Adenoviral BMP-2 Gene Transfer in Mesenchymal stem cells

Adenoviral BMP-2 Gene Transfer in Mesenchymal stem cells

(Parte 1 de 3)

Adenoviral BMP-2 Gene Transfer in Mesenchymal Stem Cells: In Vitro and in Vivo Bone Formation on Biodegradable Polymer Scaffolds

Kris Partridge,*,1 Xuebin Yang,*,1 Nicholas M. P. Clarke,* Yasunori Okubo,† Kazuhisa Bessho,† Walter Sebald,‡ Steven M. Howdle,§ Kevin M. Shakesheff,¶ and Richard O. C. Oreffo*,2

*University Orthopaedics, University of Southampton, General Hospital, Southampton SO16 6YD, United Kingdom; †Department of Oral and Maxillofacial Surgery, Graduate School of Medicine, Kyoto University, Sakyo-ku, Kyoto 606-8507, Japan; ‡Department of Physiological Chemistry I, Biocenter, University of Wurzburg, Wurzburg, Germany; and §School of Chemistry and ¶School of Pharmaceutical Sciences, University of Nottingham, University Park, Nottingham NG7 2RD, United Kingdom

Received February 12, 2002

The aim of this study was to determine the feasibility of adenoviral gene transfer into primary human bone marrow osteoprogenitor cells in combination with biodegradeable scaffolds to tissue-engineer bone. Osteoprogenitors were infected with AxCAOBMP-2, a vectorcarryingthehumanBMP-2gene.Alkalinephosphatase activity was induced in C2C12 cells following culture with conditioned media from BMP-2 expressing cells, confirming successful secretion of active BMP-2. Expression of alkaline phosphatase activity, type I collagen and mineralisation confirmed bone cell differentiation and maintenance of the osteoblast phenotype in extended culture for up to 6 weeks on PLGA porous scaffolds. In vivo implantation of adenoviral osteoprogenitor constructs on PLGA biodegradeable scaffolds, using diffusion chambers, also demonstrated bone cell differentiation and production of bone tissue. The maintenance of the osteoblast phenotype in extended culture and generation of mineralised 3-D scaffolds containing such constructs indicate the potential of such bone tissue engineering approaches in bone repair. © 2002 Elsevier Science (USA)

Key Words:osteoprogenitor;adenoviralgene-transfer; biodegradable polymer; poly(DL-lactic- co- glycolic acid); PLGA; tissue engineering.

The replacement of bone is a major clinical and socioeconomic need. Approaches include the use of autol- ogous and allogeneic bone to restore the function of traumatised or degenerated bone tissue. However, a lack of sufficient material precludes the universal use of autogenous bone while the use of allogenic bone for transplantation carries potential risks of immune responses, pathogen transmission and the necessary immunosuppression. Tissue engineering has emerged as a possible alternative strategy to regenerate bone. Three components are essential: isolation and expansion of osteoprogenitors or mesenchymal stem cells, provision of appropriate osteoinductive factors and an appropriately designed scaffold that mimics the structural environment to promote bone regeneration.

Bone marrow contains multipotential stromal stem cells which can differentiate into fibroblastic, osteogenic, adipogenic and reticular cells (1–3). A number of studies have shown that human bone marrow osteoprogenitors can be isolated using selective markers and are readily expanded while retaining their differentiation ability, indicating their potential for marrow repopulation (4–7). Individual fibroblastic colonies have been shown to give rise to an osteogenic tissue within diffusion chambers while studies in a variety of animal species and more recently in preclinical trials in humans, indicate marrow tissue is capable of extensive osteogenesis (8–15).

The osteoinductive factors of choice are the bone morphogenetic proteins (BMPs), pivotal in the process of bone formation (16, reviewed in 17, 18). Over 30 distinct forms of the BMPs exist although the most widely studied are BMP-1 through 7. The BMPs induce differentiation of multipotential mesenchymal cells (19), pluripotent murine stem cell cultures and rat bone marrow stromal cells as well as proliferation and maturation in osteoblast populations (19–23).K.P. and X.Y. contributed equally to this work.To whom correspondence and reprint requests should be ad- dressed at University Orthopaedics, MP817, Southampton General Hospital, Southampton, S016 6YD, UK. Fax: 4 2380 796141. E-mail: roco@soton.ac.uk.

