3D printed composite scaffolds with dual small molecule delivery for mandibular bone regeneration

Abstract

Functional reconstruction of craniomaxillofacial defects is challenging, especially for the patients who suffer from traumatic injury, cranioplasty, and oncologic surgery. Three-dimensional (3D) printing/bioprinting technologies provide a promising tool to fabricate bone tissue engineering constructs with complex architectures and bioactive components. In this study, we implemented multi-material 3D printing to fabricate 3D printed PCL/hydrogel composite scaffolds loaded with dual bioactive small molecules (i.e. resveratrol and strontium ranelate). The incorporated small molecules are expected to target several types of bone cells. We systematically studied the scaffold morphologies and small molecule release profiles. We then investigated the effects of the released small molecules from the drug loaded scaffolds on the behavior and differentiation of mesenchymal stem cells (MSCs), monocyte-derived osteoclasts, and endothelial cells. The 3D printed scaffolds, with and without small molecules, were further implanted into a rat model with a critical-sized mandibular bone defect. We found that the bone scaffolds containing the dual small molecules had combinational advantages in enhancing angiogenesis and inhibiting osteoclast activities, and they synergistically promoted MSC osteogenic differentiation. The dual drug loaded scaffolds also significantly promoted in vivo mandibular bone formation after 8-week implantation. This work presents a 3D printing strategy to fabricate engineered bone constructs, which can likely be used as off-the-shelf products to promote craniomaxillofacial regeneration.

Graphical Abstract

1. Introduction

Craniomaxillofacial bone defects are common and result from trauma, tumor, bone infection, or hereditary malformation [1]. Repair and regeneration of craniofacial bone defects are challenging due to the complexity of the structures involved and the sophisticated biomechanical and physiological environment [2, 3]. Current approaches for the reconstruction of craniomaxillofacial bone defects include the utilization of autografts and allografts, rigid fixation, and free tissue transfer [4]. However, these strategies are associated with many drawbacks, such as nonunion, plate extrusion, limited availability, donor site morbidity, potential immunogenic rejection, and disease transmission [5, 6]. As an exciting alternative strategy, bone tissue engineering provides novel treatment modalities for craniomaxillofacial bone defect repair and reconstruction [7, 8].

3D printing/bioprinting technologies have been widely used for fabricating engineered bone tissues with complex architectures and customized shapes in a layer-by-layer manner [9, 10]. It is thus a promising technique to accurately replicate the 3D complexity and achieve shape fidelity for surgical reconstruction and regeneration of craniomaxillofacial tissue [11, 12]. By optimizing the pore size and controlling the 3D printed scaffold microstructure, vascularization, and bone regeneration can be improved [13, 14]. However, in most cases, 3D printed scaffolds alone are not osteoinductive in nature. To address this shortcoming, stem cells, especially mesenchymal stem cells (MSCs), and growth factors are incorporated by using various strategies to promote defect regeneration [15, 16]. The autologous cells have osteogenic potential, but they require extensive isolation, manipulation, and culture processing [17]. The use of growth factors (e.g. bone morphogenic protein-2 (BMP-2) and vascular endothelial growth factor (VEGF)) can significantly improve the osteogenesis and vascularization [18]. 3D printing also enables the incorporation of multiple growth factors/peptides to generate heterogeneous tissue constructs and achieve multiple functions [19, 20]. However, several potential complications and concerns are associated with the growth factor approach, including short half-life, protein instability, undesired dose-related side effects, and higher costs [21]. Small molecules, on the other hand, usually have relatively simple structures, are easy to prepare with less cost, and have a variety of biological functions and applications [22]. Small molecules of both natural and synthetic origin are promising alternatives to biological factors [21].

Resveratrol (RSV) is a polyphenolic phytoalexin, and is regularly obtained from various plants and fruits [23]. It has been demonstrated that RSV has various pharmacological activities, including anti-inflammation [24], antioxidation [25], anticancer [26], and cardiovascular protection [27]. Importantly, RSV can significantly promote bone formation by facilitating osteogenic differentiation and the release of angiogenic factors [28, 29]. Strontium ranelate (SrRn), a strontium (II) salt of ranelic acid, is another small molecule that can increase bone formation and reduces bone resorption [30]. It has dual effects on bone remodeling, with stimulation of osteogenesis and inhibition of osteoclastogenesis [31]. To maximize bone formation, the engineered bone scaffolds are loaded with multiple bioactive small molecules in order to achieve controlled release and targeting of different cell types in the bone tissues.

In the present study, we 3D printed scaffolds consisting of a polycaprolactone/ β-tricalcium phosphate (PCL/TCP) mixture and hydrogel based bioink loaded with RSV, SrRn, or a combination of both. Subsequently, we systematically investigated the effects of drug loaded scaffolds and the released small molecules on the behavior and differentiation of various types of bone cells. We then implanted the scaffolds into a rat model with a critical-sized mandibular bone defect and evaluated in vivo bone formation.

