Growth Factor-Free Engineered Biphasic Scaffold for Enhanced Bone Regeneration

Abstract

Large-area bone regeneration remains a significant clinical challenge, as current grafts often mineralize only at the defect edges, leaving the core underdeveloped. This study introduces a biphasic, biomimetic scaffold integrating structural support with uniform bioactivity to address this limitation. The scaffold features a highly porous outer tube for mechanical strength and cell infiltration, paired with an electrospun nanofiber core enriched with decellularized extracellular matrix (dECM) to promote cell recruitment and mineralization. Twenty-five dECMs were derived from co-cultures of bone-healing cell types: osteoblasts (OB), chondrocytes (CH), mesenchymal stromal cells (MSCs), fibroblasts (FB), and endothelial cells (EC). Among them, OB+MSC-derived dECM showed the greatest osteogenic potential. This dECM was applied to an optimized nanofiber core (232 ± 87 nm from 5 wt.% solution), with a protein content of 67.9 ± 8.3 μg/mg and DNA <50 ng/mg. The outer tube exhibited 89.6 ± 5.8% porosity and a compressive modulus of 123 ± 6.7 MPa. After BSA coating and simulated body fluid immersion, scaffolds showed calcium phosphate deposition (0.28 ± 0.03 mmol/L Ca²⁺/scaffold). In a 10 mm critical-sized femoral defect in rats, scaffolds containing both CaP and OB+MSC-derived dECM significantly enhanced bone healing. Imaging and histological analyses showed a twofold increase in bone volume, mineral density, and cortical bone formation. The compressive modulus of regenerated bone was threefold higher than untreated controls and autografts. By 12 weeks, complete defect bridging and structural recovery were achieved. This biphasic scaffold design presents a promising strategy for large bone defect repair by enabling uniform tissue regeneration, combining osteoinductive cues with structural performance suited for clinical translation.

Introduction

Bone fractures, especially those causing critical-sized defects, represent a significant clinical challenge. Affecting a notable portion of both injured combat personnel (approximately 20%, costing the Department of Defense nearly $100 million annually) and placing a substantial multi-billion dollar burden on the civilian healthcare system through prolonged treatments and rehabilitation [1]. Bone grafts and graft substitutes are essential for reconstructing large-area and critical-sized bone defects. Although biological grafts like autografts and allografts [2–4] are widely used clinically, they have significant limitations, including donor site morbidity, limited availability, poor remodeling, and risk of disease transmission, leading to inconsistent bone healing outcomes. and variable success rates [2–7].

Bone graft substitutes provide an alternative for large-scale bone regeneration. They are engineered to enhance cell-material interactions and tissue ingrowth, while simultaneously addressing limitations in nutrient and oxygen supply, metabolic waste disposal, and biocompatibility [8]. Bone graft substitutes often lack bioactivity and are enhanced with biological, chemical, and mineral bioactive signals to accelerate bone repair [9]. These bioactive cues within engineered grafts enhance cell adhesion, proliferation, and osteogenic differentiation, while the porous architecture supports vascularization and tissue ingrowth, collectively promoting efficient healing [10]. Bone healing outcomes significantly improved when scaffolds were loaded with clinically relevant mesenchymal stem cells (MSCs) and bone morphogenetic protein-2 (BMP-2) [11], enhancing collagen synthesis, promoting osteogenic differentiation, and supporting new bone formation [12]. Thus, the synergy between biomaterials and bioactive molecules creates an optimal environment for effective and sustainable bone regeneration.

A wide variety of scaffolds derived from polymers, ceramics, metals and their composites as well as hydrogels and nanofiber matrices have been explored for critical-sized bone defect repair, each with unique advantages and limitations [13–16]. Hydrogels offer excellent biocompatibility and cell encapsulation but lack mechanical strength [13]. Ceramics like hydroxyapatite (HA) and β-tricalcium (β-TCP) show strong osteoconductivity but are brittle and resist to remodel [14, 15]. Metals, especially titanium and magnesium alloys, provide mechanical strength yet suffer from low ductility and stress shielding [16]. Polyesters and modified natural polymers allow creation of intricate 3D structures but often require functionalization to enhance bioactivity [17]. Despite advances the quality of the regenerated bone is suboptimal and fail to achieve complete mineralization [18, 19] and osteointegration with the host tissue [20]. Critical scaffold properties include mechanical strength, interconnected pores and pore volume, and degradation rate directly influence cell recruitment and tissue development [21].

Many scaffold-based bone tissue engineering approaches achieve gap bridging primarily at the scaffold periphery, while the core often becomes infiltrated with soft tissue that fails to mineralize [22, 23]. This is largely due to limited osteoinductivity, poor vascularization, and inadequate transport of metabolic waste, oxygen, and nutrients [23–25]. Enhancing osteoconduction and osteoinduction requires the deliberate integration of structural and biological features that promote osteogenic cell recruitment and extracellular matrix mineralization [26]. Protein–mineral interactions, particularly between collagen and HA, are essential for bone formation. To address these challenges, we previously developed spiral-structured scaffolds with nanofiber decoration and HA coatings to enhance bone healing [23–25]. However, HA alone often lacks sufficient osteoinductive capacity and is typically combined with osteogenic and angiogenic growth factors to support vascularized regeneration of large bone defects [27–30]. These strategies, however, have shown varied success due to the short biological half-life of growth factors, suboptimal release profiles, and dosing challenges [31]. Alternatively, cell-laden ECM constructs, which present a complex combination of ECM components, growth factors, and vesicles, offer strong osteoinductive potential but often lack the mechanical integrity necessary for effective bone repair [32].

Polyesters such as poly(lactic acid) (PLA), poly(glycolic acid) (PGA), their copolymer PLGA, and polycaprolactone (PCL) are commonly used in bone scaffold fabrication due to their tunable mechanical properties, degradation behavior, and ease of processing [33, 34]. These polymers are often combined with HA or β-TCP to replicate the composite nature of bone, buffer acidic degradation products, and stimulate bone formation [22, 24, 33]. Simulated body fluid (SBF) is a well-established method for depositing calcium phosphate (CaP) coatings on scaffolds. These Ca²⁺-deficient layers are bioresorbable, gradually releasing calcium ions and promoting osteogenic differentiation [24]. Moreover, the uniform calcium phosphate deposition achieved via SBF on both the implant surface and within scaffold pores enhances osteogenesis throughout the construct. Nonetheless, regenerating large bone defects remains a significant challenge despite these advances [22–24, 33]. The accumulation of acidic degradation byproducts and limited bioactivity continue to impede deep tissue infiltration and full scaffold mineralization [22–24, 33]. To address these challenges, our ongoing work focuses on engineering polyester-based scaffolds with advanced architectures, incorporating nanostructures such as nanofibers and calcium phosphate (CaP) coatings [22–24, 33]. These enhancements have demonstrated marked improvements over conventional polymeric scaffolds. However, despite these advances, bone regeneration in preclinical models remains suboptimal, often failing to achieve complete mineralized matrix formation and full defect bridging—resulting in regenerated bone that is mechanically and histologically inferior[24, 27–30, 35].

Alternative osteoinductive strategies, such as small molecule drugs, peptides, and extracellular matrix (ECM) proteins, offer distinct advantages and limitations. Successful healing of critical-sized femoral defects requires mimicking the biological complexity of the native fracture callus, which involves coordinated interactions among ECM components and diverse cell types. To replicate this environment, we utilized human and rat-derived osteoblasts (OB), mesenchymal stem cells (MSCs), chondrocytes (CH), endothelial cells (EC), and fibroblasts (FB) each representing key contributors to callus formation. Chondrocytes and osteoblasts drive cartilage formation and matrix mineralization during endochondral ossification [36], while MSCs serve as progenitors for both chondrogenic and osteogenic lineages. Fibroblasts secrete structural ECM proteins that support the connective tissue matrix, and endothelial cells promote neovascularization essential for sustained tissue regeneration. By generating ECM from these callus-relevant cells, our approach aims to develop a bioactive scaffold that recapitulates the spatial and temporal cues of native bone repair. Following decellularization (dECM) to remove genetic material, this engineered ECM may serve as an effective osteogenic platform for enhancing regeneration in critical-sized femoral defects[37].

This manuscript presents the development and evaluation of a biphasic scaffold designed to promote uniform bone regeneration in critical-sized femoral defects in rats. The scaffold integrates calcium phosphate (CaP) and decellularized extracellular matrix (dECM) as osteoinductive components. We systematically optimized nanofiber diameter to enhance proliferation of native callus cells, using human osteoblasts to identify the most supportive matrix. The optimal fiber diameter supporting maximal proliferation of human fibroblasts and chondrocytes was then used to culture rat osteoblasts, chondrocytes, and mesenchymal stem cells (MSCs), both individually and in co-culture, to maximize ECM production. These rat-derived co-cultures were used to generate 25 distinct dECMs from key bone-healing cell types: osteoblasts, chondrocytes, MSCs, fibroblasts, and endothelial cells. The dECM with the highest osteogenic potential in vitro was incorporated into the scaffold core. Structurally, the biphasic scaffold consists of a microfabricated porous polycaprolactone (PCL) outer tube for mechanical support and permeability, and an inner core of electrospun PCL nanofibers enriched with the selected dECM. The scaffold surface was further functionalized by bovine serum albumin coating and mineralized in simulated body fluid to deposit a CaP layer (Fig. 1). The biphasic scaffold offers several key advantages over conventional bone graft substitutes. The porous PCL outer tube provides mechanical stability and nutrient permeability, while the dECM-enriched nanofiber core creates a biologically instructive environment that promotes osteoinduction and uniform tissue infiltration. The CaP coating further enhances osteoconductivity and stimulates matrix mineralization. Together, these features enable predictable, spatially uniform bone regeneration, addressing limitations of traditional polymeric, ceramic, or growth factor–based scaffolds and supporting clinical translation for critical-sized defects. These osteoinductive cues, CaP and dECM, work together with the scaffold structure to encourage bone formation and ensure uniform cell and tissue infiltration. We hypothesize that the outer layered, porous tube provides mechanical stability and permeability, while the dECM-enriched core acts as a chemoattractant to enhance cell recruitment. Combined with CaP-driven osteoinduction, this design is expected to facilitate uniform tissue regeneration, improved mineralization, and superior mechanical and histological healing compared to control scaffolds. This approach presents a promising strategy to enhance the performance of FDA-approved polyester-based bone grafts for large-area bone regeneration and holds potential as an alternative to conventional bone grafts.

