| Sign In to gain access to subscriptions and/or personal tools. |
Healing Cranial Defects with AdRunx2-transduced Marrow Stromal Cells
1 Program in Oral Health Sciences, Correspondence: * corresponding author, rennyf{at}umich.edu
Marrow stromal cells (MSCs) include stem cells capable of forming all mesenchymal tissues, including bone. However, before MSCs can be successfully used in regeneration procedures, methods must be developed to stimulate their differentiation selectively to osteoblasts. Runx2, a bone-specific transcription factor, is known to stimulate osteoblast differentiation. In the present study, we tested the hypothesis that Runx2 gene therapy can be used to heal a critical-sized defect in mouse calvaria. Runx2-engineered MSCs displayed enhanced osteogenic potential and osteoblast-specific gene expression in vitro and in vivo. Runx2-expressing cells also dramatically enhanced the healing of critical-sized calvarial defects and increased both bone volume fraction and bone mineral density. These studies provide a novel route for enhancing osteogenesis that may have future therapeutic applications for craniofacial bone regeneration.
Key Words: adenovirus • bone • regeneration • gene therapy • Runx2 • stem cells
Tissue engineering combined with gene therapy represents a promising direction for bone-regenerative medicine (Bruder and Fox, 1999). The use of adult stem cells as a platform for regeneration has received particular attention. These cells are found in several tissues, including marrow, skeletal muscle, and adipose tissue (Asahara et al., 2000). Marrow stromal cells (MSCs) contain pluripotent stem cells capable of differentiating into all the major mesenchymal cell types (Bianco et al., 2001). MSCs from animal and human sources have been successfully used to heal cranial and long-bone defects (reviewed in Krebsbach et al., 1999). However, considerable variability in regenerative outcomes has been observed between and among studies. The use of MSCs for repairing bone defects may be limited by the low frequency of osteoprogenitors in marrow and their loss during repeated passaging, as well as by an age-related decline in the numbers of these cells (Oreffo et al., 1998; Muschler et al., 2001; Byers and Garcia, 2004; Derubeis and Cancedda, 2004). To overcome these limitations, investigators must develop methods to increase the osteogenic activity of MSCs. One strategy for achieving this goal is to use gene-therapy-based expression of factors capable of increasing MSC osteogenic potential. For example, bone morphogenic protein (BMP) expression vectors were successfully introduced into MSCs to increase their osteogenic activity (Turgeman et al., 2001; Park et al., 2003; Tsuda et al., 2003). However, diffusion of BMPs away from the implant results in an essentially uncontrolled osteogenic response. As an alternative to BMP gene therapy, we used the bone-specific transcription factor, Runx2, to direct MSC differentiation into osteoblasts. Exogenous expression of Runx2 in mesenchymal cells or primary MSCs stimulated osteoblast gene expression and mineralization in vitro and increased bone formation after in vivo subcutaneous implantation (Yang et al., 2003; Zhao et al., 2005b). However, the potential of this approach to regenerate bone in an orthotopic site was not evaluated. In the present study, we tested the hypothesis that Runx2 gene therapy can be used to heal a critical-sized cranial defect.
Isolation and Transduction of MSCs MSCs were harvested from bones of C57BL6 mice as described previously (Krebsbach et al., 1998). Cells were plated at a density of 50,000/cm2 for use in cell culture studies. An adenoviral vector encoding the type II Runx2 isoform (AdRunx2) was generated as described previously (Yang et al., 2003). AdLacZ was purchased from the University of Michigan Vector Core. MSCs were transduced with adenovirus and cultured as previously described (Zhao et al., 2005b).
Animal Experiments
RNA Analysis
Radiography, Histology, and Micro-CT Analysis Radiographic analysis was performed with the use of a microradiographic apparatus (Faxitron X-ray Corporation, Wheeling, IL, USA). For histology, implants were fixed in 4% paraformaldehyde for paraffin embedding and sectioning (8-µm sections). Slides of adjacent sections were stained with hematoxylin and eosin for determination of tissue morphology. Calvaria were scanned by micro-computed tomography (mCT MS8X-130; EVS Corp., London, ON, Canada). Specimens were viewed with a scanning direction parallel to the coronal aspect of calvaria. High-resolution scanning, with an in-plane pixel size and slice thickness of 24 mm, was performed. To cover the entire thickness of the calvarial bone, we set the number of slices at 400. We used GEMS MicroView® software (Gems Corp., London, ON, Canada) to make three-dimensional (3-D) reconstructions from scans. We used a threshold value of 750 to obtain the 3-D image.
