Development of Three-Dimensional Tissue Models

© 2012 The Chemical Society of Japan

Bull. Chem. Soc. Jpn. Vol. 85, No. 4, 401­414 (2012)

401

Award Accounts

The Chemical Society of Japan Award for Young Chemists for 2010

Development of Three-Dimensional Tissue Models Based on Hierarchical Cell Manipulation Using Nanofilms Michiya Matsusaki1,2 1

Department of Applied Chemistry, Graduate School of Engineering, Osaka University, 2-1 Yamada-oka, Suita, Osaka 565-0871 2

PRESTO, Japan Science and Technology Agency, 4-1-8 Honcho, Kawaguchi, Saitama 332-0012

Received June 27, 2011; E-mail: [email protected]

The creation of artificial three-dimensional (3D) tissues possessing structure and function similar to natural tissue is a key challenge for implantable tissues in tissue engineering, and for model tissues in pharmaceutical assays. This account is a summary of our current research toward this challenge. We have developed a simple and unique bottom-up approach, hierarchical cell manipulation technique, using nanometer-sized layer-by-layer films consisting of fibronectin and gelatin (FN­G) as a nano-extracellular matrix (nano-ECM). The FN­G nanofilms were prepared directly on the cell surface, and we discovered that at least 6 nm thick FN­G films acted as a stable adhesive surface for adhesion of the second cell layer. Various 3D-layered constructs consisting of single or multiple types of cells were successfully fabricated, and the higher cellular activities induced from the 3D-structures as compared to monolayer structure were observed. Furthermore, the multilayered constructs like a blood vessel wall structure indicated almost the same drug response as in vivo natural blood vessels, suggesting the possibility to use as an in vitro blood vessel model to analyze drug response. Recently, we also developed a rapid bottom-up approach by a single cell coating using FN­G nanofilms, because the fabrication of twolayers (2L) is limited through the above technique due to the time required for stable cell adhesion. This rapid approach easily provided approximately eight-layered (8L) 3D-tissues after only one day of incubation. The layer number, cell type, and location were all successfully controlled by altering the seeding cell number and order. Moreover, fully and homogeneously vascularized tissues of 1 cm width and 50 ¯m height were obtained by a sandwich culture of the endothelial cells. These hierarchical cell manipulations will be promising to achieve one of the dreams of biomedical field, in vitro creation of artificial 3D-tissue models.

1. Introduction In the body, nearly all tissue cells reside in the fibrous nanomeshwork of the extracellular matrix (ECM). The ECM is typically composed of fibronectin (FN) and collagen, and provides complex biochemical and physical signals.1­3 The in vitro development of highly organized three-dimensional (3D)engineered tissue constructs composed of not only multiple types of cells, but also ECM fiber scaffolds which possess a similar structure and function as natural tissues, is a key challenge for implantable tissues in tissue engineering, and for model tissues in pharmaceutical assays.4 Recently, the topological control of biodegradable porous scaffolds,5,6 especially nanofiber scaffolds by electrospinning7 or self-assembling amphiphilic peptides8 has attracted much attention due to their high porosity and the controlled alignment of the fibers to control cellular function and the development of 3D-engineered tissues.9 However, 3D-engineered tissues possessing precisely controlled cell type, cell alignment, and cell­cell interaction have not been developed yet. A conventional approach using

biodegradable matrices such as hydrogels or fiber scaffolds seems to have several limitations in developing 3D-tissue constructs which satisfy the above requirements. A bottom-up approach using multiple cell types as pieces of tissue is expected to solve these problems. Currently, various technologies such as cell sheets10 or magnetic liposomes11 have been reported in constructing a multilayered cell sheet. These methods are intriguing examples of a bottom-up approach, but have limitations due to the complicated manipulation of the fragile cell sheet or the remains of magnetite particles in the cells. Recently, we have developed a simple and unique bottom-up approach, “hierarchical cell manipulation technique,” which employs nanometer-sized layer-by-layer (LbL) FN­gelatin (FN­G) films as a nano-ECM (Figure 1).12 We focused on cell adhesive properties of ECM-surrounding cell microenvironment. In general in vitro cell culture method, we obtain only cellular monolayers probably due to absence of sufficient ECM on cell surfaces. If cell adhesive nanometer-sized materials like an ECM were directly fabricated on cell mem-

