The mechanical properties and microstructure of the extracellular matrix strongly affect 3D migration of cells. An in vitro method to study the spatiotemporal cell migration behavior in biophysically variable environments, at both population and individual cell levels, is described.
The ability of cells to migrate is crucial in a wide variety of cell functions throughout life from embryonic development and wound healing to tumor and cancer metastasis. Despite intense research efforts, the basic biochemical and biophysical principles of cell migration are still not fully understood, especially in the physiologically relevant three-dimensional (3D) microenvironments. Here, we describe an in vitro assay designed to allow quantitative examination of 3D cell migration behaviors. The method exploits the cell’s mechanosensing ability and propensity to migrate into previously unoccupied extracellular matrix (ECM). We use the invasion of highly invasive breast cancer cells, MDA-MB-231, in collagen gels as a model system. The spread of cell population and the migration dynamics of individual cells over weeks of culture can be monitored using live-cell imaging and analyzed to extract spatiotemporally-resolved data. Furthermore, the method is easily adaptable for diverse extracellular matrices, thus offering a simple yet powerful way to investigate the role of biophysical factors in the microenvironment on cell migration.
Migration of cells plays a key role in various physiological responses such as embryonic development, haemostasis, and immune response as well as in pathological processes such as vascular diseases, inflammation, and cancer1. Dissecting the biochemical and biophysical factors underlying cell migration is therefore fundamentally important not only to understand the basic principles of cellular functions, but also to advance various biomedical applications, such as in tissue engineering, anti-metastasis and anti-inflammatory drug development. Since in vivo observation is technically challenging, a lot of efforts has been focused on in vitro recapitulation of cell migration.
In vitro methods to study cell migration have largely been designed for assays on two dimensional (2D) surfaces, most notably the scratch or wound healing assay2. Such assays offer simple experimental setup, easy live-cell imaging, and provide useful insights into various biochemical mechanisms underlying cell migration. However, these assays do not account for extracellular matrix (ECM) architecture and remodeling, which are critical aspects in understanding in vivo migration. Recently, it has been increasingly appreciated that a 3D culture model, often in collagen-based matrices3, provides a platform that better resembles the in vivo situation. Indeed, cells exhibit migrational dynamics that are distinct from those on 2D surfaces, especially due to the different dimensionality of the environment4. Moreover, the biophysical and mechanical properties of the matrix sensitively affect cell migration5, including in the context of tumor cell invasion6.
Here, we present a method to study 3D cell migration behavior in ECM with biophysical properties that can be easily varied with preparation conditions. The cells are seeded in an “inner gel” and are allowed to escape into and invade the initially acellular “outer gel”. The method relies on the cell’s ability to recognize the presence of, and propensity to migrate into, cell-free regions in the outer gel, which is closely linked to cell mechanosensing7. In this study, we employ collagen networks as the ECMs invaded by highly invasive breast cancer cells, MD-MBA-231. The mechanical properties and microstructure of both the inner and outer gels can be tuned8 and characterized9 to achieve physiologically relevant conditions. Reconstruction and analysis of the cell tracks allow detailed quantitative examination of the spatiotemporal migration behavior at both population level and individual cell level. Importantly, the setup of the concentric gel system mimics the in vivo tissue topology faced by migrating cells, especially invading cancer cells, thus offering important insights into the physical mechanisms of cell migration and metastasis.
1. Cell Harvesting
2. Preparation of Collagen Solutions
3. Formation of Concentric Gel Culture
4. Live-cell Imaging
5. Cell Tracking and Data Analysis
The concentric gel assay presented here was performed using highly invasive breast cancer cells, MDA-MB-231, with 2.4 mg/ml inner collagen gel and a cell seeding density of = 2 × 106 cells/ml, as an example. As shown in Figure 2, typically after a few days of culture, the cells breached the inner–outer gel interface and started to invade the outer gel. The cell population spread predominantly radially outwards.
The polymerization conditions of the outer gel can be modified to study the role of the density and mechanical properties of the matrix on cell migration characteristics. Figure 3 shows the movements of individual cells in outer gels of 1.5 mg/ml, 2.4 mg/ml, and 4.0 mg/ml collagen, polymerized at pH 7.4. Confocal time-lapse images were acquired at the boundary between the inner and outer gel for 8 hr, and the displacement of the cells in the three different collagen concentrations is indicated in white arrows. Depending on the cell seeding density and proliferation rate, ~200 individual cell trajectories can typically be extracted and analyzed in each sample.
