Effects of electrospun membrane surface morphology on cellular behaviours and osteogenesis of bone marrow mesenchymal stem cells

Electrospun membranes are widely used in bone tissue engineering because of their similar bone extracellular matrix. The morphological characteristics of electrospun membranes, which include fibre diameter and alignment, play crucial roles in determining cellular behaviour and osteogenesis. Therefore, to investigate the effects of these two parameters on bone marrow mesenchymal stem cells (BMSCs), we prepared electrospun poly-L-lactic acid membranes using different diameters (nanoscale and microscale) and alignments (aligned and random) to investigate the effects of different surface morphologies on the proliferation, adhesion, migration, cell morphology, and osteogenesis of BMSCs. Our results showed that electrospun membranes with different surface morphologies have good biocompatibility and can regulate cell morphology, and the parallel aligned fibre orientation can promote cell migration. More importantly, BMSCs cultured on aligned nanofibres have a higher osteogenic potential than aligned microfibres and random fibres. Furthermore, our study shows that the surface morphology of electrospun membranes, which is one of the characteristics of biomaterials, can regulate the cellular behaviour of BMSCs, and that aligned nanofibre electrospun membranes can contribute to promoting osteogenesis, which can be used as the surface morphology of bone repair materials.


Introduction
Severe bone defects caused by trauma, tumours, and pathological diseases, remain challenging in orthopaedic treatment [1]. Additionally, 'autologous bone grafting' is the current gold standard of treatment, which is limited by insufficient donor bone and secondary trauma [2,3]; allograft bone transplantation also has the risks of immune rejection and infection [4,5]. Therefore, solving bone defects to actively search for bone transplantation alternatives is essential. Notably, the increase in bone tissue engineering based on biomaterial scaffolds has brought new hope and directions for bone defect treatment [6][7][8].
At the nanoscale, the bone extracellular matrix (ECM) comprises well-aligned collagen fibers [9], and the development of scaffolds that mimic the natural bone morphology is important to promote osteogenesis [5,10] Electrospinning has attracted considerable attention for its ability to produce micro/nanofibers, with high surface area-to-volume ratios and porosity that is similar to the ECM, and easy control of fibre diameter and alignment [11][12][13]. Some recent studies have found that physical stimulation without the delivery of cellular or biological factors can promote endogenous healing [5,9], those physical stimulation factors include mechanical stress, magnetic field, microstructure, and particle size. Based on these physical signals, they synergise to promote the repair and regeneration of damaged tissues by regulating biological behaviours, such as cell proliferation, adhesion, migration, and differentiation [14,15] The interaction between the biomaterial and host is the key to tissue repair and regeneration, and the biomaterial's surface morphology directly determines the mechanical binding and biological response to the host tissue. Therefore, surface morphology is one of the most critical physical signals for designing biomaterials. In contrast, the fibre diameter and alignment are the main parameters for an electrospun membrane [16,17]. Related studies have found that the electrospun fibre diameter plays a critical role in regulating cellular biological behaviour, and Li et al [18] demonstrated that nanofibres significantly upregulate-smooth muscle actin (-SMA), transforming growth factor and filament vascular expression compared with microns, thereby promoting fibroblast-to-myofibroblast differentiation, which is another important factor that regulates cellular biological behaviour. Gluais et al [19] co-cultured intervertebral disc fibroblasts with electrospun scaffolds and compared them with random fibres; cells travelled on aligned fibres in line with fibre orientation, were highly ordered, and formed similar AF extrastromal deposits. Because the surface morphology of aligned electrospun fibres is similar to that of bone collagen bundles, the former was employed for bone tissue repair [20]. Cells on aligned fibres were extended along the fibre direction and showed the ability to contribute to osteogenesis, which the expression of osteogenesis markers may have initiated after the cytoskeleton was stretched [21].
Bone repair and regeneration are well-known to be a dynamic and balanced process involving cells such as MSCs, inflammatory cells, osteoblasts, and osteoclasts [22]. BMSCs are commonly used as seed cells in bone tissue engineering because of their ability to self-renew and differentiate multi-directionally [23]. However, the adhesion, proliferation, migration, and osteogenesis of BMSCs are influenced by the material's surface morphology [24,25]. As an emerging bone repair material, the design of electrospun surface morphology that is suitable for the survival of BMSCs is a necessary consideration. However, the effects of the two most important parameters of the electrospun surface morphology, which are fibre diameter and alignment, on BMSCs have been poorly reported. Therefore, a systematic study of the effects of the electrospun fibre diameter and alignment on the biological behaviour and osteogenic differentiation of BMSCs is necessary.
Summarily, we aimed to investigate the effects of fibre diameter and alignment on the cellular behaviour and osteogenesis of BMSCs. Therefore, we prepared random and aligned fibres with nanoscale (600 nm) and microscale (1200 nm) diameters to investigate the effects of different electrospun membrane surface morphologies on the proliferation, adhesion, migration, morphology, and osteogenesis of BMSCs, and to provide optimal morphological parameters for developing electrospun materials.

