Remote-controlled 3D Porous Magnetic Interface towards High- throughput Dynamic 3D Cell Culture

Abstract

Mechanical stimuli have been shown to play a large role in cellular behavior, including cellular growth, differentiation, morphology, homeostasis, and disease. Therefore, developing bioreactor systems that can create complex mechanical environments for both tissue engineering and disease modeling drug screening is appealing. However, many of existing systems are restricted due to their bulky size with external force generators, destructive microenvironment control, and low throughput. These shortcomings have preceded to the utilization of magnetic stimuli responsive materials, given their untethered, fast, and tunable actuation potential at both the microscale and macroscale level, for seamless integration into cell culture wells and microfluidic systems. Nevertheless, magnetic soft materials for cell culture have been limited due to the inability to develop well-defined 3D structures for more complex and physiological relevant mechanical actuation. Herein, we introduce a facile fabrication process to develop magnetic- PDMS (polydimethylsiloxane) porous composite designs with both well-defined and controllable microlevel and macrolevel features to dynamically manipulate 3D cell-laden gel at the scale. The intrinsic stiffness of the magnetic-PDMS porous composites is also modulated to control the deformation potential to mimic physiological relevant strain levels, with 2.89 to 11% observed in magnetic actuation studies. High cell viability was achieved with the culturing of both human adipose stem cells (hADMSCs) and human umbilical cord mesenchymal stem cells (hUCMSCs) in 3D cell-laden gel interfaced with the magnetic-PDMS porous composite. Also, the highly interconnected porous network of the magnetic-PDMS composites facilitated free diffusion throughout the porous structure showcasing the potential of a multi-surface contact 3D porous magnetic structure in both reservoir and 96-well plate insert designs for more complex dynamic mechanical actuation. In conclusion, these studies provide a means for establishing a biocompatible, tunable magnetic-PDMS porous composite with fast and programmable dynamic strain potential making it a suitable platform for high-throughput, dynamic 3D cell culture.

Graphical Abstract

Introduction

Cells have been shown, especially mesenchymal stem cells (MSCs), to both sense and act upon their mechanical environment via a mechanism called mechanotransduction^1–7. This process consists of the cell being able to convert mechanical cues into intracellular biochemical signals for driving cellular growth, differentiation, morphology, homeostasis, and disease. Mechanical stimuli experienced by cells can come in both intrinsic and extrinsic forms. Current studies have looked at primarily one dependent force variable at a time to understand how each stimuli affects cellular behavior^{1, 2}. Most of these investigations utilized extrinsic forces to manipulate the intrinsic behavior of the cells through actomyosin contractility and cytoskeleton rearrangement, which can trigger biochemical signals depending on the dynamic nature of the force^3–5. There are many different platforms to mimic in vivo physiological relevant forces experienced on cell cultures including substrate stiffness^{6, 7}, micro/nanopattering^8–10, hydrostatic pressure^{11, 12}, perfusion flow^{13, 14}, shear forces^{15, 16}, compressive forces^17–20 , tensile forces^21–23, and vibration^24–26. The findings have led to the development of bioreactor systems for providing specific and complex mechanical environment used in tissue engineering and disease modeling^{27, 28}. Some developed microscale systems can overcome the complications involved with conventional bioreactor systems which are bulky, lack microenvironment control, and have low experimental throughput^{19, 25, 29, 30}. The mechanical stimuli to 3D cell cultures also can be achieved via pneumatic actuation of a PDMS membranes ^{19, 31}, dielectric elastomer actuators^{32, 33}, and magnetic elastomer actuation^34–37 in high-throughput microfluidic, micro array, or cell culture well-plate systems. However, the current existing microscale systems to date are still lacking the capacity for spatial specific, patterned interface mechanical actuation with most needing external forces to act upon an interface substrate.

Microscale actuator systems is critically needed in evaluating and developing physiological relevant actuation to cell cultures with multiple modes and ranges of actuation. Herein, we introduced a magnetic stimuli responsive materials to bridge such technology gap. Although magnetic materials have shown to provide highly tunable, predictive, complex transformation potential at both the macroscopic and microscopic level^38–40, developing magnetic elastomer actuators with defined 3D structures and porosity using PDMS (polydimethylsiloxane) for dynamic cell culture is emerging ^{6–10, 19, 34–36, 41}. Investigators have explored the development of magnetic porous composites as devices for on-demand drug delivery by the usage of sacrificial molds of salt or sugar microparticles to define the porous network^{42, 43}. Yet, there has been minimal investigation into porous magnetic composites as scaffold or porous interfaces with high-fidelity porous networks for cell culture considerations. Therefore, we developed a high- fidelity, tunable magnetic PDMS porous composite that can be actuated remotely via an external magnetic field. To our knowledge, this will be the first magnetic PDMS porous composite with well-defined 3D microstructures for full control over the porosity and micropatterns for ensuing topographical cue studies on 3D cell cultures. Due to the ability to precisely manipulate the intrinsic stiffness and magnetic particle density in the porous magnetic PDMS composites, the resulted physiological relevant strain level enables the potential control of human umbilical cord mesenchymal stem cells (hUCMSCs) differentiation. Also, the porous magnetic PDMS showed to be a suitable 3D cell culture substrate with high cell viability for both adipose-derived mesenchymal stem cells (hADMSCs) and hUCMSCs in a 3D fibrin gel. The highly interconnected porous network of the magnetic PDMS porous composites enhances free diffusion through the material unlocking the potential to use as the multiple surface contact porous constructs for more complex dynamic mechanical actuation. With this platform using remotely controllable porous materials, small and complex cell culture magnetic interfaces can be implemented into standard cell culture 96 well plate as a potential high-throughput platform for advancing dynamic 3D cell culture systems.

