The Influence of Vitamin C–incorporated Polycaprolactone on Osteogenesis in Osteoblast-Osteoclast Co-culture In Vitro

Abstract

Objectives

Implantation of biomaterials generates reactive oxygen species (ROS) in the peri-implant microenvironment. Bone loss occurs when ROS levels exceed the local antioxidant capacity. The aim of this study was to investigate the influence of vitamin C–incorporated polycaprolactone (PCL-Vit C) membrane in scavenging ROS and enhancing biomineralisation in osteoblast–osteoclast (OB-OC) co-culture system.

Methods

OB-OCs were cultured on polycaprolactone (PCL) and PCL-Vit C membranes under osteogenic conditions to mimic the bone microenvironment. ROS generation was measured using flow cytometry. ALP and the RANKL/OPG ratio were determined by colorimetric and ELISA assays, respectively. Gene expression of ALP, Col1, Runx-2 and OCN and protein expression of Runx-2, BMP-7, Col1 and OCN were determined by real-time PCR and western blotting, respectively. Activation of P38, ERK and JNK and beta-catenin expression was also analysed.

Results

OB-OC grown on PCL membrane generated a higher amount of ROS compared to those on PCL-Vit C. Colorimetric assays revealed a significantly higher ALP activity in OB-OC co-cultures on PCL-Vit C membranes. Furthermore, OB-OC grown on PCL-Vit C membrane showed significant upregulation in mRNA levels of ALP, Col1 and OCN with lower RANKL/OPG ratio and higher amounts of mineralisation nodules. Runx-2 expression was comparable in both membranes. Western blotting showed a significant increase in phosphorylation of P38 MAPK, ERK, JNK and beta-catenin expression in OB-OC maintained on PCL membrane, plausibly due to increased ROS levels, compared to PCL-Vit C.

Significance

OB-OC grown on PCL-Vit C membrane scavenged ROS and supported a higher osteogenic potential environment for OB-OC co-culture in vitro.

Graphical Abstract

Introduction

Implantation of biomaterials into the human body leads to the generation of reactive oxygen species (ROS), which is often overcome by the local tissue responses at the peri-implant microenvironment.¹ An appropriate amount of ROS is crucial for wound healing because these reactive species are now known to act as second messengers in cellular signal transduction.² However, prior presence of ROS at the implantation site resulting from a previous infection and trauma from implant surgery itself, along with the host cell–biomaterial interaction, may generate a significant amount of ROS that exceeds the local antioxidant capacity, leading to local oxidative stress and the generation of an intense pro-inflammatory environment. The individual may be having a systemic disease with high ROS levels such as diabetes mellitus that may complicate the local pro-inflammatory situation.^3,4

In bone, the generation of ROS and oxidative stress around implants has a direct impact on osteoblasts and osteoclasts, which are cells that control the continuous bone remodelling process. Whereas it is known that ROS-mediated pro-inflammatory conditions drive osteoclastogenesis and stimulate bone resorption, less is known regarding the association of ROS with osteoblasts and its coupling mechanism with osteoclasts during oxidative stress.⁵ The Mitogen Activated Protein Kinase (MAPK) pathway plays a critical role in bone formation because it can crosstalk with all molecular pathways, but the physiological role of this signalling pathway in osteoblasts remains controversial.⁶ Interestingly, ROS is known to activate MAPK pathways, but the mechanisms for these effects is still unclear.⁷ Some investigations support a stimulatory role of MAPK pathway in osteoblast differentiation whereas others suggest that this signalling is inhibitory.^{8, 9} The interaction of MAPK with Wnt signalling protein β-catenin following ROS stimulation is even less understood, although both signals are key to osteoblast differentiation. Furthermore, such biomaterials were often investigated individually on osteoblasts and osteoclasts in monoculture systems. A direct osteoblast–osteoclast co-culture model could be designed to mimic the in vivo cellular environment that allows cellular communication and the exchange of soluble factors that may demonstrate a more realistic oxidative stress mechanism.¹⁰

Manipulating the level of ROS within a wound-healing microenvironment by using antioxidants may allow modulation of the local immune system and control intracellular signalling of osteoblasts and osteoclasts that influence bone remodelling. These opportunities may improve bone tissue engineering outcomes in the future. The impact of local delivery of antioxidants in mitigating oxidative stress during biomaterial implantation in bone has not been fully investigated, and reports on its efficacy in osteoblast–osteoclast co-culture are scanty. Although antioxidants such as vitamin E, N-acetylcysteine, alginate, and xanthan gum have been applied locally, vitamin C has been chosen for scavenging ROS directly because of its extensive therapeutic use in chronic inflammatory diseases¹¹ and its established role in wound healing.¹² Furthermore, local delivery of vitamin C using suitable carriers may allow quick diffusion into target cells and tissues because it is smaller than other antioxidants.

