Curved microstructures promote osteogenesis of mesenchymal stem cells via the RhoA/ROCK pathway

Abstract

Objectives

Cells in the osteon reside in a curved space, accordingly, the curvature of the microenvironment is an important geometric feature in bone formation. However, it is not clear how curved microstructures affect cellular behaviour in bone tissue.

Materials and methods

Rat primary bone marrow mesenchymal stem cells (BMSCs) on wavy microgrooves were exposed to PDMS substrates with various curvatures to investigate alterations in cellular morphology and osteogenic differentiation. Additionally, the expression levels of RhoA and its effectors were examined by immunofluorescence and quantitative PCR to determine the mechanisms of curvature‐dependent osteogenic differentiation.

Results

Wavy microgrooves caused dramatic nuclear distortion and cytoskeletal remodelling. We detected a noticeable increase in the expression of osteogenic‐related genes in BMSCs in wavy microgroove groups, and the maximum expression was observed in the high curvature group. Moreover, immunofluorescent staining and quantitative RT‐PCR results for RhoA and its effectors showed that the RhoA/ROCK signalling pathway is associated with curvature‐dependent osteogenic differentiation.

Conclusions

Our results illustrated that curved microstructures could promote BMSC differentiation to the osteogenic lineage, and the osteogenic effects of higher curvature are more obvious. Wavy microstructures could also influence the RhoA/ROCK pathway. Accordingly, curved microstructures may be useful in bone tissue engineering.

1. INTRODUCTION

The incidence of bone diseases caused by infections, tumours and bone loss requiring bone regeneration is increasing.1 Bone tissue engineering techniques have played an indispensable role in the reconstruction of bone defects. However, tissue engineering is a complicated process influenced by the migration, recruitment, proliferation and differentiation of osteoprogenitor cells.2 The fundamental components for tissue engineering are seed cells, scaffolds, and growth factors.3 Bone tissue engineering scaffolds are normally made of porous biodegradable materials able to provide mechanical support to seed cells. Bone scaffolds are effective and feasible for the repair and regeneration of bone defects. Porous, biodegradable scaffolds with tailored properties are available owing to the emergence of novel materials and innovative techniques.4

Cell and biomaterial interactions, a principal subject in the field of biomedical materials, can be promoted using chemical,5, 6, 7 biological,8 and physical approaches. Physical and topographical cues are not only crucial factors in tissue development, but have also emerged as important parameters in the regulation of cellular behaviors.9, 10 Cells recognize and discriminate topographical signals from the extracellular matrix (ECM) at both micrometer and nanometer scales.11, 12 With the development of micro‐machining techniques, there have been great improvements in the fabrication of a variety of surface topographies. The topographical scale (micro or nano), type (pillar, pit or groove) and distribution (random or regular) possess different characteristics on affecting cellular behaviors.13

Topographical cues are extremely important with respect to matrix formation or functional bone repair. Natural bone is a sophisticated composite with a hierarchical structure. It consists of two main parts, cortical bone and trabecular bone. The osteon, also called the Haversian system, is the basic functional unit of cortical bone. Each osteon consists of 5–20 layers of concentrically arranged lamellae of the bone matrix, and a central canal (the Haversian canal). The average diameter of the osteon is approximately 100–500 μm. The osteon is formed by collagenous fibre and calcium phosphate.14, 15 Moreover, the Haversian system in the long bone and the periosteum are comprised of highly oriented collagen fibres.16, 17 The properties of natural bone are strongly dependent on the hierarchical structure and organization of the ECM.18 Thus, from the viewpoint of bionics, it may be beneficial to fabricate bone tissue‐like topographies on biomaterials for bone tissue engineering.

In the past few decades, various surface topographies have been fabricated to promote the osteogenic process. Zhang et al. reported that oriented microgrooves, which mimic the collagen fibres of the Haversian system and periosteum, have the ability to enhance fracture healing.17 Wang et al. revealed that the concentric microgrooves which mimic osteon could have significant effects on MC3T3‐E1 cell mineralization and gene expression.19, 20 We know that cells in Haversian system reside in curved space; therefore, the curvature of the microenvironment is an important geometric feature in bone formation. However, few studies have focused on curved microstructures, despite their wide distribution in bone tissue.

