1Department of Biomedical Engineering, University of Wisconsin-Madison, 2Materials Science Program, University of Wisconsin-Madison, 3Department of Neurological Surgery, University of Wisconsin-Madison, 4Carbone Comprehensive Cancer Center and Center for Stem Cell and Regenerative Medicine, University of Wisconsin-Madison
Huang, Y., Agrawal, B., Clark, P. A., Williams, J. C., Kuo, J. S. Evaluation of Cancer Stem Cell Migration Using Compartmentalizing Microfluidic Devices and Live Cell Imaging. J. Vis. Exp. (58), e3297, doi:10.3791/3297 (2011).
In the last 40 years, the United States invested over 200 billion dollars on cancer research, resulting in only a 5% decrease in death rate. A major obstacle for improving patient outcomes is the poor understanding of mechanisms underlying cellular migration associated with aggressive cancer cell invasion, metastasis and therapeutic resistance1. Glioblastoma Multiforme (GBM), the most prevalent primary malignant adult brain tumor2, exemplifies this difficulty. Despite standard surgery, radiation and chemotherapies, patient median survival is only fifteen months, due to aggressive GBM infiltration into adjacent brain and rapid cancer recurrence2. The interactions of aberrant cell migratory mechanisms and the tumor microenvironment likely differentiate cancer from normal cells3. Therefore, improving therapeutic approaches for GBM require a better understanding of cancer cell migration mechanisms. Recent work suggests that a small subpopulation of cells within GBM, the brain tumor stem cell (BTSC), may be responsible for therapeutic resistance and recurrence. Mechanisms underlying BTSC migratory capacity are only starting to be characterized1,4.
Due to a limitation in visual inspection and geometrical manipulation, conventional migration assays5 are restricted to quantifying overall cell populations. In contrast, microfluidic devices permit single cell analysis because of compatibility with modern microscopy and control over micro-environment6-9.
We present a method for detailed characterization of BTSC migration using compartmentalizing microfluidic devices. These PDMS-made devices cast the tissue culture environment into three connected compartments: seeding chamber, receiving chamber and bridging microchannels. We tailored the device such that both chambers hold sufficient media to support viable BTSC for 4-5 days without media exchange. Highly mobile BTSCs initially introduced into the seeding chamber are isolated after migration though bridging microchannels to the parallel receiving chamber. This migration simulates cancer cellular spread through the interstitial spaces of the brain. The phase live images of cell morphology during migration are recorded over several days. Highly migratory BTSC can therefore be isolated, recultured, and analyzed further.
Compartmentalizing microfluidics can be a versatile platform to study the migratory behavior of BTSCs and other cancer stem cells. By combining gradient generators, fluid handling, micro-electrodes and other microfluidic modules, these devices can also be used for drug screening and disease diagnosis6. Isolation of an aggressive subpopulation of migratory cells will enable studies of underlying molecular mechanisms.
1. BTSC's cell dissociation
BTSCs are derived from pre-existing cultures grown in serum-free stem cell medium as neurospheres. culture of which is previously described10, 11.
2. Fabrication of bi-layered su-8 master and molding of PDMS stamp (see figure 1)
We use optical lithography and soft lithography to fabricate su-8 master and PDMS stamp, which are essential for assembly of the microfluidic devices. Slightly different than the standard procedure12, our su-8 master is composed of two layers. The 3-μm-tall microchannels are casted in the first/bottom layer earlier in the process, while the 250-μm-tall seeding and receiving chambers are in the second/upper layer. In order for correct connection between two culture compartments (microchannels and chambers), the two layers have to be accurately aligned in position. However, thickness and opacity of the second layer are large enough to block the fiducial/optical aligner from accessing the bottom features. Here we design the masks with fiducial markers being remotely positioned. Thus, these markers in the first layer can be selectively shielded while spinning the second layer. As a result, both layers of features are made in one master and ready for molding in bas-relief of the PDMS stamp.
3. Assembly of microfluidic devices and cell culturing (figure 2)
In order to culture cells, feature-engraved PDMS stamp is attached to a glass coverslip to form enclosed channels. Inlets and outlets are created for loading the culture/media. Meanwhile, cleanup and other procedures to the glass substrate and PDMS stamp are necessary to ensure the cell-compatibility.
4. Long-term time-lapse imaging of BTSC migration with BioStation IM (Nikon Instruments Inc, Melville, USA).
Combining camera, software and incubator all in one box, this microscopic system makes it possible to image cell culture with no disturbance for days. In addition, its unique motor design moves the objective lens and keeps sample stage stationary in the point-visiting feature. This makes it practical to image parallel experiments and track cell locomotion in high-throughput assay.
