Collective cell migration in development, wound healing, and cancer metastasis is often guided by the gradients of growth factors or signaling molecules. Described here is an experimental system combining traction microscopy with a microfluidic system and a demonstration of how to quantify the mechanics of collective migration under biochemical gradient.
Cells change migration patterns in response to chemical stimuli, including the gradients of the stimuli. Cellular migration in the direction of a chemical gradient, known as chemotaxis, plays an important role in development, the immune response, wound healing, and cancer metastasis. While chemotaxis modulates the migration of single cells as well as collections of cells in vivo, in vitro research focuses on single-cell chemotaxis, partly due to the lack of the proper experimental tools. To fill that gap, described here is a unique experimental system that combines microfluidics and micropatterning to demonstrate the effects of chemical gradients on collective cell migration. Furthermore, traction microscopy and monolayer stress microscopy are incorporated into the system to characterize changes in cellular force on the substrate as well as between neighboring cells. As proof-of-concept, the migration of micropatterned circular islands of Madin-Darby canine kidney (MDCK) cells is tested under a gradient of hepatocyte growth factor (HGF), a known scattering factor. It is found that cells located near the higher concentration of HGF migrate faster than those on the opposite side within a cell island. Within the same island, cellular traction is similar on both sides, but intercellular stress is much lower on the side of higher HGF concentration. This novel experimental system can provide new opportunities to studying the mechanics of chemotactic migration by cellular collectives.
Cellular migration in biological systems is a fundamental phenomenon involved in tissue formation, the immune response, and wound healing1,2,3. Cellular migration is also an important process in some diseases like cancer4. Cells often migrate as a group rather than individually, which is known as collective cell migration4,5. For cells to move collectively, sensing of the microenvironment is essential6. For instance, cells perceive physicochemical stimuli and respond by changing motility, cell-substrate interactions, and cell-cell interactions, resulting in directional migration along a chemical gradient7,8,9,10. Based on this connection, rapid advancements have been made in lab-on-a-chip technologies that can create well-controlled chemical microenvironments such as the gradient of a chemoattractant11,12,13. While these lab-on-a-chip-based microfluidics have previously been used to study chemotaxis of the cellular ensemble or cellular spheroids14,15,16,17, they have been used mostly in the context of single-cell migration18,19,20,21. Mechanisms underlying a cellular collective response to a chemical gradient is still not well-understood14,22,23,24,25,26. Thus, the development of a platform that enables the spatiotemporal control of soluble factors as well as in situ observation of cells' biophysical will help to unravel the mechanisms behind collective cell migration.
Developed and described here is a multi-channeled microfluidic system that enables the generation of a concentration gradient of soluble factors that modulates migration of patterned cell clusters. In this study, hepatocyte growth factor (HGF) is chosen to regulate the migratory behavior of Madin-Darby canine kidney (MDCK) cells. HGF is known to attenuate cell-cell integrity and enhance the motility of cells27,28. In the microfluidic system, Fourier transform traction microscopy and monolayer stress microscopy are also incorporated, which allows analysis of the motility, contractile force, and intercellular tension induced by constituent cells in response to an HGF gradient. Within the same island, cells located near the higher concentration of HGF migrate faster and show lower intercellular stress levels than those on the side with lower HGF concentration. The results suggest that this new experimental system is suitable to explore other questions in fields involving collective cellular migration under chemical gradients of various soluble factors.
NOTE: Lithography of SU-8 molds for stencils (thickness = 250 μm) and microchannel parts (thickness = 150 μm), glass etching (depth = 100 μm), and cast fabrication were outsourced by sending designs using computer-aided design software to manufacturers.
1. Fabrication of polydimethylsiloxane (PDMS) stencil and microchannel
2. Preparation of bottom glass with polyacrylamide (PA) gel
3. Micropatterning of cell islands
4. Assembly of PA gel with PDMS microchannel
5. Integrated microfluidic system
6. Image acquisition
7. Data analysis
NOTE: A custom code for data analysis was developed using MATLAB, and details have been described elsewhere32,33,34,35,36.
To explore collective migration under a chemical gradient, a microfluidic system was integrated with traction microscopy (Figure 1). To build the integrated system, polyacrylamide (PA) gel was cast on custom-cut glass, and MDCK cells were seeded within micropatterned islands made by a PDMS stencil. For this experiment, twelve islands of MDCK cells (four rows by three columns, diameter of ~700 μm) were created. After cells attached to PA gels, the PDMS stencil was removed to initiate collective migration. The pre-made microfluidic channel was placed on top of the PA gels to build microfluidic channel above cellular islands (Figure 1A).
