| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a Inserm U475, Montpellier, France;
b Molecular Pathology Laboratory, Hebrew University-Hadassah Faculty of Dental Medicine, Jerusalem, Israel;
c Institut Gustave Roussy, Villejuif, France;
d Hôpital Lapeyronie, Service d’Immuno-Rhumatologie, Montpellier, France
Key Words. Chondrocytes • Osteocytes • BMP-2 • Tet-Off system • Vascular endothelial growth factor • Intra-articular • Stromal progenitor cells
Danièle Noël, Ph.D., Inserm U475, 99 Rue Puech Villa, 34197 Montpellier Cedex 5, France. Telephone: 33-04-67-63-62-74; Fax: 33-04-67-04-18-63; e-mail: noel{at}montp.inserm.fr
 |
ABSTRACT |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
INTRODUCTION |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
More recent attempts of cartilage regeneration have been made using mesenchymal stem cells (MSCs) originating from the bone marrow [2, 3]. These progenitor/stem cells are pluripotent and can differentiate into multiple lineages such as myoblasts, adipocytes, osteocytes, and chondrocytes [4]. Such MSCs have thus been successfully used to repair large full-thickness defects created in the femoral condyle of rabbit [5]. However, although the subchondral bone was completely regenerated without loss of the overlying articular cartilage, progressive thinning and incomplete integration of the repaired tissue within the host cartilage were observed at 24 weeks, suggesting the need for local growth/differentiation factors.
Bone morphogenetic proteins (BMP) belong to the transforming growth factor ß (TGF-ß) superfamily, which participates in the regulation of cell migration, adhesion, multiplication, and differentiation throughout the life span of the organism [6]. Some of these growth factors display osteoinductive properties that induce ectopic bone and cartilage formation in vivo. Among these growth factors, BMP-2, BMP-4, and BMP-7 (also called osteogenic protein-1) have been largely used as recombinant proteins for their ability to repair bone defects in different animal models [7–9]. Based on this property, MSCs have been engineered to locally deliver BMP-2 in order to induce their differentiation into bone cells. Hence, using ex vivo adenoviral gene transfer, BMP-2-expressing MSCs were shown to form bone when injected into muscle [10]. More importantly, BMP-2-producing progenitor cells can be used successfully to heal segmental femoral defects in syngeneic Lewis rats [11, 12] and radial segmental defects in mice [13]. Fibroblast-mediated growth factor gene therapy has also been used for generating cartilage repair after transplantation into knee joints of mice [14]. In that study, it was suggested that secretion of BMP-2 by ectopic fibroblasts stimulated chondroid tissue and osteophyte formation, probably acting through a local stimulus on cells present in the joint. A similar approach was used in rabbits using TGF-ß1-expressing fibroblasts [15]. Interestingly, the results indicated that the local production of TGF-ß1 induced proliferation of chondrocyte and/or chondrocyte precursors inside the joint, leading to a gradual replacement of ectopic fibroblasts by rabbit chondrocytes, and generation of a fully differentiated hyaline cartilage at 6 weeks.
BMP-2 is also proposed to exhibit an angiogenic effect both in vitro, using a chick chorioallantoic membrane assay, and in vivo, after subcutaneous implantation of BMP-2-expressing MSCs that induced formation of blood vessels in mice [16]. In adults, angiogenesis is essential for pathological processes such as tumor growth and also for physiological processes such as tissue repair and regeneration during wound healing. Among the known angiogenic growth factors, the main effectors are the vascular endothelial growth factor (VEGF), the angiopoietin family, and their receptors [17]. In juveniles, VEGF is required for endochondral bone formation and longitudinal growth [18]. It has been shown that BMPs stimulate angiogenesis through the production of VEGF by osteoblasts [19], both acting synergistically to increase endochondral bone formation [20].
