This report describes a microfluidic chip-based method to set up a single cell culture experiment in which high-efficiency pairing and microscopic analysis of multiple single cells can be achieved.
Cite this ArticleCopy Citation | Download Citations | Reprints and Permissions
He, C. K., Chen, Y. W., Wang, S. H., Hsu, C. H. Establishing Single-Cell Based Co-Cultures in a Deterministic Manner with a Microfluidic Chip. J. Vis. Exp. (151), e60202, doi:10.3791/60202 (2019).
Translate text to:
Cell co-culture assays have been widely used for studying cell-cell interactions between different cell types to better understand the biology of diseases including cancer. However, it is challenging to clarify the complex mechanism of intercellular interactions in highly heterogeneous cell populations using conventional co-culture systems because the heterogeneity of the cell subpopulation is obscured by the average values; the conventional co-culture systems can only be used to describe the population signal, but are incapable of tracking individual cells behavior. Furthermore, conventional single-cell experimental methods have low efficiency in cell manipulation because of the Poisson distribution. Microfabricated devices are an emerging technology for single-cell studies because they can accurately manipulate single cells at high-throughput and can reduce sample and reagent consumption. Here, we describe the concept and application of a microfluidic chip for multiple single-cell co-cultures. The chip can efficiently capture multiple types of single cells in a culture chamber (~46%) and has a sufficient culture space useful to study the cells' behavior (e.g., migration, proliferation, etc.) under cell-cell interaction at the single-cell level. Lymphatic endothelial cells and oral squamous cell carcinoma were used to perform a single-cell co-culture experiment on the microfluidic platform for live multiple single-cell interaction studies.
Efficient capture of different types of single cells and providing sufficient culture space are needed for single cell co-culture experiments of multiple types of single cells1. Limiting dilution is the most commonly used method to prepare the single cells for such experiments, due to the low cost of equipment required. However, due to the Poisson distribution limitation, the maximum single cell acquisition probability is only 37%, making the experimental operation laborious and time-consuming2. In contrast, using fluorescence activated cell sorting (FACS) can overcome the Poisson distribution limitation to high-efficiently prepare single cells3. However, FACS may not be accessible to some laboratories due to expensive instrumentation and maintenance cost. Microfabricated devices have been recently developed for single cell trapping4, single cell pairing5, and single cell culture applications. These devices are advantageous based on their ability to accurately manipulate single cells6, perform high-throughput experiments, or reduce sample and reagent consumption. However, performing single-cell co-culture experiments with multiple cell types with the current microfluidic devices is still challenging due the limitation of Poisson distribution1,7,8, or inability of the devices to capture more than two types of single cells4,5,6,9,10.
For example, Yoon et al. reported a microfluidic device for cell-cell interaction study11. This device uses the probabilistic method to pair cells in one chamber. However, it can only achieve the pairing of two different cell types due to geometric restrictions in the device structure. Another report from Lee et al. demonstrated a deterministic method to capture and pair single cells12. This device is increases pairing efficiency by the deterministic method but it is limited by the prolonged operation time required to pair cells. Specifically, the second cell capture can only be performed after the first captured cell is attached to the surface after 24 h. Zhao et al. reported a droplet-based microfluidic device to capture two types of a single cell13. We can found that the droplet-based microfluidic device is still limited to the Poisson distribution and can only be used on non-attached cells, and it is not possible to change the culture solution during the cultivation process.
Previously, we have developed a microfluidic "hydrodynamic shuttling chip" that utilizes deterministic hydrodynamic forces to capture multiple types of single-cell into the culture chamber and can subsequently perform cell co-culture experiment to analyze individual cell migration behavior under cell-cell interactions14. The hydrodynamic shuttling chip comprises an arrayed sets of units that each contains a serpentine by-pass channel, a capture-site, and a culture chamber. By using the difference in flow resistance between the serpentine by-pass channel and the culture chamber, and a specially designed operation procedure, different types of single cells can be repeatedly captured into the culture chamber. Notably, the ample space of the culture chamber can not only prevent the cell from being flushed during cell capture out but also provide sufficient space for the cells to spread, proliferate and migrate, allowing for observing of live single-cell interactions. In this article, we focus on the production of this device and detailed protocol steps.
1. Fabrication of a wafer mold by soft lithography
NOTE: Mask pattern data is available in our previous publication14.
