In Vivo Bone Formation by Human Bone Marrow Stromal Cells: Reconstruction of the Mouse Calvarium and Mandible

Authors

  • Mahesh H. Mankani M.D.,

    Corresponding author
    1. Division of Plastic Surgery, Department of Surgery, University of California-San Francisco, San Francisco, California, USA
    • University of California, San Francisco, Department of Surgery, 1001 Potrero Avenue, Box 0807, San Francisco, California 94143-0807, USA. Telephone: 415-206-8814; Fax: 415-206-3618
    Search for more papers by this author
  • Sergei A. Kuznetsov,

    1. Craniofacial and Skeletal Diseases Branch, National Institute of Dental and Craniofacial Research, National Institutes of Health, Bethesda, Maryland, USA
    Search for more papers by this author
  • Raymond M. Wolfe,

    1. Craniofacial and Skeletal Diseases Branch, National Institute of Dental and Craniofacial Research, National Institutes of Health, Bethesda, Maryland, USA
    Search for more papers by this author
  • Grayson W. Marshall,

    1. Preventive and Restorative Dental Sciences, University of California San Francisco, San Francisco, California, USA
    Search for more papers by this author
  • Pamela Gehron Robey

    1. Craniofacial and Skeletal Diseases Branch, National Institute of Dental and Craniofacial Research, National Institutes of Health, Bethesda, Maryland, USA
    Search for more papers by this author

Abstract

Bone marrow stromal cells (BMSCs) contain a subset of multipotent cells with the potential to repair hard-tissue defects. Mouse BMSCs, combined with a collagen carrier, can close critical-sized homologous mouse calvarial defects, but this new bone has a poor union with the adjacent calvarium. When human BMSCs are transplanted for the purpose of engineering new bone, best results can be achieved if the cells are combined with hydroxyapatite/tricalcium phosphate (HA/TCP) particles. Here, we demonstrate that transplantation of cultured human BMSCs in conjunction with HA/TCP particles can be used successfully to close mouse craniofacial bone defects and that removal of the periosteum from the calvarium significantly enhances union with the transplant. Transplants were followed for up to 96 weeks and were found to change in morphology but not bone content after 8 weeks; this constitutes the first description of human BMSCs placed long-term to heal bone defects. New bone formation continued to occur in the oldest transplants, confirmed by tetracycline labeling. Additionally, the elastic modulus of this engineered bone resembled that of the normal mouse calvarium, and our use of atomic force microscopy (AFM)-based nanoindentation offered us the first opportunity to compare these small transplants against equally minute mouse bones. Our results provide insights into the long-term behavior of newly engineered orthotopic bone from human cells and have powerful implications for therapeutic human BMSC transplantation.

Introduction

Successful repair of bone deficiencies in the craniofacial skeleton, whether arising from trauma, tumor resection, or congenital disorders, continues to be a major concern to reconstructive surgeons. The transfer of bone autograft remains the preferred reconstructive method but is inadequate for large defects [1, 2]. For larger defects, osteoconductive matrices are used in lieu of autograft, but they suffer from limited tissue incorporation, significant fracture and infection rates, and the risk of migration [3, [4]–5]. The calvarium and the mandible remain two sites of considerable clinical concern.

Bone loss in the calvarium can arise secondary to trauma, congenital craniofacial dysplasias, neoplasms, infection, and decompressive craniotomies. Patients are at significant risk of cerebral injury and of neurologic dysfunction while these defects remain untreated [6, [7]–8]. Mandibular problems occur both early and late in life. Mandibular hypoplasia is present in congenital conditions such as hemifacial microsomia, whereas mandibular atrophy is a common problem after loss of the dentition. Both problems require operative transfer of significant amounts of bone to reconstruct the loss, and appropriate healing is uncertain at best [9, 10].

As a consequence of the clinical need to develop a more effective and consistent reconstructive material, we earlier demonstrated the feasibility of closing critical-sized mouse calvarial defects with cultured homologous bone marrow stromal cells (BMSCs) [11]. BMSCs represent a population of nonhematopoietic marrow-derived cells, a subset of which has multipotent capability, that can be isolated in vitro as a consequence of their high adherence to tissue culture plastic and high proliferation potential [12, 13]. Populations of BMSCs which include osteoprogenitor cells have been expanded in tissue culture and transplanted into recipient animals to form corticocancellous bone [11, 14, [15], [16], [17], [18], [19]–20]. Bone formation by BMSCs is dependent upon their transplantation with an appropriate matrix. With human BMSCs, the best new bone can be achieved if the cells are combined with hydroxyapatite/tricalcium phosphate (HA/TCP) particles [14]. Up until now, transplantation of cultured human BMSCs in conjunction with HA/TCP particles has not been used successfully to close craniofacial bone defects. The major purposes of this study were threefold: (a) to demonstrate the feasibility of closing critical-sized calvarial defects by using human BMSCs in conjunction with an HA/TCP matrix, (b) to establish a technique for achieving bone union between the transplant and the adjacent calvarium, and (c) to demonstrate the feasibility of augmenting the mandible with engineered bone.

