The Haploinsufficient Hematopoietic Microenvironment Is Critical to the Pathological Fracture Repair in Murine Models of Neurofibromatosis Type 1

Germline mutations in the NF1 tumor suppressor gene cause neurofibromatosis type 1 (NF1), a complex genetic disorder with a high predisposition of numerous skeletal dysplasias including short stature, osteoporosis, kyphoscoliosis, and fracture non-union (pseudoarthrosis). We have developed murine models that phenocopy many of the skeletal dysplasias observed in NF1 patients, including reduced bone mass and fracture non-union. We also show that the development of these skeletal manifestations requires an Nf1 haploinsufficient background in addition to nullizygous loss of Nf1 in mesenchymal stem/progenitor cells (MSCs) and/or their progenies. This is replicated in two animal models of NF1, PeriCre+;Nf1flox/− and Col2.3Cre+;Nf1flox/−mice. Adoptive transfer experiments demonstrate a critical role of the Nf1+/− marrow microenvironment in the impaired fracture healing in both models and adoptive transfer of WT bone marrow cells improves fracture healing in these mice. To our knowledge, this is the first demonstration of a non-cell autonomous mechanism in non-malignant NF1 manifestations. Collectively, these data provide evidence of a combinatory effect between nullizygous loss of Nf1 in osteoblast progenitors and haploinsufficiency in hematopoietic cells in the development of non-malignant NF1 manifestations.


Introduction
Bone integrity is maintained by the balance between osteoclasts originated from hematopoietic stem cells and mesenchymalderived osteoblasts. Osteoclasts are multinucleated cells specialized in bone resorption, whereas osteoblasts are bone forming cells derived from the mesenchymal stem cells. The development of the osteoblast and osteoclast lineages is tightly coupled [1]. Imbalances in coupling of the two lineages leads to defective skeletal development [2] as well as abnormal hematopoiesis [3,4].
Mutations in the NF1 tumor suppressor gene cause neurofibromatosis type 1 (NF1), a common chronic debilitating disorder that affects 1 in 3,000 individuals [5,6,7,8,9]. Neurofibromin, the protein encoded by NF1, functions at least in part as a GTPase activating protein (GAP) for Ras. Loss of heterozygosity of the tumorigenic cell is required for genesis of tumors in individuals with NF1 [10,11]. However, recent studies have also identified a key role of the microenvironment in tumor progression in both plexiform neurofibromas [12] and optic gliomas [13,14] in NF1. Despite advances in understanding the cell-cell interactions in cancers affecting NF1 patients, little information of the molecular and cellular signaling of the microenvironment of the prevalent nonmalignant manifestations of NF1 exists, particularly those impacting the skeletal system that are prevalent in ,50% of NF1 patients [6,7,8,9,15,16,17,18,19,20]. Clinical studies have found that individuals with NF1 are at significant risk for generalized loss of bone mass (osteopenia and osteoporosis) [21,22,23,24,25,26] as well as focal skeletal abnormalities [25,27,28,29]. One focal skeletal abnormality, non-union of spontaneous fractures, also called pseudoarthrosis, is particularly debilitating and frequently affects the distal tibia. Despite a range of orthopedic procedures attempting to facilitate fracture healing, the typical clinical outcome is non-union of the fracture and ultimately amputation of the affected limbs.
Skeletal development and remodeling depends on the balance of interactions between osteoblasts-mesenchymal stem cell derived cells, and osteoclasts-specialized macrophages derived from hematopoietic stem cells that govern skeletal remodeling and bone resorption [30]. We have previously shown that NF1 patients and Nf1 haploinsufficient mice have elevated osteoclast activity [31]. However, the increase in lytic activity of osteoclasts alone is not sufficient to explain some of the osseous manifestations observed in the NF1 patients. Kolanczyk and colleagues observed bowing of the tibia and diminished growth in Nf1 flox/flox mice that express Prx1Cre, which is expressed in undifferentiated mesenchymal cells in the limb bud. While those studies were informative concerning limb manifestations of the disease, the axial skeletal (e.g. kyphosis) are not modeled well due to a lack of expression of PrxCre in the vertebrae [32]. Elefteriou and colleagues utilized a genetic approach to delete Nf1 in the osteoblast lineage by intercrossing Nf1 flox/flox mice with a1(I) Collagen Cre transgenic mice and found that Nf1 inactivation in osteoblasts affects osteoclast differentiation in addition to bone formation [33,34].