Fig. 1. Biphasic scaffold design featuring an outer porous shell and a core filled with dECM-functionalized nanofibers.

(A & B) Core filled with 3 rolled layers of dECM nanofibers, maintaining adequate pore structure and spacing. (C & D) Core filled with 6 layers results in dense packing, compromising pore integrity and space required for tissue infiltration. Blue coloration was applied to the dECM core for visualization. Scale bar =1mm.

Materials and Methods

Unless otherwise indicated, all chemicals and polymers used for preparation of the scaffold and cell sheet decellularization were purchased from Sigma-Aldrich (St. Louis, MO, USA). Dichloromethane (DCM) was purchased from VWR Chemicals (Radnor, PA, USA). The human mesenchymal stem cell (hMSCs; Cat# 492-05f), the human chondrocytes (hCH; Cat#: 402-05a), the primary rat osteoblasts (OB; Cat# R406-05a), the rat dermal fibroblasts (FB; Cat# R106-05n), and the primary rat endothelial cells (EC) were purchased from Cell Applications (San Diego, CA). The primary rat articular cartilage chondrocytes (CH; cat# 10930) were purchased from Innoprot (Spain). The human fetal osteoblasts (hOB), the human fibroblasts (hFB) were purchased from ATCC. The primary rat mesenchymal stem cells (MSCs) were isolated from rats. Culture reagents, including fetal bovine serum (FBS), penicillin-streptomycin, and trypsin, were purchased from Life Technologies. Culture media were purchased from the following sources: Rat Mesenchymal Basel Medium and its growth supplement (Cell Application INC), supplemented EC medium without phenol red (Cell Applications). Eagle’s Minimum Essential Medium (ATCC), Dulbecco’s Modified Eagle Medium (Gibco), Dulbecco’s Modified Eagle Medium/Nutrient Mixture F-12 (Gibco). Analytical reagents were purchased from the following sources: CellTiter-Blue Cell Viability Assay (Promega), Quant-iT^™ PicoGreen^® dsDNA (Invitrogen), and BCA assay kit (Thermo Fisher Scientific). Sprague–Dawley rats are from Charles River Laboratories (Wilmington, MA).

Fabrication of Nanofiber Matrices

Polycaprolactone (PCL, Mn = 80,000, Sigma–Aldrich) pellets were dissolved in 1,1,1,3,3,3-hexafluoro-2-propanol (HFIP, Oakwood Products) to create solutions with different concentrations: 5%, 8%, 10% and 12% (w/v). The solution was loaded into a 1 mL syringe with a 23-gauge blunt tip needle. A 15 kV (voltage) was applied via a high-voltage power supply as the polymeric solution was pumped at 0.6 mL/h. The nanofibers were collected on 18 mm square glass coverslips mounted conductive collector covered in aluminum foil. The deposition on each coverslip was maintained for 7 minutes to ensure uniform coating[38]. Random nanofibers were collected on the aluminum rotating target at 20 rpm [39] and kept desiccated until further use. Nanofiber matrix thickness and fiber diameter influence scaffold properties in terms of surface area and porosity.

Cell Seeding

Each scaffold was seeded with human or rat 25,000 cells with the supplier recommended media. Human osteoblast cells were cultured in DMEM/F12 medium supplemented with 10% FBS and 1% P/S. Fibroblast cells were cultured in an EMEM medium supplemented with 15% FBS and 1% P/S. Rat MSC cells were cultured in Rat Mesenchymal Basel Medium with a growth supplement. Chondrocyte and endothelial cells were cultured in DMEM medium supplemented with 10% FBS and 1% P/S. Human cells were used initially to optimize scaffold fabrication parameters, while rat cells were employed for subsequent coculture and functional assays to align with in vivo studies. Importantly, our previous work demonstrated that rat MSCs and hMSCs exhibit comparable proliferation patterns, supporting the translational validity of using human cells for this parameter [40]. This design ensured experimental continuity and improved translational relevance across both in vitro and in vivo phases. The basal media was switched to the osteogenic media following 48 h. The osteogenic media consisted of the basal media supplemented with 50 μg/mL ascorbic acid and 8 mM β-glycerol phosphate and replaced every 2 days for a period of 21 days. Cell proliferation, alkaline phosphatase (ALP) activity, and matrix mineralization were evaluated using standard protocols. Proliferation of cells cultured on various scaffolds was assessed using the CellTiter-Blue^® Cell Viability Assay (Promega, Madison, WI, USA) according to the manufacturer’s instructions[41]. ALP activity, a marker of early osteogenic differentiation, was measured using a colorimetric ALP assay kit (Colorimetric; Abcam, ab83369) [41]. For mineralization analysis, Alizarin Red S staining was performed to visualize calcium deposition. After staining, cetylpyridinium chloride (CPC, 10%, w/v) was used to solubilize the bound Alizarin Red dye, producing a purple solution. The absorbance of this solution was then quantified at 562 nm to determine relative calcium content, as previously described [41].

Decellularization of Cell Laden ECM on Nanofiber Matrices (dECM)

PCL scaffolds laden with fibroblasts, osteoblasts, chondrocytes, and their coculture (OB/MSCs, OB/CH, MSCs/CH, FB/CH, FB/MSCs FB/OB, EC/CH, EC/MSCs, EC/OB and EC/FB) were decellularized according to a previously published protocol [41, 42]. Briefly, scaffolds were washed thrice with PBS and placed in a 10 mM phosphate-buffered 50 mM sodium solution containing 125 mM 3-(decyl-dimethyammonio) propane sulfonate (SB3-10) overnight at 4°C under gentle agitation. After another three PBS washes, scaffolds were placed in 10 mM phosphate-buffered 50 mM sodium solution containing 0.14% sodium deoxycholate and 0.6 mM 3-(N, N-Dimethyl-palmitylammonio) propane sulfonate (SB3-16) overnight at 4°C under gentle agitation. This process was repeated twice to ensure the removal of cell debris. Finally, decellularized scaffolds were stored in PBS containing protease inhibitor cocktail at 4°C.

Evaluation of Decellularization Based on DNA and Protein Content

DNA and protein contents of dECM from PCL scaffolds laden with fibroblasts, osteoblasts, chondrocytes, and their coculture were quantified using the Quant-iT PicoGreen dsDNA assay [33] and BCA protein assay [43], respectively. For DNA extraction, samples were soaked in TE buffer, followed by thermal shock and treatment with 20% SDS/Triton X-100 (4:6) for purification. Protein was extracted by incubating samples in RIPA buffer with protease and phosphatase inhibitors at 4 °C for 30 min, then centrifuged at 12,000 rpm for 20 min. DNA and protein contents in the supernatants were measured using a BioTek Synergy HT plate reader, following manufacturers’ protocols with standard curves generated for each assay.

Biphasic Scaffold Assembly

For the animal study, layered hard-shell scaffolds were filled with nanofibers embedded with hybrid dECM to form the biphasic scaffold core, as shown in Fig. 1. These scaffolds were dimensioned to match the rat femur defect site, with a height of 5.0 mm and an inner channel diameter of 0.82 mm. Fig. 1A & B illustrate a core filled with 3 layers of rolled dECM, while Fig. 1C & D show a 6-layer configuration. These variations were evaluated to determine optimal core packing without compromising pore architecture, which is critical for tissue infiltration and nutrient diffusion. The 3-layer configuration preserved pore integrity and was therefore selected for in vivo implantation.

Composite Scaffold Assembly and Incorporation of Calcium Phosphate (Cap) Biominerals

Composite scaffolds were formed by inserting rolled dECM-laden nanofiber matrices into a tubular, layered, and highly porous outer shell. Scaffolds with or without dECM were soaked overnight in 30% (v/w) BSA. Simulated body fluid (SBF) was prepared as previously described [7], and scaffolds were incubated in SBF at 37 °C to induce calcium phosphate (CaP) mineralization. The solution was refreshed daily for up to 7 days. After incubation, scaffolds were rinsed in deionized water overnight to remove residual ions and vacuum-dried at room temperature. Calcium content was quantified using an alizarin red-based colorimetric assay, following our published protocol [24]. Briefly, scaffolds were stained in 1% alizarin red for 1 min, rinsed, and the bound dye was solubilized in 10% cetylpyridinium chloride. A 150 μL aliquot was transferred to a 96-well plate, and absorbance was measured at 562 nm. Samples were tested in triplicate (n = 5), and calcium levels were calculated using a standard CaCl₂ curve.

