Competitive interactions in heterogeneous cancer cell populations: a multimodal imaging, machine learning and computational modelling approach

大多数成人癌症起源于上皮细胞片内单个细胞的转化。然后，几轮突变和选择被认为产生了肿瘤祖细胞。在分子水平上，导致癌症进展的一系列事件正变得越来越清晰。最近的研究表明，转化的细胞与转化程度较低的细胞竞争，尽管人们对这一点知之甚少。这种竞争的结果可能取决于环境条件。该项目将使用长期显微镜、最先进的图像分析和计算模型来描述癌症竞争的特征。在这个项目中，我们建议研究微环境选择力如何影响不同细胞系之间竞争的结果。在最简单的配置中，我们将考虑两种细胞类型在癌症进展的不同阶段之间的竞争。我们将通过分析在正常和压力条件下共培养的细胞增殖来检查这种竞争的结果。我们建议研究环境条件，如营养消耗、缺氧或暴露于抗癌药物如何影响细胞类型之间的竞争。实验将在一个定制的显微镜上进行，该显微镜可以在几天内连续成像。将细胞植入圆形微模式或微流体装置中。图像分析使用定制编写的软件管道，可以跟踪单个细胞，识别细胞周期阶段，并生成谱系树。我们将用博弈论来解释我们的实验结果，以理解不同应激条件下细胞类型之间的竞争。在这里，对立的细胞类型采用策略（例如自私或调节生长），旨在最大化其在环境中的适应性。博弈论结构，如囚徒困境，已被用于对混合细胞群体的增长率进行理论预测。在这里，我们试图提供在选择压力下细胞增殖的直接实验测量，与理论预测进行比较，并提供肿瘤发生最早事件的更清晰的图像。

Most adult human cancers originate from transformation of a single cell within an epithelial cell sheet. Then, several rounds of mutation and selection are believed to give rise to tumour progenitors. At the molecular level, the sequence of events leading to cancer progression is becoming better characterised. Recent work suggests that transformed cells compete with their less transformed neighbours, although this is poorly understood. The outcome of such competition likely depends on environmental conditions. This project will use long-term microscopy, state-of-the-art image analysis and computational models to characterise competition in cancer. In this project, we propose to examine how microenvironmental selection forces affect the outcome of competition between different cell lineages. In the simplest configuration, we will consider a competition between two cell types at different stages of cancer progression. We will examine the outcome of this competition by analysing cell proliferation in co-cultures under normal and stressful conditions. We propose to examine how environmental conditions such as nutrient depletion, hypoxia, or exposure to anti-cancer drugs affect competition between cell types. Experiments will be run on a custom built microscope that enables continuous imaging over periods of several days. Cells are seeded either on circular micro-patterns or within microfluidic devices. Images are analysed using a custom-written software pipeline that can track individual cells, recognise cell cycle stage, and generate lineage trees.We will interpret our experimental results using game theory to understand competition between cell types under different stress conditions. Here, opposing cell types adopt strategies (for example selfish or regulated growth) intended to maximise their fitness within the environment. Game theoretic constructs such as the Prisoner's Dilemma, have been used to make theoretical predictions about growth rates in mixed cell populations. Here, we seek to provide direct experimental measurements of cell proliferation under selection pressure, to compare with theoretical predictions, and provide a clearer picture of the earliest events in tumourogenesis.