Statistical Analysis
Comparison of AdRunx2 Induction of Osteoblast Gene Expression in vitro and in vivo We previously characterized actions of AdRunx2 on osteoblast gene expression and mineralization of murine MSCs in vitro and showed that AdRunx2 could increase in vivo osteogenic activity after subcutaneous implantation (Zhao et al., 2005b). However, we did not examine gene expression in vivo. Because cells are exposed to very different environments in cell culture and in vivo, we compared the time-course of osteoblast marker gene expression under these two conditions. MSCs were transduced with AdRunx2 or AdLacZ (control vector) and either grown in cell culture or subcutaneously implanted into C57BL6 mice. In both groups, cells/implants were harvested at the indicated times for measurement of mRNAs by QRT-PCR. In cell culture, Ad-Runx2-transduced MSCs expressed higher levels of Runx2 mRNA than did controls (Fig. 1A  ). Runx2 mRNA levels were highest one day after transduction (90-fold induction). Although Runx2 expression decreased with time, it remained elevated relative to controls for the duration of the experiment. Ad-Runx2 transduction also increased expression of ALP, BSP, and OCN mRNAs, with ALP mRNA being expressed at early times, followed by BSP and OCN (Fig. 1B). ALP expression was higher in Runx2 MSCs for up to 6 days, and then rapidly declined to values comparable with those of the AdLacZ control. For BSP and OCN, a slower, but similar, induction profile was observed, such that values were equivalent to those of controls after 9 days (Figs. 1C, 1D).
In vivo expression profiles for implanted MSCs are shown in Fig. 2 . Overall, patterns were similar to those seen in cell culture, with some important differences. Ad-Runx2 transduction increased Runx2 mRNA 10-fold at the first timepoint examined (1 wk). Levels then declined at 2 and 3 wks, although they still remained elevated relative to those in controls. Similarly, ALP, BSP, and OCN mRNAs were increased by Runx2 after 1 wk (Figs. 2B, 2C, 2D). However, a gradual increase in expression was also observed in control implants, consistent with the known capacity of MSCs to differentiate into osteoblast in vivo in the absence of genetic modification (Krebsbach et al., 1998). After 3 wks, ALP mRNA in these cells was actually somewhat higher than that in the Runx2 group, while BSP and OCN mRNAs were equivalent.
AdRunx2-tranduced MSCs Stimulate Healing of a Critical-sized Calvarial Defect To evaluate the ability of AdRunx2-transduced MSCs to promote bone repair, we seeded cells onto a gelatin sponge and implanted them into critical-sized calvarial defects (Krebsbach et al., 1998). After 7 wks, radiographic analysis revealed that AdRunx2 MSCs had completely closed defects (Fig. 3A ), while only limited bone healing was observed with AdLacZ-transduced control cells. Previous work showed that AdLacZ and untransduced MSCs have equivalent osteogenic activity (Zhao et al., 2005b). Histological evaluation confirmed that complete bone bridging was achieved with AdRunx2 MSCs (Fig. 3B). Newly formed bone was continuous with host bone and contained a pronounced marrow space, with little residual gelatin scaffold (Fig. 3C). In contrast, only limited new bone formation was seen at the margins of defects in the control group. We used micro-CT images to evaluate the 3-D structure of the newly formed bone (Fig. 3D). Quantitative measurements demonstrated that the bone volume fraction of a standard cylindrical volume incorporating the defect increased from 41.01 ± 14.25% in controls to 55.79 ± 4.98% (p < 0.05) in the Runx2-treated group. Bone mineral density also increased from 279.3 ± 81.3 mg/cc to 347.9 ± 27.6 mg/cc (p < 0.05).
The combination of tissue engineering and stem cell therapy is a promising strategy for the regeneration of bone with well-defined morphologies (Rose and Oreffo, 2002; Warren et al., 2003a). In the present study, we showed that the osteogenic potential of MSCs was enhanced by adenoviral expression of Runx2, and that virally transduced cells can heal a critical-sized cranial defect. Osteoblast marker genes were strongly induced by AdRunx2-transduction of MSCs, regardless of whether cells were cultivated in vitro or implanted into subcutaneous sites. However, clear differences in the magnitude and duration of responses were observed. In vivo, control MSCs had a greater tendency to differentiate in the absence of exogenous Runx2. Specifically, ALP, BSP, and OCN mRNAs gradually increased over time, reaching expression levels that were equivalent to or greater than those of AdRunx2-transduced cells after 3 wks. This is consistent with reports of previous work showing that MSCs spontaneously formed bone in vivo (Krebsbach et al., 1998). Marker induction in vitro was also quite transient, and generally returned to basal levels after 9 to 12 days, while mRNA levels were more sustained in vivo. For example, in cell culture, ALP mRNA was maximally induced at 1 day and declined to control levels after 9 days. In vivo, levels also decreased to control levels after 2 wks, but these were still 40 to 50% of the levels seen at 1 wk. To explore the potential clinical relevance of this gene therapy approach, we evaluated the ability of AdRunx2 MSCs to repair a critical-sized cranial defect. X-ray evaluation of the newly formed bone revealed that Runx2-expressing cells completely repaired calvarial defects. In contrast, only partial healing was observed with control MSCs. The newly formed bone derived from AdRunx2 cells also had a higher bone mineral density. Histological examination demonstrated that newly formed bone completely covered defects implanted with AdRunx2 MSCs, while only limited new bone formation was seen in the control group. Consistent with these findings, quantitative analysis by micro-CT revealed