Several intracellular signalling pathways recruited by transmembrane receptors to regulate cellular behaviours in response to topographical cues have been identified.21, 22 In particular, RhoA, a small GTPase protein in the Ras superfamily, and its effectors serve as key mediators in mechano‐transduction via cytoskeletal remodelling.23 Moreover, RhoA and its downstream effector ROCK regulate the differentiation of stem cells into osteogenic lineages.24, 25, 26, 27 Based on these previous results, we postulated that the RhoA/ROCK signalling pathway is involved in osteogenic differentiation induced by curved microstructures.

Here, BMSCs, crucial cells in bone tissue and bone regeneration, were used as a model to evaluate cellular morphology and osteogenic differentiation. Polydimethylsiloxane (PDMS) was used as a substrate material owing its light‐emitting quality, biological compatibility and detailed reproduction ability. Microgrooved substrates with constant curvature were fabricated by replica moulding from a customized silicon wafer. Additionally, we examined the mechanisms of curvature‐induced osteogenic differentiation and the RhoA/ROCK signalling pathways. To the best of our knowledge, this is the first report of the influence of curved microstructures on osteogenic differentiation and the underlying mechanism. These findings not only improve our understanding of the effects of curved microstructures on cell behaviour, but also provide a valuable basis for the design of biomaterials for applications in bionics.

2. MATERIALS AND METHODS

2.1. Cell culture

BMSCs were isolated from Sprague–Dawley rats (2 weeks old) using the whole marrow method. Briefly, euthanized rats were sterilized with alcohol and the femur and tibia were separated from the muscle. The metaphyses were removed, followed by repeated irrigation of the marrow cavity with 10% FBS‐supplemented low‐glucose DMEM (L‐DMEM). Bone marrow cells were collected in suspension in a culture dish and incubated at 37°C in a 5% CO₂ incubator in accordance with routine culture methods. Half of the medium was exchanged at 48 h, and all medium was exchanged at 72 h. At 80−90% confluence, the cells were trypsinized and subcultured. Purified BMSCs could be obtained after cells were subcultured two times.

2.2. Fabrication of wavy microgroove substrates

PDMS was synthesized by stirring a mixture of liquid oligomeric base and Sylgard184 (Corning, Inc., Corning, NY, USA) at a ratio of 1:10. Then, the solution was poured into micro‐fabricated silicon wafers (customized by Capital Bio Corporation, Beijing, China) and maintained in a 65°C oven for 4 h to allow cross‐linking. Subsequently, substrates with a wavy microgroove surface were acquired by peeling off the cross‐linked PDMS from the customized silicon wafers. The micro‐patterned PDMS substrates were cut into pieces to fit the size of 6‐well culture plates (Corning). The polymer substrates were then treated with tris(hydroxymethyl) aminomethane (Adamas, Emeryville, CA, USA) and dopamine solution (pH 8) for 72 h to accelerate cellular adhesion. The substrates were sterilized using UV disinfection lamps for 24 h and rinsed with phosphate‐buffered saline in triplicate before cell culture. To explore the influence of curvature on cellular behaviour, BMSCs were cultured in wavy microgrooves with high and low curvatures. Flat PDMS substrates served as controls.

2.3. Osteogenic induction

The digested cells were seeded at a density of 2 × 10⁴ cells/cm² on PDMS substrates in a 6‐well culture plates (Corning). Then, 2 mL of L‐DMEM was added to each well. One day later, the complete medium was carefully aspirated from culture plates (Corning). Osteogenic Induction Differentiation Complete Medium (Cyagen, Chicago, IL, USA) was preheated to 37°C and 2 mL was added to each well of the six‐well plates.

