An In Vivo Assessment of Different Mesenchymal Stromal Cell Tissue Types and Their Differentiation State on a Shape Memory Polymer Scaffold for Bone Regeneration

Abstract

A combined biomaterial and cell-based solution to heal critical size bone defects in the craniomaxillofacial area is a promising alternative therapeutic option to improve upon autografting, the current gold standard. A shape memory polymer (SMP) scaffold, composed of biodegradable poly(ε-caprolactone) and coated with bioactive polydopamine, was evaluated with mesenchymal stromal cells (MSCs) derived from adipose (ADSC), bone marrow (BMSC), or umbilical cord (UCSC) tissue in their undifferentiated state or pre-differentiated toward osteoblasts for bone healing in a rat calvarial defect model. Pre-differentiating ADSCs and UCSCs resulted in higher new bone volume fraction (15.69% ± 1.64%) compared to empty (i.e., untreated) defects and scaffold-only (i.e., unseeded) groups (4.41% ± 1.11%). Notably, only differentiated UCSCs exhibited a significant increase in new bone volume, surpassing both undifferentiated UCSCs and unseeded scaffolds. Further, differentiated ADSCs and UCSCs had significantly higher trabecular numbers than their undifferentiated counterparts, unseeded scaffolds, and untreated defects. Although the mineral density regenerated within the unseeded scaffold surpassed that achieved with cell seeding, the connectivity of this bone was diminished, as the regenerated tissue confined itself to the spherical morphology of the scaffold pores. The SMP scaffold alone, with undifferentiated BMSCs, with undifferentiated and differentiated ADSCs, and differentiated UCSCs (29.72 ± 1.49 N) demonstrated significant osseointegration compared to empty defects (14.34 ± 2.21 N) after 12 weeks of healing when assessed by mechanical push-out testing. Based on these results and tissue availability to obtain the cells, pre-differentiated ADSCs and UCSCs emerge as particularly promising candidates when paired with the SMP scaffold for repairing critical size bone defects in the craniofacial skeleton.

1 |. Introduction

In both the civilian and military populations, severe bony defects resulting from blunt and penetrating trauma—such as those from vehicle accidents, sports injuries, gunshot, and blast injuries—continue to be a challenging repair scenario for maxillofacial surgeons. The resulting consequences include patient morbidity, prolonged inpatient care, and an increased economic burden on the healthcare system. Epidemiology studies have shown that globally in 2019, there were over 18 million fractures of the skull and facial bones [1]. In the war-fighter population, recent conflicts have resulted in 26% of the injured warfighters enduring craniomaxillofacial trauma, with 27% of those injured warfighters sustaining facial fractures [2]. The estimated annual costs for treating bony defects is in the billions of dollars [3, 4].

The most challenging of bony defects to treat are critical size defects, wherein significant bone tissue loss precludes healing without therapeutic intervention. Although autografting is the gold standard for treating such defects, it carries potential risks, including donor site morbidity and limited availability [5, 6]. In a study of patients that underwent autologous cranioplasty, for example, they noted a failure rate of 19.6% [7]. Allografts can overcome the challenge of limited availability, but they are sub-optimal as they can be immunogenic and have reduced osteo-inductive capabilities resulting from storage and sterilization techniques [8–10]. These limitations in the current standards of care underscore the need for improved reconstruction options for surgeons.

Tissue engineering represents an alternative approach wherein tissue regeneration is imparted via the use of a synthetic scaffold, and traditionally also employs exogeneous cells and/or exogeneous growth factors. To promote bone healing, the scaffold would ideally exhibit osteoconductivity (i.e., permitting cellular migration and neotissue infiltration by nature of porosity and biodegradability), osseointegration (i.e., the ability to integrate with surrounding native tissue), and osteoinductivity (i.e., the ability to promote cellular osteogenic differentiation) [11]. Paramount to osseointegration is achieving good scaffold contact with adjacent bone tissue and is a particular challenge for irregularly shape defects. A scaffold that could be easily shaped to fit within the defect would also reduce time in the surgery room [12]. Osteoinductivity is usually obtained by incorporation of a biotherapeutic such as mesenchymal stromal cells (MSCs), recombinant growth factors, exosomes, and/or cell secretomes [13–15]. Still, questions remain in regards to the optimal MSC tissue type to use for bone healing, as MSCs from different sources can have varying degrees of osteogenic potential [16, 17]. Although bone marrow-derived MSCs generally demonstrate greater osteogenic potential than other tissue-derived MSCs, they also tend to have a lower proliferation capacity and overall availability, as compared to adipose tissue or fetal-derived MSCs [18–20]. Additionally, different MSC tissue types exhibit preferences in regards to scaffold materials (i.e., one MSC tissue type may grow and/or differentiate better on one scaffold, while another MSC tissue type can grow and/or differentiate better on a completely different scaffold) [21]. Thus, an effective tissue engineering approach to heal critically size defects presents a demanding set of requirements for the scaffold and cellular components.

We have investigated the utility of a “self-fitting” scaffold prepared from biodegradable shape memory polymer (SMP), poly(ε-caprolactone) (PCL), coated with bioactive polydopamine, and MSCs incorporated [22]. Scaffolds were prepared via a solvent-casting particulate leaching (SCPL) and then coated with polydopamine (tens of nm thick) by exposure to a dopamine solution to afford high porosity (~70%), interconnected pores (average diameter ~220 μm), and non-brittle mechanical properties that include a compressive modulus of ~10 MPa. The polydopamine coating was observed to induce acellular hydroxyapatite mineralization when exposed to simulated body fluid (SBF, 1X). The shape memory nature is imparted by covalent crosslinks that serve as net-points and crystalline lamellae (T_m or “T_trans” ~55°C) that act as switching segments. Thus, exposure to warm saline (T > T_trans) causes the scaffold to become malleable such that it can be press-fitted into defects. Shape recovery causes scaffold expansion to the perimeter, and subsequent cooling to body temperature (T < T_trans) fixes the scaffold in the new shape while also restoring rigidity. In a previous study, our group demonstrated that this SMP scaffold is osteo-conductive to cultured MSCs derived from bone marrow, adipose, umbilical cord, amniotic membrane, and chorionic membrane tissue, and capable of binding those MSCs and supporting their growth and osteogenic differentiation in vitro [23]. Further, it was found that, of the five different MSC-derived tissue types, cultured bone marrow MSCs (BMSCs) demonstrated optimal growth and differentiation on the SMP scaffold, followed by MSCs derived from adipose tissue (ADSC) and umbilical cord (UCSCs).

