This protocol describes a reconfigurable membrane-based cell culture platform that integrates the open-well format with fluid flow capabilities. This platform is compatible with standard protocols and allows for reversible transitions between open-well and microfluidic culture modes, accommodating the needs of both engineering and bioscience laboratories.
Microphysiological systems are miniaturized cell culture platforms used to mimic the structure and function of human tissues in a laboratory setting. However, these platforms have not gained widespread adoption in bioscience laboratories where open-well, membrane-based approaches serve as the gold standard for mimicking tissue barriers, despite lacking fluid flow capabilities. This issue can be primarily attributed to the incompatibility of existing microphysiological systems with standard protocols and tools developed for open-well systems.
Here, we present a protocol for creating a reconfigurable membrane-based platform with an open-well structure, flow enhancement capability, and compatibility with conventional protocols. This system utilizes a magnetic assembly approach that enables reversible switching between open-well and microfluidic modes. With this approach, users have the flexibility to begin an experiment in the open-well format using standard protocols and add or remove flow capabilities as needed. To demonstrate the practical usage of this system and its compatibility with standard techniques, an endothelial cell monolayer was established in an open-well format. The system was reconfigured to introduce fluid flow and then switched to the open-well format to conduct immunostaining and RNA extraction. Due to its compatibility with conventional open-well protocols and flow enhancement capability, this reconfigurable design is expected to be adopted by both engineering and bioscience laboratories.
Vascular barriers serve as a critical interface that separates the blood compartment from the surrounding tissue. They play a critical role in preserving homeostasis by attracting immune cells, controlling molecular permeability, and shielding against the intrusion of pathogens into the tissue1,2. In vitro culture models have been developed to mimic the in vivo microenvironment, enabling systematic investigations into the factors and conditions that impact barrier properties in both healthy and diseased states3,4.
The most widely used approach for such culture models is the Transwell-like "open-well" configuration5, where a porous, track-etched culture membrane separates media-filled compartments (Figure 1A). In this format, cells can be seeded on either side of the membrane, and a wide range of experimental protocols has been developed. However, these systems are limited in their ability to provide the fluid flows essential for supporting barrier maturation and mimicking immune cell circulation seen in vivo5,6. Consequently, they cannot be used for studies requiring dynamic flows that introduce drug doses, mechanical stimulation, or fluid-induced shear stresses6,7,8.
To overcome the limitations of open-well systems, microfluidic platforms that combine porous culture membranes with individually addressable fluidic channels have been developed9. These platforms offer precise control over fluid routing, perfusion, and the introduction of chemical compounds, controlled shear stimulation, and dynamic cell addition capabilities7,10,11,12,13. Despite the advanced capabilities provided by microfluidic platforms, they have not seen widespread adoption in bioscience laboratories due to complex microfluidic protocols and their incompatibility with established experimental workflows4,10,14.
To bridge the gap between these technologies, we present a protocol that employs a magnetically reconfigurable, module-based system. This system can be easily switched between open-well and microfluidic modes based on the specific needs of the experiment. The platform features an open-well device, known as m-µSiM (modular microphysiological system enabled by a silicon membrane), with a 100 nm thick culture membrane (nanomembrane). This nanomembrane possesses high porosity (15%) and glass-like transparency, as illustrated in Figure 1B. It physically separates the top compartment from a bottom channel, allowing for molecular transport across physiological length scales15. Unlike conventional track-etched membranes, which have known challenges in imaging live cells with bright-field imaging, the nanomembrane's favorable optical and physical properties enable clear visualization of cells on either side of the membrane surface15,16,17.
The present protocol outlines the fabrication of specialized seeding and flow modules and explains the magnetic reconfiguration of the platform. It demonstrates how the platform can be employed to establish endothelial barriers under both static and dynamic conditions. This demonstration reveals that endothelial cells align along the flow direction, with an upregulation of shear-sensitive gene targets under shear stimulation.
This design can be used in various modes based on experimental requirements and the preferences of the end user. Prior to each experiment, consult the decision flow chart presented in Figure 2 to determine the necessary steps and modules for the protocol. For example, if the user intends to maintain the open-well format throughout an experiment to directly compare it with the Transwell-type system, the patterning stencil is not required for cell seeding. The core module is commercially available (see Table of Materials), and the ultrathin nanomembrane can be selected from a library of materials with different porosity and pore sizes to suit experimental needs.