Polymer-based bone tissue engineering scaffolds such as polycaprolactone, which is biocompatible with low immunogenicity, have been previously investigated as suitable carriers for vitamin C.¹³ The aim of this study was to investigate the influence of vitamin C–incorporated polycaprolactone membrane scaffold for enhancing osteogenesis in osteoblast–osteoclast co-culture system.

Materials and methods

Preparation of polycaprolactone (PCL) membrane

Preparation of the porous electrospun PCL and PCL-Vit C membrane is detailed previously.¹³ In brief, 11 wt % of PCL was initially mixed in a solution of chloroform and dimethyl sulfoxide (DMSO) at a ratio of 9:1. The resulting solution was subsequently introduced into the electrospinning machine. A needle (21-gauge) was placed at a distance of 22 cm (needle-to-collector distance). A 0.5-mL/h flow rate was maintained, and the drum speed was set at 220 rpm. For the PCL-Vit C formulation, L ascorbic acid was incorporated at 25 wt % relative to PCL. L ascorbic acid prepared in DMSO was combined with PCL solution in DMSO and chloroform. The nanofiber membrane was then fabricated as outlined earlier. Both membranes were subjected to physico-chemical characterisation and biological evaluation studies using human foetal osteoblast cells.¹⁴

Cell culture preparation of human osteoblasts

Human foetal osteoblast cells (hFOB 1.19; Addexbio), were cultured in Dulbecco's Modified Eagle Medium/Nutrient Mixture F12 (DMEM/F-12), supplemented with 10% foetal bovine serum (FBS) and 1% penicillin-streptomycin. These cells were incubated at 37 °C in a humidified atmosphere containing 95% oxygen and 5% carbon dioxide. Sub-culturing was performed using 1x trypsin-EDTA (Sigma) once the cells reached confluence, and the culture medium was refreshed every 2 days.

Cell culture preparation for differentiation of THP-1 monocyte cells into osteoclasts and characterisation by TRAP assay

THP-1 monocytes, obtained from CLS, maintained in complete RPMI-1640 medium were used in the study. To induce differentiation into M0 macrophages, 100 ng/mL phorbol 12-myristate 13-acetate (PMA) was added to THP-1 cells for 24 hours following the method used in Chanput et al.¹⁵ After the PMA treatment, cells were cultured in PMA-free medium for an additional 72 hours. Subsequently, the differentiation of THP-1 cells into osteoclasts was initiated by treating the macrophage-like cells with RANKL (receptor activator of nuclear factor κ-B ligand) and M-CSF (macrophage colony-stimulating factor) (R&D Systems), 50 ng/mL each for 8 days. The culture medium was refreshed every 3 days to ensure optimal conditions.

At the end of the differentiation process, the cells were subjected to tartrate-resistant acid phosphatase (TRAP) staining for osteoclast identification. The cells were first washed with phosphate-buffered saline (PBS) and fixed in a 10% formalin neutral buffer solution for 5 minutes at room temperature. After fixation, the cells were rinsed with deionised water and incubated with a chromogenic substrate at 37 °C for 20-60 minutes. Following incubation, the cells were washed again with deionised water and subsequently examined using an inverted microscope.

Direct co-culture of hFOB 1.19 and THP-1–derived osteoclast

To mimic the in vivo bone environment within electrospun membranes, a direct co-culture model involving hFOB 1.19 and THP-1–derived osteoclast was developed. The membranes, each measuring 1 × 1 cm, were sterilised using ultraviolet light for 30 minutes. They were then placed in 12-well plates and supported by sterile stainless-steel rings. Prior to cell seeding, the membrane was incubated with complete media for 2 hours in a CO₂ chamber. After removing the media, THP-1 cells grown in complete media supplemented with 100 ng/mL PMA were seeded onto the membrane at a density of 1 × 10⁶ cells per well in a maximum volume of 100 µL. This setup was incubated at 37 °C with 5% CO₂ for 2 hours. The culture was then replenished with an additional 1 mL of complete culture media and incubated overnight. The following day, the media was supplemented with 50 ng/mL RANKL and M-CSF to promote osteoclast formation for the next 8 days. To establish a direct co-culture, hFOB 1.19 osteoblasts were seeded (5 × 10⁴ cells) on top of the osteoclast cells. The cultures were maintained in a 1:1 mix of RPMI and DMEM F12 culture media.

