Induction of Human Osteoprogenitor Chemotaxis, Proliferation, Differentiation, and Bone Formation by Osteoblast Stimulating Factor-1/Pleiotrophin: Osteoconductive Biomimetic Scaffolds for Tissue Engineering†
University Orthopaedics, University of Southampton, Southampton, United Kingdom
The process of bone growth, regeneration, and remodeling is mediated, in part, by the immediate cell-matrix environment. Osteoblast stimulating factor-1 (OSF-1), more commonly known as pleiotrophin (PTN), is an extracellular matrix-associated protein, present in matrices, which act as targets for the deposition of new bone. However, the actions of PTN on human bone progenitor cells remain unknown. We examined the effects of PTN on primary human bone marrow stromal cells chemotaxis, differentiation, and colony formation (colony forming unit-fibroblastic) in vitro, and in particular, growth and differentiation on three-dimensional biodegradable porous scaffolds adsorbed with PTN in vivo. Primary human bone marrow cells were cultured on tissue culture plastic or poly(DL-lactic acid-co-glycolic acid) (PLGA; 75:25) porous scaffolds with or without addition of recombinant human PTN (1 pg-50 ng/ml) in basal and osteogenic conditions. Negligible cellular growth was observed on PLGA scaffold alone, generated using a super-critical fluid mixing method. PTN (50 μg/ml) was chemotactic to human osteoprogenitors and stimulated total colony formation, alkaline phosphatase-positive colony formation, and alkaline phosphatase-specific activity at concentrations as low as 10 pg/ml compared with control cultures. The effects were time-dependent. On three-dimensional scaffolds adsorbed with PTN, alkaline phosphatase activity, type I collagen formation, and synthesis of cbfa-1, osteocalcin, and PTN were observed by immunocytochemistry and PTN expression by in situ hybridization. PTN-adsorbed constructs showed morphologic evidence of new bone matrix and cartilage formation after subcutaneous implantation as well as within diffusion chambers implanted into athymic mice. In summary, PTN has the ability to promote adhesion, migration, expansion, and differentiation of human osteoprogenitor cells, and these results indicate the potential to develop protocols for de novo bone formation for skeletal repair that exploit cell-matrix interactions.
THE REQUIREMENT FOR new bone to replace or restore the function of traumatized or degenerated bone or for the replacement of lost mineralized tissue, as observed with aging in diseases such as osteoporosis, has led to the need for procedures to generate cartilage and bone for skeletal use.(1) To date, bone formation-stimulating regimens that achieve significant increases in bone density have yet to yield functional and mechanically competent bone. Bone tissue engineering of skeletal tissues for reconstruction and tissue and organ replacement may provide alternative solutions(2–6) based on the application of tissue engineering principles, namely, a rich source of osteoprogenitor cells, a scaffold conducive to cell attachment and maintenance of cell function, and appropriate selected osteoinductive growth factor(s).
