Migratory Metrics of Wound Healing: A Quantification Approach for in vitro Scratch Assays

Abstract

Metastatic dissemination generates an aggressive disease facilitated by enhanced migratory and invasive properties. Experimental approaches employ several in vitro and in vivo assays toward quantification of these functionalities. In vitro assessments of cell motility often employ endpoint assays that rely on the global efficacy of wound closure and thwart quantification of migratory phenotypes observed during metastatic dissemination. Recent studies highlight the distinct signatures associated with individual vs. collective cell migration and necessitate the incorporation of these modalities into routine analyses. Advances in live cell imaging that permit real-time visualization of pathophysiological processes can be employed toward elucidating phenotypic plasticity associated with cell migration to overcome caveats inherent to end-point assays. Herein, we corroborate live cell imaging with the in vitro scratch assay toward quantification of migratory modalities in transformed cells. Our protocol describes a step-by-step approach for live cell setup of the scratch assay, and details analyses employed toward definition of three quantitative metrics viz., displacement, velocity and number of nearest neighbors. The current protocol (from scratch induction to data acquisition) is implemented for ~30 h and provides global/single-cell resolution of migratory phenotypes as opposed to the endpoint assays. Routine application of this protocol in cancer biology can aid the design of therapeutic regimes targeting specific migratory modalities and significantly contribute to the dissection of associated molecular networks.

Keywords: cell migration; displacement; fiji; live cell imaging; migratory modalities; nearest neighbors; velocity.

Figure 1

Image processing of time lapse data. Pop-up windows generated during image processing include the (A) import of data as a gray-scale image, (B) selection of appropriate sigma radius for a Gaussian blur, (C) illumination correction with Image calculator, (D) CLAHE assisted contrast enhancement, and (E,F) definition of particles by Threshold tool.

Figure 2

MTrack2 assisted extraction of migratory data. (A) MTrack2 plugin can be applied to post-threshold images to extract the positional co-ordinates for particles based on size and velocity; (B) Schematic representation of array arrangement used for calculation of cell displacement and velocity; Schematic representation of Graphpad Prism 5 tools used for representing (C) displacement trajectories and (D) velocity.

Figure 3

Detection of nearest neighbors with BioVoxxel plugin. (A) BioVoxxel plugin application to post-threshold images extracts the nearest neighbors in particle vicinity based on definition of neighbor radius, particle size and circularity by the user; (B) Representative graph generated by BioVoxxel plugin to depict the frequency of nearest neighbors; Data from this file can be exported as.csv file (C); (D) Schematic representation of Graphpad Prism 5 tools used for representing frequency of nearest neighbors.

Figure 4

PCA analysis in MATLAB. (A) PCA analysis in MATLAB involves the import of data from a.csv file wherein the data is arranged as depicted. Import of data involves the selection of cells containing values for the migration metric and their import as a matrix (arrow in red); (B) Sample names are imported by selection of cells containing sample labels and conversion to text (blue arrow). The labels are imported as a cell array (red arrow); Output for variance (C) and scores for PC1 and PC2 (D) are obtained following execution of the command line; (E) Schematic representation of Graphpad Prism 5 tools used for representing PC1 and PC2 scores obtained from the analysis in MATLAB.

Figure 5

Derivation of quantitative metrics for migration in ovarian cancer cell lines. (A) Percent wound closure derived from in vitro scratch assays for A4, OVCA420, and OVCAR3 cells in the absence (SS) and presence of serum (Ser); (B) Trajectories depicting direction of migration for (i) A4, (ii) OVCA420, and (iii) OVCAR3 cells derived from : “x” and “y” positional co-ordinates over a 16 h duration of live cell imaging.; (C) Representative boxplots depicting mean migratory velocity for A4, OVCA420, and OVCAR3 cells; (D) Frequency of nearest neighbors for (i) A4, (ii) OVCA420, and (iii) OVCAR3 cells at 0 h (red) and 16 h (green) time points. Experiments were performed in the absence and presence of serum and altered migratory metrics were duly noted. All data are representative of experiments performed in triplicate and are depicted as mean ±SD, *p < 0.05, **p < 0.01, ***p < 0.001. Details of PCA values are provided in Table S2.

Figure 6

Validation and corroboration of quantitative metrics. (A) Percent wound closure derived from in vitro scratch assays for DA3 cells in the absence of serum (SS) and presence of HGF/SF; (B) Trajectories depicting direction of migration for DA3 cells derived from “x” and “y” positional co-ordinates over a 16 h duration of live cell imaging in the presence of HGF/SF as compared to control (SS); (C) Representative boxplots depicting mean migratory velocity for DA3 cells; (D) Frequency of nearest neighbors for DA3 cells at 0 h (red) and 16 h (green) time points. Alterations in migratory metrics were duly noted; (E) Principle component (PC) analysis of time-lapse imaging-based migration data of A4, OVCA420, OVCAR3, and DA3 cells, PC1—variance between displacement (Final Y) and velocity vs. nearest neighbors, PC2—variance between displacement and velocity, filled (blue) and empty shapes indicate presence and absence of serum, respectively; HGF/SF exposure is represented by filled yellow shape. All data are representative of experiments performed in triplicate and are depicted as mean ±SD, *p < 0.05, **p < 0.01, ***p < 0.001. Details of PCA values are provided in Table S2.