Computer-aided Design of Anti-cancer Drugs Targeting Protein Kinases

描述（申请人提供）：蛋白激酶已成为抗癌药物设计的重要靶点。几种蛋白激酶药物，如赫赛汀、格列卫和易瑞沙，已经被批准用于治疗几种癌症。这些药物作用于三种不同的蛋白激酶。由于人类有500多种蛋白激酶，它们的各种突变形式与治疗不同的疾病有关，因此应该有更多适合药物开发的蛋白激酶相关靶点。这项研究的长期目标是开发一种有效的计算模型，以帮助加速针对这些蛋白质的治疗药物的发现和开发。该应用程序的直接目标是进一步开发一种独特的五层分层计算筛选模型，平衡理论严谨性和不同层的速度，以缩短从大型化学文库中选择最有前途的候选药物所需的时间和费用，并进一步优化它们的效力、选择性和药物样性质。Specific Aim 1开发了一种复杂的连续介质溶剂模型，用于预测cLogP，并将其与Lipinski的五项规则一起用于评估化合物的药物性质。Specific Aim 2通过在分子对接中加入蛋白质的灵活性来改进虚拟筛选，并开发了一种增强的副本交换方法来改进对接结构，以获得更可靠的结合亲和力计算。Specific Aim 3扩展了我们以前的信息学/结构建模方法，以帮助指导具有较少副作用的选择性蛋白激酶抑制剂的开发。它包括建立整个人类蛋白激酶家族的结构模型，并使用增强的构象采样技术（在Specific Aim 2中开发）来完善这些模型，以便进行分子对接和结合亲和力计算。我们还将使用这一整套结构模型进行全面的比较分析，以更好地了解靶向atp结合口袋的蛋白激酶抑制剂如何实现选择性。Specific Aim 4通过对蛋白质数据库中所有蛋白激酶抑制剂复合物进行能量成分分析来评估流行药效团模型的有效性，以证明其在药物设计中的使用是合理的。Specific Aim 5应用这种改进的五层分层模型来识别新的靶向EGFR、HER2、CDK2和BCR-ABL蛋白激酶的抗癌候选药物。

DESCRIPTION (provided by applicant): Protein kinases have become important targets for the design of anti-cancer drugs. Several protein kinase drugs such as Herceptin, Gleevec, and IRESSA have already been approved for treating several forms of cancer. These drugs act on three different protein kinases. Because there are more than 500 protein kinases in human and their various mutant forms are relevant for treating different diseases, there should be more protein kinase-related targets suitable for drug development. The long-term goal of this research is to develop an efficient computational model to help speed up the discovery and development of therapeutic drugs targeting these proteins. The immediate goals of this application aim at further developing a unique five-tier hierarchical computational screening model that balances theoretical rigor and speed at different tiers to shorten the time and expense needed to select the most promising drug candidates from large chemical libraries and to further optimize them for potency, selectivity and drug-like properties. Specific Aim 1 develops a sophisticated continuum solvent model for predicting cLogP and uses it with Lipinski's Rules of Five to evaluate compounds for drug-like properties. Specific Aim 2 improves virtual screening by incorporating protein flexibility in molecular docking and develops an enhanced replica exchange method for refining docking structure for more reliable binding affinity calculation. Specific Aim 3 extends our previous informatics/structural modeling approach to help guide the development of selective protein kinase inhibitors that have fewer side effects. It involves building structural models of the whole family of human protein kinases and on using enhanced conformational sampling techniques (developed in Specific Aim 2) to refine these models well enough for molecular docking and binding affinity calculation. We will also use this whole set of structural models to perform comprehensive comparative analysis to understand better how selectivity can be achieved by protein kinase inhibitors that target the ATP-binding pocket. Specific Aim 4 evaluates the validity of a popular pharmacophore model by performing energy component analysis on all the protein kinase-inhibitor complexes in the Protein Data Bank in order to justify its use in drug design. Specific Aim 5 applies this refined five-tier hierarchical model to identify new anti-cancer drug candidates for the protein kinase targets EGFR, HER2, CDK2, and BCR-ABL.