Preparation and characterisation of electrospun membrane
Aligned and random fibres with different diameters were prepared using electrospinning (Yong Kang Le Ye Co., Ltd, Beijing, China). Briefly, PLLA was dissolved in HFIP solution and configured into a spinning solution of 16% w/v, which was stirred overnight in a magnetic stirrer. The spinning solution was placed in a syringe and adjusted by altering the receiving distance, voltage, injection speed and receiver speed. Table 1 shows the relevant parameters. The rotational speed of the aligned fibre receiver and random fibre was 2800 rpm and 100 rpm, respectively. All electrospun membranes were stored at room temperature for 3 days to remove the residual organic solvent. The aligned nanofibres and microfibres were named A-600 and A-1200, and the random nanofibres and microfibres were named R-600 and R-1200, respectively.
The electrospun membrane was cut into 1 cm × 1 cm sizes. After the samples' surfaces were sprayed with gold, the surface morphological features were scanned using a scanning electron microscope (SEM) (Hitachi Japan) at different magnifications in random areas. Next, 100 fibres were randomly obtained from each group, and the diameter and orientation angle of the fibres were calculated using ImageJ software. The hydrophilicity was determined using a water contact angle (WCA) meter (Oca20, Dataphysics Co., Ltd, Germany), the sample was fixed on a flat test bench, then 5 μl of deionized water was dropped at a height of 5 cm from the sample and immediately captured the image on the membrane surface, the WCA was measured using image analysis software (n = 3). Uniaxial tensile tests were performed using a universal testing machine (EZ-LX, Shimadzu, Japan) with a 50 N loading. All specimens were cut into rectangles (5 × 1 cm), the both ends of the sample was clamped by sandpaper and the length of the sample between the fixed ends was recorded, then the crosshead speed was fixed at 1 mm min −1 , and the stress-strain curves were recorded (n = 5). Finally, the ultimate tensile strength, failure strain, and elastic modulus were calculated based on the stress-strain curve.

BMSCs isolation and culture
Specific-pathogen-free male Sprague-Dawley rats aged 8 weeks were purchased from Kunming Medical University, euthanised, and sacrificed. BMSCs were isolated and cultured using the whole bone marrow adherence method in ɑ-MEM medium supplemented with 10% FBS and 1% penicillin-streptomycin. The culture environment was set at 37°C, 95% relative humidity, and 5% CO 2 . Next, the medium was changed every 2 days; cell abundance was 80%-90% for passaging and P3-P5 for cell experiments. The university animal ethics committee approved all animal studies [26].