Biochemical and Biophysical Research Communications 292, 144–152 (2002) doi:10.1006/bbrc.2002.6623, available online at http://www.idealibrary.com on

144006-291X/02 $35.0 © 2002 Elsevier Science (USA)

All rights reserved.

Although BMPs can induce bone formation, the inability to identify a suitable delivery system has, to date, limited their clinical use. The use of a cell delivery vehicle (gene therapy) offers a simple solution and a number of studies in recent years have examined the potential of adenoviral delivery of BMP-2 in target cell populations (24–27). These studies demonstrate bone formation in vivo; however, a number of issues regarding the use of adenovirus remain, including potential immune response, clearance, and safety of the virus, quantification and duration of protein production.

Synthetic materials, such as poly(lactic acid) (PLA), poly(DL-lactic- co- glycolic acid) (PLGA), polyglycolic acid (PGA) have emerged as potential scaffolds for cell transplantation and tissue growth. These materials have the advantage of FDA approval and are currently used as suture materials and in drug delivery vehicles. Moreover, protocols for the generation of these materials with defined porosity using super critical fluid mixing and procedures for the surface modification of these materials with biological agents, have been developed (reviewed in 28, 29). The resultant scaffolds generate attractive biomimetic scaffolds for cell growth and differentiation including isolated human osteoprogenitors (30–32).

Thus, an alternative attractive approach for skeletal repair is the selection and genetic modulation of primary human osteoprogenitor cells in combination with biodegradable polymer scaffolds, which interact and promote osteoblast differentiation and osteogenesis. These studies demonstrate bone formation, in vitro and in vivo, using human osteoprogenitors secreting active BMP-2 on a biodegradeable polymer scaffolds. The formation of mineralised bone tissue within diffusion chambers in nude mice highlight the potential to tissue- engineer bone for orthopaedic use.

Materials. Tissue culture reagents were obtained from Gibco/

BRL (Paisley, Scotland). Fetal Calf Serum was from Meldrum Ltd (Bourne End, UK). Resin support was purchased from Novabiochem (Calbiochem-Novabiochem (UK) Ltd, Beeston, Nottingham). Dexamethasone, alkaline phosphatase kits and all other biochemical reagents were of analytical grade from Sigma Chemical Company (Poole, Dorset) unless otherwise stated.

Cell culture. Bone marrows were obtained from haematologically normal patients undergoing routine total hip replacement surgery. Only tissue, which would have been discarded, was used with the approval of the Southampton General Hospital Ethics Committee. Primary cultures of bone marrow cells were established as previously described (3). In brief, marrow cells were harvested using Minimal Essential Medium - alpha modification ( MEM) from trabecular bone marrow samples and pelleted by centrifugation at 500 g for 5 min at 4°C. The cell pellet was resuspended in 10 ml MEM and passed through nylon mesh (70 m pore size; Becton–Dickinson, UK). Cells were maintained in 10% FCS in MEM. C2C12 and HEK293 cells were grown in Dulbecco’s modified Eagle medium with 10% FCS. For cell growth on PLGA scaffold, following trypsinization and resuspension, in serum-free MEM, 1 10 cells were then added to individual wells of 24-well plates containing PLGA scaffolds After 24 h, the media was removed and cultures maintained in MEM supplemented with 10% FCS for up to 6 weeks.

Infection of cells with adenovirus expressing BMP-2. Cell lines were transduced with AxCAOBMP-2, a replication-deficient adenoviral vector carrying the human BMP-2 gene or AxCALacZ (hereafter termed AdBMP2 or AdLacZ, respectively), a control vector carrying the Escherichia coli (E. coli) LacZ gene. The construction of AdBMP2 has been described previously (34). Virus was amplified in 293 cells, purified through a CsCl cushion and titrated as previously described (35). Cells were infected with the adenovirus once confluence had been reached and maintained for several days. Virus was added to the cells at various multiplicities of infection (MOI) in media containing 5% FCS. Flasks were rotated every 30 min for 1.5 h before addition of the same volume of fresh 5% FCS MEM.

