Introduction

Recently, in vitro cell culture devices made from synthetic polymers have attracted considerable attention [1,2,3,4,5,6]. Generally, adherent cells require suitable scaffolds [7,8,9] and physiologically active substances [10, 11] in a cell-culture environment to maintain their original functions. However, some external humoral factors may cause serious damage to cells [11,12,13]. Therefore, the regulation of cell functions [1, 6, 14,15,16] and the collection of cultured cells [17] by only tuning the structural and physical properties of polymeric scaffolds are important for further advances in the biotechnology and biomedical fields. A typical example of this is to prepare a cell sheet on the basis of the change in aggregation state of a thermoresponsive polymer in a scaffold [17, 18]. Poly(N-isopropylacrylamide) (PNIPAAm) starts to dissolve in water at ~305 K [19]. Thus, a layer of cultured cells on PNIPAAm can be gently detached by cooling them to this temperature, without the use of digestive enzymes or denaturing treatments. This technology contributes considerably to the regenerative medicine field combined with cell engineering using induced pluripotent stem (iPS) cells [20], resulting in the generation of various types of tissues from exfoliated cell sheets.

The physical properties, like the mechanical properties, of polymeric scaffolds are also one of the key factors regulating the functions of cells cultured on the scaffolds [1, 14,15,16, 21,22,23,24,25,26,27]. In most cases, the static elastic modulus of the polymeric scaffolds is set in the kPa range, mimicking a typical extra cellular matrix (ECM) [28, 29]. In addition, the temperature-responsive dynamic elastic change has also been used to control cellular functions [23]. Therein, the morphology of myoblasts on a cross-linked poly(ε-caprolactone) scaffold depended on the crystal-amorphous transition, ranging from a few MPa to several tens of MPa.

The mechanical properties not only in the bulk but also at the surface [30,31,32] play a key role in cell regulation. We recently studied the effect of surface mechanical properties on cellular phenomena for thin films made from typical glassy polymers of polystyrene (PS) [33] and poly(methyl methacrylate) (PMMA) [34]. The initial adhesion of L929 mouse fibroblasts on the films prepared on glass substrates was insensitive to the surface modulus, which ranged from GPa to MPa. However, in the case of bilayer scaffolds composed of glassy polymers on rubbery polyisoprene (PI), the cell adhesion and spreading were dependent on the thickness of the upper glassy layer if it became thinner than a threshold value [33, 34]. This finding could be explained in terms of the manifestation of the mechanical response from the underlying PI phase, called thinning-induced mechanical instability [33]. We further examined this issue using a film of poly[(2-methoxyethyl vinyl ether)-block-(L-lactic acid)] (P(MOVE-b-LLA)), in which the surface was covered with PMOVE [35], and found that fibroblasts were sufficiently sensitive to identify crystalline/non-crystalline regions of PLLA existing underneath the surface PMOVE layer. These results suggest that typical stiff polymers, unlike ECM, can be used as functional cellular scaffolds based on their surface mechanical properties. In this study, we propose a novel scaffold for the realization of patterned cellular behavior on a flat surface on the basis of the mechanical instability of synthetic glassy polymers. This novel methodology for cell patterning enables the potential development of further-functionalized arrays of living cells, which are applicable to cell-based sensors and drug discovery devices in the biomedical field [36, 37].

Experimental

Scaffold preparation

PS with a number-average molecular weight (Mn) of 235k and a molecular weight distribution (Mw/Mn) of 1.05, where Mw is the weight-average molecular weight, was purchased from Polymer Source Inc. (Dorval, QC, Canada). Figure 1 shows the preparation process for a cellular scaffold with 2D-patterned mechanical properties. A spin-coated PS film on a silicon wafer was removed from the wafer and then floated to the surface of ultrapure water purified by a Milli-Q system (Merck KGaA, Darmstadt, Germany) [38]. The thickness of the PS film (dPS) was adjusted from 100 nm to 1 μm. Then, the film was transferred onto an epoxy resin-based line and space (L&S) pattern substrate prepared by photolithography (Tokyo Ohka Kogyo Co. Ltd., Kanagawa, Japan). The height difference and half pitch of the L&S pattern were 2.4 and 40 μm, respectively. The PS/L&S scaffold was dried at a room temperature of approximately 298 K for 24 h under vacuum. The morphology of the L&S patterned substrate before and after the transfer of the PS film was examined by using a LEXT OLS4000 3D laser measuring microscope (Olympus Co., Tokyo, Japan). In addition, sectional views of the scaffold were observed by scanning electron microscopy (SEM, SS-550, Shimadzu Co., Kyoto, Japan).