- Dehydrate a 4-inch silicon wafer in a 120 °C oven for 15 min.
- Spin coat 4 g of SU-8 2 negative photoresist onto a 4-inch silicon wafer at 1,000 rpm for 30 s to create a 5 µm thick layer (layer #1).
- Soft bake layer #1 on a 65 °C hotplate for 1 min and then transfer layer #1 to a 95 °C hotplate for 3 min.
- Cool layer #1 to room temperature, place it onto the holder of the semi-automatic mask aligner, and align with the layer #1 chrome-plated photomask (capture-site layer).
- Expose layer #1 with 365 nm UV light at a dose of 150 mJ/cm2.
- Remove layer #1 from the aligner and post-bake on a 65 °C hotplate for 1 min. Transfer layer #1 to a 95 °C hotplate for 1 min.
- Cool layer #1 to room temperature. Immerse in a propylene glycol monomethyl ether acetate solution to wash away the uncrosslinked photoresist for 2 min. Gently dry with nitrogen gas to reveal a layer #1 alignment mark.
- Cover the layer #1 alignment mark by an adhesive tape, spin coat 4 g of SU-8 10 negative photoresist onto the layer #1 at 1,230 rpm for 30 s to create a 25 µm thick layer #2.
- Remove the tape, soft bake layer #2 on a 65 °C hotplate for 3 min, and then transfer layer #2 to a 95 °C hotplate for 7 min.
- Cool layer #2 to room temperature, place layer #2 onto the holder of the semi-automatic mask aligner, and align the layer #2 chrome-plated photomask (bypass channel layer) to the layer #1 alignment mark.
- Expose layer #2 with 365 nm UV light at a dose of 200 mJ/cm2.
- Remove layer #2 from the aligner and post-bake on a 65 °C hotplate for 1 min and transfer layer #2 to a 95 °C hotplate for 3 min.
- Cool layer #2 to room temperature, and cover the layer #1 alignment mark by adhesive tape. Spin coat 4 g of SU-8 2050 negative photoresist onto layer #2 at 1,630 rpm for 30 s to create a 100 µm thick layer #3.
- Remove the tape, soft bake layer #3 on a 65 °C hotplate for 5 min, and then transfer layer #3 to a 95 °C hotplate for 20 min.
- Cool layer #3 to room temperature, place layer #3 onto the holder of the semi-automatic mask aligner, and align the layer #3 chrome-plated photomask (culture chamber layer) to the layer #1 alignment mark.
- Expose layer #3 with 365 nm UV light at a dose of 240 mJ/cm2.
- Remove layer #3 from the aligner and post-bake on a 65 °C hotplate for 5 min. Transfer layer #3 to a 95 °C hotplate for 10 min.
- Cool layer #3 to room temperature. Immerse in a propylene glycol monomethyl ether acetate solution to washed away the uncrosslinked photoresist for 10 min, and gently dry with nitrogen gas.
2. PDMS device preparation for multiple single cell capture
- Place the wafer mold and the weighing dish containing 100 µL of trichlorosilane in a desiccator (only for silanization) and apply a vacuum (-85 kPa) for 15 min.
NOTE: Silanize the wafer surface with trichlorosilane to create hydrophobic surface properties before PDMS castingso that it can effortlessly be peeled off from the wafer PDMS mold.
- Stop the vacuum, and then silanize the wafer mold in the desiccator (only for silanization) at 37 °C for at least 1 h.
- Mix PDMS base and PDMS curing agent in a ratio of 10:1. Pour a total of 20 g of mixed PDMS onto the wafer mold in a 15 cm dish.
- Place the 15 cm dish into a desiccator and apply vacuum (-85 kPa) for 1.5 min. Then remove the 15 cm dish from the desiccator. Keep for 20 min at room temperature. Finally, remove residual air bubbles in PDMS with nitrogen gas.
- Place the 15 cm dish in an oven at 65 °C for 2-4 h to cure PDMS.
- Remove the PDMS replica from the wafer mold, and then punch a 1.5 mm inlet and a 0.5 mm outlet on the PDMS using a 1.5 mm inner diameter and a 0.5 mm inner diameter puncher (Figure 1C).
- Clean the PDMS replica and the slide surface with removable tape and then treat the surface with oxygen plasma (100 W for 14 s).
- Manually align the PDMS replicas with the slide and bring them into contact with each other.