A variety of osteoinductive factors, many present within bone matrix, have the potential to stimulate bone growth. A 136 amino acid bone growth factor rich in lysine and cysteine residues, osteoblast stimulating factor 1 (OSF-1), more commonly known as pleiotrophin (PTN) or heparin-binding growth-associated molecule, has been suggested to play a role in bone formation.(7) PTN is highly conserved across species, with more than 90% homology between chick, rat, bovine, and human sequences.(8–10) PTN shares 50% homology with the mouse midkine (MK) protein and MK-type chicken protein, retinoic acid-induced heparin-binding protein (RHIB).(11) After identification in osteoblasts, the same protein was named OSF-1 and shares the same sequence as the mitogenic and neurite outgrowth promoting protein PTN. Thus, HB-GAM, heparin-binding neurotrophic factor, heparin affinity regulatory peptide, and OSF-1 are synonyms of PTN. Expression postnatally of ptn is predominantly in bone and brain, from where it was first isolated. Transgenic mice overexpressing the human ptn gene have been shown to achieve a higher peak bone mass, which compensated for bone loss caused by estrogen deficiency(12) and to contain 10% higher bone mineral content (BMC).(13) HB-GAM is highly expressed in embryonic and early postnatal fiber pathways of the nervous system.(14) In developing and regenerating bone, PTN was identified as a matrix-bound chemotactic signaling molecule for migration of osteoblasts/osteoblast precursors.(7) Zhou et al.(15) have shown that PTN can enhance the attachment and subsequent differentiation of MC3T3-E1 cells. Recently we have shown, in agreement with Masuda et al.,(12) that PTN overexpressing transgenic mice had an increased BMC, and furthermore, that the epiphyseal function of these mice remained active for longer.(13)
Bone marrow contains multipotential stromal stem cells that can differentiate into fibroblastic, osteogenic, adipogenic, and reticular cells.(16–21) Delivery of these stem cells in appropriate matrices, which stimulate differentiation and bone conduction, has included the use of hydroxyapatite and calcium phosphate and a range of ceramic biomaterials.(22–25) Furthermore, these matrices have been widely used in cell transplantation studies as carriers for osteoinductive factors.(26–28) However, hydroxyapatite and calcium phosphate are not themselves osteoinductive, are resorbed relatively slowly, and have problems associated with biodegradability, inflammatory reactions, immunologic reactions, and in disease transmission when used as carriers for osteoinductive factors.(1) To circumvent these limitations, natural or synthetic materials and biodegradable composite scaffolds based on polylactic, polyglycolic acid, and polypropylene fumarate have been developed.(2,29–33) These materials have the advantage of Food and Drug Administration (FDA) approval, are biodegradable, can be chemically modified, and their degradation rates can be controlled. Moreover, the ability to modify the surface chemistry of biomaterials through the incorporation of signal recognition ligands or adsorption of appropriate signal molecules to mediate the molecular and cellular response to the material provides new approaches to promote osteoblast adhesion, proliferation, and differentiation. The potential of such an approach has recently been demonstrated by the generation of poly D,L-lactic acid (PLA) films coupled with adhesion peptide sequences using poly(L-lysine) for the attachment of endothelial cells(34) and modulation of human osteoprogenitor activity using PLA films and poly(DL-lactic acid-co-glycolic acid) PLGA scaffolds adsorbed with integrins and extracellular matrix proteins.(35) Recently Howdle et al.(36) have shown the potential of supercritical fluid technology in the development of porous biodegradable scaffolds such as PLGA for tissue engineering. Furthermore, the authors showed the use of super-critical carbon dioxide (scCO2) mixing technology to produce polymers of defined porosity and composition in the absence of porogens.
We have examined the efficacy of the osteotropic factor PTN to modulate human bone marrow stromal cell adhesion, chemotaxis, proliferation, differentiation, and colony formation (colony forming unit-fibroblastic [CFU-F]). Furthermore, we have examined the potential of porous biodegradable scaffolds generated using a new supercritical fluid process adsorbed with PTN to modulate bone formation in vivo using the subcutaneous and the diffusion chamber assays. The ability to enhance human osteoprogenitor activity and differentiation, through exploitation of cell-matrix interactions using three-dimensional biomimetic structures incorporating PTN, indicates the potential for the augmentation of de novo bone formation for skeletal repair.
MATERIALS AND METHODS
Tissue culture reagents were obtained from Gibco (Paisley, Scotland). Fetal calf serum (FCS) was from Meldrum Ltd. (Bourne End, UK). 1,25(OH)2D3 was a generous gift from Dr. Binderup (Leo Pharmaceutical Products Ltd., Ballerup, Denmark). Resin support was purchased from Novabiochem (Beeston, Nottingham, UK). Molecular biology reagents were purchased from Promega (Southampton, UK). Dexamethasone, alkaline phosphatase kits, and all other biochemical reagents were of analytical grade from Sigma Chemical Co. (Poole, Dorset, UK) unless otherwise stated.