X-gal histochemistry and immunohistochemistry. Expression of -galactosidase was visualised by staining with X-gal. Following 72 h exposure to AdLacZ at various MOI cells were fixed with 4% paraformaldehyde for 10 min at room temperature. Cells were stained for

FIG. 1. Cultured primary human bone marrow cells were grown to confluence and transduced with AxCALacZ at increasing MOI or mock infection. After 3 days the cells were fixed and stained for -galactosidase activity. (A) MOI 20. (B) Mock infection.

FIG. 2. C2C12 cells stained for alkaline phosphatase activity after exposure to conditioned media from AdBMP2 transduced cells or Mock transduced cells. Virus was added to confluent human bone marrow cells, washed after 24 h and given fresh medium. After 3 days the medium was removed, filtered, and applied to confluent C2C12 cells for 3 days. After fixation, the cells were stained for alkaline phosphatase activity.

Vol. 292, No. 1, 2002 BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS

1–3ha t3 7°C using a solution containing X-gal (20 mg/ml), 5 mM potassium ferricyanide, 5 mM potassium ferrocyanide, and 2 mM magnesium chloride in PBS.

C2C12 alkaline phosphatase assay. BMP-2 activity was screened by conditioned media transfer to promyoblastic C2C12 cells as previously described (36). Briefly, following 24 h exposure to AdBMP-2, the media was removed and replaced with fresh. At various time intervals conditioned media was filtered and transferred to confluent C2C12 cells in a 96-well microtitre plate. Dilutions, 1:1, 1:2, or 1:10, were prepared with fresh 2% DMEM. Cells were incubated for a further 72 h before assaying for alkaline phosphatase activity. The effect of recombinant BMP-2 was measured in a range from 10–200 nM over 72 h in DMEM supplemented with 2% FCS. Cells were washed twice with PBS then fixed with cold 95% ethanol. Cell layers were washed with phosphate-buffered saline (PBS) and stored at 70°C until assayed for alkaline phosphatase activity. For assay, the cell layer from each well was scraped into 0.2 ml 0.05% (v/v) Triton X-100. Alkaline phosphatase activity was measured using p-nitrophenyl phosphate as substrate in 2-amino-2-methyl-1-propanol alkaline buffer solution (1.5 M, pH 10.3 at 25°C). DNA content was measured using PicoGreen as per manufacturer’s instructions (Molecular Probes, Leiden, The Netherlands). Alkaline phosphatase specific activity was expressed as nanomoles of p-nitrophenol/min/ng DNA

Scaffolds. Poly(DL-lactic- co- glycolic acid) (PLGA 75:25) (Mw 22K) porous (200 m) scaffolds were used in all studies. The scaffolds were produced by a supercritical carbon dioxide method in which the polymer is plasticised at 35°C under a pressure of 1500 psig (29). On release of the pressure, pores are formed in the polymer by the escape of the carbon dioxide gas. The PLGA used in this study will dissolve in the body over 10–12 weeks and was selected on the ability of incorporated glycolic acid to allow sufficiently rapid degradation of the poly lactic acid component. Porous scaffolds were sterilized using 70% ethanol for 3 h and coated with MEM supplemented with 20% FCS for 3 h.

Alkaline phosphatase expression. Following 72 h exposure to media conditioned by AdBMP-2-infected cells, C2C12 cells were rinsed twice with PBS then fixed for 10 min in 95% ethanol at room temperature. Staining was with a Sigma alkaline phosphatase kit (No. 85) used according to the manufacturer’s instructions.

Cell viability. Adenovirally transduced human bone marrow cells were incubated with Cell Tracker green (5-chloromethylfluorescein diacetate, CFMDA) (Molecular Probes, Leiden, The Netherlands) and Ethidium Homodimer-1 (EH-I) (Molecular Probes, Leiden, The Netherlands) for 45 min to label viable and necrotic cells respectively. The medium was then replaced and the cells incubated for a further hour.

Microscopy and image analysis. Images from PLGA porous scaffolds were taken using an inverted microscope (Leica DMIRB/E), equipped with a fluorescence filter enabling fluorescent imaging. Cells labeled with CFMDA and EH-1 were recorded on a Leica Leitz DM RBE with an X50 water immersion objective. Electron microscopy was undertaken using a Hitachi S-800.

