BMP-9-Stimulated Adipocyte-Derived Mesenchymal Progenitors Entrapped in a Thermoresponsive Nanocomposite Scaffold Facilitate Cranial Defect Repair

Abstract

Due to availability and ease of harvest, adipose tissue is a favorable source of progenitor cells in regenerative medicine, but has yet to be optimized for osteogenic differentiation. The purpose of this study was to test cranial bone healing in a surgical defect model utilizing BMP-9 transduced immortalized murine adipocyte (iMAD) progenitor cells in a citrate-based, phase-changing, PPCN-gelatin scaffold. Mesenchymal progenitor iMAD cells were transduced with adenovirus expressing either BMP-9 or GFP control. Twelve mice underwent craniectomy to achieve a critical-sized cranial defect. iMAD cells were mixed with the PPCN-gelatin scaffold and injected into the defects. MicroCT imaging was performed in 2-week intervals for 12 weeks to track defect healing. Histologic analysis was performed on skull sections harvested after the final imaging at 12 weeks to assess quality and maturity of newly formed bone. Both the BMP-9 group and control group had similar initial defect sizes (p = 0.21). At each time point, the BMP-9 group demonstrated smaller defect size, higher percentage defect healed, and larger percentage defect change over time. At the end of the 12-week period, the BMP-9 group demonstrated mean defect closure of 27.39%, while the control group showed only a 9.89% defect closure (p<0.05). BMP-9-transduced iMADs combined with a PPCN-gelatin scaffold promote in vivo osteogenesis and exhibited significantly greater osteogenesis compared to control. Adipose-derived iMADs are a promising source of mesenchymal stem cells for further studies in regenerative medicine, specifically bone engineering with the aim of potential craniofacial applications.

Introduction

Critical-sized craniofacial skeletal defects preclude closure through spontaneous healing and require surgical intervention to achieve full closure.^1,2 Autologous bone grafting has long been the standard of care for these lesions, but these procedures are often accompanied by significant perioperative morbidities including pain, infection, and bleeding.^3–7 Other limitations include the potential for graft resorption and the inherently limited supply of bone available for grafting.^6–8

Polymer-based implants, bone ceramics, and porous polyethylene implants have been used as alternatives to autologous bone grafting with varying success. These materials have been associated with increased rates of failure and infection compared to autologous bone grafts.^9–13 Permanent fixation with metals or metal alloys have demonstrated similar drawbacks with the added disadvantage of limited growth adaptation potential.¹⁴ The issues of adaptation and integration render permanent rigid fixation a highly unfavorable treatment option for pediatric patients.

Regenerative medicine strategies may hold a solution to current treatment limitations. One such investigational strategy involves utilizing mesenchymal stem cells (MSCs), novel biomaterials, and tissue engineering technologies to generate new bone for the closure of critical-sized skeletal defects.^15–17

Osteoblastic differentiation can be induced by a variety of osteoinductive growth factors both in vivo and in vitro.¹⁸ While many growth factors and signaling pathways play important roles in regulating osteogenic differentiation, bone morphogenetic proteins (BMPs) are among the most potent osteoinductive factors.^19–25 BMPs are expressed by nearly all mammalian cells throughout life. BMP-9 works to regulate downstream targets within the nucleus that act as co-activators and co-repressors for osteoblastic activity and has proven to be the most effective BMP for promoting osteogenic differentiation.^25,26

For optimal osteoblastic stem cell function, an engineered osteoconductive scaffold must be used.^27–29 The ideal scaffold is biocompatible and biodegradable, supporting growth, differentiation, vascularization, and cell transport as new tissue forms. One novel scaffold that has shown promise in early studies is poly(polyethylene glycol citrate-co-N-isopropylacrylamide) (PPCN).^26,30 This thermoresponsive, biodegradable polymer supports protein loading, three-dimensional cell proliferation, and cell viability.^31–33 As a thermoresponsive material, it undergoes a liquid-to-solid phase change when heated from 4^o to 37°C.^34–39 The liquid state at cool temperatures is convenient for mixing, injecting, and storage. As the liquid approaches body temperature, it gradually solidifies, conforming to the site of implantation. The addition of gelatin to PPCN (PPCNg) was shown to enhance cell adhesion properties for cells entrapped in the scaffold and injected in vivo while promoting BMP-9-induced osteogenesis of MSCs.²⁶