ROS detection and measurement

ROS estimation kit (ROS Kit) from Abcam was used to estimate total ROS. Osteoblast–osteoclast (OB-OC) co-cultures, maintained on the electrospun membranes for 14 days, were collected and treated with an ROS assay reagent. Negative control cells were treated with hydrogen peroxide (H₂O₂). For the assay, 2 µL of the ROS reagent was added to the cell mixture suspended in ROS buffer, followed by a 1-hour incubation in a CO₂ incubator. Post-incubation, the cells underwent analysis via flow cytometry using the FACS Aria III system (Becton Dickinson).

Alkaline phosphatase Assay

An alkaline phosphatase (ALP) assay was conducted using a colorimetric assay kit (ab83369; Abcam) according to the manufacturer’s protocol. Cell lysates were obtained from control osteoblast–osteoclast co-cultures, as well as those cultured under osteogenic conditions on both electrospun membranes for up to 14 days. These lysates were then added to 96-well plates and treated with pNPP solution for 60 minutes at RT. Simultaneously, ALP standards were prepared, and absorbance was recorded at 405 nm. All experiments were executed in triplicate, using 3 independent samples.

Measurement of RANKL/OPG ratio in co-culture

To determine the effects of electrospun membranes on osteoblast mineralisation and osteoclastic activity, we determined the levels of receptor activator of nuclear factor kappa-B Ligand (RANKL) and osteoprotegerin (OPG) protein using ELISA assays on cells seeded on both membranes. Cell culture supernatants from osteoblast–osteoclast co-cultures were collected on day 14 for detection of the amount of RANKL using the Human RANKL ELISA (Abclonal) and OPG using the human Osteoprotegerin ELISA kit (Abcam). Values are represented in a ratio of RANKL/OPG.

Osteoblast Mineralization

To perform osteoblastic differentiation assays within the OB-OC co-culture system, a combination of RPMI and DMEM-F12 media in a 1:1 ratio was used. The medium was further supplemented with FBS at a concentration of 10%, antibiotics at 1%, dexamethasone (10 mM) from Sigma and beta-glycerophosphate (5 mM) from Merck Millipore. The differentiation media was refreshed every 3 days, and the cultures were sustained over a 14-day period. As a control, OB-OC co-cultures maintained in tissue culture well plates were used.

Alizarin red staining and quantification of calcium release

OB-OC co-cultures were maintained on both membranes for a period of 14 days, alongside control samples in 24-well culture plates. Matrix mineralization was assessed using Alizarin staining to measure the deposition of calcium by osteoblast cells. On the 14th day, both the control and the cells seeded on the membranes were fixed at room temperature using a 4% paraformaldehyde solution for 15 minutes. Afterward, the samples were stained with a 40 mM Alizarin stain for 20 minutes. Post-staining, the Alizarin Red S solution was carefully removed, and the samples were rinsed in deionised water. Images were acquired using an Olympus CKX 41 inverted microscope.¹⁶

To determine calcium release, a 10% solution of acetic acid (v/v) was added to the cells that were stained with Alizarin Red in controls, PCL, and PCL-Vit C membrane in cell culture plates. Alizarin dye was extracted using the following previously published protocols,¹⁶ and quantification was done at 405 nm. A calibration curve with Alizarin dye was created to measure total calcium release. Each experiment was conducted in triplicate, using 3 independent samples for precision.

RNA isolation and real-time PCR

Real-time PCR was performed to measure the expression of osteogenic genes that include Runt-related transcription factor 2 (Runx-2), ALP, collagen type I (Col 1), and osteocalcin (OCN). RNA was extracted from OB-OC co-cultures on day 14. Co-cultures grown on cell culture plates were used as controls. The RNA isolation was performed using the Qiagen RNeasy kit (Qiagen) according to the manufacturer's instructions. cDNA synthesis was achieved using the SuperScript III First-Strand Synthesis System (Invitrogen). Quantitative real-time PCR analysis was then conducted using the Hot FirePol SYBR Green Master mix (Solis Biodyne) in a StepOne Real-time PCR System (Life Technologies). Human GAPDH served as the normalisation control, and each experiment was conducted in triplicate. Details of the primers used in the reactions are listed in Table 1.

Table 1.

Primer sequences used in real-time PCR analysis.

