Grooved hydroxyapatite scaffold modulates mitochondria homeostasis and thus promotes osteogenesis in bone mesenchymal stromal cells.

Hydroxyapatite scaffolds (HASs) are widely studied as suitable materials for bone replacement scaffolds due to their chemical similarities to organic materials. In our previous study, a novel HAS with a 25‑30‑µm groove structure (HAS‑G) exhibited enhanced osteogenesis of bone mesenchymal stromal cells (BMSCs) compared with HAS, potentially by modulating the macrophage‑induced immune microenvironment. However, the exact effects of different surface patterns on the physiological processes of attached cells is not known. The present study aimed to determine the effects of HAS‑G on the osteogenesis and physiological processes in BMSCs. Cell counting kit‑8 assays and propidium iodide staining followed by flow cytometry were performed, and the results demonstrated that both in normal medium and differentiating medium, HAS‑G promoted cell proliferation by decreasing the proportion of G1/G0 cells and decreased reactive oxygen species (ROS) accumulation in BMSCs compared with HAS. Detection markers of osteogenesis revealed that compared with HAS, HAS‑G increased runt‑related transcription factor 2, osteocalcin and osteopontin protein levels and promoted osteogenesis, which was further confirmed by Alizarin Red S staining. Following JC‑1 staining, it was observed that HAS‑G maintained the mitochondrial membrane potential, similar to that achieved by N‑acetylcysteine pretreatment. In addition, compared with those of HAS, HAS‑G decreased mitochondrial ROS levels, which potentially contributed to the promotion of osteogenesis. The results also demonstrated that HAS‑G inhibited mitophagy induced by ROS accumulation and ATP synthesis compared with HAS. In conclusion, HAS‑G decreased ROS accumulation and mitophagy and thus promoted osteogenesis of BMSCs, indicating that ROS modulation of HAS‑G may serve a key role in osteogenesis.

Introduction

Hydroxyapatite [HA, Ca10(PO4)6(OH)2], an alkaline calcium phosphate, is reported to be the most biocompatible replacement biomaterial among the developed artificial bones and is a major constituent of 70% of human bone; thus, it is considered one of the most promising biomaterials for tissue replacement (1–3). Since HA possesses mechanical properties that are relevant to the anatomical site of implantation (4), HA scaffolds (HASs) are believed to be a potential bone replacement material. Several modifications of HASs have been reported to enhance the performance of HASs in osteoinduction. For example, Zhou et al (5) reported that preparing type I collagen in an eco-friendly PBS/ethanol solvent and covalently coating it onto the surface of HAS enhanced osteogenesis. Clinically, D'Agostino et al (6) produced HASs mixed with microfibrillar collagen for use in cheekbone augmentation in hundreds of patients, and satisfactory results have been obtained.

The grooved structure on the surface of alumina ceramic scaffolds induces mesenchymal stem cell osteogenesis by regulating the polarity of cells and modulating their morphology (7). This suggested that the surface of HASs may be modified by producing a grooved structure. In our previous study, HASs with a 25–30-µm groove structure (HAS-G) countered the limitations of regular HAS showing clinical inhibition of osteogenic effects (8). HAS-G significantly promoted human placenta-derived mesenchymal stem cell (hPMSC) osteogenic differentiation and the maturation of osteoblasts in vitro and in vivo compared with HAS (8). In addition, HAS-G regulated RAW cells and subsequently affected the expression and release of cytokines/chemokines (9). However, the mechanism by which the groove structure promotes osteogenesis of the attached cells is still unknown.

Bone marrow-derived stromal cells (BMSCs), which are osteogenic precursors, play a key role in bone generation (10). Due to their osteogenic capacity, BMSCs are widely considered to be candidates for bone repair in bone tissue engineering (10). The present study aimed to utilize BMSCs to study the potential osteogenic promoting effects of HAS-G and the underlying mechanism. By determining how HAS-G exerts a protective effect on BMSCs, the present study also aimed to identify the importance of mitochondria homeostasis in improving the surface modification of HASs.