1. Fabrication of the patterning stencil
NOTE: The patterning stencil serves to position cells exclusively on the porous region of the membrane chip, preventing cells from settling onto the surrounding silicon layer where they could potentially experience damage after the flow module is added16 (refer to Figure 3). Damage to the monolayer can adversely affect barrier integrity and compromise experimental outcomes. The stencil is unnecessary in an open, static culture, as there is no risk of damage.
2. Fabrication of the flow module
NOTE: The flow module shares a similar footprint with the clover-shaped well of the core module and includes a molded microchannel (width = 1.5 mm, height = 0.2 mm, length = 5 mm). The clover shape aids in aligning the channel over the porous culture region (Figure 5).
3. Fabrication of lower and upper acrylic housings
NOTE: The core module fits into the lower housing. The attraction between embedded magnets in the housings compresses and seals the flow module to the core module (Figure 6).
4. Fabrication of the flow circuit
NOTE: The closed-loop flow circuit contains two sample collection vials as reservoirs (Figure 7). The inlet reservoir has a polyvinylidene difluoride (PVDF) filter to allow the cell media to equilibrate with the CO2 concentration in the incubator.
5. Cell seeding
NOTE: Similar to conventional membrane inserts, different cell types can be cultured on the nanomembrane. A secondary cell type can also be co-cultured on the other side of the membrane in the bottom channel15.
6. Reconfiguration to microfluidic mode
7. Conducting downstream analysis in open-well format after flow introduction
NOTE: The culture time here depends on the experimental goals. Users can conduct downstream analysis (e.g., Immunocytochemistry, RNA extraction) either in the open-well or microfluidic formats based on their preference. For instance, if an open-well format is preferred, the system should be reconfigured to conduct assays based on standard protocols16,19.
The open-well core module is initially positioned within a specific cavity created by a lower housing and a coverslip, as illustrated in Figure 6A. Subsequently, the flow module, which includes a microchannel and access ports, is inserted into the well of the core module. The flow module is securely sealed against the silicon support layer of the membrane due to the magnetic attraction force between magnets embedded in the lower and upper housings, as depicted in Figure 6B. To evaluate the effectiveness of this magnetic latching mechanism, a burst pressure test was conducted, demonstrating that the system can withstand dead-ended pressures of up to 38.8 ± 2.4 kPa. This pressure tolerance significantly exceeds the typical operating pressures encountered in cell culture applications. Furthermore, the system remains free of leaks when subjected to flow rates of up to 4000 µL/min, which is equivalent to a shear stress of 74 dynes/cm2 at the culture region16.
When developing a platform that can switch between open-well and microfluidic modes, careful consideration must be given to the cell seeding approach, which is not typically a concern for conventional static open-well platforms16. Damage to the monolayer around the channel boundary could introduce complications in experimental results20. To address this issue, a removable stencil was designed that fits within the open well of the core module and provides a specific window for cells to settle preferentially on the membrane surface (Figure 3). Once the cell monolayer is patterned and reaches confluency, the user has the flexibility to continue the experiment in the open-well format or reconfigure the platform into microfluidic mode to expose the cell monolayer to physiological shear stress (Figure 3). The magnetic latching mechanism provides the ability to easily switch between the open-well and microfluidic formats as required. For instance, the device can be reverted to the open-well format after a flow stimulation, offering users the flexibility to conduct a variety of assays (such as immunostaining, RNA extraction, and molecular permeability measurements) using standard experimental protocols15,16.
In the physiological setting of the human body, the vascular barrier is exposed to flow-induced shear stress, which serves as a key biophysical cue that affects the structure and function of the barrier5,21,22. Thus, the addition of fluid flow in microphysiological systems is a key requirement. To demonstrate the versatility of the platform, an HUVEC monolayer was established in an open-well format using standard protocols. After 24 h of static culture, the platform was reconfigured into microfluidic mode to expose the cell monolayer to 10.7 dynes/cm2 shear stress for 24 h. The results indicated that cells cultured under flow aligned along the flow direction while cells cultured without flow remained randomly oriented (Figure 8A,B). After shear stimulation, the platform was reconfigured to the open-well format to extract RNA using standard protocols. The results indicated that the exposure of cells to shear stress resulted in the upregulation of Kruppel-like factor 2 (KLF2) and endothelial nitric oxide synthase (eNOS), which serve critical roles such as anti-thrombotic and atheroprotective functions in healthy blood vessels23,24(Figure 8C).