Scaffold production using super-critical fluid processing
PLGA (75:25; Mw 22K) (Alkermes, Cambridge, MA, USA) porous (50–200 mm) scaffolds were used in all studies. The scaffolds were produced by a super-critical carbon dioxide method in which the polymer is plasticized at 35°C under a pressure of 1500 psi.(36, 37) On release of the pressure, pores are formed in the polymer by the escape of the carbon dioxide gas. The PLGA used in these studies will dissolve in vivo in approximately 3 to 4 months and was selected on the ability of incorporated glycolic acid to allow sufficiently rapid degradation of the copolymer. Scaffolds were characterized using scanning electron microscopy. Porous scaffolds were sterilized using 70% ethanol for 3 h and coated with minimal essential medium-alpha modification (αMEM) supplemented with 20% FCS, or 50 ng/ml PTN in αMEM (serum-free), or αMEM alone for 3 h.
Bone marrow samples (17 patients; 8 females and 9 males; 50–83 years of age; mean age, 66.7 years) were obtained from hematologically normal patients undergoing routine total hip replacement surgery. Only tissue that 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.(38) For cell growth on PLGA scaffolds, after trypsinization and resuspension in serum-free αMEM, 1 × 106 cells were added to individual wells of 24-well plates containing PLGA scaffolds. After 15 h, the media was removed, and cultures were maintained in basal media (αMEM supplemented with 10% FCS) or osteogenic media (αMEM supplemented with 10% FCS, 100 μm ascorbate-2-phosphate and 10 nM dexamethasone) for up to 6 weeks. Optimal cell culture points were 9 days for osteogenic CFU-F analysis, 12 days for basal CFU-F analysis, 21 days for enzyme activity, and 6 weeks for cell ingrowth and mineralization on three-dimensional scaffolds. Experiments were stopped at appropriate time intervals as detailed below.
Primary human bone marrow cells were cultured in 6-well tissue culture plates (5 × 105 cell/well) in the presence and absence of PTN (5–50 pg/ml) for 12 days in basal or 9 days in osteogenic media. Cells grown in the presence of dexamethasone/ascorbate-2-phosphate were stopped on day 9 to prevent merger of the colonies.(39) Cells were fixed using 95% ethanol before examination of CFU-F number using alkaline phosphatase histochemistry as previously described.(40) As shown in previous studies, individual distinct colonies were present at day 12, and these samples provided reproducible stable colony numbers.(40) Total and alkaline phosphatase-positive colonies were counted by eye using an Anderman colony counter (Anderman and Co. Ltd., Kingston-on-Thames, UK). Colonies were determined to be alkaline phosphatase positive if any cells showed observable staining by light microscopy. All counts were performed blind without prior knowledge of the sample characteristics, and the counts were repeated to confirm reproducibility of counts obtained. Mean values for each group were derived from 3–6 samples. For time course studies, 10 pg/ml PTN was added to cultures either at days 0–6 and days 6–9 (osteogenic media) or between days 0–7 and days 7–12 (basal media).
The chemotactic potential of PTN was examined on patterned surfaces generated using electron microscopy (EM) grids (mesh: 400 μm; Bar: 80 μm from Agar Scientific, Stansted, UK) on tissue culture plastic coated with and without PTN (50 μg/ml) and irradiated with UV light (254 nm) to denature the exposed PTN as previously described.(7) After removal of the grids, the chambers were washed thoroughly with sterile distilled water. Human bone marrow cells (passage 1) were plated at 1.25 × 105 cell/cm2 and cultured in serum-free medium for up to 24 h. Chamber slides coated with αMEM supplemented with 20% FCS or 10 mg/ml bovine serum albumin (BSA) were used as control substrates.
Cell culture on scaffolds
After trypsinization and resuspension in serum-free αMEM, 1 × 106 cells human bone marrow cells were added to individual wells of 24-well plates containing PTN adsorbed and nonadsorbed scaffolds in αMEM alone. Cells were grown in parallel on PLGA scaffolds in serum-containing medium as a positive control. After 15 h, the media was removed and cultures maintained in osteogenic media for up to 6 weeks. At appropriate time points, samples were harvested and stained for alkaline phosphatase, type I collagen, cbfa-1, and osteocalcin. Cell Tracker green was added to cultures at appropriate time points (4 and 6 weeks) to examine cell viability.