In the present study, we utilized two novel murine models and established that the hematopoietic microenvironment plays a critical role in fracture healing in the NF1 murine models. Our study provides strong evidence of a non-cell autonomous mechanism for the skeletal manifestations of NF1 and demonstrates that the hematopoietic microenvironment is instrumental for bone loss and impaired fracture healing process.

Results
Detection of periostin expression in MSCs, osteoblasts, and generation of PeriCre + ;Nf1 flox/2 mice Endogenous full-length Periostin is expressed in primitive MSCs and epithelial cells [35,36]. To confirm 3.9 kb Periostin fragment promoter mediated Cre expression in adult MSCs and skeletal tissues, PeriCre + mice were intercrossed with transgenic reporter mice that have LacZ sequence knocked into the Rosa26 locus. The skeletal tissues of the intercrossed progeny were then subjected to b-gal staining. LacZ expression was abundant in the peripheral nervous system as expected [37] but was also detected in osteoblasts within the cartilaginous ribs as they begin the process of ossification, the long bone, and vertebrae cartilage of adult PeriCre + ;Rosa26 flox/flox mice (data not shown). However, LacZ staining was not observed in PeriCre 2 ;Rosa26 flox/flox littermates ( Figure S1A, c, d). The expression of LacZ in bone tissue was confirmed following LacZ staining and hematoxylin/eosin (H&E) counterstaining ( Figure S1A). Interestingly, LacZ expression within MSCs and skeletal tissues was detected only in mice 4 weeks of age or older, as PeriCre activity is initially restricted to the in utero heart valves, fetal and neonatal cardiac fibroblasts [38], and peripheral nervous system (Snider & Conway, unpublished data). Significantly, like the endogenous Periostin protein, Peri-Cre reporter expression is robustly localized to both the postnatal and adult outer periosteum, which gives rise to the bone collar covering all skeletal elements, including cancellous bone, and within the inner endosteal osteoblasts that secrete the vascularized bone-specific matrix [39]. LacZ expression was also examined in adult primary calvarial osteoblasts and MSCs in vitro following culture in specialized medium that selectively promotes the growth of these lineages. LacZ was strongly expressed in cultured adult calvarial osteoblasts and ,15% of the cultured MSCs were LacZ positive ( Figure S1B, a, b).
To investigate these skeletal phenotypes in more detail, a series of radiographic analyses were conducted to measure BMD and bone architecture. PeriCre + ;Nf1 flox/2 mice were shown to have reduced whole body BMD as compared to PeriCre + ;Nf1 flox/flox , WT, or Nf1+/2 mice ( Figure 1A) measured by peripheral Dual-Energy X-Ray Absorptiometry (pDEXA). When trabecular bone of 16week old mice were evaluated using micro-computed tomography (mCT), the bone volume (BV/TV) was similar in WT, Nf1+/2, and PeriCre + Nf1 flox/flox mice, while PeriCre + Nf1 flox/2 mice exhibited an ,23% reduction as compared to the other three genotypes ( Figure 1B PeriCre + ;Nf1 flox/2 mice have increased osteoclasts, reduced osteoblast formation, and altered bone turnover in vivo Biochemical markers of bone turnover are indicative of the overall bone metabolism and provide additional information besides the measurement of BMD [40]. Therefore, we evaluated the osteoclast activity in the trabecular bone of the four genotypes of mice by tartrate-resistant acid phosphatase (TRACP) staining and scoring the TRACP + osteoclasts surface (OcS) in the tibial trabeculae over trabecular bone surface (BS). A significantly greater OcS/BS ratio in the trabecular bone was observed in Nf1+/2 and PeriCre + ;Nf1 flox/2 mice than WT and PeriCre + ;Nf1 flox/flox mice as shown qualitatively and quantitatively in Figure 2A and Figure 2B, respectively. Furthermore, PeriCre + ;Nf1 flox/2 mice had significantly higher serum levels of TRACP5b compared to WT mice (1.26 unit60.06 vs. 1.82unit60.10, respectively, p,0.005) as measured by ELISA, suggesting an increased osteoclast development in vivo. In addition, a significantly lower number of osteoblasts was observed in Nf1+/2, PeriCre + ;Nf1 flox/flox and PeriCre + ;Nf1 flox/2 mice compared to WT mice, as evaluated by counting osteoblasts in tibial trabecular bone on MacNeal's stained sections ( Figure 2C-D). Interestingly, PeriCre + ;Nf1 flox/2 mice had reduced osteoblast numbers in comparison to PeriCre + ;Nf1 flox/flox mice given that both genotypes of mice had a similar genetic background in mesenchymal-osteoblast lineage. This suggests an additional effect of the haploinsufficient hematopoietic microenvironment on the skeletal deficits in PeriCre + ;Nf1 flox/2 mice in vivo.