Published on the web March 17, 2012; doi:10.1246/bcsj.20110194

402

Bull. Chem. Soc. Jpn. Vol. 85, No. 4 (2012)

AWARD ACCOUNTS : FN :G

LbL assembly First cell monolayers

Second cell seeding

Fabrication of nanofilms onto cell surface

LbL assembly & cell seeding

Bilayers

3D-Layered tissues

Figure 1. Schematic illustration of the hierarchical cell manipulation using layer-by-layer (LbL) assembly technique. FN and G were fibronectin and gelatin, respectively.

brane surfaces, 3D-control of cell­cell adhesion would be predicted to be possible. In fact, the FN­G nanofilms prepared directly on the cell surface acted as a stable adhesive surface for the adhesion of the second cell layer. We have taken up the challenge to develop an artificial 3D-tissue model possessing a similar structure and function as a natural tissue, and this paper describes an overview of our effort toward this goal. 2. Hierarchical Cell Manipulation Technique To develop the 3D-cellular multilayers, direct fabrication of nanometer-sized cell adhesive materials like ECM fibrous scaffolds onto the surface of a cell membrane is crucial, because the insufficient ECM is secreted onto cell surfaces in the early stage of cell culture. We focused on LbL technique which is an appropriate method to prepare nanometer-sized films on a substrate by alternate immersion into interactive polymer solutions.13,14 The preparation of nanometer-sized multilayer films composed of ECM components on the surface of the first layer of cells will provide a cell-adhesive surface for the second layer of cells. Rajagopalan et al. demonstrated a bilayer structure composed of hepatocytes and other cells by preparing a polyelectrolyte (PE) multilayer consisting of chitosan and DNA onto the hepatocyte surface.15 However, chitosan cannot dissolve in neutral buffer, and the use of PE films as a cell adhesive material is limited due to the cytotoxicity of polycations.16,17 The appropriate choice of natural ECM components for preparation of the nanofilms is important to avoid cytotoxicity, and the typical ECM presents cell-adhesive moieties such as RGD (arginine­glycine­aspartic acid) and other functional moieties.18 We selected FN and G to prepare nano-ECM films on the cell surface. FN is a flexible multifunctional glycoprotein and plays an important role in cell attachment, migration, differentiation, and so on.2,19 FN is well known to interact not only with a variety of ECM proteins such as collagens (gelatins) and glycosaminoglycans but also with the ¡5¢1 integrin receptor on the cell surface.20 We have reported FN-based LbL multilayers composed of FN and ECM components, such as gelatin, heparin, and elastin, constructed by alternate immersion into the solutions (LbL assembly).21 Although FN and G have a

negative charge under physiological conditions, they interacted with each other because FN has a collagen binding domain,2 indicating different driving force of PE films using polycations. Thus, the FN­G nanofilms are expected to provide a suitable cell-adhesive surface similar to the natural ECM for the second layer of cells, without any cytotoxicity. The fabrication of 3D-cellular multilayers composed of cells and FN­G nano-ECM films was performed according to the process shown in Figure 1. The LbL assembly of FN and G onto the cell surface was analyzed quantitatively using a quartz crystal microbalance (QCM) as the assembly substrate, and with a phospholipid bilayer membrane as a model cell membrane (Figure 2a). A phospholipid bilayer composed of 1,2-dipalmitoyl-sn-glycero-3-phosphatidylcholine (DPPC) and 1,2-dipalmitoyl-sn-glycero-3-phosphate (DPPA) was prepared onto the base layer, a 4-step assembly of poly(diallyldimethylammonium chloride) (PDDA) and poly(sodium styrenesulfate) (PSS), according to Krishna’s report.22 The mean thickness of the LbL assembly after 1, 7, and 23 steps was calculated to be 2.3, 6.2, and 21.1 nm, respectively. The top and cross sections of the confocal laser scanning microscopy (CLSM) 3D-merged images indicated a homogeneous assembly of fluorescently labeled FN­G nanofilms on the surface of mouse L929 fibroblasts. For the quantitative studies on the thickness of the multilayers on the cell surface, the fluorescence intensity of rhodamine-labeled FN (Rh­FN) was estimated by a line scan (Figure 2b). The fluorescence intensity of the Rh­FN­G nanofilms increased linearly upon increasing the LbL assembly step number, similar to the frequency shift of the QCM analysis, indicating a clear increase in the film thickness on the cell surface. These results demonstrated the fabrication of FN­G nanofilms on the cell surface. We fabricated a bilayer of mouse L929 fibroblast cells with or without FN­G nanofilms by using a cover glass as a substrate. When the 7-step-assembled FN­G nanofilms were prepared on the surface of the first L929 cell layer, the second layer cells were then observed on the first cell layer (Figures 2c and 2d). However, when the nanofilm was not prepared or the 1-step-assembled nanofilm (only FN) was assembled on the