Characterization of Mechanical and Pore Properties

To ensure reproducibility of scaffold mechanical strength and pore characteristics, we conducted evaluations following protocols established in our previously published work [43]. Briefly, tubular samples (5 mm diameter × 10 mm height) were compressed to failure at a rate of 15 mm/min using an Instron Model 5965 Materials Testing System (n=6). The compressive modulus was determined from the linear region of the resulting stress–strain curves [43]. Similarly, porosity was evaluated using the water intrusion method as described in a previous study. Samples (n=3) were submerged in distilled de-aired water and placed under vacuum for 24 h. After removing surface water, the change in mass was measured using an analytical balance to determine the volume of water retained in the pores. (W_after – W_before)/ρ_water/V * 100%. where W_after is the water-saturated mass, W_before is the dry mass, and V is the sample volume measured with calipers [43].

Osteogenesis on dECM Nanofiber Matrix

Five hybrid dECM groups such as OB/FB, OB/MSCs, OB/CH, FB/MSCs, and EC/MSCs were selected to evaluate their potential to promote osteogenesis. These matrices were seeded with MSCs and cultured as described previously. Constructs were maintained in basal medium (10% FBS, 1% P/S) for the first 7 days. Starting on day 7, the medium was switched to osteogenic differentiation medium supplemented with 100 μg/mL ascorbic acid and 5 mM β-glycerophosphate. The medium was refreshed every 2 days, and cultures were continued for a total of 28 days. [33, 44].

Alkaline Phosphatase (ALP) Assay

ALP activity was measured on days 7, 14, 21 and 28 using a commercial assay kit (n = 3). Scaffolds were washed with ice-cold PBS. Scaffolds were then transferred to microcentrifuge tubes, and cell lysis was performed using an ALP lysis buffer (0.2% Triton X-100 in Tris-based buffer). Samples were incubated on ice for 10–15 minutes and pipetted up and down to ensure thorough lysis. Supernatants (80 μL) were mixed with 50 μL of 5 mM pNPP and 10 μL of ALP enzyme in a 96-well plate, incubated at 25 °C for 60 min, and stopped with 20 μL of stop solution. Absorbance was measured at 405 nm. ALP activity (μmol/min/mL) was calculated using the formula: (B/TV)D, where B = pNPP amount from standard curve, T = reaction time, V = sample volume, D = dilution factor [33, 43].

Mineralized Matrix Deposition

Alizarin Red staining was performed on OB/FB, OB/MSCs, OB/CH, FB/MSCs and EC/MSCs groups on days 7, 14, 21 and 28 to assess calcium deposition. Scaffolds with different dECM were soaked overnight in 30% (v/w) bovine serum albumin (BSA) and washed with deionized (DI) water three times to remove extra BSA. Simulated body fluid (SBF) was prepared according to Kokubo recipe [3]. Scaffolds were then incubated in a solution of SBF and maintained at 37°C for mineral deposition. SBF was renewed every 24 hours. After being incubated for predetermined periods (7, 14, 21, and 28 days), samples were removed from the solution and immersed in DI water overnight to remove soluble inorganic ions. Samples were then stained with Alizarin Red S staining solution to check the calcium deposition, followed by images documenting. The quantification of calcium deposition was done by cetylpyridinium chloride (CPC) extraction. Briefly, scaffolds with staining were incubated in the 10% CPC solution for 30 mins with agitation to get calcium deposition fully extracted. The extraction was then read with BioTek Synergy HT plate reader at the A562 together with several gradient standards solutions [24, 33]. Alizarin red stained scaffolds were examined by Nikon SMZ1500 stereomicroscope with DS-Fi1 digital camera.

Immunostaining for Osteogenic Proteins

Two differentiation markers of MSCs, osteopontin (OPN) and type I collagen (COL-1), were used to examine the process of osteogenesis. Briefly, after 7,14, 21, and 28 days of incubation, the scaffolds were washed with cold PBS three times, fixed in 1 mL 4% formaldehyde solution for 10 mins, and permeabilized in 1 mL 0.5% Triton X-100 for 10 mins at room temperature. The scaffolds were then blocked in 1 mL solution containing 3% BSA for 30 mins. Afterward, samples were incubated with the OPN (Abcam, ab8448, from rabbit) or COL I(Abcam, ab292, from rabbit) antibodies for 1h at room temperature and subsequently visualized with Fluorescein (FITC)-conjugated AffiniPure goat anti-rabbit IgG(H+L). Next, the scaffolds were rinsed with PBS, stained for 10 mins with 4’,6-diamidino-2-phenylindole (DAPI) to label the nuclei. Finally, the scaffolds were visualized under a fluorescence microscope[43]. Imaging was performed on a Zeiss 780 laser scanning confocal microscope.

Scanning Electron Microscopy

The morphology of prepared microspheres was examined with SEM (Zeiss Auriga FIB-SEM with Leica VCT-100 cryo-system, Oberkochen, Germany). Briefly, samples with dECM were dried with increasing ethanol concentration (50, 70, 80, 90, 95, and 100%); other samples were air-dried. All samples were attached to the SEM stages with double-sided adhesive tape, then sputter-coated with gold for 1 min using a Denton Desk-1 Sputter coater. Surfaces were visualized at an accelerating voltage of 3kv using an AMray 1830-D4 equipped tungsten gun. At least five randomly selected SEM images were acquired for each type of scaffold. Images were manually measured and analyzed using ImageJ software v1.5 (NIH, Maryland, US) to calculate the average microsphere diameter and pore size[24].

Animal Care and Experimental Procedure

All animal procedures were conducted in accordance with protocols approved by the Institutional Animal Care and Use Committee (IACUC) at the University of Nebraska Medical Center (Approval No: IACUC # 24-065-01-FC) and at the University of Connecticut Health Center (Approval No: IACUC # AP-201005-0826). Sixty-three male Sprague–Dawley rats (Charles River Laboratories, Wilmington, MA), aged 10 weeks and weighing 450–500 g, were housed separately in a 12 h light/12 h dark, temperature-controlled environment (21°C) with proper airflow and were given food and water ad libitum. Rats were acclimated for at least 7 days before undergoing any procedures. Animals were monitored daily for 3 days during the post-op period. Body weight was recorded twice per week during the first postoperative week and subsequently once per week.

Animal Groups

To evaluate the individual and combined roles of biomineralized calcium phosphate (CaP) and decellularized hybrid extracellular matrix (dECM) in bone regeneration, we established seven experimental groups (Table 2). These groups leverage our primary biphasic scaffold (BS), which consists of a hard shell (HS) and a nanofiber (NF) core. For clarity, groups were organized into Test Combinations, Scaffold Component Controls, and Biological Controls. Test Combinations: These groups evaluate the performance of our BS with various bioactive additions. G1: BS+CaP + dECM, G2: BS+CaP and G4: BS+dECM. Scaffold Component Controls: These groups isolate the effects of specific scaffold components. G3: HS+CaP (W/O NF) and G5: BS (BS base control). Biological Controls: G6: Autograft (Positive Control) and G7: Untreated Defect (Negative Control). This design enabled evaluation of the individual and combined contributions of CaP and hybrid dECM to in vivo osteogenesis and scaffold-mediated bone regeneration.

Table 2.

Experimental groups for bone regeneration study

Group	Category	Composition	Description
G1	Test Combination	BS+ CaP+dECM	Biphasic scaffold with both biomineralized CaP and dECM-loaded nanofiber
G2	Test Combination	BS+CaP	Biphasic scaffold with biomineralized CaP only
G3	Scaffold Component Control	HS+CaP (W/O NF)	Hard shell with CaP, without nanofiber
G4	Test Combination	BS+dECM	Biphasic scaffold with dECM-loaded nanofiber only
G5	Scaffold Component Control	BS (base control)	Biphasic scaffold with no added CaP or dECM
G6	Biological Control	Autograft	Positive control: bone autograft
G7	Biological Control	Untreated Defect	Negative control: untreated bone defect

Critical-sized Femoral Defect in Rats

All surgical procedures were conducted under aseptic conditions and based on a previous study, with some modifications [45]. Briefly, general anesthesia was induced in a chamber using a 4–5% isoflurane/O₂ gas mixture and maintained at 1.5–2.5% isoflurane/O₂ via a nose cone with controlled ventilation. Pre-operatively, each rat received a subcutaneous injection of buprenorphine hydrochloride extended release (1-1.2 mg/kg) for analgesia. The animals were positioned in the left lateral decubitus position. The right leg was shaved and disinfected 3 alternating preps of 70% EtOH/chlorhexidine and/or betadine (Braunol, B. Braun Melsungen AG, Melsungen, Germany). A lateral approach was used to expose the femur, with a longitudinal incision made along the femur to reveal the fascia lata. A lateral approach to the femur between the biceps femoris and vastus lateralis was used for surgical exposure of the femur [33]. The periosteum of the femur was incised along the length of the femur and the soft tissue will be freed circumferentially. The k-wires were inserted to make the holes. Then k-wires were removed, and 1.1 mm shaft mini half-pins were inserted and fixed using a custom-made external fixator (IMEX veterinary Inc., Longview, Texas, USA). A mid-diaphyseal defect involving the entire circumference of cortical bone and the intramedullary contents measuring 10 mm in length was created using a high-speed burr by cutting through the thickness of the bone at two locations, allowing for the removal of the segment of the long bone. The existing marrow at the site was removed by irrigation. Cold Saline irrigation was used to minimize thermal damage to the bone at the osteotomy site. The defect was filled with a carrier that will be implanted at the defect site. Bone segments were fixed with customized external bone fixators. Wound was irrigated with sterile saline, and the incision was closed in two layers (muscular and subcutaneous) with interrupted 5–0 sutures (Vicryl^®, Ethicon^™, Johnson & Johnson^™, Somerville, NJ). The bioglue (Dermabond^®, Ethicon^™, Johnson & Johnson^™) was subsequently applied to the closed skin incision. Post-operatively, the rats were allowed to return to full weight-bearing immediately.

