Aligned cryogel fibers incorporated 3D printed scaffold effectively facilitates bone regeneration by enhancing cell recruitment and function

Abstract

Reconstructing extensive cranial defects represents a persistent clinical challenge. Here, we reported a hybrid three-dimensional (3D) printed scaffold with modification of QK peptide and KP peptide for effectively promoting endogenous cranial bone regeneration. The hybrid 3D printed scaffold consists of vertically aligned cryogel fibers that guide and promote cell penetration into the defect area in the early stages of bone repair. Then, the conjugated QK peptide and KP peptide further regulate the function of the recruited cells to promote vascularization and osteogenic differentiation in the defect area. The regenerated bone volume and surface coverage of the dual peptide-modified hybrid scaffold were significantly higher than the positive control group. In addition, the dual peptide-modified hybrid scaffold demonstrated sustained enhancement of bone regeneration and avoidance of bone resorption compared to the collagen sponge group. We expect that the design of dual peptide-modified hybrid scaffold will provide a promising strategy for bone regeneration.

Bone regeneration scaffold enables rapid cell recruitment and functional regulation.

INTRODUCTION

Now, the most commonly used clinical solutions for large cranial bone defects include the use of autologous vascularized bone graft (1, 2), polyetheretherketone (PEEK) cranial implant (3), and titanium/titanium alloy mesh (4, 5). However, these modalities have well-established limitations after years of clinical use. The autologous vascularized bone graft remains the clinical gold standard for reconstructing large cranial defects. However, several drawbacks exist. For instance, the mismatch between the graft shape and cranial defect contours can lead to poor aesthetic outcomes. Autologous grafts undergo substantial resorption over time, resulting in contour irregularities and the need for secondary renovation (1, 2). In addition, the donor site also suffers excruciating pain and infection risks. Although synthetic PEEK implants and titanium meshes can avoid donor site morbidity, they carry a high complication rate. Exposure and infection frequently necessitate removal (6), and their rigidity causes discomfort (7). Moreover, the limited osseointegration caused by nondegradation and dense structure can cause implant failure (8). Therefore, there is an unmet need for resorbable cranial bone grafts that can preclude donor site morbidity and disadvantages associated with allografts and existing cranioplasty materials.

Collagen sponge is a commonly used biodegradable bone repair material, and it has been widely used as osteoconductive bone graft substitutes to promote bone regeneration in clinics (9, 10). The highly porous structure allows cellular infiltration, vascularization, and bone regeneration (11). However, collagen sponges lack inherent osteoinductivity (9, 12). Moreover, the degradation of collagen sponge is before sufficient new bone formation, resulting in new bone formation that cannot be continuously induced, and the formed osteoid tissue may be resorbed. Ongoing challenges of biodegradable bone repair material include balancing degradation with bone ingrowth (13, 14), improving mechanical properties (15, 16), and incorporating bioactive factors to enhance overall osteogenic performance (17). In our previous study, we developed a class of cell- and therapeutic agent–free scaffolds made of polycaprolactone electrospun nanofibers with predesigned three-dimensional (3D) aligned nanotopography (18) based on fully considering the advantages and disadvantages of collagen sponge bone repair materials, which have higher porosity, longer degradation time, and can enable rapid cell recruitment either from the surrounding to the center or from the bottom to the top, resulting in rapid bone regeneration (19). It inspires us with the importance of cell recruitment in the early stages of bone repair.

Among various additive manufacturing processing technologies, 3D printing is the most widely used technology for preparing bone repair materials. However, the 3D printed scaffolds (PSs) have low cell seeding efficiency and cannot induce cell migration. Here, in the present study, we hypothesized that introducing aligned fibers within the 3D PS may further accelerate cell infiltration into the defect area (18, 20), thereby boosting tissue regeneration. Here, one purpose of this work is to study how to introduce aligned fibers inside the 3D PS (Fig. 1A). Another issue that needs to be solved is how to regulate the function of these recruited cells, such as enhancing cells’ osteoinductive ability (Fig. 1B). To regulate the cell behaviors of the recruited cells, the QK peptide and KP peptide were used to modify the aligned fibers and 3D PS, respectively. The QK peptide is an engineered vascular endothelial growth factor (VEGF)–mimicking peptide, which can enhance angiogenesis to provide more nutrients for these recruited cells during bone healing (21, 22). The KP peptide is a bone morphogenetic protein 2 (BMP-2) mimicking peptide, which can significantly enhance new bone formation (22–24). Therefore, the introduction of aligned fibers within the 3D PS and the cell recruitment ability of aligned fiber-modified 3D PS were explored. In addition, the pro-angiogenesis and pro-ossification ability of the QK peptide and KP peptide conjugated hybrid scaffold were also investigated in this study.

Fig. 1. ACFs incorporated 3D PS effectively facilitates bone regeneration by enhancing cell recruitment and function.

(A) Fabrication procedure of ACFs incorporated into a 3D PS and cell recruitment methods of PS/ACFs in the cranial bone defect. (B) PS/ACFs functionalized with KP and QK peptides facilitate critical-sized cranial bone defect repair in rats. Red arrows indicate the cell recruitment, blue arrows show the orientation of the newly formed collagenous fibers, and orange arrows indicate the osteogenic differentiation. BMSC, bone marrow mesenchymal stromal cell.

RESULTS

Fabrication and characterization of ACFs incorporated 3D PS

The aligned cryogel fibers (ACFs) incorporated a 3D PS, which was prepared by 3D bioprinting in combination with unidirectional freeze casting. Briefly, a 3D lattice scaffold was printed using 5 weight % (wt %) gelatin-methacrylamide (Gel-MA). The lyophilized PS was submerged into a solution containing 0.18 wt % Gel-MA and 0.02 wt % gelatin for cryogel fiber formation. Then, the soaked PS was transferred to a − 80°C precooling stainless steel plate for unidirectional freezing and chemical crosslinking over 24 hours (Fig. 2A). The PS exhibited a colorless, transparent lattice structure, while the PS/ACFs showed a similar morphology with some air bubbles within the lattice holes. After lyophilization, the holes of PS/ACFs were filled with white fibers in contrast to the PS scaffold (Fig. 2B). The presence of air bubbles in PS/ACFs before freeze drying corresponded to the formation of ACFs within the pores. The preparation of random cryogel fibers (RCFs) incorporating a 3D PS was similar to PS/ACFs but without unidirectional freeze-casting (fig. S1A).