Gene	Primer Sequences
GAPDH	5′- CCACTCCTCCACCTTTGACG- 3′ 5′-CCACCACCCTGTTGCTGTAG-3′
Runx-2	5’-ACGCGAGTCTGTGTTTTTGC-3’ 5’- CATGGTGCGGTTGTCGTG-3’
Alkaline Phosphatase	5’-CTGGGTGCAGGATAGCAGTC-3’ 5’-CCTCCTGGCCTGTGTTATCC-3’
Collagen Type 1	5′-ACGAAGACATCCCACCAATCA-3′ 5′-TCACGTCATCGCACAACACC-3′
	5’- CACCGAGACACCATGAGAGC-3’ 5’- CTCTTCACTACCTCGCTGCC-3’

Western blotting analysis for osteogenic proteins

Cell lysates were prepared using RIPA lysis buffer (Elab Biosciences). Protein concentrations were determined with a BCA kit (Pierce). Samples containing 40 µg of protein were heated at 95 °C for 5 minutes, run on a 10% SDS-PAGE, and then electrotransferred to a nitrocellulose membrane (Amersham). All the primary antibodies used were purchased from Abcam. Immunodetection was carried out using anti-human primary antibodies against Runx-2 (rabbit monoclonal), BMP-7 (rabbit polyclonal), collagen type 1 (rabbit polyclonal), and osteocalcin (rabbit polyclonal) to assess osteogenic lineage expression.

Western blotting analysis for MAPK and Wnt proteins

The impact of MAPK signalling pathways on the osteogenic influence of the PCL-Vit C membrane was examined via Western blot analysis. Specific rabbit monoclonal antibodies against human ERK, phospho-ERK, JNK, phospho-JNK, P38, and phospho-P38 (1:1000) as well as beta catenin (rabbit polyclonal) were employed to analyse MAPK signalling activation in OB-OC co-culture seeded in PCL and PCL-Vit C membrane. Beta actin (Cell Signalling Technologies) was used for data normalisation, and HRP-conjugated rabbit anti-goat was the secondary antibody used (Cell Signalling Technologies).

For experimental controls, osteoblast cells and OB-OC co-culture grown in cell culture plates with osteogenic differentiation media (OS+) were used. Immunodetection was visualised using Clarity Western ECL substrate (Bio-Rad Laboratories).

Statistical analysis

The data obtained represent results from a minimum of three independent experiments, each conducted in triplicate.^17,18 The outcomes of these experiments are presented as the mean ± SEM. Both one-way and two-way analysis of variance (ANOVA) were employed, followed by multiple comparison tests. A P value of less than .05 was deemed statistically significant. The statistical analyses were performed using GraphPad Prism version 9 software (GraphPad Software).

Results

Membrane morphology and physico-chemical characterisation

Both PCL and PCL-Vit C membranes showed variation in pore size and diameter. In the PCL membrane, the mean diameter of pores is 109.46 ± 35.36 nm and the mean diameter of fibres produced is 759.05 ± 10.11 nm, whereas in PCL-Vit C, the mean diameter of the pores is 93.29 ± 58.16 nm and the mean diameter of the fibre is 515.25 ± 14.07 nm. The pores for both membranes are irregularly shaped with varying depth and a wide size distribution. The addition of 25 wt% vitamin C significantly reduced both the pore size (P < .05) and fibre diameter (P < .01) of the PCL membrane¹³ and rendered the surface property more hydrophilic.

Differentiation of THP-1 monocyte cells into osteoclasts

THP-1 cells that had differentiated into macrophages under PMA stimulation for 24 hours showed a mixed population of large round cells and long spindle-shaped adherent cells. Following stimulation with RANKL and M-CSF, the cells fused further into larger multinuclear cells, with an irregular morphology (Figure 1A). TRAP staining revealed the presence of wine-red particles (TRAP-positive cells) by day 5 following stimulation with RANKL and M-CSF (Figure 1B), whereas by day 8, large multinuclear cells THP-1–derived osteoclasts were evident in culture (Figure 1C).

Fig. 1.

Osteoclast formation and TRAP staining. A, THP-1 cells differentiated into multinuclear osteoclast cells upon stimulation with 50 ng/mL RANKL and M-CSF. B, TRAP-positive cells on day 5 of RANKL and M-CSF stimulation. C, Large multinucleated TRAP-positive cells were seen by day 8 of RANKL and M-CSF stimulation. TRAP-positive cells were characterized by wine-red particles in the cytoplasm and contained 3 or more nuclei. TRAP, tartrate-resistant acid phosphatase. Magnification: 20 µm.

