A 3D prevascularized calcium phosphate cement scaffold for accelerated alveolar bone regeneration and angiogenesis in rats

Abstract

Objectives

The objective of this study is to develop a load-bearing prevascularized construct by combining calcium phosphate cement (CPC) with cells within a three-dimensional (3D) hydrogel culture system, to accelerate the regeneration of alveolar bone defects.

Methods

A 3D co-culture system was established by encapsulating human periodontal ligament stem cells (hPDLSCs) and human umbilical vein endothelial cells (hUVECs) within a gelatin methacryloyl (GelMA) hydrogel on 3D-printed porous CPC scaffolds. The mechanical properties, pore structure and angiogenic potency were determined in vitro. In vivo performance was evaluated using a nude rat subcutaneous implantation model and a rat alveolar bone defect model. Four groups were tested: (1) Blank group (surgery-only group); (2) CPC+GelMA group (non-prevascularized group); (3) CPC+GelMA-cell group (prevascularized group)༛(4) Natural Periodontium group.

Results

The novel construct had good mechanical properties and biocompatibility. The 3D co-culture in GelMA successfully induced microvascular formation in vitro. Subcutaneous implantation in nude rats showed that the CPC+GelMA-cell group exhibited markedly greater angiogenic capacity than the CPC+GelMA group after 6 weeks, with a neovascular density 1.93-fold higher than that of the non-prevascularized group. Among all groups, the CPC+GelMA-cell group exhibited the strongest capacity for repairing rat alveolar bone defects. Compared to CPC+GelMA group, CPC+GelMA-cell group significant enhanced bone regeneration in rats by 1.23–1.37 folds, and increased vascularization by 2.65 folds (p<0.05).

Conclusions

The novel 3D prevascularized CPC construct combined appropriate mechanical properties with great efficacy for alveolar bone regeneration and vascularization in vivo in an animal model.

Background

Orthodontic treatment carries the risk of periodontal tissue damage. Studies indicate that over one-third of adult patients exhibit anterior alveolar bone recession exceeding 2 mm post-treatment [1]. Improper orthodontic design may also lead to alveolar bone defects such as fenestration and dehiscence [1]. Tissue engineering scaffolds are a critical tool for reconstructing alveolar bone defects, yet their regenerative potential is critically constrained by inadequate and delayed vascularization [2]. Insufficient blood supply can cause hypoxia, impair osteogenesis, and lead to graft necrosis and failure [2, 3]. Therefore, overcoming insufficient vascularization is essential for bone-defect repair.

Prevascularized scaffolds enables their vascular networks to anastomose with the host’s vasculature after implantation, markedly reducing the risk of ischemic necrosis, and substantially increasing the success rate of bone regeneration [4].

Calcium phosphate cement (CPC) is a self-setting osteogenic scaffold material, resembling the inorganic composition of natural bone [5]. CPC regards as a highly promising scaffold material for craniofacial and dental repairs, due to its good biocompatibility, osteoconductivity, and mechanical properties [5]. Nevertheless, unmodified CPC is predominantly microporous and lacks interconnected macroporous structures, hindering vascular ingrowth and limiting its utility in large bone defects [6]. Prevascularized CPC scaffolds may significantly enhance bone repair capabilities. Currently, research on prevascularized calcium-phosphate osteogenic scaffolds primarily focuses on cranial and long bone repair [7–10]. To date, there has been no report of constructing three-dimensional (3D) prevascularized CPC scaffolds for alveolar bone tissue engineering.

Prevascularization of CPC scaffold is currently achieved through vascular implantation [7], in vivo culturing [6], or cell-based in vitro prevascularization [11]. However, vascular implantation and in vivo culturing involve invasive procedures, limiting their clinical applicability. Therefore, cell-based in vitro prevascularization is now in the spotlight [12]. Human umbilical vein endothelial cells (hUVECs) readily self-assemble into microcapillaries, yet endothelial monocultures fail to generate stable, mature vascular structures [13]. Endothelial cell migration and neovessel formation require specific pro-angiogenic factors, which are insufficiently produced by endothelial cells in monoculture [4, 14]. Mesenchymal stem cells (MSCs) enhance angiogenesis by secreting pro-angiogenic factors and stabilizing nascent vessels as pericytes [13, 15, 16]. Numerous studies demonstrate that co-culturing MSCs with endothelial cells yields stable vascular structures [16].

Human periodontal ligament stem cells (hPDLSCs) are a seed cell source that can be harvested from the extracted wisdom teeth or the teeth extracted for orthodontic purpose without additional invasive surgery for the patient. Previous studies have shown positive results in using hPDLSCs for tissue regeneration, especially in bone and periodontal tissue repairment [17, 18]. In addition, hPDLSCs can differentiate into bone, nerve, connective tissue, and cementum under specific conditions [19, 20]. Therefore, hPDLSCs are a potent cell source in stem cell delivery via scaffolds for bone regeneration, especially for alveolar bone repair. Studies showed that co-culturing hPDLSCs and hUVECs on CPC surfaces can form microvascular-like structures, suggesting their potential for CPC scaffold prevascularization [11]. To date, a literature search revealed no report on the prevascularized CPC scaffolds seeded with hPDLSCs and hUVECs for alveolar bone defect repair.

Beyond seed cell selection, scaffold structural properties also influence prevascularization efficacy. 3D printing technology enables precise control over pore size and interconnectivity [21, 22]. 3D-printed grid-like CPC scaffolds exhibit uniform, interconnected micropores that readily conducive to microvascular growth support microvessel ingrowth and anastomosis [23]. Additionally, the culture method for seed cells is crucial for prevascularization. Traditional two-dimensional (2D) culture involves seeding cells directly onto scaffold surfaces, where restricted area and contact inhibition curtail expansion and capillary morphogenesis [24, 25]. Research indicated that 3D microenvironments enhance cell responsiveness to biochemical signals during angiogenesis [24]. Compared to 2D culture, 3D culture better mimics natural cell growth conditions, promoting cell-cell and cell-matrix interactions [26]. Encapsulating cells in hydrogel-based materials can simulate extracellular matrix structures, providing 3D support [24]. Gelatin methacryloyl (GelMA) hydrogel offers excellent biocompatibility and enables 3D cell encapsulation, simulating a natural extracellular matrix conducive to microvascular formation [27]. However, hydrogels lack sufficient mechanical strength for load-bearing applications like alveolar bone repair [28]. Recent studies combine hydrogels with high-strength scaffolds to create composite materials with both biological and mechanical advantages [28].