Fig. 2. Morphological features of 3D printed scaffold embedded without/with ACFs.

(A) Schematic illustrating the preparation procedures of ACFs incorporated 3D PS by 3D printing in combination with freeze-casting. (B) Gross view of PS and PS/ACFs before and after lyophilization. (C) SEM images showing the top view and cross section of the PS and PS/ACFs. (D) False color images of ACFs indicate the angle mapping of fiber orientations. (E) Angle distribution of formed cryogel fibers. (F) Distribution of cryogel fibers’ diameter. Oct, optimal cutting temperature compound; UV, ultraviolet.

Scanning electron microscopy (SEM) images further found that the holes of PS/ACFs were filled with fibers (top view), and these fibers exhibited a vertically aligned structure (cross section) when compared with unmodified PS (Fig. 2C). The PS/RCFs were filled almost entirely with horizontal fibers and lamellas (fig. S1B). The false color images (Fig. 2D) and the distribution of directionality analysis (Fig. 2E) demonstrated that the formed ACFs had highly coherent orientation, and the median fibers’ angles were nearly 90° (Fig. 2E). The formed RCFs were random and almost perpendicular to the wall of the holes (fig. S1, C and D). In addition, the diameter of ACFs ranged from 1 to 5 μm, with a mean of 2.46 ± 0.86 μm (Fig. 2F).

Moreover, the Young’s modulus of wet PS/ACFs was higher than that of PS (fig. S2, A and B). The maximum strain of wet PS/ACFs was slightly lower than that of PS, and there was no difference in the maximum stress between the two groups (fig. S2, C and D). Both lyophilized PS and PS/ACFs could reach swelling equilibrium within 1 hour in phosphate-buffered saline (PBS) buffer at 37°C (fig. S3A). In vitro enzymatic degradation experiment showed that both PS and PS/ACFs could be rapidly and completely degraded (within 3 hours) in 0.05% collagenase solution (fig. S3B).

In vitro cell seeding efficiency and in vivo cell recruitment ability of PS/ACFs

As shown in Fig. 3A (i), human umbilical cord endothelial cells (HUVECs) suspension was dropped to the top surface of the rehydrated scaffold and incubated for 2 hours, and then the HUVECs seeded PS, PS/RCFs, and PS/ACFs were transferred to fresh medium to culture another 24 hours. The 3D reconstructed confocal images showed that HUVECs only adhered to the surface of the PS (depth less than 200 μm), and there were no cells in the hole area. In contrast, HUVECs could adhere to both PS and ACFs/RCFs areas of PS/RCFs and PS/ACFs. In addition, the cells were not only distributed on the surface but also migrated downward to the bottom of the PS/RCFs (about 300 μm in depth) and the PS/ACFs (about 600 μm in depth) (Fig. 3B). SEM images further visualized that HUVECs preferred to attach to the RCFs/ACFs area rather than 3D printed Gel-MA lines (Fig. 3C). The Cell Counting Kit-8 (CCK-8) assay indicated that the PS/ACFs and PS/RCFs had twice as attached cells as PS after 24 hours of culture (Fig. 3D).

Fig. 3. The in vitro cell seeding efficiency and in vivo cell recruitment ability of the ACFs incorporated 3D PS.

(A) (i) Schematic illustrating the in vitro cell seeding and adhesion on the PS, PS/RCFs, and PS/ACFs. (ii) Schematic illustrating the in vivo cell recruitment and tissue ingrowth induced by PS/ACFs by a subcutaneous implantation model in rats. (B and C) Confocal microscope 3D images and SEM images showing HUVECs adhesion and distribution on the PS, PS/RCFs, and PS/ACFs after 24 hours of cell seeding. (D) Quantification of attached HUVECs on the PS, PS/RCFs, and PS/ACFs after 24 hours of culture. (E to G) H&E staining and semi-quantitative analysis exhibit the cell infiltration and tissue ingrowth of PS and PS/ACFs after 1 and 2 weeks of subcutaneous implantation. (H) Trichrome staining discovers the newly formed collagenous fiber orientation of the ingrowth soft tissue in the PS and PS/ACFs groups after 2 weeks of operation. The false color images indicate the orientation of collagenous fiber. (I) Angle distribution of the newly formed collagenous fibers after 2 weeks of operation. ***P < 0.001 and ****P < 0.0001. ns, not significant.

To further explore the in vivo cell recruitment ability of PS/ACFs, it was implanted subcutaneously in rats (Fig. 3A, ii). Hematoxylin and eosin (H&E) images revealed that more infiltrated cells were found in the gap of printed Gel-MA lines of PS/ACFs compared with PS after 1 and 2 weeks of implantation. The density of newly formed tissue in the ACFs area of PS/ACFs is significantly denser than the empty area of PS, and no fibrous capsules formed around the PS and PS/ACFs (Fig. 3, E to G). In addition, we also explored the extracellular matrix (ECM) morphology of newly formed tissue by trichrome staining. Trichrome staining images found that the orientation of ECM fibers of the PS group was randomly arranged. In contrast, it exhibited an orderly arrangement in the PS/ACFs group (Fig. 3H), which consisted of the direction of ACFs. The false color images and angle quantitative analysis further verified that the orientation of collagen fibers in the PS/ACFs group was highly coherent compared with those in the PS group (Fig. 3, H and I). Furthermore, we found that the angiogenesis in the PS/ACFs group was better than that in the PS group after 1 and 2 weeks of subcutaneous implantation (fig. S4).