BMSCs proliferation and viability
Before the BMSCs were inoculated, electrospun membranes were prepared as 15 mm diameter circles, immersed in 75% alcohol for 2 h, and irradiated using ultraviolet light for 1 h for sterilisation. The proliferation of BMSCs on the electrospun membranes was detected using CCK-8 (DOJINDO, Japan). Cells were cultivated on each group of electrospun membranes at a density of 2 × 10 4 /well, and the culture medium was discarded at 1, 4, and 7 days after inoculation and washed three times with PBS. In total, 200 μl of culture medium containing 10% Cell Counting Kit-8 (CCK-8) reagent was added after incubation at 37°C for 2 h under light-free conditions, and 100 μl was transferred to a 96-well plate, and the absorbance was measured at a wavelength of 450 nm (n = 6). After co-culture for 72 h, cell activity was detected using Calcein-AM/PI reagent (Solarbio, Beijing, China), stained for 30 min at 37°C under light-free conditions, and the membranes were photographed using fluorescence microscopy(OLYMPUS, Japan) at excitation wavelengths of 490 nm and 560 nm (n = 3).

BMSCs migration
Using the method of the previous study [27], a custom-made stainless steel ring (f = 15 mm) was placed in a 24well culture plate with a 1 mm barrier in the middle to artificially create a cell-free wound-like gap area (defined as an 'artificial wound'). The cell-free wound-like gap area on the aligned fibre membrane was categorised into parallel and vertical fibre orientations. Overall, 500 μl of the medium containing 5 × 10 5 cells was inoculated onto the electrospun membrane, and the rings were removed after 24 h. After 48 h of further incubation, the cells were fixed, stained with rhodamine-phalloidin, and cell migration was observed using fluorescence microscopy (n = 3).

BMSCs morphology
BMSCs at densities of 2 × 10 4 cells and 1 × 10 4 cells for cell alignment and cell shape analyses, respectively, were inoculated onto electrospun membranes in 24-well culture plates. After culturing the cell membrane for 3 days, the cellularised constructs were washed and fixed using 4% paraformaldehyde for 30 min and permeabilised with 500 μl Triton-X100 for 10 min Subsequently, the cells were stained with 200 μl of rhodamine-phalloidin in the dark at room temperature for 30 min, and the nuclei were stained with DAPI for 10 min After washing, the BMSCs were observed using fluorescence microscopy for the alignment and morphology of a single BMSC. Based on the acquired microscopic images, the morphological characteristics of the cells were analysed using ImageJ software as in a previous study [24]. The morphological features of a single cell include the degree of deviation of the nucleus from the fibre direction, the average area of the cell spreading, the elliptical form factor (EFF) defined as the long axis divided by the short axis, and the shape factor defined as 4πA/P 2 , where A is the area of the cell, and P is the perimeter of the cell.

ALP and ARS staining
BMSCs were inoculated at a density of 5 × 10 4 /well in 24-well plates and cultured in an osteogenic medium containing 100 nm dexamethasone, 10 mm β-glycerophosphate disodium, and 50 μg mL −1 ascorbic acid. After 7 and 14 days of incubation, the cells were washed with PBS and fixed with 4% paraformaldehyde for 30 min ALP staining solution was prepared according to the manufacturer's protocol; first, the samples were stained in the dark for 30 min, and the reaction was terminated with distilled water and visualised under a stereomicroscope (OLYMPUS, Japan). Cells cultured under the same conditions were lysed with Triton-X100, the lysate was obtained after 30 min, and the supernatant was collected using high-speed frozen centrifugation. Subsequently, ALP activity was measured according to the manufacturer's instructions, and absorbance was measured at 520 nm and normalised to total protein content, which was determined using a BCA analysis kit. Finally, ALP activity was estimated according to the method described in the kit. BMSCs were cultured for 14 and 21 days under the same culture conditions and fixed with 4% paraformaldehyde for 30 min, after which the samples were washed. ARS solution was added for 2 h at 37°C after washing three times with PBS, and images were collected under a stereomicroscope (OLYMPUS, Japan).