M. Matsusaki

a)

Bull. Chem. Soc. Jpn. Vol. 85, No. 4 (2012)

b)

1500

403

80

1000 6.2 nm 2.3 nm 500

60 250

Intensity / a.u.

Intensity / a.u.

Frequency shift / Hz

21.1 nm

40

20

Phospholipid bilayer Base layer 5

9

13

17

g)

0

21

10

15

20

25

30

35

40

1

3

5

7

9

11

13

Step Number e)

10 μm

h)

20 μm

50

Distance / μm

d)

20 μm

100

5

Step Number c)

150

0

0 1

200

f)

20 μm

10 μm

i)

10 μm

20 μm

Figure 2. a) Quartz crystal microbalance (QCM) analysis of the LbL assembly of FN and G onto a phospholipid bilayer (DPPC/ DPPA 4:1). Open and closed circles show the assembly steps of G and FN, respectively. The base layer was a 4-step assembly of poly(diallyldimethylammonium chloride) (PDDA) and poly(sodium styrenesulfate) (PSS). DPPC and DPPA were 1,2-dipalmitoylsn-glycero-3-phosphatidylcholine and 1,2-dipalmitoyl-sn-glycero-3-phosphate, respectively. b) Relationship between the LbL assembly step of rhodamine-labeled FN (Rh­FN)­G on the L929 cells and the fluorescence intensity of the Rh­FN. The inset shows fluorescence intensity of the cells (green) and the 7-step-assembled Rh­FN­G nanofilms (red) by line scan. The cells were labeled with cell tracker green. c, e, g) Phase-contrast (Ph) and d, f, h) fluorescence microscope images of L929 fibroblast bilayers (c, d) with the 7-step-assembled, (e, f ) without a 7-step-assembled, or (g, h) with a 1-step-assembled FN­G nanofilms on the surface of the fist cell layer. i) Ph image of a L929 cell monolayer as a control. The nuclei were labeled with DAPI.

first cell layer, then the bilayer architecture was not observed (Figures 2e­2h). These results suggested 2.3 nm of FN film was inadequate and at least approximately 6 nm of FN­G nanofilm was required as a stable adhesive surface for the second cell layer. The four layered (4L) cellular multilayers were clearly observed after four repetitions of these steps by confocal laser scanning microscopy (CLSM) and hematoxylin and eosin (HE) staining images (Figures 3a and 3b). The thickness of the obtained multilayers estimated from 3Dreconstructed CLSM images linearly increased with increasing cell layer number (Figure 3c). This hierarchical cell manipulation technique can be applied to various types of cells, e.g., myoblast cells (Myo, Figure 4a), cardiac myocytes (CMyo, Figure 4b), smooth muscle cells (SMC), hepatocytes (Hep), and endothelial cells (EC). The FN­G nanofilms did not show any cytotoxicity,12 and the obtained 3D-cell architectures presented high intercellular adhesion to easily peel off from the substrate (Figure 4c).