Human bone marrow cells were incubated with Cell Tracker green (5-chloromethylfluorescein diacetate, [CFMDA]) (Molecular Probes, Leiden, Netherlands) and ethidium homodimer-1 (EH-I) (Molecular Probes) for 45 minutes to label viable and necrotic/late apoptotic cells, respectively. The media was then replaced, and the cells were incubated for 1 h. Images from PLGA porous scaffolds were taken using an inverted microscope (DMIRB/E; Leica, Milton-Keynes, UK), 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 ×50 water immersion objective. Electron microscopy was undertaken using a Hitachi S-800 (Wokingham, UK).
Alkaline phosphatase-specific activity
Primary human bone marrow cells were cultured in 6-well tissue culture plates (5 × 105 cell/well) in the presence or absence of PTN (1 pg/ml-50 ng/ml) in basal or osteogenic medium for 21 days. Cell layers were washed with 1× phosphate buffered saline (PBS) and stored at −80°C until required. For each assay, the cell layer from each well was scraped into 0.5-ml 0.1% (vol/vol) triton X-100, freeze-thawed twice, and total cell lysis in triton X-100 confirmed by light microscopy. 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 per manufacturer's instructions (Molecular Probes). Alkaline phosphatase-specific activity was expressed as nanomoles of p-nitrophenol/minute/ng DNA.
Histochemistry and immunohistochemistry
Before 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-μm sections were prepared. Antibody controls (omission of primary antibody) were included in all immunohistochemical studies.
Alkaline phosphatase activity
Cultures were stained using a Sigma alkaline phosphatase kit (no. 85) according to the manufacturer's instructions.
Alcian blue/Sirius red Staining
Sections were stained using Weigert hematoxylin solutions before 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% AgNO3 under ultraviolet (UV) light for 20 minutes until black deposits were visible, and after air drying, slides were counterstained with Toluidine Blue.
Type I collagen and Cbfa-1
Reactivity to type I collagen antibody (LF 67, rabbit polyclonal against human collagen type I, α1 chain carboxy-telopeptide, Dr. Larry Fisher, National Institutes of Health, Bethesda, MD, USA) and Cbfa-1 (PEBP2αA, Santa Cruz Biotechnology, Santa Cruz, CA, USA) antibody was assessed in 4% paraformaldehyde-fixed sections or 95% ethanol-fixed sections (cbfa-1). Endogenous peroxidase activity was blocked using 3% H2O2 before incubation with LF 67 (1:300 in PBS) or Cbfa-1 (1:100 in PBS) for 3 h at 4°C. Samples were incubated with peroxidase-conjugated anti-rabbit immunoglobulin G (IgG) (1:30 in PBS) and peroxidase activity was detected using 3-amino-9-ethyl-carbazole (AEC) in acetate buffer containing H2O2. Samples were counterstained with Mayer's hematoxylin.
Reactivity to the osteocalcin specific monoclonal antibody OS35 (1:100; 30 minutes at 15°C; CisBio, Cedex, France) was assessed after fixation of the cells in 4%(w/v) paraformaldehyde in PBS (pH 7.4). Goat-anti-mouse-alkaline phosphatase (1:200) was used as secondary antibody and localization of enzymic reaction produced with fast red determined microscopically. Endogenous alkaline phosphatase activity was blocked by the addition of 1 mM levamisole. Samples were counterstained with Mayer's hematoxylin.
Reactivity to the PTN antibody (a gift from H. Rauvala, University of Helsinki, Helsinki, Finland) was assessed using a rabbit polyclonal antibody raised against the synthetic N-terminal peptide of HB-GAM isolated from rat perinatal rat brain.(41) Reactivity was assessed after fixation of the cells in 95% ethanol. Goat-anti rabbit-alkaline phosphatase (1:200) was used as secondary antibody, and localization of enzymic reaction produced with fast red was determined microscopically. Endogenous alkaline phosphatase activity was blocked by the addition of 1 mM levamisole. Samples were counterstained with Mayer's hematoxylin.
Mineralization within PLGA structures
After culture in osteogenic media for 4–6 weeks, cells were supplemented with osteogenic medium containing 5 mM inorganic phosphate for the final 48 h of the culture period, and mineralization was detected by von Kossa staining as described above.