2. Materials and methods

2.1. Generation of 3D printed scaffolds

3D printed scaffolds were fabricated using a high-resolution 3D bioprinter (3D-Bioplotter^® Manufacturer Series, EnvisionTEC; Dearborn, MI), similarly to our previous method [32]. Briefly, PCL (Mw 80,000, Sigma-Aldrich) and TCP (nanocrystals, mean size 100 nm, Berkeley Advanced Biomaterials, Inc.) (1:2 w/w) were dissolved and suspended in a mixed solvent consisting of dichloromethane (DCM, Acros Organics), dimethylformamide (DMF, Acros Organics), and tetrahydrofuran (THF, Fisher Scientific) (1:1:1 v/v/v) and formed a slurry for the 3D printing of the PCL/TCP frames in each layer throughout the construct. Another printing head deposited methacrylated hyaluronic acid (Me-HA, ~1,200 kDa, NovaMatrix) and methacrylated gelatin (Me-Gel, from type B gelatin, Sigma-Aldrich)-based hydrogels between the PCL/TCP frames [33, 34], as shown in Fig. 1A. The size of the 3D printed scaffolds was 20 mm × 20 mm × 1 mm (L×W×H). For the incorporation of the small molecules, RVS (TCI, 1.0 mg/ml) was dissolved in the mixed solvent together with PCL/TCP, and SrRn (Sigma-Aldrich, 1.0 mg/ml) was added to the Me-HA/Me-Gel hydrogel. Therefore, four types of scaffolds were 3D printed, i.e. scaffold only, scaffold with RVS, scaffold with SrRn, and scaffold with both RVS and SrRn, denoted as Scaffold only, RVS, SrRn, and RVS+SrRn, respectively. The 3D printed scaffolds were dried after printing and then punched into small discs using biopsy punches and were UV sterilized for 2 h (⌀ = 6 mm for in vitro characterization and cell culture; ⌀ = 3.5 mm for in vivo implantation).

Figure 1.

(A) Schematic representation of the 3D printing of bone scaffolds incorporated with dual small molecules; (B) Overall in vitro experimental designs and methods.

2.2. Morphological characterization and release behaviors of small molecules

The scaffolds were examined via a scanning electron microscope (SEM, FEI Quanta 200) for morphological assessment. Prior to obtaining the SEM images, a thin layer of gold was sputter-coated onto the material samples. For the evaluation of the small molecule release profile, the drug loaded scaffolds were immersed in individual centrifuge tubes containing 0.5 mL of phosphate buffer saline (PBS, pH 7.4) solution, and all of the tubes were kept in a shaking water bath at 37 °C. The PBS solutions were replaced with fresh PBS solutions at predetermined time intervals, and the collected supernatants at each time point were stored at −80 °C before analysis. The RVS and SrRn concentrations in the releasates were measured using a UV spectrophotometer at the wavelengths of 305 and 318 nm, respectively [35, 36]. The drug-release studies were performed by using five samples for each of the conditions.

2.3. Mouse mesenchymal stem cell (mMSC) culture, seeding, and differentiation

The mMSCs were previously isolated from bone marrow flushes of female FVB mice, as previously described [37]. They were provided by Dr. Leah Cook and are well characterized [38]. The mMSCs were cultured in growth medium consisting of low-glucose DMEM (GE Healthcare Life Sciences), 10% (v/v) fetal bovine serum (FBS, Sigma-Aldrich), 1% penicillin-streptomycin (P/S, Gibco), and platelet-derived growth factor AA (1 μg/100 ml, PeproTech) in 5% CO₂ at 37 °C. The cells were harvested using TrypLE (Gibco), resuspended in growth medium, and seeded onto each scaffold (1×10⁵ cells in 50 μL of growth medium). After 4 h incubation for cell attachment, the growth medium was added. Differentiation was induced 3 days after seeding by replacing the growth medium with the differentiation medium consisting of growth medium plus 100 nM dexamethasone (Sigma), 10 mM β-glycerophosphate (Sigma), and 50 μM ascorbic acid (Sigma) and cultured for another 14 days [39]. Medium was replaced every second day.

The viability of mMSCs seeded on different scaffolds was evaluated by using a Live/Dead assay after 14-day culture [40, 41]. A confocal laser scanning microscope (CLSM, LSM 710, Carl Zeiss) was used to obtain fluorescent images. The cell proliferation of mMSCs on the scaffolds was examined at days 3, 7, and 14 by using an MTT assay [42, 43].

Alkaline Phosphatase (ALP) staining was conducted after 14-day differentiation by using an ALP leukocyte kit (Sigma Aldrich), according to the manufacturer’s instructions [44]. The ALP activity was measured based on our previous method [45].

