The article describes quantification of 1) the size and number of focal adhesions and 2) cell shape index and its distribution from confocal images of the confluent monolayers of MCF7 cells.
The methods presented here quantify some parameters of confluent adherent cell monolayers from multiple appropriately stained confocal images: adhesion to the substrate as a function of the number and size of focal adhesions, and cell shape, characterized by the cell shape index and other shape descriptors. Focal adhesions were visualized by paxillin staining and cell-cell borders were marked by junction plakoglobin and actin. The methods for cell culture and staining were standard; images represent single focal planes; image analysis was performed using publicly available image processing software. The presented protocols are used to quantify the number and size of focal adhesions and the differences in cell shape distribution in the monolayers, but they can be repurposed for the quantification of the size and shape of any other distinct cellular structure that can be stained (e.g., mitochondria or nuclei). Assessing these parameters is important in the characterization of the dynamic forces in adherent cell layer, including cell adhesion and actomyosin contractility that affects cell shape.
Epithelial cell monolayers act as a collective in which cell-cell and cell-substrate adhesion as well as contractile forces and tensions represent important parameters and their proper balance contributes to the overall integrity of the unit1,2,3. Thus, assessing these parameters represents a way to establish the current status of the cell layer.
The two methods described here represent a two-dimensional analysis of the confluent monolayers of adherent, epithelial cells (in this case MCF7 breast cancer cell line). The analysis is performed using confocal images (single Z-slices) from different regions on the Z-axis; basal region near a substrate for focal adhesion (FA) measurements and apical region for cell shape measurements. The presented methods are relatively simple and require standard laboratory techniques and open-source software. Confocal microscopy is sufficient for this protocol, so it can be performed without employing more specialized TIRF (Total Internal Reflection Fluorescence) microscopy. Thus, the protocol could be implemented in a relatively standard laboratory setting. Although the accuracy of the methods is limited, they can distinguish basic differences in focal adhesion and cell shape.
Both methods described here consist of the standard experimental procedures such as cell culturing, immunostaining, confocal imaging and image analysis performed using ImageJ. However, any image processing software with the appropriate functions can be used. The presented methods can track and compare changes inflicted by pharmacological treatment or minimal genetic modification. Obtaining definite values is not recommended, due to the limited precision of these methods. Two automated macros were included, to facilitate the measurements of many images.
1. Preparatory steps
2. Image analysis
NOTE: Provided macros work optimally on ImageJ version 1.50f or newer. Use for quantification only of images with a high signal-to-noise ratio and without under- or oversaturated pixels. The described methods include steps requiring manual parameter adjustment. Thus, a blind analysis/blinded experiment setup is recommended. For encrypting image file names, ImageJ plugins such as “Blind Analysis Tool” (available at: https://imagej.net/Blind_Analysis_Tools) can be used.
3. Quantification
Focal adhesion analysis
The knockdown of HAX1 gene was previously shown to affect focal adhesions6. Cells were cultured on collagen I-coated surface for 48 h. Images of the MCF7 control cells and MCF7 cells with a HAX1 knockdown (HAX1 KD) from three independent experiments stained with focal adhesion protein paxillin were obtained using a confocal microscope (image from single focal plane/Z-slice from basal region). FAs from about 2,000-2,500 cells from each cell line were quantified using the described protocol. The mean value for the smallest focal adhesion was set to 50 (pixel2). Representative images of FAs count with ImageJ, including the final, numbered outlines and the overlay of the FAs outlines with the original image, are shown for both cell lines on Figure 3A. Differences in number and size of FAs in both cell lines are presented on Figure 3B.
