We describe an approach to detect and capture invasive cell subpopulations in real-time. The experimental design uses Real-Time Cellular Analysis by monitoring changes in the electric impedance of cells. Invasive cancer, immune, endothelial or stromal cells in complex tissues can be captured, and the impact of co-cultures can be assessed.
Invasion and metastatic spread of cancer cells are the major cause of death from cancer. Assays developed early on to measure the invasive potential of cancer cell populations typically generate a single endpoint measurement that does not distinguish between cancer cell subpopulations with different invasive potential. Also, the tumor microenvironment consists of different resident stromal and immune cells that alter and participate in the invasive behavior of cancer cells. Invasion into tissues also plays a role in immune cell subpopulations fending off microorganisms or eliminating diseased cells from the parenchyma and endothelial cells during tissue remodeling and angiogenesis. Real-Time Cellular Analysis (RTCA) that utilizes impedance biosensors to monitor cell invasion was a major step forward beyond endpoint measurement of invasion: this provides continuous measurements over time and thus can reveal differences in invasion rates that are lost in the endpoint assay. Using current RTCA technology, we expanded dual-chamber arrays by adding a further chamber that can contain stromal and/or immune cells and allows measuring the rate of invasion under the influence of secreted factors from co-cultured stromal or immune cells over time. Beyond this, the unique design allows for detaching chambers at any time and isolating of the most invasive cancer cell, or other cell subpopulations that are present in heterogeneous mixes of tumor isolates tested. These most invasive cancer cells and other cell subpopulations drive malignant progression to metastatic disease, and their molecular characteristics are important for in-depth mechanistic studies, the development of diagnostic probes for their detection, and the assessment of vulnerabilities. Thus, the inclusion of small- or large-molecule drugs can be used to test the potential of therapies that target cancer and/or stromal cell subpopulations with the goal of inhibiting (e.g., cancer cells) or enhancing (e.g., immune cells) invasive behavior.
Cell invasion is an important process that allows cells to cross basement membrane barriers in response to environmental cues provided by stromal cells. It is a crucial step during several stages of development for immune responses, wound healing, tissue repair, and malignancies that can progress from local lesions to invasive and metastatic cancers1. Assays developed early on to measure the invasive potential of cell populations typically generate a single endpoint measurement or require pre-labeling of invasive cells2. The integration of microelectronics and microfluidics techniques is now developed to detect different aspects of cell biology such as viability, movement, and attachment using the electric impedance of live cells on microelectrodes3,4. Impedance measurement allows for a label-free, non-invasive and quantitative assessment of cell status3. Here we describe a three-chambered array based on the design of the Real-Time Cellular Analysis (RTCA) system that was developed by Abassi et al.5. The three-chambered array allows for the assessment of co-cultured cells on cellular invasion and recovery of invasive cells for additional analyses or expansion.
In the cell analyzer system, cells invade through an extracellular matrix coated onto a porous membrane and reach an interdigitated electrode array positioned on the opposite side of the barrier. As the invasive cells continue to attach and occupy this electrode array over time, the electrical impedance changes in parallel. The current system comprises a cell invasion and migration (CIM) 16-well plate with two chambers. The RTCA-DP (dual purpose) (called dual purpose cell analyzer henceforth) instrument contains sensors for impedance measurement and integrated software to analyze and process the impedance data. Impedance values at baseline depend on the ionic strength of media in the wells and are changed as cells attach to the electrodes. The impedance changes depend on the number of cells, their morphology, and the extent to which cells attach to the electrodes. A measurement of the wells with media before the cells are added is considered as the background signal. The background is subtracted from impedance measurements after reaching equilibrium with cells attaching and spreading onto the electrodes. A unitless parameter of the status of the cells on an electrode termed Cell Index (CI) is calculated as follows: CI = (impedance after equilibrium – impedance in the absence of cells) / nominal impedance value6. When migration rates of different cell lines are compared, the Delta CI can be used to compare cell status regardless of the difference in attachment that is represented in the first few measurements.
The newly designed three-chambered array builds on the existing design and uses the top chamber from the dual purpose cell analyzer system that contains the electrodes. The modified middle and bottom chambers are adapted to fit the assembly into the dual purpose cell analyzer for impedance measurement and analysis using the integrated software. The two major advances that the new design provides over the existing dual-chamber CIM-plate (called cell analyzer plate henceforth) are: i) the ability to recover, and then analyze invasive cell subpopulations that are present in heterogeneous cell mixes and ii) the option to assess the impact of secreted factors from co-cultured stromal or immune cells on cell invasion (Figure 1).