Figure 1: Comparison of in vitro vascular barrier models. Schematic illustration of (A) conventional Transwell-like inserts and (B) the open-well m-µSiM. Bright-field images of a confluent HUVEC monolayer highlight the difference in bright-field imaging quality between a track-etched membrane and an ultrathin nanomembrane. Scale bars = 100 µm. Adapted from Mansouri et al.16. Please click here to view a larger version of this figure.
Figure 2: Decision-making flow chart. A flow chart based on experimental needs and downstream analysis preferences. Please click here to view a larger version of this figure.
Figure 3: Experimental workflow of the platform. (A) To directly position cells on the porous membrane, a removable patterning stencil is inserted in the well of the core module (the inset shows patterned cells, yellow lines exhibit microchannel boundaries). (B) The stencil can be kept or removed in the device for static cell culture. (C) To reconfigure the platform into microfluidic mode, the stencil is replaced with the flow module. Because of the magnetic sealing mechanism, the configuration is reversible; housings and the flow module can be removed to switch into open-well mode. Scale bar = 200 µm. Adapted from Mansouri et al.16. Please click here to view a larger version of this figure.
Figure 4: Schematic illustration of the molds. (A) The stencil mold. (B) Laser-cut acrylic sheet. (C) assembled view of the stencil mold. (D) Flow module mold. (E) Laser-cut acrylic sheet. (F) Assembled view of the flow module mold. Triangle-shaped features are alignment marks to facilitate attaching acrylic sheets to the molds. Please click here to view a larger version of this figure.
Figure 5: Schematic of the clover-shape flow module. (A) The contact interface between the flow module and the membrane chip. The inlet and outlet ports for fluid flow are shown in pink. (B) 3D image of the PDMS flow module. Adapted from Mansouri et al.16. Please click here to view a larger version of this figure.
Figure 6: Magnetic assembly for device reconfiguration. (A) Schematic demonstration of components for device reconfiguration into microfluidic mode. Embedded magnets with opposite poles induce attraction for the sealing. (B) Cross-sectional view of the reconfigured device showing the vascular channel in pink and the tissue compartment in green. Adapted from Mansouri et al.16. Please click here to view a larger version of this figure.
Figure 7: Assembled view of the flow circuit. The circuit consists of a peristaltic pump, two reservoirs for supplying cell media and damping fluctuations, tubing, and an acrylic stage to hold the components in place. Adapted from Mansouri et al.16. Please click here to view a larger version of this figure.
Figure 8: Comparison of HUVECs cultured in open-well and microfluidic modes. Cells were seeded and cultured in open-well for 24 h to establish a confluent monolayer. During the subsequent 24 h period, one set of devices was reconfigured into microfluidic mode. (A) Cells cultured under flow (10.7 dynes.cm-2 shear stress) aligned along the flow direction (the inset shows actin and nuclei of cells in green and blue, respectively). (B) Cells cultured without flow in open-well format showed no alignment. The length of bars in radar plots shows the number of cells in the corresponding direction. (C) Cells cultured under flow showed higher upregulation of KLF2 and eNOS genes compared to the no-flow condition (**p < 0.01, n = 3, mean ± SD). Scale bars = 100 µm. Adapted from Mansouri et al.16. Please click here to view a larger version of this figure.
Supplementary Table 1: Shear stress on the nanomembrane surface at different flow rates. This table provides information about shear stress values on the nanomembrane surface at various flow rates. Please click here to download this File.
Supplementary Coding File 1: CAD model of the stencil mold. Please click here to download this File.
Supplementary Coding File 2: CAD model of laser cut cavities for the stencil mold. Please click here to download this File.
Supplementary Coding File 3: CAD model of the flow module. Please click here to download this File.
Supplementary Coding File 4: CAD model of laser cut cavities for the flow module mold. Please click here to download this File.