In this study, we compare the in vivo regenerative capacity of BMSCs, ADSCs, and UCSCs seeded onto the SMP scaffold in a rodent critical size defect model. In addition, we evaluated whether or not there is any distinct advantage toward pre-differentiating the MSCs toward osteoblasts on the SMP scaffold prior to implantation, as compared to implanting the SMP scaffold with undifferentiated MSCs. Based on our prior report [23], we hypothesized that pre-differentiated MSCs seeded on the SMP scaffold would result in optimal bone healing and osseointegration compared to undifferentiated MSCs.

2 |. Materials and Methods

2.1 |. Materials

PCL-diol (M_n ~10,000 g mol⁻¹), triethylamine (Et₃N), acryloyl chloride, 2,2-dimethoxy-2-phenyl acetophenone (DMP), 4-dimethylaminopyridine (DMAP), 1-vinyl-2-pyrrolidinone (NVP), sodium chloride (NaCl), dopamine hydrochloride, sodium bicarbonate (NaHCO₃), potassium chloride (KCl), potassium phosphate dibasic trihydrate (K₂HPO₄·3H₂O), magnesium chloride hexahydrate (MgCl₂·6H₂O), sodium sulfate (Na₂SO₄), calcium chloride (CaCl₂), tris-hydroxymethyl aminomethane (Tris), L-ascorbic acid-2-phophate (AsAP), and all solvents were obtained from Sigma-Aldrich (St. Louis, MO). Anhydrous magnesium sulfate (MgSO₄) and Tris buffer (2 mol L⁻¹) were obtained from Fisher Scientific (Hampton, NH). Reagent-grade CH₂Cl₂ (dichloromethane, DCM) was dried over 4 Å molecular sieves before use.

The following materials were purchased from Promocell (Heidelberg, Germany): BMSCs, ADSCs, and UCSCs, HEPES buffer, Accutase, growth media for MSCs and osteogenic differentiation media. The cell types purchased from Promocell undergo rigorous quality control by the company including validation of current differentiation lineage by assessing the presence of CD73, CD90, CD105, CD14, CD19, CD34, CD45, HLA-DR. A ORFLO Moxi Z cell counter with type M cassettes was used to determine cell counts (ORFLO, Ketchum, ID). An Orbi-shaker CO₂ (Orbital Shakers, Atkinson, NH) was used while culturing cells on scaffolds. Denatured ethanol was purchased from Decon Labs Inc. (King of Prussia, PA). Low cell attachment, 12 well plates and Dulbecco’s phosphate-buffered saline (DPBS) was purchased from Thermo Fisher Scientific (Waltham, MA).

Isoflurane (Patternson Veterinary, Loveland, CO) was used to keep the animals anesthetized. Analgesia was maintained with Ethiqa XR from MWI Animal Health (AmerisourceBergen, Conshohocken, PA). Trephine burrs with a diameter of 8 mm were purchased from Henry Schein (Melville, NY) and the sutures used for surgery came from Ethicon (Johnson & Johnson, New Brunswick, NJ). Tissues were fixed in neutral buffered, 10% formalin from Sigma-Aldrich (Burlington, MA). Cal-Ex Decalcifier was obtained from Fisher Chemical (Pittsburgh, PA).

2.2 |. Animal Study Approval

The study protocol was approved by the Institutional Animal Care and Use Committee at the 711th Human Performance Wing, Joint Base San Antonio-Fort Sam Houston, and conducted in accordance with the Guide for the Care and Use of Laboratory Animals, Institute of Laboratory Animals Resources, National Research Council, National Academy Press, 2011. All procedures were performed in facilities accredited by the Association for Assessment and Accreditation of Laboratory Animal Care International. There were 11–13 rats per treatment group. This sample size was calculated based on previous studies which evaluated bone regeneration and mechanical integrity using rat calvarial defects [24–28] using a statistical power of 0.80 and significance level (alpha) of 0.05 to detect a mean difference of 50% between untreated and treated groups with a standard deviation of 40%.

2.3 |. SMP Scaffold Fabrication

PCL-diacrylate (PCL-DA) macromer (M_n ~10,000 g mol⁻¹) was synthesized, and PCL scaffolds formed via SCPLC as previously reported [22]. Briefly, salt templates were prepared in 20 mL glass scintillation vials (I.D. = ~25 mm) by NaCl salt particles (10.0 g, 460 ± 70 μm) mixed with deionized (DI) water (7.5 wt%). The mixture was mechanically stirred, compacted with a blunt glass rod, sealed vials centrifuged (15 min, 3220 × g), opened vials air-dried (~1 h), and dried in vacuo (room temperature, RT; overnight, ON; 30 in. Hg). A PCL-DA solution (0.15 g mL⁻¹ in DCM) was prepared and combined with a photoinitiator solution (15 vol%, 10 wt% DMP in NVP). Following vortexing, aliquots of this PCL-DA macromer solution (~5 mL) were then cast into salt templated vials and sealed. The vials were subsequently centrifuged (10 min, 1260 × g) to promote macromer diffusion into the template. Next, the vials were opened and exposed to UV light (UV Transilluminator, 6 mW cm⁻², 365 nm) for ~5 min then air-dried ON. Following, SMP scaffolds were soaked in a water: ethanol mixture (1:1 vol:vol) for 24 h to remove the template.

Scaffolds were then removed from vials and soaked for an additional 4 days in the solvent mixture (changed daily) to leach out any remaining salt. Next, the PCL SMP scaffolds were air dried ON and annealed (85°C, 1 h, 30 in. Hg). After cooling, the cylindrical scaffold specimens were sliced using a Vibratome (Leica VT 1000 S) in order to achieve a uniform thickness (~1 mm). The top and bottom sections (~2 mm) were discarded.

To coat polydopamine onto the pore walls, the sliced scaffolds were degassed via syringe and submerged into a dopamine hydrochloride solution (2 mg mL⁻¹ in 10 mM Tris buffer, pH = 8.5) at 150 rpm (shaker table) for 16 h. The scaffolds were then extensively rinsed with DI water, dried in vacuo ON (RT, 30 in. Hg), and punched to 8 mm (Integra Miltex, 8 mm), yielding final scaffold dimensions of ~8 mm diameter × ~1 mm thickness. Scaffold specimens were subjected to ethylene oxide (EtO) sterilization (Anderson, Anprolene AN74i/x) as previously described [29]. Per manufacturer specifications, the unit operates at temperatures of ∼30°C–35°C and ∼35% relative humidity (Humidichip).

2.4 |. Cell Culture and Seeding Cells on SMP Scaffolds

The BMSC group was comprised of a pool of four donors (age, gender: 47, male; 62, male; 63, male; 91, female). The ADSC group was comprised of a pool of three donors at (age, gender: 24, female; 30, female; 51, female). The UCSC group is comprised of a pool of three infant donors (1 male and 2 female). The ADSCs, BMSCs, and UCSCs were cultured in MSC growth medium at 37°C and 5% CO₂. The cells were subcultured when they reached 80%–90% confluency by washing with HEPES buffer and removing from the surface with Accutase. The cultured cells were counted on a Moxi Z cell counter and expanded at 3000–4000 cells/cm2. The cells were seeded on scaffolds for implantation no later than passage five.

