Generation of microgrooved silica nanotube membranes with sustained drug delivery and cell contact guidance ability by using a Teflon microfluidic chip.

Silica nanotubes have been extensively applied in the biomedical field. However, very little attention has been paid to the fabrication and application of micropatterned silica nanotubes. In the present study, microgrooved silica nanotube membranes were fabricated in situ by microgrooving silica-coated collagen hybrid fibril hydrogels in a Teflon microfluidic chip followed by calcination for removal of collagen fibrils. Scanning electron microscopy images showed that the resulting silica nanotube membranes displayed a typical microgroove/ridge surface topography with ∼50 μm microgroove width and ∼120 μm ridge width. They supported adsorption of bone morphogenetic protein 2 (BMP-2) and exhibited a sustained release behavior for BMP-2. After culturing with osteoblast MC3T3-E1 cells, they induced an enhanced osteoblast differentiation due to the release of biologically active BMP-2 and a strong contact guidance ability to directly align and elongate osteoblasts due to the presence of microgrooved surface topography, indicating their potential application as a multi-functional cell-supporting matrix for tissue generation.

Introduction

Cell behaviors in the native microenvironment are highly guided by many nano-/micro-scale structures such as highly aligned collagen fibrils which are the main structural components of the extracellular matrix (ECM) [1]. Mimicking in vitro such an organized and fibrous structure is very important in designing ideal biomaterials presenting similar in vivo cell behavior. Nanofibrous materials have thus attracted much interest for the construction of artificial cell-supporting matrixes for tissue generation [2].

Silica-derived forms have attracted considerable attention for the development of artificial bone substitutes since the discovery of bioglass, which consists of SiO2, CaO, P2O5 and Na2O [3]. Their silanol groups spontaneously induce the deposition of bone-like apatite in the human plasma, promoting osteointegration [4, 5], and exhibit excellent biocompatibility, supporting cell attachment and proliferation [6, 7]. To date, many kinds of silica-derived forms that include spheres [8], membranes [6], xerogels [9], scaffolds [10] and nanotubes [11] have been synthesized and applied to stimulate bone regeneration. Among them, silica nanotubes have attracted special interest since they not only present hollow structural features for delivering various anti-cancer drugs [12], DNA [13] and biological growth factors [14], but also display a typical one-dimensional ECM-like fibrous structural feature to support cell attachment and proliferation [11, 14]. However, despite the prominence of silica nanotubes in biomedical applications, very little attention has been paid to the fabrication and applications of silica nanotubes with highly organized structure.

Currently, several template routes are available for the fabrication of silica nanotubes, including carbon nanotubes [15], vertical silicon nanowires [16] and reverse microemulsion [17]. However, these conventional template routes are not suitable for the fabrication of silica nanotubes with a highly organized structure. Micropatterning is a powerful technique for creating highly organized structures in biomaterials and has been used for micropatterning various materials including polymers and hydrogels [18]. Hydrogels are most popular in the fabrication of micropatterned materials since they are mechanically flexible and can be easily processed with many kinds of highly organized surface morphologies [18, 19]. Microgrooved patterns are one of the most popular micropatterned surface topographies because of their well-defined and organized structure; they are easily directed by soft-/photo-lithographic techniques and strong cell contact guidance ability [7]. Collagen fibrils are one of the organic structural components of extracellular matrices in living tissue and exhibit a strong affinity for silica species via hydrogen bonding and electrical interactions for fabricating silica nanotubes [11, 14]. Therefore, when the collagen-based hydrogels are microgrooved, it is expected that the microgrooved collagen-based hydrogels could serve as an efficient template for in situ fabrication of free-standing and microgrooved silica nanotubes.

In the present study, microgrooved silica nanotube membranes were generated by microgrooving silica-coated collagen hybrid fibril hydrogels in a Teflon microfluidic chip followed by calcination for the removal of collagen fibrils. Their drug delivery capacity for bone morphogenetic protein 2 (BMP-2) and contact guidance ability for osteoblast MC3T3-E1 cells were evaluated.