Materials and Methods

FDM 3D Printing of Sacrificial Templates

A fused filament fabrication printer (N2, Raise 3D), or fused deposition modeling (FDM), was used to print the PVA sacrificial templates. The 3D template structure was designed using a computer aided design (CAD) 3D slicer software provided by Raise 3D (ideaMaker). The 3D CAD design file then could be exported to the printer by a .STL file to perform the 3D print. The method of printing the 3D templates include the PVA filament being fed through a heated extrusion nozzle that melts the thermoplastic and then solidifies on the heated stage bed after extrusion. The .STL computer slice file guides the nozzle to execute the print layer-by-layer by stacking x-y planes on top of each other. The printing infill percentage and infill patterns were varied to tune the mechanical properties, porosity, and topography of the interface design. The infill percentage refers to the amount of filament printed per layer with 100% infill being a filled structure with no pores. Given the PVA 3D print is a sacrificial template, the infill printed represents the pores of the final construct. A 70% infill percentage was chosen as this was the highest consistent attainable porosity for the final porous composite. The infill pattern indicates the micropattern printed to form the 3D geometric structure and both the rectangular grid and gyroid patterns were investigated. The printing parameters are shown in Table S1.

3D Fabrication of Magnetic Porous Composites

The PVA printed templates were used to form the 3D magnetic porous composite structures. Carbonyl iron microparticles (Sigma-Aldrich) and PDMS (Sylgard 184, Dow Corning) were mixed in a 1:1 (w/w) ratio with varying base to curing agent ratios of PDMS (10:1, 15:1, 20:1). The carbonyl iron/PDMS precursor solution was mixed thoroughly and degassed via centrifugation at 1000 x g for 1 minute. The physics of magnetic actuation and properties were characterized previously. The magnetic precursor was then poured in a well and the PVA template was placed on top of the magnetic precursor pool. The PVA template and carbonyl iron/PDMS precursor were placed in a vacuum chamber at 400 kPa vacuum pressure for 2 hours to allow for the magnetic precursor to completely penetrate the pores of the PVA template via capillary action. The PVA templates, with infiltrated carbonyl iron/PDMS precursor, were then wiped of any excess magnetic precursor and then baked at 65°C for 4 hours to allow for the carbonyl iron/PDMS precursor to cure within the template. The PVA template with infiltrated cured PDMS was then placed in a hot stir plate (above 70°C) for 3–4 hours to allow for the PVA to completely dissolve. Once the PVA is completely dissolved, the magnetic-PDMS porous composites could be removed and left out to dry.

Mechanical Analysis of Magnetic Porous Composites

Mechanical analysis of the magnetic-PDMS porous composites was tested by performing controlled axial compressions utilizing an ElectroForce 5500 (TA Instruments). The testing parameters including a controlled axial compression at 0.01 mm/s of cylinder composites of the gyroid infill pattern (8.5mm x 7.8 mm) and grid infill pattern with 90-degree layer offset (9mm x 7.8mm). The Young’s modulus of the composites were determined by the linear portion of the stress vs strain curve which was determined to be 2% to 10% strain.

Controlled Deformation Characterization of Magnetic Composites

Controlled deformation studies were performed to assess the ability to induce strain on the magnetic- PDMS composites with varying magnetic fields. A 3D printed compartmental box was printed to hold both a phone and 96-well plate at a defined distance and location to have reproducible image collection for both qualitative and quantitative strain analysis. Grid (6.8x5mm) and gyroid (7.2x5mm) PVA templates were prepared to make magnetic composite structures to fit inside a 96-well plate. 15:1 (PDMS base: cure) grid and 20:1 (PDMS base: cure) gyroid magnetic composites were fabricated and used for controlled deformation studies. The 3D printed box was designed to have a permanent magnet insert location. 3D printed blocks were placed underneath the permanent magnet to adjust the applied magnetic field to the magnetic composites within the 96 well plate. Applied magnetic fields of 0, 240, 325, and 415 mT were performed on the magnetic-PDMS composites in a 96-well plate with DPBS media. Image collection was obtained via a 12-megapixel f/1.8 camera iPhone XR, Apple). Applied magnetic fields were measured via a gaussmeter (TD8620 Handheld Digital Gauss TeslaMeter, China).