2.4. Osteoclast induction, culture, and characterization

Osteoclasts were differentiated from primary human peripheral blood monocytes, which were provided by the Elutriation Core Facility at the University of Nebraska Medical Center. The monocytes were seeded at a density of 1×10⁵ cells/well in 24 well plates and differentiated using osteoclast induction medium, consisting of RPMI medium (Gibco), 10% FBS, 1% P/S, 25 ng/ml macrophage colony-stimulating factor (M-CSF, Stemcell), and 50 ng/ml receptor activator of nuclear factor-κB ligand (RANKL, Peprotech) for 14 days.

In order to test the response of osteoclasts to the released drugs, the scaffolds (⌀=6 mm), with or without drugs, were conditioned in the osteoclast differentiation medium at 37 °C for 3 days, and the media were collected (Fig. 1B). Then the differentiated osteoclasts were cultured in the medium (1 ml) containing released drugs for another 2 days. A tartrate-resistant acid phosphatase (TRAP) staining kit (Sigma) was used to evaluate the differentiation of TRAP-positive osteoclasts. The purple colored stained osteoclasts were randomly imaged using an inverted microscope, and the sizes of positive cells were measured and calculated using ImageJ software (at least three images were taken from each sample, and 5 samples were used for each scaffold group). TRAP activity was also measured by the conversion of p-nitrophenyl phosphate (pNPP) to p-nitrophenol (pNP) in the presence of sodium tartrate [46]. The osteoclasts were lysed with lysis buffer and subjected to a freeze-thaw cycle. A cell lysate sample was added to pNPP solution, incubated for 1 h at 37 °C, and then converted to pNP by adding 0.3 M NaOH. The absorbance at the wavelength of 405 nm was detected using a microplate reader. The total protein content was quantified by a Micro BCA Protein assay kit (Pierce). TRAP activity was expressed as μmol of pNP generated per minute per milligram of total proteins (μmol/min/mg protein).

2.5. Human umbilical vein endothelial cell (HUVEC) culture and characterization

HUVECs (Lonza) were grown in endothelial cell basal medium-2 (EGM-2 BulletKit, Lonza) used until passage 6. The cells were cultured in 5% CO₂ at 37 °C, and the medium was replaced every two days. The scaffolds were incubated in the EGM-2 medium at 37 °C for 3 days (Fig. 1B) and then collected. Tube formation and wound healing assays were conducted to evaluate the in vitro angiogenesis and migration. For the tube formation assay, each well of a 96-well plate was filled with 50 μl of cold matrigel and then incubated at 37 °C for 1 h. After the gel formation, HUVECs (1×10⁵ cells/ml, 100 μl/well) were seeded in the matrigel-coated wells and cultured in the medium containing the released drugs from the different scaffolds for 16 h. The formed tubes were observed and photographed randomly with an inverted microscope (Leica). The tube nodes formed by HUVECs in each image were calculated using ImageJ software (at least three images were taken from each sample, and 5 samples were used for each scaffold group). For the migration assay, the 24-well plate was first coated with 300 μl of diluted matrigel (1:100) at 37 °C for 3 h. Then HUVECs were seeded in the coated plates. When the HUVECs reached 70% confluency, a 200 μl sterile pipette tip was used to make a scratch on the cell layer. The un-detached HUVECs were washed three times with PBS buffer and then cultured in medium containing released drugs from different scaffolds. Images of the scratch and the migaration of HUVECs from one edge of the scratch to the other side were taken immediately (t₀ = 0 h) and after 16 h of culture (t₁₆ = 16 h) under an inverted microscope. The migration rate was calculated according to the formula: wound closure%= [(wound length at t₀ - wound length at t₁₆) / wound length at t₀] × 100 %. At least three images were randomly taken from each sample, and 4 samples were used for each scaffold group.

2.6. RNA isolation and quantitative real time polymerase chain reaction (PCR)

Total RNA was extracted from the cell-laden scaffolds or cell monolayers using QIA-Shredder and RNeasy mini-kits (QIAgen). Total RNA was transcribed into complementary DNA (cDNA) using an iScript cDNA synthesis kit (BioRad Laboratories). Real-time PCR analysis was performed in a StepOnePlus^™ Real-Time PCR System (Thermo Scientific) using SYBR Green Supermix (Bio-Rad). The cDNA samples were analyzed for the genes of interest and for the housekeeping gene 18S rRNA. The relative expression of each target gene was calculated using the comparative Ct (2 ^−ΔΔCt) method [47].

