Inverted colloidal crystals as three-dimensional microenvironments for cellular co-cultures

Abstract

Cellular scaffolds made on the basis of inverted colloidal crystals (ICC) provide a unique system for investigation of cell–cell interactions and their mathematical description due to highly controllable and ordered 3D geometry. Here, we describe three new steps in the development of ICC cell scaffolds. First, it was demonstrated that layer-by-layer (LBL) assembly with clay/PDDA multilayers can be used to modify the surface of ICC scaffolds and to enhance cell adhesion. Second, a complex cellular system made from adherent and non-adherent cells co-existing was created. Third, the movement of non-adherent cells inside the scaffold was simulated. It was found that floating cells are partially entrapped in spherical chambers and spend most of their time in the close vicinity of the matrix and cells adhering to the walls of the ICC. Using this approach one can efficiently simulate differentiation niches for different components of hematopoietic systems, such as T-, B- and stem cells.

---

Introduction

Colloidal crystals represent an exceptionally dynamic area of research capitalizing on the unique spatial organization and diffraction characteristics of sub-micron scale lattices.^1–4 These structures are designed, having in mind, primarily, applications in optics,⁵ sensors⁶ and catalysis.⁷ Inverted colloidal crystals (ICC) also open an interesting opportunity for a rather unexpected, but tremendously important area of science related to cell communication. We recently introduced the use of micron-scale ICC systems for three-dimensional (3D) cell cultures, which mimic the microenvironment of three-dimensionally organized native tissues.^8–10 Unlike other disordered cell supports, the ICC geometry affords systematic study of cell signaling in 3D, which has been proven to be fundamentally essential for proper development of tissues.^11–16

Adequate understanding and proper methods of control of cell signaling are particularly important for stem cell research. For instance, the rate and direction of the differentiation of stem cells are strongly affected by their 3D microenvironment and soluble signaling molecules.^17–22 Recent studies have shown that a 3D culture environment significantly promotes the efficiency of stem cell differentiation.^23,24 Intense cell–cell and cell–matrix interactions have been distinguished as key factors that determine the fate of individual cells by serving as important communication channels.^23,25,26 In order to reproduce the complexity and dynamics of cellular environments, various scaffold fabrication techniques have been developed.^27–31 However, the geometry of these scaffolds mainly depends on the process, and usually they have a poorly ordered or chaotic structure. Recently, rapid prototyping and 3D deposition techniques, assisted by computer-aided design and complex robotic equipment, were developed to construct more controlled 3D architectures.^27,32 These techniques allow researchers to design 3D scaffolds with desired properties such as porosity, interconnectivity and pore size. Nevertheless, besides being heavily equipment-dependent, they suffer from limited material selection and inadequate resolution. From a manufacturing standpoint, the fabrication procedure of ICC scaffolds is simple and flexible. Any precursor solution capable of undergoing a liquid-to-solid transition may potentially be used as a scaffolding material. An ordered structure, with a high degree of uniformity, can be achieved without the need for complex computer design programs and facilities.

Beyond that, several unique characteristics of ICC used as cell scaffolds make them particularly convenient for the use with stem cell cultures,^8–10,33 which can help uncover methods for successful tissue engineering from them. In this respect, ICC systems possess high surface areas with a void fraction of 76% and a regularly spaced network of pores which provides a mechanically strong, well-connected open porous geometry.³³ These features enhance cell seeding efficiency, transport of nutrients and metabolites, and the rapid and uniform distribution of soluble signaling molecules. The exceptionally uniform and three-dimensionally ordered structure of ICC scaffolds enables the development of computational models to systematically study the effect of signaling molecules, cell–cell and cell–matrix interactions, and other processes.³⁴ Until now, only single cell culture studies have been reported for ICC scaffolds.^8–10,33 Considering that this system could be a convenient discovery tool for research on cell–cell interactions, achieving the next level of complexity, i.e. the construction of a system with two or more different cell types co-populating the ICC matrix, appears to be the most essential step in this area.

In this paper, we introduce a model system combining two types of cells co-existing in an ICC scaffold, which paves the way for future systematic studies of cell evolution mechanisms. Since these interactions are of particular importance for the development of hematopoietic stem cells, the cell cultures were chosen having in mind the recreation of the 3D microenvironment of bone marrow and thymus differentiation niches.^26,35,36 The selection of particular model cell cultures was also aided by the fact that the characteristic geometry of the ICC scaffolds resembles that of bone marrow and thymus niches (i.e. stromal cells cover the surface and well intersticed sinus cavities). Human thymus epithelial cells (Hs202.Th) and human monocytes (HL-60) were used as anchorage-dependent feeder cells and suspension cells mimicking progenitors, respectively. Before using hematopoietic stem cells in our 3D culture system, we tried to use the HL-60 cell line because it is easier to deal with and has been proven a unique in-vitro model system for studying the cellular and molecular events involved in the proliferation and differentiation of normal and leukemic cells.³⁷ Cell–cell interactions within ICC scaffolds were evidenced by simplified Brownian Dynamics (BD) simulations taking advantage of the unique 3D morphology.

