Scaffold-free three-dimensional cell culture utilizing micromolded nonadhesive hydrogels

Napolitano AP, Dean DM, Man AJ, Youssef J, Ho DN, Rago AP, Lech MP, Morgan JR.
Brown University, Providence, RI 02912, USA.

Results and Discussion

Micromolds for the directed self-assembly of microtissues were flexible, transparent, and could be autoclaved and re-used (Figure 1A). Agarose hydrogels cast from various micromolds (designed for 6-, 12-, or 24-well plates) formed precise negative replicates of their respective molds. Single-cell suspensions of NHF were added to the seeding chamber of the micromolded agarose where they sank into the recesses forming a sheet that conformed to the bottom of the well within 20 min (Figure 1, B and C). Since the hydrogel surface is nonadhesive for cells, cell-to-cell binding was promoted and, in

Horizontal microscopy revealed the vertical aspect of self-assembly. Single-cell suspensions of NHF and H35 were seeded into micromolds of polyacrylamide with a single row of recesses (Figure 1, D and E, respectively). Cells settled onto the bottom of the recesses and were funneled together due to the concave design of the recess bottoms maximizing cell-cell interactions. After 24 h, both cell types formed microtissues that appeared spherical via conventional microscopy, but horizontal microscopy revealed a significant difference. NHF spheroids were near perfect spheres, while H35 spheroids were shorter in the z dimension than the x, y dimension.

To evaluate the ability to control spheroid radius as a function of cell seeding density and to measure homogeneity of spheroid radius within relatively large batches, NHF and RG2 cells were seeded, (two gels per cell type, each containing 822 recesses/gel) at various seeding densities. After 1 (RG2) or 2 days (NHF) of self-assembly, image analysis was used to determine the radius distribution (Figure 2, A and B). Spheroid radius was proportional to seeding density, and spheroids were monodispersed about the mean radius. More cells settled into recesses along the periphery of the gel seeding chambers, resulting in larger spheroids. Radius distribution was measured across the entire gel. NHFs seeded at 0.25, 0.5, and 1 × 106 cells/gel resulted in spheroids with average radii of 53.3 µm with a coefficient of variation (CV) of 15.2%; 68.8 µm with CV of 8.7%; and 85.8 µm with CV of 10.6%, respectively. RG2s seeded at 0.18, 0.48, and 0.84 × 106 cells/gel resulted in spheroids with average radii of 63.0 µm with CV of 22.8%; 87.0 µm with CV of 23.8%; and 107.9 µm with CV of 17.2%, respectively. These CVs are comparable to published results of methods in which known numbers of cells are pipeted into individual hanging drops where CVs ranged from 10% to 15% for HepG2 cells and 5% for MCF-7 cells (18). The trend toward higher CVs using our method could be due to differences of when the CVs were measured (5–16 days after seeding versus 1–2 days in the present study), which could reflect differences in the short-term processes of self-assembly versus growth in 3-D. It could also be due to differences in the behavior of different cell types (e.g., HepG2 vs MCF-7) or to partitioning of cells into individual wells by a settling process as occurs in our method. More experimentation is needed to distinguish these possibilities.

To measure the growth of microtumor spheroids, MCF-7 cells were seeded at various densities (0.5, 1.0, and 1.5 × 106 cells/gel), and growth was measured by the WST-1 assay (Figure 2, C and D). MCF-7s seeded over a range of cell densities and self-assembled for 24 h yielded a standard curve with strong linear correlation (r2 = 0.953) between cell number and optical density within the range of interest (0.5 to 8 × 106). Spheroids displayed linear growth for the first 3 days, followed by slower growth from days 3–5. The slope of the growth curve for the initial 3 days was independent of initial spheroid size. We also assessed the smallest number of MCF-7s that could form a spheroid in this mold design. Small spheroids of MCF-7 cells could be self-assembled reliably (>95% of the recesses containing a single spheroid) by seeding as few as 3.1 × 104 cells/gel (approximately 38 cells/recess).

