Blockade of adrenergic β‐receptor activation through local delivery of propranolol from a 3D collagen/polyvinyl alcohol/hydroxyapatite scaffold promotes bone repair in vivo

Abstract Objectives Activation of the sympathetic system and adrenergic β‐receptors following traumatic bone defects negatively impairs bone regeneration. Whether preventing β‐receptor activation could potentially improve bone defect repair is unknown. In this study, we investigated the effect of systematic administration and local delivery of propranolol through composite scaffolds on bone healing. Materials and methods Collagen/PVA/propranolol/hydroxyapatite（CPPH）composite scaffolds were fabricated with 3D printing technique and characterized by scanning electron microscope (SEM). Micro‐CT analysis and bone formation histology were performed to detect new bone formation. Osteogenic differentiation of bone marrow stromal cells (BMSCs) and osteoclastogenesis of bone marrow monocytes cultured with scaffolds extract were performed for further verification. Results Intraperitoneal injection of propranolol did not significantly improve bone repair, as indicated by micro‐CT analysis and bone formation histology. However, CPPH scaffolds exhibited sustained release of propranolol in vitro and significantly enhanced bone regeneration compared with vehicle collagen/PVA/hydroxyapatite (CPH) scaffolds in vivo. Moreover, in vitro experiments indicated the scaffolds containing propranolol promoted the osteogenic differentiation and migration of rat BMSCs and inhibited osteoclastogenesis by preventing β‐receptor activation. Conclusions This study demonstrates that local adrenergic β‐receptor blockade can effectively enhance the treatment of bone defects by stimulating osteogenic differentiation, inhibiting osteoclastogenesis and enhancing BMSCs migration.

this technique. 2 Therefore, the development of novel biomaterials and identification of potential therapeutic targets are urgently required to improve bone regeneration.
Major trauma and severe injury result in post-trauma stress disorder, which activates the sympathetic system and induces a catecholamine surge. [3][4][5][6] In turn, this pathophysiologic reaction leads to host damage by activating adrenergic β-receptors. 7,8 Recently, a growing body of research has revealed that activation of β-receptors negatively modulates bone remodelling under physiological conditions 9,10 and impairs bone metabolism and bone fracture healing after trauma. 11,12 Moreover, β-receptors are expressed on osteoblastic cells and osteoclast-like cells, and stimulation of β-receptors using agonists inhibits alkaline phosphatase (ALP) activity and increases osteoclastic activity in vitro. 13,14 This evidence indicates that the high levels of catecholamines induced by traumatic bone injury may activate adrenergic β-receptors and negatively impact bone healing process. Therefore, adrenergic β-receptors represent a potential therapeutic target to improve the treatment of bone defects.
Propranolol is a classic adrenergic beta-blocker that exerts well-characterized anti-hypertensive effects. However, in-depth studies have indicated that administration of propranolol enhances bone mineral density and reduces the risk of fracture in osteoporosis. 15,16 These positive clinical results promoted us to further explore the role of propranolol in the treatment of bone defects.
However, limited studies that assessed the effects of systematic administration of propranolol on bone defect repair have reached conflicting conclusions. 17,18 Drug metabolism and clearance may make it difficult to achieve effective concentrations of the drug in the defect zone after systematic administration. Taking these limitations into consideration, local delivery of propranolol could potentially be more effective in the treatment of localized bone defects.
Local delivery system employs a carrier to deliver therapeutic molecules in order to improve the therapeutic potential by maintaining effective concentrations of the molecules or drug in the treatment area. 19,20 Our group previously constructed a porous bioactive collagen/hydroxyapatite scaffold that successfully improved bone regeneration. 21 Polyvinyl alcohol (PVA) has been shown to release propranolol in a sustained manner in vitro and been employed as a release medium for the treatment of localized diseases such as infantile hemangioma. [22][23][24] In this work, we added a low concentration of PVA incorporating propranolol to the raw collagen/hydroxyapatite homogenate to generate novel collagen/ PVA/propranolol/ hydroxyapatite (CPPH) scaffolds. HPLC assays indicated the CPPH scaffolds continuously released propranolol in vitro for up to 3 weeks. Based on this stable drug-release behaviour, we further investigated and evaluated the efficacy of local delivery of propranolol from CPPH scaffolds on bone regeneration in the rat critical-sized femoral defect model. This study provides further knowledge of the effects and possible mechanisms of action of local β-receptor blockade during the repair of critical-sized bone defects in vivo.

