Cell behaviors on nano- and micropatterns
To confirm the ability of nanopillars to increase cell adhesion on generated on PDMS substrates, C2C12, an immortalized mouse myoblast, was chosen as a model cell line [34]. C2C12 is an adherent cell line and should be cultivated on a surface that is functionalized to promote cell adhesion. Interestingly, periodic nanostructures, including nanogrooves and nanopillars, have proven effective in enhancing cell adhesion even without the use of ECM proteins and peptides [35,36,37,38]. Therefore, in this study, the uniform PDMS nanopillar patterns were fabricated by laser interference lithography, as specified in Additional file 1: Fig. S1 [23]. To investigate the effects of nanopattern sizes on cell adhesion and morphology, nanohole with different sizes (e.g., 400, 600, and 800 nm) were generated on the silicon mold and were reversely replicated using PDMS, as shown in Fig. 2a. Based on the size distribution result (Fig. 2b) and high-resolution vertical SEM image (Fig. 2c), we found that all the PDMS nanopillars were uniformly fabricated. Such uniformity in size and shape is reported to be critical for facilitating cell adhesion to the desired substrates [23].
As shown in the images of the cells double-stained for F-actin and nuclei (Fig. 2d), C2C12 cells on the nanopatterned PDMS showed a more elongated and stretched morphology than those on bare PDMS, regardless of the nanopattern sizes. For intensive analysis of cell behavior on each substrate, we calculated the cell spreading area, circularity, and cell aspect ratio based on multiple actin/DAPI fluorescence cells (Fig. 2e−g). The cell spreading area on nanopatterned PDMS was increased compared with the bare PDMS substrate (Fig. 2e). Specifically, the 600 nm-sized nanopatterned PDMS showed a cell spreading area of 679.7 µm2, which was significantly greater than those of the other substrates (428.2 µm2 for bare PDMS, 616.3 µm2 for 400 nm-sized nanopatterned PDMS, and 535.0 µm2 for 800 nm-sized nanopatterned PDMS).
The cell circularity was considered next. Cells with a circularity of 1 are perfectly circular without any stretch from the surface and structural deformation. By contrast, a circularity of 0 means that the roughness of the outer area of the cell is extremely high, which is considered as cell growth. We found that the highest circularity corresponded to bare PDMS, showing a value of 0.12, followed by the 400-, 600-, and 800 nm-sized nanopatterned substrates, with values of 0.071, 0.062, and 0.064, respectively (Fig. 2f). Similar to the cell spreading results, the 600 nm-sized nanopatterned substrate showed the lowest circularity among all the groups, indicating that more branches stretch out from the cells; that is, the cells show enhanced adhesion. Unlike these two parameters, the cell aspect ratio was not affected by the nanopatterns, as shown in Fig. 2g. Taken together, we found that the 600 nm-sized PDMS nanopillar arrays are highly effective in enhancing cell adhesion and spreading.
In addition to an effort to improve cell attachment to the platform, we further investigated the effects of PDMS microgroove patterns on cell morphology. The fabrication process of the micropattern is described in Additional file 1: Fig. S2. As shown in Fig. 2h, the micropatterns on the fabricated micropatterned PDMS were uniformly aligned in one direction. The size and height of the micropatterns on the PDMS were about 20 and 2 μm, respectively, which is consistent with those of the micropatterns on the silicon wafers. After the C2C12 cells were seeded on the micropatterned PDMS, the cells were immunostained to confirm the alignment of the cells (Fig. 2i). As shown in the confocal microscopy images, aligned microstructures of the C2C12 cells were obtained on the micropatterned PDMS compared with the cells on the bare PDMS. To acquire the quantitative data on the cell alignment from the images, spatial point pattern analysis was undertaken using the J-function in the ImageJ program. An initial color survey of the cells on the substrates was performed (Additional file 1: Fig. S3). It revealed that the cells on the bare PDMS showed randomly arranged structures, and the bar graph demonstrated the absence of any dominant angle of cell alignment (Fig. 2j). By contrast, the cells on the micropatterned PDMS were aligned horizontally, and 73.6% of the intensity was focused within 30° from a dominant line that represented 0° in the graph. Therefore, we could adapt the micropatterned PDMS to fabricate the NMPA for cell alignment.