Taking into account the importance of the environmental factors on the differentiation capacities of stem cells in vivo, we investigated whether BMP-2-engineered MSCs form bone in the particular synovial/articular cartilaginous environment, as it has been reported in other ectopic sites. To this end, we used the C9 cell clone derived from the C3H10T1/2 murine mesenchymal progenitor cell line [16], expressing the human (h)BMP-2 gene under control of the Tet-Off system in an attempt to timely regulate the expression of the differentiation factor. These cells were injected into the intra-articular space of mouse knee joints, and we showed that BMP-2-expressing MSCs formed hypertrophic cartilage that was rapidly replaced by bone by a process resembling endochondral ossification. These results suggest that the endogenous factors secreted into the joint space had no detectable effect on the differentiation capacities of adult MSCs. Importantly, we showed that a short-term expression of the growth factor (BMP-2 in our experiments) is both necessary and sufficient to irreversibly induce bone formation.
|
MATERIALS AND METHODS |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
In Vitro Adenoviral Infection
First-generation E1-E3-deleted recombinant adenoviruses used were either the Ad-CO1 void vector or the Ad-K3 human serum albumin (HSA) vector expressing the angiostatin K3 HSA chimeric protein [21, 22]. Recombinant adenoviruses were expanded into 293 cells, twice purified on CsCl gradient ultracentrifugation, desalted, and frozen at -80°C in phosphate-buffered saline (PBS)-10% glycerol. Viral titers were expressed as plaque-forming unit/ml. C9 cells (5 x 106) were infected at subconfluence on 160-mm diameter culture dishes with the adenovirus vectors at the multiplicity of infection of 200. Two hours after infection, virus-containing medium was replaced with fresh complete medium. Following an 18-hour incubation, medium was withdrawn and stored at -20°C for enzyme-linked immunosorbent assay (ELISA). Cells were trypsinized and suspended in PBS for in vivo implantation experiments.
In Vitro Chondrogenesis and Osteogenesis Assays
Chondrogenic differentiation of C9 MSCs was induced by a 21-day culture in micropellet, as described elsewhere [23]. Briefly, MSCs (2.5 x 105 cells) were pelleted by centrifugation in 15 ml conic tubes and cultured in DMEM supplemented with 0.1 µM dexamethasone (Sigma), 0.17 mM ascorbic acid (Sigma), and 1% insulin-transferrin-sodium selenite supplement (Sigma), with or without addition of Dox. Osteogenesis was induced by culture of MSCs at low density (1.5 x 104 cells in a 100-mm diameter culture dish) for 21 days in complete DMEM supplemented with 10 mM ß-glycerophosphate (Sigma), 0.1 µM dexamethasone, and 0.05 mM ascorbic acid, with or without addition of Dox.
Semiquantitative Reverse Transcription Polymerase Chain Reaction (RT-PCR)
Total RNA was extracted from micropellets of C9 cells using RNA Plus (Qbiogene; Illkirch, France; http://www.qbiogene.com) according to the manufacturer’s recommendations. RT-PCR was performed on 1 µg of RNA using the GeneAmp RNA PCR Core Kit (Applied Biosystems; Courtaboeuf, France; http://home.appliedbiosystems.com). The primers for PCR were designed as follows: collagen II (forward, 5'-GCC TCGCGGTGAGCCTGATC-3'; reverse, 5'-CTCCATCTCT GCACGGGGT-3'), collagen X (forward, 5'-CCACCTGGGT TAGATGGAAAA-3'; reverse, 5'-AATCTCATAAATGGG ATGGG-3'), aggrecan (forward, 5'-TCCTCTCCGGTGG CAAAGAAGTTG-3'; reverse: 5'-CCAAGTTCCAGGGTC ACTGTTACCG-3'), GAPDH (forward: 5'-TGAAGGTCGG TGTGAACGGATTTGGC-3'; reverse: 5'-CATGTAGGCC ATGAGGTCCACCAC-3'). The primers used for detection of type II collagen allowed the amplification of two fragments corresponding to splice variants, namely, type IIA (472 bp) and type IIB (268 bp) collagens [24]. PCR products were electrophoresed in 1% agarose gel, stained with ethidium bromide, and visualized by ultraviolet transillumination.
ELISA
The secretion of mouse VEGF in culture supernatants was quantified by the DuoSet mouse VEGF sandwich ELISA kit, according to the supplier’s recommendations (R&D Systems; Minneapolis, MN; http://www.rndsystems.com). Statistical analysis was performed with the nonparametric Mann-Whitney test and analyzed by the program InStat (version 2.1 for Macintosh; GraphPad; San Diego, CA; http://www.graphpad.com). Production of the K3-HSA protein in culture supernatants or serum samples was determined using an ELISA specific for the K3 molecule, as described elsewhere [21]. The A405 was determined using a microplate reader (Bio-Rad). To evaluate the concentration of the K3HSA in samples, we postulated that the immunodetection of the fusion protein is identical to the human angiostatin used as standard in the assay.