Experimental

Colloidal crystal construction

An aqueous suspension of polystyrene (PS) spheres with a diameter of 100 µm (Duke Scientific, 3 × 10⁴ particles per milliliter and 1.4% size distribution) was changed with isopropanol solution before use. A 0.5 ml plastic centrifugation tube was glued on the plastic dish and the top of the centrifuge tube was cut and connected with a long Pasteur glass pipette. The complex unit was attached to the bottom of a glass beaker, and the glass beaker was placed on an ultrasonic bath (VWR). Two drops of the solution were released through a long Pasteur glass pipette at 15 minute intervals (25 intervals total) under gentle agitation generated by the ultrasonic bath. To reduce thermal motions of the spheres, the bath temperature was maintained below 20 °C. After the dropping was finished, the isopropanol was evaporated off overnight at 60 °C.

Scaffold fabrication

Prepared colloidal crystals were heat treated at 120 °C for 4 hours, which caused partial melting of the beads' surface. As a result, the PS microspheres fused together and the free standing colloidal crystals were easily extracted from the mold. As scaffolding materials, a poly(acrylamide) hydrogel composed of a 30 wt% acrylamide (Sigma) precursor containing 5 wt% of N,N-methylenebisacrylamide (NMBA) cross-linker was used. The precursor was infiltrated into the colloidal crystal by centrifugation at 5800 rpm for 10 min. An initiator, 1 wt% of potassium peroxide solution, and an accelerator, N,N,N′,N′-tetramethylethylenediamine (TEMED), were added. Polymerization occurred in a glass vial. After the polymerization was complete, the colloidal crystal containing the hydrogel part was cut out and soaked in tetrahydrofuran (THF) for 24 hours to remove PS beads. Finally the ICC hydrogel scaffolds were equilibrated in deionized water. Fluorescent ICC hydrogel scaffolds were prepared by adding 0.05 wt% of fluorescent monomer, Polyfluor 511 (Polyscience Inc.), to the hydrogel precursor solution.

Layer-by-layer surface coating

ICC hydrogel scaffold surfaces were coated with the sequential deposition of positively charged 0.5 wt% poly(diallyldimethylammonium chloride) (PDDA, Sigma, MW = 200 000) solution for 15 min, and a negatively charged 0.5 wt% clay platelet (average 1 nm thick and 70–150 nm in diameter, Southern Clay Products) dispersion for 15 min. Each adsorption step was followed by rinsing in deionized water for 15 min, and all processes were performed under a gentle flow generated by a stirrer. Cyclic repetition of the polymer adsorption/rinsing/clay adsorption/rinsing process was carried out 10 times.³⁸

Mechanical property testing

Compressive moduli of hydrated ICC scaffolds, where the surface was coated with 10 layers of clay/PDDA, were measured at a constant strain rate (10 µm s⁻¹) using a mechanical properties tester and a 1.1 lb load cell (TestReciurces Inc., MN).

Dynamic co-culture

Human thymus cell line Hs202.Th (CRL-7163) and human premyeloblast cell line HL-60 (CCL-240) were purchased from ATCC (Manassas, VA). Hs202Th cells were grown in Dulbecco's Modified Eagle's Medium (DMEM) supplemented with 10% Fetal Bovine Serum (FBS) (GIBCO, CA). HL-60 cells were cultured in Iscove's Modified Dulbecco's Medium (IMDM) containing 20% FBS. The cells were maintained at 37 °C with 5% CO₂ and the medium was changed twice a week until they reached a confluent or desired population on T-75 culture flasks.