To then establish a concentration gradient within the microfluidic channel, the left inlet was connected to the supply reservoir containing a fluorescent dye solution. Additionally, the middle and the right inlets were connected to the supply reservoirs containing media only. Then, the outlet was connected to the syringe pump to remove fluid from the microfluidic channel. The sandwich of custom-cut glass and the microfluidic channel was inserted onto the custom-built microscope stage. The samples along with the automated fluorescent microscope was kept within an incubator for live cell imaging. Throughout the experiments, cells were kept at 5% CO2 and 37 °C. A gradient of the intensity of the fluorescent dye was quickly established at a flow rate of 0.125 – 2 µL/min (Figure 2A). The slope of the gradient depended on the flow rate. For example, at a flow rate of 0.125 µL/min, the gradient was developed over 1.5 mm, which is the comparable size of cell islands (Figure 2B).
Next, the integrated device with cells was tested. The islands of MDCK cells were micropatterned and assembled with the microfluidic channel as described above. First, without applying any flow in the microchannel, it was ensured that the cellular island spread well within the device. Over a 12 h period, the island expanded, and the average migration speed was ~0.1–0.2 μm/min. While there were variations among islands regarding how much the island expanded, all cells looked healthy in the no-flow condition. Next, flow was applied over these islands, and it was found that at a flow rate of 0.1 µL/min, the cellular island spread well, and cells maintained an average migration speed of ~0.1–0.2 μm/min for 12 h. However, upon increasing the flow rate to 0.5 µL/min, cells did not spread well, and the average migration speed decreased to below 0.1 μm/min. Furthermore, morphologies of the cells appeared different and appeared to be detaching from the PA gels (Figure 3). Based on these data, 0.1 µL/min was chosen as the flow rate for the following experiments.
To test the effects of a chemical gradient on collective cell migration, HGF was chosen32,34. The left inlet was connected to 1) the supply reservoir holding cell culture media with HGF solution (20 ng/mL) and 2) the other 2 inlets with reservoirs holding cell culture media. With a flow rate at 0.1 µL/min, flow over the migrating cell islands (four rows by three columns) was maintained for 10 h (Figure 4). As shown in a separate experiment, the HGF concentration on the left column was close to the concentration of the suppling solution (20 ng/mL), and on the right column, this was close to zero. The HGF concentration on the middle column gradually decreased from the left half to right half. When comparing the size of cell islands after 10 h, the islands on the right column did not expand much, but those on the left column became significantly larger and expanded in all directions (Figure 4A). Such changes in island size were supported by the trajectories of cells within each island (Figure 4B). Interestingly, islands in the middle column did not expand toward the right (where the HGF concentration was low) but expanded preferentially toward the left (where the HGF concentration was high). While there was little change in average migration speed in the right column, in the left and middle columns, the average migration speed increased gradually for the first 3 h, then gradually decreased (Figure 4C).
To measure contractile force and intercellular stress, traction microscopy37,39,40 and monolayer stress microscopy were combined with the microfluidic channel33,35,38. Over the course of 10 h during application of the chemical gradient, images of fluorescent particles were taken in PA gels, and the force (per unit area) applied by cells on the substrate was analyzed using Fourier transform traction microscopy35,37. Traction was plotted in polar coordinates. Blue represents the force toward the center of the island, and red represents the force away from the center.
At time zero, all islands showed similar traction distributions, with strong inward traction on the edge and fluctuation within the island. This fluctuation was similar to what has been previously shown34,35,41. After 10 h of applying the HGF gradient, while the degree of island expansion was different in each column, the traction distributions were largely similar to time zero (Figure 5A). The average traction did not change for 10 h (Figure 5C). When calculating the monolayer stress, however, each column showed different trends. In the right column (where HGF concentration was low), the average tension within islands was maintained around 200 Pa throughout a 10 h period. In the left column (where HGF concentration was high), the average tension within islands gradually decreased from 230 Pa to 100 Pa over 10 h, as previously shown32,34. In the middle column (where the HGF concentration was high on the left half and low on the right half of the island), the average tension was maintained around 150 Pa.
Figure 1: Schematic of the integrated device preparation. (A) To fabricate the integrated device, a bottom glass with PA gel on which cell islands were patterned using a PDMS stencil was covered with a PDMS microchannel. (B) The integrated devices were fixed using a custom holder to prevent fluid leakage and component dissociation. (C) After the device filled with fluid without bubbles, reservoirs were connected to the inlets, and a syringe pump was connected to the outlet. All experimental set-ups were placed on a live cell imaging system. (D) The schematic shows the design and composition of the microfluidic channels. Please click here to view a larger version of this figure.