ROS measurement

ROS levels in OB-OC co-cultures maintained on PCL alone and PCL-Vit C membrane were higher compared to the control, but the results were not significant. Interestingly, ROS generation in OB-OC co-culture maintained on PCL-Vit C membrane was significantly lower than that on PCL-only membrane (P < .05) (Figure 2).

Fig. 2.

Flow cytometry analysis of ROS generation in osteoblast–osteoclast (OB-OC) co-culture grown on PCL and PCL-Vit C membranes at day 14 of culture. Bar graphs showing ROS positive cells (Alexa fluorophore 488 positive) OB-OC co-culture control grown on cell culture plates (A), osteoblast-osteoclast co-culture grown on the PCL-only membrane (B), and OB-OC co-culture maintained on the PCL-Vit C membrane (C). D, The bar graph represents the mean fluorescent intensity. Data represent the mean ± SEM of 3 independent experiments. *P < .05, ** P < .01.

ALP activity and measurement of RANKL/OPG ratio in OB-OC co-culture

ALP activity was enhanced in OB-OC co-cultures maintained on both PCL-only membrane and PCL-Vit C membranes (P < .0001) compared to the control. When PCL alone was compared to PCL-Vit C membrane, ALP activity was significantly higher (P < .001) in OB-OC co-culture grown on PCL-Vit C membrane (Figure 3A).

Fig. 3.

Bone formation and bone resorption in osteoblast–osteoclast co-culture. A, Quantification of alkaline phosphatase activity (ALP U/mL) at day 14 of osteoblast–osteoclast co-culture maintained on PCL and PCL-Vit C membranes compared to the control; all grown in osteogenic conditions. B, RANKL/OPG ratio in osteoblast–osteoclast co-culture maintained on PCL and PCL-Vit C membranes compared to the control reflecting bone formation and resorption at day 14 under osteogenic conditions. Data represent the mean ± SEM of 3 biological replicas. *P < .05, **P < .01, ***P < .001, ****P < .0001.

Likewise, the ratio of RANKL/OPG, increased significantly in OB-OC co-cultures maintained in both PCL (P < .001) and PCL-Vit C membranes compared to OB-OC co-culture controls (P < .05). However, in the PCL-Vit C membrane, the RANKL/OPG ratio was significantly lower than in the PCL-only membrane (P < .01; Figure 3B).

Alizarin red staining and quantification of calcium release

Alizarin staining displayed bright orange nodules in OB-OC co-cultures grown on both PCL-only membrane and PCL-Vit C membrane with a more dispersed mineral distribution compared to the OB-OC co-culture control (Figure 4). In the OB-OC co-culture control group (Figure 4A), calcified nodules were visible, indicating baseline osteogenic differentiation and mineral deposition in the extracellular matrix. The co-culture grown on the PCL-alone membrane (Figure 4B) exhibited similar patterns of calcification, although the mineral deposition appeared less pronounced than in the PCL-Vit C membrane group.

Fig. 4.

Extracellular matrix mineralization in osteoblast–osteoclast (OB-OC) co-cultures grown on PCL and PCL-Vit C membranes. A, Calcification nodules (arrows) at day 14 in the presence (+) of osteogenic supplements in the control OB-OC co-culture. B, OB-OC co-culture on the PCL membrane at 14. C, OB-OC co-culture on the PCL-Vit C membrane. Scale bar = 50 µm for all panels. D, Quantification of alizarin-stained mineralized nodules showing calcium release at day 14 in culture. Data represent the mean ± SEM of independent experiments. *P <.05, **P <.01.

The quantitative analysis of calcium deposition using Alizarin Red S staining, as shown in Figure 4D, confirmed these observations. By day 14, calcium ion release in the OB-OC co-culture grown on both the PCL and PCL-Vit C groups increased significantly showing P values of <.05 and <.01, respectively, compared to the OB-OC control group. However, there was no significant difference in Ca²⁺ ion release between the 2 membrane groups.

Quantitative real-time PCR analysis

To examine the expression levels of osteogenic markers, quantitative real-time PCR was employed. The markers analysed included Runx-2 (runt-related transcription factor 2), ALP (alkaline phosphatase), Col 1 (collagen type 1), and OCN (osteocalcin) in OB-OC co-cultures grown on both PCL and PCL-Vit C membranes for 14 days (Figure 5). Expression of osteogenic markers like ALP and Col 1 was significantly upregulated in OB-OC grown on the PCL-Vit C membrane under osteogenic (+) conditions than on the PCL-only membrane (P < .0001; Figure 5). Expression of the osteogenic transcription marker Runx-2 was comparable in both types of membranes, and there was also no significant difference in gene expression for OCN among both type membranes, although an increasing trend was seen in the PCL-Vit C membrane.