Cells were seeded on the SMP scaffolds according to a previous report [23]. The scaffolds were pre-wetted by incubating in a vacuum desiccator for 15 min in cell culture medium. Next, the scaffold was tapped lightly on a sterile blotting pad. Then, 250,000 cells were seeded across the top of the scaffold in an ultra-low binding 12-well plate containing no medium or buffer solution followed by a 15 min incubation at 37°C and 5% CO₂. Following incubation, the scaffolds were removed from the well, tapped lightly on a sterile blotting pad, and returned to the well plate so the cell-seeded side was on the bottom. Then, 250,000 cells were pipetted onto the second side of the scaffold. The scaffold was incubated for 45 min at 37°C and 5% CO₂, then 2-mL growth medium was added to each well. Scaffolds for undifferentiated groups were incubated in growth medium for 4 days then used for surgeries. Scaffolds for pre-differentiated groups were incubated in growth medium for 4 days to allow the cells time to proliferate, then they were transferred to osteogenic differentiation media for 14 days. The MSCs were pre-differentiated in osteogenic medium for 14 days to induce osteogenic commitment prior to implantation as other studies have demonstrated increased bone regeneration when using pre-differentiated MSCs prior to implantation for bone healing [27, 30–32]. All scaffolds were cultured at 37°C and 5% CO₂ at 60 rpm, using an orbital shaker with 19-mm diameter.

2.5 |. Animal Surgeries

Surgery was performed on male RNU (Crl:NIH-Foxn1^rnu) rats weighing 210–320 g. Rats were purchased from Taconic Biosciences (Germantown, NY). As RNU rats are immune-compromised, all contents of the cage were purchased sterile or were autoclaved prior to use. The rats were singly housed in individually ventilated cages (IVC) with ad libitum access to sterile food and water. Husbandry and surgery were completed in high efficiency particular air (HEPA) filtered hoods under aseptic conditions. For surgery, the rats were induced and maintained under anesthesia using Isoflurane. The rat’s head was isolated and fixed using a stereotaxic device with non-puncture ear bars coated with lidocaine jelly. The scalp was previously shaved and once fixed in place, was surgically prepared using a cotton tipped applicator soaked in iodine followed by an applicator soaked in alcohol three times. Once prepped, the surgical site was sterily draped. An incision was made down the middle of the skull to expose the periosteum. The periosteum was incised and dissected away from the bone using cotton tipped applicators. A calvarial defect was made in the skull using an 8-mm trephine burr and a dental drill between the bregma and lambda. The bone was carefully removed from the skull using periosteal elevators. The defect was then filled with an SMP scaffold treated with undifferentiated or pre-differentiated ADSCs, BMSCs, or UCSCs, with an untreated scaffold (scaffold only), or left empty (empty defect). Once the scaffold was in place, the periosteum was closed with absorbable suture (4–0 Monocryl) and the skin was closed with nonabsorbable suture (4–0 Ethilon). The rats were recovered and monitored for either 4 or 12 weeks.

At the conclusion of the study period, rats were euthanized, and the skull cap was harvested. The skull cap was removed using a dental drill with a side cutting burr cutting across the middle of the nasal bone, along the sides of the skull approximately 0.5 cm from the top of the skull, and along the rear of the skull across the occipital bone. The tissue was washed with phosphate buffered saline (PBS) and immediately transferred to 10% formalin for histological analysis (n = 3 rats/group) or wrapped in gauze soaked with 1X D-PBS and frozen at −80°C for microcomputed tomography (microCT) and the push-out test (n = 8–10 rats/group).

2.6 |. microCT

New bone tissue volume (TV), bone mineral density (BMD), and architectural patterns of the mineralized tissue regenerated within the defect were evaluated by microCT using a SkyScan 1172 (Bruker, Billerica, MA). Samples were thawed for two hours at RT, then assembled inside a custom jig holding two samples at a time. Foam was placed between the samples and around the inner perimeter of the jig to eliminate movement while scanning. The scanning parameters were: 11.86 μm voxel size, 226 μA, 45 kV, and power of 10 W with a 0.5 mm aluminum filter to correct for beam hardening. The images were reconstructed with NRecon software (Bruker, Billerica, MA). An 8-mm diameter region of interest was centered over defect site and a 1.5-mm thickness cylinder created to include the full calvarial arch defect thickness, creating a cylindrical volume of interest with CTAn software (Bruker, Billerica, MA) [33]. The reconstructed images were then evaluated in CTAn with a program to threshold mineralized bone from other tissues (determined to be 60 on a 0–255 scale) based on the Otsu method as previously described [34]. The mineral density of the regenerated bone within the defect was normalized to the mineral density of the surrounding calvarial bone of the same animal to determine BMD as a percentage to represent quality of regenerated bone. Total mineralized TV was similarly normalized to the defined TV of the defect volume of interest-bone volume/TV (BV/TV) and represented the quantity of regenerated bone. Three-dimensional architectural analysis was performed on the mineralized tissue to further determine bone surface (BS) to BV ratio (BS/BV), BS to TV (BS/TV), trabecular thickness (TbTh), trabecular spacing (TbSp), trabecular number (TbN), trabecular pattern factor (TbPf), and structural model index (SMI).

In order to assess spatial differences in bone regeneration patterns within the defect, the bone area regenerated in each 11.86-μm thick 8-mm diameter slice of the defect were calculated from the top (cranial) to the bottom (caudal side/dural side) of the 1.5-mm thick cylinder. This data are represented by the mean and standard error of bone area/tissue area % for all samples within a treatment group. Further, representative 3D images of mineralized tissue regenerated within the defect were created in pseudo color for each experiment group (using the sample with BV/TV closest to the group median) using Mimics software (v24.0, Materialise, Plymouth, MI).

2.7 |. Push-Out Test

Once the microCT was completed, a push-out test was conducted on each sample individually to determine the extent of osseointegration of new bone with the host bone on an ElectroPuls E1000 (Instron, Norwood, MA) with a 5 kN load cell [25, 35, 36]. Each sample was placed upside down with the defect site directly laying on a custom cylindrical jig centered over a 10-mm diameter hollow basin. Next, the 5.75-mm diameter probe was lowered until it was just above the defect. The force and displacement were balanced then the test was run, which caused the probe to push through the defect site at a rate of 1 mm/min. The failure load was recorded as the maximum force needed to push out the scaffold or empty defect.