Materials and methods

Fabrication of a Teflon microfluidic chip

A Teflon microfluidic chip was fabricated by the method described in our previous study [20]. A polydimethylsiloxane (PDMS) master was fabricated by curing the PDMS prepolymer onto the photoresist patterns produced with standard photolithography and tightly sealed to a glass slide. Subsequently, the semicrystallized polymer perfluoroalkoxy (Teflon PFA) substrate (Yuyisong Inc., China) was sandwiched between the PDMS master and another flat glass slide. The sandwich assembly was placed on a hot compressor (TM-101F, Taiming, Inc., China), embossed at 275 °C for 2 min and cooled to room temperature. The Teflon chip was obtained after the PDMS master and glass slide were removed.

Fabrication of a microgrooved silica nanotube membrane

Silica-coated collagen hybrid fibril hydrogels were fabricated by the method described previously [11, 14], by coating collagen fibrils with silica in a Stöber-type sol–gel system consisting of tetraethylorthosilicate (TEOS; 1 ml), ethanol (9 ml), water (9 ml) and ammonia (28%, 0.5 ml) at room temperature for 24 h. Subsequently, they were placed on either a microgrooved PDMS microfluidic chip or a microgrooved Teflon microfluidic chip (1 cm × 1 cm) and subjected to a loading force of 5 N at 50 °C for 2 h to produce the microgrooved silica-coated collagen hybrid fibril membranes. After calcination at 600 °C for 2 h, the collagen fibrils were completely removed from the silica-coated collagen hybrid fibril membranes and the microgrooved silica nanotube membranes were obtained in situ. The flat silica nanotube membranes were prepared by a similar method using the flat Teflon chips. Microstructures of the resulting silica nanotube membranes were observed under a scanning electron microscope (SEM; JEOL-6500F, JEOL, Tokyo, Japan).

BMP-2 loading and release in the microgrooved silica nanotube membrane

To evaluate the capability of microgrooved silica nanotube membranes for drug delivery, BMP-2 was selected as the model drug since it is a biological growth factor widely used for stimulating osteoblast differentiation [6, 14]. First, 100 μl of 0.5 μg ml−1 BMP-2 (Peprotech, Rocky Hill, NJ, USA) solution was added to the microgrooved silica nanotube membranes and then freeze-drying was performed to produce the corresponding BMP-2-loaded membranes containing 50 ng of BMP-2. For the BMP-2 release experiments, the BMP-2-loaded samples were soaked in phosphate buffer saline (PBS) and the amount of BMP-2 released from samples was measured via an ELISA kit (ELISA; R&D Systems).

Cell culture

Osteoblast-like MC3T3-E1 cells (RIKEN, Ibaraki, Japan) were seeded on the resulting samples in a 24-well plate at a density of 2 × 104 cm−2 and cultured in α-minimum essential medium (α-MEM) containing 10% fetal bovine serum, 100 U ml−1 penicillin, and 100 μg ml−1 streptomycin in a humidified atmosphere of 5% CO2 and 95% air at 37 °C. To induce osteoblast differentiation, cell-laden samples were cultured in a differentiation culture medium containing 10% fetal bovine serum, 100 U ml−1 penicillin, 100 μg ml−1 streptomycin, 2 mM β-glycerophosphate and 50 μg ml−1 sodium ascorbate.

WST-1 assay

Cell viability and proliferation were evaluated via a conventional water-soluble tetrazolium salts (WST-1) assay. The samples were taken out and transferred to a fresh 24-well plate. Then, 200 μl of a WST-1/culture medium (1:10) mixture was added to each well and incubated at 37 °C. After 3 h, 100 μl of the liquid was taken out and transferred to a 96-well plate. The absorbance of the liquid at 450 nm was measured using a microplate reader (MTP-880, Corona Electric Co Ltd, Japan).

Alkaline phosphatase assay

Alkaline phosphatase (ALP) is a typical marker during the early stage of bone cell differentiation [21]. To evaluate the biological activity of BMP-2 released from the microgrooved silica nanotube membranes, ALP in the cells seeded on the samples was measured via an ALP assay. The samples were taken out, transferred to a fresh 24-well plate and washed twice with 0.9% NaCl. Subsequently, 200 μl of 0.1% Triton in 0.9% NaCl solution was added to each well. After 10 min, 25 μl of the cell lysis solution was taken out, transferred to a 96-well plate and mixed with 50 μl of ALP working solution at 37 °C for 15 min. Then, 25 μl of NaOH solution were added to stop the reaction. The absorbance of the solution at 415 nm was read using the above microplate reader.

