Image based phenotypic profiling of single-cell responses to perturbations

描述(由申请人提供)：越来越多的疾病状态和对治疗的反应被认为是由细胞状态和反应的不同组合组成的。对不同细胞对扰动反应的本质的洞察立即应用于药理学和疾病治疗。识别生理或临床上重要的亚群，如癌症耐药细胞或激素不敏感细胞。朝着确定诊断生物标记物和开发靶向治疗迈出的重要一步。测量不同条件下大量单个细胞的高维生理表型，将有助于表征异质细胞对扰动的反应，并识别不同表型状态的离散亚群。因此，我们的长期目标是：1)能够检测和比较单细胞对大范围扰动的反应；2)阐明细胞扰动的生理机制；以及3)识别生理上重要的亚群。我们建议使用免疫荧光显微镜来监测系统药物治疗的复杂细胞反应。对于提议的目标，药物治疗是干扰的理想选择，因为它们是快速作用、可滴定和可靠的方法，可以引发不同的细胞反应。具体目的是：1.提高单细胞表型读数的分辨能力。独立于少量荧光标记物测量的单细胞表型可能不能区分复杂的细胞状态。我们将通过增加每个细胞的标记数量和读数质量，以及通过关联来自不同细胞的不同标记的读数来扩大读出能力。2.使用单细胞表型读数对扰动效应进行分类。对扰动的影响进行分类需要从单细胞表型读数中提取信息特征。我们将应用新的特征选择和药物特征分析方法来识别多相和非靶点药物效应，并预测药物联合治疗的反应。3.将群体异质性表示为具有不同表型状态的亚群体。我们将根据药物对这些细胞亚群的影响来描述药物的特征。

DESCRIPTION (provided by applicant): Increasingly, disease states and responses to therapeutics are seen to be comprised of a heterogeneous mix of cellular states and responses. Insight into the nature of diverse cellular responses to perturbations has immediate applications to pharmacology and disease treatment. Identifying physiologically or clinically important subpopulations, such as cancer drug-resistant or hormone-insensitive cells, is. an important step towards identifying diagnostic biomarkers and developing targeted therapies. Measuring high-dimensional physiological phenotypes of large numbers of individual cells in diverse conditions will enable the characterization of heterogeneous cellular responses to perturbations, and the identification of discrete subpopulations of distinct phenotypic states. Our long term objects are therefore: 1) enabling the detection and comparison of single-cell responses to broad ranges of perturbations; 2) elucidating physiological mechanisms of cellular perturbations; and 3) identifying physiologically important subpopulations. We propose to use immunofluorescence microscopy to monitor complex cellular responses to systematic drug treatments. For the proposed aims, drug treatments are ideal choices for perturbations as they are fast- acting, titratable, and reliable methods for eliciting diverse cellular responses. The specific aims are to: 1. Increase the discriminative capacity of single-cell phenotypic readouts. Single-cell phenotypes, measured independently from small numbers of fluorescent markers, may not distinguish complex cellular states. We will expand readout capacity by increasing the number of markers and quality of readouts per cell, and by correlating readouts of different markers from different cells. 2. Classify perturbation effects using single-cell phenotypic readouts. Classifying the effects of perturbations requires the extraction of informative features from single-cell phenotypic readouts. We will apply new methods for feature selection and drug profiling to the problems of identifying multiphasic and off-target drug effects, and predicting the response to combinations of drug treatments. 3. Represent population heterogeneity as subpopulations of distinct phenotypic states. We will characterize drugs in terms of their effects on these subpopulations of cells.