Materials and Methods

Cell Culture

Surgical specimens containing fragments of normal unaffected bone with bone marrow were obtained from two patients undergoing reconstructive surgery. The patients were girls aged 14 years undergoing iliac crest bone harvest for correction of scoliosis. Tissue procurement proceeded in accordance with institutional regulations governing the use of human subjects, including the use of informed consent. Multicolony-derived strains of BMSCs were generated from the bone marrow in a manner previously described [16]. Briefly, a single-cell suspension of bone marrow cells was cultured in medium consisting of α-minimal essential medium (Invitrogen, Carlsbad, CA, http://www.invitrogen.com), 2 mM l-glutamine, 100 U/ml penicillin, 100 μg/ml streptomycin sulfate (Invitrogen), 10−8 M dexamethasone (Decadron; Sigma-Aldrich, St. Louis, http://www.sigmaaldrich.com), 10−4 M l-ascorbic acid phosphate magnesium salt n-hydrate (Wako, Osaka, Japan, http://www.wako-chem.co.jp/english), and 20% fetal bovine serum of a preselected lot (Equitech-Bio, Inc., Kerrville, TX, http://www.equitech-bio.com). In our hands, the addition of dexamethasone and ascorbic acid 2-phosphate to the medium significantly accelerated the growth and increased the harvest of human BMSCs while not influencing their in vivo osteogenic potential following subsequent transplantation. Thus these medium ingredients were used exclusively for the growth of BMSCs. The cells were incubated at 37°C in an atmosphere of 100% humidity and 5% CO2. Cells were passaged at near confluence with Trypsin-EDTA (Invitrogen).

Upon reaching confluence at passages two through five, cells were released by trypsin-EDTA and pipetted into 1.8-ml polypropylene cryotubes (NUNC A/S, Roskilde, Denmark, http://www.nuncbrand.com), each previously loaded with a 40-mg aliquot of HA/TCP particles (Zimmer, Inc., Warsaw, IN, http://www.zimmer.com). By means of a sieve shaker (CSC Scientific, Fairfax, VA, http://www.cscscientific.com), only particles of size range 0.1–0.25 mm were isolated and used. These represented an optimum sieve size, determined previously [19]. Each tube received 1.5–2.0 million BMSCs (BMSC transplant) or no cells (sham transplant). The mixtures were incubated for 90 minutes at 37°C on a slowly rotating platform. They were then centrifuged at 200g for 60 seconds, and the supernatant was discarded. An approximately equal number of transplants were generated from cells from each of the two donors.

Transplant Preparation and Placement

Three-month-old immunocompromised Bg-Nu-Xid female mice (Harlan Sprague Dawley, Inc., Indianapolis, http://www.harlan.com) served as transplant recipients. All animals were cared for according to the policies and principles established by the Animal Welfare Act and the National Institutes of Health (NIH) Guide for the Care and Use of Laboratory Animals. Operations were performed in accordance to specifications of an approved institutional small animal protocol. Mice were anesthetized with a combination of i.p. ketamine (140 mg/kg of body weight) and i.p. xylazine (7 mg/kg of body weight). Midline skin incisions approximately 1 cm in length were made on the dorsal surface of the scalp and the ventral surface of the jaw. The mice were then placed into the following studies: (a) calvarial onlay with or without an intact periosteum, (b) filling of a calvarial defect with no transplant, with a sham transplant, or with a BMSC transplant, and (c) mandibular onlay with either a sham transplant or a BMSC transplant.

Among the first study group of mice, an effort was made to determine whether removal of the calvarial periosteum could increase bony union between BMSC transplants and the skull. The calvarial periosteum was either removed or left intact, and BMSC transplants were placed either on the periosteum over the calvarium (supraperiosteal) or on the bare calvarium (subperiosteal).

In the second study group, the periosteum was removed from the skull, and a 5-mm full-thickness cranial defect was prepared with a trephine (Fine Science Tools, Inc., Foster City, CA, http://www.finescience.com) attached to an electric motorized handpiece (Dremel, Racine, WI, http://www.dremel.com). A 2-mm margin of bone was cleared of periosteum at the margin of the defect. Care was taken to minimize trauma to the dura mater. The defects were left unfilled, filled with sham transplants, or filled with BMSC transplants.