Figure 2: Establishing the chemical gradient. (A) Fluorescence images of rhodamine B-conjugated dextran (70 kDa) visualize the concentration gradient under the fluid flow of 0.125 µL/min, 0.5 µL/min, and 2 µL/min. Dotted circles indicate the position of the patterned cell islands. (B) The intensity profile of the fluorescence image show that the concentration gradient in the device was adjustable by the flow rate. Error bars indicate standard deviation. The colored bars indicate the column positions of the patterned cell island. Please click here to view a larger version of this figure.
Figure 3: Migration of MDCK cell islands under flow. (A) The phase images at 12 h of the MDCK cell island at each flow rate (0 µL/min, 0.1 µL/min, and 0.5 µL/min) show that cell islands expanded well at a low flow rate but did not expand at a high flow rate. Dotted lines represent the boundaries at 0 h. (B) The mean and histogram of the cellular speed within each island every 10 min for 12 h. Please click here to view a larger version of this figure.
Figure 4: Migration of MDCK cell islands under flow and chemical gradient. (A) The relative intensity range of rhodamine B-conjugated dextran maintained a constant HGF concentration gradient in each column after stabilization for 2 h. (B) Phase images at 10 h (dotted line = cell island boundary at 0 h) of the MDCK cell island under the HGF gradient (from left to right: 20–0 ng/mL) show that the island expanded more where HGF concentration was higher. (C) The trajectories of migration paths (color coding indicates the path length) of cells in the MDCK cell islands. Dotted lines and dashed lines represent each cell island’s boundaries at 0 h and 10 h, respectively. (D) The mean and histogram of the cellular speed within each island every 10 min for 10 h. Please click here to view a larger version of this figure.
Figure 5: Mechanical interactions of MDCK cell islands. (A) The maps of traction (cell-substrate interaction) in the cell islands at 0 h and 10 h under the HGF gradient. The maps were plotted in radial coordination, with outward traction using warm color and inward traction using cold color. (B) The maps of monolayer tension (cell-cell interactions) in the cell islands at 0 h and 10 h under HGF gradient. (C) The plot of average traction (gray square) and tension (blue circle) within the cell islands under HGF gradient over time every 10 min for 10 h. The color bands represent the mid-quartile (~25%–75 %). Please click here to view a larger version of this figure.
Supplementary Figure S1: Proper bead distribution required for sufficient data quality during traction force analysis. Examples of good (left) and bad (right) quality fluorescent bead images with enlarged images in the white boxes are shown. Please click here to view a larger version of this figure.
Collective migration of constituent cells is an important process during development and regeneration, and the migrating direction is often guided by the chemical gradient of growth factors4,23. During collective migration, cells keep interacting with neighboring cells and underlying substrates. Such mechanical interactions give rise to emergent phenomena such as durotaxis42, plithotaxis33, and kenotaxis43. However, this research was performed done using a single concentration of growth factors, such as those in fetal bovine serum. To fill the gap, this protocol describes a new experimental platform that can measure mechanical interactions during collective cellular migration under a chemical gradient.
The new platform combines three experimental systems that have been used widely: micropatterning, traction microscopy, and microfluidics. First, micropatterning of cell islands for collective migration was performed. Second, both traction microscopy and monolayer stress microscopy was used for mechanical measurements of migrating cells. Finally, a microfluidics platform was used to establish chemical gradient over migrating cells. With this experimental system, it was demonstrated that cells within a single island migrate differently in response to different chemical gradients.
As described in Figure 2, the chemical gradient forms only over a distance of 1 mm in the middle of the microchannel. Islands in the left column become exposed to the maximum concentration, and those on the right column are exposed to the minimum concentration. In contrast, islands in the middle column are exposed to the chemical gradient. In this way, the response of cell islands is measured at maximum, minimum, and a gradient of concentrations of a soluble factor during the same experiment. Using this array of cell islands, a gradual increase in cellular migration speed was observed as HGF concentration increased.
To achieve high quality data, there are key steps in the procedure that require particular attention. The PA gels in a custom-glass substrate must be perfect regarding the distribution of fluorescent particles and surface coating of collagen. This is because without high quality PA gels, a high-quality data set for traction analysis and monolayer stress analysis cannot be acquired. Examples of high-quality bead images are shown in the Supplementary Figure 1. The microfluidic channels are very small, and bubbles are easily trapped in them. These bubbles not only disturb the flow but can also sweep cells away if they start to move along the flow. Therefore, it is crucial to fill the microfluidic channels without bubbles. To achieve a stable gradient within the microchannel, the balance of pressure within the three inlets is very important. To avoid imbalances between them, large reservoirs should be used so that any imbalances at the beginning of an experiment can easily be adjusted. This new platform will help further exploration of the mechanics of collective cell migration under chemical gradients, which is key to understanding regeneration, cancer metastasis, and development.