Labeling bones with fluorochrome markers provides a means to study the dynamic changes of bone formation [41]. Thus, we examined bone remodeling in the four genotypes of mice using fluorescent labeling. There was a 30% reduction in mineralizing surface (MS/BS), a 20% reduction in mineral apposition rate (MAR), and a 50% reduction in bone formation rate (BFR/BS) in PeriCre + ;Nf1 flox/2 mice in comparison to control mice ( Figure 3A-D). Together, these data suggest that PeriCre + ;Nf1 flox/2 mice had a significant reduction in bone turnover in comparison to WT, Nf1+/2, or PeriCre + ;Nf1 flox/flox mice. The observation that skeletal phenotypes manifest in adult PeriCre + ;Nf1 flox/2 mice, but not PeriCre + ;Nf1 flox/flox or Nf1+/2 mice, suggests that the development of skeletal defects requires nullizygous loss of Nf1 in MSCs/ osteoblasts and haploinsufficient loss of Nf1 in other cell lineages.
PeriCre + ;Nf1 flox/2 mice have delayed fracture healing Approximately 3% of young children with NF1 acquire pseudoarthrosis [42]. To assess whether the fracture healing process is altered in PeriCre + ;Nf1 flox/2 mice, we induced three point bending fractures in the control and experimental mice. WT mice healed 28 days after the fracture as evaluated by X-ray (data not shown) and histological analysis following H&E, Masson Trichrome, McNeal staining and TRACP/Von Kossa staining (Figure 4 A, B). In contrast, only 15% of PeriCre + ;Nf1 flox/2 mice had healed fractures after 28 days. While WT mice had a normal architecture in the fractured tibia, Nf1+/2 and PeriCre + ;Nf1 flox/2 tibias showed deficient trabecular and cortical bone formation and increased osteoclasts in the callus site and impairment was most severe in Peri Cre+ ;Nf1 flox/2 mice ( Figure 4A, B). These data are summarized in Figure 4C. To evaluate whether the fractures in PeriCre + ;Nf1 flox/2 mice are delayed in healing or non-union, the fracture healing time was extended to 8 weeks after the fractures. As shown in Figure 4C, 33% of the PeriCre + ;Nf1 flox/2 mice withstood the non-unioned fractures at the end of 8-weeks.
Nf1+/2 bone marrow transplantation impairs fracture healing in Nf1 flox/flox ;PeriCre+ mice To formally test the hypothesis that the hematopoietic microenvironment has a key role in fracture healing in PeriCreNf1 flox/flox mice, bone marrow mononuclear cells (BMMNCs) from either WT or Nf1+/2 donor mice, that also express an EGFP transgene, were transplanted into PeriCre + ;Nf1 flox/2 , PeriCre;Nf1 flox/flox recipient mice that had been previously treated with ionizing radiation [31]. The hematopoietic reconstitution was verified with flow cytometric analysis demonstrating that the reconstituted mice had .95% EGFP positive donor bone marrow cells and clonogenic myeloid progenitors ( Figure S4). To clarify whether the donor mesenchymal stem cells were reconstituted in the recipient bone marrow, we cultured CFU-F using the recipient bone marrow cells. Two weeks later, we scored the CFU-F under the fluorescent microscope. All the CFU-Fs were negative in GFP, suggesting the mesenchymal stem cells were originated from recipients, but not the donor cells. This data indicated that the donor hematopoietic cells are key for the skeletal changes in the recipient mice.
To determine whether the transplantation of Nf1+/2 bone marrow cells to PeriCre + ;Nf1 flox/flox mice alters fracture healing, all the recipient mice underwent 3-point bending fracture and the fracture healing process was investigated. Reconstitution with Nf1+/2 BMMNCs resulted in a significant reduction in BV/TV as assessed by mCT (Figure 6 B-C). In addition, the fracture healing in PeriCre;Nf1 flox/flox mice transplanted with Nf1+/2 marrow was significantly impaired relative to those receiving WT bone marrow cells. This impaired fracture healing was confirmed by histological analysis (Figure 6D), which revealed that transplantation of Nf1+/2, but not WT BMMNCs, led to impaired fracture healing ( Figure D). Similar results were obtained in a parallel study using Nf1+/2 BMMNCs transplanted into Nf1 flox/lfox ;Col2.3Cre+ recipients ( Figure 6E). Most importantly, transplantation of WT BMMNCs restored the fracture healing of Nf1 flox/l2 ;PeriCre+ recipients ( Figure 6 A-D) and Nf1 flox/2 ;Col2.3Cre+ recipients ( Figure 6E). Collectively, these data indicate that the haploinsufficient bone marrow microenvironment plays an essential role in the skeletal manifestations of NF1, particularly in fracture non-union.