Dynamic actuation studies of the magnetic composites were performed using a 3D printed box that was designed to develop a dynamic magnetic field profile to magnetic composites within a 96-well plate. The box was built to hold two micro linear actuators that drive a stage, encompassing permanent magnets, in the z-direction to control the frequency, amplitude, and duration of an applied magnetic field experienced by the magnetic porous composites. A cyclic, 325 mT magnetic field was used for all the magnetic actuation studies. Magnetic porous composite recovery studies were performed by imaging the height of both gyroid (20:1) and grid (15:1) under no applied magnetic field after different duration of dynamic magnetic field actuation cycles (0, 5, 25, 50, 100). On- demand precise actuation analysis of the magnetic porous composites was executed by inducing a pause function at the maximum applied magnetic field during the dynamic magnetic actuation cycle for imaging the strain levels of the gyroid (20:1) and grid (15:1) for the same actuation cycle durations studied for the recovery tests.

Cell Culture of hUCMSCs and hADMSCs

Patient derived human umbilical cord stem cells (hUCMSCs) and adipose stem cells (hADMSCs) were provided by the Midwest Stem Cell Therapy Center (MSCTC). Cells were maintained in Gibco™ DMEM, high glucose, pyruvate media supplemented with 15% fetal bovine serum and 1% penicillin- streptomycin, all purchased from Gibco™, Dublin, Ireland. Cells were cultured in a humidified incubator at 37 °C and 5% CO₂. Cells were detached by the incorporation of trypsin-EDTA (Gibco™) followed by the addition of cell culture media, and then centrifuged for 5 minutes at 1500 rpm. Passages 3–5 of hUCMSCs were used in 3D fibrin gel preparation.

Fibrin hUCMSC Laden 3D Gel Preparation

Stock solution aliquots of fibrinogen (Millipore Sigma, US) and thrombin at concentrations of 45 mg/mL and 10 U/mL were used in the preparation of the fibrin gels. The aliquots were heated up to 37°C prior to preparation. The fibrinogen stock solution is then diluted with DPBS and mixed gently to give a final concentration of 22.5 mg/ mL of fibrinogen. DPBS is also added to the thrombin stock solution to yield a final concentration of 5 U/mL thrombin solution. The pelleted cells were then suspended and mixed carefully into the thrombin solution. 50 µL of the fibrinogen preparation is then added to the top of the porous constructs followed by the addition of the 50 µL of the thrombin-cell suspension. The gel is then mixed with the pipette tip immediately given the rapid polymerization and placed in the cell incubator for 15 minutes to ensure full polymerization. Cell culture media is then added on top of the gel and placed back in the incubator. The final fibrinogen and thrombin concentrations of the gels are 11.25 mg/ mL and 2.5 U/ mL respectively with 12,000 cells per gel preparation.

3D Cell Culture for Cell Viability Testing

Grid (3x 6.8mm) and gyroid (3 x 7.2 mm) cylinder PVA structures were printed to fit at the bottom of a 96-well plate well. Porous composites were fabricated as previously described with 8 different groups for cell viability analysis. The groups were represented by the two different infill patterns (grid, gyroid), two PDMS crosslinking ratios (10:1, 15:1), and both magnetic and PDMS porous constructs. For cell viability studies, the porous magnetic and PDMS constructs were sterilized by a 70% ethanol bath followed by UV exposure. The constructs were thoroughly washed with DPBS before insertion into a cell culture plate well.

Fibrin hADMSC laden 3D Gels were prepared as described for the hUCMSC laden 3D gel, while using hADMSCs, with a fibrinogen and thrombin concentrations of 10 mg/mL and 2.5 U/mL. Fibrin gels were cultured in both magnetic and PDMS only 3D porous well constructs. The 3D porous well design consisted of a 4 mm diameter and 2 mm tall well formed in a cylinder with 3 mm thick walls. The fibrin gels were cultured for 5 days in 48 well tissue culture plates.

For qualitative and quantitative analysis of the cell viability of the different porous construct groups and the no construct control group, a live/ dead-assay was performed using a live/dead imaging kit (LIVE/DEAD® Viability/Cytotoxicity Kit, Invitrogen™). The live/dead staining was performed on the cell cultures on the fourth day under cell culture conditions. The attached hUCMSC cell-laden fibrin gels attached to the porous constructs were removed from the 96 well-plate and flipped over and placed at the bottom of the wells for imaging. Given the high porosity and thin nature of the constructs, they provided sufficient transparency for imaging without removing the construct from the gel. As for the magnetic well cell culture studies, hADMSC cell-laden fibrin gels were removed from the well constructs with a spatula and placed in the bottom a 96-well plate for imaging.