The authors have nothing to disclose.
This work was supported by the National Research Foundation of Korea (NRF) grant funded by the Korean government (MSIP) (No. NRF-2017R1E1A1A01075103), Korea University Grant, and the BK 21 Plus program. It was also supported by the National Institutes of Health (U01CA202123, PO1HL120839, T32HL007118, R01EY019696).
0.25% trypsin-EDTA (1X) | Gibco | 25200-056 | |
1 M HEPES buffer solution | Gibco | 15630-056 | |
1 mm Biopsy punch | Integra Miltex | 33-31AA-P/25 | |
100 mm petri dishes | SPL | 10100 | 100 mm diameter, 15 mm height |
14 mm hollow punch | ILJIN | 124-0571 | |
18 mm Ø Coverslip | Marienfeld-Superior | 111580 | Circular 18 mm, thickness No. 1 (0.13 to 0.16 mm) |
2% bis-acrylamide solution | Bio-Rad | 1610142 | Wear protective gloves, clothing, and eye protection. |
3-(Trimethoxysilyl)propyl methacrylate (TMSPMA) | Sigma-Aldrich | 440159-500ML | |
3-way stopcock | Hyupsung | HS-T-61N | CAUTION: do not use if previously opened. do not resterlize or resuse |
30 cm minimum volume line (for pediatric) | Hyupsung | HS-MV-30 | CAUTION: do not use if previously opened. do not resterlize or resuse |
35 mm cell culture dish | Corning | 430165 | |
40% Acrylamide Solution | Bio-Rad | 1610140 | Wear protective gloves, clothing, and eye protection. |
75 cm minimum volume line (for pediatric) | Hyupsung | HS-MV-75 | CAUTION: do not use if previously opened. do not resterlize or resuse |
acetic acid | J.T. Baker | JT9508-03 | |
Ammonium persulfate (APS) | Bio-Rad | 1610700 | |
Antibiotic-Antimycotic | Gibco | 15240-062 | |
Bottom glass chip | MicroFIT | 24 x 24 x 1 mm, custom-made, rectangular groove (6 x 12 mm, depth : 100 μm) | |
Collagen typeI, Rat tail | Corning | 354236 | |
Custom glass holder | Han-Gug Mechatronics | custom-made | |
Dulbecco's Modified Eagle's Medium (DMEM) | Welgene | LM 001-11 | |
Dulbecco's Phosphate Buffered Saline (PBS) | Biowest | L0615-500 | w/o Magnesium, Calcium |
Fetal bovine serum (FBS) | Gibco | 26140-179 | |
FluoSpheres amine-modified microspheres | Invitrogen | F8764 | 0.2 µm, yellow-green fluorescent(505/515) |
Hepatocyte Growth Factor (HGF) | Sigma-Aldrich | H1404-5UG | recombinant, human |
JuLI stage live cell imaging system | NanoEnTek In | Automated X-Y-Z stage and fluorsent imaging Incubator-compatible (37 °C and 5% CO2) | |
Madin-Darby Canine Kidney (MDCK) cell | type II | ||
Oxygen plasma system | Femto Science | CUTE-MPR | |
Pluronic F-127 | Sigma-Aldrich | P2443-250G | |
Rhodamine B isothiocyanate–dextran | Sigma-Aldrich | R9379-100MG | 70 kDa, used to estimate spatiotemporal distribution of HGF in the microfluidic channel |
Steril hypodermic needle 18 G | KOVAX | Trim the tip of the needle and bend it 90 degrees for connecting in/out ports with volume line | |
Sticky tape | 3M/Scotch | 810D | 33 m x 19 mm |
SU-8 master molds | MicroFIT | 4” diameter, custom-made | |
sulfosuccinimidyl 6-(4’-azido-2’-nitrophenylamino)hexanoate (Sulfo-SANPAH) | Thermo Scientific | 22589 | Store at -20°C. Store protected from moisture and light. |
Sylgard 184 Elastomer Kit | Dow Corning | PDMS | |
Syringe pump | Chemyx Inc. | model fusion 720 | withdraw fluid |
Syringes | KOVAX | 1, 3, 5, 10, or 50 cc for using inlet reservoir or outlet syringe pump | |
tetramethylethylenediamine (TEMED) | Bio-Rad | 1610800 | Wear protective gloves, clothing, and eye protection. |
Ultraviolet (UV) lamp | UVP LLC | 95-0248-02 | 365 nm wavelength |