2.8 |. Histology

Calvarial bone was excised to include the 8-mm diameter defects (with or without scaffold implants) and placed in 10% neutral buffered formalin. After fixation, tissue specimens were decalcified in Cal-Ex Decalcifier for 48–72 h. Coronal sections of each skull at the center of the 8 mm defect was routinely processed and embedded in paraffin. Tissue blocks were decalcified for 30 min after rough cut. Tissues sections were cut at 5 μm and stained with hematoxylin and eosin (H&E) for histopathologic evaluation. Histologic scoring was performed by a board-certified veterinary pathologist. Digital photographs were taken using an Olympus BX53 microscope and an Olympus DP74 camera.

2.9 |. Statistical Analysis

All data are represented as mean ± standard error of the mean in the text table, and as the mean with the Tukey range determining the box, and the standard error the whiskers on the box and whiskers plot, with outliers identified by color dots outside the inter-quartile range. A two-way analysis of variance (ANOVA) was used to compare the main effects (time and graft treatment type) for all quantitative data, following the Shapiro–Wilk test for normality and Brown-Forsythe test for equal variance of the experimental data set, followed by Tukey’s test for post hoc testing (multiple comparisons conducted after correction, α = 0.05). Significant differences were determined at p < 0.05. All statistical analyses were performed using SigmaStat (v14.5, Inpixion, Palo Alto, CA) and plotted using Prism (v9, GraphPad Software, Boston, MA). Significant differences found in statistical analyses are consistently indicated for both main and pairwise post hoc tests (* represents p < 0.05, ** represents p < 0.01, and *** represents p < 0.001).

3 |. Results

3.1 |. New Bone Formation by microCT

Representative microCT reconstructions for each experimental group at four and 12 weeks are shown in Figure 1A, and new bone formation was calculated as the percent of BV within a consistent TV of interest (BV/TV) from the reconstructed microCT scans (Figure 1B, Table 1). Both the untreated defect as well as the unseeded scaffold show some defect coverage at both time points, and the regenerated BV/TV increased significantly (p < 0.001) across all treatment groups from 4 weeks (5.63% ± 0.51%) to 12 weeks (8.46% ± 0.52%). Scaffolds seeded with differentiated BMSCs (10.38% ± 1.44%) regenerated significantly more bone after 4 weeks than those seeded with undifferentiated BMSCs and untreated defects (2.49% ± 1.37%, p < 0.01 and 2.80% ± 1.44%, p < 0.01, respectively). After 12 weeks, the differentiated UCSCs (18.86% ± 2.57%) and differentiated ADSCs (12.85% ± 1.75%) regenerated significantly greater bone than both the untreated defects (3.85% ± 0.90%, p < 0.001) and the unseeded scaffolds (5.13% ± 1.53%, p < 0.01). While this trend remained when scaffolds were seeded with differentiated BMSCs (9.24% ± 1.65%) after 12 weeks, the difference was not significant (p = 0.15). Further, it was observed that seeding with undifferentiated UCSCs (8.61% ± 1.22%) at 4 weeks resulted in significantly greater bone formation (p < 0.05) than seeding with undifferentiated ADSCs and BMSCs (1.76% ± 1.53% and 2.49% ± 1.37%, respectively), but no differences were observed between these treatments after 12 weeks. BMD was also normalized to the mineral density of the surrounding calvaria to account for any inter-specimen variability (Figure 1C, Table 1). A significant decrease (p < 0.05) in BMD was observed across all treatments from 4 weeks (98.40% ± 0.86%) to 12 weeks (95.95% ± 0.87%), which is often observed when there is an accompanying increase in newly regenerated bone over that time. After 4 weeks, the BMD observed in the unseeded scaffolds (112.58% ± 2.58%) was significantly higher than BMD for all other treatments (p < 0.05), but decreased significantly by 12 weeks (101.06% ± 2.31%, p < 0.001). After 4 weeks, the undifferentiated BMSCs (100.97% ± 2.31%) regenerated bone of significantly higher BMD quality than the undifferentiated UCSCs (89.56% ± 2.43%, p < 0.05), but the BMD observed in the undifferentiated BMSC seeded group also decreased significantly by 12 weeks (91.08% ± 2.58%, p < 0.01). No significant differences between BMD of regenerated bone was found between the groups after 12 weeks. This is indicative of the new bone being mineralized at levels quite close to the native calvaria and that this mineral accretion is quite physiological. No differences were noted between the various MSC sources in either BV/TV (bone quantity) or BMD (bone quality).

FIGURE 1 |.

(A) MicroCT images of the harvested calvaria. Representative 3D reconstructions show the bone regenerated in the defect volume for each group. (B) New bone formation. Supplementation of pre-differentiated stromal cells resulted in a significant increase in BV/TV compared to defect groups after 4 weeks with BMSCs, and after 12 weeks with both ADSCs and UCSCs compared to defect and scaffold only groups (**p < 0.01). A significant increase (***) was seen across all groups from 4 to 12 weeks. (C) Bone mineral density (BMD) after 4 and 12 weeks of healing. The BMD of the regenerated bone in the scaffold group without any cells was significantly higher after 4 weeks, compared to all seeded and unseeded scaffold treatments (**p < 0.01). Significant differences between treatments are indicated by (*p < 0.05; **p < 0.01, and ***p < 0.001).

TABLE 1 |.

Summary of the bone regeneration in the calvarial defects either empty or treated with scaffolds both without any cell seeding or seeded with pre-differentiated or undifferentiated stem cells of various origins (adipose derived, bone marrow derived, or umbilical cord derived).