2.4. Alignment and deformation of cell nuclei

Circularity was used as a metric of nuclear deformation. Briefly, the area of the nucleus and nuclear perimeter were estimated for the same cell using ImageJ (NIH, Bethesda, MD, USA). Cell nucleus circularity was calculated based on immunofluorescent staining images according to C =$4\mathrm{\pi }A/P^2$, where A represents to the area and P represents the perimeter. If C = 1, the shape was considered a perfect circle. As the deviation from 1 increased, distortion of the geometric figure increased.28 The mean and standard deviation were calculated. All experiments were performed at least three times.

2.5. Scanning electron microscopy (SEM)

Micro‐patterned substrates and flat PDMS were analyzed by observing the surface and vertical cross‐sections of samples based on scanning electron micrographs. After PDMS was cross‐linked and peeled away from the silicon wafers, the specimens were treated by desiccation, sputter‐coated with gold, and observed by SEM.29

2.6. Immunofluorescence staining

BMSCs on micro‐patterned PDMS were washed with phosphate‐buffered saline three times after inoculation. Then, they were treated with 4% paraformaldehyde for fixation. Samples were permeabilized with Triton X‐100. After they were blocked with goat serum for 1 h at 37°C, samples were incubated in primary antibody solution (anti‐RhoA, ab187027 and anti‐RUNX2, ab23981; Abcam, Cambridge, UK), at 4°C overnight. Subsequently, cells were treated with a donkey anti‐rabbit antibody (A21207; Invitrogen, Carlsbad, CA, USA) at 1:500 for 1 h at 37°C. Next, the samples were treated with rhodamine phalloidin and DAPI to visualize F‐actin and nuclei of BMSCs.30, 31 Immunofluorescence micrographs were obtained using a confocal laser scanning microscope (Leica TCS SP8; Wetzlar, Germany).

2.7. Real‐time quantitative PCR (qPCR)

Briefly, BMSCs were cultured on PDMS substrates with various micropatterns and cultured for 3 days. mRNA was extracted from BMSCs using TRIzol (Sigma, St. Louis, MO, USA). cDNA was prepared using a First‐strand cDNA Synthesis Kit (Fermentas, Burlington, Canada). The expression levels of target mRNAs were evaluated by qPCR. The primers are shown in Table 1. Target mRNA was amplified using SYBR Green I PCR Master Mix. Reactions were performed using the ABI 7300 Thermal Cycler (Applied Biosystems, Shanghai, China) as follows: denaturation for 30 s at 95°C, followed by 5 s at 95°C and 34 s at 60°C for 40 cycles.

Table 1.

Sequences of forward and reverse primers of selected genes designed for qPCR

Target gene (Rat)	Primer pairs
GAPDH	Forward: ACAGCAACAGGGTGGTGGAC Reverse: TTTGAGGGTGCAGCGAACTT
ALP	Forward: CCTGACTGACCCTTCCCTCT Reverse: CAATCCTGCCTCCTTCCACT
Runx2	Forward: CCTCTGACTTCTGCCTCTGG Reverse: GATGAAATGCCTGGGAACTG
OCN	Forward: GGCGTCCTGGAAGCCAATGTG Reverse: GACCAGGAGGACCAGGAAGTCCACGT
RhoA	Forward: AACAGGATTGGCGCTTTTGG Reverse: GATGAGGCACCCCGACTTTT
ROCK‐1	Forward: TTGGTTGGGACGTACAGTAAAA Reverse: CGTAAGGAAGGCACAAATGAGA
ROCK‐2	Forward:GACATTGAACAGCTTCGGTCGGA Reverse: CATTGAACAGCTTCGGTCG
Rac‐1	Forward: GAGAGTACATCCCCACCGTC Reverse: AACACGTCTGTTTGCGGGTA

2.8. Statistical analysis

All experiments were performed independently in triplicate. Statistical analyses were implemented in SPSS 19.0 (IBM, Armonk, NY, USA) using one‐way ANOVA. Multiple comparisons were performed using Student−Newman−Keuls test or Dunnett's T3 test. Differences were considered significant if the two‐tailed p‐value was less than 0.05.