Pro-osteogenic properties of the KP peptide–functionalized PS

To accelerate bone repair, we used KP peptide to modify the 3D PS to promote the differentiation of recruited bone marrow mesenchymal stromal cells (BMSCs) into osteoblasts. To verify the pro-osteogenic differentiation potential of the KP peptide–conjugated PS (PS-KP), three different concentrations of KP peptide (250, 500, and 750 μg/ml)–modified PS scaffold were prepared (fig. S5), and the in vitro and in vivo pro-osteogenic ability of different PS-KP were explored. Alkaline phosphatase (ALP) staining result showed that the rat BMSCs (rBMSCs) treated with PS-KP250, PS-KP500, and PS-KP750 groups had stronger ALP expression than that in the PS-treated group after 7 days of coculture (fig. S6, A and B). The relative expression of ALP, osteopontin (OPN), and osteocalcin (OCN) in the PS-KP–treated rBMSCs was increased compared with PS-treated rBMSCs (fig. S6, C to E). The in vitro results suggested that PS-KP could promote osteogenic differentiation of rBMSCs.

We also explored the pro-osteogenic effects of PS-KP using a subcutaneous ectopic osteogenesis model in nude mice. The suspension of rBMSCs was dropped onto the surface of the scaffolds and cocultured for 24 hours, and then subcutaneous implantation of rBMSCs loaded PS-KP in nude mice (Fig. 4A). The micro-computed tomography (Micro-CT) 3D reconstruction images of the isolated scaffold tissue showed that there was apparent new bone formation in the PS-KP500 group after 4 weeks of surgery, while there was almost no new bone formation in the PS, PS-KP250, and PS-KP750 groups (Fig. 4B). The axial, sagittal, and coronal images showed the distribution and morphology of regenerated bone. The regenerated bone in the PS-KP500 group showed a lattice shape consistent with PS morphology (Fig. 4C). The quantitative analysis of the regenerated bone volume (BV/TV, %), bone surface (BS/TS, %), trabecular number (Tb.N, 1/100 mm), and thickness (Tb.Th, μm) of the PS-KP500 group was definitely increased compared to the PS, PS-KP250, and PS-KP750 groups (Fig. 4, D to G). Trichrome (Fig. 4H) and H&E staining (fig. S7) showed that the PS-KP500 group had the most osteoid matrix distributed in clumps. The PS-KP750 and PS-KP250 groups had a small number of osteoid matrices. Almost no osteoid matrix was detected in the PS group. Immunohistochemical staining showed the expression of OPN in the PS-KP500 and PS-KP750 groups was higher than in the PS and PS-KP250 groups. The expression of OPN in the PS-KP500 group was also higher than in the PS-KP750 group (Fig. 4, I and J). Therefore, we determined KP peptide (500 μg/ml) as an effective concentration for modifying 3D printed Gel-MA hydrogel.

Fig. 4. Pro-osteogenic properties of the KP peptide-functionalized PS.

(A) Schematic illustrating the preparation procedures and verification of pro-osteogenic property of the KP peptide-functionalized PS. The ectopic osteogenesis model was used by subcutaneous implantation of rBMSCs-seeding scaffolds in nude mice (n = 5). (B) Micro-CT 3D reconstruction images exhibited ectopic new bone formation in the PS, PS-KP250 (250 μg/ml), PS-KP500 (500 μg/ml), and PS-KP750 (750 μg/ml) group after 4 weeks of operation. (C) Micro-CT axial, sagittal, and coronal views of scaffold area after 4 weeks of operation. (D to G) Regenerated bone volume (BV/TV, %), bone surface (BS/TS, %), Tb.N (1/100 mm), and Tb.Th (μm) after 4 weeks of operation. (H) Trichrome staining of the decalcified scaffold area and surrounding tissue of PS, PS-KP250, PS-KP500, and PS-KP750 groups after 4 weeks of operation. (I and J) OPN immunohistochemical staining of the decalcified scaffold area of PS, PS-KP250, PS-KP500, and PS-KP750 groups after 4 weeks of operation and semi-quantitative analysis. *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001. IOD, integral optical density.

Pro-angiogenic properties of QK peptide–functionalized PS/ACFs

As shown in Fig. 5A and fig. S8A, the QK peptide with different concentrations (100, 300, and 500 μg/ml) was conjugated to the ACFs, and the subcutaneous implantation of different PS/ACFs-QK was used to verify the pro-angiogenic properties of PS/ACFs-QK. The SEM images, false color images, and angle distribution analysis showed the vertical alignment of the cryogel fibers in the ACFs-QK100, ACFs-QK300, and ACFs-QK500 groups (fig. S8, B to D). The subcutaneous implanted scaffold with surrounding tissue was isolated 2 weeks after surgery. Trichrome staining results indicated that the newly formed blood vessels were gradually increased with the increased concentration of QK peptide from 0 to 500 μg/ml (Fig. 5, B and C). Immunohistochemical staining and semi-quantitative analysis further demonstrated that the area of newly formed blood vessels in the PS/ACFs-QK300 and PS/ACFs-QK500 groups was increased compared to the PS and PS/ACFs-QK100 groups. The area of newly formed blood vessels in the PS/ACFs-QK500 group was higher than the PS/ACFs-QK300 group (Fig. 5, D and E). Therefore, we determined QK peptide (500 μg/ml) as an effective concentration for modifying the ACFs.

Fig. 5. Pro-angiogenic properties of QK peptide-functionalized PS/ACFs.