3. Control of Cell Surface and Functions by Nanofilms and Nanocrystals For further studies of the multilayered cellular constructs, detailed evaluation and understanding of the effect of FN­G nanofilms on cell surface morphology and functions are important. Moreover, there are no reports on the effects of the various LbL nanofilms prepared on the cell membrane on cell functions, although cell functions such as adhesion, proliferation, and differentiation on varied LbL films have been reported.23­25 Accordingly, various LbL films consisting of synthetic polymers, polysaccharides, poly(amino acid)s, and proteins were prepared onto the surface of mouse L929 fibroblasts, and the cell viability, morphology, and proliferation were investigated in relation to the LbL films, because we would like to evaluate the effect of LbL films on basic cellular functions to clarify the universal effect of the LbL films prepared on the cell surface.26 Table 1 describes summary of

404

Bull. Chem. Soc. Jpn. Vol. 85, No. 4 (2012)

AWARD ACCOUNTS

a)

30

c)

b)

Thickness / μm

25

30 μm

20 15 10 5 0 0

1

2

3

4

5

Layer Number

Figure 3. a) 3D-reconstructed confocal laser scanning microscopy (CLSM) cross-section image of 4-layered (4L) L929 cells. The cells were labeled with cell tracker green. b) Hematoxylin and eosin (HE) staining image of 4L-L929 cells. c) Relationship between the L929 cell layer number and the mean thickness estimated from 3D-CLSM images (n = 3).

a)

b) 5L-CMyo

1L-Myo

1 1 Nuclei

Cytoplasm

3

4

5 50 μm

50 μm

3L-Myo

2

c)

1 2 3 50 μm

7L-Myo

7

65

4

3

2

1

50 μm

Figure 4. C2C12 cardiac process

a) HE staining images of 1-, 3-, and 7L-mouse myoblast cells (Myo). b) HE image of 5L-rat myocytes (CMyo). c) Photograph of the peeling of 4L-L929 cells from cover glass as a substrate.

the nanofilm components, cell viabilities, and cell types used in this paper. Figures 5a­5e show the morphologies of L929 fibroblasts after preparing the LbL films. PE multilayers induced a round morphology of the adhered cells, whereas each component of the multilayers had high cytocompatibility. On the other hand, FN­G and FN­dextran sulfate (FN­DS) multilayers prepared by FN-binding domain interactions (FN films) showed extended morphologies of the cells similar to that of control cells (without films). Figure 5f indicates the relationship between cell viability and thickness of the LbL films. The extremely high cell viability independent of the thickness of FN films was observed even with the thick 121 nm of FN­DS films, while

all of PE films indicated thickness dependent cytotoxicity. Furthermore, a clear difference in cell proliferation was observed for PE and FN films (Figure 5g). The cells with FN films on their surfaces showed good proliferation profiles independent of the film thickness, but cell growth was not observed using the PE films although the cells survived during the culture period. The morphologies of FN and PE films on the cell surface were observed by fluorescence microscopy and scanning electron microscopy (SEM). The nanometer-sized meshwork morphology of the FN films on the cell surface was found after 24 h of incubation, whereas the PE films showed homogeneous film morphologies on the cell surface (Figures 5h­5j).26 These nanomeshwork morphologies seemed to be similar to the fibrous structure of the natural ECM. To understand in detail the effect of PE films on cell functions, secretion of interleukin-6 (IL-6) cytokine, which is expressed from the inflamed cells, was evaluated after preparing PE or FN films (Figure 6). In the case of PE films, normal human dermal fibroblasts (NHDFs) expressed approximately 2.5-fold higher IL-6 per single cell than NHDFs without films. On the other hand, FN films did not show any effect to IL-6 production. These results suggested the PE films prepared on the cell surface induce strong inflammation to the cells probably due to the cationic cytotoxicity, and finally the films would induce cell death or growth stop dependent on thickness, cationic charge, or cationic species. We also tried to control cellular differentiation by preparing hydroxyapatite (HAp) nanocrystal on the cell surface.27 HAp, Ca10(PO4)6(OH)2, is one of the more attractive materials for artificial bone applications because of its biocompatibility, the absence of inflammatory or foreign body responses, and the fact that it is of similar crystalline phase as bone mineral.28,29 HAp crystals were also prepared in similar fashion, alternate immersion of substrates into Ca2+ and PO4¹ solutions.30 Accordingly, to indicate the versatility and expandability of our technique to not only polymer nanofilms but also nanocrystals, we tried to fabricate HAp nanocrystals onto the cell surface, like the nanofilms. HAp nanocrystals were directly prepared on the surface of mesenchymal stem cell (MSC)-based tissueengineered construct (TEC)31 by the alternate soaking process, and the formed HAp crystals were spherical with 100 nm to 1 ¯m size (Figure 7a). The analyses of the crystals using Fourier transform infrared spectroscopy (FT-IR) and wideangle X-ray diffraction (WAXD) clearly demonstrated HAp crystal form (Figures 7b and 7c). The implantation experiments of the TEC­HAp composites to rabbit osteochondral defects exhibited accelerated osteoinduciton, strongly suggesting differentiation of MSCs inside the TEC­HAp composite to osteoblast probably due to HAp nanocrystals prepared on the surface of MSCs. 4. High Cellular Properties Induced by 3D-Architectures Studies on the functions of layered cellular architectures as compared with cell monolayer are valuable not only for understanding how a 3D environment composed of cells, ECM, and signaling molecules regulates functions similar to natural tissues, but also for creating 3D-artificial tissues resembling natural tissues. Recently, some researchers have reported the functions of layered cellular architectures in vitro.15,32­35