		Defect	Scaffold	ADSC		BMSC		UCSC
				— Diff	+ Diff	— Diff	+ Diff	— Diff	+ Diff
BV/TV	4 weeks	2.80 ± 0.57	5.77 ± 1.24	1.76 ± 0.47	6.84 ± 1.08	2.49 ± 0.68	10.38 ± 1.52	8.61 ± 1.22	6.38 ± 1.03	%
	12 weeks	3.85 ± 0.90	5.13 ± 2.31	7.53 ± 1.16	12.85 ± 1.75	6.19 ± 2.37	9.24 ± 1.65	4.01 ± 0.89	18.86 ± 2.57
BMD	4 weeks	97.72 ± 1.53	112.58 ± 4.16	96.67 ± 3.81	98.37 ± 2.40	100.97 ± 3.73	97.08 ± 2.32	89.56 ± 1.95	94.27 ± 2.18	%
	12 weeks	101.06 ± 2.75	98.82 ± 2.23	97.00 ± 1.97	95.38 ± 1.37	91.08 ± 1.17	92.51 ± 1.87	93.58 ± 1.61	98.19 ± 1.64
BS/BV	4 weeks	21.67 ± 1.74	16.64 ± 2.34	30.30 ± 3.19	18.31 ± 2.27	25.67 ± 4.00	21.44 ± 2.30	25.74 ± 1.73	26.81 ± 1.31	mm−1
	12 weeks	20.29 ± 3.71	19.97 ± 3.98	18.10 ± 1.97	19.01 ± 0.95	21.34 ± 2.05	21.96 ± 1.40	25.60 ± 2.80	18.88 ± 1.62
BS/TV	4 weeks	0.53 ± 0.07	0.82 ± 0.12	0.45 ± 0.11	1.12 ± 0.14	0.37 ± 0.09	2.07 ± 0.28	2.06 ± 0.20	1.53 ± 0.26	mm−1
	12 weeks	0.56 ± 0.11	0.63 ± 0.18	1.35 ± 0.22	2.43 ± 0.33	1.16 ± 0.42	2.04 ± 0.37	0.91 ± 0.17	3.33 ± 0.36
Tb.Pf.	4 weeks	10.42 ± 1.40	9.30 ± 1.13	14.83 ± 2.14	7.36 ± 1.06	14.96 ± 1.67	8.86 ± 1.08	9.00 ± 1.28	11.88 ± 0.73	mm−1
	12 weeks	9.33 ± 1.91	13.54 ± 4.07	9.74 ± 2.08	6.00 ± 0.87	9.81 ± 1.86	7.43 ± 0.87	15.30 ± 2.58	8.45 ± 1.31
SMI	4 weeks	2.81 ± 0.20	3.55 ± 0.48	2.88 ± 0.15	2.41 ± 0.16	3.85 ± 0.25	2.53 ± 0.26	2.11 ± 0.24	2.79 ± 0.12
	12 weeks	2.74 ± 0.37	3.71 ± 0.30	3.14 ± 0.39	1.92 ± 0.29	2.68 ± 0.38	2.05 ± 0.23	3.51 ± 0.26	2.69 ± 0.29
Tb.Th.	4 weeks	208.61 ± 17.55	292.93 ± 34.06	155.81 ± 18.14	256.78 ± 29.72	224.88 ± 25.16	236.74 ± 27.57	176.83 ± 18.68	189.73 ± 17.43	μm
	12 weeks	235.85 ± 23.53	264.53 ± 36.57	266.40 ± 18.67	216.10 ± 13.29	212.47 ± 23.33	196.33 ± 14.03	194.24 ± 16.80	262.82 ± 34.15
Tb.N.	4 weeks	0.13 ± 0.02	0.19 ± 0.03	0.11 ± 0.03	0.27 ± 0.03	0.10 ± 0.02	0.45 ± 0.07	0.49 ± 0.05	0.34 ± 0.06	mm−1
	12 weeks	0.13 ± 0.02	0.19 ± 0.03	0.11 ± 0.03	0.27 ± 0.03	0.10 ± 0.02	0.45 ± 0.07	0.49 ± 0.05	0.34 ± 0.06
Tb.Sp.	4 weeks	1180.64 ± 25.98	1135.46 ± 62.76	1281.20 ± 34.19	1049.86 ± 41.09	1220.68 ± 58.58	802.58 ± 50.71	797.64 ± 32.71	903.11 ± 54.50	μm
	12 weeks	1207.54 ± 50.92	1150.73 ± 96.12	959.94 ± 88.72	740.09 ± 52.47	1033.29 ± 95.47	858.88 ± 66.81	990.08 ± 72.34	580.68 ± 40.10
Pushout Strength	4 weeks	12.60 ± 0.92	15.83 ± 1.07	12.35 ± 0.80	18.03 ± 2.22	13.27 ± 0.85	17.88 ± 1.94	15.43 ± 2.28	15.12 ± 0.94	N
	12 weeks	14.34 ± 2.21	29.12 ± 4.33	27.31 ± 2.70	32.86 ± 2.19	25.35 ± 2.06	20.96 ± 1.19	22.49 ± 1.18	32.82 ± 4.58

Note: Bone volume regenerated (quantity), bone mineral density (quality), bone morphometric parameters (architecture), and push-out strength (function) are all tabulated (average ± standard error of the mean) after 4 and 12 weeks of implantation.

The architecture and distribution of newly form bone tissue were assessed from the microCT analysis using the traditional histomorphometric parameters for trabecular bone (Figure 2A) [37]. The ratio of BS to BV (BS/BV) is typically indicative of bone turnover [38], and increases both with new dispersed mineralization as well as bone resorption. BS/BV would be expected to reduce over time if multiple fronts of new bone formation consolidate as time goes on and the active remodeling surface reduces relative to an increase in new BV. While there was a decrease observed in BS/BV from 4 weeks (23.0 ± 0.87 mm⁻¹) to 12 weeks (20.6 ± 0.87 mm⁻¹) showing increased consolidated deposition of mineralized tissue, this was not significant (p = 0.055) and there were no significant differences in BS/BV between individual treatments (p = 0.051). The SMI is a morphological parameter that ranges from spherical (4) to cylindrical (3) to more plate like (0) in terms of trabeculae of mineralization fronts in the case of regenerating bone [39, 40]. We observed that the SMI approached 4 for the unseeded scaffold group after both 4 weeks (3.55 ± 0.30) and 12 weeks (3.71 ± 0.30), very aligned with the spherical mineralization patterns within the pores of the scaffold that were unconnected to one another (Figure 2B). Undifferentiated BMSC seeded scaffolds had an SMI after 4 weeks (3.85 ± 0.27) that was significantly higher than those seeded with undifferentiated UCSCs (2.11 ± 0.28), or with pre-differentiated ADSCs or BMSCs (2.41 ± 0.28 and 2.54 ± 0.28, respectively, p < 0.05), again indicating distributed spherical new bone formation. The incidence of more plate like growth at the defect periphery reduced the SMI. Seeding with pre-differentiated ADSCs or BMSCs resulted in bone regeneration with SMIs significantly lower after 12 weeks (1.93 ± 0.27 and 2.05 ± 0.28, respectively) than bone regenerated within unseeded scaffolds (3.71 ± 0.30) or in scaffolds seeded with undifferentiated UCSCs (3.51 ± 0.28, p < 0.01) indicating more connected bone ingrowth.

FIGURE 2 |.

MicroCT based regenerated bone architectural parameters. (A) BS/TV showed no significant effect of treatment or time. (B) SMI was significantly higher for the scaffold only group, especially compared to differentiated ADSC- and BMSC-seeded groups. (C) Tb.Th. of the regenerated bone across groups was higher in scaffolds only after 4 weeks. (D, E) BS/TV and Tb.N. were significantly higher in the differentiated ADSC- and BMSC- and UCSC-seeded groups compared to the empty defect and scaffold only groups by 12 weeks. (F) Tb.Sp. decreased significantly by 12 weeks and was significantly lower in the differentiated cell-seeded groups compared to the empty defects and unseeded scaffolds. Significant differences between treatments are indicated by (*p < 0.05, **p < 0.01, and ***p < 0.001).