For staining, 20 µL of 2 mM EthD-1 and 5 µL of 4 mM calcein AM from the LIVE/DEAD® reagents were added to 10 mL of DPBS. The resulting working dye was then vortexed to ensure thorough mixing and a final concentration of 4 µM EthD-1 and 2 µM calcein AM was achieved. Cell culture media was then aspirated from the well of each sample and 100 µL of the live-dead solution was added to each well. The well-plate was then covered with aluminum foil and incubated for 25 minutes. After incubation, the live/dead solution was aspirated from each well and each sample was washed with DPBS and imaged with a fluorescent microscope (EVOS, Invitrogen & Primovert, ZEISS). Qualitative and quantitative analysis was performed by using post-processing imaging software (Fiji, ImageJ).

Diffusivity Studies

To determine the viability of the ferromagnetic interface system to be used as a 360-degree interface system, the ability for waste and nutrients to diffuse through the interconnected micropores of the magnetic porous required investigation. Therefore, a diffusivity study of the composite interfaces was performed by developing a magnetic composite reservoir with a 5 x 5 mm cylinder well with 3 mm thick porous walls. The diffusivity would be measured by the absorbance of a 5% (w/w) alginate sol gel (Alginic acid, Fischer Scientific) with 15 mg/mL of aniline blue (Fischer Scientific) in the bulk fluid. The reservoirs were first submerged in 1 mL of PBS in 24 well plates and preconditioned via physical agitation to allow for PBS to enter the pores of magnetic porous composites. The alginate/ aniline blue sol gel was then loaded in a 1 mL syringe with a 26 G needle and 100 μL of solution was injected into each magnetic composite reservoir, followed by the addition of another 800 μL of DPBS. The cumulative mass transport out of the magnetic reservoir composites was measured by removing 200 μL for absorbance analysis of the well plate bulk solution outside the composite interface at specified time points (20, 40, 60, 80, 100, 120, and 150 min), 200 μL of PBS was added after each time point. Studies were completed in triplicates for both grid and gyroid reservoirs magnetic composites.

Cumulative Release Profile

A calibration curve was found by performing seven 1:1 serial dilutions of 1.75 g/ mL of aniline blue and then measuring the absorbance at 625 nm of 200 uL of each dilution with a plate reader (Synergy H1 Hybrid Multi-Mode Reader, BioTek). Absorbance levels for PBS were also measured to allow for the calculation of relative absorbance values from the diffusivity studies. Release profiles were found by measuring the absorbance at 625 nm of the aliquots taken at each time point with a plate reader. Relative increase in the absorbance profiles for each time point were then calculated by subtracting the measured value by the previous absorbance level accounting for the addition of 200 μL after each removed sample. Cumulative release profiles were then established by calculating the cumulative sum of the relative increase values.

Results

Facile and Versatile Fabrication of 3D Porous Ferromagnetic Interface

Combining 3D printing and particulate leaching methods, we were able to fabricate high-fidelity 3D porous magnetic composites, which has not been demonstrated elsewhere to our knowledge. We used PDMS to develop a soft porous magnetic substrate actuator for micromechanical actuation of 3D dynamic cell cultures⁴⁴. Both grid and gyroid infill patterns were investigated for precise control and high-fidelity fabrication of 3D magnetic-PDMS porous constructs. The grid pattern structure was chosen given its popularity in many different tissue types in the mesoderm lineage that have anisotropic cellular network formations and align in a linear fashion. As for substrate interface considerations, rectangular grid patterns have been widely studied to induce anisotropic cell morphology in both 2D static¹⁰ and time dependent 3D cell culture surfaces⁴⁵. The gyroid infill pattern was also investigated as triply minimal surfaces in many 3D printed porous biomaterial fabrication, which has showed good mechanical properties and allow for high permeability and nutrient and waste transport^{46, 47}. The facile fabrication with simple three steps were illustrated in Figure S1 A, which can produce versatile, precisely controlled microstructure patterns. The resulting 3D constructs were shown in Figure 1 D-F, which exhibited very uniform and regular 3D microstructure pores. The entire fabrication is straightforward, and only takes about twelve hours to complete with high throughput capabilities as many 3D magnetic composites can be developed at the same time. Most importantly, the microscale porous structures can be precisely control via a 3D FDM printer with feature sizes under 200 μm, 1 μm x-y axis precision, and layer resolution around 100 μm (Figure 1 G and I) which is well suited for dynamic cell culture actuation.

Figure 1.

3D Fabrication of Smart Responsive Ferromagnetic Porous Composites. (A-C) Illustrates the process of fabricating magnetic porous composites utilizing both 3D printing and particulate leaching techniques. (A) The first step includes 3D FDM printing a PVA sacrificial template which represents the pores of the final structure. (B) The sacrificial template is placed in a magnetic Sylgard 184 (PDMS) precursor pool in a 1:1 w/w ratio (PDMS: Carbonyl Iron) and desiccated to allow for the precursor to fill the pores of the PVA sacrificial template. (C) Magnetic PDMS precursor filled with PVA template is baked (65°C) for PDMS curing. The PVA is leached out or dissolved away via a warm water bath. (D-F) Display images of resulting 3D constructs from the fabrication steps from A to C. scale bar = 3 mm. (G, I) Show microscopic brightfield images of 3D microstructures of the PVA sacrificial template and the PMDS- magnetic porous composite (F). (E) Shows the filament 3D printer used (N2, Raise 3D) with a 400 μm nozzle with submicron precision in the x-y axis.

