Modifications to the Transwell Migration/Invasion Assay Method That Eases Assay Performance and Improves the Accuracy

Abstract

Migration is a key property of live cells and critical for normal development, immune response, and disease processes such as cancer metastasis and inflammation. Methods to examine cell migration are especially useful and important for a wide range of biomedical research such as cancer biology, immunology, vascular biology, cell biology, and developmental biology. In vitro assays are excellent approaches to extrapolate to in vivo situations and study live cells behavior. The aim of this article is to discuss the existing methods for transwell migration/invasion studies, the problems associated with this assay, and proposed modifications to this methodological approach that makes it simple to perform and improve the assay accuracy. Results of our studies demonstrated that the count of cells that had grown on top of the membrane is important to accurately evaluate the percentage of migrated/invaded cells. The results also showed that the transparent transwell insert with 4′,6-diamidino-2-phenylindole (DAPI) stained cells is the best approach to ease the analysis of cell numbers on top of the membranes. In addition, the overlay of bright light (representing membrane pores) and DAPI images can further improve the accuracy of cell count. All these modifications in combination simplify the assay performance and improve the accuracy of the transwell migration assay method.

Introduction

Cell migration is one of the major cellular processes in development, tissue regeneration, wound healing, cancer, and immune function.^1–3 The impact of physical, chemical, and molecular aspects on cell motility is a challenge to understanding migratory cells’ behavior in vivo. In vitro assays are excellent approaches to extrapolate in vivo situations and study live cell behavior.

Thus, accurate and effective methods for examining cell migration in vitro are necessary for a vast range of research areas, including immunology,⁴ vascular biology,^3,5 regenerative medicine,⁶ and cancer biology.^7–10

A PubMed literature search showed that the scratch wound-healing assay and the Boyden chamber/transwell migration assay are the two main methods that have been used as the standard techniques to study cell migration.

The scratch wound-healing assay is where a scratch is generated on a confluent cell monolayer and the speed of wound closure and cell migration can be quantified by taking images with a regular inverted microscope at periodic time intervals. This assay estimates a cell's migration ability across a two-dimensional surface. Owing to the variable nature of this technique, there is a large variance of results causing a low reproducibility factor.¹⁰ Thus, although the wound-healing assay is simple and cost-effective, it lacks the complexity of modeling that transwell assays offer.

Alternatively, the Boyden chamber-based cell migration assay, also called filter membrane migration assay, transwell migration assay, or chemotaxis assay, is not only the more widely used method for cell migration studies, but also allows researchers to study cellular invasion as well. This method, developed in 1962 by Dr. Stephen Boyden,¹¹ studies a cell's migratory potential toward a specific chemoattractant through a three-dimensional matrix.

The Boyden chamber consists of a transwell cell culture insert that separates a singular well of a multiwell cell culture plate into top and bottom compartments. To conduct migration and invasion studies, cells are plated on top of a coated or uncoated membrane, and a chemoattractant is placed in the bottom compartment. Uncoated membranes are used for migration studies and coated membranes (e.g., with Matrigel basement matrix) for invasion studies.¹²

The transwell assay can be performed with either a transparent plastic membrane⁴ or a fluorescence blocking membrane (FluoroBlok™).¹³ Both membranes are produced by Corning, Inc., and are made of the same PET plastic material. The only difference between them is the opacity of the FluoroBlok membrane, which is due to a proprietary dye. As proven by the manufacturer's validation studies there is no impact on cellular adhesion or cell motility due to this difference.

Accordingly, two different varieties of dye are used with these membranes. Fluorescence dyes are used for FluoroBlok and histochemical dyes (cytosolic or nuclear) for the transparent membrane.^5,14,15 For the FluoroBlok membrane, it is not necessary to swab away the cells from the top of the membrane because cells stained with fluorescent dye cannot be visualized from the bottom of the membrane with the inverted microscope.¹⁶ In the case of the transparent membrane, in addition to stained migrated/invaded cells, brightly stained cells on top of the membrane may also be visualized. Therefore, in contrast to FluoroBlok, top-adhered membrane cells should be swabbed away to properly visualize and quantify the total percentage of migrated/invaded cells.¹⁷

The most critical omission researchers have made by conducting the transwell migration/invasion assay is the disregard for the number of nonmigrated/invaded cells on top of the membrane. This creates an incomplete story since several confounding factors such as cellular adhesion, survival, and proliferation regarding cells growing in the upper chamber are not considered. Counting the upper chamber is more important in long-term invasion experiments as they can last for 48–72 h. During that elapsed time drastic changes may occur in the upper chamber, which would not be reflected in the analysis of the lower chamber. Whereas some researchers utilize antimitiotics to inhibit proliferation of invading cells, many do not.^1,2,18–27 Failure to consider these factors can result in a drastic effect on accurate evaluation of total percentage of migrated/invaded cells.