To determine if microtumor spheroids could be grown through clonal expansion of single cells, micromolds were seeded with a low cell density of H35 cells (approximately 800 cells/gel with 822 recesses) (Figure 2E). Over a 21-day period, spheroids (300-µm diameter) readily grew from a single cell. This method is especially useful for cells that are proliferation competent in aggregate culture, such as tumor cells or possibly stem cells, because it promotes aggregate growth and allows for quantification of proliferation.

To establish their ability to guide the self-assembly of cellular structures with complex geometries, micromolds with rod, toroid, loop-ended dog bone (pair of toroids connected by a rod), and honeycomb (a lattice of toroids) recesses were seeded with H35 or MCF-7 cells (Figure 3, A and B) or NHF (see Supplementary Movie S1 available online at Time-lapse microscopy of NHF revealed the process of toroid self-assembly (see Supplementary Methods). Cells sank into the toroid recesses, conformed to the bottom of the well, and immediately began to aggregate. NHFs assembled into a torus-shaped structure that wrapped tightly around the central hydrogel peg and decreased in outer diameter with time. H35s self-assembled into rod and loop-ended dog bone structures. The loop-ended dog bones adhered tightly to the two hydrogel pegs on the outer edges. On the inner edges, however, the dog bones pulled away from the pegs forming a cleavage, indicating that the structure is under tension. MCF-7 cells readily self-assembled toroid and honeycomb structures. In micromolds with toroidshaped recesses, the microtissues were wrapped around the protruding hydrogel pegs. In micromolds with honeycomb-shaped recesses, the self-assembled honeycomb was held on the micromold by contact with the outer edges of the outermost hydrogel pegs. From these points of contact, the cellular honeycomb appeared to evenly distribute tension in the structure such that a hexagonal geometry was approximated and most of the inner pegs were not in contact with the cellular structure. After 5 days, the honeycomb microtissues were removed from the hydrogel and cultured on agarose coated plates for 3 additional days. The branches became slightly shorter and wider, as if removed from tension, but the honeycomb shape was maintained.

To demonstrate the ease with which micromolds can generate multilayered microtissues, various cell types were labeled with fluorescent tracking dyes, mixed, and seeded into hydrogels (Figure 3, C and D). NHF/H35 and NHF/HUVEC mixtures were assembled into honeycombs and loop-ended dog bones, respectively. Epifluorescent and confocal microscopy revealed that cells also self-sorted in these structures, forming multilayered microtissues. Within each of the mixtures, the NHF (red) formed the core structure, while the H35 or HUVEC (green) formed an epithelial-like outer coating.

Tritypic spheroids were generated by seeding a 1:1:1 mixture of labeled NHF (green), HUVEC (red), and H35 (blue) cells (Figure 3D). Cell sorting occurred in the trytipic spheroids with an outer layer of H35s engulfing interspersed islands of NHFs and HUVECs. The sorting of only two different cell types within a spheroid is a well-known phenomenon thought to be influenced by the relative strength of homotypic and heterotypic cell-to-cell adhesion (12,19). Although more experimentation is needed, sorting within the tritypic spheroid appears to be more complex.

We have presented an easy-to-use, cost-effective, versatile, and scalable platform for controlled production of 3-D microtissues. Beyond spheroids, we have demonstrated the directed self-assembly of microtissues with prescribed shape and cell type composition. Because the method is scaffold-free and cells are not adhered to a surface, cells spontaneously self-assemble and reach a structural equilibrium that is governed by cell-to-cell interactions. The microtissues are in a hydrogel, so their surfaces can readily receive nutrients and exchange wastes in all three dimensions, whereas, these processes are not uniform for microtissues that are bound to a substrate such as polystyrene. This system has a wide range of applications in cell biology, cancer biology, high-throughput drug screening, and tissue engineering.

3D Petri Dishes™ used in this paper (please click catalog numbers for detailed product descriptions):
Catalog # 12-256, and 24-96 are the MicroTissues, Inc products that grow small spheroids. Catalog # 12-81 and 24-35 are used to grow larger spheroids. Catalog #12-60TR and 24-24TR are used to measure the rate of self-assembly and rod contraction. Catalog # 12-36TO is used to form microtissues in the shape of a toroid and catalog #24-H is used to form a multi-cellular honeycomb structure.

Click here to read full text.


Leave a Reply