Tissue Morphological Assessment

Following euthanasia at 12 weeks post-surgery, femurs were harvested and visually examined for gross morphological changes. The surrounding bone tissue was inspected for signs of inflammation, edema, or necrosis. Particular attention was given to the defect area to identify any evidence of bone remodeling or new bone formation. The integrity of the cortical bone adjacent to the defect site was also assessed to evaluate structural continuity and potential bridging.

Compression and Mechanical Characterizations

Compression mechanical properties of the femur were measured using material testing system machine (MTS 858 Mini-Bionix, Eden Prarie, MN) and MTS TestSuite^™ Multipurpose Testing Software[33]. Samples with size of 6 mm×12 mm (diameter × length) were compressed at a compression rate of 0.1 mm/min (n=6). For each sample, stress-strain curves were recorded, and compressive modulus was calculated. Compression testing was selected as the primary mechanical assessment, as load transmission along the femoral axis during stance is predominantly compressive. Although tensile and bending tests can provide complementary information, the limited bone volume available in the rat defect model precluded performing multiple destructive tests on the same specimen.

Radiography

All rats underwent two radiographs (anteroposterior and lateral of the femur) (Kubtec, Stratford, CT, USA) under anesthesia to check the stability of the osteosynthesis (absence of femoral displacement, femoral fracture around a screw, plate breakage, screw loosening) at 12-weeks post-surgery. Radiographs were assessed based on periosteal reaction, union quality, and bone remodeling with a maximum possible score of 8 for bone fracture repair (Table 3). Briefly, the scoring evaluates three parameters periosteal reaction, bone union, and remodeling using a 0–3 scale, where higher scores indicate more complete healing. Periosteal reaction is scored as complete across the defect (3), moderate >50% (2), mild <50% (1), or none (0). Bone union is assessed as complete bony union (3), moderate >50% (2), mild <50% (1), or none (0). Remodeling is scored as complete remodeled cortex (3), mild <50% (1), or none (0). These scores were assigned independently by two blinded reviewers and averaged for each sample to quantify radiographic healing [46].

Table 3.

Radiographic scoring system for fracture healing

Categories	Scores
	3	2	1	0
Periosteal reaction	Complete across defect	Moderate (>50%)	Mild (<50%)	None
Bone union	Complete bony union	Moderate (>50%)	Mild (<50%)	None
Remodeling		Complete remodeled cortex	Mild (<50%)	None

Micro-CT

Micro computed tomography (μCT) (μCT50, Scano medical CT-50, Brüttisellen, Switzerland) was used to obtain three-dimensional visualization of the healing process at 12 weeks post-surgery. Prior to imaging, the intramedullary K-wires were carefully removed to prevent beam-hardening artifacts that could interfere with quantitative radiographic analysis. Samples were stored with PBS till analysis. Each femur was precisely positioned at the center of the gantry in a desktop μCT machine for scanning. The femoral samples were scanned with a voxel size of 8 μm, 55 kV and intensity of 145 μA and the femur samples with a voxel size of 34.4 μm, 55 kV and intensity of 145 μA. A 3D reconstruction of the obtained images was carried out using 3D software creator. Bone volume (BV), mineral density (BMD) and bone volume fraction (BV/TV) were calculated.

Histological Examination

Euthanasia was carried out via carbon dioxide overdose. After euthanasia, the right femur was harvested, washed with PBS, and fixed with 10% neutral buffered formalin at 4 °C for two days[47]. Samples (n=3) were decalcified in EDTA (changed every other day for three weeks), washed in PBS (pH 7.4), and dehydrated in graded ethanol before embedding in paraffin wax and being cut into 7 μm sections. A graded series of ethanol concentrations was used to rehydrate sections. Paraffin blocks were stained with haematoxylin-eosin (H&E), Masson-Goldner trichrome (MT) and Safranin O staining. Another set of histological sections were subjected to the immunohistochemistry for protein deposition of bone sialoprotein (BSP) and Collagen I (COL I). The slides were imaged using Aperio CS2 high-resolution digital scanner (Leica Biosystems Inc., Buffalo Grove, IL).

Statistical Analysis

For all experiments, the sample size was determined based on a power analysis to ensure statistical robustness. In vitro scaffold characterization and bioactivity assessments were designed according to our previous studies to achieve a 95% confidence level. Sample sizes of n = 6 for mechanical testing, n = 5 for porosity, and n = 6 for cell culture experiments were deemed sufficient for statistical reliability. In vivo study group sizes were similarly determined based on prior work and power analysis using one-way ANOVA with a type I error (α) of 0.05, an expected variance of 10–15%, and a power of 0.80, reflecting an 80% probability of detecting a true effect. Experimental results are presented as mean ± standard deviation (SD). Statistical analyses were performed using a two-tailed Student’s t-test for comparisons between two groups and one-way ANOVA with Tukey’s post hoc test or Kruskal-Wallis test for comparisons among multiple groups. When evaluating the effects of two independent variables, two-way ANOVA was applied, followed by Tukey’s post hoc test where appropriate. All analyses were conducted with GraphPad Prism 8.0.2 (GraphPad Software, San Diego, CA, USA), with a significance threshold set at p < 0.05.

Results

Composite Scaffold Characterization

The biphasic scaffold integrates dECM-functionalized nanofibers within a layered hard-shell structure (height: 5.0 mm; inner channel diameter: 0.82 mm) designed for in vivo bone repair in the rat femur[43]. The scaffold core was filled with a rolled dECM fiber matrix using either 3 layers (Fig. 1A & B) or 6 layers (Fig. 1C & D). Incorporation of 6 layers compromised pore space critical for tissue infiltration (colored core), whereas the 3-layer configuration maintained adequate spacing and was therefore selected for all subsequent characterizations. The assembled composite scaffold was then coated with bovine serum albumin and immersed in SBF to induce biomineralized CaP deposition. The final construct was characterized to assess mechanical strength, porosity, and the presence of CaP and dECM, ensuring consistency with previous findings and process reproducibility. The scaffold’s architecture, porosity, mechanical strength, and incorporation of bioactive components such as CaP and dECM were specifically engineered to enhance its osteoconductive and osteoinductive properties, thereby promoting homogeneous tissue distribution and effective osteointegration at the bone defect site. The outer tube of the composite scaffold, fabricated from highly porous PCL sheets with grid patterns and an open-lumen design, demonstrated significantly higher porosity 89.6 ± 5.8% compared to nonporous no-grid scaffolds, which exhibited the lowest porosity 30.4 ± 3.6% [43]. The compressive modulus of the scaffolds reflected this structural difference: nonporous no-grid scaffolds achieved 152.8 ± 7.3 MPa, while the porous grid scaffolds showed a slightly lower modulus of 123 ± 6.7 MPa, consistent with the expected decrease due to increased porosity and in agreement with our previous findings [43].The scaffolds contained 0.28 ± 0.03 mmol/L Ca²⁺ per scaffold, as quantified using an alizarin red-based colorimetric assay [24]. The developed biphasic composite scaffold’s structure, components, dimensions, physicochemical, mechanical and biological properties have been detailed in Table 1.

Table 1.

Summary of biphasic scaffold components and properties

Feature	Component	Description
Structure	Outer Tube	Ten-layered porous PCL sheets with 10 μm grid patterns, bonded using dichloromethane (DCM); tubular design (OD: 5 mm, ID: 3 mm, wall: 1 mm); open-lumen for tissue ingrowth.
	Inner Core	Rolled electrospun nanofiber sheet made of PCL:CA (80:20 wt%) with micro-nanostructured architecture.
	Overall Scaffold	Biphasic composite (outer + inner layers); coated with bovine serum albumin and biomineralized using simulated body fluid (SBF) for CaP deposition.
Porosity	Outer Tube	Highly porous structure with 89.6 ± 5.8% porosity; grid-enhanced pattern for permeability and mechanical support.
Mechanical Strength	Outer Scaffold	Compressive modulus: 123 ± 6.7 MPa (porous grid scaffold).
CaP Content	Overall Scaffold	0.28 ± 0.03 mmol/L Ca2+ per scaffold (measured via Alizarin Red assay); confirms effective mineralization.
dECM Content	Inner Core	Derived from OB/MSCs co-cultures; low residual DNA (<50 ng/mg) and high protein content (67.86 ± 8.32 μg/mg) post-decellularization.