| Fabrication of 3D collagen/PVA/ hydroxyapatite scaffolds incorporating propranolol
Vehicle CPH scaffolds and CPPH scaffolds containing two ratios of propranolol were 3D printed using a filament-free print- In our previous work, a collagen-hydroxyapatite (CHA) scaffold was printed using the inks (2 g of collagen, 4 g of HA, 10 mL of AAS) and successfully repaired bone defect of rabbits. 21 We kept this ratio and added PVA to the inks for drug delivery. The printability was poor when the mass fraction of PVA solution was over 5 wt%, and the drug release was stable when PVA was 2wt% according to releasing patterns. Briefly, 2 g PVA was dissolved in 98 mL deionized water at 95°C to prepare 2 wt% PVA solution, and then, the solution was cooled to room temperature and various doses of propranolol (0, 10, 100 mg) were added to 1 mL of PVA solution to obtain 0%, 1% or 10% w/v propranolol solutions. Collagen (18% w/v) and 36% (w/v) hydroxyapatite were individually dissolved in 0.5 M acetic acid solution (AAS) and agitated to form turbid liquids. The printing ink was made of 1.8 g collagen, 3.6 g nano-hydroxyapatite, 9 mL AAS and 1 mL PVA with propranolol. Air bubbles were removed from the CPH and CPPH (collagen/PVA/propranolol/hydroxyapatite) pastes using a vacuum, and the ink was transferred to the plotting cartridge and placed in the robotic deposition device.
The entire printing process was controlled by using the FFP software program installed in the printer. The STL profile encoding the scaffold characteristics was designed in advance and converted into G-code. The ink was precisely delivered through the pinhole of the nozzle (diameter＝400 μm) and deposited onto the stainless-steel baseplate at a constant x/y-axes printing speed of 10 mm/s. To increase porosity and stability, each 200 μm thick layer was printed twice (back and forth; XXYY model) in the z-axis to form a 400 μm story height.
Cubic scaffolds of different sizes were printed for the in vitro (10 mm × 10 mm × 3 mm) and in vivo (5 mm × 3 mm × 3 mm) tests.
The CPH scaffolds without propranolol are called vehicle scaffolds in this study. The collagen/PVA/propranolol/hydroxyapatite ratio for the CPPH scaffolds was 1.8:0.02:0.01:3.6 w/w for the L-Pro (low propranolol) scaffolds and 1.8:0.02:0.1:3.6 w/w for the H-pro (high propranolol) scaffolds. The printed scaffolds were frozen at −80℃, lyophilized in a freeze dryer (Alpha 1-2 LD plus, Christ, Osterode, Germany), immersed in 1% w/v genipin solution (Wako Pure Chemical Industries, Ltd., Kanagawa, Japan) for cross-linking, sterilized using ethylene oxide and sealed in a sterile package until use.

| Scanning electron microscopy and porosity analysis
The surface and microstructure of the 10 mm × 10 mm × 3 mm scaffolds were characterized by scanning electron microscopy (SEM, S-4800; Hitachi, Tokyo, Japan). The porosity of the scaffolds was measured using a high-resolution micro-computed tomography (micro-CT) analysis system (Yxlon) at 10.5 μm and 80 keV using VGStudio MAX software (Volume Graphics, Heidelberg, Germany).

| X-ray diffraction analysis
The crystal phases of the scaffolds were scanned by X-ray diffractometry (X' Pert MPD PRO, PANalytical BV Netherlands) at a range of 10°-80° at 2°/min and 3 kW.

| High-performance liquid chromatography (HPLC)
Vehicle, L-Pro and H-Pro scaffolds were immersed in 5 mL PBS at 37°C and subjected to sustained oscillation using a vibrating plate.
At the designated time points, aliquots of the solutions were collected and replaced with the same volume of PBS. All samples were stored at − 80℃ until the concentrations of propranolol in each solution were measured by high-performance liquid chromatography (HPLC, Agilent Technologies, CA, USA). A standard propranolol solution (0.5 mg/mL) was used to determine the concentrations.

| Biocompatibility of scaffolds
Cell viability was determined using the Cell Counting Kit-8 assay (CCK- To achieve fluorescent double staining, rats were intramuscularly injected with tetracycline (80 mg/kg; Sigma-Aldrich) 14 days before euthanasia and calcein (8 mg/kg; Sigma-Aldrich) 3 days before euthanasia (at 12 weeks after surgery).