(A) Schematic illustrating the preparation procedures and verification of pro-angiogenic property of the QK peptide-functionalized PS/ACFs. The subcutaneous implantation of QK peptide-functionalized PS/ACFs in rats was used to evaluate its pro-angiogenic properties (n = 4). (B) Trichrome staining of the scaffold area and surrounding tissues of PS/ACFs, PS/ACFs-QK100 (100 μg/ml), PS/ACFs-QK300 (300 μg/ml), and PS/ACFs-QK500 (500 μg/ml) groups after 2 weeks of operation. (C) Quantification of the number of newly formed blood vessels in the PS/ACFs, PS/ACFs-QK100, PS/ACFs-QK300, and PS/ACFs-QK500 groups after 2 weeks of operation. (D) CD31 immunohistochemical staining of the scaffold area and surrounding tissues in the PS/ACFs, PS/ACFs-QK100, PS/ACFs-QK300, and PS/ACFs-QK500 groups after 2 weeks of operation. (E) Quantification of the area of newly formed blood vessels in the PS/ACFs, PS/ACFs-QK100, PS/ACFs-QK300, and PS/ACFs-QK500 group after 2 weeks of operation. *P < 0.05 and ***P < 0.001.

The bone pro-regenerative ability of KP peptide– and QK peptide–functionalized PS/ACFs in rat critical-sized cranial bone defect model

To verify the potential bone pro-regenerative ability of KP peptide– and QK peptide–functionalized PS/ACFs, we implanted the PS-KP500/ACFs-QK500 into the critical-sized cranial bone defects (8 mm in diameter) in rats, the commercial collagen sponge was used as a positive control (Fig. 6A and fig. S9). The x-ray images (Fig. 6B) and reconstructed Micro-CT images (Fig. 6, C and D) showed that the PS-KP500/ACFs-QK500 and collagen sponge groups showed significantly more new bone formation than the control group after 4 and 8 weeks of surgery. The quantitative analysis further revealed no difference in the regenerated bone volume (BV/TV, %) and surface coverage (%) between the PS-KP500/ACFs-QK500 and collagen sponge groups after 4 weeks of treatment. While the regenerated bone volume (BV/TV, %) and surface coverage (%) of the KP500/ACFs-QK500 group were higher than the collagen group after 8 weeks of treatment (Fig. 6, E and F).

Fig. 6. The bone pro-regenerative ability of KP peptide- and QK peptide-functionalized PS/ACFs in rat critical cranial bone defect model.

(A) Schematic illustrating the implantation of experimental scaffold (PS-KP500/ACFs-QK500) and collagen sponge (positive control) to a critical-sized cranial bone defect (φ 8 mm) in rats (n = 5). (B) X-ray raw images of the control, collagen sponge, and PS-KP500/ACFs-QK500 groups after 4 and 8 weeks of operation. (C and D) Micro-CT 3D reconstruction images and coronal views of the control, collagen sponge, and PS-KP500/ACFs-QK500 groups after 4 and 8 weeks of operation. (E and F) Regenerated bone volume (BV/TV, %) and surface coverage (%) of the control, collagen sponge, and PS-KP500/ACFs-QK500 groups after 4 and 8 weeks of operation. *P < 0.05, **P < 0.01, and ***P < 0.001.

We further analyzed the microstructure of the regenerated bone of each group. The coronal section of the 3D reconstructed Micro-CT images of all groups showed that the new bone was thicker at 8 weeks than at 4 weeks. The regenerated bone volume and distribution in the PS-KP500/ACFs-QK500 group achieved the expected healing effect after 8 weeks of surgery (fig. S10A). The trabecular pattern factor (Tb.Pf) is an index used to describe the degree of concavity of the trabecular surface. A decrease of Tb.Pf indicated that the trabecula changes from rod bone to plate bone, and the negative value of Tb.Pf showed that the measured bone tissue is a cortical bone. The Tb.Pf values of the control and PS-KP500/ACFs-QK500 groups did not change over time, while the Tb.Pf value of the collagen group changed from negative to positive (fig. S10B). The spatial morphological and structural indexes of bone trabeculae [including trabecular separation (Tb.Sp), trabecular number (Tb.N), and trabecular thickness (Tb.Th)] can indicate the changes of bone microstructure.

The collagen sponge and PS-KP500/ACFs-QK500 groups both had lower Tb.Sp than control group at 4 weeks, and only PS-KP500/ACFs-QK500 group had lower Tb.Sp than the control group at 8 weeks (fig. S10C). Tb.N in the collagen sponge and PS-KP500/ACFs-QK500 groups were higher than in the control group at 4 weeks after surgery, and only PS-KP500/ACFs-QK500 group had higher Tb.N than that in the control group at 8 weeks after surgery (fig. S10D). After 4 weeks of surgery, Tb.Th in the PS-KP500/ACFs-QK500 group was higher than that of the collagen sponge and control groups, with no difference among them at 8 weeks (fig. S10E). The several indexes suggest that in the early stages of bone regeneration, PS-KP500/ACFs-QK500 could promote the formation of bone trabeculae more quickly and maintain and promote the morphological and structural remodeling of trabecular bone favoring dense rod trabeculae.

The histological observations revealed that a more organic matrix of regenerated bone was detected in the PS-KP500/ACFs-QK500 and collagen sponge groups after 4 weeks of treatment, which was consistent with the reconstructed Micro-CT images (Fig. 7, A and B). We found that the newly formed collagen fibers in the defect area of the control and collagen sponge groups were all parallel to the skull surface (direction of the arrow). In the PS-KP500/ACFs-QK500 group, a part of vertically aligned collagen fibers was detected, which was introduced by ACFs. Some decalcified organic bone matrix formed within the vertically aligned collagen fibers (direction of the arrow) (Fig. 7B). We further analyzed the arrangement direction of collagen fibers located at the edge and middle of the wound. The collagen tissues in the control and collagen sponge groups were horizontally arranged, while the collagen fibers in the PS-KP500/ACFs-QK500 group were arranged vertically (Fig. 7, C to F). After 8 weeks of treatment, H&E and trichrome staining showed more newly formed organic bone matrix in the defect area 8 weeks after surgery (Fig. 8, A and B). Immunohistochemical staining (Fig. 8C) and semi-quantitative analysis showed that the relative expression of CD31, osteocalcin (OCN), osteopontin (OPN), and runt-related transcription factor 2 (RUNX2) in the PS-KP500/ACFs-QK500 group was the highest among the three groups (Fig. 8, D to G).

