A Thermoresponsive and Magnetic Colloid for 3D Cell Expansion and Reconfiguration

A dual thermoresponsive and magnetic colloidal gel matrix is described for enhanced stem-cell culture. The combined properties of the material allow enzyme-free passaging and expansion of mesenchymal stem cells, as well as isolation of cells postculture by the simple process of lowering the temperature and applying an external magnetic field. The colloidal gel can be reconfigured with thermal and magnetic stimuli to allow patterning of cells in discrete zones and to control movement of cells within the porous matrix during culture.


DOI: 10.1002/adma.201403626
cell volumes has yet to be reported. Moreover, the fi xed internal structures of some of these materials pose technical challenges, making it diffi cult to seed cells in controllable locations or to passage cells at specifi ed time points, as well as to harvest the cells effi ciently after expansion. For example, cells are usually seeded into support matrices under static or low-fl ow conditions that rely on diffusional processes and weak adhesion forces, resulting in inhomogeneous cell localization and growth. [ 12 ] Cell seeding under dynamic fl ow conditions is possible using industrial bioreactors with built-in perfusion systems, [ 13 ] but this is more problematic when manipulating the small amounts of primary cells available from donors. Other technical challenges for cell manufacture include the requirement to use proteolytic enzymes to detach cells from support materials during passaging or after cell expansion. Enzymes such as trypsin can cause cleavage of membrane proteins and growth factor receptors thus altering the proteomic profi le in mammalian cells. [14][15][16] There is accordingly a need for new materials that can support expansion of clinically relevant cell types, which can be easily manipulated to allow cell recovery, and which can be reconfi gured on demand to enable patterning of varying cell types. We report here a dual functional colloidal gel system, which combines the advantages of allowing homogeneous cell seeding, eliminates the need to use proteolytic enzymes during passaging, and enables placement of cells in discrete environments which can be patterned magnetically. Furthermore, the methodology is inherently scalable, allowing culture in volumes ranging from those appropriate to primary cells for single patient use through to stem cell manufacture.
Approaches to fabricate 3D cell support matrices include chemical cross-linking, solvent casting, particulate leaching, gas forming, and phase-separation emulsion or freeze-drying techniques. [ 17 ] Although each method has certain advantages, there is no single technique that can be used to produce material assemblies that address all the fundamental problems linked to cell seeding, passaging, and harvesting. The confl icting requirements of a matrix that is mechanically strong enough to support cell growth and manipulation, yet easily dismantled to allow cell recovery rule out the use of a single material type. However, stimuli-responsive polymers offer a means by which materials properties can be switched, [18][19][20] and in previous papers we have shown that it is possible to grow fi broblast cells in a reversibly assembling colloidal gel, utilizing a thermoresponsive polymer to control aggregation and disassembly of a matrix around a cell population. [ 21,22 ] Nevertheless, the separation of cells from these gels was not effi cient enough over repeated passages for stem cell expansion, and there were no means to move one region The ability to harness cells as therapeutic products remains a scientifi c challenge of major signifi cance. [ 1 ] Recent studies and clinical trials have shown the potential of cell-based therapies for the treatment of a range of conditions including cardiovascular, [ 2 ] neurologic, [ 3 ] and autoimmune diseases. [ 4 ] However, to transfer cell therapies from the research laboratory toward practical application where cells are needed in high volumes, new manufacturing tools are required. A variety of techniques have been used to grow cells in vitro with increasing emphasis on 3D matrices in order to overcome the inherent diffi culties associated with 2D culture. [ 5 ] In comparison to the standard 2D systems, 3D matrices provide higher surface areas to support growth to larger cell volumes and produce microenvironments which better mimic in vivo systems. [ 6,7 ] Recent advances include the use of 3D printing technologies to create cell patterns in a controlled fashion [8][9][10][11] and there are several marketed products to allow formation of 3D cell supports such as Matrigel, Alvetex, QGel, and Biomerix. Although these matrices are widely used in research settings, the potential for their use in cell manufacturing at high of a cell-containing gel next to another of a different cell type without mechanical micro manipulation. We now show how the combination of an amphiphilic thermoresponsive material with polymer microparticles containing superparamagnetic cores can be used to create a 3D support matrix that allows not only effi cient stem cell expansion over several passages but also cell pattern formation as a fi rst step to tissue-type biological organization. The material can be rapidly and reversibly transformed from a dispersed suspension into a solid matrix by temperature elevation only, thus enabling the seeding of cells in a homogenous fashion, which can be quickly disassembled to release cells by a small reduction in temperature, and then the cells can be simply separated from the colloidal particles by a magnetic fi eld. We further demonstrate that by combining the reversible assembly with magnetic fi eld manipulation, the colloidal gel particles can be used to pattern co-cultured cells into a variety of architectures and that the 3D matrix can be used to expand stem cells in an effi cient manner while maintaining the desired stem cell lineage during multi ple cell growth and harvesting cycles.