PCL: Polycaprolactone, CA: Cellulose acetate, CaP: Calcium phosphate, dECM: Decellularized extra cellular matrix, OB: Rat osteoblasts, MSCs: Rat mesenchymal stem cells

Electrospun nanofibers fabricated with increasing PCL concentrations (5%, 8%, 10%, and 12%) showed a corresponding increase in fiber diameter (Fig. 2A). A clear inverse relationship was observed between polymer concentration and fiber thickness: lower concentrations produced thinner fibers when other electrospinning parameters such as flow rate, voltage, and working distance were held constant[40]. Among these parameters, polymer concentration had the most pronounced influence on fiber diameter. This trend was further supported by diameter distribution profiles and average diameter analysis (Fig. 2B and 2C). The mean fiber diameters were 232.10 ± 86.98 nm (5% PCL), 502.10 ± 246.70 nm (8%), 661.20 ± 179.60 nm (10%), and 1060.02 ± 117.50 nm (12%).

Fig. 2. Influence of PCL fiber morphology and cell co-culture on cell proliferation.

Fiber morphology and diameter produced by varying PCL concentrations. (A) SEM images show a progressive increase in fiber diameter with increasing PCL concentration. (B) Corresponding fiber diameter distribution histograms for each concentration. (C) Quantitative analysis of average fiber diameters confirms increased fiber thickness at higher PCL concentrations; n=3 (100 measurements). Nanofibers (5% PCL–Thin) supported cell growth better than microfibers (12% PCL–Thick). Cell viability was assessed using the CellTiter-Blue fluorescence assay, and relative fluorescence units (RFU) were used to measure cell growth for (D) human fetal osteoblasts (hOB), (E) chondrocytes (hCH), and fibroblasts (hFB); n = 3. Co-cultures on nanofibers significantly enhanced cell proliferation, indicating synergistic cell–cell interactions. (F) RFU measured by CellTiter-Blue assay showed increased growth over 7 days for co-cultures of rat osteoblasts (OB), chondrocytes (CH), and mesenchymal stromal cells (MSCs) compared to individually cultured cells. (G) MSCs/OB co-cultures exhibited significantly greater proliferation than OB or MSCs alone, sustained through 14 days; n=3. Data presented as Mean ± SD, using one-way ANOVA with Tukey’s post hoc test. ****p < 0.0001 vs 5% PCL; ****p < 0.0001 vs OB and MSCs.

Based on the CellTiter-Blue assay and RFU measurements (Fig. 2D), human fetal osteoblasts (hOB) exhibited higher proliferation on thin nanofibers (5% PCL) compared to thicker fibers (8% and 12% PCL), with intermediate growth observed on 8% and 10% PCL fibers. Overall, thinner fibers supported greater cell growth, likely due to increased porosity and surface area[48] . To confirm this trend, human chondrocytes (hCH) and normal human dermal fibroblasts (hFB) were cultured on thin (5%) and thick (12%) fibers only. Similar to hOB, both hCH and hFB showed greater proliferation on thinner fibers (Fig. 2E). Notably, the RFU values for OB were lower than those for CH and FB, reflecting inherent differences in cell type. These results collectively demonstrate that smaller fiber diameters enhance cell proliferation across multiple cell types. As a result, nanofibers produced using 5% PCL were selected for all subsequent experiments.

To enhance cell-laden ECM production, key bone-healing cell types including rat osteoblasts (OB), chondrocytes (CH), and mesenchymal stromal cells (MSCs) were seeded on PCL nanofibers either individually or in co-culture. Each scaffold received 25,000 cells for single-cell cultures and co-cultures (12,500 of each type at a 1:1 ratio). Cell proliferation was assessed by RFU using the CellTiter-Blue assay over 14 days (Fig. 2F and 2G). Initial proliferation trends through day 7 (Fig. 2F) revealed cell- and co-culture–dependent responses. Osteoblasts showed limited proliferation and poor spreading on nanofibers, adopting rounded morphology, while MSCs spread better but grew slowly. In contrast, chondrocytes alone or in co-culture (OB/CH, MSCs/CH) exhibited rapid growth, reaching near confluence by day 7 and were not cultured further. MSCs/OB co-cultures showed notably enhanced proliferation compared to either cell type alone, and were extended to day 14 (Fig. 2G). Quantitative analysis of two-week proliferation rates revealed a threefold increase for MSCs/OB co-culture (slope = 485.9) compared to MSCs-only (178.0) and OB-only (164.6), highlighting a strong synergistic effect on scaffold colonization.

dECM analysis for DNA and total protein content

Decellularization of cell-laden ECM should minimize DNA content while preserving high protein levels to ensure biocompatibility and bioactivity. DNA quantification demonstrated effective decellularization across all scaffolds. Residual DNA levels were effectively reduced in all dECM groups, remaining below the FDA-recommended threshold of 50 ng/mg. The lowest DNA content was observed in CH (18.4 ± 2.1 ng/scaffold) and EC/CH (19.7 ±3.4 ng/scaffold) matrices (Supplementary Fig. 1A). Among all groups, OB/MSCs dECM had the highest total protein content (67.86 ± 8.32 μg/scaffold), followed by FB/MSCs (65.02 ± 2.5 μg/scaffold) and FB alone (62.21 ± 1.13 μg/scaffold), likely reflecting stronger cell-material and cell-cell interactions. In contrast, EC co-cultures yielded the lowest protein content, particularly in EC/FB (20.12 ± 0.64 μg/scaffold) and EC/CH (17.29 ±0.14 μg/scaffold) groups (Supplementary Fig. 1B). Efficient decellularization, characterized by low residual DNA and preserved ECM proteins, further supports the low immunogenicity and regenerative potential of these scaffolds.

Bioactivity assessment of dECM

Simulated Body Fluid (SBF) was utilized to replicate in vivo-like mineral depositions on scaffolds. Scaffolds from the no ECM (PCL-only) group and 15 different dECM-modified groups were soaked in SBF for four weeks. Alizarin Red S staining was performed weekly to quantify calcium-based mineral deposition (Fig. 3A–E). Overall, mineralization levels generally aligned with the total protein content retained after decellularization, indicating a positive correlation between dECM protein abundance and mineral deposition. For example, the OB/MSCs group, which retained the highest protein content post-decellularization, exhibited the second-highest mineralization, while the FB/MSCs group, second in protein content, showed the highest mineralization (Fig. 3E). However, deviations were observed, particularly in groups involving fibroblasts and endothelial cells. Although the fibroblast-only dECM scaffold retained substantial protein levels, it exhibited relatively poor mineral deposition. Conversely, the EC/CH co-culture group, despite having the lowest protein yield among all dECM groups, demonstrated higher mineralization than most single-cell dECM scaffolds. Notably, coculture-derived dECMs consistently resulted in greater mineral deposition compared to their single-cell counterparts and significantly outperformed the no-ECM control. These findings underscore the superior bioactivity of hybrid dECMs and their potential to enhance biomineralization in scaffold-based bone regeneration strategies. Alizarin-stained scaffolds of selected dECM groups exhibit high mineral deposition at the end of the 4-week SBF soaking period (Fig. 3F). The red color intensity and nodules indicate the areas of calcium deposition (Fig. 3F). Compared to the no ECM control, the dECM-modified scaffolds demonstrated markedly greater mineral accumulation, with staining covering nearly the entire scaffold surface. In several dECM groups, intense dark red regions were observed, suggesting multilayered or denser mineral deposition. In contrast, the no ECM group displayed only sparse and limited red staining, indicating minimal mineralization in the absence of dECM. These results further support the role of dECM in enhancing scaffold bioactivity and promoting in vitro biomineralization.

Fig. 3. Cell-laden dECM enhanced in vitro mineralization of PCL nanofibers over time, indicating osteoinductive potential.

Relative absorbance at 562 nm, corresponding to Alizarin Red S staining, indicates calcium (Ca²⁺) deposition by cells over 4 weeks in co-cultures with (A) OB, (B) MSCs, (C) CH, (D) EC, and (E) FB. (F) Representative images show enhanced calcium accumulation in dECM scaffolds (OB/FB, OB/MSCs, OB/CH, FB/MSCs, EC/MSCs) compared to the no-ECM control. All experimental groups at weeks 1–4 were significantly higher than their preceding time points and the PCL control; however, for clarity, only week 4 significance is indicated in the figure. Data presented as Mean ± SD, n=6 using one-way ANOVA with Tukey’s post hoc test. ****p < 0.05 vs. PCL at 4^th week.