Materials and methods

Preparation of scaffolds

Groove-modified porous hydroxyapatite scaffolds were produced as previously described (8,9). Briefly, by adjusting the quantity of water on the chitin sugar ball template surface, 25–30-µm groove structures were formed on the surface of HAS, which was termed HAS-G. The scaffolds were fixed using 4% paraformaldehyde for 30 min at room temperature. The surface morphology of the produced scaffolds was identified by scanning electron microscopy (SEM) using the Quanta 200 SEM instrument (magnification, ×5,000, 20,000, 50,000 or 100,000).

Cell culture

All animal experiments were approved by the Ethics Committee of the Sichuan Provincial People's Hospital (approval no. 2018-280). Following euthanasia by 30 mg/kg sodium pentobarbital and cervical dislocation, BMSCs were isolated from the femurs of one female Sprague-Dawley rat (age, 3–4 weeks; weight, 200–220 g; Dossy Experimental Animals Co., Ltd.). The rat was housed at room temperature with 12-h light/dark cycles, 50–65% humidity, and access to standard chow and water. The bone marrow was extracted by syringe needles, as previously described (11), and cultured for 24 h at 37°C in DMEM (Thermo Fisher Scientific, Inc.) supplemented with 10% FBS (Thermo Fisher Scientific, Inc.), which was considered as normal medium (NM). To promote osteogenesis, the medium was replaced with SD Rat Mesenchymal Stem Cell Osteogenic Differentiation Basal Medium (DM; cat. no. RASMX-90021; Cyagen Biosciences, Inc.) and cells were cultured at 37°C for 14 days with 5% CO2. For the observation of autophagosomes, cells were incubated with DM supplemented with 10 µM bafilomycin-A1 (Sigma-Aldrich; Merck KGaA) for 24 h at 37°C.

Reactive oxygen species (ROS) detection

BMSCs attached onto HAS or HAS-G were detached by 0.5 mM EDTA for 30 min at 37°C, transferred to a 6-well plate and allowed to attach for 4 h at 37°C. Subsequently, 10 µM 2′,7′-dichlorodihydrofluorescein diacetate (DCFH-DA; Beyotime Institute of Biotechnology) was added and incubated for 45 min at 37°C in the dark. Following two washes with ice-cold PBS, fluorescence was detected by fluorescence microscopy (magnification, ×40) and was semi-quantitatively measured using ImageJ software (version-2.0; National Institutes of Health). To quantitatively detect ROS levels, cells were detached from HAS or HAS-G by incubating in 0.5 mM EDTA for 30 min at 37°C. The cells were collected and stained with 10 µM DCFH-DA for 45 min at 37°C in the dark, and the fluorescent signal at a wavelength of 525 nm was detected using a Lumina Fluorescence Spectrometer (Thermo Fisher Scientific, Inc.). To scavenge ROS, 10 µM N-acetylcysteine (NAC; cat. no. A9165; Sigma-Aldrich; Merck KGaA) was added into medium for 24 h and subsequently ROS detection was performed.

To detect mitochondal-related ROS accumulation, MitoSOX™ red (Thermo Fisher Scientific, Inc.) staining was performed. Cells were plated in serum-free supplemented with 1 µM MitoSOX for 20 min at 37°C. Cells were imaged using a fluorescence microscope (Olympus Corporation; magnification, ×400).

Cell cycle analysis

To analyse the cell cycle phases by quantification of DNA content, cells attached to HAS were harvested, washed with ice-cold PBS and fixed overnight at 4°C with ice-cold 70% ethanol. The fixed cells were washed with PBS three times and incubated with a final concentration of 100 µg/ml RNase A and 40 µg/ml propidium iodide (PI; Beyotime Institute of Biotechnology) for 15 min in the dark at room temperature. Cells were analysed using a three-laser Navios flow cytometer (Beckman Coulter, Inc.) and FlowJo software (version 9.7.4; FlowJo Treestar).