Another challenge associated with transwell migration/invasion assay methods is that the fluorescent and histochemical dyes are used for analysis of FluoroBlok and transparent membranes, respectively, can slightly stain the membrane pores. This may result in the inability to distinguish membrane pores from stained cells, thus affecting cell count.

There are also technical difficulties associated with the cell count of the top of the membrane for FluoroBlok inserts. For example, to be able to count the cells that grew on top of the membranes, membranes should be precisely removed without touching either side of the membrane and carefully mounted on the slides. For transparent membranes it is not viable to count cells that had grown on top of the membrane since it is not possible to distinguish top and bottom stained cells from each other. Hence, the only way to evaluate the number of migrated/invaded cells is to swab the cells from the upper surface of the membrane. However, without knowing the number of live cells that grew on top, it is impossible to accurately evaluate the percentage of migrated/invaded cells.

All these methodological pitfalls and challenges led us to modify the transwell migration/invasion assay method to ease the performance and improve accuracy of analysis.

Materials and Methods

Isolation and Expansion of Adipose-Derived Mesenchymal Stem Cells

Adipose tissue was obtained from the Medical College of Wisconsin's Tissue Bank with complete patient anonymity (MCW IRB # PRO00017015). For isolation of adipose-derived mesenchymal stem cells (AD-MSCs) from adipocytes and other cell types, adipose tissue was digested in a 1% Collagenase (Gibco) solution for 1 h then centrifuged at 130 g for 3 min. The pellet containing AD-MSCs was washed three times in 1 × phosphate-buffered saline (PBS). These cells were then plated in 25-cm² culture flasks in Alpha-Minimum Essential Medium (Gibco) (α-MEM) supplemented with 10% fetal bovine serum (FBS) (R&D Biosystems), 1% Pen/Strep (Gibco), and 1% GlutaMAX (Gibco) and incubated at 37°C in 5% CO₂. The medium was changed every 4 days until the cells reached confluence.

Collection of U87 Conditioned Medium as a Chemoattractant

U87 conditioned medium (U87-CM) was used as the chemoattractant for the transwell migration/invasion assay since it has been demonstrated that AD-MSCs exhibit high migration and invasive properties to this conditioned medium.^28,29 To collect the conditioned medium, 24 h before the assay 3 mL of serum-free media were added to a confluent 25-cm² flask of U87 culture. U87-CM was collected in a 15 mL conical and stored at 4°C until use.

Transwell Migration and Invasion Assay

For these studies, standard 24-well transwell inserts with either a FluoroBlok PET membrane (Corning, Inc.) or a transparent PET plastic membrane (Corning, Inc.) (8-mm pore size), were used to study the migration and invasion properties of AD-MSCs (see Tables 1 and 2 for detailed protocols). The difference between the migration and invasion studies was that for invasion studies the membranes were coated with (1:50) Matrigel^® (Corning, Inc.) for 24 h before experiments. Expanded AD-MSCs at 95% confluency were harvested with combination of cell dissociation media (Gibco) and 0.05% Trypsin (Gibco) and plated on top of the transwell membrane in 300 μL of α-MEM medium supplemented with 1% GlutaMAX, 1% Pen/Strep and 10% FBS at a density of 1 × 10⁴ per membrane.

Table 1.

