September 15, 2008 (Vol. 28, No. 16)

Early Detection of Diseases Could Be Substantially Aided by Metabolomics

Metabolomics is a member of an unruly menagerie of omics creatures, whose pedigrees trace back to the grandfather of all omics-genomics. Omics is now defined as an all-encompassing study of some area of the biological sciences. This includes what appears to be great stretches of linguistic invention as well as a vast capacity for broad generalization such as the speechome, the exposome, and the metallome.

Currently, there is great interest in metabolomics, which is defined as the comprehensive set of low molecular weight metabolites comprising a biological sample—such as a single organism, a group such as a population of individuals or a species—and the interactions that drive the system and provide its dynamic collection of interactions.

To invest some order to its study and evaluation, investigators at the Universities of Alberta and Calgary recently completed the first draft of the human metabolome, consisting of approximately 2,500 metabolites. This ome, or totality of the metabolome, is available at the Human Metabolome Database. It is far from complete, however, so it can be called an ome only in a partial or a relative sense.

In contrast, much more is known about the metabolomes of other organisms, especially plants. Metabolomics is quite a different undertaking from proteomics or genomics, in which one characterizes the actual genetic makeup of a living system.

The metabolome could represent drugs and other exogenous agents and an army of small molecules whose presence will change radically reflecting the particular metabolic and environmental state of the subject. The biochemical charts of metabolic pathways that identify the thousands of intermediates across the cell and their relative levels form the blueprint of the metabolome.

The state of field will be discussed at the Metabolomics Society’s annual meeting in Boston this month. Some of the speakers gave GEN a preview of their presentations.

Endogenous Metabolites

“In all stages of biomarker discovery, it is extremely important to know the identities of differentially expressed metabolic components as compared to viewing them solely as statistical objects,” states Robert Mistrik, Ph.D., and CEO at HighChem, who will discuss the challenges posed by structural profiling of endogenous metabolites.

Metabolic profiling, a widely used tool today, compares two groups of individuals —a sample of normal controls and a matched group of diseased individuals. Scanning the levels of many metabolites, profiles for individuals can be obtained and differences noted between the various groups. Consistent alterations in metabolite expression can reveal significant features of a disease process not obtainable through other approaches.

The company’s main thrust is the development of software solutions for crunching complex analytical data. The analysis is aimed at low molecular weight substances, that is, nonpeptide or proteinaceous materials. Other areas of interest include comprehensive libraries of fragmentation knowledge and experimental spectral data.

Much of the company’s efforts involve the development of an accurate mass search tool for analyzing structural libraries. The traditional approaches adopted by HighChem include spectra searches against reference libraries, multivariate pattern-recognition methods, biochemical pathway analysis, and direct spectra interpretation.

To facilitate this analysis, HighChem is pursuing an extension of its spectral library of endogenous metabolites referred to as spectral trees, according to Dr. Mistrik. Mass spectrometers are usually coupled to either liquid chromatography or gas chromatography devices, which yield data with divergent properties resulting from the different features of the two separation technologies.

HighChem software allows the spectral trees to be reconstructed from data files, extracting and comparing informative data from both LC and GC characterization. “It is possible to store and search trees in libraries, to annotate every node spectrum or to create chromatographic libraries with spectral tree components,” states Dr. Mistrik.

One of the important collaborative efforts that HighChem is involved in is the METAcancer project, a large consortium currently investigating breast cancer in the EU. The company is investigating new biomarkers for early detection of breast cancer as part of the project.

Focus on Prostate Cancer

Metabolon is developing diagnostic tests based on metabolomic profiling for prostate cancer, diabetes, and other diseases, according to Jeffrey Shuster, Ph.D., director of diagnostics development. The company has built a platform using gas and liquid chromatography in combination with mass spectrometry for identification of metabolic differences between normal and affected individuals. Ideally, the populations should be balanced with respect to age, sex, race, diet, and lifestyle, although in the real world this may not always be possible.

“Through metabolomics, we believe that we can bring new diagnostics to market faster and more efficiently by sidestepping much of the research effort required to characterize new markers by traditional methods,” Dr. Shuster explains.

Metabolon is also focused on prostate cancer. The company is developing profiles that will discriminate normal and cancer-ridden individuals on the basis of metabolites in urine and serum and also analyze the differences between indolent and aggressive tumors among the positive cases.

The primary goal of the project is to reduce the number of biopsies that are performed while at the same time producing a more accurate diagnosis. Another goal is to produce metabolic profiles that ascertain the level of aggressivity.