| Micro-CT bone analysis
The femur samples (n＝6 per group) were dissected, fixed in 80% ethanol and subjected to micro-CT scanning (Y. Cheetah; Yxlon, Hamburg, Germany). The region of interest (ROI) was a 3.5 × 3.5 × 5 mm 3 cylindrical region in the middle of the tunnel defects along the same major axis. A total of 450 micro-CT slices were acquired for each sample, and the projections were reconstructed using VGStudio MAX software (Volume Graphics, Heidelberg, Germany).

| Quantification of epinephrine and noradrenaline in plasma
Plasma samples were isolated by centrifuging whole blood at 4℃ (1400 g for 10 min)and stored at −80℃. Commercial ELISA kits (Cloud-Clone Corp., Houston, TX, USA) were used to measure the levels of noradrenaline (NE) and epinephrine (Epi) in plasma following the manufacturer's instructions; the plates were read at 450 nm.

| Cell migration assays
Cell migration assays were performed using 8 μm pore size transwell cell culture inserts (Corning, Tewksbury, MA, USA). BMSCs were serum-starved for 12 h and seeded onto the upper chamber in media without FBS; scratch wound assays were employed to objectively evaluate BMSCs migration. BMSCs were cultured until 100% confluent in six-well plates, and then, the cell layer was scarped using a 200 μL sterile pipette tip to create a linear wound and washed three times with PBS. The media was replaced with scaffold-conditioned media containing isoproterenol. After 12 h, the scratch wound closure was measured under the microscope. Cell migration between two scratch edges was calculated using Image-Pro Plus 6.0.

| Statistical analysis
Quantitative data are presented as the mean ± standard deviation (SD). All experiments were performed with at least n = 3 samples in each group. After confirming the data were normally distributed using the Kolmogorov-Smirnov test, the differences between groups were analysed using two-sided Student's tests or one-way ANOVA analysis of variance, followed by Turkey's post hoc test. The statistical analysis was calculated by SPSS 16.0 software, and the levels of significance were set at *P < .05.

| Systemic blockade of adrenergic β-receptors does not obviously improve bone repair
Collagen/PVA/hydroxyapatite (CPH/vehicle) scaffolds were implanted into distal femur defects under sterile conditions ( Figure   S1A), and the rats were intraperitoneally injected with propranolol or normal saline every day.
Micro-CT reconstructions of the defects at 12 weeks are presented in Figure 1A. New bone volume ratio (BV/TV) was slightly higher, but not significantly higher, in the propranolol-treated group than the rats treated with normal saline ( Figure 1D). Masson's trichrome staining and H&E staining confirmed that the area of The plasma levels of catecholamine and noradrenaline were lower in the injection group within the first 3 days after surgery but not significantly different between groups from 1 week until 12 weeks after surgery ( Figure 1G,H). TH + sympathetic nerve density in the defect zone was not significantly different between groups at 12 weeks ( Figure S1D,G).

| Characterization of composite scaffolds containing propranolol
Uniform (

| Release patterns of propranolol in vitro
The scaffolds were immersed in 5 mL PBS for release assays, and  BMSCs were also cultured in scaffold-conditioned media. Flow cytometry analysis revealed no significant differences in the ratios of living and apoptotic cells between groups ( Figure 4B,C). The CCK-8 assay confirmed that the BMSCs cultured in scaffold-conditioned media proliferated over time, with no significant differences between groups at any time point ( Figure 4D). High-power magnification images (100×) clearly identified Trap + osteoclasts containing more than three nuclei ( Figure S3D). As shown in Figure 6B, tyrosine hydroxylase (TH + ) adrenergic nerve fibres were rarely observed within the scaffolds at 4 weeks.

| Scaffolds containing propranolol recruit BMSCs and impair inner sympathetic innervation during bone repair in vivo
As the scaffolds degraded and new bone in-growth occurred, the total length of TH + nerve fibres inside the scaffolds significantly increased ( Figure 6B, Figure S3F). Semiquantitative analysis revealed the density of TH + nerve fibres was lower in the H-Pro (19.0 ± 1.63/ mm 2 ) and L-Pro (25.7 ± 0.94/mm 2 ) groups compared with the vehicle group (33.3 ± 2.62/mm 2 ) at 8 weeks ( Figure 6D).