Trabecular organization and active bone apposition/resorption area within the defect volume was measured through the metrics of TbTh (Figure 2C, Table 1), BS/TV (Figure 2D, Table 1), TbN (Figure 2E, Table 1), and TbSp (Figure 2F, Table 1). TbTh is the average thickness of the continuous mineralization within the 3D region, while TbSp is the average space between mineralization fronts within the region. These parameters are calculated by sphere-filling of the solid (TbTh) and void (TbSp) spaces, respectively, and a lower TbTh and higher TbSp indicate poorer regeneration outcomes with potentially smaller and thinner mineralization patterns that are disconnected and distant. The only significant differences observed in terms of TbTh was that it was significantly higher (p < 0.05) in the unseeded scaffolds (292.93 ± 25.14 μm) after 4 weeks compared to both the undifferentiated and pre-differentiated UCSC seeded groups (176.83 ± 23.71 μm and 189.73 ± 22.49 μm, respectively) and the undifferentiated ADSC seeded group (155.81 ± 25.15 μm). There were no significant changes in TbTh observed from 4 weeks to 12 weeks (p = 0.27) or between treatments at 12 weeks (p > 0.42). The trends in BS/TV, TbN and TbSp were much more profound, with significant increases in BS/TV and TbN and significant reduction in TbSp across all treatments from 4 weeks to 12 weeks (p < 0.001). TbN is representative of trabecular interconnectivity at nodes while BS/TV is a measure for total bone surface in the current analysis since TV is constant across samples and represents total active bone apposition/resorption area within the defect volume (Figure 2D, Table 1). TbSp is representative both of the length of individual trabeculae in a fully interconnected trabecular structure, but also of the distance between unconnected bone growth locations in a formative structure (as in this case) because it is calculated by spheres filling the void volume. TbSp thus decreases as bone growth becomes more interconnected and there is greater coverage. Consistently across BS/TV, TbN, and TbSp after 4 weeks, the empty defects, unseeded SMP scaffolds, and the SMP scaffolds seeded with undifferentiated BMSCs had significantly lower BS/TV (p < 0.01) and TbN (p < 0.05) and greater TbSp (p < 0.01) than the scaffolds seeded with differentiated BMSCs. By 12 weeks, the differentiated BMSCs maintained this significance difference across BS/TV (p < 0.001), TbN (p < 0.01), and TbSp (p < 0.05) with the empty defect and scaffold only groups, but it was lost with the undifferentiated BMSCs (p ≥ 0.166). By 12 weeks, the differentiated ADSCs demonstrated significantly greater BS/TV (p < 0.05) and TbN (p < 0.01) than their undifferentiated counterpart, empty defects, and scaffold only groups. For TbSp after 12 weeks, the differentiated ADSCs had lower spacing than undifferentiated ADSCs (p = 0.158), but only significantly so than the empty defect and scaffold only groups (p < 0.001). This effect was more profound in the differentiated BMSC at 4 weeks (p < 0.001) and the differentiated UCSC seeded group after 12 weeks (p < 0.001) compared to their undifferentiated counterparts. For the scaffolds seeded with UCSCs, a significantly higher active BS/TV (p < 0.01) and higher connectivity as measured by significantly increased TbN (p < 0.05) and significantly decreased TbSP (p < 0.01) was observed for undifferentiated UCSC seeded treatment after 4 weeks compared to empty defects and scaffold only groups, but by 12 weeks, this effect was lost (p ≥ 0.139). While differentiated UCSCs did not exhibit drastic effects by 4 weeks, they did demonstrate significantly higher BS/TV (p < 0.001), TbN (p < 0.001), and significantly lower TbSp (p < 0.001) than their undifferentiated counterparts, empty defect, and scaffold only groups after 12 weeks. Across these trends, the scaffolds seeded with differentiated ADSCs and differentiated UCSCs showed the most new bone coverage and connectivity after 12 weeks of evaluation (Figure 2D–F). TbSp in the undifferentiated MSC seeded groups after both 4 weeks (1097.60 ± 48.95 μm) and 12 weeks (994.27 ± 47.43 μm) was much higher compared with the TbSp in the differentiated MSC seeded groups (917.97 ± 33.58 μm and 727.03 ± 37.02 μm, respectively). While TbSp in all groups reduced over time, no changes were observed in the untreated defect and unseeded SMP scaffold groups, which had significantly higher TbSp at both time points evaluated (1171.16 ± 29.67 μm) indicating that bone mineralization fronts if any in the defect site remained far apart.

Quantitative difference in the depth profiles of bone regeneration within the defect thickness was determined between all the treatment groups after 4 and 12 weeks (Figure 3). It was clearly evident that the overall BV regenerated increases significantly across all groups from four to 12 weeks and that for most treatments, the location of this regenerated bone within the defects also changes during this time. The scaffolds seeded with differentiated MSCs of all types performed the best after both 4 and 12 weeks, though the relative differences between the differentiated MSC type change with time. Most of the treatment groups, including the defect and empty scaffold showed a bias in bone regeneration after 4 weeks to the lower third in depth of the defect site toward the dural membranes. After 12 weeks though, the differentiated cell seeded groups especially demonstrated more uniformly distributed bone with high bone area in the center (by depth) of the defect suggesting patterns expected with consistent peripheral bone in-growth from the calvarial diploe at the margins of the circular defect sites and distributed bone regeneration in the scaffold interior.

FIGURE 3 |.

Spatial patterns in bone regeneration after four and 12 weeks. The cross-sectional regenerated bone area (mean ± SEM) was plotted every 11.86 μm across the 1.5 mm space created from the inferior dural side (0 mm) to the superior (1.5 mm) side to assess depth profiles of bone regeneration and remodeling trends for the eight treatment groups after four and 12 weeks in vivo implantation. Untreated defects showed a superior bias in terms of regenerated tissue, differentiated ADSC- and BMSC-seeding showed an inferior regeneration bias after 12 weeks, while groups with high regeneration such as the differentiated BMSC-seeded (at 4 weeks) and the differentiated UCSC-seeded scaffolds (at 12 weeks) did not show any bias.

