May 1, 2011 (Vol. 31, No. 9)
K. John John Morrow Jr. Ph.D. President Newport Biotech
Field Advances in Quest to Improve Disease Diagnosis and Predict Drug Response
As an important discipline within systems biology, metabolomics is being explored by a number of laboratories for its potential in pharmaceutical development. Studying metabolites can offer insights into the relationships between genotype and phenotype, as well as between genotype and environment. In addition, there is plenty to work with—there are estimated to be some 2,900 detectable metabolites in the human body, of which 309 have been identified in cerobrospinal fluid, 1,122 in serum, 458 in urine, and roughly 300 in other compartments.
Guowang Xu, Ph.D., is a researcher at the Dalian Institute of Chemical Physics. He is investigating the causes of death in China and how they have been changing over the years as the country has become a more industrialized nation. A major issue is the increase in the incidence of metabolic disorders such as diabetes, which has grown to affect 9.7% of the Chinese population.
This emergent risk is still identified by relatively primitive methods, and there is currently no means of early diagnosis of insulin-resistant subjects through blood or urine samples. These considerations are foremost in Dr. Xu’s attempt to develop a metabolomic-based strategy for identification of at-risk individuals.
Dr. Xu, working in collaboration with the laboratory of Rainer Lehman, Ph.D., of the University of Tübingen, Germany, compared urinary metabolites in samples from healthy individuals with samples taken from prediabetic, insulin-resistant subjects. Using mass spectrometry coupled with electrospray ionization in the positive mode, they observed striking dissimilarities in levels of various metabolites in the two groups.
“When we performed a comprehensive two-dimensional gas chromatography, time-of-flight mass spectrometry analysis of our samples, we observed several metabolites, including 2-hydroxybutyric acid in plasma, that could be identified as potential diabetes biomarkers,” Dr. Xu explains.
In other, unrelated studies, Dr. Xu and the German researchers used a metabolomics approach to investigate the changes in plasma metabolite profiles immediately after exercise and following a 3-hour and 24-hour period of recovery. They found that medium-chain acylcarnitines were the most distinctive exercise biomarkers, and they are released as intermediates of partial beta oxidation in human myotubes and mouse muscle tissue.
Traditionally, metabolites have been used as markers of inborn errors of metabolism, and diseases such as phenylketonuria were recognized as curable through introduction of appropriate nutritional intermediates as far back as the 1930s. Today, the availability of advanced technologies will make possible treatment of diseases with a much more complex genetic and environmental basis.
“I believe we are entering into a new era of phenotype-based diagnosis based on groups of biomarkers,” Dr. Xu says. “The traditional approach of assessment based on a singular biomarker is being superseded by the introduction of multiple marker profiles.”
Evaluation of individual metabolic profiles is becoming increasingly important in researching new pharmaceutical molecules, according to Rima Kaddurah-Daouk, Ph.D., associate professor at the Duke University Center for Pharmacometabolomics.
“The field is based on scaling up our knowledge of biochemistry to encompass not only the metabolism of the host but also that of the gut bacteria and other factors that may influence drug uptake and processing,” she explains. “We are focusing our studies of personalized medicine on psychiatric disorders and cardiovascular diseases.”
Typical of the studies under way by Dr. Kaddurah-Daouk and her colleagues is a recently published investigation highlighting the role of an SNP variant in the glycine dehydrogenase gene on individual response to antidepressants. It was determined that patients who do not respond to the selective serotonin uptake inhibitors citalopram and escitalopram carried a particular single nucleotide polymorphism in the GD gene.
“These results allow us to pinpoint a possible role for glycine in selective serotonin reuptake inhibitor response and illustrate the use of pharmacometabolomics to inform pharmacogenomics. These discoveries give us the tools for prognostics and diagnostics so that we can predict what conditions will respond to treatment.
“This approach to defining health or disease in terms of metabolic states opens a whole new paradigm. By screening hundreds of thousands of molecules, we can understand the relationship between human genetic variability and the metabolome.”
Dr. Kaddurah-Daouk talks about statins as a current model of metabolomics investigations. Whereas, these drugs were originally developed to lower cholesterol levels in individuals at risk for cardiovascular disease, it is now known that they have widespread effects, altering a range of metabolites. To sort out these changes and develop recommendations for which individuals should be receiving statins will require substantial investments of energy and resources into defining the complex web of biochemical changes that these drugs initiate.