2.7. Critical-size rat mandibular defect model

The animal experiment was approved by the University of Nebraska Medical Center (UNMC) Animal Care and Use Committee. Sprague Dawley rats (male, weight 280 – 330 g) purchased from Charles River were used. They were anesthetized with an intramuscular administration of ketamine and xylazine. The surgical areas were shaved, and the skin was sterilized before surgical incision. A longitudinal incision was made along the left mandibular angle. The branches of facial nerves were protected from injuries. The masseter muscle was detached from the mandibular angle, and the outside of the mandibular bone was visible. A full-thickness 4-mm circular defect was drilled through in the mandibular angle with a trephine bur. The 3D printed scaffold was placed into the defect. The muscle and skin incision were closed. Five rats were included in the bone defect only group, six rats in the scaffolds only group, and another six rats in the scaffolds containing both RVS and SrRn group. Routine inspection was carried out in accordance with the protocol. At week 8 after surgery and implantation, the rats were euthanized. The left mandibular bone was collected, and bone formation was evaluated using micro computed tomography (CT) and histological staining.

2.9. Micro CT evaluation

The mandibular bones were fixed with paraformaldehyde and scanned with a high-resolution X-ray micro-CT scanner (SkyScan 1172) using a tube voltage of 55 kV, current of 181 μA, and a slice thickness/slice increment of 8.89 μm. The analysis of the micro-CT data was performed using CT analyzer software (Brucker micro-CT) [48]. From the multiple scan slices, a 3D reconstruction was performed using Micro-CT 3D visualization software. For evaluating new bone formation, the areas of the defect and 3D printed scaffold were selected as the region of interest (ROI). Bone volume (BV) and total volume (TV) in the ROI were measured, and the percentage of bone regeneration was calculated by dividing the bone volume to the total volume (BV/TV).

2.10. Histological staining

After micro-CT scanning, the bone samples were decalcified, paraffin embedded, sectioned into 6-μm slices, and subjected to hematoxylin-eosin (H&E) and Masson’s Trichrome staining [49]. Histological staining images were acquired with an optical microscope.

2.11. Statistical analysis

All quantitative data are expressed as the mean ± standard deviation (SD). Statistical analysis was performed using ANOVA with Scheffé post-hoc tests. A p-value of < 0.05 was considered statistically significant.

3. Results

3.1. 3D printing of small molecule loaded scaffolds and their characterization

We integrated a hydrogel-based bioink together with composite PCL/TCP to generate 3D bone constructs. RVS is hydrophobic and can be easily dissolved in the organic solvent with PCL/TCP. SrRn is freely soluble in acidic aqueous solutions, but it is slightly soluble in pure water (~3.7 mg/ml) [50]. In this study, we used a relatively low concentration (1 mg/ml) of SrRn in the Me-HA/Me-Gel based bioink. These two inks, with or without small molecule drugs, were printed in an alternating order (Fig. 1A). PCL/TCP was printed first, followed by 3D printing of the hydrogel in the groove created by the first material on the same layer. Then, layer-by-layer, we printed the initial large constructs with a square shape. The thickness is comparable to the mandibular bone thickness in the rat model. The punched scaffolds with various drugs were shown in Fig. 2A. The scaffolds were porous, and the scaffolds with RVS showed slightly yellowish color. To confirm the morphology, the 3D printed scaffolds, with or without drugs, were observed by SEM. Fig. 2B shows that all of the scaffolds had the PCL/TCP strut size of approximately 400 μm and that the hydrogels were successfully printed next to PCL/TCP, with a width of approximately 400 μm. The incorporation of small molecules did not affect the printing process or the printed scaffold structure.

Figure 2.

Characterization of 3D printed scaffolds. (A) Representative images of different 3D printed scaffolds after punching: (a) scaffold only; (b) scaffold with RVS; (c) scaffold with SrRn; (d) scaffold with RVS+SrRn; (B) Typical SEM micrographs of 3D printed scaffolds with and without the incorporation of small molecules (scale bars: 1 mm for the general view and 300 μm for the close view); (C) Cumulative release of RVS and SrRn from 3D printed scaffolds after incubation in PBS at 37 °C.

The 3D printed scaffolds loaded with RVS and SrRn were subjected to in vitro release studies for 3 weeks. The small molecule loading amount was calculated to be 1.01 ± 0.05 μg/mg scaffold for RVS and 0.58 ± 0.02 μg/mg scaffold for SrRn. Since RVS loaded PCL/TCP were printed as the frame structure and SrRn loaded hydrogels were printed in between, more PCL/TCP, and thus RVS, were in the printed scaffolds. The cumulative release profiles of RVS and SrRn from the 3D printed scaffolds were shown in Fig. 2C. In vitro release studies demonstrated that SrRn had a significant burst release, resulting in more than 70% of the total drug being released at day 1. After the burst release stage, SrRn had sustained release for at least 21 days. In contrast, the release of RVS was very slow, and less than 30% of RVS was released throughout the 21-day release period. It is well known that PCL has a very slow degradation rate. Although TCP has a faster degradation rate compared to hydroxyapatite, the overall degradation rate of the composite scaffold is also slow. After the 21-day release study, all scaffolds were observed to not shrink, curl, or delaminate.