3. RESULTS

3.1. Wavy microgrooved PDMS substrates

Micro‐patterned substrates were fabricated by replica moulding from a customized silicon wafer based on the SU‐8 polymer. This silicon wafer was generated using the following steps: cleaning, drying, coating, soft‐baking, exposure, post‐baking and developing. A pre‐designed photomask was used to define micro‐patterns on the silicon wafer, and SU‐8 was selected as an optimal photoresist for the fabrication of the high‐aspect‐ratio structure. A mixture of liquid oligomeric base and Sylgard184 (Corning) at a ratio of 1:10 was stirred to synthesize the PDMS substrate. Figure 1A illustrates the general steps in the fabrication process for micro‐patterned PDMS substrates.

Figure 1.

(A) Schematic illustration of micropatterned substrates fabrication. (B) and (C) contact angle analysis of poly(dopamine)‐coated PDMS. For the contact angle measurement, the substrates were coated with poly(dopamine) for 72 h. PMDS modified by dopamine showed better hydrophilicity. The contact angle decreased from 117.9.4 to 41.8 after modified. Data presented as means ± SD, *P < 0.05

The oxidative polymerization of dopamine results in the production of a poly‐(dopamine) ad‐layer on the PDMS surface.32 Static contact angles were obviously reduced from 124.4° to 34.2° after coating, indicating poly‐(dopamine) functionalization on PDMS surfaces (Figure 1B,C). Based on SEM images of the surface (Figures 2A) and vertical cross‐sections (Figures 2B) of the samples, the proposed method could be used to obtain clear, smooth, and highly uniform substrates with very few defects. The wavy microgrooves consisted of several semicircles of equal size. The inner and outer diameters were 20 and 40 μm for high curvature and 100 and 120 μm for low curvature. The groove width was 10 μm, and the depth was 5 μm. As indicated in our previous study, PDMS substrates had a roughness of 55.67 nm according to atomic force microscopy.33

Figure 2.

The surface and vertical cross‐sections of the sample via scanning electron micrographs (A) SEM images of PDMS substrates. Top row: flat PDMS control Middle row: high curvature wavy microgrooves Bottom row: low curvature wavy microgrooves. (B) The depth of wavy microgrooves is 5 μm. (The vertical cross‐sections of the micropatterned sample)

3.2. Cell nucleus distribution and morphology

To explore the influence of curvature on cellular behaviour, BMSCs were cultured in wavy microgrooves with high and low curvatures. Flat PDMS substrates served as controls. DAPI staining was used to investigate the distribution of BMSC nuclei and morphology at 3 days post‐seeding. As shown in Figure 3A, BMSC nuclei were predominantly distributed in microgrooves. BMSC nuclei on the flat control were randomly distributed, while those on micro‐patterned surfaces were located along the longitudinal axis of micro‐grooves, showing a consistent orientation along the groove. Micro‐patterned PDMS could dramatically regulate the distribution and morphology of BMSC nuclei. As shown in Figure 3B, compared with the flat control, the circularity of cell nuclei decreased significantly in wavy microgrooves. Greater BMSC nuclear deformation was observed for wavy microgrooves with high curvatures than for those with low curvatures (P < 0.05).

Figure 3.

BMSCs nucleus distribution and morphology on plate and micropatterned PDMS. (A) Fluorescent images of BMSCs nucleus (blue) on micropatterned and plate PDMS at day 3 post‐seeding. Scale bars are 100 μm. (B) Circularity of BMSCs nucleus, calculated from the fluorescent images in (A). Data presented as means ± SD, * means P < 0.05

3.3. Cell morphology and cytoskeletal organization on wavy microgrooved substrates

The distribution and morphology of BMSCs on different substrates were studied by immunofluorescence (Figure 4). BMSCs were randomly arranged on flat PDMS surfaces. By contrast, cells were oriented in the direction of wavy microgrooves. Significant differences in the F‐actin distribution were observed among substrates. On the flat control, F‐actin was assembled in a normal manner and spread more widely than the spread observed on patterned substrates. On wavy micropatterned substrates, the depolymerized and poorly formed actin stress fibres were observed. The shape and area of F‐actin spread were controlled by the micro‐structures. As the curvature of microgrooves increased, the degree of deformation based on F‐actin staining also increased (Figure 4).