The fi rst part of the strategy involved the synthesis of a thermoresponsive polymer and magnetic microparticles, to be mixed together to form the 3D scaffold. We chose to fabricate the microparticles using polystyrene, as this material is already widely used in cell culture assays. Dispersion poly merizations of styrene (PS) with divinylbenzene (4 mol%) generated lightly cross-linked PS. [ 23 ] Embedding of iron oxide (Fe 3 O 4 ) in the surface of the microparticles was carried out by further polymerizations of styrene, in the presence of Fe 3 O 4 powder stirred with the particles, and then by a fi nal polymerization of styrene to cap the microparticles. We targeted the sizes of the magnetic polystyrene microparticles (MPSMs) to be larger than 1 µm to avoid possible internalization by cells. Scanning electron microscopy (SEM) images of the resultant Fe 3 O 4 -loaded polymers showed spherical microparticles with diameters in the range of 2-2.5 µm ( Figure S1, Supporting Information). Thermogravimetric analysis showed an approximate iron oxide content of 15% (w/v) ( Figure S2, Supporting Information).
For the thermoresponsive polymer, a dodecanethiol chain transfer agent was used in conventional free radical polymerization of 2-(2-methoxyethoxy) ethyl methacrylate (MEO 2 MA). [ 24 ] Proton nuclear magnetic resonance ( 1 H NMR) and gel permeation chromatography (GPC) confi rmed the structure of the polymer ( Figure S3, Supporting Information), with average number molecular weight of ( M n ≈ 20 kDa). The provision of a hydrophobic dodecyl chain end was designed to enhance adsorption of the otherwise hydrophilic poly(MEO 2 MA) to the hydrophobic MPSMs. To form the temperature reversible 3D colloid, the dodecyl-terminated polymer (DD-pMEO 2 MA) was physically mixed with MPSMs at mass ratios which enabled the colloids to be handled with ease at room temperature and yet be self-supportive gels within a few seconds of heating to 37 °C ( Figure 1 ).
Following the synthesis steps, the ability to seed and culture cells within a 3D matrix formed by the self-assembling colloidal gel was evaluated. Ratios of MPSM (25% w/v in culture media) and (DD-pMEO 2 MA) (4% w/v in culture media) were found to provide the best balance of mixing and rapid reversible gelation properties. We fi rst seeded immortalized bone marrow-derived human mesenchymal stem cells (hMSCs), which were genetically modifi ed to express green fl uorescent protein (GFP hMSCs), in order to visualize the localization of the cells within the 3D gel and also to monitor cell proliferation qualitatively. Two methods were used to seed the cells onto the matrix, in order to simulate use by individuals in a research setting, or for a case in which automated systems for highthroughput manufacture would be used. For the fi rst method, a suspension containing approximately 2 × 10 5 GFP hMSCs was gently mixed with a known volume of the matrix in liquid state, followed by dropwise addition into prewarmed cell culture media (37 °C), ( Figure 2 a, top and bottom). In the second method, the mixture of cells and particles was fi rst added onto a dry slide (coated with polytetrafl uoroethylene (PTFE) to produce a superhydrophobic surface), to form spherical droplets, and allowed to set in an incubator at 37 °C for 1 min before transferring into prewarmed media. As apparent from Figure 2 , these methods enabled the size of the cell-support matrices to be readily manipulated by controlling the initial volumes, for example 10, 20, 50, and 100 µL (Figure 2 b, top and bottom). The schematic representations (Figure 2 c, top) were confi rmed by the fl uorescent images showing a homogeneous distribution of cells throughout the matrices (Figure 2 c, bottom).