3.2. Effects of scaffolds on proliferation and osteogenic differentiation of mMSCs

We first seeded mMSCs onto various scaffolds to evaluate cell viability and the effects of drug release on mMSC osteogenic differentiation. The cells cultured on all of the scaffolds, with or without small molecules, showed high cell viability, and very few dead cells were observed (Fig. 3A). All the cells were well attached and showed a normal spreading morphology. An MTT assay was performed to determine the proliferation of mMSCs on the scaffolds. The results showed that the cells proliferated well on all of the scaffold groups with the increase of the culture time (Fig. 3B). The proliferation rate on the scaffolds with SrRn was highest throughout 14-day culture and was significantly higher than that on the scaffold alone and scaffolds with RVS at day 7. However, the MTT absorbance for mMSCs on all of the scaffold groups were comparable after 14-day culture.

Figure 3.

Effects of 3D printed scaffolds, with and without the incorporation of small molecules, on mMSC viability, proliferation, and osteogenic differentiation. (A) Typical live/dead staining images (scale bars: 100 μm, red: dead cells, green: live cells) of mMSC on different scaffolds; (B) MTT assay for mMSC proliferation seeded on various scaffolds (n=5, bars that do not share letters are significantly different from each other (p<0.05), # indicates a significant difference compared to the scaffold only group); (C) ALP staining images of the various scaffolds with mMSCs after osteogenic induction (scale bars: 500 μm); (D) ALP activity test (n=5, ***p<0.001); (E) qPCR analysis of ALP, Runx2, OCN, and Col 1A1 genes for mMSC seeded on different scaffolds after 14 day differentiation (n=3; ***p < 0.001, bars that do not share letters are significantly different from each other (p<0.05)).

After standard osteogenic induction, mMSCs on all of the scaffolds were positive to ALP staining (Fig. 3C). The 3D printed scaffolds loaded with both RVS and SrRn showed significantly higher ALP activity compared to the other three scaffold groups (Fig. 3D). Consistently, the ALP gene expression for mMSCs on RVS and SrRn loaded scaffolds was also significantly upregulated, as shown in Fig. 3E. The qPCR results also showed that the incorporation of SrRn within the scaffold statistically increased the expression of Runt-related transcription factor 2 (RUNX 2), osteocalcin (OCN), and collagen 1A1 (Col 1A1) (Fig. 3E). These results demonstrated that the combination of RVS and SrRn within the 3D printed scaffolds significantly promoted mMSC osteogenic differentiation.

3.3. Effects of released small molecules on monocyte induced osteoclast behaviors

In order to determine the effects of 3D printed scaffolds, with or without small molecules, on osteoclast activities, we incubated the scaffolds in osteoclast induction medium for 3 days and then utilized the released medium to treat human primary monocyte induced osteoclasts for another 2 days. As shown in Fig. S1, osteoclast-like cells with multiple nuclei and positive expression of F-actin ring and cathepsin-K were successfully differentiated from monocytes after 14-day induction. After 3-day incubation, the released RVS and SrRn were calculated to be 2.15 ± 0.13 μg/ml and 7.59 ± 0.40 μg/ml, respectively. We first evaluated the effects of the released small molecules on osteoclast morphology and size. Fig. 4A shows that the osteoclasts were positive to TRAP staining, and the size and number of the osteoclasts were decreased after incubation in the medium containing released small molecules. The quantitative measurement of osteoclast sizes confirmed that all of the treatments with released RVS, SrRn, or their combination significantly decreased osteoclast size, compared to the scaffold alone group (Fig. 4B). The osteoclasts had the smallest size after culture in the media with released SrRn. The TRAP activity was also evaluated, as shown in Fig. 4C. No significant difference was found between the scaffold alone and scaffold with RVS groups. However, compared to these groups, the other two groups with released SrRn showed significantly decreased TRAP activity.

Figure 4.

Effects of released small molecules on human monocyte-derived osteoclast morphology, size, TRAP activity, and gene expression. (A) TRAP staining images of osteoclasts treated with released media with RVS and SrRn (scale bars: 200 μm); (B) Measurement of osteoclast size based on TRAP staining images (at least three images were taken from each samples, and 5 samples were used for each scaffold group; ***p<0.001, n.s. indicates no significant difference); (C) TRAP activity (*p<0.05); (D) qPCR analysis of NFATc1, MMP-9, and CTSK genes for osteoclasts (n=3; *p<0.05, ***p < 0.001).