Fig. 1
figure 1

A schematic illustration of the preparation of a PS/L&S scaffold

Full size image

Cell culturing

Suspensions of mouse fibroblast L929 cells (Cell Engineering Division, RIKEN BioResource Center, Tsukuba, Japan) at 5.0 × 104 cells/well were seeded onto the PS/L&S scaffolds placed on the base of 24-well culture dishes filled with cell culture medium, normal Roswell Park Memorial Institute (RPMI) 1640 medium (Life Technologies Japan Ltd., Tokyo, Japan), with/without 10% fetal bovine serum (FBS, Life Technologies Japan). The cultures were maintained at 310 K (37 °C) in a humidified atmosphere containing 5% CO2. After 4 h of culturing, the scaffold surfaces were rinsed to remove cells floating in the culture medium. The number and morphology of cells on the PS/L&S scaffolds at a given culturing time were evaluated by phase contrast microscopy (BZ-8100, Keyence Co., Osaka, Japan). Additionally, cell proliferation under serum conditions was observed as a function of the culturing time in the same manner. The projected area of cells was quantified by using ImageJ software (National Institutes of Health, USA).

Results and discussion

Scaffold morphology

Figure 2a, b shows 3D-confocal laser scanning microscopic (CLSM) images of the L&S patterned substrate before and after picking up a 100 nm-thick PS film. Although the film was very thin, the top surface was quite flat without any defects, as shown in Fig. 2b. Figure 2c, d shows sectional views of the L&S patterned substrate before and after the transfer of the 100 nm-thick PS film observed by SEM. As mentioned above, the manifestation of thinning-induced mechanical instability for thin polymer films is strongly dependent on the underlying region [32]. Considering that the underlying region of the PS film is composed of either the epoxy resin with an elastic modulus of 3.0 GPa or air, hereafter referred to as the L- or S-regions, respectively, it can be posited that this scaffold possesses 2D-patterned mechanical properties.

Fig. 2
figure 2

a, b 3D-CLSM images of the L&S patterned substrate a before and b after the transfer of a 100 nm-thick PS film. The half pitch of the L&S patterned substrate was 40 μm, and the widths of the L- and S-regions were the same. The height difference between the L- and S-regions was 2.4 μm. c, d Sectional views of the L&S patterned substrate from SEM c before and d after the transfer of the thin PS film

Full size image

Cell adhesion under serum-free conditions

First, cell culturing on the PS/L&S scaffold was performed under serum-free conditions, that is, the integrin-independent interaction between the cells and polymer scaffolds was examined [39,40,41]. Figure 3a, b shows the phase-contrast images of L929 cells cultured on the PS/L&S scaffolds with different dPS values for 4 h. Although the dividing lines between the L- and S-regions can be discerned in the images, the scaffold geometry is also shown for easy understanding. While the cell adhesion on the 1 μm-thick PS scaffold was homogeneous over the 2D mechanical pattern, as shown in Fig. 3b, more cells were attached to the L-regions than to the S-regions of the 100 nm-thick PS scaffold, as shown in Fig. 3a. Figure 3c shows the dPS dependence of the number of cells adhered on the L- and S-regions (NL and NS), and the ratio of the mean NL to mean NS values is shown in Fig. 3d. When the PS film was thicker than 800 nm, the ratio was independent of the underlying L&S pattern. On the other hand, when the dPS fell short of  400 nm, NL was greater than NS, and their difference showed a tendency to increase with decreasing dPS. These points make it clear that L929 cells can recognize the 2D mechanical pattern of the scaffolds.

Fig. 3
figure 3

Phase-contrast images of L929 fibroblasts attached to the PS/L&S scaffolds with dPS values of a 100 nm and b 1 μm, respectively, under serum-free conditions. Scale bars correspond to 50 μm. Schematic illustrations of the scaffolds accompany the images. c The number of cells adhered on the L- and S-regions (NL and NS). Data are the mean values with standard deviations. *p < 0.05 (Student’s t-test). d The dPS dependence of the ratio of mean NL to mean NS values (NL/NS)

Full size image

Figure 4a, b shows the projected area of cells attached to the PS/L&S scaffolds with dPS values of 100 nm and 1 μm, respectively. The peak position and full width at half maximum (FWHM) obtained from the normal distribution curves are summarized in Table 1. While the cell spreading was not dependent on the underlying L&S pattern for the thicker PS scaffolds, cells attached onto the L-regions of the thinner scaffolds spread well in comparison with those attached to the S-regions. Thus, these results demonstrate again that the cell adhesion and spreading at the initial stage under serum-free conditions was somehow controlled by the L&S pattern if the upper PS film was sufficiently thin.