Osteogenesis gene expression
Briefly, BMSCs were inoculated in 6-well plates at a density of 5 × 10 5 /well. Next, total RNA was extracted using the TRIzol reagent (Invitrogen), and reverse transcription was performed using the RevertAid™ First Strand cDNA Synthesis Kit (Thermo Fisher US). Real-time quantitative polymerase chain reaction (qPCR) was performed using an SYBR Premix Ex Taq kit (Clontech) with cDNA as a template. Table 2 presents the primer sequences used in this study. The reaction comprised an initial denaturation at 95°C for 5 min, 40 cycles of 95°C for 10 s, and 60°C for 31 s (ABI7300 Real-Time Quantitative PCR System). Finally, the value of target gene expression was normalised to GAPDH, and the relative expression level was calculated using the 2 −ΔΔCt method.

Statistical analysis
Data analyses were performed using the Origin 2021 software (OriginLab, Massachusetts, USA), and all experimental data are presented as mean ± standard deviation. One-way analysis of variance followed by Tukey's post-hoc test was used for statistical comparisons between groups. Statistical significance was considered at p < 0.05.

Results
Surface morphology and physical properties of the electrospun membrane The surface morphology of the electrospun membrane was observed using SEM. The SEM images showed that the surfaces of the four groups of fibres were smooth and uniform, and the fibre orientations of R-600 and R-1200 were random (figures 1(a), (c), (e), and (g)), whereas that of A-600 and A-1200 were aligned (figures 1(b), (d), (f), and (h)). The diameters of R-600, A-600, R-1200, and A-1200 were 601.49 ± 33.65 nm, 599.62 ± 33.88 nm, 1205.84 ± 73.32 nm, and 1203.71 ± 101.24 nm, respectively. No significant difference was found in fibre diameter between R-600 and A-600, R-1200, and A-1200 (p > 0.05) ( figure 1(q)). However, the diameter distributions in the four groups were approximately normally distributed (figures 1(i)-(l)). The fibre angle distributions of A-600 and A-1200 were centralised (figures 1(n) and (p)), whereas those of R-600 and R-1200 were irregularly distributed (figures 1(m) and (o)). The hydrophilicity of the electrospun membrane was analysed using a WCA meter. Specifically, the smallest was 111.07 ± 3.69°in the A-600 group, which was statistically different from the other three groups (p < 0.05); 151.77 ± 8.17°in the R-600; 146.33 ± 5.06°in the R-1200; and 143.77 ± 5.99°in the A-1200 groups, respectively, without statistical difference between the three groups (p > 0.05) ( figure 1(r)). The mechanical properties of the electrospun membranes were determined using uniaxial tensile tests. Figure 2(a) presents a typical tensile stress-strain curve of the electrospun membranes. These curves follow Hooke's law, for the initial stages of stretching and plastic deformation occur after the yield point is reached. The ultimate tensile strength, strain failure, and Young's modulus of the membranes showed similar trends, which implies that the aligned fibres were stronger than the random fibres. The ultimate tensile strength of A-600 and A-1200 was 18.97 ± 2.50 MPa and 18.86 ± 1.06 MPa, respectively, greater than those of R-600 and R-1200 at 2.15 ± 0.34 MPa and 1.83 ± 0.23 MPa, respectively, with a statistical difference (p < 0.01) ( figure 2(b)). The strains at breakage were 231.14 ± 18.35 MPa and 229.96 ± 6.69 MPa for A-600 and A-1200, respectively, which were greater than those of 121.11 ± 12.29 MPa and 130.12 ± 13.55 MPa for R-600 and R-1200, respectively (p < 0.01) (figure 2(c)). The Young's modulus for A-600 and A-1200 was 225.00 ± 20.01 MPa and 231.35 ± 14.22 MPa, respectively, greater than that of 27.32 ± 6.32 MPa and 21.80 ± 6.36 MPa for R-600 and R-1200, respectively (p < 0.01) (figure 2(d)).