We further examined gene expression of nuclear factor of activated T cells 1 (NFATc1), matrix metalloproteinase-9 (MMP-9), and cathepsin K (CTSK). NFATc1 is an important transcription factor that is required for sufficient osteoclast differentiation [51]. MMP-9 is a type IV collagenase and is highly expressed in osteoclasts [52]. MMP-9 plays an important role in extracellular matrix degradation and cell migration. Both NFATc1 and MMP-9 are upregulated at the early stage of osteoclastogenesis [53]. CTSK is predominantly expressed by multinucleated osteoclasts and can degrade collagen and other matrix proteins during bone resorption. The qPCR results showed that the released SrRn significantly downregulated the expression of NFATc1 and MMP-9, but it did not change the expression of CTSK (Fig. 4D). In contrast, RVS increased the expression of NFATc1 and had no effects on MMP-9 and CTSK expression. The combination release of SrRn and RVS only downregulated NFATc1expression but not MMP-9.

3.4. Effects of released small molecules on HUVEC tube formation and migration

Released medium (EGM medium) with RVS, SrRn, or their combination was also collected to condition HUVECs. A tube formation assay was conducted to evaluate in vitro angiogenesis of HUVECs in response to various released small molecules. In this assay, HUVECs were seeded onto the Matrigel (containing basement membrane matrix) and developed a vascular tubular structure (Fig. S2A). The numbers of branch sites/nodes were quantified. As shown in Fig. 5A, the node density for HUVECs was significantly increased after incubation in all the three types of release media, compared to the scaffold alone group. No statistical difference was seen among the three groups with various small molecules. We also determined how the release medium affected HUVEC migration by using a wound healing assay. The “wound” was created on the confluent monolayer of HUVECs by the mechanical removal of a portion of cells using a pipette tip. Then the HUVECs started migrating to heal the “wound” (Fig. S2B). After 16 h migration, we found that HUVECs treated with RVS medium showed the highest wound closure ratio, which was significantly higher than in scaffold alone and SrRn groups (Fig. 5B).

Figure 5.

Effects of released small molecules on HUVEC tube formation, migration, and gene expression. (A) Node density measured based on the formed tubes (n=4, **p<0.01); (B) Wound closure percentage by measuring the distance of HUVEC migration in the wound scratch assay (n=4, *p<0.05, **p<0.01); (C) qPCR analysis of VEGFA, PECAM, and vWF genes for HUVEC (n=3; bars that do not share letters are significantly different from each other (p<0.05)).

We also examined the gene expression of vascular endothelial growth factor A (VEGFA), platelet-endothelial cell adhesion molecule (PECAM)-1, and von Willebrand factor (vWF), all of which are important for angiogenesis and HUVEC functions. The qPCR results demonstrated that HUVECs had the highest levels of VEGFA, PECAM, and vWF gene expression in the RVS release medium, compared to those in other release medium conditions (Fig. 5C). Moreover, the medium with both RVS and SrRn also significantly upregulated angiogenesis related gene expression, compared to scaffold alone and SrRn groups.

3.5. 3D printed scaffolds with dual small molecules promoted in vivo bone formation

To evaluate bone regeneration, we implanted the 3D printed scaffolds into rats with critical-size mandibular bone defects for 8 weeks (Fig. 6A). Since the incorporation of both RVS and SrRn showed promising combinational effects on osteogenesis, angiogenesis, and osteoclast inhibition in vitro, we implemented the scaffolds with both small molecules and the scaffold alone as the control for the mandibular implantation and compared their in vivo bone formation capacity. The harvested mandibular bone samples were first scanned using micro-CT, to evaluate the bone healing. Representative 3D reconstructions of the micro-CT images for the various implant groups are shown in Fig. 6B. After 8-week implantation, the defect only group showed unfilled defects with very limited bone regeneration. In the scaffold only group, some new bone formation at the periphery of the bone defects was observed, but the center of the defect was still filled with scaffold rather than bone tissue. The 3D printed scaffolds with the incorporation of RVS and SrRn facilitated bone formation (Fig. 6B). The formed new bones were detected not only around the scaffold struts but also on the top of the scaffold. A quantitative measurement of new bone volume percentage (BV/TV ratio), based on micro-CT analysis, confirmed that the 3D printed scaffolds with both RVS and SrRn elicited remarkable new bone formation in comparison to the other two groups (Fig. 6C) After 8-week implantation, all of the composite scaffolds were still detectable, as expected.

Figure 6.

Implantation of 3D printed scaffolds with and without small molecules into the rat model with critical-size mandibular defect. (A) Schematic drawing to show the implantation location; (B) Representative images of the animals with mandibular defects and with implantation of 3D printed scaffolds loaded with RVS and SrRn; (C) Representative micro CT reconstruction images at 8 weeks after surgery in different groups; (D) Micro CT quantification of bone volume to the total volume of the former defect area (BV/TV, n=5–6, *p<0.05, ***p<0.001).