Having established the ability to seed cells uniformly into the 3D matrices, we investigated the use of the colloidal gel to support repeated expansion of cells. After seeding 5 × 10 4 cells into 50 µL of colloidal gel (formed on the PTFE surfaces), cell growth was monitored by recording fl uorescence images up to day 16 of incubation, using rhodamin-labeled MPSMs to provide contrast against the GFP-expressing cells. Figure 3 indicates that GFP hMSCs expanded progressively as compared to initial cell seeding. Furthermore, to investigate whether the cells grew equally throughout the outer and the core region of the gel, the matrices were sectioned after day 16 ( Figure S4, Supporting Information). The fl uorescence images from different sections of the matrices revealed uniform cells growth, most likely due to the large pores created by the particles which allows for rapid provision of oxygen and nutrient or removal of cellular wastes in and out of the matrix.
Having demonstrated cell proliferation using GFP-labeled MSCs, we sought to expand the investigation to use hMSCs. The same method to seed hMSCs was used as mentioned above; however, in this case cell expansion was monitored using a standard method (PrestoBlue). It can be seen from Figure 3 g that the hMSCs (25 × 10 3 cells in 25-µL matrix) showed a linear proliferation profi le up to day 10, at which point the experiment was terminated to draw a comparison with HMSCs cultured on 2D plastic surfaces. Equal amounts of cells were initially seeded on both the 3D matrices and 2D culture plastic; however, the cell growth on the 2D was comparably lower than in the 3D gel. As expected, the total cell numbers on the 3D matrices were higher compared to cells grown on the 2D plastic. For example, the total number of cells grown on the 3D matrices was in the range of 119 × 10 3 (4.7-fold increase) and 200 × 10 3 (eight-fold increase) at day 5 and day 10, respectively. However, for the same seeding density on the 2D culture plastic, the total cell numbers were 45 × 10 3 (1.8-fold increase) and 87 × 10 3 (3.48-fold increase) at day 5 and day 10, respectively. These data suggested that the 3D colloidal gel not only supported proliferation of stem cells but also generated a high cell volume compared to classical 2D plastics, thus establishing a key practical advantage of using a 3D matrix system.
A potentially signifi cant second advantage of a reversibly assembling colloidal gel is the ability to passage cells without using a trypsinization process or other biochemical cell detachment method, a required step to detach cells from most substrates including current 2D and 3D matrices. As indicated earlier, the colloidal 3D matrices can reversibly be transformed from liquid-suspension to solid-like matrices through a simple and also rapid temperature modulation. Cell growth and passaging experiments were carried out using the same cell seeding and incubation process as before; however, the HMSCs were passaged at day 5 of incubation, and were Adv. Mater. 2015, 27, 662-668 www.advmat.de www.MaterialsViews.com Figure 2. The generation of 3D matrix seeded with cells. a) Suspension of cells and matrix in liquid state (below 37 °C) mixed and added dropwise into a prewarmed cell culture media (37 °C); b) suspension of cells and matrix in liquid state added dropwise on a dry slide (coated with polytetrafl uoroethylene to produce superhydrophobic surface), forming droplet-like cell-seeded matrix, which was then allowed to set in incubator at 37 °C for 1 min before transferring into prewarmed media. c) The size of cell-seeded 3D matrix was readily manipulated by controlling the initial volumes of 10, 20, 50, and 100-µL depositions, which resulted in c) homogeneous cell distribution of GFP-labeled ihMSCs within the matrix. Scale bar = 100 µm.
then further incubated for an additional 5 days. As shown in Figure 3 h, the total number of cells measured at day 5 was 119 × 10 3 , and these cells were then passaged. It is important to note that cells did not need to be separated from the matrices at the passaging point. The passaging step was performed in a sequential manner. First, the nutrient media were removed from the cell culture "parent matrix" by gentle aspiration. Then the parent matrix gel was returned to the fl owable colloidal suspension state by cooling the system to room temperature. An aliquot of freshly prepared particle suspension (containing no cells), equal in volume to the parent matrix, was added and the whole suspension was gently mixed. Following this addition, the resulting cell/particle-mixed suspension was split into two equal volume aliquots or "daughter matrices", such that the concentration of the cells in the suspension was half that of the original parent matrix. Finally, the mixed matrices, still in liquid phase, were split into two equal volumes to create daughter matrices, and these were further incubated as described above.