Adipose tissue is emerging as a promising source of MSCs. Adipose tissue harbors progenitor cells termed adipose-derived mesenchymal stem cells (AD-MSCs).^40–43 AD-MSCs have shown early osteogenic potential, but studies that evaluate their ability to form bone in vivo are lacking. In this study, we investigate cranial bone healing in a surgical defect model utilizing BMP-9-transduced immortalized murine adipocyte (iMAD) progenitor cells embedded within a PPCNg scaffold. We hypothesize that BMP-9-transduced iMADs will exhibit significantly greater osteogenesis in vivo when compared to green fluorescent protein (GFP)-transduced iMAD controls in a mouse model.

Materials and Methods

Cell Culture and Chemicals

293pTP cells were obtained by modifying HEK-293 cells for efficient adenovirus packaging and amplification.⁴⁴ The iMAD cells are previously characterized mouse adipocyte-derived progenitor cells.⁴⁵ Both 293pTP and iMAD cells were maintained in Dulbecco’s Modified Eagle Medium (DMEM), as described.^44,46–49 Unless indicated otherwise, all chemicals were purchased from Sigma-Aldrich (St. Louis, MO) or Thermo Fischer Scientific (Waltham, MA).

Synthesis of PPCN Gel

PPCN (polyethylene glycol citrate-co-N-isopropylacrylamide) was synthesized as previously described.^26,33 For each newly synthesized batch, the chemical, biodegradable, and thermoresponsive features were tested as previously reported.³³ PPCN was subsequently mixed with a 0.2% gelatin/PBS (phosphate-buffered saline) at a 1:1 volume ratio to form PPCN-gelatin, referred to as PPCNg. Prior to surgery, iMADs were transduced with recombinant adenovirus expressing BMP-9 or GFP (Ad-BMP-9 or Ad-GFP) for 48h. The transduced cells were then collected and suspended in the PPCNg liquid mixture.

Generation and Amplification of Recombinant Adenovirus Expressing BMP-9 and GFP

Recombinant adenoviruses expressing human BMP-9 and green fluorescent protein (GFP) were created using the AdEasy system.^50–52 The coding regions of human BMP-9 and GFP were PCR amplified and cloned into adenoviral shuttle vectors and subsequently used to generate recombinant adenoviruses, Ad-BMP-9 and Ad-GFP, in 293pTP cells.⁴⁴ For all adenoviral transductions, polybrene (4–8 μg/ml) was added to enhance transduction efficiency as reported.⁵³

Establishment of Murine Critical-Sized Defect Model

All animal studies were conducted by following the NIH guidelines approved by Institutional Animal Care and Use Committee (ACUP #71745). Eight-week-old male, athymic, nude mice (n=12) were obtained from Harlan Laboratories (Indianapolis, IN). There was no significant difference between the mean weight of the control group (Ad-GFP+iMADs+PPCNg) weighing 27.33g and the experimental group (Ad-BMP-9+iMADs+PPCNg) weighing 26.00g (p=0.13). Mice were treated with 2.5% isoflurane delivered with 100% oxygen to induce anesthesia and maintained thereafter with 1.5–2.5% isoflurane. Additional pain control was provided with Meloxicam (1.0 mg/kg) injected subcutaneously prior to incision and every 24 hours for the first 48 hours postoperatively.