Discussion
Although skeletal manifestations collectively affect up to 50% of NF1 patients [44,45,46,47], the cellular mechanism(s) underlying the development of skeletal defects in NF1 patients are incompletely characterized. Previous attempts using other murine models have identified single lineage and phenotypic alterations that do not fully recapitulate the pathological phenotypes observed in individuals with NF1 [32,33,48,49]. The genetic intercrosses conducted in the present study provide novel murine models (PeriCre + ;Nf1 flox/2 mice and Col2.3Cre + ;Nf1 flox/2 mice ) that closely recapitulate multiple skeletal phenotypes observed in NF1 patients, particularly the pathological non-union of fractures affecting young children.
In the present study, we generated two novel murine models using Cre-mediated recombination under the direction of either the MSC/osteoblast-selective periostin fragment promoter or the osteoblast-specific Col2.3 Cre promoter. In these mice, one Nf1 allele was systemically deleted from all tissues (global knockout), while the remaining allele was conditionally deleted in mesenchymal stem/progenitor cells and/or osteoblasts. The two models allow us to compare the impact of the deficiency of Nf1 at different stages during MSC and osteoblast development of the skeletal deficits. PeriCre + ;Nf1 flox/2 mice and Col2.3Cre + ;Nf1 flox/2 mice show increased osteoclast development compared to control mice in vivo. In addition, these mice display reduced bone mineral density (BMD) and impaired fracture healing compared to control mice.
Recent studies have begun to elucidate the interactions between haploinsufficient and nullizygous Nf1 cells within the slow growing tumors found in individuals with NF1. Consistent with Nf1 being a tumor suppressor gene, Zhu, et al have shown that nullizygosity of Nf1 in the tumorigenic Schwann cell is required for plexiform neurofibroma progression [15]. However, in this model which results in a small fraction of Nf12/2 Schwann cells that likely appear analogous to the human situation, haploinsufficiency in non-tumorigenic lineages is also required [15]. Furthermore, Daginakatte, et al. have shown a similar phenomenon in optic gliomas in the NF1 murine model [13,14]. Recently, we reported that haploinsufficiency of Nf1 in the hematopoietic microenvironment specifically has a critical role in plexiform neurofibroma formation [12]. Thus, while there is a structural framework that explains the development of slow growing tumors in NF1 patients, no such pathogenic framework to explain the genesis of nonmalignant conditions exists. In the present study, we showed that the full development of skeletal manifestations in the NF1 murine model requires both the loss of Nf1 in osteoblast precursors/ osteoblasts and the haploinsufficiency of Nf1 in the hematopoietic microenvironment. This data is consistent with a single case report by Stevenson and colleagues who found that localized nullizygous loss of NF1 in pseudarthrosis tissue is acquired by micro-dissection and NF1 haploinsufficiency in peripheral blood [29].
Given that our adoptive transfer study indicates that the Nf1 haploinsufficient bone marrow cells are important for non-union to occur and WT bone marrow cells restores fracture healing, these observations lead to opportunities of clinical and therapeutic significance. Here, we show that PeriCre + ;Nf1 flox/2 mice transplanted with Nf1+/2, but not WT bone marrow, exhibit impaired fracture healing. Most importantly, WT transplantation of bone marrow restored fracture healing of PeriCre + ;Nf1 flox/2 and Col2.3Cre + ;Nf1 flox/2 mice. We previously reported a key role for Nf1 haploinsufficiency in the elevated bone resorptive activity in both murine and human NF1 osteoclasts [31]. This lineage is likely the key effector of the microenvironment. To our knowledge, this study provides the first evidence of a non-cell autonomous mechanism for the pathologic development of non-malignant NF1 pathologic manifestations. As such, these models provide an excellent platform to further study the cellular and molecular mechanism(s) underlying skeletal manifestations in NF1 patients. Future research investigating the specific lineage and molecular targets in the microenvironment that are critical for the non-union fracture healing in NF1 is necessary.