A range of infill volume percentages (30%, 50%, 70%) were also studied to analyze the potential bounds in controlling the porosity, mechanical properties, interface dimensions and geometries to influence cellular behavior. Microscopic images of 3D printed PVA sacrificial templates were shown in Figure 2A which exhibited the unique and precisely defined 3D microstructures. The casted magnetic PDMS composites replicate such 3D porous microstructures with excellent precision as demonstrated sequentially in Figure 2B. Figure 2C is the overview of resulting 3D magnetic PDMS composites with different porous geometries. We observed the PVA template with 70% infill volume could lead to robust and reproducible fabrication of 3D magnetic PDMS composites regardless of the porous geometries. In contrast, 50% and 30% infill volume both led to a certain degree of defects on both microscale (Figure 2B) and macroscale (Figure 2C) levels during the casting process, and 30% infill volume form PVA template showed much more structural and fractal abnormalities for both grid and gyroid constructs, making them unable to develop a fully intact 3D construct, particularly for gyroid-shaped constructs. Thus, these results indicate that the infill volume range in PVA template for the development of reproducible high-volume magnetic PDMS porous constructs is between 50–70%, which casts sufficient PDMS material for maintaining the mechanical strength in support 3D microstructures. Although higher infill volume in PVA template is attainable with larger magnetic- PDMS porous constructs (diameter > 1cm), we are more interested in developing 3D constructs suitable as the insert in 96 well plates for 3D cell culture employed in high throughput drug screening.

Figure 2.

Investigation on the percentage of infill volumes (30%, 50%, 70%) along with the infill patterns (grid and gyroid) in PVA sacrificial template for influencing on resulting 3D constructs. (A) Shows microscopic brightfield images of the microstructures of PVA templates and resulting magnetic-PDMS porous composites (B). The scale bar is 400 μm. (C) Macroscale images showing the overview of the magnetic-PDMS porous. The scale bar is 3 mm.

Due to the increased distance between strands, the liquid- air surface tension and adhesive forces were not sufficient for inducing capillary action to infill the PVA template volume with magnetic- PDMS precursor at lower infill percentages (< 50%). The inability to establish the full porosity range exhibited by Mohanty, S. et al. was likely due to the higher density of the carbonyl iron and PDMS precursor compared to the PDMS only precursor (980 vs. 4420 kg/ m³). This result can be described by the equation for capillary action Eq. (1), as the height that the liquid precursor can travel (h) will decrease with both an increase in distance between the strands (r) and density (ρ) with the assumption that there is little or no change in the liquid- air surface tension (у) and contact angle (θ).

$$h= \frac{2\gamma cos\mathrm{\theta }}{\rho gr}$$ (1)

However, this result is tolerable given higher percentage infill PVA templates lead to higher porous constructs. This is desired for interface and scaffolding in 3D cell culture systems, due to the ability for higher nutrient and waste transport from higher interconnected microchannels. The higher infill PVA templates also allow for casting smaller strand widths of the magnetic porous composites, which is advantageous when considering topographical contact points as micro and nanopattering has been shown to control the alignment of cells to induce the anisotropic cell formation experienced in particular tissue sets such as bones, muscles, and nerves. Such tissue sets have been extensively studied in vitro but is usually limited to a means for static control of 3D dynamic cell culture^8–10. Our developed platform supplies suitable topographical contact points and mass transport for interfacing 3D cell culture with remote-control of dynamic conditions, which could advance deep understanding in a variety of tissue engineering models. Future developments will study the usage of smaller FDM nozzles to explore the ability to fabricate 3D structures with smaller strand gaps to yield both more porous and smaller strand width magnetic porous composites.

Owing to the power of 3D printing, many complicated macroscopic designs can be achieved to cast the 3D magnetic PDMS porous composites with well-defined microstructures and high fidelity. These structural examples were included in Figure 3, which proves that broad applications can be developed. Small feature and high aspect ratio characteristics were obtained which is illustrated in the 2mm by 4mm well developed in a cylinder (Figure 3A) and the ability to form a void space within a cylinder with a reservoir construct (Figure 3B). Complex structures are also attainable with highly specific features shown in the hollow cylinder (Figure 3C) and KU logo (Figure 3D). These presented designs showcase the ability to develop complex structures at both the micro and macro scale that can be translated to the development of organ specific interfaces and scaffolds.

Figure 3.

Development of Complex Magnetic-PDMS Porous Composites with High Microlevel and Macrolevel Fidelity. The structures shown include (A) cell culture well, (B) reservoir on left with PDMS only reservoir on right shown to visualize the reservoir filled with an aniline blue dye hydrogel, (C) hollow cylinder, (D) KU logo.