The mice in the calvarial studies underwent concurrent participation in the mandibular onlay study. The skin at the midline of the lower jaw was incised, and the mandibular bodies were exposed along their inferior surfaces. The periosteum was elevated for a distance of 7 mm bilaterally; onto these bare surfaces, the exposed bones received one of two onlays, sham transplants or BMSC transplants. The soft tissues were closed in a single layer using surgical staples. All incisions were closed with stainless steel surgical staples. Thirty-four mice underwent a calvarial procedure, and 22 of these also received a mandibular procedure. Tetracycline (200 mg/kg i.p.) was given 5 days before sacrifice. The mice were sacrificed at time points ranging from 2 to 96 weeks postoperatively with inhaled CO2, and their calvaria, mandibles, and transplants were harvested.

Histologic Preparation

The tissues were fixed in 4% phosphate-buffered formalin freshly prepared from paraformaldehyde (Sigma-Aldrich). After an overnight fixation at 4°C, the tissues were suspended in phosphate-buffered saline (PBS) (Invitrogen). Most were completely demineralized in buffered 10% EDTA (Quality Biological, Inc., Gaithersburg, MD, http://www.qualitybiological.com) prior to embedding, whereas the remainder were left mineralized for undecalcified processing. The decalcified tissues were embedded in paraffin so that their largest cut surfaces were sectioned; the sections were deparaffinized, hydrated, and stained with H&E. Separate unstained slides, incorporating 5-μm-thick sections, were prepared for in situ hybridization, described below. The undecalcified tissues were dehydrated in ethanol and embedded in methylmethacrylate; 5-μm-thick sections were obtained and stained with Goldner's modified trichrome or left unstained.

Estimation of Bone Formation and Bone Union

The H&E-stained sections were examined histologically, and the extent of bone within each transplant was scored on a semiquantitative, logarithmic scale by three independent, blinded observers in a manner similar to that described previously [16, 19, 20]. Each observer was an investigator in our laboratory who had been trained to evaluate the histologic characteristics of the transplants. Transplants were scored on a scale of 0–4; a score of 0 corresponded to no bone formation, whereas a score of 4 was given to transplants with abundant bone formation occupying greater than one half of the section (Table 1; Fig. 1).

Table Table 1.. Semiquantitative scale for the estimation of bone formation
original image
Figure Figure 1..

HA/TCP and BMSC transplants that have formed varying amounts of bone. (A): Transplant exemplifying a bone score of 0. No bone formation present. Rather, particles are widely separated by connective tissue. (B): Transplant exemplifying a bone score of 1. Only one bone trabecula present. (C): Transplant exemplifying a bone score of 2. Weak bone formation, with only a few trabeculae present. New bone does not bridge adjacent particles. (D): Transplant exemplifying a bone score of 3. Bone formation is appreciable but less than one half of the transplant. (E): Transplant exemplifying a bone score of 4. Abundant bone formation. Bone bridges adjacent particles. Magnification: ×10. Stain: Hematoxylin and eosin; paraffin embedding after demineralization. Abbreviations: b, bone; BMSC, bone marrow stromal cell; f, fibrous connective tissue, h, hematopoietic tissue; HA/TCP, hydroxyapatite/tricalcium phosphate; p, particle.

Additionally, the degree of bony union between transplant and adjacent mouse bone was evaluated. Gross adherence was evaluated by attempting to move the transplant manually and observing the extent to which it could be displaced from its original location. A union was scored on a scale of 0–4; a score of 0 corresponded to the absence of any union between transplant and adjacent mouse bone, whereas a score of 4 was given to transplants that were grossly secure and where bony union was evident in greater than one half of the histologic sections (Table 2).

Table Table 2.. Semiquantitative scale for the estimation of transplant-bone union
original image

Bone formation of the calvarial sections was also quantitatively estimated using a histomorphometric image analysis system (Bioquant II; R&M Biometrics; Nashville, Tennessee, http://www.bioquant.com) and a Zeiss microscope at ×2.5 magnification. Within each section, measurements were made of the entire transplant, the portion of the transplant that had formed bone, the portion of the transplant that formed no bone, the length of calvarial defect that was closed with bone-forming transplant, and the length of calvarial defect that remained open.

Identification of Donor Cells

The origin of bone-forming cells within the transplants was confirmed through in situ hybridization. Unstained paraffin-embedded sections of tissues were obtained as described above. Control slides consisting of human and mouse tissues underwent in situ hybridization with the probe during each hybridization run, serving as positive and negative controls, respectively.