Surgical protocol was conducted by a trained member of the research team to minimize variability and error. A sagittal incision was made from the orbital ridge to 5 mm behind the ears. Skin flaps on either side were retracted to expose the calvaria. A Dremel (Racine, WI) MultiPro Cordless handheld drill fitted with a stainless-steel trephine drill bit was used to create a full-thickness 4 mm diameter calvarial defect on the left parietal bone of each mouse. Sterile phosphate buffered saline (PBS) with 1000 IU/ml penicillin and 1 mg/ml streptomycin (35 ml/kg) was used for irrigation during drilling. Care was taken to avoid damaging cranial sutures or the underlying dura mater. All resultant defects were grossly uniform, circular, and did not involve cranial sutures. Six mice were treated with PPCNg prepared with Ad-GFP-transduced iMADs to serve as controls (Figure 1A). Six mice were treated with PPCNg prepared with Ad-BMP-9-transduced iMADs (Figure 1B). P200 pipets were used to fill the cranial defects with 20μl of the respective scaffold/cell combination and each was allowed to solidify for 5 minutes after placement. Importantly, all aliquots were kept on ice until just prior to injection to prevent premature solidification of scaffold and maximize iMAD viability. Representative images of surgical defects before and after scaffold placement are shown in Figure 2B. The skin flaps were closed over the surgical defect with 5-O monofilament non-absorbable (Nylon) sutures. These were later removed between 10 and 14 days post-operatively.

Figure 1:

Experimental Design Schematic Representation. (A) Murine adipose cells (MADs) were immortalized via retrovirally introduced SV40 large T antigen to produce iMADs. iMADs then transduced with BMP-9 via adenoviral vector and mixed with PPCN gel scaffold. (B) Implant Components of Experimental and Control Groups.

Figure 2:

Cranial Defect Model and MicroCT Analysis. (A) Baseline Defect volume. (B) Creation of surgical defect prior to scaffold placement (left) and following placement of cell-laden scaffold (right). (C) Axial microCT image demonstrating surgical defect site prior to 3D reconstruction. Arrow points to defect site. (D) MicroCT images following 3D reconstruction. Image on left demonstrates cranial defect prior to analysis. Image on right demonstrates microCT reconstruction following fitting of measurement graphic for defect size calculation.

MicroCT Imaging and Data Analysis

Bone regeneration and healing were assessed serially with microCT imaging utilizing the GE Triumph (GE Healthcare) trimodality pre-clinical imaging system. Baseline imaging and defect volume calculations (using volumetric reconstructions generated in the computer-based program Amira®) were performed at 48 hours (Week 0) postoperatively and served as the standard for all subsequent residual defect volume measurement comparisons (Figure 2C). Subsequent imaging was performed at 2, 4, 6, 8, 10, and 12 weeks postoperatively. At each time point, imaging data were collected for defect volume size (mm³), percentage of defect healed, and percentage change in defect size over time for the Ad-BMP9 and Ad-GFP groups (Figure 2D). During scanning, the mice were sedated using the same isoflurane delivery procedure previously described and were monitored via an SA Instrument MRI monitoring system that tracked heart rate, respiratory rate, and temperature. The images were analyzed with Amira® 5.3 software (FEI, Hillsboro OR). Scale bars were used to standardize the images.

Histologic Evaluation

Following the final 12-week microCT imaging, the mice were euthanized with CO₂ inhalation followed by cervical dislocation. The skull samples containing defect sites were retrieved, fixed, decalcified in 10% buffered formalin, and embedded in paraffin. Serial sections of the specimens were stained with hematoxylin and eosin (H&E) and trichrome staining as previously described.⁵⁴

Statistical Analysis

Statistical analysis was carried out with the Microsoft Excel program. Data were expressed as mean ± standard deviation. Bone regeneration volume was scored using a one-way analysis of variance and student’s t-test. A value of p<0.05 was considered statistically significant. Two-tailed-t-tests were used to assess defect size relative to baseline (Week 0) at differing time intervals and to compare defect filling between the experimental and control groups.