Fig. 7. Histological observations of bone regeneration after 4 weeks of treatment.

(A and B) H&E and trichrome staining of the decalcified cranial bone of the control, collagen sponge, and PS-KP500/ACFs-QK500 groups after 4 weeks of operation (n = 5). (C and E) Trichrome staining of newly formed soft tissues in the edge and center area of the control, collagen sponge, and PS-KP500/ACFs-QK500 groups after 4 weeks of operation. The false color images indicate the orientation of the regenerated soft tissue. (D and F) Angle distribution of the regenerated collagenous fibers in the edge and center of the regenerated soft tissue after 4 weeks of operation. NB, new bone; C, collagen sponge; PS, printed scaffold.

Fig. 8. Histological observations of bone regeneration after 8 weeks of treatment.

(A and B) H&E and trichrome staining of the decalcified cranial bone of the control, collagen sponge, and PS-KP500/ACFs-QK500 groups after 8 weeks of operation (n = 5). (C) CD31, OCN, OPN, and RUNX2 immunohistochemical staining of the regenerated bone tissue of the control, collagen sponge, and PS-KP500/ACFs-QK500 groups after 8 weeks of operation. (D to G) Quantification of integral OD of expressed CD31, OCN, OPN, and RUNX2 in the control, collagen sponge, and PS-KP500/ACFs-QK500 groups after 8 weeks of operation. *P < 0.05, **P < 0.01, and ****P < 0.0001.

DISCUSSION

In this work, our strategy for repairing large-area skull defects is to first recruit many cells involved in bone regeneration to the defect area through the aligned cyrogel fibers. Then, the conjugated QK peptide promotes blood vessel formation to support nutrients to help the recruited cells perform their functions. Meanwhile, the conjugated KP peptide promotes osteogenic differentiation and accelerates the formation of new bone. The ACFs incorporated 3D printed scaffold effectively facilitates bone regeneration by enhancing cell recruitment and function.

In this work, the most innovative point of the designed scaffold is the introduction of vertically ACFs into the 3D printed frame. Directional freeze casting is a typical method for preparing scaffolds with oriented structures (25, 26). However, the obtained scaffold is a lamellar structure with directional arrangement of channels. On the basis of the directional freeze casting technology, we successfully transformed the lamellar structure to ACFs by controlling the solution concentration and freezing polymerization temperature. Low concentrations of GelMA solutions (around 0.2%) could form ACFs after freeze polymerization and lyophilization. In addition, the dimension, density, and orientation of cryogel fibers can affect local cell recruitment. For example, the ACFs can promote cell migration compared to the RCFs (Fig. 2B). In addition, we also tried to control the density of ACFs, and we found that temperature is the key point. The fibrous ACF scaffold formed under −20°C exhibited significantly low density in comparison to the scaffold formed under −190°C. Low-density fiber scaffolds are more conducive to cell migration. Moreover, the dimension of the formed cryogel fibers is in micron size (1 to 4 μm). We are still trying to get aligned cryogel nanofibers. The nanoscale cryogel fibers can further promote cell adhesion, migration, and proliferation because of the regulatory role of nanotopography cues.

Our previous study demonstrated that the 3D aligned nanofiber scaffolds could enhance cell penetration and subsequently accelerate granulation tissue formation (18, 27). In the present research, incorporating ACF into a 3D PS enhanced granulation tissue formation in the defect area. The mechanism of repairing large-area bone losses is endochondral ossification. The granulation tissues first form in the defect area, and then cartilage forms within the granulation tissue and gradually transforms into bone (28, 29). Our study found several dissociative new bones in the vertically ACF area of the hybrid scaffold group. This may be one of the reasons that our materials can accelerate bone regeneration.

We noticed that the regenerated bone volume in the collagen sponge group at 8 weeks after surgery was not significantly higher than that at 4 weeks after surgery. However, the immunohistochemistry results related to ossification were better than those in the control group. We speculated that bone resorption in the collagen sponge group increased at week 8. We further performed tartrate-resistant acid phosphatase (TRAP) staining of the new bone area to test the hypothesis on week 8. TRAP staining images showed many osteoclasts existed around and inside the new bone of the collagen sponge group, while those in the control and PS-KP500/ACFs-QK500 groups were mainly inside the new bone (fig. S11). Our findings were consistent with many reported studies (30, 31). Our study solved bone resorption due to the prolonged in vivo degradation time and enhanced ossification induced by KP peptide (32, 33). These effects allow new bone to continuously form and gradually mature. Also, this is the main reason our bone repair materials improve the quality of bone regeneration.

In summary, we reported a hybrid 3D PS with modification of QK peptide and KP peptide for effectively promoting endogenous cranial bone regeneration. The hybrid 3D PS consists of vertically ACFs that guide and promote cell penetration into the defect area in the early stages of bone repair. Then, the conjugated QK peptide and KP peptide further regulate the function of the recruited cells to promote vascularization and osteogenic differentiation in the defect area. The regenerated bone volume and surface coverage of the dual peptide-modified hybrid scaffold were significantly higher than the positive control group (collagen sponge) after 8 weeks of implantation. In addition, the dual peptide-modified hybrid scaffold demonstrated sustained enhancement of bone regeneration compared to the collagen sponge group from 4 to 8 weeks. We expect that the strategy of introducing ACFs and incorporating functional peptides will inspire the design of next-generation biomaterials for effective endogenous bone regeneration with the desired quality.