CCK-8 assay

Cell viability was determined by the CCK-8 assay. Briefly, BMSCs (1×104) were cultured in 96-well plates at 37°C for 24 h. Subsequently, 10 µl CCK-8 solution (Beyotime Institute of Biotechnology) was added to each well and incubated at 37°C for 2 h. The absorbance was measured at a wavelength of 620 nm using a Multiskan spectrum microplate reader (Thermo Fisher Scientific, Inc.) and cell viability is presented as optical density-620.

Immunofluorescence staining

Cells attached to HAS were suspended by culturing in a 0.5 mM EDTA/PBS solution for 15 min at 37°C and reseeded in 6-well plates (5×104 cells/well) for 24 h attachment at 37°C. The cells were fixed with 4% paraformaldehyde in 1X PBS containing 0.1% Triton X-100 for 30 min at room temperature, and non-specific binding was blocked with 10% normal goat serum (Beyotime Institute of Biotechnology) in 1X PBS for 1 h at room temperature. The cells were incubated at room temperature for 1 h with the following primary antibodies targeted against osteogenesis-associated proteins: Runt-related transcription factor 2 (RUNX2; cat. no. ab23981; 1:1,000; Abcam) and osteocalcin (OCN; cat. no. ab93876; 1:1,000; Abcam), osteopontin (OPN; cat. no. ab192143; 1:1,000; Abcam). Subsequently, the cells were incubated with an AlexaFluor 488-conjugated secondary antibody (cat. no. ab6702; 1:5,000; Abcam) for 1 h at room temperature. For TOM20 staining, an Alexa Fluor647-conjugated anti-TOM20 antibody was used (cat. no. ab209606; 1:200; Abcam) at room temperature for 2 h. The cells were mounted in Vectashield® and stained with Hoechst 33342 (Vector Laboratories, Inc.; Maravai LifeSciences) at room temperature for 10 min. Cells were imaged using an X71 (U-RFL-T) fluorescence microscope (Olympus Corporation; magnification, ×100).

GFP-LC3B expression

Lentiviral constructs expressing GFP-LC3B were synthesized by Guangzhou RiboBio Co., Ltd. At 60% confluence, BMSCs were incubated with lentivirus-containing medium (1×106 titre/ml) overnight at 37°C with 5% CO2. Subsequently, the media was replaced with fresh DM and incubated for 48 h at 37°C with 5% CO2 prior to observation of formed autophagosomes. To detect GFP-LC3B, cells were observed using an X71 (U-RFL-T) fluorescence microscope (Olympus Corporation; magnification, ×40).

Western blotting

Western blotting was performed as previously described (12,13). Total protein was isolated from cells attached to HAS using a SoniConvert® sonicator (DocSense, Inc.) and quantified using the bicinchoninic acid assay (cat. no. BCA1-1KT; Sigma-Aldrich; Merck KGaA). Briefly, HAS or HAS-G was suspended in 500 µl SDS loading buffer (Beyotime Institute of Biotechnology) and lysed using the SoniConvert® cell lysis system, according to the manufacturer's instructions. Following brief centrifugation at room temperature at 1,000 × g for 5 min, the lysis sample was denatured at 100°C for 10 min. Total protein (20 µg) was separated via 10% SDS-PAGE and transferred to a PVDF membrane, which was blocked in PBS containing 5% BSA (Beyotime Institute of Biotechnology) at room temperature for 30 min. The membrane was incubated at room temperature for 1 h with the following primary antibodies: Anti-RUNX2 (cat. no. ab23981; 1:1,000), anti-OCN (cat. no. ab13421; 1:2,000), anti-OPN (cat. no. ab8448; 1:1,000), anti-microtubule-associated proteins 1A/1B light chain 3B (LC3B; cat. no. ab51520; 1:1,000), anti-nucleoporin p62 (p62; cat. no. ab109012; 1:1,000), anti-translocase of the inner membrane 23 (TIM23; cat. no. b230253; 1:1,000), anti-translocase of the outer membrane 20 (TOM20; cat. no. ab56783; 1:1,000), anti-heat shock protein 60 (HSP60; cat. no. ab46798; 1:1,000), anti-optineurin (OPTN; cat. no. ab23666; 1:1,000) and anti-β-actin (cat. no. ab8227; 1:5,000). All primary antibodies were purchased from Abcam. Following primary incubation, the membrane was incubated with a horseradish peroxide-labelled goat anti-rabbit IgG H&L antibody (cat. no. ab7090; 1:10,000; Abcam) at room temperature for 1 h. Protein bands were visualized using an ECL detection system (Thermo Fisher Scientific, Inc.). Bands were quantified by densitometry using Image J software (version 2.0; National Institutes of Health) with β-actin as the loading control.