For application to stem cell manufacture, expansion on a support matrix must maintain the desired phenotypic signature, characterized, for example, by the specifi c immunophenotypic surface markers expressed by MSCs. Accordingly, the surface markers CD90, CD73, and CD105 were used for positive hMSCs identity and CD45, CD34, CD19, CD11b, and HLA-DR were used as negative markers. [ 25 ] Cells were analyzed before and after 5-days incubation in the 3D gels, using standard fl ow cytometry protocols. Analysis indicated (Table S1, Supporting Information) that the surface markers remained unchanged before and after inoculation on 2D tissue culture plastic (TCP) and the 3D matrix, and cells were adherent to standard culture plastic. For both the 2D TCP and the 3D matrix, as well as both pre-and postinoculation, >99% of the cell population were positive for CD90, CD73, and CD105, while the numbers of cells displaying the markers CD45, CD34, CD19, CD11b, and HLA-DRA were <2%. These results suggest that the 3D matrix may be advantageous for the expansion of clinically important cell type such as MSCs.
In order to accomplish the complete process from cell seeding to harvesting, we examined whether the expanded HMSCs could be effi ciently separated from the matrices to Adv. Mater. 2015, 27, 662-668 www.advmat.de www.MaterialsViews.com Figure 3. The expansion of hMSCs in 3D matrix and enzyme-free passaging. a-f) Cells expansion of GFP-tagged hMSCs seeded within rhodaminlabeled 3D matrix, fl uorescent images were recorded on a) day 1, b) day 3, c) day 5, d) day 10, e) day 13, and f) day 16, showing progressive increase in cell proliferation, compared to initial cell seeding at day 1. g) Proliferation of unlabeled hMSCs as monitored by standard PrestoBlue assay up to day 10, cells proliferated on 3D matrix (solid line) compared to tissue-cultured plastic (dashed line) for the same initial cell seeding density and incubation time. h) Proliferated cells on 3D matrix were passaged at day 5 (without using any enzymes) by liquefying the "parent" matrix, followed by addition of equal volume of new matrix (without cells) into the parent matrix and subsequently formed two new "daughter" matrices (formed on PTFE slides). The daughter matrices were further incubated to day 10; at this point, the total cell numbers in daughter matrices were equal to the patent matrix before passaging. Error bars are ± standard deviations. Scale bar = 100 µm. recover cell populations in high numbers at the end point. The separation and recovery process is as illustrated in Figure 4 a, with sequentially recorded digital images in Figure 4 b, showing that the matrix transformed from a gel to a liquefi ed suspension from which the cells were separated by applying an external magnetic bar in close proximity. The total recovery of hMSCs was also quantifi ed at days 5, 10, and 15 using the PrestoBlue assay (Figure 4 c). This was an essential step not only to quantify the cells but also to ensure the cells retained full viability. These data demonstrated a consistently high cell recovery of 93.5%, 95.6%, and 92.1% from the matrices at days 5, 10, and 15, respectively. By contrast, poor cell recoveries were observed when a conventional centrifugation process was applied to the gel (data not shown). These data indicated the critical importance of combining the magnetic core component of the particles with the thermoreversible shell; in this way, the colloidal gel could be easily separated from the cells using a low shear stress magnetic separation step after cooling, thus enabling recovery of cells rapidly and in high volumes.
