Extracellular Proteolysis as a Molecular Stratification Tool for Cancer

描述（由申请人提供）：功能基因组策略已被广泛应用于定义癌症的独特分子亚型，以预测表型特性，如转移潜力和对治疗化合物的敏感性。然而，基因和蛋白质表达水平的变化可能是间接的，因此在疾病的发展中不起作用。侵袭性癌症的特征之一是其逃离细胞环境并扩散到新组织的能力，这一过程部分由细胞外蛋白酶的活性介导。蛋白酶活性受到亚细胞定位、内源性蛋白酶抑制剂的存在以及从无活性前体形式的必要转化的严格调节。因此，在这些情况下，仅仅知道蛋白酶表达水平是不够的。我们建议，全球概况的细胞外蛋白酶活性可能会成为一个强大的功能工具，癌症的分子分层。Craik实验室开发了一种新的基于质谱的筛选技术，该技术可以通过采用合理设计的肽底物的小型多样性库来识别蛋白酶单独和复杂生物混合物中的全局底物特异性和动力学效率。该技术被称为多重底物质谱分析（MSP-MS），通过允许无偏和同时检测给定样品中的所有蛋白酶活性，标志着蛋白酶分析的重大突破。在该提案中，Craik实验室将与Sali实验室合作开发和测试基于蛋白酶特异性对癌症样本进行分类的计算模型，目标是为亚型特异性成像建立蛋白酶激活诊断。将使用MSP-MS测定来确定来自日益复杂的乳腺癌和前列腺癌样品的细胞外蛋白酶活性的总体概况。同时，机器学习算法将用于开发基于特异性的分类模型，这些模型将与肿瘤侵袭性的已知指标相关。代表所鉴定的主要分类组的肽序列的子文库将有助于设计蛋白酶可激活的成像探针，该探针将通过实验测试亚型选择性。探针切割序列将被迭代地改进，以通过将切割速率并入建模策略和针对靶蛋白酶的3D结构的肽对接来提高选择性。在项目的下一阶段，开发的新型试剂将应用于临床相关性并进一步优化。我们预计，我们为癌症的功能分析生成定制诊断的策略将促进疾病的识别和监测， 有助于发现癌症生物标志物。

DESCRIPTION (provided by applicant): Functional genomic strategies have been widely implemented to define unique molecular subtypes of cancer in order to predict phenotypic properties such as metastatic potential and sensitivity to therapeutic compounds. However, changes in the level of gene and protein expression can be circumstantial and therefore play no functional role in the development of the disease. One of the hallmarks of aggressive cancer is its ability to escape the cellular milieu and spread to new tissues, a process that is mediated in part by the activity of extracellular proteases. Protease activity is tightly regulated by subcelluar localization, the presence of endogenous protease inhibitors, and requisite conversion from inactive precursor forms. Therefore, in these circumstances, it is not enough to know protease expression levels alone. We propose that global profiles of extracellular protease activity may emerge as a powerful functional tool for the molecular stratification of cancer. The Craik laboratory has developed a novel mass spectrometry-based screening technology that can identify the global substrate specificity and kinetic efficiency of proteases alone and in complex biological mixtures by employing a small, diverse library of rationally designed peptide substrates. This technology, referred to as Multiplex Substrate Profiling by Mass Spectrometry (MSP-MS), marks a significant breakthrough in protease profiling by allowing for the unbiased and simultaneous detection of all protease activities in a given sample. In this proposal, the Craik laboratory will partner with the Sali laboratory to develop and test computational models that classify cancer samples on the basis of protease specificity with the goal of building protease-activatable diagnostics for subtype-specific imaging. Global profiles of extracellular protease activity from increasingly complex breast and prostate cancer samples will be determined using the MSP-MS assay. In parallel, machine learning algorithms will be used to develop specificity-based classification models that will be correlated to known metrics for tumor aggressiveness. Sub-libraries of peptide sequences that represent the major classification groups identified will aid in designing protease-activatable imaging probes that will be tested experimentally for subtype selectivity. Probe cleavage sequences will be iteratively refined to improve selectivity through both incorporation of cleavage rates into the modeling strategy and peptide docking against the 3D structures of the target proteases. The new class of reagents developed will be applied to and further optimized against clinical correlations in the next phase of the project. We anticipate that our strategy for generating tailored diagnostics for the functional profiling of cancer will advance the identification and monitoring of disease as well as aid in cancer biomarker discovery.