We quantitatively analyzed the cell trajectories in the 3 different collagen concentrations in terms of mean net displacement, distance travelled, directional persistence, and mean speed over 8 h of imaging (Figure 4). It was observed that the mean displacement and distance were highest for 4.0 mg/ml concentration at 58 µm and 141.5 µm, respectively, although there was no significant difference in the displacement and distance between cells in 2.4 mg/ml and 4.0 mg/ml gels. In 1.5 mg/ml gel, the mean displacement and distance were smallest. This observation may appear to contradict recent reports that show that invasion efficiency decreases in gels with increasing collagen concentration11,12. Note, however, that the displacements and distances in our assay are measured from all individual cell tracks within predominantly radially outward migrating population. These parameters are therefore much less susceptible to the fastest-moving cell subpopulation and are not completely dominated by stochastic movements.
In all cases, the total distance covered (Figure 4B) was greater than the net displacement (Figure 4A). Defining persistence simply as the ratio between displacement to distance, we obtained persistence of ~0.4 across all gel conditions (Figure 4C). This relatively low persistence reflects the intrinsically weak directional migration of the cells in our assay. We anticipate that the presence of directional chemotactic stimuli or biochemical gradients will increase the persistence, potentially in collagen concentration-dependent manner. Furthermore, we also observed that the mean speed of cell migration did not significantly change with collagen concentration (Figure 4D), likely because of the interplay between matrix stiffness (which increases with collagen concentration) and pore size (which decreases with collagen concentration)10, as well as the spatiotemporal variation of migration characteristics within the cell population7. The mean speed varied between 31 µm/hr and 37.5 µm/hr, with a minimum individual cell speed of 7 µm/hr and a maximum individual cell speed of 114 µm/hr.
Figure 1. Schematic of the steps involved in setting up the concentric gel 3D migration assay. (A) Formation of the inner collagen gel at the center of the well in the glass-bottom dish, containing the cells. (B) Formation of the (acellular) outer collagen gel encapsulating the inner gel. (C) Immersion of the gels in cell culture medium. The cells are then allowed to equilibrate and attach to the surrounding fibrous matrix. This step can take a few hours up to 1 day after the initiation of gel polymerization. (D) Live-cell imaging, monitoring the spreading of cell population into the outer gel.
Figure 2. Spreading of cell population. (A) After 12 days of culture, MDA-MB-231 cells had breached the inner–outer gel interface (green dashed line) and invaded outwardly into the originally acellular outer gel (scale bar = 200 µm). The brighter areas are regions occupied by cells, whereas the darker areas are cell-free regions in this phase contrast image. The yellow rectangles represent the VoV’s of interest, where the 3D cancer cell invasion was monitored. The cells migrated predominantly radially outward, as indicated by the arrows. (B) Overlay of confocal fluorescence image of the migrating MDA-MB-231 cells (red; yellow when overlaid with green pseudo-colored collagen) and reflectance image of the collagen matrix (green). The arrows point to the migration direction of individual cells within a time span of 5 h. Scale bar = 50 µm. (C) From the time-lapse images, tracks of the cells can be obtained. Here the cell tracks are color-coded for time, as indicated by the color bar (timescale = 8 hrs). Adapted from Sun et al7,10. Please click here to view a larger version of this figure.
Figure 3. Cell movement in different collagen gel densities. The movement of MDA-MB-231 cells in outer gels with three different outer collagen concentrations: 1.5 mg/ml, 2.4 mg/ml, and 4.0 mg/ml. Cells were labeled with live-cell tracker CMTMR (red). The white arrows indicate the net displacement of individual cells within 8 hrs of imaging. Scale bar = 100 µm. Please click here to view a larger version of this figure.
Figure 4. Analysis of cell migration trajectories. The migration trajectories of the individual MD-MBA-231 cells over 8 hrs were quantified in terms of (A) net displacement, (B) distance, (C) persistence, and (D) mean speed for different outer gel collagen concentrations: 1.5 mg/ml, 2.4 mg/ml, and 4.0 mg/ml. Time-lapse images were acquired near the interface of the inner and outer gels. Data are mean ± standard error of the mean (s.e.m.) obtained from more than 200 cell trajectories in 3 independent experiments in each condition. Asterisks (*) denote statistically significant difference (p-value < 0.01).
In this protocol we describe an in vitro assay to study the 3D migrational behavior of cells in matrix environments that topologically resemble ECMs encountered in vivo. There are three main strengths of this assay as compared to other currently available methods. First, this assay allows one to simultaneously examine the cell migration mechanisms at both population level and individual cell level. This opens up possibilities of studying collective cell migration13, which has to date been largely limited to 2D assays14, while at the same time observing the behavior of the individual cell within the population. For cells that do not naturally migrate in groups or clusters, the assay offers a way to analyze the migratory properties in detail from a large number of cells with diverse local microenvironment, which can provide important insights into the biophysical origins of migration efficiency and directional persistence7,15.