Cell shape analysis
Manual assessment: MCF7 cells were cultured for 24 h, the medium was exchanged for the same fresh medium (untreated) or the medium with 0.1 μM paclitaxel (PTX) – to induce cell rounding – and cultured for another 24 h. Images of the confluent MCF7 monolayers stained with anti-plakoglobin antibody and DAPI were obtained using a confocal microscope (single Z-slice from apical region). About 200-400 cells (2-3 fields of view) from each experiment (untreated/treated) were documented (40x objective) and the images were assessed using ImageJ image processing software (Figure 4A). Representative regions of each cell layer with outlined cells are shown in Figure 4A (untreated and PTX-treated cells). All cells were categorized according to their CSI values (10 intervals, bin width 0.1) and presented in the respective histograms (Figure 4B), which show an increase in the last bin (0.9-1) and the flattening of the main peaks (0.6-0.9) for the PTX-treated cells, comparing to the untreated control. Differences in the cumulative distribution of CSI were calculated for statistical significance using Kolgomorov-Smirnov test (K-S), a nonparametric test of the equality of one-dimensional probability distributions. Frequency distribution histograms and cumulative distribution plots (Figure 4C) were generated with software.
Automated assessment: MCF7 cells were cultured for 24 h, fixed and stained with fluorophore-conjugated phalloidin to visualize actin. Images were taken using a confocal microscope (single Z-slice from apical region). The grayscale images of the monolayers were analyzed using the attached macro file, according to the included protocol (Figure 5A-C). Overall, 512 cells from 12 fields of view were quantified. Results were plotted as a histogram presenting distribution of circularity (Figure 5D).
Figure 1: Actin staining with fluorophore-conjugated phalloidin, two different Z-slices from the same field of view. (A) The apical region with cortical actin. (B) The basal region with actin stress fibers. Bar: 10 μm. Please click here to view a larger version of this figure.
Figure 2: Examples of different Cell Shape Index (CSI) values for the shapes with distinct perimeters, but the same area. The very elongated shape on the left has CSI close to 0, while the ideal circle on the right has CSI of 1. Please click here to view a larger version of this figure.
Figure 3: Focal adhesion quantification. (A) Representative images of MCF7-based cell lines: CONTROL and HAX1 KD stained with paxillin; from left to right: (1) paxillin and nuclei (2) outlined and numbered FAs (3) overlay of FAs outlines and the original image. Bar: 20 μm. Insets (below each picture) show zoomed boxed areas. (B) Plots showing the difference in FAs size and number between the two cell lines. Statistical significance was assessed using Student t-test. Please click here to view a larger version of this figure.
Figure 4: Changes in cell shape induced by paclitaxel (PTX) in MCF7 cell layer; manual count. (A) Representative regions of MCF7 monolayers, cells untreated and treated with PTX, stained with junction protein plakoglobin and DAPI, bar: 20 μm. The panel on the right shows the image processed in ImageJ; cell edges are outlined and all counted cells are numbered. (B) Histograms showing CSI distribution in untreated and PTX-treated cells, 200-400 cells in each experiment (untreated/treated), bin width: 0.1. Plots were generated using frequency distribution analysis with relative frequencies as a fraction. (C) The plot of cumulative distribution for the untreated and PTX-treated monolayers. Statistical difference calculated using Kolgomorov-Smirnov nonparametric test (KS). Please click here to view a larger version of this figure.
Figure 5: Cell shape in MCF7 monolayers assessed using the automated method. (A-C) An example of the automated image analysis illustrating subsequent steps executed by the attached macro. (A) Input image; cortical actin; grayscale, scale bar 10 μm (for informative purposes; scale bars should not be embedded into analyzed images). (B) Cell layer after segmentation; cell borders outlined; cells without complete borders eliminated. (C) Overlay of the cell outlines on the original image. (D) A histogram showing the distribution of CSI in the analyzed dataset, 512 cells. Please click here to view a larger version of this figure.
Supplemental Figure S1, S2: MCF7 cells, paxillin and DAPI-stained to visualize FAs and nuclei. Provided as a training dataset for the FAs.ijm macro.
Figure S1: Please click here to download this figure.
Figure S2: Please click here to download this figure.
Supplemental Figure S3, S4: MCF7 cells, plakoglobin and DAPI-stained to visualize cell-cell junctions and nuclei. Provided as a training dataset for the CSI.ijm macro.
Figure S3: Please click here to download this figure.
Figure S4: Please click here to download this figure.