Fig. 5.

Quantitative real-time PCR analysis for the expression of osteogenic markers in OB-OC co-culture grown on PCL and PCL-Vit C membranes for 14 days. Bar graphs representing the fold change in the levels of expression of Runt-related transcription factor 2 (Runx-2), alkaline phosphatase, collagen type 1 and osteocalcin (OCN). Data represent the mean ± SEM of 3 independent experiments. ***P < .0001.

Western blotting analysis of osteogenic marker expressions

For western blotting studies, both osteoblast cells and OB-OC co-culture grown in cell culture plates in osteogenic conditions (+) were used as experimental controls while OB-OC co-culture grown on both membranes were the test groups.

At day 14, there was a significant reduction in Runx-2 expression in both PCL and PCL-Vit C groups compared to the control, with the PCL-Vit C group showing a significant reduction (P < .05) compared to the OB-OC control group (Figure 6A and B). OB-OC co-culture in both membranes exhibited significantly higher expression of BMP 7 compared to the OB (P < .01) and OB-OC (P < .05) control group, whereas the BMP 7 expression in the PCL-Vit C membrane was statistically lower than that of the PCL-only membrane (P < .05; Figure 6A and C). Although Col 1 expressions seem comparable among the 4 study groups (Figure 6A and D), a significant reduction (P < .01) in osteocalcin expression was observed in OB-OC cells grown in the PCL-only membrane (Figure 4.5A and E). Interestingly, PCL-Vit C-grown OB-OC co-culture cells showed a significant increase (P < .05) in OCN expression compared to the PCL-only membrane (Figure 6D).

Fig. 6.

Western blotting and data quantification of osteogenic markers. A, Western blot analysis of the expression of osteogenic markers such as Runx-2, bone morphogenic protein-7 (BMP-7), collagen type 1 (Col 1) and osteocalcin (OCN). Beta-actin expression was used for data normalization. Densitometric analysis of blots obtained for expression of (B) Runx-2, (C) BMP-7 (D) Col 1 and (E) OCN osteogenic markers. OB, osteoblast cells; OB-OC, osteoblast–osteoclast co-culture representing culture in osteogenic conditions. PCL and PCL-Vit C represent an OB-OC co-culture grown in osteogenic conditions in PCL-only and PCL-Vit C membranes. Data represent the mean ± SEM of three independent experiments. *P < .05, **P < .01.

Western blotting analysis of MAPK and β-catenin signalling protein expressions

Western blot results showed that OB-OC co-culture grown on PCL-Vit C can attenuate P38, ERK, and JNK in the MAPK signalling cascade compared to that grown on the PCL-only membrane (Figure 7). The ratio of the relative intensities of the phosphorylated protein to the corresponding total protein of the MAPK signalling pathways are shown (Figure 7B-D). Expression of β-catenin is also significantly decreased in OB-OC co-culture grown on the PCL-Vit C compared to PCL-only membrane (Figure 7E). The ratio of p-P38/P38 was significantly increased in OB-OC control, PCL-only, and PCL-Vit C membranes compared to control OB cells alone (P < .0001), while the intensity ratio was significantly higher (P < .0001) in PCL-only compared to the one in the PCL-Vit C membrane (Figure 7A and B).

Fig. 7.

Western blotting and data quantification of proteins in the mitogen-activated protein kinase (MAPK) pathway. A, Representative western blots of total and phosphorylated MAPK family proteins P38, ERK, JNK and Wnt signalling protein β-catenin. Beta-actin expression was used for data normalization. The bar graphs show the relative phosphorylation ratio of the phosphorylated protein to the total level of respective proteins such as (B) p-P38, (C) p-ERK, (D) p-JNK and (E) β- catenin. OB, osteoblast cells; OB-OC, osteoblast-osteoclast co-culture representing culture in osteogenic conditions. PCL and PCL-Vit C membranes represent an OB-OC co-culture grown in osteogenic conditions in PCL-only and PCL-Vit C membranes. Data represent the mean ± SEM of three independent experiments. **P <.01, ***P < .001, ****P < .0001.