Cell laden dECM on the nanofiber matrices provide a 3D structural and biochemical environment that regulates key cellular behaviors such as adhesion, proliferation, and differentiation [49]. To assess their osteoinductive potential, cell proliferation, ALP activity, and mineralization, were evaluated over 28 days (Fig. 4A–C). Based on prior protein quantification and biodeposition data, five hybrid dECM groups OB/FB, OB/MSCs, OB/CH, FB/MSCs, and EC/MSCs were selected for further evaluation of their potential to promote osteogenesis. MSCs were seeded onto cell laden dECM matrices to evaluate their bioactivity in terms of proliferation, ALP expression and mineralized matrix synthesis over four weeks (Fig. 4A–C). During the first week, all groups showed comparable cell numbers, confirming consistent initial seeding density. By week 3, a marked increase in MSCs proliferation was observed in all groups except EC/MSCs. In week 4, proliferation plateaued in the OB/MSCs, FB/MSCs, and OB/CH groups, as reflected by a reduced slope (Fig. 4A), which is likely due to the cells reaching confluency. Notably, while the EC/MSCs group previously demonstrated strong mineralization in biodeposition assays, its dECM composition appeared less supportive of MSCs proliferation, highlighting potential differences in bioactivity depending on cell source combinations. The ALP activity assay results mirrored the trend observed in mineralization assay (Fig. 4B). While the differences among groups were less pronounced than in the mineralization analysis, the OB/MSCs group consistently exhibited the highest ALP activity by week 4. As shown in Fig. 4B, the rate of increase in ALP activity for the OB/MSCs group surpassed that of the other groups, indicating a more rapid progression of osteogenic differentiation. These findings further support the conclusion that the OB/MSC hybrid dECM enhances MSCs differentiation and promotes osteogenesis more effectively than other dECM formulations. Mineral deposition by MSCs on different scaffolds was assessed using Alizarin Red S staining, starting from week 2 (Fig. 4C). The OB/FB and EC/MSCs groups exhibited consistently low levels of mineralization throughout the culture period. Notably, while the OB/MSCs group showed an initially comparable mineralization to the OB/FB and EC/MSCs groups, its mineral deposition significantly increased over time, surpassing all other groups by week 4. This trend is clearly illustrated in Figure 3B, where the OB/MSCs group displayed the steepest slope, indicating the most robust increase in calcium deposition. In contrast, the relatively flat slopes of the OB/FB and EC/MSCs groups suggest limited osteogenic potential. These findings diminish the suitability of OB/FB and EC/MSCs hybrid dECM formulations for promoting MSCs osteogenesis, while highlighting OB/MSCs as a superior candidate for constructing osteoinductive dECM-based soft-shell scaffolds.

Fig. 4. Cell-laden dECM matrices promoted in vitro osteogenic differentiation of MSCs.

(A) RFU measured by CellTiter-Blue assay indicate MSCs proliferation, (B) ALP activity reflects early osteogenic differentiation, and (C) mineralized matrix formation confirms late-stage osteogenesis. All groups showed significant increases from week 1 to week 4, consistent with progressive osteoblast phenotype development. OB/MSCs exhibited significantly higher proliferation and ALP activity, while mineralized matrix deposition at week 4 was not significantly different from OB/CH and FB/MSCs. For clarity, only selected comparisons are shown; full statistical details are provided in the text. Data are presented as Mean ± SD (n = 6) and analyzed using one-way ANOVA with Tukey’s post hoc test.

The expression of bone-specific proteins collagen I (COLI), an early marker of osteoblast differentiation, and osteopontin (OPN), a mid-to-late marker associated with early mineralization was qualitatively assessed by immunofluorescence staining to evaluate the osteogenic differentiation of MSCs cultured on cell-laden dECM matrices (Supplementary Fig. 2A). COL1 staining was conducted from week 2, while OPN staining began at week 3. At week 2, minimal differences in COL1 expression were observed among groups, although EC/MSCs and OB/FB groups exhibited notably weaker signals (Supplementary Fig. 2A). By week 3, COL1 expression intensified in OB/MSCs, FB/MSCs, and OB/CH groups, while EC/MSCs and OB/FB remained weak for both COL1 and OPN (Supplementary Fig. 2B). Importantly, OB/MSCs and FB/MSCs began to show strong OPN expression at this stage. By week 4, EC/MSCs and OB/FB groups demonstrated improved COL1 expression, although this response lagged other groups. In contrast, OB/MSCs, FB/MSCs, and OB/CH groups exhibited robust and widespread staining for both COL1 and OPN, along with a morphological shift in MSCs to a star-like shape indicative of osteoblastic differentiation. These findings align with the mineralization and ALP data, reinforcing the superior osteoinductive potential of the OB/MSCs hybrid dECM, which was ultimately selected as the core material for the biphasic scaffold.

In vivo bone healing efficacy of biphasic scaffolds in a rat critical sized-femoral defect

To evaluate the individual and combined roles of biomineralized calcium phosphate (CaP) and decellularized hybrid extracellular matrix (dECM) in bone regeneration, we established seven experimental groups (Table 2). These groups leverage our primary biphasic scaffold (BS), which consists of a hard shell (HS) and a nanofiber (NF) core. For clarity, groups were organized into Test Combinations, Scaffold Component Controls, and Biological Controls. Test Combinations: These groups evaluate the performance of our BS with various bioactive additions. G1: BS+CaP+dECM, G2: BS+CaP and G4: BS+dECM. Scaffold Component Controls: These groups isolate the effects of specific scaffold components. G3: HS+CaP (W/O NF) and G5: BS (BS base control). Biological Controls: G6: Autograft (Positive Control) and G7: Untreated Defect (Negative Control). As shown in Fig. 5A, the dECM-optimized nanofibrous scaffolds were tightly rolled and inserted into the lumen of the porous tubular PCL scaffold. The assembled biphasic construct was then coated with CaP. The surgical procedure and overall scaffold structure are illustrated in Fig. 5B(a–f), where a 1 cm long bone defect was repaired using these scaffolds, and healing was evaluated relative to controls. After 12 weeks, regenerated bone tissue was harvested (Fig. 5C) and analyzed for quality and functionality using mechanical testing and histological evaluation. Tissue morphology demonstrated that all harvested bone remained fixed except the defect group. Compressive mechanical strength of the regenerated femoral bone is presented in Fig. 5D. Compressive testing revealed significant differences in mechanical strength among the groups, with G1 (BS+CaP+dECM ) exhibiting the highest strength (19.1 ± 0.3 MPa), followed by G2 (BS+CaP ) (14.5 ± 0.6 MPa) (Fig. 5D). These values were significantly higher than the negative control (defect-only G7) (7.3 ± 0.4 MPa) (p < 0.0001), indicating the importance of both CaP and dECM in enhancing mechanical integrity. G1 significantly outperformed all other test and control groups (p < 0.0001), demonstrating the synergistic effect of CaP and dECM.

Fig. 5. Bone healing efficacy of bioactive hybrid scaffolds in a rat critical-sized femoral defect model.

Seven groups (G1–G7) were evaluated, with G5 (BS Base control), G6 (Autograft), and G7 (Untreated defect) serving as controls, while test groups G1–G4 contained varied bioactivity components. (A) Schematic of the biphasic scaffold. (B) Surgical procedure: (a) Prone positioning, (b) Muscle incision exposing the femur, (c) External fixator secured with K-wire, (d) Created 10 mm segmental defect, (e) Scaffold placement (f) Sutured muscle and skin with external fixator in place. (C) Tissue morphology of bone samples at 12 weeks. (D) Compressive modulus of regenerated bone 12-weeks post-surgery. Data presented as Mean ± SD, n=6 using one-way ANOVA with Tukey’s post hoc test. ####p < 0.0001 vs G3, !!p < 0.01, !!!p < 0.001 vs G4, %p < 0.05, %%p < 0.01 vs G7, ¥¥¥¥p<0.0001 vs G2, G3, G4, G5, G6, G7.

Radiographic and micro-CT evaluation of bone regeneration

At 12 weeks post-implantation, radiographic analysis showed new bone formation along the defect margins in all groups, though some animals exhibited Kirschner wire misalignment and bone deformation (Fig. 6A). While all implant groups integrated with native bone, the defect-only group (G7) failed to achieve full fusion, likely due to insufficient osteogenic activity. Bone healing outcomes were evaluated using a composite scoring system that assessed three key parameters in radiography: periosteal reaction, bone union, and remodeling. Each parameter was scored on a scale (0–3 for periosteal reaction and bone union; 0–2 for remodeling), with a maximum total score of 8, indicating optimal healing (Table 4) [46]. The periosteal reaction score reflected the extent of new bone formation across the periosteum, bone union represented the degree of bridging at the defect site, and remodeling assessed cortical bone maturation. Groups 1 (BS+CaP+dECM ), 4 (BS+dECM), and 6 (Autograft) achieved the maximum total score of 8, signifying complete periosteal reaction and bone union with moderate remodeling. These results suggest that both dECM scaffolds provide strong osteogenic support comparable to the gold standard treatment, autograft. The consistent performance of Groups 1 and 4 emphasizes the synergistic effect of dECM scaffolds when combined with CaP-based systems. In contrast, Groups 2 (BS+CaP ) and 3 (HS-CaP) scored 5, showing moderate periosteal reaction and bone union but only mild remodeling. This indicates that BS or HS-CaP alone provides limited osteogenic enhancement. Group 5 (BS) scored slightly higher (6), suggesting some benefit from NF when combined with HS, although still inferior to the dECM formulations. This highlights the necessity of active biomaterial intervention for effective bone regeneration. Overall, the findings clearly demonstrate that hybrid nanofiber scaffolds significantly enhance bone healing (Table 4). To further the assessment of callus formation, micro-CT was utilized. G1 exhibited the highest bone volume (BV: 475.8 ± 83.8 mm³), significantly greater than G5 (186.9 ± 29.3 mm³, p < 0.001) and G7 (242.0 ± 62.1 mm³, p < 0.05) (Fig. 6B–C). G4 and G6 (Autograft) also showed significantly greater BV than G5 (p < 0.05 and p < 0.01, respectively). Bone mineral density (BMD) was highest in G1 (589.6 ± 35.6 mg/ccm HA), significantly exceeding G5 (366.5 ± 80.2 mg/ccm HA, p < 0.001) and G7 (490.3 ± 20.8 mg/ccm HA, p < 0.01). Groups containing both CaP and dECM (G1–G4) demonstrated nearly complete defect regeneration with aligned cortical bone, while G7 exhibited incomplete bridging and persistent gaps (Fig. 6D). BV/TV analysis confirmed structural and mineralization improvements in G1, nearing autograft (G6) levels (Fig. 6E). Scaffolds without dECM (G2, G3, G5) bridged the defect but had inferior mineralization.

