Embedded Files
Biological Context in Which Drugs Act Variability in drug actions may reflect genetically determined changes not only in metabolism or in the drug target but also in the complex biological milieu in which drugs interact with their target molecules. An example that has received considerable attention is drug-induced prolongation of the QT interval and susceptibility to the ventricular tachycardia torsades de pointes, a common cause for drug withdrawal or relabeling (39). Although the length of the QT interval is determined by many ion-channel and other genes, only 1 (HERG [KCNH2]) is the target for QT-prolonging drugs; as a consequence, new drug candidates are now routinely screened for KCNH2-blocking activity (40). Of importance, surveys of patients with drug-induced torsades de pointes occasionally identify predisposing mutations not only in the drug target (the KCNH2 channel) but also in other genes that control the QT interval (41, 42); that is, variability in the physiologic control of the QT interval can contribute to variable effects of KCNH2 blockers. A key first step in any pharmacogenetic study, and indeed in all contemporary translational medicine, is precisely defining responders and nonresponders to therapy. Failure in this effort ensures failure for any subsequent genetic analysis. There is as much tension between well-defined clinical (often termed “hard”) end points, such as death, and biomarkers and surrogate end points, such as altered tumor size, change in blood pressure, or altered serum concentration of a disease-related protein, in pharmacogenetics as there is in all translational research. To the extent that biomarkers and surrogate end points can serve as relevant “intermediate phenotypes” that define important subsets of disease or drug responses, their definition is a critical component of minimizing confounding variables in any genetic-based analysis. A related challenge is the informatics problem of developing uniform methods to represent phenotypic data, such as drug responses, as is now the case with relatively standardized methods for representing genetic sequence and variation data. Evolving Approaches to Identify Genes That Modulate Drug Responses One approach in contemporary pharmacogenetics is to first use an understanding of pharmacokinetics and pharmacodynamics (see Glossary) to identify candidate genes that may determine the action of a drug. Then, the candidate genes are screened for variants, and the variants are screened for biological activity. This “genotype to phenotype” approach can lead not only to associations (see Glossary) between functional variants and drug responses but also to new insights into the basic biology of these proteins (43, 44). The converse strategy, “phenotype to genotype,” first defines variable drug response phenotypes and then searches in single genes, in pathways, or across the genome for sources of that variability. Initial studies in pharmacogenetics often focused on small numbers of drug response outliers, studying candidate genes identified on the basis of an understanding of the underlying physiology and culminating in identification of variants associated with the outlier phenotype. This approach is intuitively appealing, especially if variants display abnormal function in vitro or in animal models that could plausibly explain the outlier drug response phenotype. CYP2D6, UGT1A1, and KCNH2 variants that mediate unusual responses to drug therapy are examples of this approach. Outlier drug responses that may lend themselves to such a strategy include the problems of rare and clinically severe adverse drug reactions, such as torsades de pointes, the lupus reaction, or hepatotoxicity (39, 45). Because these reactions often occur in an apparently unpredictable fashion, they have been termed “idiosyncratic,” and a genetic predisposition is often invoked. Accrual of patient sets in which to test this idea will require multi-center collaborations (46), and the initial focus of genetic analysis has been on a small number of genes chosen because they may be associated with disease susceptibility Disease Phenotypes An obvious potential mechanism underlying variability in therapy for such common
Google Sites
Report abuse