Cell-cell and cell-substrate adhesion constitute inherent attributes of the epithelial cells and play the critical role in tissue morphogenesis and embriogenesis. In adult tissues the proper regulation of mechanical properties of the cell layer is crucial in maintaining homeostasis and preventing pathological responses like tumor progression and metastasis. The size and number of focal adhesions depend on the strength of cell-substrate adhesion, while cell shape depends on contractile forces and is related to the status of cell-cell-contacts.
Here, we describe two simple methods of quantitative analysis of the area, number and shape of cellular structures stained by immunofluorescence, in this case focal adhesions and the whole cells in the cell layer. However, the proposed tools can be repurposed for the quantification of any chosen structure. The key issue for these analyses is the quality of immunofluorescent staining and confocal imaging. These methods can be implemented in any standard laboratory equipped with a cell-culture unit and a confocal microscope. They are designed to compare cell lines, especially when the differences (natural or induced by specific treatment) in the measured parameters are substantial. They are not recommended to measure minute differences or to establish absolute measurements, because they are sensitive to minimal changes in the initial arbitrary settings, especially in the case of focal adhesions. This method of FAs quantification is inferior to more advanced and specific methods like TIRF microscopy, but it has an advantage of not requiring sophisticated equipment.
Similar methods of focal adhesion measurements were described before7,8,9,10. Here, we specified the settings and created a free ImageJ macro, with several options, to facilitate the measurements.
Cell shape analysis was described many times, including very complex and detailed methods11,12. Here, we present a simple method to track the changes in epithelial cell monolayers, which could be very important for comparing cell morphology or developmental changes. The manual method of cell shape analysis in the monolayers presented in this protocol was described in a previous report6. The CSI formula as a way to describe and compare the shape of an object(s) is widely used in different disciplines13,14, including geology from which it originated. Presentation of the results in a form of a histogram and/or as a cumulative distribution function is commonly used for comparisons of distributions of any kind8,10,15.
Notably, we present here a tool for the automated cell shape analysis based on ImageJ plugin MorhpLibJ (https://imagej.net/MorphoLibJ). We provide a macro file, which can perform this analysis quickly and efficiently. This method does not always recognize the correct cell borders, but the percentage of faulty measurements (which are present in every automated analysis) is minimal and should not significantly affect the final result, especially if enough number of cells is analyzed. Cells without complete borders are eliminated. Automated cell shape analysis has undeniable advantages and we present this method so that it can be appreciated by the scientific community.
The authors have nothing to disclose.
This work was supported by the Polish National Science Center under grant no. 2014/14/M/NZ1/00437.
Alexa Fluor 594 | ThermoFisher Scientific | A32740 | goat anti-rabbit, 1:500 |
Ammonium chloride | Sigma | A9434 | |
BSA | BioShop | ALB001.500 | |
Collagen from calf skin | Sigma | C9791-10MG | |
DAPI | Sigma | D9542 | 1:10000 (stock 1 mg/mL in H2O), nucleic acid staining |
DMEM + GlutaMAX, 1 g/L D-Glucose, Pyruvate | ThermoFisher Scientific | 21885-025 | |
FBS | ThermoFisher Scientific | 10270-136 | |
Junction plakoglobin | Cell Signaling | 2309S | rabbit, 1:400 |
Laminar-flow cabinet class 2 | Alpina | standard equipment | |
MCF7-basedHAX1KD cell line | Cell line established in the National Institute of Oncology, Warsaw, described in Balcerak et al., 2019 | MCF7 cell line withHAX1knockdown | |
MCF7 cell line (CONTROL) | ATCC | ATCC HTB-22 | epithelial, adherent breast cancer cell line |
Olympus CK2 light microscope | Olympus | ||
Paxillin | Abcam | ab32084 | rabbit, 1:250, Y113 |
PBS | ThermoFisher Scientific | 10010023 | |
Phalloidin-TRITC conjugate | Sigma | P1951 | 1:400 (stock 5 mg/mL in DMSO), actin labeling |
PTX | Sigma | T7402-1MG | |
TBST – NaCl | Sigma | S9888 | |
TBST – Trizma base | Sigma | T1503 | |
Triton X-100 | Sigma | 9002-93-11 | |
Zeiss LSM800 Confocal microscope | Zeiss |