To address the challenges of slow vascularization and inadequate mechanical strength in current bone grafts, we aim to design a CPC-based, cell-laden 3D hydrogel composite. The primary goals are to engineer a scaffold with improved load-bearing capacity and to establish a preformed vascular network within it, ultimately boosting the regenerative outcomes in alveolar bone defects.

Methods

Harvesting hPDLSCs from extracted teeth

Periodontal ligament (PDL) tissues were harvested from healthy premolars extracted from patients aged 18–26 years for orthodontic purpose. The hPDLSCs were isolated as described previously [29]. The procedures were approved by the Medical Ethics Committee of Beijing Stomatological Hospital, Capital Medical University (NO. CMUSH-IRB-KJ-PJ-2024-28). The written informed consent was obtained from each participant before the study. The study was carried out in accordance with the Declaration of Helsinki.

The PDL tissues were enzymatically digested with collagenase type I (Gibco BRL, Grand Island, NY, USA) and dispase (Gibco BRL). After digestion, the cell suspension was collected and transferred to culture dishes (Costar, Cambridge, MA, USA) with dulbecco’s modified Eagle’s medium (DMEM, Gibco BRL) supplemented with 1% penicillin/streptomycin (P.S, Gibco BRL) and 20% fetal bovine serum (FBS, Gibco BRL). Upon reaching 70–80% confluence, cells were passaged using 0.25% trypsin-EDTA (Gibco BRL). Cells at passages 3–5 were used in subsequent experiments.

Identification of hPDLSCs

The expression of surface antigen profiles (CD34, CD45, CD90, CD105, and STRO-1) of passage 3–5 cells were analyzed by flow cytometry as described previously [29]. PE-conjugated antibodies against CD34, CD105, and STRO-1 (Thermo Fisher Scientific, Rockford, IL, USA), along with FITC-labeled CD45 and CD90 antibodies (Thermo Fisher Scientific), were employed for immunophenotyping. Cells were digested, centrifuged, resuspended and incubated with antibody. After being washed and resuspended, cell surface antigen expression was tested using a BD Vantage flow cytometer (BD Biosciences).

Culturing of hUVECs

Primary hUVECs were acquired from Sciencell (Carlsbad, CA, USA). The cells were cultured in 10 cm culture dishes with endothelial cell medium (ECM, Sciencell) under standard culture conditions [30]. Subsequent passages were performed when confluency reached 70% to 80%. Cells between passage 2–3 were used in this study.

Fabrication of 3D prevascularized CPC scaffold

The CPC paste was manufactured by InnoTERE GmbH (Radebeul, Germany) and fabricated with the BioScaffolder 2.1 (GeSiM mbH, Radeberg, Germany) operated in a laminar flow workbench [31]. CPC paste was plotted in 60° configuration (the layer orientation changed after every second layer by 60°) and a filament diameter of 300 μm with designed pore size of 300 μm. Scaffolds used for a rat alveolar defect model were plotted a rectangular external structure with a size of 4 mm×2 mm×1 mm. A cylindrical outer geometry with a height of 3 mm and a diameter of 12 mm were used for the following experiments. Three groups were set:

1. Control group (Co-culture only): hPDLSCs and hUVECs were co-cultured in ECM in the absence of scaffold materials, serving as a blank control;

2. CPC-only group (Co-culture + CPC extract): co-cultured cells were maintained in ECM medium supplemented with CPC extract;

3. CPC-GelMA group (Co-culture + CPC-GelMA scaffold): co-cultured cell were seeded in CPC-GelMA scaffold that had undergone 21 days of in vitro maturation to form a prevascularized structure.

The hUVECs and hPDLSCs were detached and mixed at a ratio of 3:1 (hUVECs: hPDLSCs) as described previously [30]. The co-cultured cells were suspended into 5% GelMA solution at a total density of 1 × 10⁶ cells/ml in dark. Then, the suspension was placed in a 1 ml syringe (with 100 μm inner diameter) and injected into the pores inside the 3D printed CPC scaffold. Expose the construct to a 405 nm wavelength light source to facilitate its curing process. The 3D co-culture scaffold was cultured with the ECM for 21 days.

Scanning electron microscopy (SEM) of 3D printed CPC scaffolds

After 21 days of culturing, the composite constructs were fixed overnight. The next day, each cylindrical scaffold was bisected through its mid-plane and the resulting cross-sections were dehydrated and examined under a scanning electron microscope (Quanta 200, FEI, Hillsboro, OR, USA). Filament widths and pore sizes were measured from representative micrographs. Six specimens were analyzed as described previously [19].

Mechanical properties

Three-point flexural tests were performed on the CPC-only group and CPC-GelMA group for mechanical testing [29]. The span is 20 mm and the displacement speed of test head is 1 mm/min. Flexural strength (σ) and elastic modulus (E) were derived from the load–displacement curves. σ = 3FmaxL/2bh², E=(F/d) (L³/4bh³), where F_max is the peak load, L is the span, b and h are the specimen width and thickness, respectively, and F/d represents the slope of the linear-elastic region. Six specimens were tested.

Viability of encapsulated hDPLSCs

After 1, 4, 7 and 14 days of culturing, cellular viability was evaluated using a live/dead viability assay kit (Sigma-Aldrich). Epifluorescence microscope (Sigma-Aldrich) was used for observation. The percentage of live cells was calculated by Image J software (NIH) as described previously [29].