This technology can be useful in studying the subpopulations of cells with different invasive capacities. That includes (a) invasive cancer cells that invade surrounding tissues or blood and lymphatic vessels or extravasate at metastatic seeding sites in distant organs, (b) cells from the immune system that invade tissues to tackle pathogens or diseased cells, (c) endothelial cells that invade tissues to form new blood vessels during tissue reorganization or wound healing, as well as (d) stromal cells from the tumor microenvironment that support and invade along with cancer cells. The approach allows the inclusion of stromal cross-talk that can modulate cell motility and invasion. The feasibility studies shown here use this modified array focused on cancer cell invasion and the interaction with the stroma as a model system, including endothelial invasion in response to differential signals from cancer cells. The approach can be extrapolated to isolate cancer cells and other cell types such as subpopulations of immune cells, fibroblasts, or endothelial cells. We tested invasive and non-invasive established breast cancer cell lines as a proof of principle. We also used cells from patient-derived xenograft (PDX) invasion in response to immune cells from human bone marrow to show feasibility for future use also in clinical diagnostic settings. PDX are patient tumor tissues that are implanted in immunocompromised or humanized mice model to allow for studying of growth, progression, and treatment options for the original patient7,8.
The study was reviewed and considered as "exempt" by the Institutional Review Board of Georgetown University (IRB # 2002-022). Freshly harvested bone marrow tissues were collected from discarded healthy human bone marrow collection filters that had been de-identified.
1. New chamber design (Figure 2)
2. Cell culture (MDA-MB-231, DCIS, DCIS-Δ4, J2-fibroblasts)
3. Patient-derived xenograft dissociation
4. Bone marrow cell extraction
5. Cell seeding and assembly
6. Background and impedance measurement
7. Detachment and cell collection
8. 3D cell propagation and retrieval
NOTE: Due to the small number of cells collected, seed the cells in 3D using an extracellular matrix (ECM) to enhance viability. That said, 2D culture is also an option at this point, especially if the cells used are from established cell lines.
Using the newly designed three-chambered array, invasion of the cells was tested in the presence or absence of stromal cells such as fibroblasts. MDA-MB-231 cell invasion was enhanced when irradiated Swiss 3T3 fibroblasts (J2 strain) were seeded in the bottom chamber, allowing for the exchange of factors between the two cell lines. Interestingly, MDA-MB-231 invasion increased when 3T3-J2 cells were doubled in number (Figure 3A). On the other hand, the invasion rate of an invasive clone of MCFDCIS cells (DCIS-Δ4)9 appears to be inhibited by the cross-talk with 3T3-J2 cells (Figure 3B). This data shows the useful application of the three-chambered array to measure varying effects of the stroma, in this case, fibroblasts, on cell invasion.
Next, to monitor the change in endothelial cells motility and invasion in response to signals from either invasive (MDA-MB-231)10,11 or non-invasive (DCIS)12 cancer cells, human umbilical vein endothelial cells (HUVECs) that represent endothelial cells lining the walls of blood vessels were used. HUVECs were more invasive in response to factors secreted by MDA-MB-231 cells, unlike those secreted by DCIS cells (Figure 4). This is consistent with the ability of invasive tumors to recruit endothelial cells for blood vessel formation and later dissemination into the circulation.
The data above demonstrate the ability of the cell analyzer system to monitor different invasion rates of cell lines; when the invasion starts, progresses and plateaus. This allows the user to choose the time point of interest, for example, after the first 2-3 h of invasion, to capture the pioneer cells that initiate invasion and are distinct from the follower cells that invade thereafter by collective invasion.
While the above data demonstrates a solid proof of principle for the use of the three-chambered array to observe invasion in a co-culture setting, we wanted to test the potential usability of this array in clinical and diagnostic settings. For that, the invasion of cell suspensions from patient-derived xenografts (PDX)8 co-cultured with immune cells from human bone marrow samples were monitored. Total human bone marrow immune cells (BM) were seeded onto the bottom chamber with or without serum. PDX cells invasion from the top chamber increased in response to co-cultured BM immune cells (Figure 5). Interestingly, the presence of 2% serum in the bottom chamber with the BM cells was essential for PDX invasion.
Figure 1: Workflow of the cell invasion monitoring and cell collection system. (A) Stroma cells are added to the bottom chamber that is mounted on a middle chamber, which serves as a cell barrier that is permeable for soluble factors only. (B) The top chamber with impedance biosensors receives the cell line to be monitored. Real-time invasion is recorded until a user-defined timepoint for cell collection is reached. (C) The dismantled top chamber is inverted, and cells are harvested using a cell lifter. Please click here to view a larger version of this figure.
