Automated Podosome Identification and Characterization in Fluorescence Microscopy Images
Automated Podosome Identification and Characterization in Fluorescence Microscopy Images
复制标题
DOI：
10.1017/s1431927612014018
复制
发表时间：
2013-02-01
影响因子：
2.8
通讯作者：
van den Dries, Koen
中科院分区：
文献类型：
--
作者：
Meddens, Marjolein B. M.;Rieger, Bernd;van den Dries, Koen
粒体是参与基质降解和侵袭的细胞粘附结构，由肌动蛋白核心和细胞骨架衔接蛋白环组成。它们最常通过鬼笔环肽染色来鉴定，鬼笔环肽结合F-肌动蛋白，因此使核心可视化。然而，不仅是podosome，而且许多其他细胞骨架结构含有肌动蛋白，这使得通过自动图像处理分割podosome变得困难。在这里，我们已经开发了一种定量图像分析算法，该算法被优化以识别用鬼笔环肽染色的典型样品内的足状体核心。通过连续的局部和全局阈值，我们的分析确定了高达76%的不包括其他F-肌动蛋白为基础的结构的podosome核心。基于重叠的podosome识别和量化的podosome数，我们的算法表现同样好相比，三个专家。使用我们的算法，我们显示了肌动蛋白聚合和肌球蛋白II抑制对肌动蛋白强度的影响，在两个podosome核心和相关的肌动蛋白网络。此外，通过扩大核心分割，我们揭示了一个以前不受重视的差异分布的细胞骨架适配器蛋白的podosome环内。这些应用程序表明，我们的算法是一个有价值的工具，快速，准确的大规模分析的podosomes，以增加我们的理解，这些特性的粘附结构。
Podosomes are cellular adhesion structures involved in matrix degradation and invasion that comprise an actin core and a ring of cytoskeletal adaptor proteins. They are most often identified by staining with phalloidin, which binds F-actin and therefore visualizes the core. However, not only podosomes, but also many other cytoskeletal structures contain actin, which makes podosome segmentation by automated image processing difficult. Here, we have developed a quantitative image analysis algorithm that is optimized to identify podosome cores within a typical sample stained with phalloidin. By sequential local and global thresholding, our analysis identifies up to 76% of podosome cores excluding other F-actin-based structures. Based on the overlap in podosome identifications and quantification of podosome numbers, our algorithm performs equally well compared to three experts. Using our algorithm we show effects of actin polymerization and myosin II inhibition on the actin intensity in both podosome core and associated actin network. Furthermore, by expanding the core segmentations, we reveal a previously unappreciated differential distribution of cytoskeletal adaptor proteins within the podosome ring. These applications illustrate that our algorithm is a valuable tool for rapid and accurate large-scale analysis of podosomes to increase our understanding of these characteristic adhesion structures.