Similarly, the ratio of pERK/ERK was significantly higher in OB-OC control (P < .0001), PCL-only membrane (P < .0001), and PCL-Vit C membrane (P < .01) than in OB control cells (Figure 7A and C). The pERK/ERK ratio was also significantly higher in OB-OC grown on the PCL-only membrane group than on the OB-OC control and OB-OC grown on the PCL-Vit C group (P < .001). Furthermore, the ratio of pJNK/JNK was significantly higher (P < .001) in OB-OC grown on the PCL-only membrane group than in OB-OC grown on the PCL-Vit C membrane group (Figure 7A and D). Notably, β-catenin and JNK expression were negligible in OB and OB-OC controls (Figure 7A).

Discussion

ROS are produced from both internal cellular reactions and external causes^9,19,20 following biomaterial implantation. Many clinical steps are taken by clinicians to minimise generation of ROS during routine biomaterial implantation operative procedures such as adhering to strict sterilisation protocols, refining surgical procedures to minimise tissue trauma, reduce inflammation using anti-inflammatory medications and control of infection with antibiotics. Systemic administration of antioxidants to patients is also widely practised to control ROS and encourage wound healing.²¹ However, it may be more effective to scavenge the ROS if an advanced biomaterial can be designed to mitigate excessive ROS around the surgical injury site and release their payloads that include antioxidants such as vitamin C at the cellular level, downregulating oxidative stress and promoting a conducive environment for tissue regeneration.

In this study, a polycaprolactone membrane with and without vitamin C incorporation was used to support OB-OC co-culture and intended to scavenge ROS and enhance biomineralisation in the extracellular matrix over a 14-day period. While this synthetic polymer is an established scaffold for bone tissue regeneration, the addition of vitamin C was designed to render higher protection to the regenerative cellular activity and confer improved biocompatibility property. In the current study, the PCL-Vit C membrane demonstrated a favourable substrate that scavenged ROS and facilitated a greater amount of biomineralisation than the PCL-only membrane. Co-cultures grown on the PCL-Vit C membrane supported an increase in ALP and collagen type 1 expression, highlighting higher osteoblast differentiation and high-quality extracellular matrix formation, while increased OCN expression suggested a high degree of the biomineralisation process. These favourable osteogenic activities are associated with a decrease in the RANKL to OPG ratio, indicating a shift in the regenerative balance towards bone formation. Furthermore, besides the ROS-scavenging effects, vitamin C alone may promote early differentiation of osteoblasts and enhance the synthesis of extra-cellular matrix,²² as shown by the previous work of Franceschi et al.²³ Vitamin C release from the PCL-Vit C membrane used in this study was reported in our previous work,¹³ which demonstrated similar release mechanism to the work of others.^24,25 Its timely release in this co-culture study could have modulated ROS levels and influenced the crosstalk between osteoblasts and THP-1-derived osteoclast cells. While monoculture systems provide mechanistic insights into osteoblast or osteoclast behaviour in isolation, they fail to capture the dynamic bidirectional communication essential to bone remodelling. The OB-OC co-culture model established in the current study mimics this physiological complexity by direct cell–cell interactions that are critical in modulating oxidative stress responses. Notably, we observed that vitamin C–mediated ROS scavenging in co-culture settings impacted both osteoblastic differentiation (evidenced by ALP, Col1, and OCN expression) and osteoclast activity (via reduced RANKL/OPG ratios), which would not have been as effectively interpreted in monocultures.

The differential expression of BMP-7 in the PCL-Vit C membrane and the PCL-only membrane seen in our results may be attributed to variations in the level of ROS generation in each membrane type. The observed lower BMP-7 levels in the PCL-Vit C membrane compared to the PCL-only membrane in this study may be associated with the scavenging effects of vitamin C, potentially influencing ROS levels and subsequently affecting MAPK signalling transduction, resulting in comparatively lower BMP-7 levels in the PCL-Vit C membrane group. In the PCL-only membrane, Vit C absence may have resulted in higher ROS levels and created a pro-inflammatory environment that could have driven BMP-7 expression.²⁶ Furthermore, Sanchez-de-Diego et al. showed that there is an interplay between ROS and BMP signalling, whereby BMPs could stimulate ROS production through activation of NOX expression.²⁶ These findings suggest a delicate balance between ROS-scavenging mechanisms and BMP activities in cells, which demand an optimal ROS-induced BMP signalling while maintaining manageable ROS levels. Indeed, BMPs and ROS modulate overlapping signalling pathways such as PI3K/AKT and MAPK.²⁶