Immunochemical assay

To observe nuclei and F-actin in the cells on the samples, the cell-laden samples were fixed in 3.7% paraformaldehyde for 10 min at 37 °C, permeabilized in 0.1% Triton X-100 for 3 min, and stained with 4′,6-diamidino-2-phenylindole for nuclei and Alexa Fluor 594 phalloidin for F-actin filaments. To quantify cell alignment and elongation on the samples, the angle between the elliptic axis of the nuclei and the major axis of the microgroove/ridge was measured using NIH ImageJ software. Cell alignment was defined as an angle of less than 10° [22]. To evaluate cellular elongation, circularity of the nuclei was used as the shape index of each individual cell and measured using the same software. Circularity was defined as 4π × area/(perimeter)2, which has a maximum value of 1 for a circle.

Statistical analysis

All the results in triplicate were expressed as mean and standard deviation (SD) and analyzed using one-way analysis of variance with a significance level of p < 0.05.

Results and discussion

PDMS microfluidic chips are most commonly used in microfabricating various polymers and hydrogels because of their excellent flexibility in mechanical properties [23]. After collagen hydrogels were soaked in the sol–gel system for a coating of silica, the resulting silica-coated collagen hybrid fibril hydrogels remained soft. Therefore, it is expected that they could be microgrooved using the PDMS chip. When the silica-coated collagen hybrid fibril hydrogels were microgrooved and removed from the PDMS chip, it was found that some silica-coated collagen hybrid fibrils remained in the microgroove of the PDMS chip (figure 1(a)). This indicates that PDMS chips have a strong affinity for the silica species and this affinity resulted in mass loss of samples. To overcome this problem, the conventional PDMS chips were replaced with novel Teflon chips for microgrooving silica-coated collagen hybrid fibril hydrogels [24]. From figure 1(b), no silica-coated collagen fibrils were found in the microgroove of the Teflon microfluidic chip, indicating that the Teflon microfluidic chip was much more suitable for microgrooving silica-based materials.

Figure 1.

Phase contrast images of (a) a conventional PDMS microfluidic chip and (b) a Teflon microfluidic chip after both were used for microgrooving silica-coated collagen hybrid fibril hydrogels. Arrows indicate the presence of silica-coated collagen hybrid fibrils.

Figure 2 shows SEM images of silica nanotube membranes produced on (a) a flat Teflon chip, (b) a microgrooved PDMS chip and (c) a microgrooved Teflon chip. It was found that the surface morphology of the samples depended on the type of microfluidic chips. In the case of a flat Teflon chip, the silica nanotube membranes exhibited a similar flat and smooth surface (figure 2(a)) with a porous and fibrous structure (inset) owing to the template role of the collagen fibrils, as was found in the previous studies [11, 14]. In the case of a microgrooved PDMS chip, no clear microgroove structure was found in figure 2(b). This is because PDMS is mechanically flexible and deforms under pressure. Therefore, it cannot effectively maintain its surface microstructure under pressure for further microgrooving silica-coated collagen hybrid fibril hydrogels. In contrast, the silica nanotube membranes in figure 2(c) exhibits a typical microgrooved/ridge surface topography (about 50 μm microgroove width and about 120 μm ridge width), further showing that the Teflon chip is much better than the PDMS chip in fabricating silica-based materials because of their excellent mechanical stability [24]. The inset in figure 2(c) shows that the microstructure of silica nanotubes is not influenced by microgrooved patterns and all of them exhibit a typically fibrous and hollow structure. In addition, it was found that collagen fibril hydrogels could not be directly microgrooved in the present case because of their rapid biodegradation.

Figure 2.

SEM images of silica nanotube membranes produced on (a) a flat Teflon chip, (b) a microgrooved PDMS chip and (c) a microgrooved Teflon chip.