Second, by modifying the gel conditions of the inner and outer gels, one can directly study the effect of matrix mechanics and microstructure on 3D cell migration behavior10. This is especially important, as it is known that cell migration and cancer invasion are highly influenced by matrix mechanics and porosity16,17, and that soft tissues in our body have a very diverse topology18. Moreover, the method is readily adaptable for matrices other than collagen. This is advantageous, as not all cell lines can spread and migrate in collagen-based matrices. A critical step that must be noted is that, to ensure physical connectivity of the gel interface, the formation of the outer gel must be initiated before complete polymerization of the inner gel. Since this is strongly dependent on the specific polymerization kinetics of the gels (also for collagen prepared under different conditions, e.g., concentration, polymerization pH, temperature, cross-linking8), an optimal polymerization duration should first be systematically identified. Imperfections in the interfacial connectivity between the inner and outer gel usually results in a gap and collection of the spreading cell population in the interface. Note, however, that perfect interfacial connectivity can introduce a practical problem: the boundary between the inner and outer gels becomes less obvious when imaging. If necessary, sparse fiducial markers, such as inert beads, can be used to distinguish the inner and outer gels.
Third, the possibility of long-term 3D culture (up to 3 weeks in our laboratory) and live-cell imaging allows a wide application of the assay. It is now possible to obtain spatially-resolved viscoelastic measurements of the local matrices surrounding the cells using microrheological approaches19. Furthermore, our recent study10 reveals that the efficacy of various anti-migratory drugs (administered on the 7th day of the experiment) can be strongly affected by the architecture and stiffness of the matrix (i.e., outer gel). In the future, we plan to use these experimental frameworks to further understand how such effects can arise from the basic molecular mechanisms of cell migration (e.g., cytoskeletal contractility20, extracellular matrix reorganization21, and pericellular proteolysis22).
One limitation of this assay is the asymmetry of the mechanical support: the top surface of the gel is not attached to any solid surface, while the bottom surface adheres to the glass surface of the dish. While edge effects can be avoided simply by excluding cells near the edges, the lack of mechanical support at the top surface can potentially lead to in-plane preferential migration and therefore introduce artificial flattening of the 3D cell trajectories7. One way to solve this issue while still maintaining oxygen and nutrition supply to the cells is by topping the gel with holed glass plate23. Additionally, to further suppress any wall effect, a thin layer of collagen gel can be first made on the glass surfaces, before the formation of the inner and outer gels. This ensures complete geometrical encapsulation of the gels.
To obtain meaningful and accurate description of cell migration behavior, it is important to consider the time scale and the nature of migration, which can vary considerably between cell lines. Preliminary experiments must be systematically performed to understand these factors, and then the interval time between image captures and live-cell imaging duration can be adjusted. The interval time should be no larger than the characteristic time scale of cell turning. It is also important to consider that cell migration behavior may change both spatially and temporally, depending on the biophysical conditions of the local microenvironment7. It is always a good idea to simultaneously visualize the cells and the fibrous matrix, for example using multi-color confocal fluorescence or multimodal (e.g., reflectance, second harmonic generation) imaging. After imaging, post-processing of the time-lapse 3D image, construction of the cell tracks, and analysis of the trajectories can in principle be done with any image processing software.
In conclusion, the concentric gel assay described here provides a simple and useful experimental system for quantitatively studying 3D cell migration, mechanosensing and mechanotransduction in diverse microenvironments.
The authors have nothing to disclose.
The authors thank W. Sun and K. Jansen for the critical discussions, and acknowledge support by the Nano Biomechanics Lab at the National University of Singapore. N.A.K. acknowledges support by a Marie Curie IIF Fellowship.
Cell culture incubator | Fisher Scientific Pte Ltd | Model: 371, S/No 318854-6055 | |
Confocal microscope | Nikon A1R | Inverted confocal laser scanning microscope equipped with incubator chamber | |
Dulbecco's Modified Eagle's Medium (DMEM) | Life Technologies | 11965-092 | |
Fetal Bovine Serum (FBS) | Life Technologies | 10082-147 | |
Fluorescent CellTracker dye CMTMR | Life Technologies | C2927 | |
Glass-bottom dish | IWAKI Cell Biology | 3931-035 | 35 mm diameter dish with 12 mm diameter glass-bottom well |
Hemocytometer | iN CYTO | DHC-N01 (Neubauer Improved) | |
Microprocessor pH meter | Hanna Instruments | pH 211 | |
Nutragen Collagen | Advanced BioMatrix | #5010-D | Acid-solubilized bovine collagen type I (stock pH ~ 2) |
Objective lens | Nikon | CFI Super Plan Fluor ELWD ADM 20XC, W.D. 8.2-6.9mm, NA 0.45. | |
Penicillin-Streptomycin | Life Technologies | 15140-122 | |
pH meter | Sartorius | S/No 29153352 | Basic pH Meter PB-11 |
Trypsin-EDTA | Life Technologies | 15400-054 |