Total ATP synthesis

Cells attached to HAS were resuspended in buffer containing 0.22 M sucrose, 0.12 M mannitol, 40 mM Tricine, pH 7.5, and 1 mM EDTA, as described previously (9) to a final concentration of 1×105 cells/ml. Cells were analysed using an Optocomp I BG-1 luminometer (GEM Biomedical, Inc.) and the ATP Bioluminiscent Assay kit (Sigma-Aldrich; Merck KGaA), according to the manufacturer's instruction.

Mitochondrial membrane potential assay

The mitochondrial membrane potential was measured by JC-1 staining (Thermo Fisher Scientific, Inc.), according to the manufacturer's instructions. Briefly, 1×105 cells attached to HAS-G were washed with ice-cold PBS twice and incubated with the JC-1 working solution at 37°C in the dark for 30 min. Subsequently, the supernatant was removed, the cells were washed twice with PBS and imaged using a fluorescence microscope (Olympus Corporation; magnification, ×100).

Alizarin Red S (ARS) staining and quantification

For Alizarin Red S staining, cells attached to HAS-G (1×105) were washed twice with PBS and fixed with 4% formaldehyde in PBS for 30 min at room temperature. Following a brief wash with PBS, the cells were stained for 30 min at room temperature with the ARS staining solution (Beyotime Institute of Biotechnology) supplemented with 40 mM ARS (pH 4.2). The cells were rinsed five times with PBS to reduce non-specific staining. Stained cells were observed using an X71 (U-RFL-T) fluorescence microscope (Olympus Corporation; magnification, ×40). Using Metamorph imaging software (version 6.1; Universal Imaging), osteogenic differentiation was quantified by measuring the area stained with Alizarin Red S. Measurements were performed in duplicate for each experiment, and the experiments were repeated three times.

Statistical analysis

Statistical analysis was performed using SPSS for Windows (version 19; IBM Corp.). Significant differences were determined using either the Student's t-test for two independent groups or one-way ANOVA for comparing results across multiple groups followed by a Tukey's post hoc test. Data are presented as the mean ± standard deviation. P<0.05 was considered to indicate a statistically significant difference. All experiments were repeated at least three times.

Results

HAS-G surface pattern promotes proliferation in BMSCs

HAS-G and HAS were prepared as previously described (8), and the surface patterns were confirmed by SEM (Fig. 1). Multiple interconnecting windows on the walls of each macropore were observed, indicating a well-prepared HAS. To determine the attachment of BMSCs on different surfaces, cells seeded on top of the HAS or HAS-G were observed by SEM. As presented in Fig. 1B, BMSCs attached to both the HAS and HAS-G, and no differences were observed in cell morphology or numbers.

Figure 1.

Identification and physical characterization of the HAS and the HAS-G. (A) The surface micromorphology of the HAS and the HAS-G was detected by scanning electron microscopy. (B) The morphology of BMSCs attached to the surface of the HAS and the HAS-G were detected by scanning electron microscopy. BMSCs, bone marrow-derived stromal cells; HAS, hydroxyapatite scaffold; HAS-G, hydroxyapatite scaffold with a groove structure.