Having established some practical advantages for stem cell expansion, we investigated the possibilities for patterning cells within discrete zones in 3D, as a fi rst step to tissue modeling in vitro. Cultured 3T3 fi broblasts expressing RFP (red) and GFP (green) were incorporated into the 3D matrices, at cell densities of 1 × 10 6 per 20 µL of the colloidal particle suspension. By moving a magnetic bar underneath free fl owing bead suspensions in culture media, followed by in situ gelation, it was possible to form patterned gels in layers in partial mimicry of the organization of cells in normal tissue. As shown in Figure 5 a (left and right), patterns were generated comprising: i) multi ple regions of GFP 3T3 cells with a single RFP 3T3-seeded matrix in the center (left) and ii) a two-layered pattern with alternating zones of RFP and GFP-expressing fi broblasts (right). We also investigated the reconfi guration of cell patterns, using the magnetic fi eld as a noninvasive method, and were able to demonstrate dynamic mixing of the cells after culture. RFP 3T3 and GFP 3T3 cells were incorporated into magnetic and nonmagnetic matrices, respectively. We then aligned the matrices together and applied repeated magnetic fi elds bar by sequential movement of a magnetic bar from side to side. The fl uorescence microscopy images revealed a progressive reconfi guration of the matrices from discrete colored zones into a single region containing both cell types. Cell viability was unaffected by the reconfi guration process as determined by Presto-Blue assay (data not shown).
These experiments show that the combined thermoresponsive and magnetic colloid can be used not only to expand clinically important cell types such as MSCs, enabling passaging without using animal derived products such as trypsin and achieving high cell recoveries, but also to pattern cells into discrete regions and subsequently reconfi gure these patterns in a noninvasive manner. The colloidal gels are formed from inexpensive and readily available materials, and are compatible with standard cell culture assays. We believe these materials should fi nd use in manufacturing of cells through providing architectures/enviro nments similar to those occurring in vivo, and also in tissue modeling, where specifi c placement of different cell types is necessary to recapitulate the cell organization of normal tissue. Further development of this technology could defi ne a new platform 3D matrix, adaptable for a range of automated cell manufacturing, tissue modeling, and in vitro in vivo correlation assays, as desired.

Experimental Section
Materials : 2-methoxyethoxy)ethyl methacrylate polymer (DD-pMEO 2 MA) and MPSMs were prepared according to literature procedures as detailed in the Supplementary Information. All other chemicals were purchased from Sigma-Aldrich or Fisher Scientifi c and used without further purifi cation. All the solvents were HPLC grade, purchased from Sigma-Aldrich and used without further purifi cation.
Adv. Mater. 2015, 27, 662-668 www.advmat.de www.MaterialsViews.com www.advmat.de www.MaterialsViews.com Cells and Cell Culture : Bone marrow-derived human mesenchymal stem cells (hMSCs) were purchased from Lonza. The cells were received at passage 2 and were expanded until passage 4, and were then used for experimentation. For expansion, cells were cultured in Lonza hMSC medium at 37 °C in a humidifi ed atmosphere of 5% CO 2 . To visualize proliferation of hMSCs within the 3D matrix (Figure 3 ), the cells were immortalized and GFP tagged as described previously. [ 26 ] Similarly, mouse 3T3 fi broblasts, of the NIH3T3 strains, were also modifi ed separately to express RFP and GFP, for use in cell patterning experiments ( Figure 5 ), as previously reported, [ 26 ] whereas hMSCs, unlabeled, were used to investigate cell expansion, passaging, and recoveries (Figure 3 g,h and Figure 4 ).
Matrix Sterilization : The DD-pMEO 2 MA polymer was predissolved in cell culture medium (4% w/v) and fi ltered through a 0.2-μm sterile fi lter at ≈5 °C. The MPSMs were contained in a glass vial, placed on a rotating shaker, and sterilized under UV light (260-nm wavelength) for 60 min.
Matrix Preparation : Following the sterilization process, the matrix was prepared by mixing MPSM (25% w/v) in DD-pMEO 2 MA polymer (4% w/v) (predissolved in cell culture media). The components were thoroughly mixed and then refrigerated until further use.
