Membrane ruffles are actin-rich protrusions of the plasma membrane that can be observed on the surface of many cell types. Phase contrast and fluorescent microscopy are widely used in the analysis of ruffles, which are commonly identified in cells stained withfluores-cently labeled phalloidin. Currently, comparison of cellular ruffle formation under different experimental conditions is generally qualitative or semiquantitative. Ruffle structures are often defined using manual tracing and thresholding methods. Here, we report the rapid and accurate segmentation of ruffles from two-dimensional confocal projections of cells using an automated method based on well-established image processing and analysis methods. Line-shaped ruffles were detected using line detectors and were then separated from the filtered images. Automated categorizing of the segmented line structures enabled accurate quantification of the ruffles. This automated approach is efficient and reliable and hence can serve as a powerful tool in studies of the mechanism of ruffle formation.
Membrane ruffles are thin cell surface membrane protrusions that are enriched with filamentous actin (F-actin). These structures have been widely observed in cultured eukaryotic cells, including both stationary and migrating fibro-blasts (1,2,3). F-actin-rich membrane ruffles have also been observed in vivo, for example in embryonic cells (4). Formation of ruffles in migrating cells is known to be closely related to establishing cell polarity; in general, the formation of ruffles is restricted to the leading edge of migrating cells (5). The polarized distributions of cellular components, such as matrix-metal-loproteinase 2 (MMP2) (6), aquaporin 1 (AQP-1) (7), high-affinity αvβ3integrin (8), phosphatidylinositol-4, 5-bisphosphate [PI(4,5)P2] (9), and endocytosed epidermal growth factor receptors (EGFRs) (10) have also been observed to correlate with sites of ruffle formation. While recent studies indicate that ruffles are compartments of inhibited actin turnover that correlate with inefficient lamellipodia adhesion (11), the biochemical mechanisms that regulate ruffle formation are not completely understood.
In analyses of ruffles, phase contrast and fluorescent microscopy are widely used techniques. Ruffles are commonly identified through their unique morphology, observed in cells stained with fluorescently labeled phalloidin to reveal F-actin-containing structures. In dorsal views of two-dimensional (2-D) cell images, the morphology of F-actin-rich ruffles is quite consistent between different cell lines, including Chinese hamster ovary-K1 (CHO-K1), COS-7, NIH 3T3, and embryonic fibro-blasts, and can be described as sharp sinuous line-shaped structures (1). Currently, comparison of cellular ruffle formation under different experimental conditions is generally qualitative or semiquantitative. Ruffle structures are often defined using manual tracing and thresholding methods (12,13,14). These manual operations are time-consuming and thus limit the number of images that can be analyzed. The repeatability of these manual operations can be poor and may also reduce the accuracy of the analyses.
Here, we report that rapid and accurate segmentation of ruffles from 2-D images of cells can be achieved through an automated method based on well-established image processing and analysis methods. Line-shaped ruffles were detected using line detectors, and the ruffles were accurately separated from surrounding cellular structures through thresholding methods. Automated categorizing of the segmented line structures enabled accurate quantification of the ruffles. This automated approach is efficient and reliable and hence can serve as a powerful tool in studies of the mechanism of ruffle formation.Materials and Methods Cells, Transfection, and Reagents
CHO-K1 cells were maintained in Dulbecco's modified Eagle's medium (DMEM; Sigma-Aldrich, St. Louis, MO, USA) containing 10% fetal bovine serum (FBS; Sigma-Aldrich), at 37°C and 5% CO2. Phorbol-12-myristate-13-acetate (PMA) was purchased from Sigma-Aldrich. pEGFP-N1 vector was purchased from Clontech Laboratories (Palo Alto, CA, USA). Rhodamine-phalloidin and 4′6-diamidino-2-phenyl-indole (DAPI) were purchased from Molecular Probes (Eugene, OR, USA), and fluorescent mounting medium was obtained from DakoCytomation (Carpinteria, CA, USA).Cell Stimulation and Immunofluorescent Labeling
To induce ruffle formation, cells were washed with cold phosphate-buffered saline (PBS) three times and detached with 5 mM EDTA in PBS. Detached cells were then plated subcon-fluently on 10 µg/mL fibronectin-coated glass coverslips and allowed to attach and spread on the coverslips for 3 h in serum-free DMEM at 37°C with 5% CO2. Plated cells were treated with 500 nM PMA in DMEM or DMEM alone for 10 min, to induce ruffles or serve as control, respectively. Cells were fixed with 4% paraformaldehyde at 4°C for 30 min and permeabilized with 0.1% Triton® X-100 in PBS for 10 min. Fixed and permeabilized cells were washed three times with PBS and incubated in 5% skim milk/PBS blocking solution for 1 h. After three washes in PBS, cells were incubated with 0.04 U/mL rhodamine-phalloidin for 30 min. Cells were washed three times and then stained with 300 nM DAPI for 5 min to label nuclei. After two more washes, cells were mounted on glass slides using fluorescent mounting medium.Image Acquisition and Outlining Cell Boundaries
Labeled cells were viewed using a Leica TCS SP2 confocal imaging station equipped with a Leica DM-IRE2 inverted microscope and a Leica 63× oil immersion lens (Leica Microsystems, Heidelberg, Germany). Cells were optically sectioned along the z-axis from ventral to dorsal surfaces. The voxel size was 0.23 µm in each of the x, y, and z dimensions. From the acquired z-series, a projection image, which offered a high-resolution 2-D overview, was generated for each cell. The projection image was generated by overlaying the optical sections on top of each other (the ZProject/Sum Slices function in ImageJ software).