Histological and immunofluorescent stainings were performed after micro-CT analysis to confirm the extent of bone formation and vascularization. For H&E staining, connective soft tissues, rather than mineralized tissues, were observed in the defect only group without any scaffold implantation (Fig. 7A). In contrast, new bone tissue formation partially occurred in the defect implanted with 3D printed scaffolds. Notably, new bone formation in the defect implanted with RVS and SrRn loaded 3D printed scaffolds occurred at the periphery as well as the center of the defect. In Masson’s Trichrome staining (Fig. 7B), blue represents mineralized collagen tissue and red indicates muscle fibers. The staining results also confirmed that more new bone formation was observed in defect sites treated with dual small molecule loaded 3D printed scaffolds.

Figure 7.

Histological staining and observations of the newly formed bone in the three groups. (A) H&E staining; (B) Masson’s Trichrome staining. Scale bar: 1 mm for general view, 400 μm for close view. White arrows indicate soft tissue, * indicate scaffolds, black arrows indicate newly formed bones.

4. Discussion

It is well known that bone is a metabolic organ that possesses properties of osteoclastic bone resorption and osteogenic bone formation [54]. The bone marrow derived MSCs, osteoblasts, and osteoclasts play important roles in modulating mineral homeostasis [55]. Vascularization can have a great effect on bone regeneration by increasing the delivery of nutrients, oxygen, growth factors, and stem cells[56]. Our current study aims to fabricate 3D printed scaffolds with the incorporation of multiple small molecules to target multiple bone cell types (i.e. MSCs, osteoclasts, and endothelial cells) for promoting bone regeneration.

We implemented RVS and SrRn to demonstrate the feasibility and in vivo efficacy. RVS is soluble in organic solvents, while SrRn is water soluble. These two small molecule drugs were thus incorporated into a PCL/TCP-based ink and a Me-HA/Me-Gel-based bioink, respectively. Ideally, the incorporated small molecules should be released in a stable and sustainable manner in order to provide continuous bioactivities on targets [57]. In our study, RVS had a very limited burst release, while SrRn showed a significant burst release in first 24 hours in PBS solution. Both RVS and SrRn had sustained release over 3 weeks. A study from Tian et al. also showed a similar release profile of SrRn from a chitosan film coated onto a titanium surface [58]. We have used a similar strategy to fabricate 3D printed scaffolds loaded with multiple antibiotics to mitigate an established biofilm and clear craniotomy-associated infection [32].

RVS was reported to increase ALP activity via SIRT1 activation of BMP2 [59, 60], induce osteoblastic differentiation [61], and promote mineralization of mMSCs [62]. RVS also activated the Wnt/β-catenin signaling pathway, upregulated lncRNA KCNQ1OT1, and inhibited polymethylmethacrylate particle-induced osteolysis in a mouse model [63]. In late stages of osteogenic differentiation, RVS activated Notch signaling to promote differentiation of osteoblasts into osteocytes [64]. Previous studies showed that RVS could also inhibit the osteoclastogenesis to reduce alveolar bone loss [65] and improved angiogenesis [66] in rodent models. Our results are consistent with reported studies and demonstrate that the 3D printed scaffolds with RVS promoted osteogenic differentiation of MSCs. The pharmacological effects of RVS on cell proliferation are closely related to drug dose/concentration [67, 68]. RVS was found to enhance the proliferation of human umbilical cord derived MSCs (hUC-MSCs) at concentrations from 0.1 μM to 2.5 μM. However, increased doses of RVS (from 5 μM to 10 μM) conversely decreased the growth of hUC-MSCs [69]. A higher concentration of RVS (25 μM) was reported to have a negative impact on MSCs proliferation but promote osteogenic differentiation [70]. In our study, the controlled release of RVS did not significantly affect mMSC proliferation, demonstrating its safety. The effects of RVS on endothelial cells are also dose dependent [71]. At a low concentration, RVS was reported to increase HUVEC migration and tube formation, whereas a high concentration of RVS inhibited HUVEC angiogenesis [72]. Our results demonstrated that the concentration of sustained-release RVS was in a safe and effective range to promote HUVEC angiogenesis in vitro.

