How Genetic Variation in Protein Kinases Affects Drug Response

描述（由申请人提供）：人类因单核苷酸多态性（snp）而彼此不同，共有1000万个snp（等位基因频率bb0.1 %）。只有一小部分snp被认为具有功能或表型后果。与疾病相关的snp通常是功能中性的，仅由于连锁不平衡而与特定表型相关。相比之下，对疾病机制的了解主要是通过对功能变异的了解来提供的。系统地将功能性遗传变异与疾病和药物反应联系起来是基因、环境和健康倡议的一个主要目标。已知的致病突变绝大多数是编码的，选择的测量显示非同义snp （nssnp）的显著部分是进化有害的。然而，大多数nssnp对分子功能和表型的影响可能有限。利用序列和进化信息的结合，已经开发了几种方法来区分功能性和非功能性nssnp。虽然这些预测可以作为指导，但它们目前还不能代替直接实验。在这里，我们研究了人类编码变异对蛋白激酶功能的影响。蛋白激酶已被充分研究，并与许多疾病相关，通常是因果关系。不恰当激活激酶的突变会导致癌症等疾病的发生。在前期工作中，我们已经了解到外源性蛋白激酶的表达可以在培养的人细胞中产生显著的药物反应。我们的第一个目标是生成约3000个开放阅读框的资源，捕获535种人类蛋白激酶的变体。所有克隆将被克隆成慢病毒表达载体，以允许转染到广泛的人类细胞和组织。该资源将包含野生类型和变体的深度采样，重点是常见和预测有害的变体。每个克隆体的序列都经过验证，并有唯一的条形码。我们的第二个主要目标将是确定激酶及其变体的表达如何改变药物反应。在组织培养中，激酶的表达可以定量地、可重复地增强或阻断药物的作用。例如，我们已经证明，对EGFR药物易瑞沙的反应被许多激酶的过表达阻断，包括那些在EGFR途径下游起作用的激酶。我们将使用功能测试来确定激酶变体如何改变对三种常见癌症化疗药物的反应。一旦建立，所提出的框架可以经济地适用于培养的人类细胞中具有定量表型的任何药物的作用，并且一般适用于其他类别的基因和编码变体。遗传变异是人类健康的一个主要因素，但仍然知之甚少，它改变了疾病的易感性、耐药性和对治疗的反应。在这个应用程序中，我们测试了一类重要基因的序列变异，蛋白质激酶，以及如何影响对重要抗癌药物的反应。我们试点了一个经济的大规模框架，将个体遗传差异与治疗药物反应联系起来，同时提供了一个实验途径来研究任何在培养的人类细胞中诱导可量化反应的药物。我们将构建一个3000个克隆的表达文库，代表一个条形码慢病毒主干中535种蛋白激酶的变体，这将对研究界具有很高的价值，我们将生成一套以前无法获得的关于人类细胞代谢的激酶依赖性调节的大量数据。

DESCRIPTION (provided by applicant): Humans vary from one another by single nucleotide polymorphisms (SNPs), with 10 million SNPs being common (allele frequency > 1%). Only a small fraction of SNPs are thought to have functional or phenotypic consequences. SNPs associated with diseases are often functionally neutral, correlated with a particular phenotype only due to linkage disequilibrium. By contrast, insight into disease mechanism is primarily provided by knowledge of the functional variant. Systematically relating functional genetic variation to disease and drug response is a major goal of the Genes, Environment and Health Initiative. Known disease-causing mutations are overwhelmingly coding and measures of selection show a significant fraction of non-synonymous SNPs (nsSNPs) to be evolutionarily deleterious. Nevertheless, the majority of nsSNPs probably have limited effect on molecular function and phenotype. Several methods have been developed to discriminate functional from non- functional nsSNPs, exploiting a combination of sequence and evolutionary information. Although, these predictions can serve as a guide, they are not currently a substitute for direct experiment. Here we investigate the effects of human coding variants on protein kinase function. Protein kinases are well studied and have been associated, often causally, with many diseases. Mutations that inappropriately activate kinases can drive the development of diseases such as cancer. In preliminary work we have learned that exogenous expression of protein kinases can drug response in cultured human cells dramatically. Our first goal will be to generate a resource of ~3000 open reading frames capturing variants of 535 human protein kinases. All clones will be cloned into lentiviral expression vectors to allow transfection into a wide range of human cells and tissues. This resource will contain wild type and a deep sampling of variants, with emphasis on common and predicted-deleterious variants. Each clone will sequence verified and uniquely bar-coded. Our second major goal will be to determine how expression of kinases and their variants modify drug response. Kinase expression can quantitatively and reproducibly enhance or block the action of drugs in tissue culture. For example, we have shown that response to the EGFR drug Iressa is blocked by overexpression of a number of kinases including those that act downstream in EGFR pathway. We will use functional tests to determine how kinase variants change response to three common cancer chemotherapy drugs. Once established, the proposed framework can be economically adapted to the action of any drug with a quantitative phenotype in cultured human cells and general to other classes of genes and coding variants. Genetic variation is a major but still poorly understood contributor to human health, altering disease susceptibility, resistance, and response to treatment. In this application we test how sequence variation in one important class of genes, the protein kinases, and affects response to important anti-cancer drugs. We pilot an economical large-scale framework for relating individual genetic differences to therapeutic drug response, while offering an experimental avenue to study any drug that induces a quantifiable response in cultured human cells. We will construct a 3000-clone expression library representing variants of 535 protein kinases in a bar-coded lentivirus backbone that will be of high value to the research community, and we will generate a large set of previously unobtainable data on kinase-dependent regulation of human cellular metabolism.