The human-specific repetitive alu sequence, which comprises approximately 5% of the total human genome, was applied for identification of human cells [21]. We used in situ hybridization for the alu sequence to study the origin of tissues formed in the transplants. The digoxigenin-labeled probe specific for the alu sequence had been prepared by polymerase chain reaction (PCR), including 1× PCR buffer (Perkin Elmer, Foster City, CA, http://www.perkinelmer.com), 0.1 mM dATP, 0.1 mM dCTP, 0.1 mM dGTP, 0.065 mM dTTP, 0.035 mM digoxigenin-11-dUTP (Roche Diagnostics, Indianapolis, http://www.roche.com), 10 pmol of specific primers, and 100 ng of human genomic DNA. The following primers were used on the basis of previously reported sequences [22]: sense, 5′-GTGGCTCACGCCTGTAATCC-3′, and antisense, 5′-TTTTTTGAGACGGAGTCTCGC-3′. The method for in situ hybridization of HA/TCP-containing transplants has been described previously [16]. Sections deparaffinized with xylene and ethanol were immersed in 0.2 N HCl at room temperature for 7 minutes and then incubated in 1 mg/ml pepsin in 0.01 N HCl at 37°C for 10 minutes. After washing in PBS, the sections were treated with 0.25% acetic acid containing 0.1 M triethanolamine (pH 8.0) for 10 minutes and prehybridized with 50% deionized formamide containing 4× standard saline citrate (SSC) at 37°C for 15 minutes. The sections were then hybridized with 1 ng/μl digoxigenin-labeled probe in hybridization buffer (1× Denhardt's solution, 5% dextran sulfate, 0.2 mg/ml salmon sperm DNA, 4× SSC, 50% deionized formamide) at 42°C for 3 hours after a denaturation step at 95°C for 3 minutes. After a washing with 2× SSC and 0.1× SSC, digoxigenin-labeled DNA was detected by immunohistochemistry using antidigoxigenin alkaline phosphatase-conjugated Fab fragments (Roche Diagnostics).

Mechanical Testing of Calvarial Transplants

A set of BMSC-HA/TCP transplants was embedded undecalcified in methylmethacrylate and cut into 5-μm-thick sections. These sections were stained with Goldner's modified trichrome for the collection of architecture data with the light microscope. The methylmethacrylate-embedded BMSC-HA/TCP blocks were further polished on one side with progressively finer grades of diamond paste (down to 0.1 μm) until a smooth bone surface was exposed (approximate nanometer roughness).

For topographic imaging and discrete mechanical properties determination of individual trabeculae, a modified atomic force microscope (AFM) (Nanoscope IIIa; Digital Instruments, Inc., Buffalo, NY, http://www.digitalinstruments.com) was used. The modification consisted of replacing the cantilever/tip assembly of the microscope with a transducer-driven head and tip (Triboscope Micromechanical Test Instrument; Hysitron, Inc., Minneapolis, http://www.hysitron.com) that allowed the microscope to operate both as an imaging and an indentation instrument as previously described [23]. A sharp diamond Berkovich indenter with a radius of curvature less than 100 nm was fitted to the transducer. The AFM piezo and respective control systems were used to image the surface of the sample to find a specific site of interest after which the load-displacement transducer was used to indent the sample while collecting the load displacement data. All indentations were performed with a trapezoidal load profile of 0.3 mm/s in time to a 150-μN maximum load. Elastic modulus and hardness were calculated from the unloading force/displacement slope at maximum load and the projected contact area at this load following the method of Doerner and Nix [24]. After indentation, the AFM piezo was used to scan the indented area. However, because of the texture of the sample surface, it was difficult to distinguish the indents.

The AFM measurements were performed on different trabeculae on each specimen; specimens from three mice were analyzed. The elastic modulus and hardness were obtained by indentation at four different sets of sites on each specimen; each site underwent nine indentations, with an interval of 5 μm between successive indentations.

Statistical Analyses

Statistical analyses between groups of transplants used a nonparametric unpaired t test, performed using InStat version 3.06 (GraphPad Software, San Diego, http://www.graphpad.com).

Results

Gross Morphology of Transplants

Calvarial defects that were left unfilled all remained visibly patent (Fig. 2A). Calvarial and mandibular transplants that included HA/TCP but no BMSCs were vascularized and were tightly adherent to the overlying soft tissues but only loosely adherent to the underlying or adjacent mouse bone (Fig. 2B). HA/TCP particles were poorly bound to each other and could be easily disassociated. In contrast, calvarial and mandibular transplants that included BMSCs were well vascularized (Fig. 2C, 2D). The transplants were tightly adherent to the underlying mouse bone, and individual particles were tightly bound to each other.

Figure Figure 2..

Gross images of mouse transplants. (A): Critical-sized calvarial defect, 20 weeks postoperatively. The defect remains patent. (B): Calvarial defect filled with sham (only hydroxyapatite/tricalcium phosphate) transplant, 86 weeks postoperatively. (C): Calvarial defect filled with bone marrow stromal cell (BMSC) transplant, 66 weeks postoperatively. (D): Mandible covered with BMSC transplant, 81 weeks postoperatively. The transplant is tightly adherent to the underlying bone.