Cell Seeding into 3D Matrix : We devised two protocols to seed cells into the 3D matrix. In the fi rst method, cells were counted and resuspended in media (20 µL). The cells were gently mixed with the DD-pMEO 2 MA/MPSM suspension and added dropwise into prewarmed media (37 °C), instantly forming a cell-seeded 3D colloidal gel matrix. In the second method, the cells and matrix mixture were added dropwise onto the surface of PTFE-coated slides, forming a droplet that, when briefl y incubated (1 min) at 37 °C, set in place to form the cell-seeded colloidal gel. This gel was subsequently transferred into prewarmed cell culture media. The two complementary methods were designed to demonstrate applicability both for research use in a lab setting and for highthroughput-automated culture formats appropriate for cell manufacture.
Cell Viability and Proliferation : Cell viability and proliferation were measured using a standard PrestoBlue assay according to the manufacturer's instructions (Invitrogen). Briefl y, hMSCs were seeded into the 3D colloidal gel at 25 × 10 3 cells per 25-µL volume of the matrix (formed on the PTFE surfaces as described above). To quantify the number of proliferated cells, the media were fi rst removed from the matrix, and the cells were released from the matrix by cooling to room temperature. The cells were mixed with 10 µL of PrestoBlue reagent and incubated for 30 min. A standard calibration curve was used to quantify the total cell numbers. To draw a comparison between tissue culture plastic (TCP) and the 3D matrix, hMSCs were also seeded (25 × 10 3 cells per well, 24 well plate) and quantifi ed following the protocol as described above. The absorbance of the colloidal gel components (cell free) was also measured with the PrestoBlue reagent as a further control reading.
Enzyme-Free Cell Passage : The proliferated cells in the matrix were passaged by removing the cell culture media and cooling briefl y to room temperature to allow the colloidal particles to fl ow and thus to release the cells. At this point, an equal volume of new colloidal particle suspension (without cells) was added to the "parent" mixture and gently mixed. Following this addition, two new cell-seeded "daughter" matrices were formed following the same protocol as described above (formed on the PTFE surfaces).
Cell Harvesting : Following proliferation, the cell-seeded matrices were briefl y cooled to room temperature to release the cells. Subsequently, the cell and matrix mixture was drawn into a glass tube and exposed to an external magnetic fi eld (10-mm bar magnet) in close proximity. The magnetic microparticles were separated, leaving cells only in the culture media. Finally, the cells were centrifuged at 200 g for 5 min to produce cell pellets and resuspended in fresh media, ready for subsequent analysis.
Flow Cytometry : Immunophenotypic analysis of the hMSCs was determined by fl ow cytometry before and after proliferation in the colloidal gel. This was performed using a BD Stemfl ow hMSCs analysis kit and BD LSR II fl ow cytometer. Cells were prepared for analysis following the manufacturer's instructions (BD-Biosciences, UK). In the case of cells proliferated in the 3D matrix, cells were fi rst magnetically separated as described above. Mouse antihuman monoclonal antibodies CD90 FITC, CD105 PerCP-Cy5.5, and CD73 APC were used to target cell surface receptors for positive identifi cation of hMSCs, while CD34, CD45, CD19, CD11b, and HLA-DR were used for negative Figure 5. The cell patterns and reconfi guration in 3D. Cultured 3T3 fi broblasts expressing RFP (red) and GFP (green) were seeded into the 3D matrices, at 1 × 10 6 cells per 20 µL of the matrix. a) Fluorescent images showing cell patterns in 3D formed as multiple regions of GFP 3T3-seeded matrices with a single RFP 3T3-seeded matrix in the center (left), (right) two-layered patterns with alternating zones of RFP-and GFP 3T3-seeded matrices. b,c) Reconfi guration of cells patterns shown in b) bright fi eld and c) fl uorescent images, in which reconfi guration of the cell patterns was demonstrated by seeding RFP-and GFP-3T3 cells into magnetic and nonmagnetic 3D matrices, respectively. The matrices were sequentially aligned in green-red-green patterns, and external magnetic bar was applied and repeatedly swiped from side-to-side. Fluorescent images were taken at 0, 10, 20, and 100 swipes, showing progressive reconfi guration of the GFP-and RFP-seeded matrices from discrete colored zones into a single region containing both cell types. Scale bar = 2 mm.