Co-culture was carried out in 10 ml rotary cell culture vessels (RCCS-4D, Synthecon Inc.). Scaffolds were sterilized by soaking in 70% EtOH for one hour followed by washing in phosphate buffered saline (PBS) for 15 min twice. 2 × 10⁵ Hs202.Th cells were placed in a culture vessel, which subsequently was filled with the medium. The rotation speed was set at 12 rpm for the first 12 hours and later it was decreased to 8 rpm, the normal speed. The medium was replaced once every three days. On day six, both Hs202.Th and HL-60 were stained with fluorescent dyes followed by a five day co-culture period. The HL-60 cells were stained with 5 µM chloromethyl derivatives fluorescent dye (CMRA, Molecular Probes) diluted in PBS buffer following the protocol provided by the vendor. Hs202.Th cells on the scaffold were stained with carboxyfluorescein diacetate succinimidyl ester (CFDA-SE, Molecular Probes). The culture medium was replaced with 10 ml of 5 µM CFDA-SE diluted in PBS buffer, and the culture was incubated at 37 °C for 20 min. After that, the medium was changed to IMDM supplemented with 20% FBS, and pre-stained 1 × 10⁶ HL-60 cells were seeded.

Cell culture characterization

Scaffolds were carefully removed from the culture vessel together with the medium and moved to a glass bottom culture dish (MatTek Corporation). The scaffolds were kept immersed in the medium during the process. Co-cultured scaffolds were investigated utilizing a Leica SP2 confocal microscope with 10× and 20× objective lenses. Thymic epithelial cells stained with CFDA-SE were imaged using a 470 nm filter while excited with 457 nm laser light (green). CMRA stained monocytes were excited at 543 nm and fluorescent lights were detected at 576 nm (red). Cross-sectional images were obtained in the same manner by taking pictures at 1 µm depth increments down to 160 µm.

Scanning Electron Microscope (SEM) observations were performed with a Philips XL30 SEM at 5.0 KV. Before imaging, hydrogel scaffold samples were first fixed in 2% cacodylate-buffered glutaraldehyde for 2 hours and washed three times in 0.1 M cacodylate buffer for 30 min. Fixed hydrogel scaffolds were dehydrated through a series of ethanol solutions concentrations of 50, 70, 90, 95, and 100% for 10 min. Dehydrated samples were freeze dried overnight utilizing a Labconco FreeZone (Labconco), and then were coated with gold for 180 s using a sputter coater (Desktop 2, Denton Vacuum Inc.). Cross-section images of the internal architecture were obtained after cutting the sample with a razor blade.

Results and discussion

ICC scaffold construction

The diameter of the spheres commonly used as colloidal crystals is around 100–1000 nm for the purpose of matching the optical band-gap in the visible region. Various methods such as electrophoretic deposition,³⁹ solvent evaporation,⁴⁰ dipping,⁴¹ agitation^42,43 and most recently spin coating⁴⁴ have been developed to construct highly ordered colloidal crystal structures. In order to utilize the unique geometry of the inverted colloidal crystals as a scaffold, the sphere size has to be increased to the 10–1000 µm range. However, it is difficult to obtain the same degree of order with micron scale beads using the above methods of synthesis, mainly due to their larger volume and heavier mass. Fortunately, micron-sized beads offer two advantages over nanosize spheres. First, the agitation of beads by shear force works more effectively because of their larger volume. Previously, we reported the construction of colloidal crystals by employing a gentle agitation method.⁹ The second advantage is a faster sedimentation rate due to their greater mass. However, the sedimentation rate was often too fast for them to self-assemble into a closely-packed ordered array. The opposite problem, viz., how to retard the sedimentation rate, was solved by introducing a Pasture glass pipette before the beads entered into the mold. The pipette extended the sedimentation distance and worked as a thin funnel, which caused a bottleneck effect for precipitating beads (Fig. 1a).

Fig. 1 (a) Schematic diagram of the experimental setup used for assembling micron-range polystyrene beads in 3D ordered structure. (b–d) SEM images of the colloidal crystal structure made from 100 µm polystyrene beads, showing the bottom (d), and internal structure at different magnifications (b, c). Internal images were taken after cutting the colloidal crystal with a razor blade. The small white spots on each PS sphere in (b) are contact points between beads, which later become channels.

Once beads precipitated at the bottom of the mold, gentle agitation generated by an ultrasonic bath assisted the movement of beads and positioned them at the lowest energy spots. This led to a highly packed and ordered array of spheres. When the bottom area was covered with beads, their rugged surface served as a template for the formation of the second layer. Since structural defects accumulated from the bottom area, incomplete layers and less ordered arrays were usually observed on the top area.

The sedimentation rate was controlled further by adjusting the concentration of beads in the solution and the time interval between injections. For example, decreasing the amount of beads and increasing the interval period provided more time for the repositioning of precipitated beads. The use of isopropanol guaranteed that the agitation was not too violent to destroy the whole structure, while its buoyancy made it easier for the PS beads to rearrange. Generated colloidal crystal structures were 8 mm diameter and 1–1.5 mm in thickness and SEM investigations revealed a highly ordered hexagonally close-packed structure (Fig. 1b–d).