Ion

94.3 « 3.2e)

PO4¹ (negative) Ion

Cell viability 1d) Component 2 Species /% (Charge) 86.3 « 7.6 G Protein (negative) 86.3 « 7.6 DS Naturally (negative) polymer 77.0 « 3.6 ¾-Lys Naturally (positive) polymer occurring 77.0 « 3.6 DS Naturally (negative) polymer occurring 77.0 « 3.6 PSS Synthetic (negative) PSS Synthetic polymer 73.3 « 5.9e) (negative) polymer 6.5 « 1.1 PAA Synthetic (negative) 96.8 « 5.3e)

100­1000h) Electrostatic interaction

Cell viability 2d) Total thicknessg) Interaction /% /nm 97.7 « 8.0 8.9 « 0.9 FN-binding domain interaction occurring 91.3 « 2.3 121.1 « 6.0 FN-binding domain interaction occurring 86.3 « 7.6 11.0 « 1.3 Electrostatic interaction occurring 91.3 « 2.3 12.0 « 0.9 Electrostatic interaction polymer 96.7 « 9.3 9.1 « 0.6 Electrostatic interaction polymer 96.7 « 9.3 17.2 « 1.2 Electrostatic interaction polymer 99.0 « 7.8 24.4 « 2.7 Electrostatic interaction

L929 L929, NHDF L929 L929, NHDF

Nanofilm Nanofilm Nanofilm Nanofilm Nanocrystal

MSC

L929, NHDF

Nanofilm

Nanofilm

L929, NHDF, HUVEC, SMC, HAEC, AoSMC L929, NHDF

Nanofilm

Nanostructure Cell typei)

a) Abbreviations: FN, fibronectin; G, gelatin; DS, sodium dextran sulfate; ¾-Lys, ¾-poly(lysine hydrochloride); PSS, poly(sodium styrenesulfate); PDDA, poly(diallyldimethylammonium chloride); PAH, poly(allylamine hydrochloride); PAA, poly(acrylic acid). b) Concentration of each solution used for LbL assembly is 0.2 mg mL¹1 in 50 mM Tris-HCl (pH 7.4). c) The living cell percentage was estimated from positive control sample as 100% which was not performed any treatment. d) Cell viability was determined after immersion in 0.2 mg mL¹1 of each component solution for 15 min by using WST-1 reagent. e) Concentration of PDDA soultion is 0.02 mg mL¹1 in 50 mM Tris-HCl (pH 7.4) with 0.035 M CaCl2. The concentration of Ca2+ and PO4¹ were 10 in 50 mM Tris-HCl (pH 7.4). f ) PAH is nagetive control as a cytotoxic component. g) The total thickness at 13 step on a phospholipid bilayer prepared on 2.3 « 0.4 nm of (PDDA-PSS)1.5 base layer is 4.1 « 1.1 nm. h) These values are crystal sizes. i) The cell types used for nanofilm or nanocrystal preparations. Abbreviations: L929, mouse L929 fibroblast; MSC, mesenchymal stem cell; NHDF, normal human dermal fibroblast; HUVEC, human umbilical vein endothelial cell; SMC, smooth muscle cell; HAEC, human artery endothelial cell; AoSMC, human aortic smooth muscle cell.