A cell counting kit (CCK-8 assay, Dojindo, Tokyo, Japan) was used to evaluate cell viability at 1, 4, 7, 14 and 21 days. The working solution was prepared with ECM containing 10% CCK-8 solution, followed by a 1-hour incubation at 37 °C. The cell proliferative rate was determined by measuring the absorbance at an optical density of 450 nm using microplate reader (SpectraMax M5, Molecular Devices, Sunnyvale, CA) as described previously [29].

Observing hUVECs via CD31 immunofluorescence staining

After 1, 4, 7 and 14 days of culturing, the resulting microvascular-like structures were then visualized and assessed by CD31 (PECAM-1) immunofluorescence staining as described previously [11]. The samples were fixed and incubated with CD31 mouse mAb (1:500, Cell signaling technology, Pudong District, Shanghai, China) overnight. After washing with PBS, goat anti-mouse IgG (1:1000, goat anti-mouse Alexa Fluor 488, green fluorescence, Cell signaling technology) was applied, followed by DAPI counterstaining (1:1000, Beyotime) at room temperature. The samples were observed with confocal laser scanning microscopy (OLS5100, Olympus, Tokyo, Japan). For each time-point, three random regions from specimens were recorded. Image J (National Health Institute, Bethesda, MA, USA) was used to obtain the vessel length per area and junction number per area (n = 6). Quantification was performed in duplicate by a double blinding protocol.

Rat complete periodontal defect model

The animal protocol was approved by the Committee for Animal Experiments of Beijing Stomatological Hospital, Capital Medical University (NO. KQYY-2026-01-019). Male Sprague-Dawley rats (8 weeks old, 200–250 g, SPF Biotechnology, Haidian District, Beijing, China) were anesthetized with an intraperitoneal injection of Zoletil 50 (Virbac, Carros, France) at 50 mg/kg body weight (n = 6). All surgical procedures were performed under strict aseptic conditions. After sterilization of rats, a critical-size alveolar defect (4 mm × 2 mm × 1 mm) was prepared buccal to the mesial root of the mandibular second molar to establish a rat alveolar bone defect model as described previously [32]. The mesial root was exposed, and all residual periodontal ligament and cementum were meticulously curetted. Scaffolds were press-fit into the defects. Additionally, the rats received flunixin meglumine (2 mg/kg, s.c.; Shanghai Yuanye Bio-Technology, Shanghai, China) and penicillin G benzathine (24,000 IU/kg, i.m.; Pengdi, Henan, China) once daily for 3 consecutive days. Animals were monitored daily for signs of pain, wound integrity, and normal ambulation. Four groups were set:

1. Blank group: underwent the surgical procedure without any scaffold implantation;

2. CPC+GelMA group: received an acellular scaffold, which was preconditioned in ECM for 21 days prior to implantation, serving as the non-prevascularized control;

3. CPC+GelMA-cell group: received the experimental intervention—a prevascularized scaffold seeded with a co-culture of hPDLSCs and hUVECs and matured in ECM for 21 days;

4. Natural Periodontium group: consisted of the contralateral non-operated sites, provided the baseline native tissue.

At 4 weeks post-surgery, six rats per group were euthanized. After deep anesthesia with an overdose of intraperitoneal pentobarbital (150 mg/kg), animals were sacrificed by CO₂ asphyxiation followed by exsanguination via bilateral thoracotomy. The implants were then retrieved and immediately fixed in 4% paraformaldehyde at 4 °C for 24 h.

Micro computed tomography (Micro-CT)

Micro-CT (SkyScan 1276, Bruker BioSpin, Germany) was employed to scan the bone-defect regions and perform 3D reconstructions [29]. The ROI was defined as the buccal alveolar bone encircling the mesial root of the mandibular second molar and encompassed a volume of 4 mm×2 mm×1 mm. Bone thickness, bone volume fraction and scaffold degradation rate was calculated to quantify the percentage of newly formed bone within the defect. Additionally, a cross-sectional slice at the mid-root level of the mesial root was selected for qualitative evaluation.

Histomorphometric analyses

Specimens were decalcified in 10% ethylene diamine tetraacetic acid (EDTA) (Solarbio Science & Technology, Beijing, China) for 2 months, processed routinely, and embedded in paraffin. The central part of the implant and defect was cut into 5 μm-thick sections for hematoxylin and eosin (H&E) staining, Masson’s staining and immunohistochemistry (IHC) staining.

H&E staining

The samples were decalcified and embedded in paraffin. Serial 5 μm-thick sections were prepared and stained with H&E. New bone area, total defect area and the number of new vessels were quantified in each section by image J as described previously [19]. New bone area fraction was calculated by dividing the area of new bone with the area of the total defect. New vessels density was expressed as the number of new vessels divided by total defect area (n = 6).

Masson’s staining

Deparaffinized sections were stained with Weigert’s iron hematoxylin, sequentially incubated in Biebrich scarlet-acid fuchsin and phosphomolybdic/phosphotungstic acid, and finally differentiated in aniline blue to visualize collagen. Blue-stained bone area and total defect area were measured in each section [30]. Osteoid area fraction was calculated by dividing the area of blue-stained bone with the area of the total defect (n = 6).

IHC staining

Immunodetection of human CD31 and Runx2 was performed on 5 μm paraffin sections. Sections were incubated with rabbit anti-human CD31 (1:500, Abcam) and anti-human Runx2 (5 µg/mL, Abcam), followed by HRP-conjugated secondary antibody (1:500, Abcam). Signals were developed with DAB and counterstained with hematoxylin. For histomorphometry, one mid-sagittal section per animal was analyzed (n = 6). The density of TRAP-positive cells was calculated. CD31 and Runx2 expression was assessed as integrated optical density (IOD) from 6 random fields per section as described previously [32].

Subcutaneous transplantation model in nude rats

For subcutaneous transplantation, CPC+GelMA group and CPC+GelMA-cell group were selected. After induction of general anesthesia and sterile preparation, scaffolds were implanted into subcutaneous pockets on the back nude rats (8 weeks, 200–250 g) as described previously [33]. After 6 weeks of implantation, all scaffolds were obtained and fixed for stereomicroscope observations (Olympus, Tokyo, Japan) and H&E staining. Vessels were enumerated in three randomly selected fields per section, and vascular density was calculated as vessel number per unit area.