Results

BMP-9-Transduced iMADs in PPCNg Increased Osteogenesis vs. Control

All mice survived the duration of the study until planned euthanasia following the final imaging. Additional sutures were placed for one mouse following breakdown of post-operative stitches at 24 hours. This mouse was also given topical antibiotics for a scab formed from scratching his medication injection site. There were no cases of infection or implant extrusion. There were no cases of scaffold loss or displacement.

There was no significant difference in mean initial defect size (Figure 2A) between the experimental and control groups (p=0.21). At each time point, the BMP-9 group demonstrated smaller defect size, higher percentage healed, and higher percentage changed with respect to time (Table 1, Figure 3A and B). At the end of the 12-week period, the BMP-9 group demonstrated mean defect closure of 27.39%, compared to 9.89% defect closure in the GFP group (Figure 3C).

Table 1.

Defect Size Data Over Time

	Ad-GFP+iMADs+PPCNg	Ad-BMP9+iMADs+PPCNg	p value
Number of mice	6	6
Avg. mouse mass (g)	27.33 ± 1.63	26.00 ± 2.19	0.13
Defect size (mm3)
Week 0	3.53 ± 0.21	3.76 ± 0.37	0.21
Week 2	3.35 ± 0.19	3.23 ± 0.33	0.47
Week 4	3.27 ± 0.24	3.14 ± 0.31	0.43
Week 6	3.27 ± 0.22	3.04 ± 0.31	0.18
Week 8	3.25 ± 0.23	2.98 ± 0.28	0.10
Week 10	3.27 ± 0.24	2.87 ± 0.35	0.06
Week 12	3.18 ± 0.19	2.72 ± 0.28	<0.05
% Defect healed
Week 2	5.10 ± 1.55	13.93 ± 3.84	<0.05
Week 4	7.28 ± 3.06	16.36 ± 3.11	<0.05
Week 6	7.35 ± 3.58	18.87 ± 5.55	<0.05
Week 8	8.00 ± 3.69	20.46 ± 5.25	<0.05
Week 10	8.25 ± 3.25	23.69 ± 4.77	<0.05
Week 12	9.89 ± 3.09	27.39 ± 4.82	<0.05
% Change defect size
Week 2	5.10 ± 1.55	13.93 ± 3.84	<0.05
Week 4	2.19 ± 3.15	2.43 ± 2.22	0.88
Week 6	0.07 ± 2.08	2.51 ± 2.70	0.11
Week 8	0.65 ± 0.63	1.60 ± 2.93	0.46
Week 10	0.14 ± 1.67	3.23 ± 3.84	0.13
Week 12	2.49 ± 0.96	3.70 ± 2.64	0.36

Figure 3:

Results. (A) Defect Size (mm³) (B) Percent defect healed over time (%) (C) Representative schematic comparing percent defect healed at 12 weeks

BMP9-Transduced iMADs in PPCNg Increased Mature Bone Formation vs. Control

The data provided by microCT volumetric analysis do not allow interpretation of the quality of newly formed bone. As numerous bone engineering studies have previously shown, new “bone” on gross inspection or imaging analysis may prove to contain significant non-osseous material such as fat, cartilage, and other fibrous elements when viewed at the microscopic level. At times, imaging studies demonstrate features of poor quality bone (decreased volume, porous quality, etc.) but this is not always the case. Histologic analysis is necessary to evaluate the quality of newly formed bone. Histology from skull microsections harvested 12 weeks post-treatment demonstrate increased maturity and quality of bone in the Ad-BMP9-iMADs-PPCNg group compared to the Ad-GFP+iMADs+PPCNg group (Figure 4). All histology obtained shows more robust bone in the experimental group compared to control, consistent with the representative histology images shown in Figure 4. Histology of newly formed bone in the control group consistently demonstrated increased defects, discontinuity, and non-osseous elements.

Figure 4:

Histologic Analysis. (A) Trichrome (left) and H&E (right) stain of bone tissue harvested from Ad-BMP9-PPCNg group. (B) Trichrome (left) and H&E (right) stain of bone tissue harvested from Ad-GFP-PPCNg group. Both Trichrome and H&E stains indicate increased quantity and maturity of bone in the BMP-9 group. All sections show no trace of PPCN-gelatin scaffold, indicating complete resorption.