MATERIALS AND METHODS

Materials

Type A porcine skin gelatin (catalog number: V900863), methacrylic anhydride (catalog number: 276685), ammonium persulfate (APS) (catalog number: 248614), tetramethylethylenediamine (TEMED) (catalog number: T9281), and 2-hydroxy-4′-(2-hydroxyethoxy)-2-methylpropiophenone (photoinitiator 2959) (catalog number: 410896) were purchased from Sigma-Aldrich (Shanghai, China). Dulbecco’s modified Eagle’s medium (DMEM), high glucose (catalog number: 11965092), minimum essential medium (α-MEM) (catalog number: 12571063), fetal bovine serum (FBS) (catalog number: 10099141C), and penicillin-streptomycin (P-S) (catalog number: 15140148) were obtained from Gibco (Shanghai, China). CD31 polyclonal antibody (catalog number: 28083-1-AP) and osteocalcin rabbit polyclonal antibody (catalog number: 23418-1-AP) were purchased from Proteintech (Wuhan, China). Anti-osteopontin antibody (catalog number: ab63856) and anti-RUNX2 antibody (catalog number: ab236639) were acquired from Abcam (Shanghai, China). TRAP kit (catalog number: G1492) was purchased from Beijing Solarbio Science & Technology Co. Ltd. (Beijing, China). The commercial collagen sponge was purchased from Beijing Pashionbio Co. Ltd. (Beijing, China). The BMP-2 knuckle epitope-derived peptide (KP) [N to C sequence: Alloc-KIPK(Ac) ASSVPTELSAISTLYL] and VEGF-derived peptide (QK) [N to C sequence: Alloc-KLTWQELYQLK(Ac)YK(Ac)GI] were commercially synthesized and identified by Nanjing TGpeptide Biotechnology Co. Ltd. (Nanjing, China).

Preparation of highly substituted Gel-MA

Highly substituted Gel-MA was prepared according to the previous literature. Briefly, 10 g of type A porcine skin gelatin was added to 100 ml of carbonate bicarbonate buffer (0.25 M) to be dissolved at 55°C under continuous stirring. Then, 2 ml of methacrylic anhydride was added into the gelatin solution using a micropump. After that, the reaction solution continued to react for 1 hour at 55°C while pH was kept at 9.4. The final pH of the reaction solution was adjusted to 7.4 to terminate the reaction. The final solution was dialyzed in deionized water (changed three to four times a day) at 37°C for 4 days and lyophilized to obtain Gel-MA. The product was stored at −20°C.

Preparation of 3D PSs embedded with ACFs

First, 0.025 g of photoinitiator 2959 was dissolved in 5 ml of deionized water at 50°C away from light. Then, 0.25 g of Gel-MA was added into the above solution at 50°C under continuous stirring away from light to attain 5 wt % Gel-MA solution. The hydrogel scaffold was produced by a 3D bioprinter (BioMaker4, SunP Biotech, Beijing, China) and crosslinked and cured under ultraviolet (UV) light (OmniCure S2000, Excelitas Technologies Corp., Canada). The printing parameters were set as the following: printer’s needle, 25 gauge; model height, 1.0 or 2.0 mm; layer height, 0.2 mm; line distance, 1.2 mm; the lines of the adjacent two layers crossed (90°); printing speed, 6 mm/s; and extrusion speed, 1 mm³/s. The cured hydrogel was cut into cylinders with a diameter of 8 mm, lyophilized, and stored at −20°C. The preparation method of the KP peptide–modified Gel-MA hydrogel scaffold was the same as above. Bio-ink containing different concentrations (250, 500, and 750 μg/ml) of KP peptide was prepared with 5 wt % Gel-MA solution mixed with KP peptide at 37°C under continuous stirring away from light.

The aligned cryogel fiber (ACF) was embedded in the lattices of a 3D PS by unidirectional freeze casting technique. A 90 mg of Gel-MA and 10 mg of gelatin were dissolved in 50 ml of deionized water at 50°C under continuous stirring away from light to obtain a presolution containing 0.18 wt % Gel-MA and 0.02 wt % gelatin. Then, 4 ml of presolution was mixed with 0.2 ml of APS (10 wt %) and 5.2 μl of TEMED in an ice bath. The lyophilized 3D PS was immersed in the above mixture for 1 min in an ice bath and then quickly transferred to the stainless steel plate precooled at −80°C while ensuring that the scaffold’s bottom fell to the plate simultaneously. The scaffold continued to be placed at −80°C for 24 hours for chemical crosslinking of Gel-MA. The product was lyophilized and stored at −20°C. In subsequent experiments, gelatin helps ACFs maintain its vertical fiber structure in PBS at low temperatures. The preparation method of QK peptide-modified ACF was the same as above. QK peptide was dissolved in the presolution in an ice bath under continuous stirring away from light to attain different concentrations (100, 300, and 300 μg/ml) of QK peptide. The scaffold continued to be placed at −80°C for 24 hours for chemical crosslinking of Gel-MA and QK peptide.

Characterization of biomaterials

Gross photographs of the biomaterials were taken with a stereoscopic microscope (Olympus SZ61). The microstructure was observed by SEM (Hitachi SU8010, Japan). The directionality of aligned fibers was analyzed using the “Directionality” and “OrientationJ” plugins in Fiji software. The aligned fibers’ diameter distribution was obtained by measuring 200 fiber segments using ImageJ software.

Compressive mechanical properties of the wet PS and PS/ACFs after crosslinking were obtained by an electronic universal testing machine (Instron 5944, USA). In compression parameter settings, the model was cylindrical (diameter, 8 mm; height, 2 mm) and the moving speed was 1 mm/min.

In the swelling test, the freeze-dried scaffold was weighed (M0). It was immersed in 37°C PBS buffer and weighed (M1) at a predetermined time. The swelling ratio was calculated as M1/M0.

In vitro enzymolysis test, the freeze-dried scaffold was weighed after rehydration in PBS buffer at 37°C for 1 hour (M0). Then, the rehydrated scaffold was immersed in 0.05% collagenase I (Gibco) solution at 37°C (enzyme activity about 100 U/ml) and weighed (M1) at a predetermined time point. The remaining mass ratio was calculated as M1/M0.