Following sedimentation, the colloidal crystals were heat treated which resulted in partial melting of the spheres.²⁹ This step allowed the beads to stick together and on subsequent cooling (re-solidification), junctions were created between the spheres setting the structure in place. The resulting free standing colloidal crystals were strong enough to be easily handled and removed from the mold. The formation of the junctions later prevented breakage of the crystal lattice during the infiltration of scaffolding material and ensured the connectivity between spheres and continuity of the chain of pores in the final scaffold. The channel diameter was determined at this stage, because the size of the melted area depended on the annealing temperature. Increasing temperature enlarged the melting spot, but it caused shrinkage of the colloidal crystal. Usually it led to the cracking of the crystal structure and/or incomplete precursor solution infiltration. For 100 µm diameter PS beads, heat-treatment at 120 °C for 4 hours gave the optimal result.

The diameter of pores was dictated by the details of the co-culture system. Although monocytes and trypsinized thymic epithelial cells have similar dimensions, epithelial cells stretch out after attachment to the surface. Based on 2D characterization, the size of the elongated thymic epithelial cells was around 80–160 µm. Kotov et al. studied the pore size effect of ICC scaffolds on a 3D cell culture utilizing three different sizes of beads: 10 µm, 75 µm and 160 µm. The 75 µm pore diameter favored bone marrow stromal cells nesting, while the 10 µm pore size was too small for even a single cell, and 160 µm diameter pores were too large to effectively entrap cells.¹⁰ Also, Zinger et al. investigated osteoblast-like cell cultures on well-defined 2D cavities which were analogous to ICC scaffolds, and found that 100 µm cavities favored osteoblast attachment and growth.⁴⁵ For the entrapment and transport of suspension cells, the channel diameter, which is determined by the size of the microspheres, was the most important parameter. PS beads which had a 100 µm diameter made 25–30 µm diameter channels after annealing at 120 °C. The diameter of the suspension cells (approximately 15–20 µm) was small enough to enable them to pass through the channels.

As a scaffolding material, we selected poly(acrylamide) hydrogel. Hydrogel is a broadly used scaffolding material because of its biocompatibility, mechanical strength, and transparency.^46,47 The transparency of the hydrogel makes it easier to monitor cell migration and growth deep inside the scaffold using optical microscopy. Recently the observation of cell growth at a depth greater than 250 µm ⁹ and real time cell migration via a channel³³ were reported. In addition, the hydrogel exhibited another feature that facilitated its use in ICC work. At low viscosity of the precursor solution, it completely infiltrated to the colloidal crystal, and the whole structure of the crystal template was transferred intact (Fig. 2). The monomer concentration was set low enough to prevent incomplete infiltration due to increased viscosity, and simultaneously to prevent deformation of the geometry during solvent extraction.

Fig. 2 SEM images of the hydrogel scaffold after dehydration: (a) bottom structure image, (b) internal structure image taken after cutting the scaffold with a razor blade. The dehydration process caused shrinkage of the ICC hydrogel scaffold, which led to some deformation of the structure. Confocal images of a fluorescent hydrogel scaffolds: (c) a 3D reconstruction of serial z-section images taken in 0.5 µm steps showing the organization of main pores and interconnected channels of a hydrogel ICC scaffold without shape deformation, and (d) 3D overlapping images of serial z-sectional images of 160 µm interval with 5 µm step size.

The compressive moduli of hydrated and LBL coated ICC scaffolds were 189.4 ± 5.89 KPa (Fig. 3). Compared to the mechanical strengths of other porous hydrogel substrates, it showed stronger mechanical stability.^33,48 This was mainly due to the higher content of polymer and the highly ordered structure of the hydrogel ICC scaffolds. The achieved compressive modulus was within the range of normal articular cartilage.⁴⁹ This degree of mechanical property was adequate to construct artificial supports of targeted soft tissues.

Fig. 3 A compressive stress–strain curve from the mechanical property test.