Discussion

We have demonstrated that BMP-9-transduced iMADs combined with a PPCN-g scaffold promote in vivo osteogenesis. Our work is most significant in the larger scope of utilizing adipose-derived progenitor cells for the purposes of bone engineering. We have confirmed that adipose-derived stem cells hold promise for bone tissue engineering by showing that iMADs can be induced to repair a surgical defect in mice. AD-MSCs have been investigated with the goal of regenerative medicine applications for several years, but only a small portion of these have focused on the aim of bone engineering.^42,55–59 The ability to achieve comparable results in bone engineering using AD-MSCs in lieu of other MSC lineages would be a major accomplishment. Adipose tissue is more available and attainable than many other commonly used sources of MSCs, such as bone marrow.

Despite robust historical efforts, the efficiency in engineering bone from AD-MSCs has yet to be optimized, and our study confirms this. The defect closure of roughly 27% at 12 weeks is significantly less than levels achieved in similar studies utilizing other MSC lineages.⁶⁰ The specific etiologies for the decreased osteogenic potential of AD-MSCs compared to more osteogenic MSC lineages has not been fully elucidated. Future studies comparing the detailed microenvironments of various MSC lineages and the specific interplay between transcriptional regulators, growth factors, and other proteins involved in the process of osteogenic differentiation may hold answers to these questions. Drawing on studies that comparatively assess the osteogenic potential of multiple MSC lineages, we can infer that MSC sources further along the differentiation pathway, towards osteocyte formation, more robustly favor osteogenesis.^61–63 With this consideration, it is not surprising that AD-MSCs are less potent than MSC lineages closer to the pre-osteoblast and osteoblast stages. Other potential etiologies for the decreased potency of AD-MSCs can be found in studies comparing them with bone marrow derived MSCs (BM-MSCs). Across several studies, BM-MSCs overwhelmingly demonstrated increased osteogenic potential compared to AD-MSCs.⁶³ Reasons included higher expression of osteogenic promoters in the BM-MSC groups and a degree of heterogeneity found within the AD-MSC groups that included several AD-MSC subpopulations with diminished osteogenic potential.^62,63

To optimally influence AD-MSC osteogenic differentiation, more precise manipulation of the pathways that drive them towards osteogenesis is needed. One method of doing this is to optimize the interplay of multiple signaling pathways to maximize their osteogenic potential. We have recently accomplished this by demonstrating that Notch signaling activation augments BMP-9-induced osteogenic differentiation of iMADs.⁶⁴ Another important goal will be to prevent alternative paths of differentiation. Since AD-MSCs are further removed from the targets of differentiation than other MSC lineages, this is especially pertinent. We recently demonstrated an example of this by showing that the addition of NEL-Like Molecule-1 (Nell1) increased osteogenic differentiation of MSCs by inhibition of adipogenic differentiation.⁶⁵ Further investigating the complex array of interactions that promote osteogenesis of AD-MSCs, by incorporating new developments such as these with Notch signaling and NELL1, is a promising next path for inquiry.

There are limitations of this study to note. The sample size of 12 mice divided into 2 groups is not particularly large. Additionally, the small physical size of mice presents unique challenges for a cranial defect model. Experience and skill are required to consistently fill the surgical defect. The mouse skull is extremely thin at baseline and can easily be overfilled once the defect is created. Larger, thicker skulls, a rabbit or pig, for example, would allow for closer approximation of the craniofacial anatomy of humans.

In conclusion, BMP-9-transduced iMADs combined with a PPCN-g scaffold promote in vivo osteogenesis. We found that bone volume, maturity, and quality were improved by utilizing this strategy for bone engineering. iMADs are a promising source of mesenchymal stem cells for further studies in regenerative medicine. BMP-9-transduced iMADs combined with a PPCN-gelatin scaffold is a novel and encouraging approach to current bone engineering challenges.