NEWS AND VIEWS
Personalized medicine progresses
© 2006 Nature Publishing Group http://www.nature.com/naturemedicine
John N Haselden & Andrew W Nicholls A new approach to personalized drug treatment emerges in a study examining the metabolic profile of rats. The profile, which is a measurement of small molecules such as sugars and amino acids, is used to predict the response to drugs that are toxic to the liver. This study proposes the extension of this concept into humans as a way of predicting the outcome of a therapy for a given profile.
Conventional medical practice, in most cases, is based on a reactive assessment of a patient’s medical state, determined through a combination of clinical signs (where does it hurt?) and biochemical and functional measurements (such as blood tests). Based on such assessments, the physician makes a diagnosis and prescribes an appropriate course of treatment. This might be seen as a broad-spectrum treatment approach based on population responses, in that such a treatment works for most people most of the time. However, it may not reflect the best treatment for the individual, which requires a definition of the broader scope of the patient’s medical state. This is determined not only by an individual’s predetermined genetic fingerprint, but also by his interaction with the environment, or lifestyle. Understanding the potential for an individual to exhibit specific medical events based on analysis of his genetic and environmental signature would enable a patient to be prescribed the most effective medication. Such an ideal of personalized medicine has started to receive significant attention, in particular from the US Food and Drug Administration1. In a recent issue of Nature, Clayton et al.2 edge us closer to this ideal. They propose a concept of metabolically phenotyping individuals. Their approach is the latest entry in the field of metabonomics (or metabolomics), which uses nuclear magnetic resonance spectroscopy–or mass spectrometry–based technologies to generate a distinctive signature of small molecules, such as sugars, amino acids and lipids. In studies on rats, the preexisting metabolic profile of each individual animal was able to predict the extent of liver damage induced by acetaminophen. The authors were also able to tease out which aspects of the metabonome were John N. Haselden is in Metabolic Profiling and Cellular Pathology/Toxicology Groups in Safety Assessment, and Andrew W. Nicholls is in the Metabolic Profiling Group, Safety Assessment, GlaxoSmithKline Research & Development, Hertfordshire, SG12 0DP, UK. E-mail:
[email protected]
510
associated with more severe damage. Higher predose levels of taurine (a sulfur-containing amino acid) in rats seemed to protect against liver damage, which is consistent with the role taurine has in detoxification reactions. Although metabonomics has been around for some time, this study is the first to highlight—and show in animals, albeit in a simplistic manner—the new concept of predicting a biological outcome based on the individual phenotype as defined by the metabolic profile. The extrapolation of this concept from laboratory rats to humans, however, is not straightforward and, as the authors allude to, has some way yet to go. The diversity of the human population from a genetic standpoint is vastly different from inbred experimental animals, as are the dietary, lifestyle and environmental factors—all of which massively impact our understanding of the metabolic profile or signature. To attain some understanding of what ‘normal’ humans look like, we need very large epidemiological-size background studies covering all ethnicities, and not least capturing all data pertaining to diet, exercise, previous medical history and the like. All is not lost, however—the key strengths of metabonomic analyses are that they are relatively simple and cost-effective, which means these clinical studies are a practical reality. Although there is limited evidence in the literature of such studies going on3,4, we are aware of one such investigation currently underway—the London Life Sciences Population (LOLIPOP) study, where an array of analyses is being carried out, and from which preliminary data are now emerging5–7. This is a prospective study of potentially up to 24,000 subjects (UK-based Indians and whites) primarily investigating cardiovascular risk factors of certain subpopulations. The study will assess these factors in relation to the subjects’ genotypes and metabolic phenotypes (the pharmacometabonome, if you like). Pharmacometabonomics has huge potential in personalized medicine as an adjunct to the accepted use of pharmacogenomics, which relies on gene expression profiles. Combining the two approaches can define the impact
Figure 1 The genome remains constant—the pharmacometabonome is a moveable feast, as indicated by the many factors that impact it.
of environment and lifestyle on the genetic blueprint. Although pharmacogenomic tests are coming into clinical practice to predict outcomes for individuals with breast cancer, predictions associated with other disease states could well be augmented by reference to the pharmacometabonome. In addition, a logical extension of the concepts tested in rats would be the use of pharmacometabonomics to preselect volunteers at key stages of the clinical drug development process. This would enable stratification of subjects into cohorts, which could minimize the risk of adverse events, or focus on those individuals with a characteristic disease phenotype for assessment of efficacy. Although the concept of personalized medicine is almost universally approached from the aspect of patient benefit, other issues arise that require recourse to extensive ethical debate. Understanding this provides the potential for classification into risk groups which, in some cases, is a threat to the provision of health care. With an aging population, the economics of public health care is in apparent crisis, so the use of ‘molecular triage,’ based on the cost of treatment for an individual with an unfavorable disease phenotype, could lead to denial
VOLUME 12 | NUMBER 5 | MAY 2006 NATURE MEDICINE
NEWS AND VIEWS
© 2006 Nature Publishing Group http://www.nature.com/naturemedicine
of treatment. In some respects, this already occurs, whereby certain lifestyle factors such as smoking change the priority of an individual’s access to certain treatments. This study has demonstrated that in laboratory animals, scientists are able to predict a biological outcome based on an individual’s preexist-
ing metabolic profile. This concept highlights the potential that pharmcometabonomics, when combined with pharmacogenomics, could lead to major changes in how medicine is practiced. 1. FDA Consum. 36 (2005). 2. Clayton, T.A. et al. Nature, 440, 1073–1077 (2006). 3. Lenz, E.M. et al. J. Pharm. Biomed. Anal. 36, 841–
849 (2004). 4. Bijlsma, S. et al. Anal. Chem. 78, 567–574 (2006). 5. Chambers, J., Lim, E., Jain, P., Singh, D., Elliott, P. & Kooner, J. Heart 92 Suppl 2, A4–A121 (2006). 6. Chambers, J., Lim, E., Singh, D., Jain, P., Elliott, P. & Kooner, J. Heart 92 Suppl 2, A4–A121 (2006). 7. Lim, E., Lightstone, L., Chambers, J., Roderick, P., Mullee, M. & Kooner, J. Heart 92 Suppl 2, A4–A121 (2006).