Collectively, these data demonstrate a critical role of the hematopoietic microenvironment in fracture healing in the NF1 population. Our study provides strong evidence of a non-cell autonomous mechanism for the skeletal manifestations of NF1 and demonstrates that the hematopoietic microenvironment is instrumental for bone loss and impaired fracture healing process.

Ethics statement
Animal care and experiments were conducted according to the guidelines established by the Indiana University Animal Care and Use Committee (IACUC). This study was approved by IACUC via Protocol Number 0000003401 originally approved on 6/19/ 09.

Animals
Nf1+/2 mice were obtained from Dr. Tyler Jacks at the Massachusetts Institute of Technology (Cambridge, MA) [50]. Nf1 flox/flox mice were provided by Dr. Luis Parada at the University of Texas Southwestern Medical Center [15]. 3.9kb peri-EGFP/Cre (PeriCre) transgenic mice were generated as described elsewhere [38,51]. To examine the lineage distribution of Cre-Mediated recombination, adult PeriCre transgenic mice were crossed with ROSA26 reporter mice. Col2.3Cre transgenic mice were generated as described elsewhere [52]. All studies were approved by the Institutional Animal Care and Use Committee.

Histochemical detection of b-gal activity
To examine the Cre recombinase expression pattern in MSC and skeletal tissues, LacZ expression was determined as described previously [52]. Histochemical detection of b-gal activity was   Longitudinal in vivo peripheral dual-energy X-ray absorptiometry (pDEXA) To evaluate the bone mass in each genotype of mice, bone mineral density (BMD) of the tibial diaphysis was evaluated in vivo using pDEXA (PIXImus II; GE-Lunar Corp., Madison, WI) [48]. Adult mice were anesthetized via inhalation of 2.5% isoflurane (IsoFlo; Abbott Laboratories, North Chicago, IL) mixed with O 2 (1.5 liter/min). The mice were placed in a prone position on the specimen tray and scanned. The head was excluded from total body scans. The region of interest included the central 50% of the whole tibia.

Micro (m) computed tomography
Bone volume and microarchitecture in the distal femoral metaphysis were evaluated using a high-resolution desktop microcomputed tomography imaging system (mCT-20; Scanco Medical AG, Basserdorf, Switzerland). For evaluation of the trabecular envelope at the distal femoral metaphysis, each specimen was scanned with a slice increment of 9 mm. CT images were reconstructed, filtered (s = 0.8 and support = 1.0), and thresholded (22% of maximum possible gray scale value) as previously described [53]. Scanning for the femur was started at 15% of the total femur length measured from the tip of the femoral condyle and extended proximally for 200 slices. The area for trabecular analysis was outlined within the trabecular compartment, excluding the cortical and subcortical bone. Every 10 sections were outlined manually, and the intermediate sections were interpolated with the contouring algorithm to create a volume of interest. Parameters of microarchitecture for both skeletal sites included bone volume (BV, mm 3 ) and bone volume fraction (BV/TV, %), as well as trabecular number (Tb.N, mm 21 ), trabecular thickness (Tb.Th, mm), and trabecular separation (Tb.Sp, mm).

Fluorochrome bone labeling and histomorphometric measurements
Fluorochrome labeling of the bones was performed in 8 weekold mice by intraperitoneal injections of calcein (20 mg/kg, Sigma Chemical, St. Louis, MO, USA) and alizarin (20 mg/kg, Sigma) 8 and 3 days before sacrifice, as previously described [54]. After sacrifice, femurs were fixed in 10% neutral buffered formalin for 48 hours, dehydrated in graded ethanols, and embedded undecalcified in methylmethacrylate. Frontal sections (4 mm thick) were cut from the distal femur using a motorized microtome equipped with a tungsten-carbide knife (Leica Inc, Deerfield Il). The sections were reacted with tartrate-resistant acid phosphatase (TRACP) for identifying and measuring the osteoclast surface or mounted unstained for fluorochrome-derived bone formation parameters. One section per tibia was viewed at 1606 magnification on a Leitz DMRXE microscope (Leica Mikroskopie und System GmbH, Wetzlar, Germany) and the image captured using a SPOT digital camera (Diagnostic Instruments, Inc., Sterling Heights, MI). The measurement area for the distal femoral metaphysis was determined by a region beginning 0.5 mm proximal to the midpoint of the growth plate, non-inclusive of cortical bone, and extending proximally for a total area of approximately 2.8 mm 2 . Trabecular bone turnover was assessed by measuring the extent of single label (sL.Pm), double label (dL.Pm) and the area of bone (dL.Ar) between the calcein and alizarin labels using Image Pro Plus version 4.1 software (Media Cybernetics, Silver Spring, MD). Derived histomorphometric parameters included mineralizing surface (MS/BS, %), a measure of active bone-forming surface, calculated as follows: the [K sL.Pm+dL.Pm]/Tt.Pm * 100; mineral apposition rate (MAR, mm/ day), a measure of the rate of radial expansion of new bone, calculated as follows: dL.Ar/dL.Pm/4 dy; and bone formation rate, an overall measure of bone formation that combines MS/BS and MAR, calculated as follows: MS/BS * MAR * 3.65.