Statistical analysis

All statistical analyses were conducted using SPSS 22.0 (IBM Corp., Armonk, NY, USA). Data are expressed as mean ± standard deviation (SD). Group comparisons were performed by one-way analysis of variance (ANOVA) followed by Tukey’s post-hoc test for multiple comparisons. p-value < 0.05 was considered statistically significant.

Results

Identification of hPDLSCs

Figure 1. A plots flow cytometry result of isolated hPDLSCs. CD105, CD90, STRO-1 were highly expressed to 99.6%, 76.4% and 88.6%. While, CD34 and CD45 were weakly expressed to 0.3% and 0.6%, respectively.

Fig. 1.

A Flow cytometry results of the isolated hPDLSCs: 99.6% CD105-positive cells, 76.4% CD90-positive cells, 88.6% STRO-1-positive cells, 0.3% CD34-positive cells, 0.6% CD45-positive cells. B-C 3D-printed CPC scaffold. D 3D-prevascularized CPC scaffold

Fabrication of 3D prevascularized CPC scaffold

Figure 1. B, C illustrates the fabrication of a 3D grid-like CPC scaffold. Figure 1. D plots the injection of an hPDLSCs–hUVECs-laden GelMA hydrogel into the 3D-printed CPC scaffold to generate a 3D prevascularized CPC construct.

Physical properties of scaffolds

Figure 2. A shows CPC filaments intersecting at defined angles. Cross-sectional images (Fig. 2. B, C) reveal spherical crystalline structures and abundant pores within the CPC matrix. Figure 2. D, E reveal the morphology and distribution of co-cultured cells (hPDLSCs and hUVECs) within the pores. The letter “c” indicates the cells, which are seen to be firmly attached to the scaffold. Quantitative measurements yielded filament diameters of 285.21 ± 1.4 μm, pore sizes of 302.31 ± 3.4 μm, and filament intersection angles of 60.12 ± 0.4° for the 3D-printed CPC scaffold.

Fig. 2.

Physical properties of 3D-printed CPC constructs. A-E SEM images of 3D-printed CPC constructs. C is a higher magnification image of the red dotted frame in (B). E is a higher magnification image of the blue dotted frame in (D), and cells (labeled as C) attached well to the scaffold. F Flexural strength. G Elastic modulus. All data are presented as mean ± SD (n = 6)

The flexural strength (Fig. 2. F) and elastic modulus (Fig. 2. G) of both the CPC-only group and the CPC-GelMA group were significantly higher than those of cancellous bone. However, no significant difference was observed between the two experimental groups. Specifically, the CPC-only group exhibited a flexural strength of 6.98 ± 0.24 MPa and an elastic modulus of 0.40 ± 0.02 GPa, while the CPC-GelMA group showed values of 6.78 ± 0.68 MPa and 0.41 ± 0.02 GPa, respectively.

Viability and cell proliferation of cells co-cultured within GelMA-CPC constructs

Live/dead staining images for the encapsulated cells in CPC-GelMA group (Fig. 3. A) represented numerous live cells (green staining) and a few dead ones (red staining) at each time points tested. At day 1, cells were embedded within the hydrogel, predominantly spherical and clustered, with relatively low cell numbers visible. From day 4 to 7, cells began to aggregate, the number of released cells increased continuously, which extended well, showing spindle or polygonal shape. At day 14, cells were evenly distributed throughout the hydrogel. At day 21, all cells had fully extended into long spindle shapes, forming pseudopodia and interconnecting into branched, network-like structures. As shown in Fig. 3. B, the percentage of live cells in the CPC-GelMA group exceeded 80% on day 1 and significantly increased to approximately 90% from day 4 to 21. From day 7 to 21, no significant difference in cell viability was observed among the three groups. The CCK-8 assay (Fig. 3. C) further revealed that co-cultured cells in the CPC-GelMA group exhibited a 9.2-fold increase in proliferation from day 1 to 21. Notably, on day 21, the proliferative capacity of cells in the CPC-GelMA group was significantly higher than that in the other two groups (p < 0.05).

Fig. 3.

Viability and cell proliferation of co-cultured cells versus time. A Live/dead staining images of cells encapsulated in the CPC-GelMA group at various time points. Viable cells are stained green, while dead cells appear red. B Quantitative analysis of live cell percentages over 21 days. C Cell proliferation assessed by CCK-8 assay over 21 days. All data are presented as mean ± SD (n = 6)

CD31 immunofluorescence staining

Co-cultures of hPDLSCs and hUVECs from all groups were subjected to CD31 immunofluorescence staining for visualization (Fig. 4. A–F). HUVEC membranes were stained green for the endothelial marker CD31, while nuclei were counterstained blue with DAPI. Branch-like structures gradually increased from day 14 to day 21. The co-cultured cells within the GelMA hydrogel progressively formed vessel-like networks with prolonged culture. Both the number of vessel junctions (Fig. 4. G) and the cumulative vessel-like branch length (Fig. 4. H) in the CPC-GelMA group increased from day 14 to day 21. Moreover, on day 21, these parameters were significantly higher than those in the other two groups (p < 0.05). Specifically, the junction density in the CPC-GelMA group reached 36.53 ± 2.20 junctions/mm² on day 21, representing a 2.3-fold increase from day 14. Meanwhile, the cumulative vessel length showed a 4.5-fold increase over the same period, reaching 26.25 ± 1.87 mm/mm² by day 21.

Fig. 4.