Fig. 6. Representative radiograph and micro-CT images illustrating bone regeneration and fixation in a femoral defect model 12-weeks post-surgery.

Seven groups (G1–G7) were evaluated, with G5 (BS Base control), G6 (Autograft), and G7 (Untreated defect) serving as controls, while test groups G1–G4 contained varied bioactivity components. (A) Radiograph images of the defect site stabilized with metallic implants, depicting fracture alignment, fixation stability, and bone healing progression over time. (B) Micro-CT images showing progressive scaffold integration and new bone formation. (C) Quantitative analysis of bone regeneration, including Bone Volume (BV), (D) Bone Mineral Density (BMD), and (E) Bone Volume Fraction (BV/TV), demonstrating the structural and mineralization changes during healing. Data presented as Mean ± SD, n=6 using one-way ANOVA with Tukey’s post hoc test. $p<0.05, $$p<0.01, $$$p<0.001 vs G5. %p<0.05, %%p<0.01 vs G7.

Table 4.

Radiographic image scoring

Groups	Periosteal Reaction	Bone Union	Remodeling	Total Score (Max 8)
G1 (BS+Cap+dECM)	3 (Complete)	3 (Complete)	2 (Moderate)	8 (Best outcome)
G2 (BS+CaP)	2 (Moderate)	2 (Moderate)	1 (Mild)	5
G3 (HS+CaP)	2 (Moderate)	2 (Moderate)	1 (Mild)	5
G4 (BS+dECM)	3 (Complete)	2 (Moderate)	2 (Moderate)	8 (Best outcome)
G5 (BS)	2 (Moderate)	2 (Moderate)	1 (Mild)	6
G6 (Autograft)	3 (Complete)	3 (Complete)	2 (Moderate)	8 (Gold standard)
G7 (Untreated defect)	1 (Mild)	1 (Mild)	1 (Mild)	3

The study compares seven groups based on periosteal reaction, bone union, and remodeling, with a total score of 8 reflecting overall bone healing quality.

Histological and immunohistochemical evaluation

Histological and immunohistochemical analyses were performed to evaluate bone regeneration and tissue integration. Hematoxylin-eosin (H&E), Masson-Goldner trichrome (MT), and Safranin O staining were used to assess new bone formation, collagen deposition, and endochondral ossification, respectively. Additionally, immunostaining for bone sialoprotein (BSP) and collagen type I (COL I) examined bone matrix protein expression and maturation. H&E-stained sections revealed varied degrees of bone regeneration among the groups (Fig. 7A). Distinct histological differences were evident between newly formed (black stars) and residual bone (red stars), reflecting different stages of remodeling. Low-magnification views illustrated overall defect filling and tissue architecture, while high-magnification images highlighted differences in matrix organization and cellular distribution. The most prominent osteoid formation, an early marker of bone regeneration was observed in G1 (BS+CaP+dECM) and G4 (BS+dECM), compared to G5 (base control) and G7 (Untreated defect) indicating active osteogenesis. Quantitative histomorphometry confirmed that G1 and G4 had significantly higher osteoid area percentages compared to G2, G3, and all control groups (p < 0.05) (Fig. 7B). These results suggest that the combined presence of CaP and dECM in these scaffolds enhances early-stage bone regeneration by promoting osteoid matrix deposition.

Fig. 7. Both G1 & G4 exhibited excessive production of osteoid that was comparable to autograft and significantly higher than defect control.

Seven groups (G1–G7) were evaluated, with G5 (BS Base control), G6 (Autograft), and G7 (Untreated defect) serving as controls, while test groups G1–G4 contained varied bioactivity components. The inset images in each panel provide a lower magnification view (scale bar: 3 mm), illustrating the overall defect area, while the main images show higher magnification details (scale bar: 100 μm). (A) Histological analysis of bone regeneration in different experimental groups using H&E staining. Black stars indicate areas of newly formed bone, while red stars highlight regions of old bone. (B) Quantification of osteoid area percentage was performed using ImageJ software. Data presented as Mean ± SD, n=3 using one-way ANOVA with Tukey’s multiple comparisons test. *p < 0.05 vs G2. #p < 0.05 vs G3. $p < 0.05 vs G5. %p < 0.05, %%p < 0.01 vs G7.

Masson’s Trichrome staining revealed enhanced mature collagenous bone tissue in G1 (BS+CaP+dECM) and G6 (Autograft), indicating superior bone remodeling and regeneration (Supplementary Fig. 3A). Quantitative analysis confirmed a significantly higher collagen-rich area in G1 and G6 compared to G2 (p < 0.0001), G5 (p < 0.01), and G7 (p < 0.05) (Supplementary Fig. 3B). Collagen appeared prominently in blue and osteoid tissue in red, with both features more pronounced in G1 and G6. Despite the lack of complete bone union in G7, mature collagenous bone tissue observed similar to G3, which can be attributed to the formation of callus. This callus, though less mineralized, indicates active bone formation and ongoing remodeling at the defect site. These findings highlight the critical role of incorporating both CaP and dECM in scaffolds to create a structurally and biochemically supportive environment for bone healing.

Safranin O staining showed increased cartilage and callus formation in G1 (BS+CaP+dECM) and G4 (BS+dECM ), indicating enhanced early-stage bone repair and osteochondral regeneration (Fig. 8A). Quantitative analysis confirmed a significantly greater cartilage/callus area in G1 and G4 compared to G2 (p < 0.05), and even more so versus G3, G5, and G7 (p < 0.01) (Fig. 8B). Proteoglycan-rich cartilage (red/pink) and mineralized bone (blue/green) were more prominent in G1 and G4 than in controls, reflecting superior regenerative activity.

Fig. 8. G1 demonstrated cartilage and callus formation comparable to autograft and significantly higher than defect only.

Seven groups (G1–G7) were evaluated, with G5 (BS Base control), G6 (Autograft), and G7 (Untreated defect) serving as controls, while test groups G1–G4 contained varied bioactivity components. Insets provide a macroscopic overview of the defect site (scale bar: 3 mm), while larger images depict microscopic structural details (scale bar: 100 μm). (A) Representative images of tissue sections stained with Safranin O to assess cartilage formation and bone regeneration. Red/pink regions indicate proteoglycan-rich cartilage, while blue/green regions correspond to collagen and mineralized bone matrix. Each group with high-magnification views highlighting cellular organization, tissue integration, and matrix deposition. (B) Quantification of cartilage/callus area percentage was performed using ImageJ software. Data presented as Mean ± SD, n=3 using one-way ANOVA with Tukey’s multiple comparisons test. *p < 0.05, vs G2. ##p < 0.01 vs G3. $p < 0.05, $$p < 0.01 vs G5. %p < 0.05, %%p < 0.01 vs G7.

Immunohistochemical analysis showed significantly elevated expression of collagen type I (COL I) and bone sialoprotein (BSP) in G1 (BS+CaP+dECM) and G4 (BS+dECM ), indicating enhanced bone matrix production and mineralization (Supplementary Fig. 4A). COL I stain (brown) revealed abundant collagen deposition, while BSP staining confirmed active osteogenesis. Quantitative analysis demonstrated significantly higher COL I (Supplementary Fig. 4B) and BSP (Supplementary Fig. 4C) expression in G1 and G4 compared to G2, G3, and controls G5–G7 (p < 0.01 to p < 0.0001). These findings suggest that incorporating dECM into the scaffold enhances osteoblastic activity and extracellular matrix formation, leading to improved bone regeneration (Supplementary Fig. 4D). Histological scoring aligned with radiographic and micro-CT findings, confirming significant differences in bone regeneration among the groups. G1 (BS+CaP+dECM) and G4 (BS+dECM) showed the most favorable outcomes, closely resembling the autograft group (G6). Immunohistochemistry reinforced these observations: strong BSP and COL I expression in G1 indicated high osteoblastic activity and organized matrix formation. In contrast, the defect-only group (G7) showed minimal BSP and excessive COL I expression hallmarks of fibrosis and impaired healing. These results underscore the synergistic role of dECM’s biochemical cues and CaP’s structural support. Groups lacking dECM (G2, G3, G5) showed moderate bone formation but poorer trabecular organization and incomplete cortical bridging.

Discussion

This study presents a comprehensive evaluation of a biphasic composite scaffold designed to address critical challenges in bone tissue engineering by integrating mechanical strength with bioactive functionality. By combining a CaP-reinforced porous PCL outer structure with a dECM-laden nanofibrous core, the scaffold was engineered to replicate key architectural and biochemical features of native bone. Design optimization, in vitro characterization, and in vivo assessments collectively demonstrate that the synergistic integration of CaP and dECM significantly enhances scaffold performance promoting osteoconductivity, cellular infiltration, matrix remodeling, and functional bone regeneration across large defects.