Calvarial Onlay

Of six mice receiving transplants as calvarial onlays, all survived. All transplants exhibited bone scores of 4. The three subperiosteal transplants had excellent union with the calvarium (mean union score of 4), whereas the three supraperiosteal transplants had mean union scores of 0.67 (p < .05). (Fig. 3A, 3B). Bone union in the subperiosteal transplants was characterized by extensive bone bridging between the transplant and the underlying calvarium. Regenerating periosteum did not interfere with the union. In contrast, the supraperiosteal transplants had extensive bone formation within the transplants but no histologic evidence of bone between the transplants and the calvarium; instead, this zone had extensive fibrovascular tissue.

Figure Figure 3..

Images of mouse histologic sections. (A): Human BMSC and HA/TCP calvarial onlay, which is superficial to the periosteum (supraperiosteal), 96 weeks postoperatively, shows no bony attachment of the new bone to the underlying calvarium. Hematopoiesis is evident in the transplant as well as the normal calvarium. (B): Human BMSC and HA/TCP calvarial onlay, which is deep to the periosteum (subperiosteal), 92 weeks postoperatively, shows good bony attachment of the new bone to the underlying calvarium. Arrowheads indicate calvarium-transplant interface. (C): Mouse calvarial defect without BMSC or HA/TCP transplant, 35 weeks postoperatively. Defect remains grossly open. (D): Calvarial defect filled with only HA/TCP particles, 86 weeks postoperatively. Minimal bone formation at defect edges, but none in the center. (E): Calvarial defect filled with BMSC and HA/TCP transplant, 6 weeks postoperatively. Excellent bone formation throughout defect. (F): Mandibular onlay of BMSC and HA/TCP transplant, 44 weeks postoperatively. Excellent bone formation within the transplant and good union with the underlying mandible. (G): Calvarial defect filled with BMSC and HA/TCP transplant, 7 weeks postoperatively. Particles are interspersed with cortical bone, with minimal hematopoietic tissues. (H): Calvarial defect filled with BMSC and HA/TCP transplant, 76 weeks postoperatively. In contrast to the young transplants seen in (G), these mature transplants have increased hematopoiesis and resorption of the HA/TCP particles. (I): Continued bone formation in a BMSC and HA/TCP transplant, 96 weeks postoperatively, visualized with tetracycline labeling (green). (J): Confirmation of the donor origin of the newly formed bone in a 6-week-old BMSC-containing transplant. In situ hybridization with an ALU-specific probe targets BMSCs of human origin in this mouse calvarial defect filled with a human BMSC and HA/TCP transplant. (K): Strongly mineralized BMSC-derived bone in conjunction with HA/TCP particles, 96 weeks postoperatively. Magnifications: ×4 (A–E), ×10 (F–J), ×20 (K). Stains: Hematoxylin and eosin; paraffin embedding after demineralization (A–H), methyl methacrylate embedding (I), paraffin embedding after demineralization (J), Goldner's trichrome; methyl methacrylate embedding (K). (L): Confirmation of the persistence of donor-derived BMSCs in a 96-week-old BMSC-containing calvarial transplant. In situ hybridization with an ALU-specific probe targets BMSCs of human origin in this mouse calvarial defect filled with a human BMSC and HA/TCP transplant. Magnification: ×20; paraffin embedding following demineralization. (M): Clusters of donor-derived signal in the bone marrow of a 96-week-old BMSC-containing transplant, suggestive of proliferating hematopoiesis-supportive stromal cells generated by human BMSCs. In situ hybridization with an ALU-specific probe targets BMSCs of human origin in this mouse calvarial defect filled with a human BMSC and HA/TCP transplant. Magnification: ×10; paraffin embedding following demineralization. Abbreviations: b, bone within bone marrow stromal cell transplant; BMSC, bone marrow stromal cell; br, mouse brain; c, mouse calvarium; f, fibrous connective tissue; h, hematopoietic tissue; HA/TCP, hydroxyapatite/tricalcium phosphate; m, mouse mandible; p, particle.

Calvarial Defect and Mandibular Onlay

Twenty-eight mice lived at least 2 weeks after undergoing a calvarial defect procedure. Of these, four mice received calvarial defects without placement of a calvarial transplant, seven mice received a sham (BMSC-free) transplant, and 17 received BMSC transplants. Twenty-two of these mice also received a mandibular transplant: seven sham (BMSC-free) and 15 containing BMSCs.