In vitro CD31 immunostaining of hPDLSCs–hUVECs co-cultured in all groups versus time. A–F Representative immunofluorescence images of branch-like structures formed by co-cultured cells at day 14 and day 21. HUVECs were identified by positive CD31 staining (green) on the cell membrane, while cell nuclei were counterstained with DAPI (blue). HPDLSCs are visible as nuclei stained with DAPI only, without membrane green staining. G Quantitative analysis of cumulative vessel length in the co-culture system at day 14 and day 21. H Quantitative analysis of the number of vessel junctions at day 14 and day 21. All data are presented as mean ± SD (n = 6)

Complete periodontal tissue regeneration in vivo

At 4 weeks post-surgery, rat alveolar bone defects were examined by Micro-CT, including both cross-sectional views and 3D reconstructions (Fig. 5. A). Figure 5. B illustrates the surgical procedure for establishing the rat alveolar defect model. The buccal bone thickness adjacent to the mesial root of the second molar in all groups were observed. As shown in Fig. 5. C, the CPC+GelMA-cell group (0.549 ± 0.01 mm) showed the greatest new formed bone thickness, which were 1.3 folds that of CPC-GelMA group (0.421 ± 0.02 mm) and 2.57 folds that of the blank group (0.214 ± 0.02 mm). The new bone area fraction (Fig. 5. D) was also highest in the CPC+GelMA-cell group (44.33 ± 3.01%), reaching 1.23 folds that of the CPC-GelMA group (35.92 ± 2.52%) and 1.43 folds that of the blank group (30.99 ± 1.78%). At week 4, as shown in Fig. 5. E, no significant difference in degradation rate was observed between the CPC-GelMA group (11.57 ± 0.93%) and the CPC+GelMA-cell group (11.71 ± 0.87%).

Fig. 5.

Micro-CT evaluation of rat alveolar-bone defect repair. A Cross-sectional views of the defect region in each group. White line displays defect boundary. B Schematic illustration of the surgical procedure for establishing the rat periodontal defect model. C Newly formed bone thickness within the defect. D New bone area fraction within the defect. E Degradation rate of scaffolds. All data are presented as mean ± SD (n = 6)

For H&E staining, all groups newly formed bone (NB) and blood vessels (V) in defects, with no evident inflammation or immune reaction (Fig. 6. A). New bone with a typical organized bone morphology was formed. New blood vessels were observed around the new bone. Osteoblasts with blue cytoplasm and round-to-oval nuclei lined the surfaces of the new bone. Compared with the Natural Periodontium group, the blank group showed only sparse alveolar bone regeneration on the buccal side of the mesial root of the second molar. The CPC+GelMA group exhibited moderate, irregularly distributed new bone and vessels. The CPC+GelMA-cell group displayed the most abundant and irregularly distributed new bone and vessels, with active osteoblasts. As shown in Fig. 6. B, CPC+GelMA-cell group (57.99 ± 4.78%) formed the highest new bone area that was 1.37 folds that of the CPC+GelMA group (42.26 ± 5.48%) and 2.29 folds that of the blank group (25.33 ± 5.25%). The new vessel density (Fig. 6. C) in CPC+GelMA-cell group (41.32 ± 2.95 vessels/mm²) were about 2.65 folds that of CPC-GelMA group (15.59 ± 1.47 vessels/mm²) and 6.2 folds that of the Blank group (6.66 ± 1.07 vessels/mm²).

Fig. 6.

Complete periodontal tissue regeneration in vivo. A H&E and Masson staining images of alveolar bone defects in rats. R: root, NB: newly formed bone, V: vessel. B Semi-quantitative analysis of fraction of new bone areas. C Semi-quantitative analysis of new vessel density. D Semi-quantitative of osteoid bone. All data are presented as mean ± SD (n = 6)

Masson staining effectively reveals the maturity of collagen within bone tissue: mature collagen stains red, whereas newly formed bone or immature bone stains blue (Fig. 6. A). Relative to the Natural Periodontium group, all three intervention groups exhibited immature bone within the defect. The Blank group showed sparse blue trabeculae embedded in abundant fibrous tissue. In contrast, the CPC+GelMA-cell group and the CPC+GelMA group produced extensive, irregularly arranged immature bone that appeared predominantly blue, with scattered foci of mature red-stained matrix. The area of blue-stained immature bone (Fig. 6. D) in CPC+GelMA-cell group was highest (56.38 ± 3.47%), which was 1.29 folds that of CPC+GelMA group (43.59 ± 2.77%) and 2.38 folds that of Blank group (23.70 ± 1.33%).

For immunohistochemical images of rat alveolar bone defects, both angiogenic marker CD31and osteogenic marker Runx2 were expressed in the periodontal defects in all groups at 4 weeks after transplantation, compared with the negative-control group (Fig. 7. A). The proportion of CD31-positive cells in the CPC+GelMA-cell group (2.39 ± 0.03 per slice) was also much higher than those in the CPC+GelMA (2.06 ± 0.04 per slice) and Blank (1.64 ± 0.03 per slice) groups, reaching 1.16-fold and 1.46-fold higher values (Fig. 7. B). Moreover, the newly formed periodontal tissues in the CPC+GelMA-cell group contained the highest proportion of Runx2-positive cells (2.471 ± 0.022 per slice), being 1.17-fold and 1.91-fold greater than in the control (2.112 ± 0.038 per slice) and blank (1.675 ± 0.027 per slice) groups, respectively (Fig. 7. C).

Fig. 7.

Complete periodontal tissue regeneration in vivo. A Immunohistochemical staining of CD31 and Runx2 implantation. B Semi-quantitative analysis of CD31-positive cell numbers per slice of (A). C Semi-quantitative analysis of Runx2-positive cell numbers per slice of (A). All data are presented as mean ± SD (n = 6)

Subcutaneous transplantation in vivo

Figure 8. A plots the images of stereomicroscopic view and H&E staining after subcutaneous implantation in nude rats at 6 weeks. Neovessels were visible on the surface of both groups. Histological evaluation revealed newly formed vessels within the macropores of both scaffolds. The vessels density of CPC+GelMA-cell group was markedly higher than CPC+GelMA group. Morphometric analysis (Fig. 8. B) demonstrated that the vessels density in the CPC+GelMA-cell group reached 81.33 ± 3.54 vessels/mm², a 1.93-fold increase over the 42.21 ± 5.84 vessels/mm² measured in the CPC+GelMA group (p < 0.05).

Fig. 8.