Figure 4.

Immunostaining images of cell nuclei (blue) and F‐actin(green) on PDMS substrates at day 3 post‐seeding. Scale bars of control group are 10 μm. Scale bars of high and low curvature group are 7.5 μm

3.4. Higher curvature induced increased osteogenic gene expression in BMSCs

As mentioned, we confirmed the cytoskeletal organization of BMSCs on wavy microgrooved substrates. We further investigated the osteogenic characteristics of cultured BMSCs under bone induction conditions. RUNX2, also named CBF‐alpha‐1, is a crucial transcription factor related to osteogenic differentiation. Immunofluorescence staining showed differences in RUNX2 expression on three substrates (Figure 5A). In a semi‐quantitative analysis of the mean optical density of RUNX‐2 immunofluorescence, we detected a noticeable increase in expression in wavy microgroove groups compared to the control group, and we observed the maximum expression in the high curvature group (Figure 5B). We also evaluated the levels of osteogenic markers (OCN, OPN and ALP) by qRT‐PCR. In concordance with the RUNX2 expression results, the expression levels of OCN, OPN and ALP (Figure 5C) were significantly up‐regulated on wavy microgroove substrates compared with flat control substrates, and the highest expression levels were detected in the high curvature group.

Figure 5.

Higher curvature induce more osteogenic gene expression of BMSCs (A) Immunostaining images of RUNX2(red), cytoskeleton (green) and nucleus (blue) after 2 days plated on PDMS substrates. Scale bars of control group and high curvature groups are 25 μm. Scale bars of low curvature group are 15 μm. (B) Mean optical density of immunostaining images of RUNX2 in Figure 4A. (C) ALP, OCN, OPN gene expression in BMSCs measured on different substrates. GAPDH levels set as internal normalized control. The samples were collected on Day 2. The results shown are representative of three different experiments (n = 3). Data presented as means ± SD, *P < 0.05

3.5. Role of the RhoA/ROCK pathway in osteogenic differentiation

We hypothesized that the RhoA/ROCK signalling pathway is involved in osteogenic differentiation induced by cytoskeleton reorganization. To evaluate this hypothesis, the expression of RhoA was examined by immunofluorescence staining (Figure 6A). Based on the mean optical density of RhoA immunofluorescence (Figure 6B), RhoA exhibited greater expression on wavy microgroove surfaces than on flat surfaces. As the curvature increased, the expression of RhoA was dramatically up‐regulated. The mRNA levels of RhoA, ROCK‐1, Rac‐1, and Vinculin were remarkably higher in the high curvature group than in the low curvature group, further establishing the association between the RhoA/ROCK signalling pathway and cytoskeletal reorganization‐mediated osteogenic differentiation.

Figure 6.

Wavy microstructures could also influence the RhoA/ROCK pathway (A) Immunostaining images of RhoA (red), cytoskeleton (green) and nucleus (blue) after 2 days plated on PDMS substrates. Scale bars of control group are 15 μm. Scale bars of high and low curvature group are 10 μm. (B) Mean optical density of immunostaining images of RhoA in Figure 5A. (C) RhoA,ROCK1,RAC1,Vinculin gene expression in BMSCs measured on different substrates. GAPDH levels set as internal normalized control. The samples were collected on Day 2. The results shown are representative of three different experiments (n = 3). Data are presented as means ± SD, *P < 0.05

4. DISCUSSIONS

Microgrooves with a 10‐μm width, 5‐μm depth, and constant curvature were successfully fabricated by replica moulding from a customized silicon wafer. BMSCs were cultured in wavy microgrooves with different curvatures and on a flat control surface. We observed distinct patterns of nuclear distortion and cytoskeletal remodelling of BMSCs in wavy microgrooves with high curvature. Likewise, we detected higher expression of osteogenic‐related genes in BMSCs in the high curvature group than in other groups. Moreover, immunofluorescent staining of RhoA and quantitative RT‐PCR results for RhoA, ROCK1, Rac1, and Vinculin showed that the RhoA/ROCK signalling pathway is involved in curvature‐dependent osteogenic differentiation. Hence, our results indicated that curved microstructures could promote BMSC differentiation to the osteogenic lineage via the RhoA/ROCK pathway. Accordingly, these results highlight the potential application of curved microstructures in the field of bone tissue engineering.