CM-DiI Labeling
Stock solution of CM-DiI (Molecular Probes; Eugene, OR; http://www.probes.com) was reconstituted in DMSO at 1 µg/µl. Cells were trypsinized, washed with PBS, and suspended at the concentration of 107 cells/10 µg CM-DiI in 5 ml PBS prepared immediately before use. Cells were labeled by incubation at 37°C for 5 minutes and 4°C for 15 minutes, in the dark. Unincorporated fluorescent dye was then removed by centrifugation at 300 g for 5 minutes and two washes in PBS. Cells were resuspended at the final concentration of 106 cells/5 µl PBS and maintained at 4°C until injection.
Animals
CB17-severe combined immunodeficient (SCID)bg mice were purchased from Harlan (Gannat, France; http://www.harlan.com) and cared for according to the Laboratory Animal Care guidelines. For cell injection, mice were anesthetized with pentobarbital at the concentration of 50 mg/kg. For intra-articular injections, a skin incision was performed at the top of the knee to visualize the tendon that covers the joint. The cell suspension (10E6 cells, otherwise mentioned, in 5 µl of PBS) was delivered into the joint cavity of the knee using a 10-µl syringe with a 25-gauge needle (Hamilton, VWR International; Strasbourg, France; http://www.vwr.com). For intramuscular injections (10E6 cells in 50 µl of PBS), cells were administered into the tibialis anterior of mice. When necessary, 1 week before cell injection, Dox (Franvet Laboratory; Segré, France) was added to the drinking water (0.2 mg/ml) and changed every 2 days. At least five mice were included per group of treatment. Knees were dissected from euthanized mice at day 21, otherwise noted, and immediately examined by magnified x-ray and then processed for histology.
Histological Staining
Knee tissues were fixed in 4% formaldehyde solution and decalcified in acid-based DC3 solution (Labonord; Templemars, France; http://www.labonord.com) for 24 hours. The tissues were then dehydrated through a gradient of alcohol and toluene, embedded in paraffin, sectioned at 5 µm, and mounted on glass slides. Frontal sections were stained with hematoxylin and eosin before examination by light microscopy. Cartilage formation was visualized by Safranin O staining as described elsewhere [25]. Slides were then mounted in Eukitt glue (CML; Nemours, France).
Immunohistochemistry
Paraffin-embedded 5-µm sections were deparaffinized, hydrated, and antigen unmasking was performed in a bath of citrate buffer (C6H5O7Na3-2H20 0.1 M, C6H8O7 0.1 M) at 97°C for 40 minutes. Immunostaining was then performed using the UltraVision Mouse Tissue Detection System kit from Lab Vision Corporation (Fremont, CA; http://www.labvision.com) according to the manufacturer’s recommendations. Type II collagen was detected using a 1:50 dilution of a mouse monoclonal antibody (Interchim; Montluçon, France; http://www.interchim.com/) and type X collagen using a 1:10 culture supernatant dilution of a mouse monoclonal antibody (Quartett; Berlin, Germany; http://www.quartett.com).
|
RESULTS |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
). During differentiation, the type II procollagen mRNA switches from type IIA in undifferentiated MSCs to type IIB splice variant in chondrocytes [24]. In contrast, in C9 cells cultured with Dox and in control C3H10T1/2 cells, genes related to chondrogenic differentiation markers were undetectable (aggrecan) or slightly expressed (type X collagen) and the type IIA collagen splice form was predominant. Thus, C9 MSCs have undergone differentiation toward chondrocytes and secreted the components of the cartilaginous matrix under chondroinductive culture conditions.
|
). The injection of immortalized mesenchymal cells (NIH-3T3) led to tumor growth that invaded the joint and formed a protuberance at the knee surface but neither cartilage nor bone tissue (data not shown). We then tested the behavior of increasing numbers of C9 MSCs (104, 105, or 106 cells) in the articular environment of mice maintained without Dox diet to induce BMP-2 expression. On day 21 after cell injection, histological analysis showed that whatever the number of cells, both cartilage and bone could be detected in injected knees. The size of the newly formed tissue was related to the number of injected cells. The new chondroid/bony tissue was preferentially localized in the recesses of the synovial membrane or in the suprapatellar region. The neocartilage showed characteristic chondrocytic cells surrounded by large lacunae and revealed a high content in proteoglycans as shown by strong safranin O staining (Fig. 2C). Intense remodeling of the matrix that began to calcify was also evidenced, and mineralization was confirmed by x-ray (data not shown). The bone trabecula had developed and bone marrow had filled the medullary cavity.
|
, 2E). As control, mesenchymal-like tissue was observed in mice receiving C9 cells combined with Dox diet (Fig. 2B), suggesting a tight regulation of the Tet-Off system. In control contralateral knees, formation of new cartilage or bone structures was never observed (data not shown).