5. Representative results:
Examples of visual inspection and characterization of cancer cell migration by using a compartmentalizing microfluidic device are illustrated in Figure 3 and Figure 4. The cell-line is BTSC. With our current setup, phase images of cell culture can be continuously recorded for as long as 5 days (Figure 3) and as frequent as every 2 seconds (Figure 4). Beyond 5 days, culture media need to be replaced for nutrient replenishment and waste removed. Our long-term time-lapse identifies a revolving sequence of six cell stages during its migration based on the morphological changes. As illustrated in Figure 3, we describe them as (i) pre-migration, (ii) initiation, (iii) path-exploration, (iv) cruising, (v) destination-exloration and (vi) post-migration. In the receiving chamber, spindle-shaped cells stay in stage (i) as in bulk tumors that are able to grow, divide and gradually migrate. As they approach the microchannel entrance, a few cells proceed to stage (ii) when they expand and generate adhesive protrusion. Only one of them is able to occupy the entrance and proceed to stage (iii), which can explore the migration direction. Once the cell determines the migration direction, it proceeds to stage (iv) to cruise through the microchannel in a steady and high speed, which is carried through the entire microchannel. At the end of microchannel, cell proceeds to stage (v) to explore the open space of receiving chamber, and then to stage (vi). In microchannel, migrating power is mostly generated by blebbing activity as illustrated in figure 4, similar to that of amoeboid cell. Cell blebbing and membrane deformation are fully recorded here.
Figure 1. Schematics of fabrication of microfluidic device. The entire process is carried on a 3-inch silicon handle wafer through a modified technique of soft lithography. 3-μm-tall microchannels and alignment markers are made as in the first layer. Then 250-μm-thick su-8 is spin-coated, while the alignment markers are covered with scotch tape, such that they can later be accessible to mask aligner without sight-blocking by thick photoresist. Thus, chamber features can be made as the second layer and accurately aligned to the first layer. PDMS stamp is eventually molded off the resulting master.
Figure 2. Schematics of device assembly and cell seeding process. Enclosed by PDMS and glass coverslip, the 3D culturing space is composed of chambers, reservoirs and microchannels. The seeding chamber is filled with cell culture, while the receiving chamber is initially filled with cell-free media. The microchannels that connect between them provide cell migrating trail from seeding side to receiving side.
Figure 3. BTSC migration through microchannel. Upper: snapshots of time-lapse show a single cell (highlighted in green) migrates over 400 μm from the seeding side to the receiving side. The entire journey takes about 2 days, involving a sequence of cell morphological changes. The scale bar is 40 μm. Lower: a cartoon representation of six migration stages. Pre-migration (i): In seeding chamber, spindle-shaped and dividable cells gradually migrate along each other. The cells near microchannel entrance race with each other by polarizing cell body and exerting lamellipodia protrusion. Initiation (ii): The migratory cell with the highest capacity occupies the channel entrance while its competitors will retreat. Path-exploration (iii): The migratory cell changes to an amoeboid mode with little polarity. This migration mode is extremely motile and able to explore path in all directions by blebbing small membrane protrusions. Cruising (iv): Once the migration path is determined, the cell transforms into an adapted amoeboid mode by maintaining an additional large protrusion heading forward. This way, cell upholds a high motility as well as a steady direction and speed. Destination-exploration (v): Upon path-end, the cell slows down and develops filopodias to explore the receiving chamber for any invasion target. Post-migration (vi): After entering the receiving chamber, cell turns into star-shaped and retains high motility but no determined direction.
Figure 4. Snapshots of time-lapse show the cycle of blebbing activity and membrane deformation of a BTSC: (A-B). initiation; (B-D). Expansion; (D-F). Retraction. The arrows and curve-lines represent the direction and location of membrane deformation, respectively. The snapshots are collected every 8 seconds.
The microfluidic device and image-recording technique presented here enables a visual characterization of cellular morphology during migration. Compared to existing conventional methods, the microfluidic platform features advantages of cost-effectiveness, high throughput and design flexibility. The presented microscopic visualization system permits study and recording of long-term live cell migration. The lens-motorized feature makes it possible to track multiple migration paths in a high image-resolution without disturbing the sample.
During assembly of the device, the PDMS stamp is non-permanently bonded to the glass coverslip for the convenience of PDL coating (step 3.4). It is critical to achieve a good seal for the success of this protocol. Contamination introduced from the device fabrication, such as air bubble in the stamp or dust/debris on the substrate, can jeopardize the bonding and result in fluid leaking. Therefore, treating the device in a dust-free manner is important to prevent device failure. Prior to the PDL-coating, the glass coverslips are nitric acid treated and the PDL solution is centrifuged to eliminate dust/debris. We find Scotch tape, alcohol rinse, and water bath very helpful in removing dust/debris from the PDMS stamp, if a clean room is not accessible. After the PDMS stamp making contact with the glass coverslip, tapping the stamp gently with tweezers facilitates the seal.