Furthermore, Dr. Kaddurah-Daouk asserts that, “genetics only encodes part of the phenotypic response. One needs to take into account the net environment contribution in order to determine how both factors guide the changes in our metabolic state that determine the phenotype.”
Clare Daykin, Ph.D., is a lecturer at the University of Nottingham, U.K. Her field of investigation encompasses “interactive metabolomics,” a tool she has developed through research funded by the Engineering and Physical Sciences Research Council. She defines the approach as “the study of the interactions between low molecular weight biochemicals and macromolecules in biological samples such as blood plasma, without preselection of the components of interest.
“Blood plasma is a heterogeneous mixture of molecules that undergo a variety of interactions including metal complexation, chemical exchange processes, micellar compartmentation, enzyme-mediated biotransformations, and small molecule–macromolecular binding.”
Many low molecular weight compounds can exist freely in solution, bound to proteins, or within organized aggregates such as lipoprotein complexes. Therefore, quantitative comparison of plasma composition from diseased individuals compared to matched controls provides an incomplete insight to plasma metabolism.
“It is not simply the concentrations of metabolites that must be investigated, but their interactions with the proteins and lipoproteins within this complex web. You have to look at the other components and how they influence one another,” Dr. Daykin explains.
Rather than targeting specific metabolites of interest, Dr. Daykin’s metabolite–protein binding studies aim to study the interactions of all detectable metabolites within the macromolecular sample. Such activities can be studied through the use of diffusion-edited nuclear magnetic resonance (NMR) spectroscopy, in which one can assess the effects of the biological matrix on the metabolites. “This can lead to a more relevant and exact interpretation for systems where metabolite–macromolecule interactions occur.”
Diffusion-edited NMR experiments provide a way to separate the different compounds in a mixture based on the differing translational diffusion coefficients (which reflect the size and shape of the molecule). The measurements are carried out by observing the attenuation of the NMR signals during a pulsed field gradient experiment.
Pushing the Limits
It is widely recognized that many drug candidates fail during development due to ancillary toxicity. Uwe Sauer, Ph.D., professor, and Nicola Zamboni, Ph.D., researcher, both at the Eidgenössische Technische Hochschule, Zürich (ETH Zürich), are applying high-throughput intracellular metabolomics to understand the basis of these unfortunate events and head them off early in the course of drug discovery.
“Since metabolism is at the core of drug toxicity, we developed a platform for measurement of 50–100 targeted metabolites by a high-throughput system consisting of flow injection coupled to tandem mass spectrometry.”
Using this approach, Dr. Sauer’s team focused on the central metabolism of the yeast Saccharomyces cerevisiae, reasoning that this core network would be most susceptible to potential drug toxicity. Screening approximately 41 drugs that were administered at seven concentrations over three orders of magnitude, they observed changes in metabolome patterns at much lower drug concentrations without attendant physiological toxicity.
The group carried out statistical modeling of about 60 metabolite profiles for each drug they evaluated. This data allowed the construction of a “profile effect map” in which the influence of each drug on metabolite levels can be followed, including off-target effects, which provide an indirect measure of the possible side effects of the various drugs.
“We have found that this approach is at least 100 times as fast as other omics screening platforms,” Dr. Sauer says. “Some drugs, including many anticancer agents, disrupt metabolism long before affecting growth.”
Furthermore, they used the principle of 13C-based flux analysis, in which metabolites labeled with 13C are used to follow the utilization of metabolic pathways in the cell. These 13C-determined intracellular responses of metabolic fluxes to drug treatment demonstrate the functional performance of the network to be rather robust, leading Dr. Sauer to the conclusion that the phenotypic vigor he observes to drug challenges is achieved by a flexible make up of the metabolome.
Dr. Sauer is confident that it will be possible to expand the scope of these investigations to hundreds of thousands of samples per study. This will allow answers to the questions of how cells establish a stable functioning network in the face of inevitable concentration fluctuations.
Is Now the Hour?
There is great enthusiasm and agitation within the biotech community for metabolomics approaches as a means of reversing the dismal record of drug discovery that has accumulated in the last decade. While the concept clearly makes sense and is being widely applied today, there are many reasons why drugs fail in development, and metabolomics will not be a panacea for resolving all of these questions. It is too early at this point to recognize a trend or a track record, and it will take some time to see how this approach can aid in drug discovery and shorten the timeline for the introduction of new pharmaceutical agents.
K. John Morrow Jr., Ph.D. (firstname.lastname@example.org), is president of Newport Biotech and a contributing editor for GEN.