We then monitored the differentiation process at earlier time points (Fig. 3). As soon as day 10, calcification could be easily detected by x-ray in mice injected with 106 cells and by histology with 105 or 104 injected MSCs (data not shown). Moreover, the neotissue formation increased over time as observed by x-ray at day 60 (Fig. 3).
|
, 4B). On day 6, some mesenchymal condensations were still present and cartilaginous tissue resembling hypertrophic cartilage was observed as characterized by large round cells included in large lacunae (Fig. 4C). On day 8, large areas of hypertrophic-like cartilage were seen in the intra-articular space (Fig. 4D). On day 10, bony tissue was detected, with bone marrow invasion in some cases (data not shown). Immunohistological analysis revealed the expression of type II and type X collagens in the developing tissue as early as day 6 after C9 implantation. The type II collagen was poorly secreted, while type X collagen, specific for hypertrophic cartilage, was highly expressed in the chondroid tissue at day 8 (Fig. 4E, 4F). This differentiation program resembles the process of endochondral ossification in which a cartilaginous matrix, constituted of hypertrophic chondrocytes, is progressively replaced by a mineralized matrix.
|
). On the contrary, BMP-2 expression for 6 days induced formation of hypertrophic cartilage and bone filled with bone marrow (Fig. 5B). Similar results were obtained when BMP-2 was expressed for 8 days. In most cases, the labeled C9 cells were mainly localized in the hypertrophic cartilaginous tissue and, more rarely, in the bony tissue without any preferential localization. Some areas of the new bone were completely devoid of fluorescent cells (data not shown). Altogether, these observations showed that, in this model, a 5–6 day period of BMP-2 expression was sufficient to trigger MSC differentiation and to lead to bone formation.
|
We then compared VEGF secretion when MSCs were induced toward osteogenic or chondrogenic differentiation by BMP-2 expression. In both cases, VEGF production dramatically decreased shortly after induction of the differentiation. During osteogenesis, we observed a dramatic decrease to undetectable levels by day 10, whereas during chondrogenesis, a weak but sustained production of VEGF in the range of 15 pg/pellet/24 hours was detected during the overall differentiation period (Fig. 6). Because dexamethasone is known to repress VEGF secretion [26], we tested the behavior of MSCs in the absence of dexamethasone in terms of differentiation potential and VEGF production. In osteogenic conditions, VEGF levels were maintained around 40 pg/105 cells/24 hours, but MSCs rapidly stopped proliferating and stayed scattered (data not shown). In the case of chondrogenesis, VEGF secretion was comparable to that obtained in the presence of dexamethasone and stayed stable over time (data not shown). In both conditions, differentiation occurred (data not shown). Independently of the presence of dexamethasone, induction of osteogenesis and chondrogenesis highly decreased VEGF secretion by differentiating MSCs.
|
). The K3HSA fusion molecule was secreted in supernatants of transduced C9 MSCs at the level of 47 ± 12 ng/106 cells/24 hours but not in supernatants of nontransduced or control vector-transduced C9 cells (Fig. 7A). We then implanted the transduced C9 cells in the knee joints of mice and measured the quantity of K3HSA present in the serum (Fig. 7B). Stable expression of the exogenous protein could be determined throughout the experiment in the serum of injected mice (up to 700 ng/ml). No secretion of K3HSA could be detected in mice that received either nontransduced C9 cells or cells transduced with a void adenoviral vector (data not shown). Both radiological and histological analysis confirmed that transduced C9 cells had undergone differentiation, and ossification of the knee joints was observed. Immunohistochemical analysis of joint sections with an anti-smooth actin antibody did not reveal variation of blood vessel number in the new cartilaginous and bony tissues or in the surrounding tissues (data not shown). These results showed that in our conditions, an antiangiogenic agent, such as the K3HSA molecule, had no detectable influence on bone generation from engineered MSCs.