In vivo subcutaneous transplantation in rats at 6 weeks (A) Stereomicroscopic view and H&E staining images of CPC+GelMA and CPC+GelMA-cell group. B Semi-quantitative analysis of the vessel density of CPC+GelMA and CPC+GelMA-cell groups. All data are presented as mean ± SD (n = 6)

Discussion

The present study is the first to combine a 3D printed CPC framework with a GelMA-hPDLSCs-hUVECs system to fabricate a prevascularized CPC scaffold for alveolar bone regeneration. The hypotheses were proven that the novel construct had good mechanical properties, pore structure, biocompatibility and angiogenic capability in vitro. Compared with non-prevascularized CPC controls, the prevascularized scaffold significantly enhanced new bone and vessels formation in vivo.

Our results showed that the isolated cells highly expressed STRO-1, CD90, and CD105, with low expression of CD34 and CD45, consistent with the characteristics of MSCs and thus considered as hPDLSCs [34]. Yeasmin et al. [35] found that hPDLSCs can secrete angiogenesis-related factors and provide stable support for endothelial cell networks. These findings further demonstrated that hPDLSCs, as seed cells, have significant vascularization potential in co-culture systems.

MSCs not only provide trophic support to ECs but are also reciprocally activated by EC-derived signals, leading to further upregulation of their own growth factor secretion [35]. Monocultured MSCs are unable to self-assemble into microvascular structures, whereas co-culture with ECs rapidly initiates capillary-like network formation [10]. In the present study, hUVECs and hPDLSCs were co-seeded at a fixed ratio of 3:1. It has been demonstrated that a 1:1 ratio of hUVECs to MSCs generates stable vascular networks [15]. Conversely, hPDLSCs have been reported to proliferate significantly faster than hUVECs [11]. To offset this imbalance, initial EC proportions have been raised to 3:1 or even 5:1 (hUVECs: hPDLSCs) in prior work [36]. Longitudinal profiling confirms a progressive decline in the relative abundance of hUVECs during extended co-culture [37]. Our previous study has claimed that a 2D prevascularization network was successfully established on CPC scaffolds using a 3:1 ratio of hUVECs to hPDLSCs, which proved essential for maintaining the osteogenic potential of the scaffold material [11]. The critical importance of this specific ratio has been further validated by subsequent studies demonstrating that hPDLSC-hUVEC co-cultures at the same 3:1 proportion effectively promoted vascularized bone regeneration in critical-sized calvarial defect models [30], thereby confirming that this cellular configuration concurrently supports both angiogenesis and osteogenesis. Therefore, the 3:1 seeding ratio chosen here is expected to counterbalance the differential expansion rates and stabilize the two populations as the assay proceeds. Nevertheless, systematic optimization studies are still required to definitively establish the ideal hUVECs: hPDLSCs numerical ratio for maximal vasculogenic output in co-culture systems.

GelMA hydrogels provide a cytocompatible extracellular environment, yet their bioactivity is dictated by GelMA concentration, photoinitiator content, and UV dose [38, 39]. GelMA below 5% (w/v) remains high biocompatibility but lacks adequate mechanical integrity [40]. Comparative studies reveal that 5% (w/v) GelMA yields larger pores than its 10% (w/v) counterpart, facilitating Ca²⁺ flux and diffusion without compromising viability [41]. Therefore, 5% (w/v) GelMA hydrogel was selected as the cell-laden matrix in the present study. CCK-8 assays revealed a robust and sustained proliferative profile of the co-cultured cells within the 5% GelMA matrix. Additionally, the live cell percentage was 80% on day 1. The brief drop in live-cell number on day 1 is likely attributable to residual unreacted functional groups or small-molecular by-products released during GelMA cross-linking [42]. Future work should therefore optimize cross-linking parameters such as lowering photoinitiator concentration or shortening UV exposure to curb radical formation and improve early survival. Moreover, GelMA’s intrinsic drug-loading capacity and interconnected porosity enable sustained release of bioactive cargos [43]. With this benefit, future studies should encapsulate pro-angiogenic factors within GelMA-based co-culture systems to enhance vascularization.

Nevertheless, the inherently low mechanical strength of GelMA hydrogel limits its standalone use in load-bearing alveolar defects [29]. Consequently, recent studies have focused on integrating hydrogels with high-strength scaffolds to generate composites that unite biocompatibility with sufficient mechanical competence [29]. CPC paste used in the present study has been extensively validated as a biocompatible and osteoconductive bone substitute [44–48]. Therefore, we combined GelMA hydrogel co-encapsulating hPDLSCs and hUVECs with a 3D-printed CPC scaffold to achieve sufficient mechanical competence to withstand physiological masticatory loads. Flexural strength and elastic modulus are critical mechanical indices for evaluating scaffold performance [49]. In human cancellous bone, flexural strength is approximately 3.5 MPa and the elastic modulus is 0.30 GPa, respectively [50]. Three-point bending test indicates that both CPC-only and CPC-GelMA group showed better flexural strength and elastic modulus than those of cancellous bone. Thus, CPC offers adequate mechanical strength and dimensional stability for load-bearing alveolar repairs.

In 3D printed scaffolds designed for vascular regeneration, filament diameter, pore size and filament intersection angle are determinants of pro-angiogenic performance [51]. Pores between 200 and 400 μm allow efficient cell migration and nutrient exchange [52]. Pores below 100 μm hinder cell infiltration and capillary ingrowth [53]. Intrinsic micropores of CPC are typically less than 50 μm and poorly interconnected, which limiting the migration and interaction of cells [54, 55]. Given that human micro-vessels range from 5 to 200 μm in diameter [51], the 302.31 ± 3.4 μm pore size generated in the printed CPC scaffold readily accommodate nascent micro-vessels and their interconnection, facilitating 3D vascular network assembly. Notably, 300 μm pores have been specifically identified as optimal for alveolar bone regeneration [19]. Moreover, scaffolds of 300 μm-pore size honeycomb architecture concurrently drive bone and vessel formation in vivo [56]. Moreover, similar filament diameter and pore size have been shown to homogenize stress distribution and markedly improve scaffold strength [57]. In the present work, the printed CPC displays comparable filament and pore sizes, providing a regular architecture conducive to cell attachment and growth. In addition, the intersection angle between filaments influences both mechanical integrity and cell alignment [23]. Previous study demonstrated that a 60° intersection angle yielded higher compressive modulus and strength than a 45° pattern [58]. Scaffolds printed with a 60° filament intersection angle likewise exhibited superior mechanical performance in rat femoral defects [44]. This advantage is attributed to the angle’s balanced stability, which simultaneously supports multidirectional cell signaling and nutrient exchange [44]. The CPC scaffolds described here were printed at 60.12 ± 0.4°, closely matching the reported optimum. Nevertheless, the definitive intersection angle for 3D-printed CPC remains to be established and demands systematic exploration.