Layer-by-layer surface modification

Hydrogel matrices rarely support adherent cell adhesion without surface modification, because acrylamide polymer chains do not have cell adhesion receptors, and the hydrophilic nature of the hydrogel inhibits adsorption of cell binding proteins on the gel surface.⁴⁷ To render the surface bioactive, we selected a layer-by-layer (LBL) deposition technique instead of the commonly used covalent coupling of specific peptide sequences such as RGD or an entire ECM protein to the polymer.⁴⁷ The driving forces for LBL coating are the electrostatic, van der Waals, and hydrogen bonding interactions between oppositely charged polyelectrolytes dissolved in aqueous solution.⁵⁰ This unique feature of the LBL technique allows a complex porous 3D geometry, such as the intricate and convoluted ICC surface, to be coated as long as fluid transport in and out of the sample is not severely constrained. It has been reported that 2D polyelectrolyte multilayers supported anchorage dependent cell attachment without using adhesive proteins.⁵¹

In our system, we used clay nanoparticles/poly(diallyldimethylammonium chloride) (PDDA) multilayers.³⁸ The clay particles are biocompatible and their flat shape effectively covered the hydrogel surface. Coated clay nanoparticles created nanoscale roughness, increased charging on the surface, and created much stiffer films than hydrogel. An increase of Young’s modulus was shown to be the primary factor determining the adhesion of cells to materials.^52,53 These synchronous effects promoted cell adhesion.^54,55 Ten layers of PDDA/clay easily changed the surface property from cell repulsive to cell adhesive, and thymic epithelial cells could attach to the hydrogel scaffold.

Dynamic co-culture

Human thymic epithelial cells and human premyelote monocytes were co-cultured in a rotary cell culture bioreactor. Rotary motion induced convective flow, and the scaffold geometry utilized this flow as a continuous driving force for the cell movement.⁵⁶ After five days of co-culture, the hydrogel ICC scaffolds were imaged through a confocal microscope. Different emission ranges of fluorescent dyes were used to stain the thymus cells and monocytes with green and red, respectively. Thymic epithelial cells attached to the cavity were observed as green circles, and floating monocytes were imaged as red spots (Fig. 4). Although many monocytes were diffused out of the scaffold during the sample preparation, some of them remained entrapped inside the pores. Confocal cross-sectional images reveal that suspended cells were distributed uniformly throughout the scaffold (Fig. 4b).

Fig. 4 Confocal images (a–c) of co-cultured ICC hydrogel scaffolds; thymic epithelial cells (green) and monocytes (red). (a) Bottom area image shows the surface of the scaffold was densely covered with thymic epithelial cells. Most of the monocytes around the edge of the scaffold were released. (b) A cross-sectional image after cutting the co-cultured ICC scaffold with a razor blade shows decreasing thymic epithelial cell density moving into the inside of the ICC scaffold. Monocytes were distributed through the whole ICC scaffold and a similar number of cells were entrapped at each pore. (c) A lateral section image of 80 µm in depth. SEM images (d–f) of co-cultured hydrogel scaffolds. (d) Cross-sectional image of the scaffold's interior. (e) Entrapped monocytes. (f) Thymic epithelial cells covering pores and channels.

Co-cultured hydrogel ICC scaffolds were dehydrated and observed under an SEM. The dehydration process deformed the structure, which is the reason for the dimensional differences between the two rows of images in Fig. 4. It was found that the scaffold exterior was covered densely with thymic epithelial cells, and their population reduced the inward movement of other epithelial cells (Fig. 4a, b, d). Secondly, epithelial cells migrated between pores through interconnected channels, and some colonies expanded over several pores (Fig. 4f). Thirdly, a few suspension cells were trapped inside when they were observed at the interior of the scaffold in SEM cross-sectional images (Fig. 4e). It suggests that monocytes travel deep into the ICC scaffolds.

Modeling

To study the interaction of a floating cell with the scaffold quantitatively, we constructed a simplified Brownian Dynamics model. The cell was treated as a hard sphere of radius a_cell that is suspended in an ICC geometry, composed of hollow spherical chambers of nominal radius R connected by channels of radius b.³⁴ To simplify the treatment, we assumed that the fluid inside the scaffold was quiescent and that the motion of the cells was purely diffusive. Under the assumptions stated above, the motion of the cell results from a balance between the drag force and the random Brownian force:

ζdr/dt = F_B (1)

where ζ = 6πηa_cell is the hydrodynamic drag exerted by the solvent of viscosity η on a cell of radius a_cell, and r is the position vector of the center of mass of the cell. The Brownian force, F_B, satisfies the fluctuation–dissipation theorem⁵⁷ which necessitates <F_B> = 0, and <F_BF_B> = 2k_BTζI. Here, I is the unit tensor, k_B = 1.38 × 10⁻²³ J K⁻¹ is Boltzmann's constant, and T is the absolute temperature.