3.2 |. Osseointegration by the Push-Out Test

Push-out tests were utilized to assess osseointegration (Figure 4). There was limited osseointegration after 4 weeks as noted by a very low push-out strength across all treatment groups (15.06 ± 0.78 N), with no significant differences observed between groups (p > 0.62). However, a significant increase (p < 0.001) in push-out strength at 12 weeks (25.65 ± 0.79 N) was observed across all groups. After 12 weeks, the groups treated with pre-differentiated or undifferentiated ADSCs (32.86 ± 2.10 N and 27.31 ± 2.35 N, respectively), pre-differentiated UCSCs (32.82 ± 2.21 N), undifferentiated BMSCs (25.35 ± 2.35 N), as well as empty scaffolds (29.12 ± 2.35 N) showed significantly greater push-out strength than the defect groups (14.34 ± 2.10 N) which had no scaffold placed (p < 0.001). Additionally, the pre-differentiated ADSCs and UCSCs seeded scaffolds also demonstrated significantly greater osseointegration after 12 weeks than the pre-differentiated BMSC seeded scaffolds (20.96 ± 2.21 N, p < 0.01). Only the empty defect group and the scaffolds seeded with differentiated BMSCs did not have a significant increase in push-out strength from four to 12 weeks. The untreated defects had the lowest change between time points and were the lowest in overall push-out strength based osseointegration after 12 weeks of healing, as expected in a negative treatment control. Furthermore, sub-group analysis comparing the maximum force within each cell type with undifferentiated versus pre-differentiated MSCs indicated that there were no significant differences for the ADSCs and BMSCs but differentiated UCSC seeding significantly improved osseointegration after 12 weeks compared with undifferentiated UCSCs (p < 0.05).

FIGURE 4 |.

Osseointegration as indicated by the push-out Test. The x-axis shows the treatment type and the y-axis shows the maximal Force (kN) after four and 12 weeks of healing. Significant increases were seen across treatments from 4 weeks to 12 weeks (***p < 0.001). Scaffold treatment, and supplementation with differentiated ADSCs and UCSCs, as well as undifferentiated ADSCs and BMSCs were significantly more integrated compared to the untreated defect. Significant differences between treatments are indicated by (*p < 0.05, **p < 0.01, and ***p < 0.001).

3.3 |. Histology

Angiogenesis appeared more robust at 4 weeks with many new vessels observed in all of the experimental groups, except for BMSC undifferentiated (Figure 5). At 12 weeks, few new vessels were observed in all of the experimental groups. In general, the presence of angiogenesis early in the wound healing process is appropriate. Normal wound healing kinetics stimulates angiogenesis early in the wound healing process and, as the wound matures, the stimulation for angiogenesis subsides. New immature woven bone was observed for most groups analyzed. The empty defect contained no osteoblastic cellular ingrowth at either assessed timepoint, indicating no osseointegration. The empty defect at 4 and 12 weeks contained only fibrous connective tissue with variable vascular ingrowth and proliferation (i.e., angiogenesis). In contrast, cellular ingrowth in the defect with SMP scaffold treated with pre-differentiated UCSCs is prominent in both the 4 and 12-week groups, evidenced by basophilic (purple)-staining osteoblastic, and osteocytic cell nuclei within and surrounding the eosinophilic (pink) new woven bone. Such tissue ingrowth is consistent with the increased force required in the push-out test for the scaffold containing groups (Figure S1). However, a significant portion of the scaffold remained in the defect site in all groups even at 12 weeks. There was no difference seen in fibrovascular tissue.

FIGURE 5 |.

Histological H&E staining of calvarial defects. (A–P) Representative histological sections of the calvaria stained with hematoxylin and eosin demonstrated new woven bone visible as eosinophilic fragments within the pores of the scaffold and at the margins (arrows; not all new woven bone fragments are annotated); as well as extant native bone at edge of each scaffold. (C–P) Scaffolds themselves are framed by bars in images. All images were acquired at a 20× magnification (scale bar = 1 mm).

4 |. Discussion

Studies have shown that the addition of MSCs to scaffolds can enhance bone formation [41, 42], but the optimal MSC tissue source and differentiation state for the SMP scaffold had not yet been determined. Previous reports comparing the bone regenerative ability of undifferentiated MSCs on various scaffold biomaterials have revealed varying outcomes. For example, studies showed similar bone healing results with multiple tissue sources of MSCs, either BMSC and UCSC, on a calcium phosphate cement scaffold or ADSC, BMSC, and umbilical cord blood-derived MSCs on a β-tricalcium phosphate scaffold [43, 44]. Conversely, BMSCs caused greater bone formation in a sheep critical size tibia defect model than ADSCs when implanted onto a mineralized collagen sponge [44, 45]. The current study aimed to determine the optimal tissue source and differentiation state of MSCs on the SMP scaffold to promote bone healing in a critical size defect. Beneficial properties are afforded by the scaffold design, including self-fitting (for osseointegration), pore features (for osteoconductivity), and bioactive polydopamine coating (for mineralization and osteoinductivity). Viability and osteogenic differentiation were previously confirmed on the SMP scaffold in vitro [23]. For the first time, this study assessed MSC-seeded SMP scaffolds in vivo, and revealed a clear impact on bone defect healing with pre-differentiated UCSCs providing the most robust response. While undifferentiated UCSCs exhibited more bone formation than any other undifferentiated MSC group at 4 weeks, there were no significant differences in new bone formation at 12 weeks.

During the normal course of bone healing, MSCs migrate to the injury site and then differentiate into osteoblasts to begin laying down new bone [46]. Thus, it stands to reason that pre-differentiating MSCs prior to implantation could further enhance the rate of bone healing, as observed by Xu et al. [47] Indeed, our study revealed that pre-differentiating MSCs for 14 days prior to implantation increased new bone formation on SMP scaffolds. By 12 weeks, pre-differentiated ADSCs and UCSCs had significantly more new bone formation than the untreated defect and scaffold only groups. UCSCs were the only group in which pre-differentiation also significantly increased new BV at this time point compared to undifferentiated UCSCs. In terms of bone organization, the same trends were observed across trabecular organization and connectivity patters where pre-differentiated MSCs of all types showed improved connectivity and increased thickness with reduced spacing between mineralization locations compared to undifferentiated MSCs and the control groups. This was especially stark for the unseeded SMP scaffold group where there was bone formation, but it was limited to individual pores as spherical unconnected islands within the defect space (based on SMI). In contrast, scaffolds seeded with differentiated MSCs produced regenerated bone that was more uniformly and centrally distributed across the scaffold depth after 12 weeks (based on microCT). The SMP scaffold encouraged osseointegration, as demonstrated by the significant increase in force required in the push-out test as compared to the untreated defect at 12 weeks. Osseointegration results were similar at 12 weeks for undifferentiated and pre-differentiated MSCs. All groups improved in graft push-out stability from four to 12 weeks with the highest levels seen in pre-differentiated ADSCs, UCSCs, and scaffold only groups. Thus, as with the undifferentiated MSCs, pre-differentiated MSCs exhibited differences in bone regeneration with the SMP scaffold. This relationship appeared to be dependent on scaffold biomaterial and pre-differentiated MSC type. For instance, in in vivo studies that likewise utilized MSCs pre-differentiated for 14 days prior to implantation, Lin et al. reported no benefit of pre-differentiating synovial fluid MSCs on polyetherketoneketone (PEEK) [48] while Annamalai et al. reported that pre-differentiated ADSC led to more bone formation on chitosan-collagen microtissues [49].