- Place the PDMS slide in a 65 °C oven for 1 day.
NOTE: Permanent bonding between the slide and the PDMS replica is achieved to form the device.
- Immerse the PDMS device in a container filled with phosphate buffered saline and place into a desiccator. Then apply vacuum (-85 kPa) for 15 min to remove air bubbles.
- Place the PDMS device in a cell culture hood and sterilize the device with UV light (light wavelength: 254 nm) for 30 min.
- Replace the PDMS device buffer with medium (DMEM-F12 basal medium containing 1% antibiotic and 10% fetal bovine serum) and incubate the PDMS device at 4 °C for 1 day. This prevents cells from adhering to the PDMS surface.
3. Preparation of a single-cell suspension
NOTE: Cell types include human lymphatic endothelial cells (LECs), human OSCC TW2.6 cells expressing WNT5B-specific shRNA (WNT5B sh4) and vector control (pLKO-GFP) which were obtained from our previous study15. Please refer to our previous publication for detailed cultivation steps.
- Remove the culture medium when the cells achieve 70-80% confluence. Then gently wash the cells with 5 mL of sterile PBS three times.
- Add 1 mL of DMEM-F12 medium containing 1 µM fluorescent dye into WNT5B sh4 and pLKO-GFP cells (use MV2 medium for LECs) and then incubate the cells for 30 min at room temperature.
NOTE: LECs were stained with green chloromethylfluorescein diacetate (CMFDA) Dye, WNT5B sh4 cells were stained with blue 7-amino-4-chloromethylcoumarin (CMAC) Dye and pLKO-GFP cells were stained with red Dil fluorescent dye.
- Gently wash the cells with 5 mL of sterile PBS three times.
- Remove the PBS and add 2 mL of 0.25% Trypsin-EDTA (0.05% Trypsin-EDTA for LECs).
- Incubate the cells for 4 min at room temperature and then gently tap the tissue culture dish to promote cells detachment.
- Add 4 mL of DMEM-F12 medium to disperse WNT5B sh4 and pLKO-GFP cells (For LECs use 3 mL of MV2 medium and 1 mL of trypsin neutralizer solution). Then transfer the cells into a 15 mL tube, and centrifuge at 300 x g for 3 min.
- Remove the supernatant, and resuspend the cell pellet in 1 mL of DMEM-F12 medium gently. Count the number of live cells in a hemocytometer by using the standard Trypan Blue exclusion method16. Prepare 1 mL of cell suspension at 3 x 105 cells/mL concentration in DMEM-F12 medium, and then keep cells on ice to prevent cell aggregation.
NOTE: In order to improve single-cell capture efficiency, careful preparation of the single cell suspension with well-dissociated is required.
4. Multiple single-cell capture and triple single-cell culture
- Connect a poly-tetrafluoroethene (PTFE) tube between the outlet of the device and syringe pump. Remove the medium and add 1 µL of cell suspension at a concentration of 3 x 105 cells/mL into the inlet of the PDMS device.
- Load the cell suspension into the device by a syringe pump at a flow rate of 0.3 µL/min (Figure 2A). Flow direction is from the inlet to the outlet.
NOTE: Load immediately after adding the cell suspension into the inlet to prevent cell sedimentation.
- Add 1 µL of DMEM-F12 medium into the inlet of the PDMS device after step 4.2.Load the DMEM-F12 medium into the device by a syringe pump at a flow rate of 0.3 µL/min (Figure 2B). Flow direction was flowing from the inlet to the outlet.
- Load 0.3 µL of DMEM-F12 medium into the device by a syringe pump at a flow rate of 10 µL/min (Figure 2C). Flow direction was flowing from the outlet to the inlet.
- Repeat steps 4.1 to 4.4 to load other cell types into the device.
- After completing the cell capture, use a microscope with 4x lens to image each culture chamber.
NOTE: The fluorescence emissions of the cells was used to identify and count the number of individual cells in each culture chamber.
- Remove the PTFE tube and seal the inlet and the outlet with polyolefin tape to create a closed culture system.
- Move the PDMS device to a 10 cm culture dish and add 10 mL of sterile PBS around the PDMS device to avoid evaporation of the medium from the device.
- Transfer the culture dish to an incubator (37 ° C, 5% CO2 and 95% humidity) for triple single-cell culture.
- Microscopically observe and photograph cell growth every 12 h.