Histochemistry and immunohistochemistry. Prior to immunocytochemical and histochemical analysis, PLGA scaffold samples were fixed with 4% Paraformaldehyde or 95% ethanol, dependent on the staining protocol and, as appropriate, processed to paraffin wax and 5- m sections were prepared. Controls (omission of primary antibody) were included in all studies.

Alcian blue/Sirius red staining. Sections were stained using

Weigert haematoxylin solutions prior to staining with 0.5% Alcian blue. After treatment with 1% molybdophosphoric acid, samples were stained using 0.1% Sirius red.

Toluidine blue and von Kossa staining. Samples were stained with 1% AgNO under UV light for 20 min until black deposits were visible and after air drying, slides were counterstained with toluidine blue.

Type I collagen. Reactivity to Type I collagen antibody (LF 67,

Dr. Larry Fisher, NIH, U.S.A.) was assessed after fixation in 4% Paraformaldehyde for 3 h. Endogenous peroxidase activity was blocked using 3% H O prior to incubation with LF 67 (1:300 in PBS) for3ha t4 °C. Samples were incubated with peroxidase-conjugated anti-rabbit IgG (1:30 in PBS) and peroxidase activity was detected using 3-amino-9-ethyl-carbazole in acetate buffer containing H O . Samples were counterstained with Mayer’s haematoxylin.

In vivo studies. Adenovirally labeled human bone marrow cells were cultured in MEM medium containing 5% (v/v) FCS prior to intraperitoneal implantation (2 10 cells/chamber) using diffusion chambers (130 l capacity; Millipore, UK) in MF1-nu/nu mice. The diffusion chamber model provides an enclosed environment within a host animal to study the osteogenic capacity of skeletally derived cell populations which resolves the problems of host versus donor bone tissue generation. Cells were recovered by collagenase (Clostridium histolyticum, type VII; 25 U/ml) and trypsin (0.05% trypsin and 0.2% ethylenediaminetetraacetic acid in PBS, pH 7.4) digestion. Diffusion chambers were implanted intraperitoneally into athymic MF1-nu/nu mice (Harlan UK Ltd; 20–24 g, 4–5 weeks old, Harlan UK Ltd). Control chambers contained human bone marrow samples on PLGA scaffolds alone. After 6–10 weeks the mice were killed, chambers removed, examined by X-ray analysis and fixed in 95% ethanol at 4°C. Samples were processed undecalcified and sectioned at 5 m and stained using toluidine blue and for type I collagen expression as well as mineralization by von Kossa.

Adenoviral Infection of Primary Human Bone Marrow Cells

The effectiveness of gene transfer into osteoprogenitor cells was examined using primary human bone marrow cells infected with AdLacZ at an MOI of 6.25, 12.5, 25, 50 and 100. After 72 h, the cells were stained for -galactosidase activity. Efficiency of transduction was observed to be 98% and independent of MOI. X-Gal staining intensity was not affected by increasing MOI. Mock infected cells showed negligible -galactosidase activity (Fig. 1).

Secretion of BMP-2 Following Adenoviral Infection

Having confirmed that osteoprogenitor cells could be infected with adenoviral constructs, human bone marrow cells were transduced with AdBMP-2 using an MOI of 20 (titre 4 109). The adenoviral vector AdBMP-2 has been shown previously by one of us to produce active BMP-2 protein by Western Blot and biochemical assay (34). To confirm secretion of active BMP-2 by the marrow stromal cells, media from transduced cultures were transferred to confluent promyoblastic C2C12 cells and alkaline phosphatase activity examined by histochemical and biochemical analysis. C2C12 cells exposed to the conditioned media from transduced cell lines showed strong staining for alkaline phosphatase while C2C12 cells exposed to conditioned media from mock infected cells showed no stain-

Vol. 292, No. 1, 2002 BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS

ing (Fig. 2) confirming the secretion of active BMP-2 by osteoprogenitor cells transduced with AdBMP-2.