Timing and Extent of Bone Formation

Mice were harvested from 2 to 96 weeks postoperatively. Four mice whose calvarial defects had been left unfilled were examined as late as 53 weeks (Fig. 3C). All of these defects remained patent, and all had bone scores and union scores of 0 (Fig. 4A, 4B). Seven mouse cranial defects were filled with HA/TCP particles alone. Minimal bone formed among these transplants, the latest of which was examined at 96 weeks after transplantation (Fig. 3D). Bone scores ranged from 0 to 2 in this group (Fig. 4C). Seventeen mouse defects were filled with BMSC-containing transplants. Thirteen of these 17 transplants formed significant bone (bone scores of 3 or 4) (Fig. 3E, 4E). The remaining four transplants had bone scores of 2. The number of transplants with poor bone formation relative to the number with good bone formation was nearly equal for the two BMSC donors. The extent of bone formation did not vary significantly between early-harvest and late-harvest transplants.

Figure Figure 4..

Bone formation and union among three different calvarial groups. Calvarial defects without a transplant (A, B), calvarial defects with a sham transplant (hydroxyapatite/tricalcium phosphate [HA/TCP] particles only [C, D]), and calvarial defects with a bone marrow stromal cell (BMSC) transplant (E, F). Please note that among certain groups, bone score is 0 whereas union score is 1. This reflects a fibrous union in all histologic sections. Evaluation of bone formation via histomorphometry (G and H), showing significant increases in the volume of bone-containing transplant and length of transplant closure with BMSC-containing transplants.

Significant bone formed as early as 6 weeks; it persisted without signs of degeneration, senescence, or sarcomatous transformation in transplants harvested up to 96 weeks. Bone scores among calvarial transplants with BMSCs ranged from 1 to 4 (mean 3.3), whereas bone scores among transplants without BMSCs ranged from 0 to 2 (mean 0.7). Differences between transplants with BMSCs and those without BMSCs were significant to the p < .001 level.

The degree and distribution of bony union among calvarial transplants mirrored the extent of bone formation. The four calvarial defects without a transplant failed to close as late as 35 weeks postoperatively and therefore had union scores of 0 (Fig. 4B). Of the seven cranial defects filled with particles without BMSCs, three had no evidence of union, and the remaining four had evidence of poor union (union score of 1 or 2) (Fig. 4D). In contrast, the 17 calvarial defects filled with BMSC-containing transplants demonstrated good bone union in 12 and poor bone union in five (Fig. 4F). Union scores among calvarial transplants with BMSCs ranged from 1 to 4 (mean 2.9), whereas union scores among transplants without BMSCs ranged from 0 to 1 (mean 0.6). Differences between transplants with BMSCs and those without BMSCs were significant to the p < .001 level. Note that among certain groups, bone score was 0, indicating no bone formation, whereas union score was 1, indicating a fibrous union in all histologic sections. Examination of the histologic sections of the calvarial transplants using histomorphometry demonstrated significant increases in bone volume and in the lengths of the bone-containing portions of the transplants when they included BMSCs (Fig. 4G, 4H).

Bone formation among mandibular transplants was not as consistently strong as among calvarial transplants. Seven mice that had received HA/TCP onlays without BMSCs on the mandible were examined as late as 96 weeks postoperatively. With the exception of one transplant that had formed a small amount of bone at 86 weeks, no bone formed in this group of transplants (Fig. 5A). In contrast, BMSC-containing transplants formed new bone as early as 6 weeks (Fig. 3F). Their bone scores ranged from 0 to 3, averaging 1.9 (Fig. 5C). Bone union scores among mandible transplants mirrored the bone scores (Fig. 5B, 5D).

Figure Figure 5..

Bone formation and union among two different mandibular groups. Mandibular onlays with a sham transplant (hydroxyapatite/tricalcium phosphate [HA/TCP] particles only [A, B]) and mandibular onlays with a bone marrow stromal cell (BMSC) transplant (C, D). Please note that among certain groups, bone score is 0, whereas union score is 1. This reflects a fibrous union in all histologic sections.

Transplant Histology

Histologically, calvarial and mandibular transplants without BMSCs contained aggregates of HA/TCP particles separated by fibrovascular tissue. Fibrovascular tissue also lay at the transplant-mouse bone interface (Fig. 3D). In contrast, BMSC-containing calvarial transplants contained abundant new lamellar bone firmly attached to both particles and the adjacent mouse bone (Fig. 3E). They also had associated hematopoietic tissue and occasional adipocytes. Mandibular transplants produced less overall new bone than did calvarial transplants. However, the bone was qualitatively similar to newly formed bone in the calvarium; it formed a lamellar structure oriented around individual particles (Fig. 3F). Good hematopoiesis was seen in the calvarial transplants, especially those with good bone formation, whereas only occasional hematopoiesis was seen in the mandibular transplants (Fig. 3E, 3F).