Example Transwell Invasion Assay with FluoroBlok Membrane

Step	Parameter	Value	Description
1	Conditioned media	24 h	Starving media for U87
2	Coat membrane	24 h	4°C overnight
3	Plate cells	300 μL	10,000 AD-MSC cells
4	Incubation time	5 h	37°C and 5% CO2
5	Starving media	900 μL	No FBS present
6	Incubation time	3 h	37°C and 5% CO2
7	Chemoattractant	600 μL	U87-CM bottom chamber
8	Incubation time	48 h	37°C and 5% CO2
9	Fixation	10 min	4% paraformaldehyde
10	DAPI staining	10 min	1:1,000 (1 μg/mL) in 1 × PBS
11	Wash	1 min	1 × PBS
12	Mounting	40 μL	ProLong Diamond Anti-fade
13	Imaging (top)	5 × /0.15	Membrane center
14	Imaging (top)	10 × /0.25	5 nonoverlapping sections
15	Imaging (bottom)	5 × /0.15	Membrane center
16	Imaging (bottom)	10 × /0.25	5 nonoverlapping sections
17	Cell count	—	ImageJ
18	Statistical analysis	—	See notes

Step Notes

1. 3 mL media, 37°C 5% CO₂.

2. Use 1:50 Matrigel.

3. Cells in upper chamber and bring to 300 μL, 600 μL complete media in bottom chamber.

4. Cells adhere.

5. Change media in top (300 μL) and bottom (600 μL) chambers to starving media.

6. Starve the cells.

7. 600 μL U87-CM in the bottom chamber.

8. Run assay.

9. Fix cells at room temperature.

10. Avoid light.

11. Rinse excess DAPI staining with 1 × PBS.

12. 20 μL on both sides being mounted.

13–16. BL and DAPI images.

17. Manual or automated cell counts with ImageJ.

18. Total invaded cells = cells on bottom of membrane.

Total number of cells = sum of top and bottom of membrane.

$Total\%invasion=\frac{Totalinvadedcells}{Totalnumberofcells}\times 100.$

DAPI, 4′,6-diamidino-2-phenylindole; FBS, fetal bovine serum; PBS, phosphate-buffered saline; U87-CM, U87 conditioned medium.

Table 2.

Example Transwell Migration Assay with Transparent Membrane

Step	Parameter	Value	Description
1	Conditioned media	24 h	Starving media for U87
2	Plate cells	300 μL	10,000 AD-MSC cells
3	Incubation time	5 h	37°C and 5% CO2
4	Starving media	900 μL	No FBS present
5	Incubation time	3 h	37°C and 5% CO2
6	Chemoattractant	600 μL	U87-CM bottom chamber
7	Incubation time	48 h	37°C and 5% CO2
8	Fixation	10 min	4% paraformaldehyde
9	DAPI staining	10 min	1:1,000 (1 μg/mL) in 1 × PBS
10	Wash	1 min	1 × PBS
11	Imaging (total)	5 × /0.15	Membrane center
12	Imaging (total)	10 × /0.25	5 nonoverlapping sections
13	Cell removal	—	Cotton swab
14	Imaging (bottom)	5 × /0.15	Membrane center
15	Imaging (bottom)	10 × /0.25	5 nonoverlapping sections
16	Cell count	—	ImageJ
17	Statistical analysis	—	See notes

Step Notes

1. 3 mL media, 37°C 5% CO₂.

2. Cells in upper chamber and bring to 300 μL, 600 μL complete media in bottom chamber.

3. Allow cells to adhere.

4. Change media in top (300 μL) and bottom (600 μL) chambers to starving media.

5. Starve the cells.

6. 600 μL U87-CM in the bottom chamber.

7. Run assay.

8. Fix cells at room temperature.

9. Avoid light.

10. Rinse excess DAPI staining with 1 × PBS.

11&12. BL and DAPI images.

13. Remove nonmigrated cells from membrane using two dry cotton swabs.

14&15. BL and DAPI images.

16. Manual or automated cell counts excluding cells on the edges.

17. Total migrated cells = cells on bottom of membrane (after top cell removal).

Total number of cells = sum of top and bottom of membrane (all cells on membrane).

$Total\%invasion=\frac{Totalmigratedcells}{Totalnumberofcells}\times 100.$

BL, bright light.

This gentle dissociation method was used to ensure the preservation of the cells’ chemoattractant receptors as well as the proteins that promote cellular invasion. The bottom chambers were filled with 600 μL of the same medium and cells were left to adhere to the membrane at 37°C and 5% CO₂. After 4 h media were gently aspirated and replaced with starving media and incubated for 3 h at 37°C and 5% CO₂. Afterward, the bottom chamber medium was replaced with U87-CM and the assay was allowed to run for 24 h for migration studies and 72 h for invasion studies at 37°C and 5% CO₂. Cells were then fixed in 4% paraformaldehyde for 15 min and stained with (1 μg/mL) 4′,6-diamidino-2-phenylindole (DAPI) (Invitrogen) for 10 min at room temperature. Chambers (top and bottom) were washed with PBS and used for imaging studies.