3D Magnetic PDMS Porous Composites enable the Remote-controlled Actuation

The non-destructive, remotely controlled actuation in the field of tissue engineering is emerging, especially for inducing dynamic mechanical strain which could play a large role in the fate of mesenchymal stem cells^{35, 36, 48, 49}. Carbonyl iron microparticles were chosen to be embedded within PDMS given their high magnetization level with a low hysteresis which leads to a strong translational force in the direction of an applied field⁵⁰. When this combinational force is greater than the intrinsic stiffness of the PDMS then this will lead to a deformation of the bulk PDMS, which is why we observed that actuation capacity of our developed 3D magnetic PDMS porous composites is directly correlated to their intrinsic mechanical properties. We were able to control such mechanical properties by altering the stiffness of PDMS composites casted from 70% infill grid and gyroid PVA templates which possessed high-fidelity in small constructs with higher porosity to increase the strain potential. Three different base-to-cure cross linking ratios (10:1, 15:1, 20:1) were studied to yield less stiff PDMS composites at higher ratios. Strain-dependent uniaxial compression of the magnetic composites was performed to assess the Young’s modulus from all the samples shown in Figure 4A. The compressive data reveals the capacity to tune the mechanical stress of the porous magnetic PDMS composites from 10.17 ± 2.31 to 44.01 ± 1.08 kPa by altering the base to curing ratio from 20:1 to 10:1. There was no statistically significant difference found between the grid and gyroid patterned porous composites with the same base to curing agent ratio, which indicates that the porous composite stiffness is independent from these two infill patterns.

Figure 4.

Mechanical and Remote Actuation Analysis of 3D Magnetic PDMS Porous Composites. (A) Controlled compression analysis of 3D magnetic PDMS porous composites (1:1 w/w) under different base-to-cure cross linking ratios and microstructural patterns (mean ± SD, n=5). The 3D magnetic PDMS porous composites were casted from 70% infilled PVA template with 90-degree offset grid and gyroid patterns. (B-C) The microscopic side view of the controlled deformation degree of 3D magnetic PDMS porous composites in (B) grid pattern with 15:1 base: cure ratio and (C) gyroid pattern with 20:1 base: cure ratio. The external magnetic field is applied consistently under different strengths. (D) Quantitative strain data of both grid (15:1 base: cure ratio) and gyroid (20:1 base cure ratio) were obtained across a range of applied magnetic fields (mean ± SD, n=5).

The ability to alter the intrinsic stiffness of the magnetic porous composites should allow for a wide mechanical strain range, with the less stiff composites having the largest strain potential. Thus, we investigated the remotely controlled deformation on the lowest stiffness attainable grid (15:1) and gyroid (20:1) magnetic composites, respectively in Figure 4B and 4C. A 3D printed compartmental box was developed to have both a controlled applied magnetic field and reproducible image (Figure S2A and S2B). Such integrated setup would allow for both qualitative and quantitative magnetically actuated deformation analysis given the control over the position of the camera and magnetic composite along with the applied magnetic field. Both the grid (15:1) and gyroid (20:1) 3D constructs were fabricated to fit 96 well-plate and were submerged in DPBS to mimic actuation under cell culture conditions. The magnetic composites were then placed under applied magnetic fields of 0, 240, 325, 415 mT. As expected, macrolevel strain of magnetic composites in both grid and gyroid patterns across all the magnetic field levels were actuated, with higher strain levels while increasing applied magnetic fields. Shown in Figure 4D, the grid construct experiences 2.89 (+/− 0.41) %, 5.87 (+/− 0.36) %, and 7.71 (+/− 0.20) % strain actuation, while the gyroid constructs encountered 3.67 (+/− 0.22) %, 6.57 (+/− 0.20) %, and 11.00 (+/− 0.43) % strain actuation along with increased magnetic field intensities (240, 325, and 415 mT). The higher actuation potential of the gyroid constructs could be correlated to its lower stiffness at a 20:1 ratio, although the magnetic volume fraction of embedded magnetic particles was kept constant in all magnetic- PDMS composite groups.

Additional dynamic magnetic composite actuation studies were then evaluated to establish the on- demand and precise actuation capacity of the magnetic porous composites. A dynamic actuation compartmental box was developed to utilize two linear actuators to control the z-directional movement of a stage, encompassing an array of permanent magnetsm to control the dynamic magnetic field experienced by a magnetic porous composite (seen in Figure S2C). The first investigation studied the capability of both a gyroid (20:1) and grid (15:1) to reverse back to its original height after multiple different dynamic actuation cycles (5, 25, 50, and 100) with an applied magnetic field of 325 mT. The results shown in Figure S3A and S3B reveal that both the gyroid and grid were able to reverse back to within 0.1% of their original height across all the different dynamic actuation cycle durations. Also, when it came to assessing the precise control of the magnetic composite strain levels after numerous actuation cycle time points, the composites revealed a consistent strain response with only an average standard deviation of 0.11 and 0.16, respectively for the gyroid and grid structures strain levels. Cyclic strain actuation of a gyroid construct at 325 mT can be observed across a range of different frequencies (0.25, 1, and 2 Hz) in Supporting Information, Movie S1-3. In addition, hUCMSC laden fibrin gels were polymerized on the top of gyroid (20:1) magnetic composites, that were the same height (3 mm) of the porous composites used for cell viability studies, and actuated with a 325 mT to show the ability for the composites to induce a strain that could be translated to the 3D fibrin gel placed above. Figure S4 demonstrates that the magnetic composites deform (6.00% and 6.33%) with a cell laden fibrin gel under dynamic magnetic actuation conditions.