The diffusivity, D, was obtained from the hydrodynamic drag via the Einstein relation,⁵⁷D = k_BT/ζ. In accordance with microscopy measurements, we took R = 50 µm, b = 12.5 µm, and a_cell = 7.5 µm. Thus, ζ = 6π(1 cP)(7.5 µm) = 1.414 × 10⁻⁴ g s⁻¹, and D = k_BT/ζ = 2.91 × 10⁻² µm² s⁻¹. We used the algorithm outlined by Larson⁵⁷ to implement the BD simulation, choosing the simulation time step, dt, so that √6Ddt ≈ 0.05a_cell. We employed reflecting boundary conditions to model collisions between the cell and the scaffold.

Grigoriev et al. considered a dimensionless Brownian particle trapped inside a spherical chamber of volume V.⁵⁸ They estimated that the time, t*, that it takes for the particle to escape from a small circular hole of radius b on the surface of the chamber is given by t* = V/4bD, where D is the diffusivity of the particle. We adapted the expression for t* to obtain a crude estimate for the escape time of a Brownian particle of finite size from an ICC scaffold as:

t*_ICC = (π/3ZD)(R − a_cell)³/(b − a_cell) (2)

where Z = 12 is the co-ordination number of the ICC lattice. From eqn (2), we obtained t*_ICC ∼ 5.5 × 10⁵/12 s ∼ 12 hours. Thus, the ICC geometry provides very suitable geometry for cell interactions due to partial entrapment of the cells in the cavity.

We simulated the dynamics of the cell in the ICC scaffold using BD, and recorded its trajectory from t = 0 to t = 1000 days. Over this period, the cell visited several chambers. From the simulation, we observed that by the time the cell vacated a chamber by escaping through the interconnecting channel to another chamber, it thoroughly, and uniformly, sampled the whole chamber. In other words, the amount of time the cell spent in any region of the chamber was proportional to the volume of that region. Fig. 5a shows a cross section of a spherical chamber that has been divided into shells of equal thickness, ΔR = a_cell. These shells do not have the same volume. For illustration, if we approximate the volume of a shell by ΔV_shell = 4πR_i²ΔR, where R_i is the inner radius of the shell, we can see that the volume of the outer shells is greater than that of the inner shells. As mentioned previously, the center of mass of the cell resides in a shell, in proportion to the volume of that shell. Thus, it spends a significant fraction of time (about 41%, see Fig. 5a) in the outermost shell, where the distance between the surface of the cell and the inner surface of the chamber is less than or equal to the radius of the cell. Thus the ICC geometry fosters contacts between the cell and the matrix surface or between the suspension and adherent cells in a co-culture.

Fig. 5 Radial probability distribution of a finite-sized Brownian particle of radius 7.5 µm diffusing in a spherical ICC chamber obtained from BD simulations, when the chamber is divided into shells of the same (a) thickness, and (b) volume. In (a), the dotted arc and the disc represent the inner surface of the chamber, and the cell which is modeled as a hard sphere, respectively. The thickness of each shell is equal to the radius of the cell, a_cell. From (b), it can be seen that the cell spends the same amount of time in each of the equi-volume shells, whereas in (a), it spends more time in the exterior shells due to their greater volume.

Conclusions

We have demonstrated a 3D co-culture model system within a LBL surface modified ICC scaffold in both experimental and modeling works. Clay/PDDA multilayer successfully formed on a complex porous ICC scaffold walls, and significantly improved adhesion of epithelial cells. The unique geometry of ICC scaffolds can accommodate two different types of cells within a same chamber. Modeling results indicated that cells were effectively trapped in the spherical chambers and that entrapped suspension cells spent a significant fraction of time in the vicinity of the ICC chamber wall or the epithelial coating in co-cultures.

Well controlled multiscale structures which can build real-size organ systems and generate the essential subcellular morphology are a key factor for the successful investigation of cell–molecule and cell–cell interactions.^12,59 It is obvious that the full function of the tissues and organs cannot be recovered without rebuilding the ultrastructure of the tissue itself. Proposed ICC scaffolds and surface modification utilizing a LBL technique will be excellent approaches for this purpose. ICC scaffold structure generates super- and cellular-scale microenvironments for intense cell contacts with other types of cell or matrix. On this surface, various insoluble signaling molecules such as ECM components, membrane bound receptors and ligands can be incorporated through a LBL method which can produce a subcellular, nanoscale resolution environment for cellular receptor–molecular interactions. In particular, this could greatly facilitate the study of B and T cell development from stem cells which requires understanding and controlling precise 3D molecular interactions.