The device has a three-layer structure as shown by the cross-section photograph of a cut PDMS device (Figure 1A). The first layer contains a capture-site (6.0 µm in width and 4.6 µm in height) that connects the culture chamber and the by-pass channel. The difference in flow resistance between the culture chamber and the by-pass channel causes the cells to flow into the capture position and fill the entrance of the small path. After a cell is captured at the capture position, the flow resistance of the small path is increased, causing the next incoming cell to go toward and through the by-pass channel to the next downstream capture-site. The serpentine design of the by-pass channel (25.0 µm in width and 26.4 µm in height) of the second layer is used to increase its flow resistance. The dimensions of the culture chamber in the third layer (500.3 µm in diameter and 111.3 µm in height) are designed to provide sufficient space for cell culture experiment, and to reduce the flow rate in the chamber to keep the cells from being flushed out of the chamber.
The tools required for the operation procedure (Figure 2) are common in general laboratories, including microscope, syringe pump, glass syringe, centrifuge tube, tubing, and pipette. The device is approximately 1/5 of a coverslip with a 2 mm thickness and is suitable for observing under a microscope with 4x to 20x lens. With this method, one type for single cells can be captured in the culture chambers under 7 min, so the total operation time for triple single cells was less than 21 min. The single cell capture efficiencies of the demonstrated three cells were 70.83% ± 15.42% (LECs), 73.96% ± 14.09% (WNT5B sh4) and 78.13% ± 3.13% (pLKO-GFP), respectively. The triple single cell capture efficiency was 47.92% ± 7.86% (Figure 3B). The reflux operation step is used to release the cells from the capture sites simultaneously and flow the cells into the culture chambers (Figure 2C). During this process, if a cell is not released from its capture-site, the released cell at its downstream capture-site will not enter the culture chamber due to by-pass channel having a lower flow resistance than the culture chamber. This is the major reason why the HSC has a triple single cell capture efficiency of only 47.92 ± 7.86%, which is still significantly greater than the probability of a Poisson distribution (~5%).
The cell culture results showed that multiple single-cell co-cultures of lymphatic endothelial cells and squamous cancer cells can be performed for 24 h, and the cells' proliferation and morphology can be observed under microscope (Figure 4, Supplementary Videos 1-3). In the presence of pLKO-GFP cells, WNT5B sh4 cells and LECs showed better proliferative capacity and showed that the morphology approached lamellipodia. These results demonstrate the ability of this device to high-efficiently capture multiple types of single cells in a culture chamber and provide a sufficient culture space useful for studying multiple single cell type interactions.
Figure 1: Photograph and microscope Images of hydrodynamic shuttling chip. (A) Microscope image of the PDMS sectional view. (B) A single cell trapping unit of magnified view. (C) Appearance of the chip containing 48 units. Scale bar: 100 µm. Please click here to view a larger version of this figure.
Figure 2: Hydrodynamic shuttling chip operation procedure. (A) Due to the high hydrodynamic resistance of the by-pass channel, a single cell is trapped in the capture-site. (B) After the first cell is trapped and has occluded the capture-site, the following cells flow toward the by-pass channel due to the increased flow resistance. Use medium to wash remaining cells in channel. (C) Reflux the cell into culture chamber. Scale bar: 100 µm. Please click here to view a larger version of this figure.
Figure 3: Single cell capture efficiency of LECs, WNT5B sh4 and pLKO-GFP in the hydrodynamic shuttling chip. (A) Microscope image of captured triple single cells in culture chamber. (B) The green, blue, and red bars represent the capture efficiency of individual cells, respectively, and the yellow bars represent the capture efficiency of triple single-cell. Scale bar: 100 µm. Please click here to view a larger version of this figure.
Figure 4: Paired single cell co-culture and triple single cell co-culture in the hydrodynamic shuttling chip for 24 h. (A) Microscope image of triple single cells co-culture for 0, 12 and 24 h. (B) Microscope image of pLKO-GFP and WNT5B sh4 cells co-culture for 0, 12 and 24 h. LECs were stained with green CMFDA Dye, WNT5B sh4 cells were stained with blue CMAC Dye and pLKO-GFP cells were stained with red DiI fluorescent dye. Scale bar: 100 µm. Please click here to view a larger version of this figure.
Supplementary Video 1: Cell Loading. Please click here to view this video (Right click to download).