|
|
DISCUSSION |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
To test the effect of the joint environment on MSC differentiation, we chose the C9 cells that derive from the murine progenitor MSC line C3H10T1/2 and express the BMP-2 growth factor, a highly potent inducer of both chondrogenesis and osteogenesis. The immortalized, nontumorigenic C3H10T1/2 cells [31] are widely used both in vitro and in vivo as a reliable and relevant cell culture model for differentiation studies. Although they cannot totally mimic primary MSCs from various donors, they represent a suitable alternative. Our data show that C9 MSCs grown in vitro in micropellets were able to differentiate specifically into chondrocytes, but in vivo they formed hypertrophic cartilage and bone in the joint and muscle (see also [32]). These observations are in agreement with previous studies in the mouse showing that MSCs genetically modified to express BMP-2 were able to form bone when implanted in subcutaneous and intramuscular sites [10, 33] or in bone segmental defects [13, 16].
Other members of the BMP family, such as BMP-7, were also shown to convert human fibroblasts toward osteoblasts that were able to form bone in vivo [34, 35]. Interestingly, in the joint, the size of the osteochondral neotissue increased with time. If labeled C9 cells could be easily detected inside the tissue, other nonfluorescent cells were also observed. This could be due to the progressive loss of the fluorescent labeling during in vivo cell division or to the recruitment of host cells that might participate to the neotissue formation. This is supported by the study showing that BMP-2 is a potent chemotactic factor [36]. Contrary to C9 cells, nonengineered MSCs remained undifferentiated when implanted in the joint, whereas NIH-3T3 cells proliferated and formed tumors. Whether tumor formation was due to long-term culture conditions and subsequent phenotypic or genotypic alterations remains unclear. Thus, a potential effect of endogenous synovial/cartilaginous soluble mediators was not observed in our conditions as they were neither sufficient to induce the differentiation of naïve MSCs nor effective to regulate the differentiation of BMP-2-expressing MSCs toward chondrocytes. It is also possible that the presence of defects in the articular cartilage is necessary to induce the expression of molecules involved in migration, adhesion [37], or differentiation of MSCs. The behavior of naïve MSCs implanted in the joints of mice that develop osteo- or rheumatoid-arthritic lesions has still to be tested.
Because BMP-2 is finely tuned during limb bud embryogenesis [38], we explored the possibility that a short-term production of this growth factor could influence the lineage specificity. We thus took advantage of the regulable Tet-Off system to investigate the effects of the repression of BMP-2 secretion following various times of initial induction of MSC differentiation in vivo. We clearly show that 6 days of BMP-2 expression are enough to irreversibly trigger bone formation in vivo after generation of a cartilaginous matrix. Shorter exposure of MSCs to BMP-2 is not sufficient to trigger differentiation and induce osteogenesis. This suggests that short-term expression of the growth factor is necessary to initiate the early steps of MSC differentiation and is required to switch from the mesenchymal condensations to the cartilaginous tissue. This supports the notion that BMP signaling is involved in the early steps of chondrogenesis and osteogenesis of adult stem cells, as it has been shown in embryonic limb bud skeletogenesis [38, 39]. In the same way, a recent study showed in vitro that BMP-2 mediates initiation of the chondrogenic lineage of C3H10T1/2 cells but that further differentiation may be BMP-2 independent [40]. The autocrine loop induced by BMP-2 is likely activated by expression of endogenous BMP-2 and/or other BMPs, and could explain why late stages of chondrogenesis become independent of exogenous BMP-2. Our results thus confirm in vivo that a short exposure of MSCs to BMP-2 is sufficient to trigger osteogenesis.