Acknowledgements

This work was supported by a grant from DARPA and VaxDesign Inc. We thank Dr Michael Solomon and Tesfu Solomon (University of Michigan, Ann Arbor) for assistance with the laser scanning confocal microscope.

References

1. T. Prasad, R. Rengarajan, D. M. Mittleman and V. L. Colvin, Opt. Mater., 2005, 27, 1250.
2. D. Wang, V. Salgueirino-Maceira, L. M. Liz-Marzan and F. Caruso, Adv. Mater., 2002, 14, 908.
3. Y. Wang and F. Caruso, Adv. Funct. Mater., 2004, 14, 1012.
4. C. E. Reese, A. V. Mikhonin, M. Kamenjicki, A. Tikhonov and S. A. Asher, J. Am. Chem. Soc., 2004, 126, 1493.
5. N. Tetreault, H. Miguez and G. A. Ozin, Adv. Mater., 2004, 16, 1471.
6. Y. Y. Song, D. Zhang, W. Gao and X. H. Xia, Chem.–Eur. J., 2005, 11, 2177.
7. R. C. Schroden, C. F. Blanford, B. J. Melde, B. J. S. Johnson and A. Stein, Chem. Mater., 2001, 13, 1074.
8. Y. Liu, S. Wang, J. W. Lee and N. A. Kotov, Chem. Mater., 2005, 17, 4918.
9. Y. Zhang, S. Wang, M. Eghtedari, M. Motamedi and N. A. Kotov, Adv. Funct. Mater., 2005, 15, 725.
10. N. A. Kotov, Y. Liu, S. Wang, C. Cumming, M. Eghtedari, G. Vargas, M. Motamedi, J. Nichols and J. Cortiella, Langmuir, 2004, 20, 7887.
11. A. Abbott, Nature, 2003, 424, 870.
12. S. Kale, S. Biermann, C. Edwards, C. Tarnowski, M. Morris and M. W. Long, Nat. Biotechnol., 2000, 18, 954.
13. Y. Xie, S. T. Yang and D. A. Kniss, Tissue Eng., 2001, 7, 585.
14. C. Trojani, P. Weiss, J. F. Michiels, C. Vinatier, J. Guicheux, G. Daculsi, P. Gaudray, G. F. Carle and N. Rochet, Biomaterials, 2005, 26, 5509.
15. H. J. Evans, J. K. Sweet, R. L. Price, M. Yost and R. L. Goodwin, Am. J. Physiol., 2003, 285, H570.
16. D. Ferrera, S. Poggi, C. Biassoni, G. R. Dickson, S. Astigiano, O. Barbieri, A. Favre, A. T. Franzi, A. Strangio, A. Federici and P. Manduca, Bone, 2002, 30, 718.
17. J. Zhang, C. Niu, L. Ye, H. Huang, X. He, W. Tong, J. Ross, J. Haug, T. Johnson, J. Q. Feng, S. Harris, L. M. Wiedemann, Y. Mishina and L. Li, Nature, 2003, 425, 836.
18. E. Fuchs, T. Tumbar and G. Guasch, Cell, 2004, 116, 769.
19. A. Spradling, D. Drummond-Barbosa and T. Kai, Nature, 2001, 414, 98.
20. T. Imamura, L. Cui, R. Teng, K. Johkura, Y. Okouchi, K. Asanuma, N. Ogiwara and K. Sasaki, Tissue Eng., 2004, 10, 1716.
21. S. Ding and P. G. Schultz, Nat. Biotechnol., 2004, 22, 833.
22. A. W. Duncan, F. M. Rattis, L. N. DiMascio, K. L. Congdon, G. Pazianos, C. Zhao, K. Yoon, J. M. Cook, K. Willert, N. Gaiano and T. Reya, Nat. Immunol., 2004, 6, 314.
23. H. Liu and K. Roy, Tissue Eng., 2005, 11, 319.
24. M. C. Poznansky, R. H. Evans, R. B. Foxall, I. T. Olszak, A. H. Piascik, K. E. Hartman, C. Brander, T. H. Meyer, M. J. Pykett, K. T. Chabner, S. A. Kalams, M. Rosenzweig and D. T. Scadden, Nat. Biotechnol., 2000, 18, 729.
25. E. K. F. Yim and K. W. Leong, Nanomedicine, 2005, 1, 10.
26. P. Bousso, I. R. Bhakta, R. S. Lewis and E. Robey, Science, 2002, 296, 1876.