It is interesting to note that in our in vitro experiments, ADSCs, BMSCs, and UCSCs all had good growth and viability on the SMP scaffold [23]. It was observed that while ADSCs and BMSCs demonstrated strong mineralization and alkaline phosphatase activity on the SMP scaffold after induction with osteogenic differentiation medium for up to 21 days, the UCSCs only had increased in alkaline phosphatase activity, but did not undergo any mineralization on the SMP scaffold. Yet, in the current study, the highest overall amount of new bone formation was observed with the pre-differentiated UCSCs at 12 weeks. Previous reports have suggested that implanted MSCs may not contribute directly to bone healing by becoming osteoblasts, but indirectly by secreting factors that enhance bone healing [50–53]. The lack of mineralization observed in vitro with UCSCs, compared to the efficacy in generating new bone by pre-differentiated UCSCs observed in vivo, does provide potential support to this hypothesis. Differences in the environment must be recognized as it is possible that there were factors involved in vivo, which were not present in vitro, that pushed the UCSCs down the osteogenic path to become permanent osteoblasts. Future studies focusing on the final fate of implanted pre-differentiated MSCs after 12 weeks of bone healing in vivo could provide additional evidence needed to answer these questions. Overall, the results of this study demonstrated that all types of pre-differentiated MSCs exhibited increased new bone formation compared to the empty defect by 12 weeks, while only the pre-differentiated UCSCs demonstrated a significant increase in new bone formation compared to its undifferentiated counterpart, as well as the empty defect and scaffold only treatments. This increase in new BV with the pre-differentiated UCSCs compared to undifferentiated UCSCs was accompanied with a significant increase in BS/TV and TbN and a significant decrease in TbSp, but no significant changes in TbTh or other morphometric parameters after 12 weeks, indicating that the bone formed was uniformly bridging across pores and more connected across the volume after 12 weeks, but not necessarily a singular in-growth of a bone front from the periphery. The significant reduction in BV fraction from 4 to 12 weeks (p < 0.05) in the undifferentiated UCSC seeded group was also similarly accompanied by a decrease in BS/TV and TbN, and an increase in TbSp over time, reflecting that without pre-differentiation, the initial new bone regenerated does not persist across bone remodeling cycles.

In addition to efficacy, the selected MSC type used for bone healing must also consider tissue availability and possible immune rejection. For example, ADSCs can be harvested in much greater quantities than BMSCs [54–56]. ADSCs can also be extracted from the patient and used as autologous MSCs, thus mitigating immune rejection even in their differentiated state [57]. On the other hand, UCSCs have the greatest tissue availability amongst the three MSC tissue types evaluated, but are generally used as an allogenic source [58]. Herein, we determined that using pre-differentiated UCSCs on the SMP scaffold induced the greatest bone formation in vivo. However, studies have shown that while undifferentiated MSCs have low immunogenicity and can be immune-privileged, these traits may be lost or reduced once they begin to differentiate [59–63]. Therefore, the use of allogenic pre-differentiated UCSCs may have some limitations with respect to immune rejection. Autologous UCSCs are possible as private fetal tissue cryopreservation banking is gaining popularity and may provide a solution to immune rejection complications by cryopreserving one’s own umbilical cord tissue after birth [64, 65]. However, the patient’s parents must have banked the patient’s UCSCs at birth to ensure availability. Thus, it may be prudent to increase education and outreach to potential new parents to increase awareness on the benefits of cryopreserving their children’s UCSCs.

Finally, the rate at which scaffold degradation occurs is important to bone healing. Ideally, the degradation rate of a scaffold should match that of neo-tissue formation so as to not impede the healing process [66, 67]. PCL is known to slowly degrade (> ~2 years) owing to its hydrophobicity and crystallinity [68]. As such, histology at 12 weeks confirmed a lack of appreciable degradation of the PCL SMP scaffold. Bone regeneration was thus limited to the available individual pores as observed in microCT morphometrics. Considering that the pore size of the SMP scaffolds was ~220 μm which is close to the threshold for bone infiltration reported in vivo [69], the compliance of the scaffold seems to have supported robust new bone growth within pores (on average 278 ± 18 μm in this study), but limited bridging and growth across pores (as indicated by almost spherical bone morphology within unseeded scaffolds). A lack of degradation at 16 weeks was also observed in a recent rabbit study using an uncoated, unseeded PCL SMP scaffold [70]. Thus, the absence of scaffold degradation is thought to have played a significant role in inhibiting even greater bone formation that might have otherwise formed due to the incorporation of MSCs. Thus, future studies will focus on leveraging emergent PCL-based SMP scaffolds that exhibit faster rates of degradation, including PCL-poly(L-lactic acid) semi-interpenetrating polymer networks (PCL-PLLA semi-IPNs) and those based on PCL macromers having a star architecture [71, 72].

5 |. Conclusions

The in vivo regenerative capacity of undifferentiated and pre-differentiated BMSCs, ADSCs, and UCSCs seeded onto the SMP scaffold was evaluated in a rodent critical size defect model. Bone regeneration was assessed by microCT, pushout tests, and histology, all at 4 and 12 weeks. Compared to the untreated defect, the SMP scaffold alone (i.e., unseeded) was osteo-conductive and supportive of bone regeneration and osseointegration. Given this, as well as its capacity to conformally fit in irregular bone defects, these SMP scaffolds hold promise. Importantly, the regenerative potency of such scaffolds may be enhanced by seeding with MSCs. Depending on MSC type and differentiation status, seeding the scaffold produced marked effects in bone regeneration. Pre-differentiating MSCs on this SMP scaffold was found to significantly enhance bone healing at 12 weeks compared to the controls. At 12 weeks, pre-differentiated UCSCs exhibited increased new BV versus undifferentiated UCSCs; however, this trend was not observed for ADSCs or BMSCs. Overall, at 12 weeks, pre-differentiated UCSCs produced the greatest extent of bone regeneration versus all other groups. In terms of clinical translation, other practical factors must also be considered. The use of autologous UCSC is limited by the requirement for prior planning with respect to tissue banking. Alternatively, autologous ADSCs can currently be obtained in larger quantities from patients, including at higher levels than BMSCs. Thus, pre-differentiated ADSCs may be the optimal autologous MSC tissue type if pre-differentiated UCSCs are unavailable. Based on the bone regeneration potential of this system, future work will also investigate refining the SMP scaffold formulation to ensure the degradation rate of the SMP scaffold aligns more closely with the rate of new bone formation, allowing for improved regeneration outcomes at the defect site.

Supplementary Material

Additional supporting information can be found online in the Supporting Information section.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.