To determine the cellular basis of the effects of the HAS-G on the physiological processes of BMSCs, the effect of the HAS-G on the proliferation of BMSCs was evaluated. As demonstrated in Fig. 2A, in both NM and DM, the HAS-G exerted a promoting effect on cell viability of BMSCs compared with that of the HAS. Cell cycle phases were analysed by flow cytometry following PI staining, which demonstrated that in both types of medium, compared with cells attached to the HAS, the proportion of cells attached to the HAS-G in the G1/G0 phase was decreased, whereas that of cells in the S phase was increased, without altering the G1/M phase, indicating that the HAS-G may promote cell proliferation by accelerating the cell cycle at the G1/G0 phase (Fig. 2B). Considering the aim of bone replacement, implanted biomaterials induce increased production of ROS at tissue-implanted biomaterials (14–16), leading to inflammation, which is known as a major cause of implant failure. Therefore, the ROS levels in BMSCs were analysed. In both types of medium, cells on the HAS-G exhibited decreased the ROS levels compared with those of cells on the HAS, which may potentially explain why the HAS-G promoted BMSC proliferation (Fig. 2C).

Figure 2.

Surface pattern of the HAS-G modifies BMSC proliferation and decreases ROS accumulation. (A) Cell viability of BMSCs attached to the surface pattern of the HAS or the HAS-G was measured by performing a cell counting kit-8 assay. (B) The proportion of BSMCs in each cell cycle phase attached to the surface pattern of the HAS or the HAS-G was measured by flow cytometry following PI staining. (C) ROS accumulation was detected in BMSCs attached to different surface patterns. *P<0.05 vs. HAS. BMSCs, bone marrow-derived stromal cells; ROS, reactive oxygen species; HAS, hydroxyapatite scaffold; HAS-G, hydroxyapatite scaffold with a groove structure; PI, propidium iodide.

HAS-G promotes osteogenesis of BMSCs

As a proliferation-promoting effect of the HAS-G on BMSCs was identified, the present study further investigated whether the HAS-G modified the processes of osteogenic differentiation. Firstly, the differentiation of BMSCs in NM and DM were investigated at the molecular level by evaluating the hallmarks of osteogenesis, including RUNX2, OCN and OPN. As demonstrated in Fig. 3A, immunofluorescent staining presented that in DM, the HAS-G promoted the immunofluorescent signal of RUNX2, OCN and OPN compared with that of cells on the HAS. Additionally, the promoting effects of the HAS-G on their expression were further confirmed by performing semi-quantitative western blotting (Fig. 3B). However, in NM, compared with that of cells attached to the HAS, the HAS-G exhibited no notable promotion on the expression of these three proteins without the presence of a differentiation promotor, indicating that the HAS-G promoted osteogenesis by decreasing proliferation inhibition. The results of cultured cell staining with ARS revealed dark red staining of calcium depositions in the cells that were grown on the HAS-G and cultured in DM Fig. 3C. Weaker staining was observed in the cells growing on HAS compared with HAS-G, which further supported the promoting effect on osteogenesis of the HAS-G.

Figure 3.

Surface pattern of the HAS-G promotes the osteogenesis of BMSCs. (A) After 14-day culture on different patterns, the expression of RUNX2 and the production of bone-related proteins OCN and OPN were imaged using immunofluorescence staining. (B) Western blotting was performed to detect the protein levels of RUNX2, OCN and OPN. (C) To detect the effects of different patterns on the production of mineralized extracellular matrix, cells were stained with Alizarin Red S on day 14. *P<0.05 vs. HAS/DM. HAS-G, hydroxyapatite scaffold with a groove structure; BMSCs, bone marrow-derived stromal cells; RUNX2, runt-related transcription factor 2; OCN, osteocalcin; OPN, osteopontin; HAS, hydroxyapatite scaffold; DM, differentiating medium; NM, normal medium.