Cell behavior on various GO-modified substrates
GO have been used as composing materials for various bio-integrated platforms due to biocompatible characteristics and highly adhesive nature of GO [39, 40]. To attach the C2C12 cells to the substrates stably and further control the myogenesis, GO sheets were coated on the substrates through a transfer technique, as previously reported [41]. The GO could be simply modified on the desired substrate via the drop-coating method. However, the large stacked GO sheets would cover the entire polymer surface and could thus eliminate the nanopattern effects. Therefore, three different types of GO, including LGO, 10-sGO, and 5-sGO, were applied to the substrates. The transfer of the GO onto the PDMS substrates is described in Additional file 1: Fig. S4. As shown in Fig. 3a and Additional file 1: Fig. S5, all types of GO were successfully transferred to the PDMS substrates regardless of their size. As a proven cytophilic material, the spreading area of C2C12 cells was increased on all GO-coated substrates (Fig. 3b and c). Specifically, the cell spreading area was calculated to be 633.1, 978.3, 1039.5, and 1154.6 µm2 for bare PDMS, 5-sGO, LGO, and 10-sGO substrates, respectively. Moreover, the circularity values were 0.13, 0.13, and 0.17 for the LGO, 10-sGO, and 5-sGO groups, respectively (Fig. 3d), indicating that GO induces branch-like formations that stretch out from the outer cell membrane area due, in part, to the enhanced filopodia and lamellipodia formation. No significant differences were observed in the cell aspect ratios among all tested groups, proving the excellent cytophilic property of GO that is similar to the periodic nanopatterns.
To further study the cell adhesion strength on GO coated NMPA, the cell adhesion was intentionally weakened via trypsin treatment and centrifugation as chemical and mechanical methods, respectively (Fig. 3f). The trypsin treatment time and rotating speed (i.e., centrifugal force) were optimized as 80 s and 8 min, respectively (Additional file 1: Fig. S6). The centrifugal force applied to the cells was calculated to be 0.269 mN based on the following equation:
$$\text{F}=m\times {\omega }^{2}r$$
(1)
After 80 s of the trypsin treatment, 52.7% of the cells remained on the bare PDMS substrate, while 68.2%, 71.7%, and 69.6% of the cells remained attached to the LGO-, 10-sGO-, and 5-sGO-coated PDMS substrates, respectively (Fig. 3g). In the mechanical detachment experiments, 74.0%, 92.0%, 92.8%, and 95.5% of the cells were found to be attached to the bare PDMS, LGO-, 10-sGO-, and 5-sGO-coated PDMS substrates, respectively, when the centrifugal force was applied for 8 min (Fig. 3h). Taken together, the results indicate that GO is effective in not only enhancing the cell spreading (i.e., cell adhesion area) but also the cell adhesion on the artificial polymer substrates, both of which are incredibly important for in vitro skeletal muscle cell differentiation.
Cell behaviors on GO coated NMPA
After confirming the excellent properties of each core component, including PDMS nanopillars and sGO for improving cell adhesion, and the microgrooves for cell morphology modulation, we attempted to fabricate all combined sG-NMPA (Additional file 1: Fig. S7). As shown in Fig. 4a, the 600 nm-sized nanohole patterns were successfully fabricated on the 20 μm-sized microgrooved Si wafer. Then, PDMS was used to reversely replicate the structure of the nanohole pattern-modified Si wafer. The height of fabricated nanopillars on microgroove was 121.98 nm (Additional file 1: Fig. S8). Afterward, LGO, 10-sGO, and 5-sGO were modified on each nano-, micro-patterned PDMS, and NMPA substrate via the contact-printing method under high humidity. As shown in Fig. 4b, owing to the size of LGO (average size: 0.5−10 μm), they were heavily stacked on the surface, resulting in uniform G-band (i.e., in-plane vibration of sp2-bonded carbon atoms) distribution in the Raman spectrum for all the substrates regardless of their topographical differences. For 10-sGO and 5-sGO, the structural morphology of both microgrooves and nanopillars resulted in variations of G-band intensity (Fig. 4b). Among microgrooves and nanopatterns, the PDMS nanopillars showed superior GO absorption efficiency over the microgrooves-only substrate due to increased hydrophobic property, which lowered the surface energy level. After confirming successful GO transfer to NMPA, C2C12 cells were seeded on each platform and stained with phalloidin and Hoechst to visualize F-actin expression and the nucleus shape, respectively (Fig. 4c). We specifically focused on the alignment and direction of the C2C12 cells, which are important indicators to guide skeletal muscle cell differentiation through cell morphology manipulation. As hypothesized, cells on the bare PDMS substrate showed random orientation and growth with limited cell size. However, the patterned PDMS substrates forced cells to exhibit a unidirectional morphology with a close cell-to-cell network that mimics the structure of in vivo muscle tissue (Fig. 4d). To better compare the cell alignment tendency on each substrate, the degree of alignment was converted to numerical values based on the following equation: [42], when σcell is the standard deviation of cell angles.