Bone morphology changed little between early and late transplants; by 8 weeks, the newly laid bone had assumed a morphology that was retained as late as the 96-week harvest date. Likewise, there was no evidence of bone degradation among late transplants. The most significant difference between early and late transplants was a visible reduction in the volume of residual HA/TCP particle among the late transplants and a corresponding increase in the amount of hematopoiesis, suggesting resorption of the particles and replacement with hematopoietic tissue (Fig. 3G, 3H).

Continued Bone Formation Among Transplants

BMSC transplants as late as 96 weeks after transplantation continued to demonstrate bone formation, as seen with tetracycline bone labeling (Fig. 3I). The bone that had formed was appropriately mineralized (Fig. 3K).

Determination of Bone Origin

The Alu gene sequence was used to follow the fate of the transplanted cells. Alu gene served as a marker for donor cell activity because it is not present in the recipient mouse cells. Tissue sections from a 6-, 81-, and 96-week-old transplants were evaluated using in situ hybridization with a probe raised against the Alu sequence. Presence of Alu was detected in osteoblasts and osteocytes within both the cortical and trabecular components of the new bone, confirming that the osteogenic cells were of donor origin rather than originating from the local microenvironment (Fig. 3J). Analysis of the late-harvest transplants demonstrated fewer ALU-positive cells than in the early-harvest transplants (Fig. 3L). With one exception, Alu immunoreactivity was restricted to the new bone and was not present in endothelial cells or in the peritransplant tissues, suggesting that the BMSCs did not migrate outside the confines of the transplant. The only exception to this observation were clusters of signal in the bone marrow of one late-harvest transplant, suggestive of proliferating hematopoiesis-supportive stromal cells generated by human BMSCs (Fig. 3M). No signs of inflammation or dysplasia were detected in the peritransplant region of any of the implants analyzed.

Mechanical Testing of Calvarial Transplants

The mechanical properties of three sets of specimens from the calvarial defect group were tested using AFM-based nanoindentation (Fig. 6). Areas of normal calvarial bone were easily distinguishable from human BMSC transplant by their location in the specimen. Care was taken to avoid sampling residual HA/TCP particles, whose elastic modulus was two- to threefold greater than the bone's. Elastic modulus values of the mouse bone and BMSC-associated bone were 24.1 (± 3.6) and 25.4 (± 4.7) GPa, respectively, whereas hardness values were 1.16 (± 0.26) and 1.13 (± 0.26) GPa, respectively. An unpaired t test with a 95% confidence level indicated no significant difference (p = .05) in elastic modulus and hardness between mouse calvarium and human BMSC-generated bone under dry conditions.

Figure Figure 6..

Mechanical data in BMSC transplants and calvarial bone. Comparison of elastic modulus (A) and hardness (B) from normal mouse calvarial bone and adjacent human bone marrow stromal cell (BMSC) transplant bone in three mice. No significant difference in modulus or hardness exists among the three sets of specimens.

Discussion

Calvarial reconstruction serves as a unique model for studying regenerative bone physiology. The calvarium is an anatomic area of limited mechanical stress and is quite unlike the axial skeleton, which is subject to long periods of compressive load. The relative stability of the surrounding structures, including intact calvarial margins, underlying dura mater, and overlying temporalis and frontalis muscles, creates a protected environment in which it is possible to study interactions between new bone constructs and in situ bone. In contrast, the mandible can offer a challenging microenvironment in which bone-bone interactions are complicated by movement of the jaw and its supportive muscles during mastication. Likewise, surgical reconstructions of the calvarium and the atrophic mandible are technically straightforward endeavors that can become complicated by large defect size or compromise of the overlying soft tissues by radiation or scarring.

In our previous work, we demonstrated that murine BMSCs were capable of closing critical-sized murine calvarial defects when placed in conjunction with a collagen carrier [11]. That study had limitations that this one was designed to address. Those transplants failed to form a union with the adjacent mouse skull; we attributed that failure to interference by the periosteum. The collagen carrier used in the previous study is technically easier to place into a bone defect than are the HA/TCP particles, but it is inappropriate for bone formation by human BMSCs, which require a mineral matrix, a BMP-expressing nonmineral matrix, or production of BMP-2 by genetically engineered human BMSCs for optimum bone formation [14, 25, [26], [27]–28]. Thus, this study was designed to test the practicality of placing HA/TCP particles into bone defects. Additionally, that study represented a short-term evaluation of the new bone, and we wanted to assess the behavior of the transplants long after placement in the animal.