CD31 immunostaining was used to quantify in vitro angiogenesis of co-cultured hPDLSCs-hUVECs. Cumulative vessel length and junction number are standard metrics of neovascularization: length reflects network expansion, whereas junction density mirrors architectural complexity and functional anastomosis [59]. Confocal imaging revealed microcapillary-like networks that elongated and interconnected over time. By day 21, the 3D CPC-GelMA co-culture system demonstrated significantly higher cumulative vessel length and junction number compared to the 2D co-culture systems (CPC-only and Control groups), highlighting the enhanced pro-angiogenic capacity of the 3D architecture. Previous study demonstrated that prevascularized scaffolds can set up a functional blood supply by speeding up vessel formation and stably connecting with host blood vessels, which greatly improving implant performance [60]. In vitro prevascularization offers a minimally invasive, surgeon-friendly approach without secondary surgery or donor-site morbidity, making it more clinically attractive [61]. Researchers first co-cultured endothelial cells with osteoprogenitor cells and generated a 3D vascular network in vitro [62]. Inspired by this, we built a prevascularized scaffold in a 3D co-culture system and tested its angiogenic potential by subcutaneous implantation in nude rat. Because nude rats are immunocompromised, xenograft rejection is sharply reduced, which improving graft survival and the likelihood of vessel formation [63]. At week 6, vessel density within the scaffolds reached 81.33 ± 3.54 vessels/mm². This finding aligns with previous reports showing that fully prevascularized implants achieve approximately 80% vascular coverage by 6 weeks, compared to only 50% in non-prevascularized controls [60]. The enhanced perfusion is attributed to pre-formed vessels acting as “highways” that guide and accelerate capillary sprouting, ultimately leading to a denser and more stable microvascular network throughout the scaffold [60].

Alveolar bone regeneration was evaluated in all groups in vivo. According to Raposo-Amaral [64], the selection of 4 weeks as the observation endpoint for animal alveolar bone defect experiments is sufficient to capture early key events in alveolar bone healing and validate the pro-angiogenic effects of the scaffolds. Therefore, the present study used a sacrifice time point of 4 weeks. By week 4, new bone was evident in every defect. However, the CPC+GelMA-cell group produced significantly more bone than either CPC+GelMA or Blank groups. Micro-CT gave a mineralized-bone fraction of 44.33 ± 3.01% in the CPC+GelMA-cell group—23% higher than CPC+GelMA group. Similarly, histomorphometric analysis of H&E-stained sections showed that the new bone area in the CPC-GelMA-cell group was 1.37-fold greater than that in the CPC-GelMA group. The discrepancy reflects the fact that Micro-CT records only mineralized tissue, whereas H&E staining also captures unmineralized osteoid [65]. Nevertheless, H&E cannot distinguish immature bone from mature lamellar bone, so Masson staining was additionally performed to verify the maturity of the regenerated tissue [66]. Masson staining reliably visualizes collagen remodeling in bone tissues [66]. Newly formed bone is rich in type I collagen and stains blue. As mineralization proceeds, the stained color shifts toward red. This color transition is routinely exploited to gauge maturation, with the nascent-to-mature conversion typically occurring between 4 and 8 weeks post-operatively [66]. We therefore harvested specimens at week 4 and used the area fraction of blue-stained bone as a surrogate for new bone formation. Masson images revealed that the CPC+GelMA-cell group contained the largest amount of immature bone, indicating that the defect area was undergoing active osteogenesis and that the experimental intervention accelerated this process. Due to Masson and H&E staining provide limited information on cellular and vascular components, IHC staining was employed for a more comprehensive evaluation. CD31, an endothelial-specific marker, was used to quantify angiogenesis [67], whereas Runx2, the master transcription factor driving osteoblast differentiation, was used to assess osteoblastic activity and maturity [68]. Semi-quantitative IHC demonstrated that Runx2 expression in the CPC+GelMA-cell group was 1.17-fold that of CPC+GelMA group, corroborating the superior osteogenic potential observed histologically. Similarly, a previous study demonstrated that scaffold pre-seeded with cells generated significantly more new bone than its cell-free counterpart, at 4 weeks in a rat cranial bone defect model [69].

Moreover, a significant positive correlation between osteogenesis and angiogenesis was evident [69]. H&E morphometry revealed a neovessel density of 41.32 ± 2.95 vessels/mm² in the CPC+GelMA-cell group, 2.65-fold that of CPC+GelMA group, accompanied by a 16% increase in CD31 immunostained area. These data corroborate earlier reports that bone formation rate rises in parallel with vascular ingrowth [70, 71].

This coupling is likely initiated by the hypoxic milieu that develops immediately after scaffold implantation. Under low oxygen tension, co-cultured cells activate hypoxia-inducible factor-1α (HIF-1α) pathway, which transcriptionally up-regulates vascular endothelial growth factor (VEGF) and fibroblast growth factor-2 (FGF-2). While promoting angiogenesis, these cytokines simultaneously deliver the nutrients and osteoinductive signals required for effective bone repair [72, 73]. Additionally, PI3K/Akt [74] and MAPK [75] signaling pathways participate in hypoxia-induced angiogenesis, which can modulate cell proliferation, survival, and differentiation to accelerate bone repair and regeneration.