Membrane Imaging

For imaging the FluoroBlok and transparent membranes an inverted microscope (Axio VertA1, Zeiss) was used. Images were taken with DAPI fluorescent filter and with 5 × /0.15 and 10 × /0.25 objectives.

With regard to the FluoroBlok membrane, first, membranes were carefully cut out and mounted to a slide, bottom side down with ProLong™ Diamond Anti-fade (Invitrogen). A glass cover slip was then mounted onto the top of the membrane with additional anti-fade. Next, the cells sitting on top and bottom were imaged separately, the images were taken with 5 × (from the center of the membrane) and with 10 × (from five nonoverlapping sections) objectives. Images taken using these objectives captured ∼80% of the membrane. In addition to the DAPI images, another set of images were taken with bright light (BL) to visualize the pores of the membrane. This was done immediately after taking the DAPI image without moving the stage.

To take images of the transparent membranes, the 24-well cell culture plate with transparent transwell inserts were mounted on the microscope and images were captured with 5 × /0.15 and 10 × /0.25 objectives. Owing to the thickness of the transparent membrane and aperture of the objectives, the cells located on top and bottom of membrane were visible and all in focus and, therefore, represents the total number of cells. As in the case of the FluoroBlok membranes, an additional set of images was taken with BL to be able to distinguish the migrated cells from the membrane's pores. Next, the nonmigrated cells on top of the membrane were removed using a cotton swab and the imaging process was repeated for the bottom of the membrane (migrated/invaded cells).

Image Analysis

All images were processed using ImageJ software.³⁰ DAPI and BL images were assigned pseudocolors (green and red filters, respectively) and then overlaid to show the difference between membrane pores and cells. Cells were counted manually using the cell count plug-in. Automated cell counts were performed using the built-in function after parameter optimization for cells such as threshold, particle size, and particle circularity. During manual and automated cell counts with higher magnification (10 × ), cells on the edges were omitted. Each cell count was compared with its counterpart (top and bottom) and a total percent migration/invasion based on these values was calculated.

Results

In this study, several changes had been made to the transwell migration/invasion assay method to ease the performance and improve the accuracy of the data collection and analysis.

For the method that utilized FluoroBlok transwell inserts the following changes had been made. After fixation and DAPI staining, the membranes were manually removed from the insert, mounted on a microscope slide, then DAPI and BL images were obtained for both sides of the membrane. The images taken with BL (marked in red) represented the pores of the membrane, and DAPI images (marked in green) represented fluorescently stained nuclei of the cells.

The reason DAPI's traditional blue color was assigned a pseudo-green color was to ease the recognition of yellow stained pores after overlay of green and red. Overlaying the DAPI and BL images allowed for clear distinction between the membrane pores and the stained cells (Fig. 1a). Owing to the opacity of the FluoroBlok membrane only one side could be counted at a time (Fig. 1b). The sum of cells on both sides of the membrane was considered the total number of cells. To evaluate the total percentage of migrated/invaded cells the following formula was used:

Fig. 1.

(a) Fluorescence (white) and BL (grey) images of invaded cells taken with 10 × /0.25 objective from the bottom of the Fluoroblok PET membrane were overlaid and analyzed by ImageJ software, (b) drawing generated from the ImageJ automated cell count. BL, bright light.

Modifications to the transparent transwell insert method were made as well. As with FluoroBlok, the membranes were fixed and stained with DAPI. However, to evaluate the total number of cells, the stained cells were counted without membrane removal. Instead, cells were counted by mounting the 24-well plate with stained inserts on the stage of the inverted microscope and images were taken for DAPI and BL. Because of the thickness of the transparent membrane and aperture of the objectives the cells located on top and bottom of the membrane were visible under the same focus and thus represented the total number of cells.

Next, cells on top of the membrane were swabbed away and new DAPI and BL images were taken with the same objectives (Fig. 2a). Automated cell count was performed similarly as FluoroBlok (Fig. 2b) with an accuracy of ∼98% for both membrane types. To evaluate the percentage of migrated/invaded cells the same formula was used.