Also, in our study, SrRn was incorporated to target MSCs/osteoblasts and osteoclasts. SrRn was previously reported to possess properties that affect bone metabolism, including inhibition of osteoclastic resorption and promotion of osteogenesis [73–76]. SrRn has been approved in Europe since 2004 for the treatment of osteoporosis [77]. SrRn has been demonstrated to enhance trabecular thickness and bone volume in both pre-clinical models and human studies [77, 78]. The fundamental mechanisms have been extensively investigated. It was reported that the osteogenic improvement by Sr ions involved the overexpression of the ROCK1 [79] and Setd2 genes [80], Wnt/β-catenin pathway activation [81], and NF-kB signaling inhibition [82]. SrRn also enhanced angiogenesis [83] via activating the PI3K/AKT/mTOR signaling pathway [84] and regulating the inflammatory response during in vivo bone formation [85]. Similar to RVS, the effects of SrRn also are also dose dependent [85, 86]. For example, 5% SrRn mixed with PCL/polydiisopropyl fumarate (PDIPF) was more likely to increase inflammation and inhibit osteogenesis, while PCL/PDIPF with 1% SrRn could stimulate new bone formation [87]. In our study, we demonstrated that SrRn promoted osteogenic differentiation of mMSCs and HUVEC migration and tube formation. Most importantly, SrRn significantly decreased osteoclast size, activity, and gene expression.

As previously discussed, many studies noted that both RVS and SrRn had effects on various bone related cells, but very few studies have evaluated their comparative pharmacological activities and evaluated their potential synergistic effects on bone regeneration. In our current study, the 3D printed scaffolds with the incorporation of both RVS and SrRn combined the features of RVS and SrRn for promoting angiogenesis and inhibiting osteoclast activities at the same time. The released RVS and SrRn also showed synergistic effects on osteogenic differentiation of MSC. We further implemented a rat model with a critical-size mandibular bone defect and implanted the 3D printed scaffolds with and without the dual small molecules. Our 3D printed scaffolds with combined release of RVS and SrRn significantly improved bone formation. There are multiple factors contributing to the enhanced bone regeneration in our study. The sustained release of low dose RVS may promote angiogenesis and enhance bone formation by facilitating osteogenesis of progenitor cells and stem cells. The released SrRn may temporally inhibit osteoclastogenesis and thus may decrease bone resorption. The effects of RVS and SrRn may be synergistic or additive, and many signaling pathways may potentially be involved.

In order to enhance the regeneration efficacy of bone tissue engineering scaffolds, growth factors have been incorporated into the scaffolds, often resulting in improved outcomes [88]. However, despite reported success in animals, the clinical use of growth factor-impregnated scaffolds is still hampered by a number of challenges, including stability, cost, and control of the release rate [89]. Many small molecules have been identified to possess bioactivity similar to growth factors and could mimic their function in vivo [90]. The application of small molecules within scaffolds is becoming a potential alternative to growth factors, as they can overcome many disadvantages related to growth factors, particularly with stability, convenience, and immunogenicity. Addressing the interdependence of angiogenesis and osteogenesis, multiple growth factors have been loaded into scaffolds to simultaneously promote angiogenesis and osteogenesis in order to improve bone regeneration [91, 92]. However, this strategy is highly complicated, and the dose and delivery of different growth factors need to be tailored carefully to reduce the potential for untoward side effects. Our studies show that the combined use of the small molecules such as RVS and SrRn loaded into bioprinted scaffolds can improve bone formation in vivo and is a promising alternative to multi-growth factor approaches previously reported.

One limitation of the current study is that SrRn is not released in a well-controlled manner due to its water solubility. Although SrRn can affect osteoclastogenesis at a relatively low concentration, the medium extracted after 3-day release may not have any effect on osteoclasts and HUVECs. For the in vivo implantation, the composite scaffolds may have a slower release rate in the bone environment compared to under aqueous in vitro conditions. A more stable and sustained release pattern might improve the pharmacological effects. In order to reduce or altogether eliminate the burst release of loaded SrRn, a valid and stable delivery system should be established. For example, we can fabricate microspheres loaded with SrRn in the future and incorporate the SrRn loaded microspheres into the hydrogel-based bioink for the scaffold fabrication.

5. Conclusion

In summary, we implemented multiple bioinks to 3D print bone scaffolds loaded with two small bioactive molecules, RVS and SrRn. RVS had a more sustained release profile, while SrRn had an initial burst release then a sustained release for three weeks. The release of RVS and SrRn from the 3D printed scaffolds targeted multiple cells, including MSCs, endothelial cells, and osteoclasts. The released RVS significantly promoted HUVEC migration, tube-like structure formation, and angiogenic gene expression in vitro. The 3D printed scaffolds with SrRn also greatly increased mMSC proliferation and osteogenic differentiation. Both RVS and SrRn decreased osteoclast size, and the released SrRn also significantly decreased osteoclast activities and related gene expression. The bone scaffolds with both RVS and SrRn had combinational advantages in enhancing angiogenesis, inhibiting osteoclast activities, and synergistically promoting mMSC osteogenic differentiation. We further demonstrated that the 3D printed scaffolds with small molecules significantly promoted bone formation after implantation in rats with critical-sized mandibular bone defects.