27. J. M. Williams, A. Adewunmi, R. M. Schek, C. L. Flanagan, P. H. Krebsbach, S. E. Feinberg, S. J. Hollister and S. Das, Biomaterials, 2005, 26, 4817.
28. X. Liu and X. P. Ma, Ann. Biomed. Eng., 2004, 32, 477.
29. P. X. Ma and J. W. Choi, Tissue Eng., 2001, 7, 23.
30. T. B. F. Woodfield, J. Malda, J. de Wijn, F. Peters, J. Riesle and C. A. van Blitterswijk, Biomaterials, 2004, 25, 4149.
31. S. Yang, K. F. Leong, Z. Du and C. K. Chua, Tissue Eng., 2002, 8, 1.
32. X. Wang, Y. Yan, Y. Pan, Z. Xiong, H. Liu, J. Cheng, F. Liu, F. Lin, R. Wu, R. Zhang and Q. Lu, Tissue Eng., 2006, 12, 83.
33. A. N. Stachowiak, A. Bershteyn, E. Tzatzalos and D. J. Irvine, Adv. Mater., 2005, 17, 399.
34. S. Shanbhag, J. W. Lee and N. Kotov, Biomaterials, 2005, 26, 5581.
35. I. R. Lemischka and K. A. Moore, Nature, 2003, 425, 778.
36. M. M. Davis, M. Krogsgaard, J. B. Huppa, C. Sumen, M. A. Purbhoo, D. J. Irvine, L. C. Wu and L. Ehrlich, Annu. Rev. Biochem., 2003, 72, 717.
37. S. J. Collins, Blood, 1987, 70, 1233.
38. Z. Tang, N. A. Kotov, S. Magonov and B. Ozturk, Nat. Mater., 2003, 2, 413.
39. A. L. Rogach, N. A. Kotov, D. S. Koktysh, J. W. Ostrander and G. A. Ragoisha, Chem. Mater., 2000, 12, 2721.
40. J. P. Hoogenboom, C. Retif, E. de Bres, M. van de Boer, A. K. Langen-Suurling, J. Romijn and A. van Blaaderen, Nano Lett., 2004, 4, 205.
41. S. H. Im, M. H. Kim and O. O. Park, Chem. Mater., 2003, 15, 1797.
42. O. Vickreva, O. Kalinina and E. Kumacheva, Adv. Mater., 2000, 12, 110.
43. M. Sasaki and K. Hane, J. Appl. Phys., 1996, 80, 5427.
44. P. Jiang and M. J. McFarland, J. Am. Chem. Soc., 2004, 126, 13778.
45. O. Zinger, G. Zhao, Z. Schwartz, J. Simpson, M. Wieland, D. Landolt and B. Boyan, Biomaterials, 2004, 26, 1837.
46. Y. Luo and M. S. Shoichet, Nat. Mater., 2004, 3, 249.
47. J. L. Drury and D. J. Mooney, Biomaterials, 2003, 24, 4337.
48. S. H. M. Soentjens, D. L. Nettles, M. A. Carnahan, L. A. Setton and M. W. Grinstaff, Biomacromolecules, 2006, 7, 310.
49. P. Kiviranta, J. Rieppo, R. K. Korhonen, P. Julkunen, J. Toyras and J. S. Jurvelin, J. Orthop. Res., 2006, 24, 690.
50. G. Decher, Science, 1997, 277, 1232.
51. S. Kidambi, I. Lee and C. Chan, J. Am. Chem. Soc., 2004, 126, 16286.
52. M. C. Berg, S. Y. Yang, P. T. Hammond and M. F. Rubner, Langmuir, 2004, 20, 1362.
53. H. Zheng, M. C. Berg, M. F. Rubner and P. T. Hammond, Langmuir, 2004, 20, 7215.
54. G. B. Schneider, A. English, M. Abraham, R. Zaharias, C. Stanford and J. Keller, Biomaterials, 2004, 25, 3023.
55. S. Kay, A. Thapa, K. M. Haberstroh and T. J. Webster, Tissue Eng., 2002, 8, 753.
56. P. A. Plett, S. M. Frankovitz, R. Abonour and C. M. Orschell-Traycoff, In Vitro Cell. Dev. Biol.: Anim., 2001, 37, 73.
57. G. R. Larson, The Structure and Rheology of Complex Fluids, Oxford University Press, New York, 1999.
58. I. V. Grigoriev, Y. A. Makhnovskii, A. M. Berezhkovskii and V. Y. Zitserman, J. Chem. Phys., 2002, 116, 9574.
59. V. Vogel and M. Sheetz, Nat. Rev. Mol. Cell Biol., 2006, 7, 265.

---