The biphasic scaffold’s design successfully integrates dECM-functionalized nanofibers within a hard-shell structure, which balances mechanical strength and porosity essential for bone tissue engineering [43]. The biomineralized CaP deposition achieved via SBF immersion confirms successful biofunctionalization, supporting osteoconductive potential [24]. The outer tube’s porous PCL design, with a porosity near 90%, achieved a desirable compromise between mechanical integrity and permeability. The observed decrease in compressive modulus relative to nonporous scaffolds is consistent with increased porosity, as expected from prior studies [43]. Electrospinning parameters notably influenced fiber diameter, with polymer concentration being the predominant factor[40]. The inverse relationship between fiber thickness and polymer concentration, as evidenced by SEM images and diameter distributions, reflects electrospinning dynamics under constant flow and voltage conditions. The preference for thinner fibers (5% PCL) in cell proliferation assays across multiple cell types suggests enhanced surface area and porosity facilitate greater cell attachment and growth [48]. Co-culture studies emphasize the interactive effects between MSCs and OBs, where the synergistic proliferation observed (Fig. 2) likely results from paracrine signaling and complementary ECM production [50, 51]. Overall, these findings confirm that scaffold design, fiber morphology, and cell-cell interactions critically influence bone repair efficacy. The well-spread cell morphology with extended filopodia anchoring to the nanofibers indicate effective adhesion and integration [50]. Nanofiber scaffolds mimic natural ECM and offer high surface area and porosity ideal for cell growth and osteogenesis under induction [44–46]. These matrices provide a supportive microenvironment for OB/MSCs co-cultures, facilitating effective ECM deposition [50, 51].

Effective decellularization was achieved, as indicated by low residual DNA content across all dECM scaffolds, with levels well below the FDA safety threshold. The particularly low DNA quantities in CH and EC/CH matrices (Supplementary Fig. 1) suggest efficient removal of nuclear material, reducing the risk of immunogenicity. The protein content trends reflect the influence of cell type and co-culture dynamics on ECM deposition. Hybrid dECM scaffolds, especially those derived from OB/MSCs and FB/MSCs, exhibited significantly higher total protein levels than their single-cell counterparts. This likely results from enhanced cell–matrix and cell–cell interactions within cocultures, which are known to stimulate greater ECM secretion [52]. In contrast, the minimal protein deposition in EC-involved cultures, especially EC/CH and EC/FB, may indicate limited matrix production capacity or weaker interactions. The preserved ECM components collagen, elastin, fibronectin, and growth factors are critical to maintaining scaffold functionality and bioactivity [46]. These proteins contribute to mechanical integrity, cell adhesion, and signaling pathways essential for regeneration. The decellularized scaffolds retained key ECM components such as collagen, elastin, fibronectin, and growth factors, which are essential for structural integrity and regenerative function [53]. The significantly higher total protein content in hybrid dECM scaffolds compared to those derived from alone suggests synergistic cell–cell and cell–matrix interactions that enhance ECM deposition [54].

The bioactivity of dECM scaffolds, as measured by SBF-induced mineralization, MSC proliferation, ALP activity, and osteogenic marker expression, demonstrated the significant influence of cell origin and co-culture combinations on scaffold function. The strong correlation between total protein content and mineral deposition suggests that dECM richness enhances scaffold bioactivity. Coculture-derived dECMs outperformed single-cell-derived matrices, indicating enhanced bioactivity due to synergistic cell–cell and cell–matrix interactions. Particularly, OB/MSCs dECM consistently supported superior MSC proliferation, ALP activity, and mineral deposition, suggesting a potent osteoinductive environment. Overall, the 3D architecture and bioactive composition of dECM including structural proteins, growth factors, and adhesive ligands create an instructive microenvironment that supports dynamic remodeling, cell differentiation, and tissue integrity [53, 55].

The in vivo results confirm that combining calcium phosphate (CaP) and decellularized extracellular matrix (dECM) within a biphasic scaffold significantly enhances both mechanical and biological performance. The G1 group (BS+CaP+dECM) showed the highest compressive strength, highlighting the synergistic effect of CaP’s structural reinforcement and dECM’s bioactivity. CaP contributed to load-bearing capacity by mimicking bone’s mineral phase, while dECM promoted osteogenesis and matrix remodeling. G1 even outperformed the autograft control, underscoring its clinical potential. These findings validate the importance of integrating both bioactive and mechanical components for effective bone tissue regeneration. Scaffolds containing both CaP and dECM showed higher compressive modulus than those lacking either component, underscoring their contribution to mechanical reinforcement and in vivo stability [56, 57]. The ability of G1 to withstand physiological loads while supporting cell activity and matrix remodeling confirms the functional benefit of this composite design. The compression testing was used as the primary mechanical assessment due to the fact that load transmission along the femoral axis during stance is largely compressive. This approach ensured physiological relevance of the measurements. Although tensile and bending tests could have provided complementary insights, the restricted bone volume inherent to the rat defect model limited the feasibility of performing multiple destructive tests on the same specimen. Consequently, prioritizing compression testing allowed us to capture the most functionally meaningful parameter while preserving the integrity of the experimental design. These findings are consistent with previous studies showing that CaP improves scaffold strength by mimicking bone’s mineral matrix, while dECM enhances bioactivity and promotes osteogenesis [58, 59]. Overall, CaP and dECM integration significantly improves scaffold performance, supporting their application in load-bearing bone tissue engineering.

The radiographic and micro-CT results demonstrate that dECM-containing scaffolds (G1 and G4) provide robust osteogenic support comparable to autograft (G6), the clinical gold standard [49]. The maximum healing scores in these groups reflect complete periosteal reaction, bone union, and moderate remodeling, highlighting the importance of bioactive matrix components in promoting bone regeneration. Groups lacking dECM (G2, G3, G5) showed less effective bone remodeling and mineralization, emphasizing the critical role of dECM in enhancing osteogenic activity. The improved bone volume and mineral density in G1 suggest a synergistic effect between CaP mineralization and the bioactive nanofiber core, leading to superior structural and biological outcomes. These results underscore the importance of dECM in promoting osteogenesis and mineral deposition, supporting its role as a bioactive substitute for autografts [13, 52, 60]. The synergistic integration of CaP and dECM in G1 yielded the highest BV and BMD, validating its potential for segmental bone repair. Enhanced cortical bone architecture in G1, aligning with prior reports that CaP enhances scaffold osteoconductivity and mechanical strength[53, 57, 61], while dECM provides bioactive cues for cell adhesion, proliferation, and differentiation[58, 59,61]. Together, these components significantly improve scaffold performance, offering a promising biomaterial strategy for clinical translation[53, 57, 61].

The histological and immunohistochemical analyses demonstrate that scaffolds combining CaP and dECM (G1 and G4) significantly enhance bone regeneration by providing both structural support and bioactive cues. Increased osteoid formation and elevated expression of COL I and BSP indicate active early-stage osteogenesis and matrix mineralization. Masson’s Trichrome staining confirmed superior collagen maturation in these groups, closely resembling physiological bone remodeling. Safranin O staining further revealed enhanced cartilage and callus formation, highlighting the dECM’s role in promoting endochondral ossification, essential for effective repair of large bone defects. In contrast, groups lacking dECM (G2, G3, G5) showed limited bone formation, poorer trabecular organization, and incomplete cortical bridging despite some mineral deposition, emphasizing dECM’s critical function in matrix maturation and osteogenic signaling. These results suggest that dECM promotes endochondral ossification by supporting a cartilage intermediate, a key step in repairing large bone defects[62]. This dual action facilitating endochondral pathway highlights the robust osteogenic potential of dECM-containing scaffolds [61].

Micro-CT and radiographic analyses further validated these histological findings. Groups incorporating CaP and/or dECM scored highest for callus formation, bony union, and cortical remodeling. Notably, G1 outperformed the gold-standard autograft (G6) in several regenerative metrics, likely due to its superior mimicry of native bone’s biochemical and structural environment[7]. In contrast, G7 consistently showed the poorest outcomes across all analyses minimal callus, no cortical bridging, and low histological scores echoing prior reports on untreated defects favoring fibrous tissue formation over ossification[63].

This study addresses key limitations of current bone graft substitutes namely, insufficient mechanical strength, poor nutrient diffusion, and limited bioactivity which are critical for supporting uniform tissue ingrowth and mineralization in large bone defect repair [58, 59, 64, 65]. The BS+CaP+dECM scaffold, which integrates synthetic components (PCL and nanofibers) with biological cues (dECM), significantly enhanced healing, integration, and mineralization compared to conventional approaches. These findings support the use of biomimetic biphasic scaffolds that combine mechanical stability with biological signaling to improve regenerative outcomes. Future studies should focus on longterm in vivo evaluation of scaffold integration, functional recovery, and degradation. Incorporating additional bioactive agents such as osteoinductive factors, autologous or allogeneic cells, small molecule drugs, or immunomodulators may further enhance therapeutic efficacy.

Overall, this study characterized a composite biphasic scaffold consisting of a mechanically robust, highly porous layered PCL tube coated with CaP and a bioactive nanofibrous core loaded with decellularized dECM, engineered to replicate native bone architecture and support osteogenesis. The hybrid dECM derived from OB and MSCs co-cultures retained its bioactivity and more effectively promoted in vitro osteogenic differentiation of rat mesenchymal stem cells compared to individual dECMs. The scaffold facilitated uniform tissue infiltration and mineralization across large bone defects. The synergistic effect of CaP and dECM significantly enhanced osteogenesis, cellular infiltration, callus formation, cortical bridging, and marrow regeneration, outperforming autografts. This biomimetic construct supports both early repair and long-term remodeling, showing strong potential for clinical use in load-bearing bone repair. Future studies should evaluate scaffold degradation, long-term efficacy, and incorporation of osteoinductive factors to further optimize performance.

Availability of data and material

All relevant data is presented within the manuscript. Any additional metadata will be made available from the corresponding author upon reasonable request.