Figure 2: Images of the array chambers and modifications. (A) The three chambers used to build the array. No modification was made on the top chamber harboring the electrodes. (B) From the middle chamber wells, a height of 2 mm has been shaved off and a membrane attached to the open bottom; longitudinal slits (1.5 mm x 5.6 mm) were added to each side. (C) Lower chamber (72 mm x 18 mm) fabricated to replicate the 16-well design; wells are 4.8 mm deep, and 4.75 mm in diameter, triangle ridges (1.5 mm horizontal x 1.4 mm vertical) are added along the sides to click into the middle chamber slits. Please click here to view a larger version of this figure.
Figure 3: The effect of co-cultured fibroblasts on cancer cell invasion. (A) Real-Time Cellular Analysis of MDA-MB-231 cell invasion, alone or in co-culture with 3T3-J2 fibroblasts (bottom chamber). 3T3-J2 fibroblasts were seeded at either 30,000 or 60,000 cells per well. (B) Real-Time Cellular Analysis of DCIS-Δ4 cell invasion, alone or in co-culture with 3T3-J2 fibroblasts (bottom chamber). The solid circles represent the mean; the thin dotted lines represent the standard deviation. Please click here to view a larger version of this figure.
Figure 4: The effect of cancer cells on endothelial cell invasion. Real-Time Cellular Analysis of HUVEC invasion, alone, in co-culture with the invasive MDA-MB-231 cells (bottom chamber) or the non-invasive DCIS cells (bottom chamber). The solid circles represent the mean; the thin dotted lines represent the standard deviation. Delta cell index normalized to the impedance at time = 1 h. Please click here to view a larger version of this figure.
Figure 5: Cell invasion of patient-derived xenografts (PDX) co-cultured with bone marrow immune cells. Single cells disintegrated from PDX (top chamber) were co-cultured with human bone marrow cells (bottom chamber), and their invasion was monitored over time in the presence or absence of serum. The solid circles represent the mean; the thin dotted lines represent the standard deviation. Delta cell index normalized to the impedance at time = 1 h and 36 min. Please click here to view a larger version of this figure.
Media | Constituents | Concentration/proportion |
MDA-MD-231 media | DMEM | |
Fetal Bovine Serum (FBS) | 10% | |
J2 Fibroblasts media | DMEM | |
Fetal Bovine Serum (FBS) | 10% | |
DCIS media | DMEM F12 | |
Horse serum (HS) | 5% | |
Epidermal growth factor (EGF) | 20 ng/mL | |
Insulin | 10 μg/mL | |
Hydrocortisone | 0.5 μg/mL | |
Cholera toxin | 100 ng/mL | |
PDX media | DMEM F12 | |
Fetal Bovine Serum (FBS) | 2% | |
HEPES | 1 M | |
Insulin Transferrin Selenium Ethanolamine (ITS) | 10 μg/mL | |
Hydrocortisone | 0.5 μg/mL | |
Bovine serum albumin (BSA) | 1 mg/mL |
Table 1: Cell culture media composition. The table lists the compositions of MDA-MD-231 media, J2 Fibroblasts media, DCIS media, and PDX media.
We have modified the design of a dual-chambered array to include a third chamber for monitoring cell invasion in real-time in the presence of stromal cells. We have observed distinct effects of co-cultured fibroblasts on invasive and non-invasive cancer cells indicating that the array can be used to distinguish between cancer cell subpopulations that respond differently to factors produced by co-cultured stromal cells. The array was also used to monitor endothelial cell invasion into stromal tissues, a critical step during blood vessels sprouting toward an angiogenic stimulus in the presence of cancer cells of varying invasive potential. These experiments show the versatile use of the array for cancer cell or other cell isolation.
It is recommended to optimize the number of cells to be added to each chamber. From our experience and the manufacturer's recommendation, 30,000-50,000 cells are optimal. Since two cell types can be co-cultured in this array, one of which may be a primary culture of stroma cells, it is suggested to monitor the effects of different media on cells used to maintain viability. We found that starvation of cells to be monitored reduces the variation between replicates. If serum-free growth conditions are harmful to the cells, low serum and shorter starvation times can be used. The addition of a low serum amount (1%-2%) may be critical for the survival of stromal cells (primary cells) in the bottom chamber, yet make sure to include the proper control conditions for data interpretation (i.e., 2% serum with and without stromal cells). If invasion rates are low, more wells per the experimental condition can increase the number of viable cells collected for downstream analyses. When analyzing patient biopsies, the disintegration of tissue into single cells is crucial before testing on the cell analyzer plate. Optimizing disintegration conditions to maintain cell viability is an important step before performing the invasion analysis. It is also possible to study additional aspects of invasion, such as sensitivity to drug treatment or the effect of extracellular basement matrix (ECM) components on invasion rate. ECM is an essential component of the microenvironment and has been reported to play a major role during cell invasion13. Several published studies have used ECM to coat the top chamber before invasive cells are added to monitor their interaction with various ECM components14,15. The three-chambered array is a useful tool to study co-culture interactions between invasive and stroma cells. While the proposed design guarantees a cell collection specific to the invasive cells without any stromal cells, this setup may not be optimal if the cross-talk between cells requires the physical interaction of the different cell types. Additionally, non-adherent cells that grow in suspension may not be collected in the proposed methods here (scrapping), yet a different approach in which media in the disassembled chambers containing the non-adherent cells may be collected to harvest the non-adherent cells.