Stroke Models

“In genomics and proteomics, data is quite species specific,” says Hans-Peter Deigner, Ph.D., director of biomarker research at Biocrates, “but in metabolomics investigations, studies on animal models translate well to humans.”

According to Dr. Deigner, metabolomics measures metabolite differences in biological fluids and tissues, providing the closest link to the various phenotypic responses. The goal is to span the gap between genotype and phenotype through an analysis that joins the individual biochemical reaction to the critical factors of drugs, nutrition, and environment. Since metabolites are the quantifiable molecules with the closest link to phenotype, toxic responses to a drug or disease prevalence are predicted by differences in the concentrations of the relevant metabolites.

The Biocrates team is investigating approximately 20 lipid analytes that vary in response to bleeding and occlusion brought on by stroke. “In addition, we follow the effects of aging in our mouse model, since the reaction to stroke changes as a function of age.” Using the mouse model, the investigators are able to sample brain tissue as well as serum from affected animals, and correlate these observations with data from patients.

Dr. Deigner and his colleagues are well aware of the limitations of omics studies. “We have learned from other’s failures and have built a robust statistical framework and analytical process for biomarker validation,” he says. “Because the number of individuals needed for a meaningful clinical trail is critical, we are working closely with our partners to ensure that we have a large candidate base sufficient to properly validate our studies.”

Markers for AIDS

In the course of an AIDS infection, the HIV virus can infiltrate the central nervous system and bring about dementia and encephalitis. With antiretroviral therapy, patients are living longer and may encounter significant neurocognitive disorders. According to William Wikoff, Ph.D., research associate at Scripps Research Institute, metabolomic profiling has the potential to provide early diagnosis of neural involvement and to predict the progression of the condition to a full-blown CNS disease. Moreover, it may eventually be possible to determine response to treatment and provide insights into the pathogenesis, treatment, and prevention of AIDS.

Although there have been numerous proteomic investigations of urine, serum, and CNS fluid from patients, these studies have provided few insights, driving the search for alternative screening tools. For neuroinfectious, neurodegenerative, and psychiatric disorders, the cellular and phenotypic complexity of the brain has hindered biomarker identification. Dr. Wikoff and his colleagues have invoked a new strategy, focusing on metabolomic studies to reveal, until now, undiscovered markers.

The Scripps group uses SIV infection of rhesus macaque monkeys as a model for human AIDS. The animals show the same pattern of symptoms, including dementia, CNS degeneration, and encephalitis observed in human AIDS patients. The experimental design that the team followed involved collecting cerebrospinal fluid before and after viral infection.

“The screening platform used a nontargeted, mass-based metabolomics approach with online data base screening to identify metabolites,” explains Dr. Wikoff.

Because the blood-brain barrier is partially compromised during HIV infection, the concentration of albumin and many metabolites such as fatty acids and phospholipids are increased. For instance, carnitine and acylcarnitines are elevated in infection as are various other unknown metabolites.

The Scripps researchers have searched for molecules that are specifically altered in monkeys showing encephalitis in the course of their SIV infection. They subsequently performed a microarray analysis to track down changes that relate to lipids or lipid processing, and identified increased expression of phospholipases in the brains of animals with encephalitis. Through the use of in situ hybridization and immunohistochemistry of brain tissue from infected animals it was determined that elevated phopholipases occurred in astrocytes and in the ventricles of the choroid plexus.

“We believe the identification of specific metabolites illustrates the potential of mass-based metabolomics to elucidate neurodegenerative diseases,” Dr. Wikoff concludes.

Various omics approaches have been used to develop databases of biomarkers for a number of years, yet it has proven extremely difficult to assemble consistent bodies of information that can be used to develop new disease biomarkers. In numerous cases, studies cannot be replicated and putative biomarkers cannot be validated. There are a number of reasons for these failures, but inadequate sample size, failure to control methodological variation, and inadequate quality assurance have been contributing factors.

Success within the omics field is absolutely essential for the development of a robust biomarkers. If we have learned anything from the advance of biomedical science in the last half century it is that diseases that are virtually untreatable by any available technology when they reach an advanced stage can be cured by essentially trivial and low-tech intervention during their early stages. It is virtually assured that no significant progress will be made until rapid, sensitive, cheap, and noninvasive methods are available for early detection.

K. John Morrow Jr., Ph.D., is president of Newport Biotech and a contributing editor for GEN. Web: Phone: (513) 237-3303. Email: [email protected].

Previous articleCV Therapeutics Receives $70M from Menarini for Rights to Angina Treatment
Next articleBiovitrum Proffers $150M for Three Approved Biologics from Amgen