Supplementary Video 2: Cell Refluxing. Please click here to view this video (Right click to download).
Supplementary Video 3: Second cell type reflux into the culture chamber. Please click here to view this video (Right click to download).
The intercellular interactions of various cells in the tumor microenvironment play an important role in the progression of the tumor17. In order to understand the mechanism of cell-cell interactions, co-culture systems are used as a common analytical method. However, multiple cell types and the heterogeneity of the cells themselves have led to experimental complexity and analytical difficulties.
The hydrodynamic shuttling chip allows multiple single-cell loading in the culture chamber by a deterministic method, without being limited by the Poisson distribution limitation in the dilution method and the microwell platform. By providing a high triple single cell capture efficiency (greater than 45%, the Poisson distribution method is 5%) and demonstrating that in the culture chamber, space is sufficient for cell growth and proliferation (Figure 4). Due to its ability to efficiently perform multiple single-cell captures and live cell culture observations with simple setup and protocol, we envision these microfluidic devices as useful tools for an extensive range of applications, including cell-cell interactions between multiple cells18, drug screening19, and cancer biology20. On the other hand, the device structure is moldable, and the structure and size of the capture-site can be changed and applied to other fields such as microorganisms and plant cells. In theory, our method is also adaptable to establish and track microbial co-cultures (e.g., bacteria, yeasts, etc.).
The main limitation of this approach is that the precision level required for the device fabrication is high. This is mainly because the flow resist of the smallest channel can be dramatically changed if its fabricated dimensions are slightly offset. The control of resistances of the microchannels are crucial for the high-efficiency single-cell capture of the device. On the other hand, during cell culture, there is no closure between the chambers. Therefore, paracrine secretion of cells in the chamber may spread into other chambers to affect other cells. Finally, attention must be paid to the cleanliness of the channel and tubing during the preparation of the device. The culture medium and any buffer used in the experiment need to be filtered to prevent particles and debris from blocking the channel.
The authors declare that they have no competing financial interests.
This work was supported by a grant from the Ministry of Science and Technology (105-2628-E-400-001-MY2), and the Ph.D. Program in Tissue Engineering and Regenerative Medicine, National Chung Hsing University and National Health Research Institutes.
|3M Advanced Polyolefin Diagnostic Microfluidic Medical Tape||3M Company||9795R|
|AutoCAD software||Autodesk||AutoCAD LT 2011||Part No. 057C1-74A111-1001|
|CellTracke Blue CMAC Dye||Invitrogen||C2110|
|CellTracker Green CMFDA Dye||Invitrogen||C7025|
|Conventional oven||YEONG-SHIN company||ovp45|
|Desiccator||Bel-Art Products||F42020-0000||Space saver vacuum desiccator 190 mm white base|
|DiIC12(3) cell membrane dye||BD Biosciences||354218||Used as a cell tracker|
|Endothelial Cell Growth Medium MV 2||PromoCell||C-22022|
|Fetal bovine serum Hyclone||Thermo||SH30071.03HI|
|Hamilton 700 series Glass syringe ( 0.1 ml )||Hamilton||80630||100 µL, Model 710 RN SYR, Small Removable NDL, 22s ga, 2 in, point style 2|
|Harris Uni-Core puncher||Ted Pella Inc.||15075||with 1.5mm inner-diameter|
|Harris Uni-Core puncher||Ted Pella Inc.||15071||with 0.5mm inner-diameter|
|Msak aligner||Deya Optronic CO.||A1K-5-MDA|
|Oxygen plasma||NORDSON MARCH||AP-300|
|Plasma cleaner||Nordson||AP-300||Bench-Top Plasma Treatment System|
|Polydimethylsiloxane (PDMS) kit||Dow corning||Sylgard 184|
|Poly-tetrafluoroethene (PTFE)||Ever Sharp Technology, Inc.||TFT-23T||inner diameter, 0.51 mm; outer diameter, 0.82 mm|
|Removable tape||3M Company||Scotch Removable Tape 811|
|Silicon wafer||Eltech corperation||SPE0039|
|Spin coater||Synrex Co., Ltd.||SC-HMI 2" ~ 6"|
|Stereomicroscope||Leica Microsystems||Leica E24|
|SU-8 10 negative photoresist||MicroChem||Y131259|
|SU-8 2 negative photoresist||MicroChem||Y131240|
|SU-8 2050 negative photoresist||MicroChem||Y111072|
|SU-8 developer||Grand Chemical Companies||GP5002-000000-72GC||Propylene glycol monomethyl ether acetate|
|Syringe pump||Harvard Apparatus||703007|
|Trichlorosilane||Gelest, Inc||SIT8174.0||Tridecafluoro-1,1,2,2-tetrahydrooctyl. Hazardous. Corrosive to the respiratory tract, reacts violently with water.|
|Trypsin Neutralizer Solution||Gibco||R-002-100|
- Goers, L., Freemont, P., Polizzi, K. M. Co-culture systems and technologies: taking synthetic biology to the next level. Journal of the Royal Society, Interface. 11, (96), 20140065 (2014).