The concentration of BMP-2 produced by the human bone marrow cells transduced with AdBMP-2 was determined from quantitative alkaline phosphatase activity assays in C2C12 cells using standard curves prepared with rhBMP2 at concentrations of 10–200 nM (data not shown). Calculation of the concentration of BMP-2 in the media of the infected cells from 4 patients ranging in age from 14–72 give values of 10, 1.8, 109, 165.4 nM although no correlation with patient age was observed. A range is to be expected given theinherentvariabilityin humanbonemarrowsamples.

Secretion of active BMP2 was followed over 8 days.

Flasks of identically prepared cells were transduced with AdBMP2 or mock. Virus was washed from the cells after 24 h and fresh media applied (Day 0). At 4, 6, and 8 days, the media was removed, filtered and fresh media provided. BMP2 activity was measured as for rhBMP2 (Fig. 3). Mock transduced cells displayed low alkaline phosphatase activity reflecting low constitutive expression. The peak of protein production was observed between days 4 and 6 when levels were higher over this 2 day period than from days 0–4. By day 8, production had decreased to control levels.

Growth of Adenoviral-Infected Osteoprogenitors on PLGA Scaffolds

Human bone marrow cells transduced with AdBMP2 were seeded onto biodegradable porous PLGA (75:25) (200 m) scaffolds in basal media (media supplemented with 10%fetal calf serum alone). Cell adhesion and extensive cell in-growth were observed following culture for 4 weeks as observed by scanning electron microscopy (Fig. 4A). Cell viability was examined using Cell Tracker green and EH-1 (Fig. 4B). The presence of intense fluorescent staining of the human osteopro- genitor cells as a consequence of label incorporation confirmed cell viability of transduced cells after 4 weeks on the porous scaffolds. The lack of cell necrosis was evidenced by the absence of ethidium homodimer-1 staining. These results indicate that adenovirally transduced cells are capable of adhering and proliferating on PLGA scaffold in vitro.

Differentiation of the cells to the osteoblast phenotype was confirmed by histochemical staining for type I collagen, Alcian blue/Sirius red and evidence of extensive mineralisation (Fig. 5). Immunostaining for Type I collagen was observed after extended in vitro culture for 4 weeks (Fig. 5A) and extracellular matrix formation was observed by Sirius red and alcian blue staining (Fig. 5B). No immunostaining was observed in type I collagen) controls (omission of primary antibodydata not shown). Further evidence of matrix formation and mineralisation of the porous scaffolds was confirmed by von Kossa staining (Figs. 5C and 5D). Mineralisation was not observed on control bone marrow cells cultured on tissue culture plastic alone (data not shown). The results are consistent with osteoblastic differentiation of the transduced bone marrow cells.

Differentiation of Adenovirally Transduced Human Bone Marrow Cells in Vivo

Adenovirally transduced human bone marrow cells expressing BMP2 cultured in basal media only, in the absence of ascorbate and dexamethasone, were injected into diffusion chambers containing PLGA scaffolds and implanted intraperitoneally in 5 nude mice. The chambers were removed after 4 and 7 weeks and examined for bone formation by X-ray analysis and histochemical analysis. X-ray analysis indicated the presence of bone tissue (Fig. 6A) after only 4 weeks in 3 of 5 mice. On histological analysis, extensive cell growth on the PLGA scaffolds was observed and morphological evidence for the formation of cartilage and bone tissue by adenovirally transduced human bone marrow cells expressing BMP-2 (Figs. 6B–6I). Immunolocalisation and histochemical analysis showed expression of alkaline phosphatase (Fig. 6C), type I collagen and extensive matrix formation (Figs. 6D–6H). Metachromatic staining with toluidine blue in combination with von Kossa staining indicated the presence of bone and cartilage respectively (Figs. 6B–C and 6F– H). New bone formation was confirmed using polarising light microscopy to demonstrate, by birefringence, new collagen formation (Figs. 6E and 6H). In diffusion chambers containing human bone marrow cells alone, in the absence of BMP-2, no evidence of bone and cartilage formation was observed (data not shown). These results indicate osteoblastic differentiation and bone formation in cells transduced with an adenoviral vector expressing human BMP2 in vivo.

(Parte 1 de 3)

Comentários