In situ hybridization studies
Subcloning of pt:A 315 bp fragment of the human PTN coding region (EMBL accession number D90226, nucleotides 362–677) was amplified by polymerase chain reaction (PCR) with primers designed to create SacI and SalI restriction sites (forward: 5′TGAAGACCCAGAGCTCTAAGATCCCCTGCA; reverse: 5′CGGATCCTGT TTGTCGACGTCCTTTTTATG). Using standard protocols,(42) the fragment was subcloned into the SacI and SalI sites of pBluescript (Stratagene, Amsterdam, Netherlands). Sense and antisense probes were generated using T7 and T3 RNA polymerase and labeled with digoxigenin-uridine triphosphate (UTP) (Roche, Welwyn Garden City, UK) according to manufacturer's instructions.
Human bone marrow cells, cultured in osteogenic medium for 12 days, were fixed in 4% paraformaldehyde in diethyl pyrocarbonate (DEPC)-phosphate buffered saline. Endogenous alkaline phosphatase activity was quenched by incubating the cells with 0.2N HCl and cell membranes were permeablized using 0.2% Triton X100 in 4% PFA in DEPC-phosphate buffered saline. Cells were incubated with diluted riboprobes (1:9) overnight at 55°C and washed using 5× SSC and 50% formamide and 2× SSC. Unhybridized riboprobes were digested with 20 μg/ml RNAse for 30 minutes at 37°C. Anti-DIG-alkaline phosphatase Fab fragments from sheep were used for digoxigenin detection. After removal of the excess unbound Ab-conjugate, color development was facilitated by incubating with the 5-bromo-4-chloro-3-indolyl phosphate (BCIP)/nitroblue tetrazolium (NBT) liquid substrate system in the dark for 15 minutes. Samples were mounted in Aquamount and photographed on a Leica Leitz DM RBE.
In vivo studies
Primary human bone marrow cells were cultured in osteogenic media until confluent (day 21) before subcutaneous implantation or intraperitoneal implantation (2 × 106 cells/chamber) using diffusion chambers (130 μl capacity; Millipore, Watford, UK) in MF1-nu/nu mice (Harlan, Loughborough, UK; 20–24 g, 4–5 weeks old). 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/EDTA (0.05% trypsin and 0.2% EDTA in PBS, pH 7.4) digestion. Chambers were implanted intraperitoneally, and after 8–10 weeks the mice were killed, chambers removed and examined by X-ray analysis, and fixed in 95% ethanol at 4°C. Polymer samples were processed undecalcified and sectioned at 5 μm and stained with toluidine blue, type I collagen, osteocalcin, and mineralization by von Kossa. For subcutaneous implant studies, confluent primary human bone marrow cells were trypsinized and seeded (in plain media) onto PLGA scaffolds adsorbed with PTN(50 ng/ml) in αMEM or αMEM alone for 15 h. After 15 h, constructs were placed in osteogenic media for a further 3 days, before subcutaneous implantation into MF1-nu/nu mice. After 4–6 weeks, the mice were killed, and specimens were collected and fixed in 95% ethanol for histochemical analysis.
Values are expressed as the mean ± SD. Experiments were performed at least three times, and results of representative experiments are presented except where otherwise indicated. Statistical analysis was performed using one-way ANOVA with Tukey-Kramer Multiple Comparisons post-test using GraphPad Instant Software (GraphPad Software Inc., San Diego, CA, USA).
Effects of PTN on human osteoprogenitor chemotaxis
The chemotactic ability of PTN was examined on patterned surfaces generated using EM grids on tissue culture plastic coated with PTN and irradiated with UV light (Fig. 1). Human bone marrow stromal fibroblasts, which were randomly distributed after 30 minutes (Fig. 1A), were clearly recruited to areas of intact PTN, coated initially at a concentration of 50 μg/ml, after 24 h (Figs. 1B–1D) of culture. No migration of human bone marrow cells was observed on cell culture plastic in the absence of PTN.
