Tracking Cancer Stem Cell Evolution

描述（由申请人提供）：目前乳腺癌的干细胞假说认为，突变发生在正常干细胞谱系中，乳腺的各种分化细胞类型来源于该谱系。这些癌症干细胞不对称地分裂，导致干细胞群体的扩增或分化为子细胞。根据这一假设，当突变和表观遗传修饰的组合破坏正常的控制机制时，乳腺癌就会发生在干细胞谱系中。这种机制暗示了肿瘤内的异质性，因为突变相对于干细胞群体的生长零星发生。我们建议跟踪单个肿瘤中肿瘤内遗传异质性的演变，并利用这些数据通过致癌树来模拟肿瘤的发展。这个致癌树上的大多数分支将对应于不进展为侵袭性癌症的异常发育途径。然而，至少有一个分支终止于侵袭性疾病，即癌症干细胞谱系。连接该末端节点与树根（即正常干细胞）的路径追踪进展途径。进一步分析与该途径交叉的分支中的基因组序列将揭示对应于肿瘤进展的突变的时间序列。我们将开发生成致癌树的有效方法，并测试几种肿瘤中几个突变的时间序列。这将通过广泛的显微切割和测量肿瘤广泛取样区域的杂合性缺失（洛）来完成。高通量显微切割已经解决，但没有足够高通量的方法用于分析洛缺失。这个问题将通过将微切割片段的基因组转换为低复杂性表示（LCR），在载玻片上点出这些LCR，并用针对单核苷酸多态性的探针进行探测，以同时评估数千个微切割样品中的等位基因不平衡来解决。然后，通过使用洛缺失数据来建立肿瘤遗传树，可以将肿瘤的每个显微切割片段与每个其他片段相关联。这棵树报告了肿瘤的历史，早期的洛缺失分支靠近树的基部，而后期的洛分支更靠近树的“叶子”。使用这棵树，我们将能够确定肿瘤发展过程中突变、差异表达和表观遗传变化发生的时间顺序。我们将研究在乳腺癌中经常观察到的几种突变的时间。从长远来看，专注于肿瘤的小区域可以避免遗漏小但重要的突变，而由于肿瘤内的异质性，高通量发现平台（如微阵列）无法在背景之上检测到这些突变。这种方法可能会导致诊断和预后标志物或治疗目标的不同阶段的进展。我们提出了几个高通量的突变检测方案，这将使我们能够建立必要的效率，使致癌树的方法实用的致癌树。公共卫生相关性：只有一小部分乳腺癌肿瘤能够引起转移性疾病，这些“癌症干细胞”很难识别和表征。通过追踪突变在肿瘤中发展和传播的历史，我们将能够确定诊断和预后标志物，以及渐进性更具侵袭性的疾病的潜在治疗靶点。

DESCRIPTION (provided by applicant): A current stem cell hypothesis for breast cancer holds that mutations occur in the normal stem cell lineage from which the various differentiated cell types of the breast derive. These cancer stem cells divide asymmetrically, leading either to expansion of the stem cell population, or to differentiated daughter cells. According to this hypothesis, breast cancer arises when a combination of mutations and epigenetic modifications that subverts normal control mechanisms accumulate in one of the stem cell lineages. Intratumoral heterogeneity is implied by this mechanism because the mutations occur sporadically with respect to the growth of the stem cell population. We propose to track the evolution of intratumoral genetic heterogeneity in single tumors, and use this data to model tumor development by means of oncogenetic trees. Most of the branches on this oncogenetic tree will correspond to aberrant developmental pathways that do not progress to aggressive cancer. At least one branch, however, terminates in aggressive disease, that is, the cancer stem cell lineage. The path connecting this terminal node with the root of the tree (i.e. normal stem cells) traces the progression pathway. Further analysis of genomic sequences in branches that intersect with this pathway will reveal the temporal sequence of mutations corresponding to tumor progression. We will develop efficient ways of generating oncogenetic trees, and test the temporal sequence of a few mutations in several tumors. This will be done by using extensive microdissection and measurement of loss of heterozygosity (LOH) in widely sampled regions of the tumor. High- throughput microdissection has been solved, but there is no sufficiently high-throughput method for analysis of LOH. This problem will be solved by converting the genomes of microdissected pieces into low complexity representations (LCRs), spotting these LCRs on glass slides, and probing with probes directed toward Single Nucleotide Polymorphisms, to assess allelic imbalance in thousands of microdissected samples, simultaneously. Every microdissected piece of the tumor can then be related to every other piece by using the LOH data to build an oncogenetic tree. This tree reports on the history of the tumor, in that early LOH branch near the base of the tree, while later LOH branch closer to the "leaves" of the tree. Using this tree, we will be able to determine the temporal order in which mutations, differential expression, and epigenetic changes occur during tumor development. We will examine the timing of several mutations that are frequently observed in breast cancer. In the longer term, focusing on small regions of a tumor avoids missing small but important mutations that high-throughput discovery platforms such as microarrays cannot detect above background, due to intratumoral heterogeneity. This approach may lead to diagnostic and prognostic markers or therapeutic targets for different stages in progression. We propose several high-throughput mutation detection schemes that will allow us to build oncogenetic trees with the necessary efficiency to make the oncogenetic tree approach practical. PUBLIC HEALTH RELEVANCE: Only a small part of a breast cancer tumor is capable of causing metastatic disease, and these "cancer stem cells" are difficult to identify and characterize. By tracing the history of mutations as they develop and disseminate through the tumor, we will be able to identify diagnostic and prognostic markers, as well as potential therapeutic targets for progressively more aggressive forms of the disease.