In vitro biological properties of biomaterials

Cell seeding and adhesion on biomaterials

3D PS embedded with or without ACF (diameter, 8 mm; height, 1 mm) was freeze-dried and sterilized using UV light. HUVECs with red fluorescent protein (RFP) fluorescence were cultured in high-glucose DMEM medium supplemented with 10% FBS and 1% P-S at 37°C and 5% CO₂. PS/ACFs were immersed in PBS at 4°C for 4 hours before biological experiments to remove the residual APS, TEMED, and their by-products. Scaffolds were put in a 48-well plate with one scaffold per well, and a complete medium was added. The scaffolds were rehydrated at 37°C for 1 hour for swelling balance, and excess liquid was discarded. HUVECs-RFP suspension (50 μl; 2 × 10⁶ cells/ml) was dropped to the top surface of the scaffold and cultured at 37°C and 5% CO₂. After 2 hours, the fresh complete medium was supplemented, and the cells continued to be cultured for 24 hours. Then, the cell-loaded scaffolds were transferred to a new 48-well plate for the subsequent experiments.

The cell-loaded semi-quantitative analysis

The culture medium of the new 48-well plate was replaced with CCK-8 solution according to the manufacturer’s protocol (Beyotime Biotechnology) and incubated for 1 hour away from light. Absorbance [optical density (OD) value)] at 450 nm was determined with a Microplate Spectrophotometer (Epoch2, Bio-Tec Instruments, USA).

CLSM and SEM observation of cell adhesion and distribution on biomaterials

For confocal laser scanning microscope (CLSM) observation, the cell-loaded scaffold in the new 48-well plate was fixed with 4% paraformaldehyde for 15 min and washed three times with PBS buffer. The cell adhesion and distribution on the biomaterials were observed using CLSM (Nikon A1, Japan) with the z-stack range from 0 to 600 μm and an interval of 10 μm. For SEM observation, the cell-loaded scaffolds fixed with 4% paraformaldehyde were dehydrated by gradient alcohol (70 to 100%, 15 min for each class) and vacuum-dried at 37°C for 24 hours. The scaffolds were sputtered with platinum (Leica EM ACE600, Germany) and observed under SEM (Hitachi SU8010, Japan).

Subcutaneous implantation in rats

The Sprague-Dawley rats [specific pathogen–free (SPF) grade, male, 10 weeks, 300 g] were purchased from the Experimental Animal Center of Zhejiang Province. This animal experiment protocol was approved by the Animal Research and Ethics Committee of Wenzhou Institute of the University of Chinese Academy of Sciences (approval number: WIUCAS22072901). The experimental groups used are 3D PS, 3D PS embedded with ACF (PS/ACFs), 3D PS embedded with QK peptide–modified (100, 300, and 500 μg/ml) ACF (PS/ACFs-QK100, PS/ ACFs-QK300, and PS/ ACFs-QK500). There were eight samples (four rats) in each group, and four samples in two rats were collected 1 and 2 weeks after surgery, respectively. The operation procedure was as follows: The rat was anesthetized with 3% isoflurane in oxygen for about 2 min, and then the back hair of the rat was shaved, disinfected with iodophor twice, and cleaned with 75% ethanol gauze once. The skin and subcutaneous tissue were cut to make a subcutaneous cavity, and the scaffold was sent into the subcutaneous cavity, with one scaffold in each cavity. The incision was closed and disinfected. The rat was placed on a heating pad during the surgery to maintain its body temperature. Euthanasia by an overdose of anesthesia was performed on the rats after 1 and 2 weeks of surgery. The obtained subcutaneous tissue samples were washed with PBS buffer and soaked in 4% paraformaldehyde for 1 day, then underwent alcohol gradient dehydration and paraffin embedding, and then cut into 5-μm-thick slices for histological staining.

Ectopic osteogenesis in nude mice

Nude mice (SPF grade, male, 6 to 8 weeks, 18 to 25 g) were purchased from the Experimental Animal Center of Zhejiang Province. This animal experiment protocol has been approved by the Animal Research and Ethics Committee of Wenzhou Institute of the University of Chinese Academy of Sciences (approval number: WIUCAS22122001). Experimental groups used are 3D PSs and 3D PSs modified with KP peptide (250, 500, and 750 μg/ml) (PS-KP250, PS-KP500, and PS-KP750). Six samples in three mice were collected in each group at 4 weeks after surgery. The procedures for cell seeding on scaffolds were as follows: First, the freeze-dried and UV-sterilized scaffolds were immersed in a complete medium (α-MEM, 10% FBS, and 1% P-S) at 37°C for 1 hour for rehydration and then transferred to a 48-well plate. P3 rBMSC suspension (50 μl; 4 × 10⁷ cells/ml) was dropped to the top surface of each scaffold and cultured at 37°C and 5% CO₂. After 2 hours, the fresh complete medium was supplemented, and the cells continued to be cultured for 24 hours. Then, the cell-loaded scaffolds were for the subsequent subcutaneous implantation. The procedures were as follows: The nude mouse was anesthetized with 2% isoflurane in oxygen for about 2 min, the back skin was disinfected with iodophor twice, and the back skin was cleaned with ethanol gauze once. The skin and subcutaneous tissue were cut to make a subcutaneous cavity, and the cell-loaded scaffold was sent into the subcutaneous cavity, one scaffold in each cavity. The incision was closed and disinfected. The mouse was placed on a heating pad during the surgery to maintain its body temperature. Euthanasia by an overdose of anesthesia was performed on the mice after 4 weeks of surgery. The subcutaneous tissue samples were washed with PBS buffer, soaked in 4% paraformaldehyde for 3 days, and then transferred to 70% ethanol, followed by a Micro-CT scan, decalcification (10% EDTA, 2 weeks, room temperature), and histological evaluation.