Ca2+ (positive)

Component 1 Species (Charge) FN Protein (negative) FN Protein (negative) FN Protein (negative) ¾-Lys Naturally (positive) polymer ¾-Lys Naturally (positive) polymer Synthetic PDDA (positive) Synthetic PAH (positive)f )

Table 1. Summary of the Nanofilm Components and Cell Types Used in This Paper a),b),c)

M. Matsusaki Bull. Chem. Soc. Jpn. Vol. 85, No. 4 (2012) 405

406

Bull. Chem. Soc. Jpn. Vol. 85, No. 4 (2012)

a)

b)

Control (without films)

FN-G: 8 nm

20 μm

AWARD ACCOUNTS

c)

d)

e)

FN-DS: 121 nm

FN-ε-Lys: 17 nm

ε-Lys-PSS: 18 nm

20 μm

f)

g) ) )

100 80 60

PDDA-PSS ( ε-Lys-DS ( FN-ε-Lys ( ε-Lys-PSS ( PAH-PAA (

40 20

20 μm

25

FN specific

Cell number / x104 cells

FN-G ( FN-DS (

120

Cell viability / %

20 μm

20 μm

) ) ) ) )

Electrostatic interaction

Control ( ) FN-G: 8 nm ( ) FN-DS: 121 nm( )

20

15 10 Film preparation 5

FN-ε-Lys 17 nm ( ) ε-Lys-PSS 18 nm ( )

0 0

5

10

15

20

40

80

120

140

0 0

Thickness / nm

24

48

72

96

Culture time / h

h)

i) FN-G, 24h

FN-DS, 24h

20 μm

5 μm

j) FN-ε-Lys, 24h

20 μm

Figure 5. a­e) Ph images of the L929 fibroblasts (a) without or with (b) 8 nm thick FN­G, (c) 121 nm thick FN­DS, (d) 17 nm thick FN­¾-Lys, and (e) 18 nm thick ¾-Lys­PSS nanofilms on the cell surfaces after 24 h of incubation. DS and ¾-Lys were sodium dextran sulfate and ¾-poly(lysine hydrochloride), respectively. f ) Relationship between cell viability and thickness of various nanofilms prepared on the cell surfaces after 24 h of incubation (n = 3). PAH and PAA were poly(allylamine hydrochloride) and poly(acrylic acid). g) Cell proliferation versus various nanofilms prepared on the L929 cell surfaces during 72 h of incubation. h) SEM image of the L929 cells with 121 nm thick FN­DS films after 24 h of incubation. Fluorescence microscopic images of i) 21 nm thick Rh­FN­FITC­G and j) 17 nm thick Rh­FN­FITC­¾-Lys films prepared on L929 cells after 24 h of incubation.

However, the basic properties induced by 3D-cellular structures, such as the layer number or the cell types, have not been clarified yet. We evaluated the structural stability of layered constructs consisting of NHDFs and human umbilical vein endothelial cells (HUVECs) in relation to their layer number.36 The two types of the layered constructs, 1L-NHDF/ 1L-HUVEC and 4L-NHDF/1L-HUVEC, were prepared using

approximately 6 nm of FN­G nanofilms. To evaluate general effects of 3D-layered structures on cellular stress or inflammation, we purposely fabricated biologically meaningless layered structures consisting of endothelial cells and fibroblasts. Interestingly, the HUVECs adhered homogeneously on the surface of 4L-NHDFs, and tight-junction formation was widely observed at the centimeter scale, while heterogeneous HUVEC

M. Matsusaki

Bull. Chem. Soc. Jpn. Vol. 85, No. 4 (2012)

IL-6

407

b) Factor IIIV

a) HE

40 μm

c) 1L-NHDF/1L-HUVEC

40 μm

d) Magnified of c)

(IL-6/pg mL−1)/(Total DNA contents/μg)

PE films 8 *p