Embedded Files
diseases as cancer, arrhythmia, or asthma is that these phenotypes represent an overlapping spectrum of mechanistically distinct diseases at the molecular or genetic level (51–53). In some cases, understanding the basis for a monogenic disease can inform and rationalize drug therapy; thus, mutations in KCNJ11, a component of a sulfonylurea-sensitive potassium channel, cause neonatal diabetes that is especially sensitive to sulfonylureas (54, 55). Disease-associated genetic variants (see Glossary) may be inherited (present in the germline) or arise later; such somatic cell mutations are often found in cancer, where they may arise as a consequence or a cause of disease and may contribute to variability in drug response. The epidermal growth factor receptor inhibitors gefitinib and erlotinib display highly variable responses in lung cancer, and dramatic tumor shrinkage now seems to be characteristic of a small group of patients whose tumors have somatic mutations in the tyrosine kinase domain of the receptor (56, 57). A screening test to identify such responders has now been commercialized (58). Another example of the relationship between disease and drug response genetics is shown by a large study that implicated specific alleles in the 5-lipoxygenase gene as risk factors for myocardial infarction and stroke (59). On the basis of this finding, drugs to inhibit the 5-lipoxygenase–activating protein have been developed, and the first trial studied only patients selected for the at-risk genetic variant (60). This study paradigm shows how screening very large gene sets, up to the whole genome, can inform new biology and drug development, an advantage that the candidate gene approach generally lacks. Genome-Wide Approaches Although the approach of analyzing candidate genes has considerable intuitive appeal, it suffers from the criticism that it fails to consider a potential contribution of other genes, including those whose function is not yet well understood. Increasingly robust technologies are being developed to interrogate sets of single-nucleotide polymorphisms (SNPs) across the entire genome, and such whole genome association approaches have been applied to identify new disease susceptibility genes and pathways (61–63). This “unbiased” whole genome approach may also be applicable to the problem of identifying new loci-modulating variability in drug response. Comparing the frequencies of hundreds of thousands of polymorphisms in patients with or without a specified genotype will inevitably generate a very large number of false-positive associations. Thus, one way to interpret genome-wide association studies is that they may be able to identify genomic loci, which can then be studied in other patient subsets (that is, they are hypothesis-generating tools). Despite the very large number of polymorphisms in the human genome, it has also become apparent that these are not independent. Knowledge of the genotype at 1 locus informs genotype at other loci that are in linkage disequilibrium (64). The definition of such haplotype (see Glossary) structure across the genome and of “tag SNPs” that can thereby provide genotype information at multiple loci is a major goal of ongoing studies within industry and at the international HapMap consortium (65, 66). Such information will in turn be an enabling resource for large-scale genome-wide association studies by reducing the number of SNP sites for study. The Concept of Drug Response “Pathways” An appealing intermediate step between the single candidate gene approach and the nascent genome-wide approach is to consider that variable disease and drug response phenotypes reflect perturbations in function of increasingly well-characterized physiologic systems comprising dozens or hundreds of interacting gene products. Thus, studies of response to corticosteroids or β2-agonists in patients with asthma consider variations not only in corticosteroid or β2-adrenergic receptors but also in the downstream and interacting signaling pathways they modulate. In 1 such study, genotyping at 131 SNP sites in 14 genes in the steroid pathway identified and subsequently reproduced an association between response to
Google Sites
Report abuse