Critical-sized rat cranial bone defect model

The Sprague-Dawley rats (SPF grade, male, 12 weeks, 350 g) were purchased from the Experimental Animal Center of Zhejiang Province. This animal experiment protocol was approved by the Animal Research and Ethics Committee of Wenzhou Institute of the University of Chinese Academy of Sciences (Approval number: WIUCAS22121603). Experimental groups used are control group (no scaffold), collagen sponge group (positive control), and KP peptide–modified 3D PS embedded with QK peptide–modified ACF (PS-KP500/ACFs-QK500) group. Each group had 10 samples (10 rats), and 5 samples in 5 rats were collected 4 and 8 weeks after surgery, respectively. The operation procedures were as follows: The rat was anesthetized with 3% isoflurane in oxygen for about 2 min, and then the rat calvarial hair was shaved, disinfected with iodophor twice, and cleaned with 75% ethanol gauze once. Sagittal incision was made at the top of the skull of the rat. The skin, subcutaneous tissue, and periosteum were sequentially incised to expose the calvaria. A critical-sized cranial bone defect (8 mm) was created by using a trephine bur mounting on a dentist drill. The normal saline solution was continuously added to reduce heat damage. The wound was cleaned with a normal saline solution, and different treatments were applied to the bone defects. The periosteum, subcutaneous tissue, and skin were sutured, and the incision was disinfected with iodophor. The rat was placed on a heating pad during the surgery to maintain its body temperature. Euthanasia by an overdose of anesthesia was performed on the rats after 4 and 8 weeks of surgery. The rat cranium samples were isolated, washed with PBS buffer, soaked in 4% paraformaldehyde for 3 days, and then transferred to 70% ethanol, followed by a Micro-CT scan, decalcification (10% EDTA, 4 weeks, room temperature), and histological evaluation.

Histological observation

After alcohol gradient dehydration and paraffin embedding, the tissue samples were cut into 5-μm-thick slices for H&E staining, trichrome staining, and immunohistochemical staining. Trichrome staining was carried out according to the standard protocol of the manufacturer (catalog number: MST-8003, Fuzhou Maixin Biotechnology Development Co. Ltd., China). During immunohistochemical staining, the slices were deparaffinized and rehydrated, followed by antigen retrieval in heated citrate buffer for 15 min (citrate buffer solution, pH 6.0 at 90°C). Nonspecific antibody binding was prevented with 10% goat serum albumin solution. The slices were incubated overnight at 4°C after adding the primary antibody. Then, the corresponding secondary antibodies were added and incubated at room temperature for 1 hour. Tissue color development using the DAB Horseradish Peroxidase Color Development Kit and observation were carried out according to the standard protocol of the manufacturer (catalog number: ZLI-9018, Beijing Zhongshan Jinqiao Biotechnology Co. Ltd., China). At least five randomly selected fields were examined for each group at each time point and used to assess the average positive intensity. The positive expression of OPN, OCN, RUNX2, and CD31 in the new bone was determined using Fiji software and was presented as the integral OD.

Micro-CT scanning and analysis

The ectopic osteogenesis and rat cranium samples were scanned using high-resolution Micro-CT (SKYSCAN 1275, Bruker). Scanning parameters included voltage (50 kV), current (60 μA), and image pixel size (9.0 μm). CTAn software (Bruker) was used to analyze bone volume, bone area, Tb.Th, Tb.N, and Tb.Pf. The Micro-CT 3D images of the isolated samples were reconstructed by CTvol (Bruker). The surface coverage of regenerated bone was calculated using ImageJ software.

KP peptide–modified 3D PS promotes osteogenic differentiation of rBMSC in vitro

ALP staining

P3 rBMSCs were seeded on a 48-well plate with a cell density of 1 × 10⁴ cells per well and cultured in α-MEM containing 10% FBS and 1% P-S at 37°C and 5% CO₂ condition. The medium was replaced by an osteogenic induction medium [α-MEM containing 10% FBS, 1% P-S, ascorbic acid (50 μg/ml), 10 mM sodium β-glycerophosphate, and 10 nm dexamethasone] 24 hours later. A scaffold was added to each well. The scaffolds were 3D PS modified with KP peptide at different concentrations (250, 500, and 750 μg/ml) (PS-KP250, PS-KP500, and PS-KP750) after lyophilization and UV sterilization. They were immersed in an osteogenic induction medium at 37°C for 1 hour for rehydration. The osteogenic induction medium was changed every 2 to 3 days. The cells were cultured for 7 days. ALP staining was performed according to the instructions (BCIP/NBT Alkaline Phosphatase Color Development Kit, Beyotime Biotechnology, China).

Real-time quantitative polymerase chain reaction

Total RNA of rBMSCs cocultured with 3D PSs modified with different concentrations of KP peptide in osteogenic induction medium for 7 days was extracted, according to the standard protocol of the manufacturer (Nanjing Vazyme Biotech Co.,Ltd.). Then, the RNA was reverse-transcribed into cDNA using SYBR PremexExTaq (Takara, China). The mRNA levels of ALP, OPN, and OCN in rBMSCs were detected by real-time quantitative polymerase chain reaction. Relative gene expression was normalized to housekeeping gene glyceraldehyde-3-phosphate dehydrogenase (GAPDH). The primers used in this study are as follows: ALP (forward primer: CAACCTGACTGACCCTTCCC; reverse primer: CGATGGCCTCATCCATCTCC), OPN (forward primer: TGAGTTTGGCAGCTCAGAGG; reverse primer: TCGTCGTCGTCATCATCGTC), OCN (forward primer: CAAGCAGGAGGGCAGTAAGG; reverse primer: GAAGCCAATGTGGTCCGCTA), and GAPDH (forward primer: GCGAGATCCCGCTAACATCA; reverse primer: CTCGTGGTTCACACCCATCA).

Statistical analysis

All data were presented as means ± SD. Statistical analysis was performed using GraphPad Prism 7.0 software. The difference between the two groups of independent samples was evaluated using a t test. Differences among groups were assessed by one-way analysis of variance (ANOVA) followed by Tukey’s multiple comparison test. The values of P < 0.05 were considered statistically significant. The values of P < 0.01 were considered statistically very significant.

Supplementary Materials

This PDF file includes:

Figs. S1 to S11