$${\text{Cell alignment index = }}\frac{{\left( {180/\sqrt {12} } \right)}}{{{\sigma _{{\text{cell}}}}}}$$
(2)
As shown in Fig. 4e, the calculated cell alignment index of cells on patterned substrates was 2.06, 2.41, 2.21, 2.07, and 2.50 for microgrooves, NMPA, LGO-coated NMPA (LG-NMPA), 10-sGO-coated NMPA (10-sG-NMPA), and 5-sGO-coated NMPA (5-sG-NMPA), respectively. The results indicate that microgrooves are critical in guiding cellular morphology into an elongated shape, while sGO modification does not harm the guiding ability of microgrooves, especially 5-sGO. Regarding the effects of GO modification, cell spreading increased on the bare PDMS nanopattern and the GO-coated substrates, while the circularity decreased for cells from all the substrates compared with bare PDMS, which is consistent with the results we obtained without microgrooves (Figs. 2e−g and 3c−e). Taken together, we can conclude that the developed sG-NMPA are highly effective in both manipulating cell morphology into an elongated shape and enhancing cell adhesion, which is advantageous for the generation of skeletal muscle cell differentiation.
Enhanced skeletal muscle cell differentiation on GO-coated NMPA
To confirm the ability of the fabricated hybrid platform to guide myogenic differentiation, C2C12 cells were finally cultured. As hypothesized, cells on bare PDMS were detached immediately due to the low cell adhesion strength (Additional file 1: Fig. S9, Additional file 2: Video S1). The detachment of C2C12 cells on flat PDMS surfaces usually occurs owing to the high contractility of the cells compared to their adhesion strength to the substrates [43, 44]. Remarkably, by contrast, cells on the GO-coated nanopillar-modified microgrooves were observed to be stable throughout the differentiation. After 10 days of differentiation, cells remaining on PDMS surfaces showed a circular shape, having a low morphological aspect ratio and high circularity. However, the cells on the GO-coated NMPA substrates showed higher cell spreading area and more elongated morphology than the cells on the bare PDMS. After that, the differentiated cells were further stained with two different myogenesis markers, α-actinin and MHC (Fig. 5b). α-Actinin is an actin-binding cytoskeletal protein critical for focal adhesion [45]. MHC is one of the motor proteins in muscle filaments [46]. As shown in Fig. 5b, both α-actinin and MHC were highly expressed in the GO-modified nanopillar groups, especially in sGO-coated substrates. To better compare the skeletal differentiation based on the immunofluorescence images, the mean intensities of MHC expressed as red color were divided by the number of nuclei and plotted. The mean values of the differentiated cells were calculated to be 0.118, 0.178, 0.197, and 0.248 for bare PDMS, LG-NMPA, 10-sG-NMPA, and 5-sG-NMPA, respectively (Fig. 5c). Although the variations were found to be high for all the groups, sG-NMPA showed 10.7% (5 nm) and 39.3% (10 nm) higher MHC expression levels than the LGO-coated nanopillar-modified microgrooves. Based on this observation, we concluded that sGO conserves the distinct nanopillar and microgroove co-existing structure of the fabricated PDMS substrate and also enhances cell adhesion, which results in enhanced skeletal muscle cell differentiation (Fig. 5d).