Overcoming insulin resistance with CNTF Rexford S Ahima Ciliary neurotrophic factor (CNTF) enhances fatty-acid oxidation in muscle and reduces insulin resistance in obese, diabetic mice. Because skeletal muscle is the major site of insulin-mediated glucose uptake, this action of CNTF could benefit individuals with diabetes (pages 541–548). The ability to maintain normal blood glucose levels involves a complex interplay between insulin secretion by pancreatic beta cells and responsiveness of organs, in particular liver, adipose tissue and skeletal muscle1. Type 2 diabetes is commonly associated with obesity and results from insufficient insulin production and reduced sensitivity to insulin1. Drugs used to treat type 2 diabetes include sulfonylureas, which stimulate insulin secretion, and metformin and thiazolidinedines, which enhance insulin action in liver and muscle. When patients no longer respond to these drugs, insulin is administered to control blood glucose and prevent complications of diabetes, but large amounts of insulin are needed because patients are insensitive to its action1. Moreover, insulin increases weight; therefore, the search is on for drugs capable of enhancing insulin’s effect without increasing weight in obese individuals who have type 2 diabetes. When researchers discovered in the mid1990s that the hormone leptin could suppress appetite, they hoped that they would be able to harness the hormone to counteract obesity2,3. But it soon became clear that most obese individuals are resistant to the effects of leptin2,3. In two studies, one in this issue and another in Endocrinology, Watt et al.4,5 examine one way to circumvent the problem of leptin resistance—administration of ciliary neurotrophic factor (CNTF). CNTF was originally named for its role in neuronal survival, but since has been shown The author is in the Department of Medicine, Division of Endocrinology, Diabetes and Metabolism, University of Pennsylvania School of Medicine, Philadelphia, Pennsylvania 19104, USA. E-mail:
[email protected]
to have a role in reducing weight by suppressing appetite and increasing metabolic rate6–8. The studies by Watt et al. demonstrate unique properties of CNTF in overcoming leptin resistance and enhancing insulin action in obese animals through direct action on muscle. Insulin resistance can arise from infiltration of fat by inflammatory cells, mitochondrial dysfunction that impairs fat breakdown, and accumulation of fat in liver and muscle cells resulting from increased uptake of fatty acids and formation of triacylglycerol1,2. This process, known as steatosis, leads to production of ceramide and various harmful substances that disrupt insulin action2. Fat tissue secretes hormones, such leptin and adiponectin, that control glucose and lipid metabolism and prevent steatosis2,3. In normal individuals, leptin acts on the hypothalamus to suppress feeding, increase metabolic rate and decrease weight3. Leptin prevents steatosis by inhibiting enzymes involved in lipid synthesis such as stearoylCoA desaturase 1 (SCD1)2,3. Leptin also controls AMP-activated protein kinase (AMPK), an enzyme that integrates nutritional and hormonal signals in the brain and peripheral organs9. A rise in leptin inhibits AMPK in the hypothalamus, while stimulating AMPK in muscle directly, and indirectly through sympathetic activity9. Activation of AMPK in muscle and liver limits deposition of fat by inhibiting acetyl CoA carboxylase (ACC) and stimulating malonyl CoA decarboxylase (MCD), which leads to a reduction in malonyl CoA, a precursor needed for lipid synthesis. Activation of AMPK also increases the activity of carnitine palmityl acyltransferase 1 (CPT-1), which transfers fatty acids into mitochondria to be oxidized.
NATURE MEDICINE VOLUME 12 | NUMBER 5 | MAY 2006
These actions of AMPK prevent steatosis and maintain insulin response in liver and muscle2,4 (Fig. 1). Besides the adipocyte hormones leptin and adiponectin, insulin-sensitizing drugs (that is, metformin and thiazolidinediones) activate AMPK in liver and muscle9. Although uptake by skeletal muscle accounts for the bulk of insulin-mediated glucose uptake and this tissue could potentially be targeted for treatment of diabetes, its value as a drug target is compromised in obesity because of fat accumulation in muscle cells2. Furthermore, ‘leptin resistance’ occurs in obesity—partly as a result of elevated levels of suppressor of cytokine signaling 3 (SOCS3), a molecule that inhibits leptin’s ability to activate signal transducer and activator of transcription 3 (STAT3)3. Since the discovery of leptin, it has become clear that the signaling pathways for CNTF and leptin have similarities. For instance, the CNTF receptor complex shares structural and functional similarities with the leptin receptor10. Binding of CNTF and leptin to their respective receptors induces phosphorylation and activation of STAT3 (ref. 10). The leptin receptor and the CNTF receptor are also both localized on hypothalamic neurons involved feeding and weight control. Leptin and CNTF inhibit feeding by suppressing regulatory peptides in the hypothalamus (neuropeptide Y (NPY) and agouti-related peptide (AGRP))3,6,7. The action of CNTF persists for several days after cessation of treatment6,7. Recently, this prolonged action of CNTF has been associated with appearance of new neurons in the hypothalamus8. The cells induced by CNTF show normal activation of STAT3 by leptin, indicating that CNTF and leptin have overlapping neuronal targets in the brain that mediate energy balance8.
511