By virtue of the highly organized and fibrous ECM-like structure, microgrooved silica nanotube membranes (figure 2(c)) have potential application as a cell-supporting matrix. Biological factors released from the cell-supporting matrix strongly stimulate cell differentiation in the tissue regenerative field [1]. We evaluated the microgrooved silica nanotube membrane for delivery of BMP-2. Figure 3 shows a time-course release profile for BMP-2 from the microgrooved silica nanotube membranes after soaking in PBS up to 2 weeks. The amount of BMP-2 released from the sample was 13 ± 2% at day 1, 28 ± 1% at day 3, 45 ± 3% at day 5, 57 ± 4% at day 7 and 75 ± 3% at day 14. The BMP-2 exhibited a sustained release behavior, and its amount increased as the soaking time increased over 2 weeks, indicating that the microgrooved silica nanotube membranes could serve as a sustained drug delivery system, as was found for the flat silica nanotube membranes in previous studies [14].

Figure 3.

Release profile of BMP-2 from the microgrooved silica nanotube membranes after soaking in PBS for 2 weeks.

In vitro biocompatibility of biomaterials is critically important in supporting cell attachment and proliferation. To evaluate in vitro biocompatibility and bioactivity of BMP-2, osteoblast MC3T3-E1 cells were seeded on the flat silica nanotube membrane and microgrooved silica nanotube membrane with and without BMP-2. Figure 4 shows the WST-1 assay result of osteoblast MC3T3-E1 cells seeded on those samples after 1 and 7 days of culturing. As the culturing time increased, the absorbance value increased for each sample, indicating that all samples supported cell attachment and proliferation and exhibited good biocompatibility. However, cell viability depended on the type of sample. At day 1, the cell viabilities of all samples were similar, despite the difference in surface topography and presence of BMP-2. At day 7, the cell viabilities on both flat and microgrooved samples without BMP-2 were very similar and higher than that on microgrooved samples with BMP-2, indicating that the difference in surface topography had no significant effect on cell viability and proliferation and the presence of BMP-2 inhibited cell proliferation with increasing culturing time.

Figure 4.

Result of the WST-1 assay for osteoblast MC3T3-E1 cells seeded on the flat and microgrooved silica nanotube membranes without and with BMP-2 (error bars: ±SD; ∗p < 0.05).

Figure 5 shows the ALP activity of osteoblast MC3T3-E1 cells seeded on microgrooved silica nanotubes without and with BMP-2 after 1, 7 and 14 days of culturing. As the culturing time increased, the absorbance value for each sample increased, indicating that both samples supported osteoblast differentiation. At day 1, both samples exhibited a similar ALP value and no significant difference was found. At day 7, the ALP activity was much higher due to the presence of BMP-2, indicating that the osteoblast MC3T3-E1 cells started to differentiate and their proliferation was slowed down. This is consistent with the results in figure 4. At day 14, a significant difference in ALP activity was observed and this confirmed that BMP-2 induced cell differentiation. That is, the BMP-2 released from the silica nanotube membrane exhibited in vitro bioactivity and the microgrooved silica nanotube membranes maintained the bioactivity of BMP-2.

Figure 5.

ALP activity in the osteoblast MC3T3-E1 cells seeded on the microgrooved silica nanotube membranes without and with BMP-2 (error bars: ±SD; ∗p < 0.05).

To evaluate cell contact guidance ability, the alignment and elongation of an osteoblast on the flat and microgrooved silica nanotube membranes were evaluated. To observe the cell size and morphology, the viable osteoblast MC3T3-E1 cells were stained with calcein-acetoxymethyl ester (calcein-AM) and excited green fluorescence. Figure 6 shows fluorescence microscopy images of MC3T3-E1 cells on the samples after 1 and 3 days of culturing. Note that viable cells were attached to all samples, and they proliferated, indicating that the resulting silica nanotube membranes had good biocompatibility, despite the difference in surface topography, and could serve as substrates to support cell attachment and proliferation. This result is in agreement with the results of figure 4. However, a significant difference in surface morphology was found. At day 1, the cells on the flat ones were randomly distributed, while that on the microgrooved ones were highly aligned and elongated, indicating good contact guidance ability of the microgrooved silica nanotube membranes. At day 3, the number of cells on all samples significantly increased and the cells grew on the whole surface including the ridge and microgroove, further demonstrating that the samples supported cell proliferation (figure 4). Moreover, the cells maintained a similar difference in size and morphology as on day 1, i.e., randomly distributed on flat ones, and highly aligned and elongated on microgrooved ones. This was evidence of the good contact guidance ability of microgrooved silica nanotube membranes.