BTSCs can be enriched from human operating room specimens via sphere culture in stem cell medium, similar to culture of normal neural stem cells10. BTSCs enriched in this manner demonstrate stem cell-like properties, such as expression of stem cell markers (CD133, nestin), differentiation to multiple neural lineages (glial and neuronal), and tumor initiation in an orthotopic immunodeficient mouse model with as few as 100 cells18. Although some debate exists about purification and maintenance of BTSCs1, 19-21, sphere-grown BTSCs better maintain the phenotype and genotype of the parental tumor22-24.
The compartmentalized space mimics the physiological environment for cell migration. In our device, BTSC exhibit a strong capacity of migration through a size-constrained space by regulating cellular morphology. Our characterization of the cell morphological changes indicates mode-transformations in a six stage sequence that may model a possible new detailed description of BTSC invasion of adjacent brain. In early stages, the cells gain a significant amount of polarity and generate adhesive protrusions to effectively anchor themselves inside the microchannel. Once the cell occupancy in the microchannel is established, it converts to a cruising mode and maintains this mode for the entire journey through a microchannel. At this stage, BTSC maintain high motility and consistent direction but conserve energy. Interestingly, it appears that one microchannel is restricted to one cruising cell at a time, such that when a single cell proceeds to this stage, other cells retreat from the microchannel. This mechanism likely ensures effectiveness of tumor spreading and prevents overconsumption of nutrients in microchannel. After completing migration across the microchannel, a cell regains its propensity for multidirectional protrusions and further exploration for invasion targets or new migration paths. The compartmentalizing microfluidic device offers a novel in vitro means of creating a microenvironment for studying BTSC infiltration of brain parenchyma.
This method can easily be adapted to the study of cancer stem cell (CSC) and migratory cell lines derived from other tumor types. Culturing cells in the microfluidic device can be performed in any modern biology laboratory that is equipped with an incubator and tissue culture expertise. Equipped with a su-8 master, PDMS casting and device assembly are feasible with some fundamental training in soft lithography. In addition, this platform may also be extended for other applications with integrating other functional modules such as gradient mixer, surface patterning, fluid control and microelectrode.
No conflicts of interest declared.
PAC is partially supported by a NIH T32 grant to University of Wisconsin Stem Cell Training Program (P. A. Clark). JSK was partially supported by the HEADRUSH Brain Tumor Research Professorship, Roger Loff Memorial GBM Research Fund, and the UW Foundation/ Neurosurgery Brain Tumor Research and Education Fund. JCW and YH are partially supported by NIH grant NIBIB 1R01EB009103-01.
|Dulbecco’s modified eagle’s medium (DMEM), high glucose||GIBCO, by Life Technologies||11965||Brain tumor stem cell (BTSC) culture supplies|
|Ham’s F12||GIBCO, by Life Technologies||31765||Brain tumor stem cell (BTSC) culture supplies|
|B27 supplement minus vitamin A||GIBCO, by Life Technologies||12587-010||Brain tumor stem cell (BTSC) culture supplies|
|Antibiotic-Antimycotic (PSA)||GIBCO, by Life Technologies||15240||Brain tumor stem cell (BTSC) culture supplies|
|Epidermal Growth Factor (EGF), human recombinant||GIBCO, by Life Technologies||PHG0313||Brain tumor stem cell (BTSC) culture supplies|
|basic Fibroblast Growth Factor (bFGF) , human recombinant||GIBCO, by Life Technologies||PHG0021||Brain tumor stem cell (BTSC) culture supplies|
|Heparin sodium salt, from porcine intestinal mucosa||Sigma-Aldrich||H1027-250KU||Brain tumor stem cell (BTSC) culture supplies|
|Laminin (natural mouse)||GIBCO, by Life Technologies||23017-015||Brain tumor stem cell (BTSC) culture supplies|
|Accutase||EMD Millipore||SCR005||Brain tumor stem cell (BTSC) culture supplies|
|Stem cell medium||
|Basic Fibroblast Growth Factor (bFGF)/ Heparin||
|Epidermal Growth Factor (EGF)||
|su-8 photoresist||MicroChem Corp.|
|Silicon handle wafer||WRS Materials||3P01-5SSP-INV|
|PDMS Sylgard 184||Dow Corning|
|laminin||BD Biosciences||50 μg/ml in PBS buffer for the final concentration|
|Biostation IM||Nikon Instruments|