HAS-G maintains the mitochondrial membrane potential by decreasing ROS levels

To investigate whether the decreased ROS level by the HAS-G protected the cells from mitochondrial dysfunction, the mitochondrial membrane potential was detected with or without the presence of an ROS scavenger NAC, which was demonstrated to scavenge ROS level in BMSCs (Fig. 4A). Confocal microscopy detection of J-aggregates and J-monomers was used in BMSCs attached to different surface patterns following staining with JC-1. As demonstrated in Fig. 4B, fluorescence was observed in small subcellular particles in the cellular space distinct from the nuclei labelled by Hoechst33342. Aggregates and monomers were observed in the cells in the same locations in the HAS-G group, indicating the maintained mitochondrial membrane potential. By contrast, the fluorescence level of aggregate labelling of the HAS-attached BMSCs was lower compared with that of monomers. To further confirm that the HAS-G induced mitochondrial ROS, MitoSOX™ Red was employed. As presented in Fig. 4C, in the Mock condition, mitochondrial superoxide levels were decreased in the HAS-G group compared with the HAS group.

Figure 4.

HAS-G maintains mitochondrial potential by decreasing ROS levels. (A) ROS levels were measured in BMSCs attached to the HAS and the HAS-G. (B) JC-1 staining was performed to detect the mitochondrial membrane potential. Red aggregates, normal mitochondrial membrane potential; green monomers, low mitochondrial membrane potential. (C) Photographs of MitoSOX Red fluorescence staining. #*P<0.05. ROS, reactive oxygen species; BMSCs, bone marrow-derived stromal cells; HAS, hydroxyapatite scaffold; HAS-G, hydroxyapatite scaffold with a groove structure; NAC, N-acetylcysteine; Mock, cells seeded on HAS/HAS-G without disturbance of ROS accumulation.

HAS-G inhibits mitophagy/autophagy and enhances mitochondrial function

ATP synthesis of BMSCs was analysed to determine whether mitochondrial function was promoted in cells seeded on the HAS-G. As demonstrated in Fig. 5A, cells seeded on the HAS-G produced significantly more ATP compared with those seeded on the HAS. Of note, in BMSCs pre-incubated with NAC, BMSCs seeded on the HAS still produced less ATP compared with cells seeded on the HAS-G without disturbance of ROS accumulation (Mock), indicating that mitochondrial dysfunction was present. To determine whether ROS accumulation induced mitophagy/autophagy and led to mitochondrial dysfunction, the markers of mitophagy/autophagy were detected by western blotting. The results demonstrated that HAS-G decreased the expression levels of LC3-2 and OPTN, but increased those of p62, TIM23, TOM20 and HSP60 (Fig. 5B) protein expression compared with the HAS. Double labelling of the mitochondria (red; TOM20) and the autophagosome (green; enhanced green fluorescent protein-LC3B) also presented that, following pretreatment with 10 µM Bafilomycin A1 (Baf A1), colocalization of TOM20 and LC3B was observed in cells attached to the HAS, which was decreased in BMSCs attached to the HAS-G (Fig. 5C), indicating that BMSCs on the HAS-G presented weaker mitophagy/autophagy compared with those on the HAS.

Figure 5.

HAS-G inhibits mitophagy compared with HAS. (A) ATP production was measured in HAS- or HAS-G-attached BMSCs pretreated with or without 10 µM NAC. (B) The hallmarks of mitophagy, including LC3 maturation, optineurin, p62, TIM23, TOM20 and HSP60 were measured by western blotting. (C) Confocal laser scanning microscopy colocalization images of TOM20 and EGFP-LC3B staining. *P<0.05 vs. HAS. HAS, hydroxyapatite scaffold; HAS-G, hydroxyapatite scaffold with a groove structure; BMSCs, bone marrow-derived stromal cells; NAC, N-acetylcysteine; LC3, microtubule-associated proteins 1A/1B light chain 3B; p62, nucleoporin p62; TIM23, translocase of the inner membrane 23; TOM20, translocase of the outer membrane 20; HSP60, heat shock protein 60; OPTN, optineurin; EGFP, enhanced green fluorescent protein.