Demonstration of PTN expression in human osteoprogenitor populations
After the demonstration of chemotaxis using PTN, we determined whether the cell preparations used expressed PTN using immunocytochemistry and in situ hybridization. As shown in Fig. 2, primary human bone marrow cells (day 12 cultures) expressed the PTN protein as assessed by immunocytochemistry (Fig. 2A). In addition, hybridization to the antisense PTN riboprobes demonstrated PTN expression throughout the human bone marrow cultures (day 12; Figs. 2B and 2C). Appropriate sense probes showed no specific staining in all samples (Fig. 2D).
Effects of PTN on CFU-F formation
PTN (10 pg/ml) addition resulted in a significant stimulation of total CFU-F colony formation as well as alkaline phosphatase-positive colony formation (Fig. 3) compared with control cultures. In osteogenic media, colony number and alkaline phosphatase-positive CFU-F colonies (Fig. 3B) were significantly stimulated above levels observed in the basal culture. Addition of PTN in basal conditions enhanced CFU-F number and alkaline phosphatase-positive CFU-F colony number (Fig. 3A). In osteogenic media supplemented with PTN, a stimulation of total CFU-F colony formation by around 30% and alkaline phosphatase colony formation by approximately 80% was observed at 10 pg/ml (Fig. 3B).
The effects of PTN on modulation of total colony formation in basal media and osteogenic media were dependent on time of addition, with increased colony formation during the late phase of culture (Figs. 4A and 4B). Thus, in basal cultures, addition of PTN from days 7–12 significantly enhanced total colony formation as well as alkaline phosphatase positive CFU-F formation (Fig. 4A), and similarly, addition of PTN in osteogenic conditions (between days 6–9) significantly enhanced total colony formation (Fig. 4B) and alkaline phosphatase CFU-F formation, suggesting an effect on late osteoprogenitors.
Effects of PTN on alkaline phosphatase activity
PTN produced a dose-dependent stimulation of alkaline phosphatase-specific activity with a significant increase in activity compared with cultures in basal media or in osteogenic media at concentrations as low as 10 pg/ml (Fig. 5A). In osteogenic media, the basal level of enzyme activity was increased approximately 10-fold compared with that in basal media, and addition of PTN resulted in further enhancement by 55% of alkaline phosphatase activity (Fig. 5B). As observed in the colony formation studies, high concentrations of PTN (>1 ng/ml) had no effect on alkaline phosphatase activity in basal or osteogenic conditions.
Growth on PTN-adsorbed scaffolds in vitro
After demonstration of the potential of PTN to stimulate chemotaxis, adhesion, and differentiation of human osteoprogenitor cells, we examined the potential of PTN adsorbed on PLGA scaffolds to induce adhesion, differentiation, and mineralization in three-dimensional scaffold templates. Cell viability after 4–6 weeks of static culture was examined using Cell Tracker green fluorescent probe and EH-1 expression to detect live or necrotic cells (Figs. 6A and 6B). Cell adhesion, spreading, and growth were observed on PLGA scaffolds adsorbed with PTN as examined by fluorescent and confocal microscopy (Figs. 6A and 6B), scanning electron microscopy (SEM) (Fig. 6C), and immunocytochemistry (Figs. 6E–6L). Increased cell attachment and proliferation was observed on PTN (50 ng/ml)-adsorbed scaffolds with few necrotic cells, as evidenced by the scarcity of EH-1 positive cells (Figs. 6A–6C). In contrast, poor cell attachment, migration, growth (Fig. 6D), and negligible matrix synthesis were observed on PLGA scaffolds alone (data not shown).
PTN stimulated differentiation and bone matrix synthesis by human bone marrow cells on porous PLGA scaffolds cultured in static culture over 4–6 weeks, as visualized by extensive expression of alkaline phosphatase, PTN, Cbfa-1, osteocalcin, and type I collagen in the matrix between the scaffold material (Figs. 6E–6L). Further evidence that PTN stimulated bone matrix synthesis was provided after visualization by Sirius red staining and Alcian blue for bone matrix and cartilaginous proteins, respectively (Figs. 6I and 6J). Collagen deposition and new matrix synthesis was confirmed by birefringence microscopy (Fig. 6J). von Kossa staining, after culture in osteogenic conditions, identified areas of mineral deposits indicative of expression of the mature osteogenic phenotype (Figs. 6K and 6L). Negligible mineralization was observed on PLGA scaffolds alone (data not shown).