Supplementary Coding File 5: CAD model of the upper housing. Please click here to download this File.
Supplementary Coding File 6: CAD model of the lower housing. Please click here to download this File.
Supplementary Coding File 7: CAD model of the acrylic stage. Please click here to download this File.
The aim of this protocol is to develop a practical method for incorporating flow capabilities into an open-well platform featuring an ultrathin nanomembrane. In this design, a magnetic latching approach is utilized, allowing switching between open-well and fluidic modes during experiments and combining the advantages of both approaches. Unlike conventional permanently bonded platforms, magnetic latching allows the platform to be disassembled at convenient points during the experimental workflow16,25,26,27. In this platform, cells are seeded, and the barrier is established in the familiar open-well format, and then the system is reconfigured to a fluidic mode by simply adding a flow module and magnetically sealing it. This reconfiguration offers two key benefits: first, it enables the simulation of physiological conditions by exposing the cell monolayer to fluid-induced shear stress; second, it facilitates the introduction of secondary components like immune cells under flow conditions16. The magnetic latching feature offers a tool-free assembly method, enabling on-demand switching between open-well and microfluidic formats. Consequently, downstream analyses, such as immunocytochemistry and RNA extraction, can be conducted in the familiar open-well format using standard techniques or using microfluidic techniques as desired16.
Magnetic sealing is a critical parameter of the design that enables the reconfigurability of the platform, and thus, several considerations are required for its accurate function: (1) The height of the lower housing must match the height of the core module (tolerance: ±0.1 mm) to enclose the module properly. (2) The height of the flow module must be slightly taller than the well of the core module (0-0.1 mm). (3) This design is adaptable to a wide range of materials and doesn't mandate that the flow module be constructed from elastomeric materials. The crucial element is to incorporate a soft, gasket-like material at the interface between the flow module and the membrane chip to establish a reliable seal. (4) The diameter of the laser-cut cavities in the housings for press-fitting magnets must be optimized based on the instrument used28. Magnets might detach and lead to flow leakage if the cavity diameter is too big, or housings might crack during magnet insertion if cavities are too small. By following the aforementioned considerations, the magnetic assembly approach introduced here can be applied to reconfigure other open-well systems into microfluidic mode.
To have the reconfiguration capability, it is necessary to use a patterning stencil that allows precise positioning of cells onto the membrane surface. The stencil ensures that no cells are outside the porous culture region (i.e., on the silicon support region), which can be damaged upon the insertion of the flow module. This aspect becomes particularly crucial in co-culture models because cells inadvertently seeded in unintended locations cannot actively participate in the development of the barrier tissue. Nonetheless, these misplaced cells are usually retrieved during the lysis process and can potentially introduce bias in gene expression studies that investigate interactions between cells across the membrane20. Open-well cell seeding using the stencil enhances seeding efficiency by mitigating cell losses along the flow path, which are inherent in conventional microfluidic platforms4,16. This platform is also capable of incorporating multiple cell types and ECM materials15,16,17. For instance, a cell-laden hydrogel can be injected into the lower channel of the device to mimic the tissue compartment while an endothelium is established on top of the membrane.
Although the magnetic modules and components can be reused, maintaining effective sealing over several uses can be an issue due to the magnets loosening in the housing and decreasing the sealing force. This problem can be mitigated by making new laminated modules for each experiment or mass-producing components using injection molding or 3D printing techniques once designs have been established. In the method described in this protocol, the flow module is manually placed in the core module, and the multiplexing and automation capability of the system is limited. To improve experimental throughput, the flow module and upper housing can be integrated into one functional module containing internal routing channels and standardized tubing connections that simultaneously perfuse several modules in parallel.
Components of the open-well core module are mass-produced and commercially available. Other components of the system, such as the flow module, housings, and patterning stencil, can also be fabricated on a large scale. Furthermore, the reconfigurability of the platform allows the addition of sensors and actuators (e.g., transendothelial electrical resistance module, electroporation module) for real-time stimulation and measurements. Overall, this platform combines conventional and microfluidic approaches to help support widespread use outside of engineering laboratories.
The authors have nothing to disclose.
This research was funded in part by the National Institute of Health under award numbers R43GM137651, R61HL154249, R16GM146687, and NSF grant CBET 2150798. The authors thank the RIT Machine Shop for aluminum mold fabrication. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.