Discussion

The mechanisms by which the HAS-G promotes osteogenesis have not been reported previously. It has been reported that HAS-G significantly promoted hPMSC osteogenic differentiation and the maturation of osteoblasts both in vitro and in vivo, but the exact mechanism is not completely understood (8). In the present study, a HAS with a 25–30-µm microgroove was used as the growing surface of BMSCs. The results demonstrated that the HAS-G promoted cell proliferation and osteogenesis, potentially by decreasing ROS accumulation and maintaining the mitochondrial membrane potential.

Groove structure is widely used to modify several biomaterials, including titanium alloy substrates (17), polystyrene (18), grapheme (19) and photonic crystals (20). The effects of cell surface interactions on physiological properties have been studied in detail. Kaiser et al (17) have reported that key factors in normal physiological processes, including embryogenesis, morphogenesis and osteogenesis, are affected by grooved structures by modifying the cell shape, migration angle α, cell orientation β and migration velocity. Rajnicek et al (21) have reported that grooved substrates control corneal and lens epithelial cell morphology, and induce these cells to become longer and thinner. Feng et al (22) have also revealed that grooved micropatterns of fibres affect the adhesion and proliferation of human vascular endothelial cells. In our previous study, the grooved structure of the HAS was also presented as a surface pattern that promoted hPMSCs (8). Compared with the effects of a regular HAS, the HAS-G significantly promoted osteogenic differentiation and the maturation of osteoblasts in DM, which was confirmed by measuring osteogenic proteins in vitro and new bone formation and vasculogenesis in vivo (8). Interestingly, the results of the present study suggested that, in NM, cultured BMSCs exhibited little osteogenic differentiation on the HAS or the HAS-G, indicating that the HAS-G promoted osteogenic differentiation and that osteoblast maturation under stimulating conditions mainly exerted protective effects.

ROS accumulation in implantology is considered one of the major causes of implant failure by inducing cell dysfunction and biomaterial corrosion (14–16). Tsaryk et al (15) have reported that endothelial cells grown on titanium substrate present decreased cell viability and metabolic activity due to the high accumulation of endogenous ROS. In addition, oxidative stress H2O2 treatment induces markedly higher ROS accumulation in the titanium substrate compared with that of the control material (15). Human neutrophil responses to aluminium oxide membranes with different pore sizes have been evaluated, and the results have revealed that ROS accumulation is the major inducer of cell death; strategies against ROS accumulation improved cell-cell interactions, wound healing and tissue regeneration (14). Poly(ε-caprolactone), which is considered to be a potential substrate for widespread medical applications, also leads to ROS accumulation in attached L929 mouse fibroblasts, and oxidative stress notably controls cells at several time points (16). In the present study, cells grown on the HAS-G displayed a higher proliferation rate compared with cells on the HAS, which was reversed by the addition of the ROS scavenger NAC. This indicated that the HAS-G may provide a more suitable environment for cell proliferation by inhibiting ROS accumulation. The results of the present study also demonstrated that compared with the HAS, the HAS-G inhibited the conversion of LC3B I to II, which indicated its inhibitory effect on autophagy/mitophagy. However, the present study failed to perform transmission electron microscopy (TEM) due to the structure of the HAS and the HAS-G. In further studies, PTEN-induced kinase 1 and parkin RBR E3 ubiquitin protein ligate may be detected to validate the inhibition of autophagy/mitophagy by the HAS-G.

In the present study, the HAS induced ROS accumulation in attached BMSCs, whereas the HAS-G exhibited lower ROS levels in BMSCs. Lower levels of ROS accumulation led to improved mitochondrial function and less mitophagy/autophagy, which potentially promoted osteogenesis. The present study also demonstrated that compared with BMSCs attached to the HAS, those attached to the HAS-G exhibited fewer mitochondrial-specific ROS. Total and mitochondria-specific ROS accumulation was measured, but the present study failed to confirm whether intracellular ROS was induced (data not shown), which may be investigated in further studies.

In conclusion, the grooved structure of the HAS presented improved growing properties compared with the HAS with a smooth surface. This structure enhanced cellular physiological processes, promoted cell viability and maintained the mitochondrial membrane potential, which may be the main reason for the protective effects observed in the attached cells.