For histological analysis of cell populations, after sacrifice, tibiae were dissected from the mice and were fixed in 10% neutral buffered formalin for 48 hours, dehydrated in graded ethanols, and embedded undecalcified in methylmethacrylate. Frontal sections (4 mm thick) were cut from the tibiae using a motorized microtome equipped with a tungsten-carbide knife (Leica Inc, Deerfield Il). The sections were reacted with TRACP for identification and measuring of osteoclast surface. MacNeal and /or von Kossa staining were utilized for identifying osteoblasts and calcified bones. The sections were viewed at 2006 magnification on a Leitz DMRXE microscope (Leica Mikroskopie und System GmbH, Wetzlar, Germany) and the images were captured using a SPOT digital camera (Diagnostic Instruments, Inc., Sterling Heights, MI). The quantification of osteoblast cell number and osteoclast surface were normalized to the bone surface in millimeters (mm), to control for the total amount of trabecular bone surface within each histological section.
Generation of fracture and X-ray scan A single tibial fracture was induced as described previously [55,56,57]. With the animal lying supine on the operating table, an intramedullary rod (a 30-G needle) was surgically inserted via the proximal end of the tibiae into the medullary canal of the tibia, so as to stabilize the bone post-fracture and permit load-bearing. A 200-gram weight was dropped from a height of 30-cm to fracture the tibia between the middle to lower 1/3 of the bone. Mice were X-rayed following the fracture to verify standardization of the portion of each femur has been removed for the metaphyseal architecture analysis. Deficiencies in the trabecular network are indicated by the red arrows. (D) Quantitation of 3D reconstruction and BV/TV ratio analysis. Trabecular bone formation is decreased in both Nf1+/2 (*p,0.05) and Col2.3Cre + ;Nf1 flox/2 (**p,0.01) mice as compared to WT and Col2.3Cre + ;Nf1 flox/flox mice. Four month old WT (n = 5), Nf1+/2 (n = 5), Col2.3Cre + ;Nf1 flox/flox (n = 5), and Col2.3Cre + ; Nf1 flox/2 (n = 5) mouse femur samples were processed for analysis. TRACP staining (E, F) and MacNeals' staining (G, H) were performed. (E) Representative photographs of the trabecular bone following TRACP staining (upper panel 46   fractures. The mice were allowed to recover on a heated pad and then placed in recovery cages. Fracture repair was monitored by weekly radiography (pixRay-100, Bioptics, Phoenix, AZ) and cases where the internal fixation failed, (due to pin slippage, bending or breakage) the affected mouse was euthanized and excluded from subsequent analysis.

MSC and committed osteoprogenitor culture in murine specimens
MSCs were generated from each experimental group of mice as previously described [49]. Briefly, BM mononuclear cells (BMMNCs) were separated by low density gradient centrifugation from 6,8-week old mice. BMMNCs were then resuspended and cultured in mouse MesenCult basal medium containing MesenCult Supplement (Stem Cell Technologies Inc.) at a concentration of 2610 6 cells/mL in a 10-cm tissue culture plate as previously described. When the cultures reached 80-90% confluency, cells were trypsinized and replated at a concentration of 5610 5 cells/ 75 cm 2 . MSCs at passage 5 to 10 were used for the following experiments. To culture committed osteoprogenitors, calvaria or femur distal metaphyses were resected and chopped finely with a scalpel. Bone chips were exposed to 0.25% trypsin for 1 hour and then passed through a 70-mm cell strainer (Falcon, Franklin Lakes, NJ, USA) to remove debris and matrix remnants. Cells were cultured in a-MEM medium and supplemented with 10% fetal bovine serum (FBS) and 1% penicillin-streptomycin at 37uC, 5% CO 2 .
Statistical analysis. One way ANOVA and post t-test was used for all the statistical analysis in this study. P values less than 0.05 are considered as statistically significant.