The properties of natural bone are strongly dependent on the hierarchical structure and well‐organized ECM and cells.18 The curvature of the microenvironment is an important geometric feature in bone formation. Thus, from the viewpoint of bionics, two wavy microgrooves consisting of several semicircles of equal size were designed. The average diameter of the osteon is approximately 100–500 μm.14, 15 In the low curvature group, the inner and outer diameters of semicircles were 100 and 120 μm, respectively, to simulate the curvature of the natural osteon. Inner and outer diameters of 20 and 40 μm were used to represent a higher curvature than that of bone tissues (Figure 2A).

PDMS substrates with wavy microgrooves were successfully fabricated by replica moulding from a silicon wafer. The PDMS substrates are non‐wetting surfaces. Thus, we used the oxidative polymerization of dopamine to form a poly‐(dopamine) ad‐layer on the PDMS surface, which not only improves the hydrophilicity of PDMS (Figure 1B,C), but also retains the accuracy of the patterns.

Cells could recognize and be regulated by topographical stimuli of the ECM at both micrometer and nanometer scales.11, 12, 34 However, cell–ECM interactions can be activated only when the size of topographies are appropriate proper. Some suggest that deeper (10 μm and narrower grooves (7.5 μm) could align and elongate the cytoskeleton and nuclei most effectively.19 Past research has also shown that cells cannot sense topographical cues on microgroove substrates if the ridge is wider than 5 μm and the depth of the groove is not more than 1 μm.35 In our study, 5‐μm‐deep and 10‐μm‐wide wavy microgrooves (Figure 2A,B) permitted a BMSC response to a topographical stimulus. BMSC nuclei on the flat control surface were randomly distributed, while those on micro‐patterned surfaces located along the longitudinal axis of micro‐grooves, showing an obvious orientation. These results are consistent with previous results in support of the theory that cells tend to locate in the microgrooves that match their sizes.19

In addition, different curvature types are expected to have different effects on BMSC nuclei. As shown in Figure 3B, compared with the flat control, the circularity of cell nuclei decreased significantly in wavy microgrooves. Wavy microgrooves with high curvature had a more noticeable influence BMSC nuclear deformation than that of microgrooves with low curvature (P < 0.05) (Figure 3A,B). These results indicated significant cell nuclear distortion due to confinement in curved spaces.

Among a variety of micro‐ and nano‐patterns, the groove is the most basic and typical. Cell elongation and orientation along the direction of micro‐ or nano‐grooves via contact guidance has been confirmed for many cell types.4, 28 Moreover, contact guidance was also observed on wavy microgroove substrates. BMSCs exhibited a random arrangement on flat PDMS control substrates. By contrast, cells were oriented in the direction of wavy microgrooves. As the curvature of wavy microgrooves increased, the degree of deformation also increased (Figure 4). The cytoskeleton consists of microfilaments, microtubules and intermediate filaments. Microfilaments are mainly composed of the contractile protein actin and make up a large portion of the cytoskeleton.36 F‐actin is the key component of microfilaments in the cytoplasm of eukaryotic cells. Moreover, the cytoskeleton is involved in mechano‐sensing and mechano‐transduction37, 38 and is a major factor in differentiation processes.10 Topographical cues could be recognized by focal adhesion, which is a direct linkage of topographical cues and the cytoskeleton‐signalling network, thus regulating cell shape and differentiation.39, 40, 41 Therefore, to determine the influence of various wavy microgrooves on cytoskeleton remodelling, BMSC differentiation was evaluated in detail.