Effect of electrospun membrane on proliferation and biocompatibility of BMSCs
Excellent biocompatibility is a desirable property of biomaterials. The effect of the electrospun membranes on the proliferation of BMSCs was measured using CCK-8. On the first day, no significant difference was found in the impact of the four electrospun membranes on the proliferation of BMSCs (p > 0.05). On the fourth day, the A-600 group had the best proliferation activity, which was significant compared with the other three groups (p < 0.05), and the absorbance value of the R-600 group was higher than that of the R-1200 group (p < 0.05). Furthermore, on the 7th day, the absorbance value of the A-600 group was higher than that of the R-1200 and A-1200 groups (p < 0.05), and the R-600 group was higher than that of the R-1200 group (p < 0.05) ( figure 3(b)).
After co-culturing the cells and electrospun membranes for 72 h, the BMSCs' activity on the electrospun membranes was examined using live/dead staining. Most of the cells survived well, and only a few dead cells were observed on the electrospun membranes of the different groups ( figure 3(a)). The live/dead staining results were consistent with the CCK-8 assay results, with all four groups of cells showing good activity, confirming the excellent biocompatibility of the electrospun membranes.

Effects of fibre diameter and orientation of electrospun membrane on the migration of BMSCs
The effect of electrospun membrane fibre diameter and alignment on the migration of BMSCs was assessed by observing the migration of BMSCs onto defined 'artificial wound' areas and analysing the cell-free areas on different electrospun membranes ( figure 4(a)). In the group with an average diameter of 600 nm, the largest number of BMSCs migrated parallel to the aligned fibre, with the smallest area not covering the 'artificial wound' area, followed by the number of cells which migrated with random fibre. The migration perpendicular to the aligned fibres was the smallest, and the uncovered 'artificial wound' area was the largest. A significant difference was found among the three groups (p < 0.05), and the same results were observed in the group with an average diameter of 1200 nm. However, no significant difference was found in the migration of BMSCs on spinning membranes with the same fibre alignment but different fibre diameters (p > 0.05) ( figure 4(b)).

Effects of fibre diameter and orientation of electrospun membrane on adhesion of BMSCs
The effects of the fibre diameter and alignment of the electrospun membranes on cell adhesion were assessed using CMFDA-labelled BMSCs. Over time, the number of BMSCs adhering to the electrospun membranes increased and exhibited a cell morphology associated with fibre alignment, with cells orderly and disorderly arranged on aligned and random fibres, respectively ( figure 5(a)). No significant difference was found in the number of adherent cells with different morphologies simultaneously (p > 0.05) ( figure 4(b)).

Effects of fibre diameter and orientation of electrospun membrane on the morphology of BMSCs
The morphology of BMSCs after co-culturing with the electrospun membranes was observed using fluorescence microscopy ( figure 6(a)). In contrast to the disordered cells' arrangement on the random fibre, BMSCs were orderly arranged on the aligned fibrous membrane, aligning with the fibre orientation. Quantification of cell orientation (θ) indicated that the cells tended to centralise on aligned fibres but tended to disperse on random fibres ( figure 6(b)). Single-cell morphological analysis revealed that single cells displayed a unidirectionally oriented spindle shape on aligned fibres compared with cells on random fibres ( figure 6(c)). Cell spreading, shape, and elongation were quantified by measuring the cell area, shape factor, and EFF (figures 6(d)-(f)), which spread over larger areas on random fibres (p < 0.05) and had a higher shape factor index compared with the A-600 group (p < 0.05). However, the aligned fibres had higher cell elongation (p < 0.05). No significant differences were found in single-cell morphology among the different diameters (p > 0.05).
Effects of fibre diameter and orientation of electrospun membrane on osteogenesis of BMSCs As an early marker of osteogenesis, ALP staining and semi-quantitative analysis were performed on days 7 and 14, respectively, to detect the osteogenesis of BMSCs on the different electrospun membranes. On day 7, the intensity of ALP staining was significantly higher in the R-600 and A-600 groups than in the other two groups, whereas on day14, it was higher in the A-600 group than in the other three groups and higher in the R-600 than in the R-1200 and A-1200 groups. However, the intensity of ALP staining between the R-1200 and A-1200 groups was not significantly different at the two-time points (figure 7(a)), and the semi-quantitative assay showed similar results (p < 0.05) (figure 7(c)).
Calcium salt deposition, which is a marker of late osteogenesis, was analysed using ARS and semiquantitative analysis on days 14 and 21. The ARS staining results on days 14 and 21 were consistent; specifically, the A-600 group had the highest ARS staining intensity, which was higher than that in the other three groups, while the R-600 group had a higher ARS staining intensity than the R-1200 and A-1200 groups. No significant difference was found between the R-1200 and A-1200 groups ( figure 7(b)). The results of the semi-quantitative experiment were consistent with the staining intensity (p < 0.05) ( figure 7(d)).