Figure 6.

Fluorescence microscopy images of osteoblast MC3T3-E1 cells (green) seeded on the flat and microgrooved silica nanotube membranes.

Cell alignment is highly correlated with the alignment of F-actin in the cells [22]. Figure 7 shows representative images of the F-actin distribution in cells on the flat and microgrooved silica nanotube membranes after 1, 7 and 14 days of culturing. F-actin is strongly associated with cell morphology and was observed on both samples, indicating that there was a strong interaction between cells and silica nanotubes. However, there was a significant difference in F-actin distributions between both samples. At day 1, F-actin was randomly distributed on flat ones, but highly aligned and elongated on microgrooved ones. At day 7, much more F-actin was produced in the cells for both samples; it was widely distributed on flat ones and highly aligned and elongated on microgrooved ones. At day 14, F-actins exhibited a similar distribution to those at day 1 and day 7. These results indicated that the microgrooved silica nanotube membranes maintained their alignment for more than 2 weeks, which is further evidence of their strong contact guidance ability.

Figure 7.

Fluorescence microscopy images of F-actin (red) in osteoblast MC3T3-E1 cells seeded on the flat and microgrooved silica nanotube membranes.

Figure 8 also shows the aligned degree and circularity of the nuclei in the cells on flat and microgrooved silica nanotube membranes after 1 day of culturing. The alignment degree (figure 8(a)) was 10 ± 1% for the flat ones and 45 ± 4% for the microgrooved ones, while the circularity (figure 8(b)) was 0.67 ± 0.01% for the flat ones and 0.45 ± 0.05% for the microgrooved ones. In contrast, the presence of the microgrooved structure resulted in a significant increase in cell alignment and elongation. Figures 8(c) and (d) display histograms of cell alignment angles in 10° increments. These histograms further demonstrate that the microgrooved silica nanotube membranes exhibited better cell alignment and elongation in comparison with the flat silica nanotube membranes.

Figure 8.

Alignment and elongation of osteoblast MC3T3-E1 cells seeded on the flat and microgrooved silica nanotube membranes after 1 day of culturing. (a) Mean percentage of aligned cell nuclei (within 10° of the preferred nuclear orientation); (b) mean nuclear shape index; (c) and (d) histograms of cell alignment angles in 10° increments (error bars: ±SD; ∗∗p < 0.01).

Microgrooved patterns have been widely used to direct cellular alignment and elongation. It is found that the dimension of the microgrooved patterns could significantly affect this behavior. Nagamine et al [25] reported that cellular alignment and elongation could be directed on the microgrooved fibrin hydrogels with the microgroove width ranging from 100 to 250 μm. In our previous study, we have demonstrated that microgrooved patterns consisting of 100 μm groove width and ridge width ranging from 50 to 100 μm could direct the cellular alignment and elongation [26]. The dimension of the microgrooved silica nanotube membranes belongs to those dimensions and the cellular alignment and elongation could be thus observed on both the ridge and the microgroove. Although the specific mechanism for such a cellular behavior is still unclear, it is reported that the microgrooved patterns provided a mechanical restriction on the formation of certain linear bundles of microfilaments to push cellular alignment and elongation [27, 28].

Conclusions

In summary, microgrooved silica nanotube membranes were successfully fabricated using a Teflon microfluidic chip. Silica-coated collagen fibrils were synthesized, microgrooved in a Teflon microfluidic chip, and then calcined for in situ fabrication of the microgrooved silica nanotube membranes due to the removal of collagen fibrils. Compared with the conventional PDMS microfluidic chip, the Teflon microfluidic chip was much better in producing microgrooved silica nanotubes. The resulting microgrooved silica nanotube membranes exhibited not only a sustained release behavior for BMP-2, but also a strong contact guidance ability to induce alignment and elongation of osteoblasts. Our results indicated that these microgrooved silica nanotube membranes can potentially be used as multifunctional biomaterials.