With the proposed magnetic porous composite bottom surface interface, the magnetic porous composite will induce a tensile strain utilizing similar design principles utilized by the commercially available Flexcell® Tissue Train® 3D Cell Culture System and other 3D tensile strain platforms^{51, 52}, which use fixed constraints to anchor the cell culture on select surfaces and have one or more stretchable interface substrates to induce a stretching event. For the particular system studied, the angular surface of the well plate acts as the anchor surface with the PDMS interface acting as the deformable surface that induces a uniaxial stretching event. The height of the 3D fibrin gel interfaced with the magnetic composite can also be modified to control the translatable strain experienced by the 3D gel as a thinner gel will result in a higher strain. The reproducible magnetic porous composite strain levels studied here are comparable with the compression and tensile mechanical stimuli used in controlling the differentiation of MSCs in 3D gels from reported actuation studies (2–20% strain) ^{19, 53, 54}, which indicates the feasibility for using our developed 3D magnetic PDMS composites as dynamic interface to induce physiological mechanical actuation.

3D Porous Ferromagnetic Interface Enables the 3D Dynamic Culture of hMSCs

Although micropatterning has been shown to control the alignment of hMSCs in 2D culture^{45, 55, 56},the 3D patterned dynamic interfaces have not been well studied due to the lack of capable dynamic micropattern platforms which is critically needed as an improved 3D culture system in mimicking in vivo MSCs cellular behavior. Our developed 3D magnetic PDMS porous composites are desired as both scaffolding and interface materials for precise remote-control of the dynamic mechanical environment in a 3D cell culture system. Figure 4 already demonstrated the potential using as the interface material by offering seamlessly comparable mechanical actuation to the 3D cell culture scaffolding environment. Herein, we further studied the biocompatibility and cell viability for interfacing with 3D cell culture system. The studies were implemented by placing a 3D cell-laden gel on the top surface of the magnetic porous materials to assess any influences on hampering the cell growth behavior due to the porous surface contacts or magnetic particle contaminations. Different porous construct groups (Figure S5), in terms of magnetic loading weight fractions to PDMS, infill patterns, and PDMS base to curing agent ratios, were used for cell viability studies shown in Figure 5. The cell-laden fibrin gel was used as the control group as the natural 3D scaffolding material.

Figure 5.

Cell viability of 3D cultured human umbilical cord-derived mesenchymal stem cells (hUCMSCs) with 3D porous ferromagnetic interface. The hUCMSCs were seeded in a 3D fibrin gel directly on (A) 3D magnetic-PDMS porous composites and (B) PDMS only 3D porous constructs with different patterns (grid and gyroid) and base to curing agent ratios (10:1 and 15:1). A 3D fibrin gel was cultured on a well plate surface as a positive control. Live/ dead staining was performed on the fourth day of cell culture (green- live; red- dead). (C) Quantitative cell viability (n=3) was accessed using imaging processing software (Fiji, Image J). Scale bar = 100 µm.

Patient derived human umbilical cord stem cells were chosen as the cell source owing to their sensitive response to various mechanical loading in culturing environment and be able to differentiate into adipose, bone, tendon, muscle, and cartilage linages depending on both the intrinsic and extrinsic mechanical environments provided^{6, 53, 54}. Fibrin gels with final concentrations of 2.5 U/mL thrombin and 11.25 mg/mL fibrinogen were polymerized on the top of magnetic composites and PDMS porous constructs. The live- dead two- color cell viability assay was performed after four days of cell culture. High cell viability (> 84%) was observed across all the experimental interface groups and were comparable to the positive fibrin control group (Figure 5C). Most importantly, both of the 10:1 magnetic porous composite groups exhibited cell viabilities above 89%. The results seemed to indicate that no carbonyl iron contaminates or detrimental effects due to the interface pattern were observed revealing a biocompatible, 3D porous ferromagnetic interface with remotely controlled, on-demand dynamic strain actuation function. Further studies will consist of observing the cellular behavior response due to both the mechanical actuation of the magnetic porous actuation and the anisotropic interface that is presented to the surface of the 3D cell culture.