- Collins, D. J., Neild, A., deMello, A., Liu, A. Q., Ai, Y. The Poisson distribution and beyond: methods for microfluidic droplet production and single cell encapsulation. Lab on a Chip. 15, (17), 3439-3459 (2015).
- Leong, K. G., Wang, B. E., Johnson, L., Gao, W. Q. Generation of a prostate from a single adult stem cell. Nature. 456, 804 (2008).
- Roman, G. T., Chen, Y., Viberg, P., Culbertson, A. H., Culbertson, C. T. Single-cell manipulation and analysis using microfluidic devices. Analytical and Bioanalytical Chemistry. 387, (1), 9-12 (2007).
- Frimat, J. P., et al. A microfluidic array with cellular valving for single cell co-culture. Lab on a Chip. 11, (2), 231-237 (2011).
- Rettig, J. R., Folch, A. Large-Scale Single-Cell Trapping And Imaging Using Microwell Arrays. Analytical Chemistry. 77, (17), 5628-5634 (2005).
- Gracz, A. D., et al. A high-throughput platform for stem cell niche co-cultures and downstream gene expression analysis. Nature Cell Biology. 17, (3), 340 (2015).
- Tumarkin, E., et al. High-throughput combinatorial cell co-culture using microfluidics. Integrative Biology. 3, (6), 653-662 (2011).
- Hong, S., Pan, Q., Lee, L. P. Single-cell level co-culture platform for intercellular communication. Integrative Biology. 4, (4), 374-380 (2012).
- Chung, M. T., Núñez, D., Cai, D., Kurabayashi, K. Deterministic droplet-based co-encapsulation and pairing of microparticles via active sorting and downstream merging. Lab on a Chip. 17, (21), 3664-3671 (2017).
- Chen, Y. C., et al. Paired single cell co-culture microenvironments isolated by two-phase flow with continuous nutrient renewal. Lab on a Chip. 14, (16), 2941-2947 (2014).
- Hong, S., Lee, L. P., Pan, Q. Single-cell level co-culture platform for intercellular communication. Integrative Biology. 4, (4), 374-380 (2012).
- Segaliny, A. I., et al. Functional TCR T cell screening using single-cell droplet microfluidics. Lab on a Chip. 18, (24), 3733-3749 (2018).
- He, C. K., Chen, Y. W., Wang, S. H., Hsu, C. H. Hydrodynamic shuttling for deterministic high-efficiency multiple single-cell capture in a microfluidic chip. Lab on a Chip. 19, (8), 1370-1377 (2019).
- Wang, S. H., et al. Tumour cell-derived WNT5B modulates in vitro lymphangiogenesis via induction of partial endothelial-mesenchymal transition of lymphatic endothelial cells. Oncogene. 36, 1503 (2016).
- Strober, W. Trypan Blue Exclusion Test of Cell Viability. Current Protocols in Immunology. 111, (1), A3.B.1-A3.B.3 (2015).
- Kim, J. B., Stein, R., O'hare, M. J. Three-dimensional in vitro tissue culture models of breast cancer-a review. Breast Cancer Research and Treatment. 85, (3), 281-291 (2004).
- Shin, Y., et al. Microfluidic assay for simultaneous culture of multiple cell types on surfaces or within hydrogels. Nature Protocols. 7, (7), 1247-1259 (2012).
- Wu, M. H., Huang, S. B., Lee, G. B. Microfluidic cell culture systems for drug research. Lab on a Chip. 10, (8), 939-956 (2010).
- Hong, J. W., Song, S., Shin, J. H. A novel microfluidic co-culture system for investigation of bacterial cancer targeting. Lab on a Chip. 13, (15), 3033-3040 (2013).