In addition to the contribution of co-cultured cells, the intrinsic properties of the scaffold itself critically influence osteogenesis and angiogenesis [76]. In the present study, the CPC+GelMA group exhibited substantially more de novo bone than the empty defect (Blank group). This effect is attributable to the osteoinductive capacity of CPC, which sequesters endogenous growth factors and directs multipotent stem cells toward an osteoblastic phenotype [77]. In earlier work, traditional 2D CPC placed in cranial defects for 8 weeks regenerated only 13.89 ± 2.95% new bone [78], while traditional 2D CPC packed into alveolar socket yielded less than 5% repair by week 12 [79]. In contrast, micro-CT of our 3D CPC scaffold revealed 35.92 ± 2.52% new alveolar bone within only 4 weeks. This pronounced disparity is presumably attributable to differences in defect site and internal scaffold structure [80].

In addition, porogen-leavened macroporous CPC [81], collagen-sponge CPC [79] and CPC fortified with SP [79] or rhBMP-2 [81] have respectively doubled to quintupled new bone output within 4 weeks. Thus, compositional and architectural refinement can readily enhance CPC-mediated alveolar regeneration.

This study has several limitations that should be considered when interpreting the results. First, the absence of a hUVECs-only control group limits our ability to fully attribute the observed angiogenic effects to cell–cell synergy. Nevertheless, our previous study has consistently demonstrated that hUVECs alone exhibit lower angiogenic capacity compared to hUVECs-hPDLSCs co-cultures, and that stable vascular structures are rarely formed by hUVECs alone under identical conditions [11]. Future studies should include both single-cell control groups to more precisely delineate the respective contributions of each cell type. Second, in this study, we employed nude rats (immunodeficient model) for subcutaneous implantation, primarily to avoid interference from immune cells during the angiogenesis process, thereby enabling a clearer analysis of the direct contributions of the scaffold material and co-cultured cells to vascular network formation. However, this model choice also presents limitations for clinical translation. Due to the lack of T cell-mediated adaptive immune responses in nude rats, the angiogenesis induced after scaffold implantation may not undergo normal immune screening and remodeling processes. Consequently, the rapid vascularization and new bone formation observed in the immunodeficient model may be attenuated or even reversed in immunocompetent clinical environments due to immune rejection responses or inflammatory osteoclast activation. Future studies should focus on validation in immunocompetent large animal models or humanized immune reconstruction models to more comprehensively evaluate the pro-angiogenic capacity, osseointegration efficacy, and long-term functional remodeling of this bone repair strategy under normal immune microenvironments. This will provide more reliable and representative data to support clinical translation.

Although CD31 immunofluorescence staining and histological sectioning are standard methods for angiogenesis assessment, they cannot provide 3D spatial information of vascular networks. Future studies should utilize micro-CT angiography for more comprehensive quantitative evaluation.

In the alveolar bone defect experiment, 4 weeks was selected as the observation endpoint. While this time point was sufficient to capture early healing events, this single time point design does not allow assessment of the long-term stability of newly formed bone, complete scaffold degradation, or bone remodeling outcomes. Future studies, particularly in large animal models, should incorporate multi-timepoint designs (e.g., 2, 4, 8, and 12 weeks) with long-term follow-up to establish complete scaffold degradation profiles and to comprehensively evaluate the dynamic process of alveolar bone regeneration.

Moreover, although this study ensured consistency in the initial seeding ratio (hUVECs: hPDLSCs = 3:1), dynamic tracking and quantitative evaluation of the actual cell ratio during and after co-culture were not performed. Differences in proliferation rates, apoptotic sensitivity, and responses to the microenvironment between the two cell types may lead to ratio shifts, thereby affecting the efficiency and stability of vascular network formation. Future studies should employ fluorescent labeling techniques combined with flow cytometry or live-cell imaging systems to monitor real-time changes in cell ratio, spatial distribution, and interactions within the co-culture system, providing more precise guidance for optimizing cell ratios in prevascularized constructs.

Despite these limitations, our in vitro and in vivo data provide valuable preliminary evidence supporting the potential of prevascularized CPC-GelMA scaffolds for alveolar bone repair. Future research will systematically address these limitations to further validate and optimize this strategy.

Conclusion

This study developed a novel 3D printed CPC scaffold to establish a prevascularized 3D construct for alveolar bone engineering. The resulting construct combined adequate mechanical strength for load-bearing oral applications with an interconnected porous architecture favorable for alveolar bone tissue engineering. The delivered hPDLSCs-hUVECs exhibited excellent cell proliferation and angiogenic potential in GelMA hydrogel in vitro. When implanted in vivo, the prevascularized CPC+GelMA-cell group elicited significantly greater neovascularization and accelerated alveolar-bone regeneration compared with the cell-free CPC+GelMA control. The intrinsic properties of CPC also enhance osteogenesis and angiogenesis. Therefore, the novel CPC-GelMA-hPDLSCs-hUVECs construct is highly promising for concurrent bone and vascular regeneration in dental applications.

Authors’ contributions

Y. Sun conducted the investigation, performed visualization and validation. Z. Zhao contributed to methodology and data curation, and drafted the original manuscript. Q. Qiao was responsible for visualization and methodology. conceptualized the study. W. Yu contributed to validation. Y. Bai participated in writing - review and editing, project administration, funding acquisition. All authors read and approved the final manuscript.

Funding

This study was supported by National Natural Science Foundation of China (No. 82301117), the Innovation Research Team Project of Beijing Stomatological Hospital, Capital Medical University (Grant No. CXTD202203) and Beijing Stomatological Hospital, Capital Medical University Young Scientist Program (No. YSP202510).

Data availability

All data generated or analysed during this study are included in this published article [and its supplementary information files].

Declarations

Ethics approval and consent to participate

Capital Medical University Ethics Committee approved this study (CMUSH-IRB-KJ-PJ-2024-28). All participants agreed to participate in this study and signed an informed consent form. All animal experiments were approved by the Committee for Animal Experiments of Beijing Stomatological Hospital, Capital Medical University (NO. KQYY-2026-01-019).

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.