0.5 x 0.86 Micro Flow tubes | Langer Instruments | WX10-14 & DG Series | |
1 mm Disposable Biopsy Punches, Integra Miltex | VWR | 95039-090 | |
1x PBS 7.4 pH | ThermoFisher Scientific | 10010023 | |
20 GAUGE IT SERIES DISPENSING TIP | Jensen Global | JG20-1.5X | |
21 GAUGE NT PREMIUM SERIES ANGLED DISPENSING TIP | Jensen Global | JG21-1.0HPX-90 | |
3M 467 MP Pressure senstitive adhesive (PSA) | DigiKey | 3M9726-ND | |
3M 468 MP Pressure senstitive adhesive (PSA) | DigiKey | 3M9720-ND | |
AlexaFluor 488 conjugated phalloidin | ThermoFisher Scientific | A12379 | |
Applied Biosystems TaqMan Fast Advanced Master Mix | Thermo Fisher Scientific | 4444556 | |
Bovine Serum Albumin (BSA), Fraction V, 98%, Reagent grade, Alfa Aesar, Size = 10 g | VWR | AAJ64100-09 | |
Clear Scratch- and UV-Resistant Cast Acrylic Sheet | McMaster-Carr | 8560K171 | 12" x 12" x 1/16" |
Clear Scratch- and UV-Resistant Cast Acrylic Sheet | McMaster-Carr | 8589K31 | 12" x 12" x 3/32" |
Clear Scratch- and UV-Resistant Cast Acrylic Sheet | McMaster-Carr | 8560K191 | 12" x 12" x 7.64" |
Corning Fibronectin, Human, 1 mg | Corning | 47743-728 | |
Cover Glasses, Globe Scientific, L x W = 24 x 60 mm | VWR | 10118-677 | |
DOW SYLGARD 184 SILICONE ENCAPSULANT CLEAR 0.5 KG KIT | Ellsworth Adhesives | 4019862 | |
EGM-2 Endothelial Cell Growth Medium-2 BulletKit | Lonza | CC-3162 | |
Fixture A1&A2 | SiMPore Inc. | NA | |
Fixture B1&B2 | SiMPore Inc. | NA | |
High Capacity cDNA Reverse Transcription Kit with RNase Inhibitor | Thermo Fisher Scientific | 4374966 | |
Human umbilical vein endothelial cells (HUVEC) | ThermoFisher Scientific | C0035C | |
LIVE/DEAD Cell Imaging Kit (488/570) | Thermo Fisher Scientific | R37601 | |
Molecular Probes Hoechst 33342, Trihydrochloride, Trihydrate | Thermo Fisher Scientific | H3570 | |
Nickel-plated magnets (4.75 mm diameter, 0.34 kg pull force) | K&J Magnetics | D31 | 3/16" dia. x 1/16" thick |
Paraformaldehyde, 4% w/v aq. soln., methanol free, Alfa Aesar | Fisher Scientific | aa47392-9M | |
Peristaltic Pump | Langer Instruments | BQ50-1J-A | |
Photoresist SU-8 developer solution | Fisher Scientific | NC9901158 | |
PVDF syringe filters | PerkinElmer | 2542913 | |
Silicon wafer | University wafer,USA | 1196 | |
SU-8 3050 | Fisher Scientific | NC0702369 | |
Target gene: eNOS (Hs01574659_m1) | ThermoFisher Scientific | 4331182 | |
Target gene: GAPDH (Hs02786624_g1) | ThermoFisher Scientific | 4331182 | |
Target gene: KLF2 (Hs00360439_g1) | ThermoFisher Scientific | 4331182 | |
Thermo Scientific Pierce 20x PBS Tween 20 | Thermo Fisher Scientific | 28352 | |
Transport Tube Sample White caps, 5 mL, Sterile | VWR | 100500-422 | |
TRI-reagent | ThermoFisher Scientific | AM9738 | |
Ultrathin Nanoporous Membrane Chip | SiMPore Inc. | NPSN100-1L | The design is compatible with all of SiMPore membranes |
uSiM component 1 | SiMPore Inc. | NA | |
uSiM component 2 | SiMPore Inc. | NA |