BMSCs have great potential for proliferation and multi‐directional differentiation ability. They can differentiate into bone cells, fat cells, cartilage cells, and others cells in appropriate circumstances in vivo or in vitro.42, 43 In addition to soluble factors, the degree of cytoskeletal tension plays an important role in directing MSC differentiation.15 Past research has shown that BMSCs on parallel microgroove substrates could have osteogenic differentiation ability owing to cytoskeletal changes.44 The effects of micro/nano‐patterns on osteogenic differentiation are related to microtubules, actin cytoskeleton, and Rho‐kinase. Some researchers have depolymerized microtubules with nocodazole and revealed that different micropatterns result in the same osteogenic properties. However, disrupting the actin cytoskeleton with cytochalasin D and Rho‐kinase with Y‐27632 lead to the loss of osteogenesis effects.10, 44 Accordingly, we hypothesized that curved microstructures, such as wavy microgrooves, could facilitate BMSC osteogenic differentiation by increasing cytoskeletal tension. Our results indicating that high curvature stimulates increased expression of osteogenic genes in BMSCs confirmed our hypothesis (Figure 5). These results have implications for biomaterial surface design. For example, in the field of oral implantation, wavy microstructures can be designed on the surface of the implant to accelerate the process of bone integration. Laser micromachined micro‐grooves have already been applied in dental implants to improve bone and connective tissue attachment, while inhibiting epithelial down‐growth. In addition, the surface of the barrier film used in guided bone regeneration (GBR) can also be designed to promote these properties.

RhoA, a small GTPase protein in the Ras superfamily, and its effectors are crucial in mechano‐transduction via the remodelling of the cytoskeleton.23 The mechanical actions of the ECM can be detected by mechano‐sensory systems and converted to RhoA signalling pathways to stimulate actin phosphorylation. Activated Rock, an effector of RhoA, increases actin polymerization and actin filament stabilization. The polymerization of actin at the cell periphery is controlled by Rac.45, 46 RhoA‐mediated stress fibre formation is related to the localization of talin and vinculin to focal adhesions,47 which arise from Rac1‐mediated focal contacts.46, 48 Additionally, ROCK can phosphorylate myosin II to increase cell tension and adds binding sites for vinculin.23, 49, 50

RhoA and its downstream effectors ROCK are also crucial in the regulation of the differentiation of human MSCs into osteogenic lineages.24, 25, 26, 27 Researchers have found a positive correlation between the activity of Rac1 and osteoblast differentiation on rough surfaces.27, 46 However, the mechanisms by which curvature induces change sin cytoskeleton tension and related cellular responses are not entirely clear.

We further tested the hypothesis that the RhoA/ROCK signalling pathway is involved in the curvature‐induced osteogenic differentiation process. RhoA displayed greater expression on wavy microgroove surfaces than on flat surfaces. As the curvature increased, the expression of RhoA was also dramatically up‐regulated. Furthermore, the mRNA levels of RhoA, ROCK‐1, Rac‐1, and Vinculin in the high curvature group were remarkably higher than those in the low curvature group, supporting the association between the RhoA/ROCK signalling pathway and cytoskeleton reorganization‐mediated osteogenic differentiation (Figure 6).

In conclusion, this study focused on a crucial geometric feature of the osteon, i.e, curved microstructures. We exposed rat primary BMSCs to wavy microgroove PDMS substrates with various curvatures to investigate alterations in cellular behaviour. Our results illustrated that microstructures with high curvature promote greater BMSC osteogenic differentiation compared with microstructures with lower curvature. Moreover, wavy microstructures could also influence the RhoA/ROCK pathway. Accordingly, curved microstructures have potential applications in tissue engineering as well as in the optimization of biomaterials, such as implants and biological barrier membranes in oral implantology, based on the observed osteogenic ability.

However, our study had some limitations. Previous studies have confirmed that modifications of the actin cytoskeleton occur in response to change in substrate topography.51, 52 Accordingly, additional in‐depth analyses of the actin polymerization dynamics in this curve‐dependent osteogenic process are needed. The direct measurement of actin polymerization dynamics in the mechanotransduction process should be performed in future studies.

Zhang Q, Lin S, Zhang T, et al. Curved microstructures promote osteogenesis of mesenchymal stem cells via the RhoA/ROCK pathway. Cell Prolif. 2017;50:e12356 10.1111/cpr.12356