The process of bone formation, as described in this study, greatly resembles endochondral ossification during which chondrocytes proliferate, become hypertrophic, and finally undergo apoptosis. Hypertrophic chondrocytes upregulate expression of VEGF and stimulate blood vessel invasion into the growth plate, necessary for recruitment of cells involved in bone morphogenesis [18]. Our in vitro studies reveal that A) the C3H10T1/2 cells constitutively secrete VEGF; B) the production of VEGF by MSCs in proliferating culture conditions is only modestly affected by BMP-2 expression, and C) the secretion of VEGF is highly reduced in the early phases of in vitro osteogenic and chondrogenic differentiations of MSCs. Contrary to a previous study reporting an increase of VEGF expression during osteoblastogenesis [19], the dramatic decrease in VEGF secretion is likely due to the presence of dexamethasone in the inductive medium used here. So, suppression of dexamethasone led to a low but stable level of VEGF during the overall period of culture without inhibiting cell differentiation. In our osteogenic culture conditions, dexamethasone seemed to be indispensable for proliferation but not for differentiation. However, if dexamethasone totally inhibited VEGF secretion during osteoblastogenesis, its effect was less pronounced during chondrogenesis. In fact, VEGF levels slightly rose as chondrocytes became mature and hypertrophic by the end of the culture. In these conditions, VEGF levels were probably underestimated on a per-cell basis due to the loss of viable cells, since it has been shown that approximately 40% of the cells undergo apoptosis in the course of micropellet cultures [41].
As previous data in vivo showed an impaired bone formation induced by a VEGF antagonist [18, 20], we investigated whether angiogenesis block could inhibit or delay osteogenesis in our murine model of C9 cell intra-articular injection. Due to the poor in vitro effect of BMP-2 expression on VEGF secretion by C9 cells, we chose to transduce C9 cells with the angiostatin K3 domain, which is not a specific inhibitor but acts by favoring antiangiogenic pathways. The K3HSA fusion protein was chosen based on previous results demonstrating the great impact of HSA genetic coupling on both the in vivo bioavailability of the antiangiogenic factor and pathological angiogenesis inhibition [21]. We showed that the local production of the K3HSA antiangiogenic agent had no effect on the fate of BMP-2-induced MSC differentiation. The lack of effect of this treatment could be due to a suboptimal concentration of the K3HSA in the joint, even though the local concentration is expected to be higher than the systemic measured levels (around 400 ng/ml). Thus, very recently, similar blood levels of the K3HSA molecule were reported after plasmid electrotransfer in the muscle of mice. In this case, no effect on tumor growth was observed [42], whereas a 10-fold higher concentration of the factor obtained after adenoviral gene transfer inhibited tumor growth [21]. Alternatively, although this molecule plays a role in counterbalancing the effects of proangiogenic factors, the use of an inhibitor that specifically targets one angiogenic factor might be more efficient in our model, as it has been shown with Flt-1 [20].
Similar to previous data on the implantation of BMP-2-expressing MSCs in ectopic sites in mice, the intra-articular injection of engineered MSCs reproduced the in vivo steps of differentiation toward the osteochondral tissue. The differentiation of adult MSCs induced by BMP-2 is thus independent of soluble factors present in the local environment of the synovial/cartilaginous tissues. Importantly, we demonstrate in vivo that a short-term expression of the BMP-2 growth factor is sufficient to irreversibly induce bone formation. This suggests that a stable genetic modification of MSCs may not be required in stem cell-based bone/cartilage engineering. An important issue yet to be solved for cartilage engineering will be to identify differentiation factors more specific for chondrogenesis. Such a specific factor has been recently identified [40], and the use of the microarray technology will make it possible to compare the fold changes in the levels of expression of multiple genes during differentiation [41]. More systematic approaches should allow analysis of the cellular and molecular events taking place during chondrogenesis and indicate specific molecules.
|
ACKNOWLEDGMENT |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
REFERENCES |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  V. Fritz, D. Noel, C. Bouquet, P. Opolon, R. Voide, F. Apparailly, P. Louis-Plence, C. Bouffi, H. Drissi, C. Xie, et al. Antitumoral Activity and Osteogenic Potential of Mesenchymal Stem Cells Expressing the Urokinase-Type Plasminogen Antagonist Amino-Terminal Fragment in a Murine Model of Osteolytic Tumor Stem Cells, November 1, 2008; 26(11): 2981 - 2990. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E. J. Koay, G. M. B. Hoben, and K. A. Athanasiou Tissue Engineering with Chondrogenically Differentiated Human Embryonic Stem Cells Stem Cells, September 1, 2007; 25(9): 2183 - 2190. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. Hou, K. Adriane, and P. K. Feng Bone in Growth in bFGF Versus BMP Combined with bFGF Saturated PLA-PEG-PLA: Experiment in Surgically Created Rabbit Mandibular Defects Journal of Bioactive and Compatible Polymers, September 1, 2005; 20(5): 455 - 474. [Abstract] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| STEM CELLS | THE ONCOLOGIST | CME | ALPHAMED PRESS JOURNALS |