that bone formed by AdRunx2 MSCs filled a greater fraction of the defect volume and had a higher mineral density than did controls. Interestingly, in the calvarial model, the newly formed bone contained marrow that was not observed in the previously described subcutaneous model (Zhao et al., 2005b). There are several possible explanations for this result. First, it is known that the dura mater underlying the cranium contains adult stem cells (Warren et al., 2003b). In response to surgical trauma, these cells may migrate into the defect and secrete growth factors (FGFs, BMPs) that are missing in the subcutaneous site. Alternatively, the cranial region is known to have a rich blood supply. The infiltration of new blood vessels into the scaffold could bring more oxygen and growth factors to the center of the scaffold. This would facilitate MSC survival, scaffold resorption, and marrow formation. Last, the thickness of the scaffold may also play a role in marrow formation (Krebsbach et al., 1998). As pointed out in the INTRODUCTION, considerable variability has been observed in the ability of MSCs to heal cranial defects. In our studies, AdLacZ-transduced control MSCs showed lower activity than was reported in certain studies where unmodified MSCs were used (Krebsbach et al., 1999). The basis for these differences is not currently known. However, it is unlikely to be explained by suppression of osteogenic activity by AdLacZ, which did not affect the osteogenic activity of MSCs after subcutaneous implantation (Zhao et al., 2005b). Retroviral vectors have the advantage of stable transgene expression and no immune response. However, potential risks associated with these vectors include insertional mutagenesis (Noguchi, 2003) and possible tumor formation associated with prolonged Runx2 expression (Blyth et al., 2001). Unlike retroviruses, adenovirus transgene expression is quite brief (St George, 2003). In the present study, AdRunx2 MSCs expressed recombinant protein for only approximately 3 wks, with maximal induction after 1 wk. This expression pattern is consistent with the results reported from a previous study that examined adenovirus expression of the Til-1 isoform of Runx2 (Kojima and Uemura, 2005). Short-term Runx2 expression may be too brief to have major deleterious effects. This hypothesis is strongly supported by the data of Kojima and Uemura (2005), who failed to observe significant changes in oncogene expression after Runx2 overexpression. Short-term Runx2 expression may also have certain advantages in terms of skeletal gene therapy. We found that transient expression of Runx2 was sufficient to induce complete bridging of a critical-sized calvarial defect. In contrast, sustained transgenic over-expression of Runx2 reportedly leads to osteopenia (Liu et al., 2001; Geoffroy et al., 2002). Recently, other groups reported on the use of Runx2 gene transfer to enhance the osteogenic potential of MSCs. However, equivocal outcomes were reported. Using a retroviral vector, Byers and co-workers assessed the ability of Runx2 overexpression in rat MSCs to heal a calvarial defect (Byers et al., 2006). After seeding cells in defects using a polycaprolactone scaffold, they observed no significant differences between the group implanted with MSCs expressing Runx2 and those implanted with empty scaffold. Mineral deposition occurred mainly in discrete local areas that mirrored the shape of the scaffold, and no integration with host bone was observed. In contrast, bone formation in our studies occurred mainly at the interface between the host and the implant, leading to complete integration of new and host bone. There are several possible explanations for these disparate results. First, differences between mouse and rat MSCs may contribute to the different healing outcomes (Krebsbach et al., 1997; Osyczka et al., 2004). Second, the biological responses of MSCs may be influenced by the different scaffolds used in these experiments (Boyan et al., 1996). The porous structure of the gelatin scaffold we used is known to facilitate in vitro cell seeding and in vivo infiltration by cells from surroundings sites (Takahashi et al., 2005). Additionally, the type I collagen used in gelatin scaffolds is known to induce osteoblast differentiation (Lynch et al., 1995; Xiao et al., 1998). Last, differences in Runx2 transgene expression patterns and durations may contribute to the different outcomes. As mentioned previously, different gene delivery vectors were also used in these studies. The retroviral vector used in the Byers study directs sustained Runx2 expression that may promote osteoclastogenesis and interfere with new bone formation (Geoffroy et al., 2002). Collectively, these results highlight the importance of cell-sourcing, scaffolds, and gene therapy vectors in the regeneration response. In conclusion, our results show that the lineage of MSCs can be directed toward osteoblasts by genetic manipulation of the Runx2 transcription factor. Furthermore, AdRunx2-transduced MSCs exhibited increased ability to heal a cranial defect when compared with AdLacZ-transduced control MSCs.
This work was supported by NIH Grant DE13386 (to R.T.F.), and by a Rackham Predoctoral Fellowship (to Z.Z.). The authors thank Drs. Jeong-Tae Koh and Di Jiang for technical support. This work was conducted in partial fulfillment of the requirements for the PhD degree in Oral Health Sciences at the University of Michigan (Z.Z.). Received for publication November 29, 2006. Revision received July 26, 2007. Accepted for publication August 6, 2007.
Journal of Dental Research, Vol. 86, No. 12,
1207-1211 (2007) This article has been cited by other articles:
|
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
 |
|
|
 |
|
Sign In to gain access to subscriptions and/or personal tools.
TOP
ABSTRACT
INTRODUCTION
