabolome is its tissue specificity. It has been known for many years that certain molecules are only found in certain cells, tissues or organs. It is also accepted that there is tissue specificity among the genes that are transcribed in a specific cell or tissue. This leads to tissue specificity in the proteome and also in the metabolome. Certain highly specialised, terminally differentiated cells may express only a small subset of the entire metabolome, making the measurements and analysis even more simplified.

Use of metabolomics in medicinal chemistry

As a practical example of how metabolomics may be used in drug development, we have carried out a case study with a group of antibiotic compounds. These compounds are directed against a target enzyme found in bacteria but not in humans. In experiments where either sub-lethal or lethal concentrations of the drug are added to bacterial and mammalian cell cultures, a certain metabolite was dramatically reduced in bacteria, as was expected (see Figure 3).

In the mammalian cells, there was no effect of the drug on the same metabolite, again as expected, since the target enzyme is not present in the mammalian cells. However, when other metabolites are evaluated in mammalian cells, some of the drugs were shown to be free of unanticipated metabolic effects while others clearly have unanticipated activities.

In this case, the metabolomic information has been used to direct the synthesis of drugs that lacked the unanticipated side effects. The expectation is that drugs with fewer unanticipated side effects will have a higher likelihood of safety.

Challenges moving forward

Challenges for the future advancement of metabolomics range from reproducible sample preparation to data visualisation. The amount of data pouring from mass spectrometers is daunting. The way the data are presented to the end-user, the coordination of multiple data streams, quality assurance and quality control and standardisation are all hurdles to establishing a useful methodology. Once measurements are made and the data are processed, the volume of data that must be analysed in a study can occupy thousands of pages of spreadsheets, so we need reliable data visualisation tools.

Finally, there is the significant issue of biological variability among individuals in populations or among independent biological samples. Dealing with this variability will likely create a need for the application of sophisticated statistical methodologies that are foreign to many biologists. Finally, as we gather more metabolomic data on varied applications, we may uncover new scavenging and shunting pathways that have not been previously described.

Outlook

The emerging science of metabolomics is certainly in its early days. Metabolon has established a robust, quantitative, high-throughput methodology that can identify many of the molecules in a sample and project that information back to biochemical pathways. The availability of this technology should have a dramatic effect on drug discovery and drug development by increasing the efficiency and safety of drugs, and decreasing the time and cost for the approval process.

* Dr John Ryals is the president and chief executive officer of Metabolon, Inc. He has 25 years of experience in the biotechnology industry. Prior to joining Metabolon, he founded Paradigm Genetics, Inc, a pioneering systems biology company. Previous positions include vice-president of research for Novartis Crop Protection, Inc, and head, agricultural biotechnology, Ciba-Geigy Corporation. Dr Ryals has published over 100 scientific papers and is an inventor on over 30 issued US patents. He is an Adjunct Professor at North Carolina State University and serves on the Board of Directors of the university's College of Physical and Mathematical Sciences.