Effects of fibre diameter and orientation of electrospun membrane on osteogenic gene expression in BMSCs
To further investigate the osteogenic gene expression of BMSCs on different morphologies of the electrospun membrane, the effects of fibre diameter and alignment regulation on osteogenic gene expression were examined using qRT-PCR (figure 8). After 7 days of co-culture of cells and electrospun membranes, ALP, COL-I, Runx2, and OPN gene expressions were higher in the A-600 group than in the other three groups (p < 0.05). ALP and Runx2 gene expressions were higher in the R-600 group than in the A-1200 group (p < 0.05), whereas COL-I and Runx2 gene expressions were higher in the R-600 group than in the R-1200 group (p < 0.05). Similarly, on day 14, ALP, COL-I, Runx2, and OPN gene expression were higher in the A-600 group than in the other three groups (p < 0.05). COL-I, Runx2, and OPN expressions were higher in the R-600 group than in the R-1200 and A-1200 groups (p < 0.05).

Discussion
Electrospinning enables the easy preparation of microscale/nanoscale fibers [28]. The fibres' diameter and orientation are controlled by adjusting these parameters [29], which can be categorised into solution, process, and environmental parameters. The solution parameters included the concentration, viscosity, relative molecular mass, conductivity, surface tension, and solvent type; the process parameters included the applied voltage, flow rate, and distance from the injection device to the collection device; and the environmental parameters were temperature and humidity. Notably, appropriate adjustment of these parameters permits the preparation of micro/nanofibers with suitable diameters, homogeneous and bead-free [30,31]. In this study, electrospun membranes with different fibre diameters and orientations were prepared using a 16% PLLA electrospinning solution and by adjusting the applied voltage, flow rate, receiving distance, and type of receiving device.
The mechanical and hydrophilic properties of biomaterials are important physicochemical properties that directly affect their biological properties. The fibre orientation and diameter significantly influenced the mechanical properties and hydrophilicity of the electrospun membranes. Jin et al [32] prepared three groups of electrospun membranes with aligned, random, and lattice topographies and found that the mechanical properties of the aligned fibres were superior to those of the other two groups. He et al [33] also found that the mechanical properties of PLLA electrospun membranes with aligned fibres were superior to those with random fibres. Similarly, our findings are consistent with those of previous studies, where aligned fibres showed a higher tensile strength. This may be explained by the orderly arrangement of the aligned fibres, which limited the tensile movement. Previous studies found that the hydrophilicity of aligned fibre membranes is better than that of random fibre membranes [16,34], and it has also been observed that the fibre diameter is negatively correlated with hydrophilicity, with larger fibre diameters being less hydrophilic [35]. In our study, because PLLA is a hydrophobic material, the WCA of the four groups of electrospun membranes exceeded 90°; however, the WCA of the A-600 group was smaller than that of the other three groups, possibly because of the small diameter and ordered arrangement of the fibres.
The surface morphology of electrospun membranes acts as a mediator of contact reactions with cells by regulating cell adhesion, migration, and morphology [29]. Adhesion is critical for the survival, growth, and proliferation of cells anchored to a material. Different morphologies affect cell adhesion differently, mainly associated with surface roughness, height, and lateral spacing of nanoscale features [36]. Park et al [37] found that MSCs adhere to oriented titanium dioxide nanotubes (with diameters ranging from 15 to 100 nm) heavily depending on the tube diameter. The cells showed good adhesion at sizes of 15-30 nm; however, the adhesion and spreading of the cells were significantly reduced on nanotubes >50 nm in diameter. In this study, the four morphologies had no significant effect on the adhesion of BMSCs, probably because of the insufficient significant differences in the roughness and lateral spacing between fibres, and cell migration was guided by material morphology, Yi et al [38] found