Fig. 2.

(a) Fluorescence (white) and BL (grey) images of invaded cells taken with 5 × /0.15 objective from the bottom of the transparent PET membrane were overlaid and analyzed by ImageJ software, (b) drawing generated from the ImageJ automated cell count.

Discussion

The transwell assay approach is the most frequently used method for cell migration and invasion studies. Two membrane types, transparent membranes and FluoroBlok membranes, have been used for the transwell assay. For staining and counting cells, the histochemical dyes such as crystal violet¹⁴ or hematoxylin⁵ are commonly used for transparent membranes, and fluorescent dyes such as Calcein AM¹³ are used for FluoroBlok membranes.

FluoroBlok cell culture inserts are an opaque membrane that blocks the transmission of light between 400 and 700 nm.¹⁶ Fluorescently labeled cells present in the top chamber of the insert are shielded from fluorescent microscopy. For the transparent transwell inserts it is impossible to distinguish cells labeled with histochemical dyes grown on top and bottom of the membrane since light can be transmitted through.

In most published studies, regardless of membrane type, to evaluate the percentage of migrated or invaded cells, the number of cells grown on top of the membrane was ignored and only the cells grown on the bottom of the membrane were counted. Researchers assumed that the number of cells plated on top of the membranes of all experimental transwell inserts were completely identical and ignored the possibility of several factors that can affect cell adhesion, survival, and growth of plated cells.

This may be partially true for experiments where the goal was to compare the migration or invasion of one cell type with different chemoattractants. However, this approach cannot be used for the experiments where the goal is to compare the migration/invasion ability of different cell types or modified cell types with the same chemoattractant since they may exhibit different adhesive, survival, and growth properties. In this scenario the count of cells grown on top of the membrane is an absolute necessity. Particularly it is even more important for invasion studies where experiments can last 48–72 h.

However, counting the cells on both sides of the FluoroBlok membrane is a difficult procedure since the membranes should be carefully cut and mounted on slides.

To solve the challenges associated with membrane removal and mounting for cell count on top and bottom, we tested the hypothesis that DAPI stained cells can be counted on top and bottom of transparent transwell inserts without membrane removal. To this end, images from the transparent membrane were taken using 5 × /0.15 and 10 × /0.25 objectives. The thickness of the membrane and aperture of 5 × and 10 × objectives allowed for visualization of cells growing on top and bottom of the membrane under the same focus.

Thus, captured images represented the total number of cells. Despite taking images using both objectives, it is not necessary to do so. As a matter of fact, 5 × is usually sufficient but 10 × should be used in experiments with high confluency cultures to distinguish individual cells more easily. This drastically simplified the transwell assay method since it eliminates the need for physical manipulation of membrane removal and images can be captured by mounting the 24-well culture plate with transparent transwell inserts on microscope stage.

Another improvement to the transwell assay method that we suggested with this study is the capture of additional images with BL and their overlay with DAPI images. This technique can significantly improve accuracy of cell count since some fluorescence dyes can partially stain the pores of the membranes making counting difficult. We also suggested that DAPI is very reliable dye for such studies because of simplicity of staining procedure.

Conclusion

To simplify the assay performance and improve the accuracy of the transwell migration assay method, the following suggestions have been made: (1) count of the cells that had grown of the top of the membranes is important to accurately evaluate the migration and invasion of cells, (2) the transparent transwell insert with DAPI stained cells is the best approach to ease the evaluation of cell numbers on top of membranes, and (3) the overlay of BL and DAPI images can further improve the accuracy of cell count.

Abbreviations Used

AD-MSC

adipose-derived mesenchymal stem cell

DAPI

4′,6-diamidino-2-phenylindole

FBS

fetal bovine serum

PBS

phosphate-buffered saline

U87-CM

U87 conditioned medium

α-MEM

Alpha-Minimum Essential Medium (Gibco)

Authors’ Contributions

A.R.A. contributed to study motivation, experimental design, and writing the article. H.M.S. provided technical support and edited the article.

Disclosure Statement

No competing financial interests exist.

Funding Information

This study was supported by the NIH (National Cancer Institute) #1R43CA221490-01A1 and Cell Reprogramming & Therapeutics LLC.