We also implemented the specific 3D microstructure designs which is imperative for improving the waste and nutrient mass transport and exchange through the interconnected networks of the 3D porous ferromagnetic constructs. We introduced a reservoir interface design which could provide a dynamic micromechanical environment to all surfaces of a 3D cell-laden gel for highly controlled actuation. Both grid and gyroid patterned reservoirs were designed with 3mm thick, and porous walls surrounding a 5 × 5 mm reservoir space for housing an alginate sol gel (5%, w/w). Mass transport analysis was then determined by the ability for fluid to move from the gel through the exterior porous walls to the bulk fluid. The mass transport out of reservoir composite structures was quantified by incorporating aniline blue dye into the gels and measuring the absorbance of the bulk fluid at different time points shown in Figure 6A. As expected, an initial lag phase was observed as the dye in the gel needs to both dissolute and transport through the interconnected network of the reservoir walls to the bulk fluid. The gyroid structures emerged as the best micropattern design for waste and nutrient transport by showing a faster mean release profile than that of the grid pattern in Figure 6A. Because of the minimal surface designed in gyroid structures, the higher permeability is expected given the single connected and infinite domains. As a result, optimal interconnectivity is achieved given there are no sealed cavities in the structure⁴⁶. Although there are different initial penetration profiles, both gyroid and grid designs showed the capability for usage as interface and scaffold materials given their suitable interconnected networks for improving the mass transport. It is worth mentioning that fluid penetration through PDMS porous structure is very challenging at the initial penetration phase due to the limited wettability. However, mechanical actuation could improve this process to ensure the good fluid penetration before the loading of the gels. The collected release data was standardized to account for the gel loading variability. Through the compression based actuation, any unwanted pressure waves can be avoided for seamless transport¹⁹. Most importantly, the 3D sacrificial template allows for casting versatile pore shapes and precisely controlling interconnected porosity which could substantially improve the permeability and mass transport within the 3D constructs.

Figure 6.

Characterization of 3D porous ferromagnetic interface for 3D Dynamic Culture of hMSCs. (A) Analysis of the cumulative mass transport over time of an injected 3% percent alginate sol (15 mg/mL aniline blue) into a magnetic- PDMS composite reservoir with a 5 × 5 mm cylinder well and 3 mm thick porous walls. The relative cumulative release was measure on the absorbance (625 nm) of the PBS buffer outside the reservoir as a function of time. (B) Fluorescence microscopic images showing the cell viability after 3D stem cell culture inside of a magnetic-PDMS porous well composites or PDMS only porous well (C). Human adipose-derived mesenchymal stem cells (hADMSCs) were seeded in a 3D fibrin gel directly into the wells of both magnetic-PDMS (D) and PDMS only (E) well structures. Live/ dead staining was performed on the fifth day of cell culture (green- live; red- dead). Scale bar = 100 µm.

A 3D cell culture and viability study were further performed in this 3D well construct using human derived adipose mesenchymal stem cells (hADMSCs) which is a different type of mesenchymal stem cell source and more available than hUCMSCs used in Figure 5. The structure consisted of a 4 mm diameter and 2 mm tall well, which encompassed the 3D gel fibrin gel, and had 3mm thick, porous walls and the top of the 3D fibrin gel was exposed to the bulk fluid. A live- dead two- color cell viability assay was performed after five days of cell culture to further demonstrate the suitable cell viability of the 3D cell cultures in the well constructs due to excellent waste and nutrient transport. There is no noticeable difference in the cell viability amongst the cells cultured in the magnetic (Figure 6B) or PDMS-only constructs (Figure 6C). Thus, the high cell viability under the well construct culture conditions indicate the feasibility and potential for dynamic 3D cell culture.

Conclusion

Magnetic stimuli responsive materials have been emerging as the next generation of dynamic cell culture platforms. These materials allow for fast, highly tunable actuation at both the macroscopic and microscopic scale without any external pressure systems and tethered electronics, leading to the ease of integration into high-throughput systems such as 96 cell culture well plates and microfluidic chips. However, current existing bioreactor systems using such magnetic responsive materials are still lacking the precision in microenvironment control, or in low throughput and bulky size. In this research, we overcame such limitations by developing a novel magnetic elastomer cell culture platform and using a porous magnetic PDMS composite with the full tunability of microstructures. The fabrication of the 3D magnetic PDMS composites are simple and straightforward with high-fidelity control of microlevel and macrolevel structures, which can reproduce strain levels suited in 3D stem cell culture and differentiation. The porous magnetic PDMS composite also demonstrated biocompatibility given the high cell viability experienced for 3D culturing both hUCMSC and hADMSCs. The high interconnected porosity of the porous structure also supports the development of a multiple surface contact design for more complex mechanical actuation given improved fluid permeability and mass transport. Our results reveal a promising biocompatible, remotely controlled magnetic PDMS composite interface material with full control over the porosity, microlevel and macrolevel structure, mechanical properties, and dynamic strain for tunable dynamic cell culture studies at the scale. For future studies, multiple embodiment designs will be established to look at both tensile and compression actuation and the ability to control the differentiation of hUCMSCs. The magnetic actuation device will be developed to remotely control the strain levels of an array of porous magnetic-PDMS composites in a cell culture plate to induce a well-defined dynamic mechanical environment in high throughput employed in drug screening.