that huvECs migrated more favourably along parallel aligned fibres than in the vertical direction. In contrast, Qu et al [39] investigated the effect of silk fibroin fibres with different arrangements on the migration of MSCs compared with random fibres of the same diameter. MSCs exhibited higher migration efficiency on the aligned fibres. Our study also showed that BMSCs had the highest migration rate parallel to the aligned fibres, followed by the disordered fibres, while they had the lowest migration rate in the vertically aligned fibres. The cell morphology is regulated by the surface morphology of the contact material. Yuan et al [40] found that the aligned PLLA fibres could modulate the vSMC to enter the elongated morphology and ordered arrangement and that the hyaluronic acid coating on the fibre surface resulted in a more elongated and well-arranged. Wang et al [41] found that the cells exhibited spherical morphology and elongated structure on membranes with smaller and larger pore areas, respectively. In contrast, on aligned fibres, the cells remained elongated along the orientation of the fibres at different gaps. Zheng et al [42] prepared Anisotropic micro/nano composite topography scaffolds of functionalized multi-walled carbon nanotubes, to regulate the oriented growth of Schwann cells and promote axon extension neurons. In our study, BMSCs were irregularly arranged on aligned fibres with a larger spread of a single cell, whereas they were well arranged on ordered fibres with more elongated single cells, possibly because of the cytoskeleton, which is regulated by different fibre morphology Natural bone is highly anisotropic because of the ordered arrangement of collagen fibres, and there is evidence that MSCs differentiate more efficiently into an osteogenic phenotype when restricted to this arrangement [9].Gao [21] showed that aligned fibres induced stronger osteogenic differentiation of hMSCs than random fibres. In a similar study, Xie et al [16] found that the osteogenesis of BMSCs on nanofibres was stronger than that on microfibres and that the osteogenic induction ability of aligned nanofibres was stronger than that of random nanofibres. Similarly, in this study, ALP and ARS were used as early and late indicators of BMSC osteogenesis, respectively. ALP and ARS production was greater in the A-600 group than in the R-600 group, suggesting that the aligned fibres positively regulate the osteogenic differentiation of BMSCs. Regarding fibre diameter, the intensity of ALP and ARS in the A-600 and R-600 groups were higher than that in the A-1200 and R-1200 groups, respectively, suggesting that the nanoscale diameter is another crucial factor in the osteogenic differentiation of BMSCs.
The expression of osteogenesis genes is sensitive and can be used to explore materials that induce the osteogenic differentiation of BMSCs. Among the four genes examined in this study, ALP promoted cell maturation and calcification in the early stages of osteogenesis. COL-I is the major collagen in the osteogenic stage and is the most important collagen component in the bone matrix. OPN is a bone matrix salivary protein associated with the mineralisation and absorption of the bone matrix [43], whereas Runx2 is essential in regulating bone formation and remodeling [44]. Our experimental results showed that, compared with the R-600 group, the A-600 group upregulated the expression of osteogenesis genes; it was similarly upregulated in the A-600 group compared with the A-1200 group. These results indicate that the fibres' diameter and orientation can regulate the expression of osteogenic genes in BMSCs, and the aligned nanofibres can upregulated the expression of osteogenic genes. Finally, the qRT-PCR results for osteogenic gene expression were consistent with those for ALP and ARS.

Conclusion
We successfully prepared PLLA electrospun membranes using different morphologies of electrostatic spinning to study the effects of fibre diameter and alignment on the cellular behaviour and osteogenesis of BMSCs. By coculturing BMSCs with electrospun membranes of different morphologies, we found that aligned nanofibres were more advantageous than random fibres and aligned microfibres in promoting the proliferation, migration, and osteogenic differentiation of BMSCs. Therefore, our findings provide a reference and motivation for designing the surface morphology of bone repair biomaterials.