| Effect of propranolol scaffold-conditioned media on osteoclasts formation in vitro
Trap staining indicated that propranolol inhibited osteoclast formation by preventing activation of β-receptors ( Figure 8A,C, Figure   S5A,B). Western blotting also revealed when the β-receptors were activated, and Trap and Ctsk expressions were lower at 5 days in the L-Pro and H-Pro groups compared with the control and vehicle groups ( Figure 8E). In addition, the inhibition of osteoclastogenesis in L-Pro and H-Pro groups was associated with reduced ROS levels, rather than inhibition of BMMs proliferation ( Figure 8B,D, Figure   S5C,D). found that systematic administration of propranolol could enhance endochondral bone formation and osteointegration of implants. The conflict between our results and those studies may be due to the different modes of repair in those bone defect models, which can self-repair without implantation of scaffolds. 17,28 Delivery of a drug directly to the defect site using a local delivery system may improve the treatment effects in bone regeneration. 29,30 To prolong the release of propranolol, we constructed collagen/ PVA/propranolol/hydroxyapatite scaffolds via a 3D printing technique. This novel scaffold exhibited good mechanical properties and biocompatibility with BMSCs. Encouragingly, after functionalization with the natural cross linker genipin, in vitro kinetic release assays showed the scaffolds led to sustained release of propranolol for at least 3 weeks. 31,32 The concentrations of propranolol released into 5 mL PBS after 24 h by the L-Pro and H-Pro scaffolds were approximately 0.12 and 2.18 μM, respectively, and a previous study reported that 1 μM propranolol reversed the inhibitory effects of isoprenaline on the osteogenic differentiation of BMSCs in vitro. 33 These results indicate that the scaffolds release pharmacologically relevant concentrations of propranolol in vitro (Scheme 1).
Larger areas of new bone tissue formed inside the scaffolds containing propranolol compared with vehicle scaffolds. Trap staining in vivo demonstrated that locally released propranolol strongly inhibited osteoclastogenesis, which is critical for effective bone remodelling and repair. 34 As shown in Figure 6A, L-Pro and H-Pro scaffolds contained higher numbers of leptin receptor + cells than the vehicle scaffolds at 4 weeks, but not at 8 and 12 weeks post-implantation.
Recruitment of endogenous mesenchymal stromal cells to the trauma site is also required to initiate bone regeneration. 35 Leptin receptor + mesenchymal stromal cells, a major subpopulation of BMSCs derived from the periosteum, migrate to the fracture zone and proliferate rapidly in response to fracture or injury . 36  To our surprise, scaffolds loaded with propranolol were innervated with fewer sympathetic nerve fibres. These results suggest that locally blocking β-receptors may modulate chemorepulsive guidance cue production such as semaphorin 3A (Sema3A) and thus prevent sympathetic innervation. 39 The in vitro results indicated the ability of propranolol to affect osteoblast differentiation was related to antagonism towards isoprenaline. Majeska et al (1992) found that culturing osteoblasts with isoprenaline massively increased cyclic adenosine monophosphate (cAMP) levels and inhibited ALP activity, while propranolol counteracted these effects of isoprenaline by more than 60%. 13 In terms of osteoclast formation, isoprenaline is well characterized to increase ROS generation, which directly promotes osteoclast formation. 40,41 Consistent with increased LepR + BMSCs recruitment to L-Pro and H-Pro scaffolds in vivo, blocking β-receptor activation using propranolol scaffold-conditioned media also accelerated the migration of BMSCs in vitro. Based on our in vitro assays, we conclude that local blockade of adrenergic β-receptors using propranolol leads to a combination of effects during bone reconstruction, including increased osteogenesis and BMSCs recruitment and reduced osteoclastogenesis.
However, there are some limitations to this study. The in vitro assays showed the propranolol scaffolds promoted bone regeneration by competing with isoprenaline stimulation. Therefore, the role of exogenous isoprenaline during bone repair needs to be studied in vivo. Moreover, the density of adrenergic β-receptors and catecholamine levels in the defect zone should be quantified to clarify how propranolol alters the sympathetic system inside scaffolds. Finally, sympathetic nerve innervation varies in different types of bone and between species, so further studies on different bones and in larger animals are necessary to validate our findings.
In conclusion, this study demonstrates that-in contrast to systemic administration-local adrenergic β-receptor blockade can effectively enhance the treatment of bone defects and excellent therapeutic effects can be achieved using 3D-printed composite scaffolds. In the future, we believe that scaffolds incorporating factors that block sympathetic activation will hold significant potential for bone tissue engineering applications.

ACK N OWLED G EM ENTS
This study was financially supported by the National Natural Science

CO N FLI C T O F I NTE R E S T
The authors have no competing financial interests to declare.

DATA AVA I L A B I L I T Y S TAT E M E N T
All data that support the findings of this study are available from the corresponding author upon reasonable request.