The purposes of this study were (a) to demonstrate the feasibility of closing critical-sized calvarial defects by using human BMSCs, (b) to determine whether the removal of the in situ periosteum would enhance union between the new and pre-existing bones, and (c) to demonstrate the feasibility of creating onlays onto the craniofacial skeleton. Additionally, we were interested in determining whether transplants maintain their integrity long after bone formation has occurred, whether bone formation continues to occur at these late timepoints, and whether the mechanical characteristics of these long-term bones are consistent with the adjacent mouse calvarium.

Consistent with our three major aims, we determined that human BMSCs can form cortico-cancellous bone that closes a critical-sized calvarial defect and forms a union with the adjacent mouse calvarium. The removal of periosteum from the calvarium enhanced union with the transplant. Similar transplants placed onto the mandible also formed bone and bone unions but to a lesser degree than was seen in the calvarium. Addressing our three minor aims, we found that transplants underwent a gradual transformation in their morphology over time, with resorption of the HA/TCP particles and steady replacement with greater amounts of hematopoiesis. New bone formation continued to occur as late as 96 weeks after transplantation, confirmed by tetracycline labeling. The elastic modulus and hardness of the bone at these later timepoints resembled those of the normal mouse calvarium, and our use of AFM-based nanoindentation offered us the first opportunity to compare these small transplants against equally minute mouse bones.

Several issues are worth remarking upon. First, we saw a slight increase over time in the amount of bone in the transplants, with rare exception; up until 24 weeks, there was a corresponding and nearly parallel increase in bony union between the transplants and calvaria. These two trends closely tracked each other because a bony union nearly always occurred when new bone formed, and the absence of a bony union was often attributable to the absence of new bone in that same location. Second, the HA/TCP matrix maintained its overall shape and silhouette after placement, not requiring fixation to stay in place. Particles did not migrate away from the transplant, suggesting that these particles maintain a cohesion. Third, the poorer bone formation and bone union among the mandible onlay grafts defy easy explanation. The mandible and calvarial transplants were prepared simultaneously and were randomly assigned to one or the other site. We would suggest either that the motion of the mouse mandible and the loose skin in this area preclude steady transplant/mandible contact during the critical early phase of bone formation or that the particles we chose may have been too large to stably sit on the relatively convex surface of the mouse mandible; both mechanisms could result in disruption of capillary growth into the transplants and subsequent bone formation. Fourth, the HA/TCP transplants that were devoid of BMSCs failed to form significant bone, even as late as 96 weeks. The bone that did form was likely secondary to osteo-conduction or creeping substitution, involving migration of mouse osteo-progenitor cells from the calvarial margins into the HA/TCP matrix. The bone formation occurring through this process is generally slow, and as we expected, the amount of bone in the transplant should be much less than that achieved through the transplantation of the BMSCs.

Of note, bone formation has been observed when human BMSCs are transplanted with polymer scaffolds. Our own experience and review of the literature suggests, however, that although polymer scaffolds per se promote good bone formation by rodent BMSCs, they do not promote bone formation by human BMSCs. Polymer-associated bone formation by human BMSCs can be achieved by the additional placement of extra osteogenic substances, such as BMPs or dexamethasone with ascorbate, either as an addition to the matrix or as a product of BMSC transfection [26, [27], [28]–29]. We think that this additional level of complexity makes polymer scaffolds less practical than ceramic matrices for human bone tissue engineering.

In this study, BMSCs from two young donors were used, so the question may arise whether our results are applicable to the older patients who might be in the greatest need of bone tissue engineering. The question of age-related changes in BMSC osteogenic capacity has not been answered conclusively; while rabbit and rat BMSCs demonstrated an age-related decrease in their osteogenic potential both in vivo and in vitro [30, 31], comparable in vitro results for human BMSCs have been controversial [32, 33], and a decrease in the in vivo osteogenic potential of aging human BMSCs has not been convincingly demonstrated. The finding that human BMSCs from all age groups form bone following in vivo transplantation, although with a variable frequency [34], demonstrates that while some adjustments in cell number or other technical aspects may be necessary the technique of human BMSC transplantation may be successfully applied for bone tissue engineering in all age groups.

Conclusion

This study represents a unique model of bone formation in an orthotopic location using human BMSCs in conjunction with a mineral matrix and addresses a number of the questions that would be expected during eventual clinical BMSC transplantation.

Disclosures

The authors indicate no potential conflict of interest.

Acknowledgements

We thank Zimmer (Warsaw, Indiana) for its gift of HA/TCP. We also thank Sandra Choe, Dr. Bernard Halloran, Katherine Huang, and Dr. Vijay Tiwari for technical assistance. This research was supported in part by the University of California-San Francisco Research Evaluation and Allocation Committee and in part by the Intramural Research Program of the National Institute of Dental and Craniofacial Research, NIH, Department of Health and Human Services.

Ancillary