While this array can be utilized for multiple areas of research that monitor cell invasion in the presence of a stromal component, here we focused on cancer cell invasion and how this approach can uncover malignant cancer cell subpopulations present in heterogeneous biological samples. The new capacity of the three-chambered array used here provides a functional assay to isolate invasive subpopulations from heterogeneous cancer cell mixtures that contain more and less invasive cancer cells. Analysis of invasive and outcome-relevant cancer cell subpopulations is essential for appropriate mechanistic studies and molecular insights not obtainable or biased by the analysis of a mixed cell population. The co-culture chamber for stromal cells provides insights into cell-cell cross-talk in the tumor microenvironment during progression to invasive disease.
As a step toward application to tissue samples, e.g., tumor biopsies from patients, we used cancer cells that were isolated from patient-derived tumor xenografts (PDXs) and tested the impact of human bone marrow cells on the invasion of tumor cells. We were able to collect the invasive cells present in the PDXs for downstream analysis, i.e., RNA-seq. Assaying PDXs is the initial step toward analyzing cell subpopulations present in the heterogeneous mix of cancer cells in tumor biopsies obtained from patients. The ultimate goal will be to use such tumor biopsies and isolate subpopulations of cancer cells that invade and thus drive poor outcomes due to their potential for metastatic spread. Identification of the molecular features and testing the sensitivity of these invasive subpopulations to drug treatment are future applications.
The authors have nothing to disclose.
We would like to thank Dr. Alana Welms, Huntsman Cancer Institute, University of Utah, for providing us with the patient-derived xenografts (HCI-010). This work was supported by NIH grants R01CA205632, R21CA226542, and in part, by a grant from Agilent Technologies.
0.05% Trypsin-EDTA | Thermofisher | 25300-054 | |
Adhesive | Norland Optical Adhesive | NOA63 | |
Bovine serum albumin (BSA) | Sigma | A9418 | |
Cell lifter | Sarstedt | 83.1832 | |
Cholera Toxin from Vibrio cholerae | Thermofisher | 12585-014 | |
CIM-plate | Agilent | 5665817001 | Cell analyzer plate |
Collagenase from Clostridium histolyticum | Sigma | C0130 | |
Dispase | StemCell | 7913 | |
DMEM | Thermofisher | 11995-065 | |
DMEM-F12 | Thermofisher | 11875-093 | |
Fetal Bovine Serum (FBS), Heat Inactivated | Omega Scientific | FB-12 | |
HEPES | Thermofisher | 15630106 | |
Horse serum (HS) | Gibco | 16050-122 | |
Human EGF | Peprotech | AF-100-15 | |
Human umbilical Vein endothelail cells (HUVEC) | LONZA (RRID:CVCL_2959) | C-2517A | |
HUVEC media | LONZA | CC-3162 | |
Hydrocortisone | Sigma | H4001 | |
Insulin Transferrin Selenium Ethanolamine (ITSX) (100x) | Thermofisher | 51500056 | |
Insulin, Human Recombinant, Zinc Solution | Sigma | C8052 | |
J2 Fibroblasts | Stemcell (RRID:CVCL_W667) | 100-0353 | |
LymphoPrep | Stemcell | 7851 | Density gradient medium for the isolation of mononuclear cells |
Matrigel | Corning | 354230 | Basement membrane matrix |
MCFDCIS.com cells ( DCIS) | RRID:CVCL_5552 | ||
MDA-MB-231 cells | RRID:CVCL_0062 | ||
Milling machine | Bridgeport Series 1 Vertical | ||
Phosphate-buffered saline (1x) | Thermofisher | 10010049 | |
Polyethersulfone (PES) membrane | Sterlitech | PCTF029030 | |
RBC lysis solution | Stemcell | 7800 | |
RNeasy Micro Kit | Qiagen | 74004 | |
RTCA DP analyzer | Agilent | 3X16 | Dual purpose cell analyzer |
Trypsin | Sigma | T4799 |