Elevated levels of ROS in the absence of an exogenous antioxidant agent may serve as secondary messengers that activate MAPK signalling cascades.²⁷ This activation stimulates cellular proliferation and differentiation processes, which are necessary for osteogenesis but may also contribute to an inflammatory response. In contrast, the incorporation of Vit C into PCL membrane in this study appears to scavenge ROS and attenuate the MAPK signalling pathway to low levels. Lower ROS levels inhibit MAPK pathways, because ROS are known to serve as upstream activators of P38, ERK, and JNK signalling.²⁸ The reduction in oxidative stress by Vit C is likely contributing to the lower phosphorylation levels observed in p-P38, p-ERK, and p-JNK in the PCL-Vit C membrane in comparison to the PCL-only membrane. However, further mechanistic validation would be required. These findings support the work of Choi et al. who showed that antioxidants like vitamin C can downregulate the MAPK pathway by reducing ROS levels.²⁹ They further demonstrated that vitamin C mitigates P38 and JNK activation in osteoblasts by lowering intracellular ROS, thereby promoting a more controlled differentiation process without triggering excessive inflammatory responses.

The MAPK and Wnt/β-catenin pathways are both critical signalling pathways involved in osteoblast differentiation and function, particularly in bone remodelling. ROS influences both MAPK and Wnt signalling pathways,^28,29,30 and the crosstalk between both pathways is essential for coordinating the processes of bone formation and resorption.³¹ Our results demonstrated a much lower phosphorylation of β-catenin in the PCL-Vit C membrane than in the PCL-only membrane, which also resembled a similar lower trend with other phosphorylated MAP kinases in PCL-Vit C. The MAPK pathway can influence the stability of β-catenin by regulating glycogen synthase kinase 3β (GSK-3β), a kinase that phosphorylates β-catenin and marks it for degradation.³² Both pathways may converge on the osteogenic transcription factor RUNX2, a master regulator of osteoblast differentiation. In this study, the level of gene transcription factor RUNX2 and its protein expressions in OB-OC co-culture were comparable on both PCL-only membranes but were lower compared to the OB-OC control group. This finding may suggest that hFOB 1.19 cells had achieved osteogenic differentiation by day 14 under the present experimental condition. Although ERK signalling and p38 MAPK activation^33,34 on osteoblast can enhance RUNX2 phosphorylation, increasing its transcriptional activity, while Wnt signalling stabilises β-catenin to promote the transcription of RUNX2 target genes, the precise impact of vitamin C on these pathways remains elusive. It is plausible that vitamin C influenced these signalling pathways, leading to the observed RUNX2 expression pattern; however, mechanistic confirmation is warranted. Furthermore, low RUNX2 expression may also result from biofeedback of bone remodelling mechanisms in achieving a balanced regenerative process, since both MAPK and Wnt pathways modulate RANKL expression to promote osteoclast differentiation as well, and the balance of these signals ensures proper coupling between bone formation and resorption.

Conclusions

Taken together, the results of this study showed that OB-OC co-culture grown on the PCL-Vit C membrane was able to scavenge ROS, reduce RANKL/OPG ratio, upregulate bone formation markers, and produce a higher amount of mineralisation nodules in the ECM compared to the PCL-only membrane. The physico-chemical characteristics of the membrane, including its porosity and hydrophilicity, support this favourable outcome. Furthermore, the interplay between reduced ROS levels and modulation of MAPK signalling emphasises the intricate role of vitamin C in creating an optimal microenvironment for bone regeneration.³⁵ The role of vitamin C extends beyond its antioxidant capacity; it supports collagen synthesis for extracellular matrix formation and enhances osteoblast differentiation. These findings deepen our understanding of an already investigated area by introducing a scaffold-based ROS modulation model in co-culture, which underscores its potential to create a conducive regenerative environment through the local presence of antioxidants during biomaterial implantation. While absolute concentrations may differ from those in vivo, the relative shift in redox balance captured by our model provides meaningful insights into the oxidative mechanisms contributing to bone loss. However, many challenges and limitations need to be addressed and acknowledged, such as that there are no mechanistic interventions (e.g. inhibitors, knockdowns) or targeted analyses to confirm causal relationships. Moreover, osteoclast activities are based solely on RANKL/OPG ratios without functional assays such as TRAP quantification or resorption activity. Other challenges are quantifying the amount and type of ROS present in the biomaterial implant site and determining the appropriate dose of vitamin C needed to control oxidative stress. Furthermore, the instability of vitamin C and its hydrophilic nature have limited its practical applications.³⁶ Thus, further research to improve the delivery system is needed to overcome these limitations, possibly through better encapsulation within the biomaterial.

Conflict of interests

None declared.