Fig. 4
figure 4

The abundance of the projected area of cells adhered on the L- and S-regions for scaffolds with dPS values of a 100 nm and b 1 μm. Normal distribution curves are superimposed for reference

Full size image
Table 1 Peak position and full width at half maximum (FWHM) with the standard deviation for the normal distribution curves in Fig. 4
Full size table

Cell adhesion and proliferation under serum conditions

Second, cellular behaviors were monitored for a longer time under serum conditions to examine the impact of surface mechanical properties on cell proliferation as well as adhesion [11]. Figure 5 shows the phase-contrast images of L929 fibroblasts cultured on the PS/L&S scaffolds with different dPS values for (a, b) 4 h and (c, d) 72 h. In the case of the 4 h cell culture, while the cell adhesion was not dependent on the 2D mechanical pattern for the 1 μm-thick PS scaffold, it was slightly dependent for the 100 nm-thick PS one. This trend became clear after 72 h, as shown in Fig. 5c, d. Figure 5e, f shows the number of cells adhered to the L- and S-regions after 4 h and 72 h of culture, respectively. Figure 5g shows the culturing time dependence of NL/NS. In the case of 4 h, the NL/NS ratio was almost unity for the scaffolds with a PS layer thicker than 800 nm and was slightly higher for the scaffolds with a thinner PS layer. This cell adhesion behavior at the initial stage was similar to that observed under serum-free conditions. When the culturing time increased to a few days for the scaffolds with a PS layer thicker than 800 nm, both NL and NS increased to the same amount, that is, L929 cells uniformly attached to and proliferated on the scaffold surface. Conversely, when the upper PS layer became thinner than 400 nm, NL was clearly greater than NS, and this difference became more remarkable as the culturing time increased. These points make it clear that fibroblast proliferation and migration as well as adhesion could be regulated by the 2D mechanical pattern of the scaffold.

Fig. 5
figure 5

Phase contrast images of L929 fibroblasts attached to the PS/L&S scaffolds with dPS values of a, c 100 nm and b, d 1 μm cultured for a, b 4 h and c, d 72 h, respectively, in medium containing 10% FBS. Scale bars correspond to 50 μm. e, f The dPS dependence of the number (N) of cells adhered on the L- and S-regions (NS and NL) for a culturing time of e 4 h and f 72 h, respectively. Data are the mean values with standard deviations. *p < 0.05; **p < 0.005 (Student’s t-test). g Culturing time dependence of NL/NS

Full size image

It has been reported that integrin-mediated cell adhesion causes conformational changes in mechanosensitive proteins owing to the induced tension, which are involved in changing the connections between the ECM and actin cytoskeleton, so-called actin remodeling [42,43,44,45]. As a consequence, the signal transduction pathway is activated to alter cellular behaviors such as proliferation and migration. Taking this into account, a possible explanation for the cellular behaviors observed here is that the difference in the mechanical response between the S- and L-regions of the scaffold could lead to a different level of activation for actin remodeling on these regions. Another possible explanation is the gentle oscillation of the PS thin film on the S-region, which is always in contact with the culture medium. This oscillation does not occur on the L-region due to the foundation of the epoxy resin. Such a situation may also contribute to the different level of activation for actin remodeling. A more conclusive study of this issue with finite element analysis is currently underway.

Conclusions

A novel concept for cellular scaffolds with 2D-patterned mechanical properties is proposed. Initial cell adhesion and spreading under serum-free conditions on the PS/L&S scaffold were suppressed on the S-regions compared to those on the L-regions once the upper PS film became thinner than a threshold value. In addition, under serum conditions, cell adhesion and proliferation were affected by the underlying L&S pattern with a thinner PS layer. Cellular behaviors on scaffolds composed of a common glassy polymer can be regulated on the basis of the physical properties of the scaffold, especially near the outermost region. We believe that this novel concept of cellular scaffolds will lead to the development of highly functional cellular scaffolds composed of synthetic polymers.