In vivo studies
Subcutaneous implant mode
Primary human bone marrow cells were impregnated onto PLGA (75:25) porous scaffolds adsorbed with or without recombinant human PTN(50 ng/ml) as described in the Materials and Methods section. Cell/growth factor constructs were placed subcutaneously into athymic mice. After 4–6 weeks, subcutaneous implants were removed and analyzed for new bone formation (Fig. 7). As observed in the in vitro studies, negligible cell attachment was observed on PLGA scaffolds in the absence of PTN, and no evidence of cell growth differentiation, mineralization, and bone formation was observed in these constructs after subcutaneous implantation (data not shown). Human bone marrow cell differentiation and bone matrix synthesis on/within PTN absorbed PLGA scaffolds showed morphological evidence of new bone matrix synthesis as detected by Sirius red staining, expression of type 1 collagen, and expression of osteocalcin (Figs. 7A–7D). Evidence of organized new woven bone was confirmed by birefringence of collagen using polarized microscopy (Fig. 7B).
Diffusion chamber model
PTN-adsorbed constructs showed morphologic evidence of new bone matrix and cartilage formation by human bone marrow cells within diffusion chambers as evidenced by X-ray analysis (Fig. 7E) and staining for bone and cartilage matrix using Alcian blue and Sirius red (Figs. 7F–7K) after 10 weeks. Cartilage formation was observed within PLGA scaffolds, confirming penetration of human osteoprogenitors through the scaffold constructs (Figs. 7F–7H). Metachromatic staining was observed using Toluidine Blue as well as expression of type I collagen (data not shown). In addition, formation of organized new woven bone was confirmed by birefringence of collagen (Figs. 7H, 7J, and 7K) and extensive alkaline phosphatase expression and mineralization (Fig. 7L) in four of nine chambers. No bone formation was observed in human bone marrow/scaffold constructs alone (data not shown). Quantitation of mineral formation from X-ray analysis showed 4.89 ± 3.84% (range, 1.32–8.95%) of the diffusion chambers were filled with mineral, whereas no mineral was observed in human bone marrow/scaffolds constructs alone.
These studies show the successful use of human osteoprogenitor cells in combination with PTN-adsorbed PLGA scaffolds to provide an appropriate microenvironment to promote cell-matrix interaction facilitating human bone precursor cell attraction, adhesion, growth, and differentiation. These studies indicate PTN recruited human osteoblasts and osteoblast precursors and promoted cell adhesion, growth, and differentiation on the PLGA scaffolds, resulting in mineralized structures ex vivo. PTN was chemotactic to human osteoprogenitors and significantly stimulated total and alkaline phosphatase-positive colony formation, as well as significantly stimulating alkaline phosphatase-specific activity in basal conditions by 93% and by 55% in osteogenic conditions compared with controls. Expression of the bone markers cbfa-1, alkaline phosphatase, type I collagen, osteocalcin, and subsequent mineralization, after 3 weeks, confirmed differentiation of the osteoprogenitor cells along the osteogenic lineage. In vivo studies confirmed the ability of PTN to drive primed human osteoprogenitor cells, after culture in osteogenic conditions, to form bone after subcutaneous implantation or within the enclosed environment of the diffusion chamber. The diffusion chamber providing unequivocal demonstration of the new bone formation by the implanted cells opposed to host cells. The inability to show cartilage and bone formation on PLGA scaffolds in the absence of PTN indicate the potential of coupling osteogenic matrix proteins, such as PTN, to porous biodegradable three-dimensional scaffolds to provide an appropriate biomimetic structure for bone growth differentiation, development, and ultimately, the augmentation of de novo bone formation.
The capacity of bone for growth, regeneration, and remodeling has been attributed to a variety of chemotactic signals, such as the TGF-β superfamily, pivotal in the recruitment and differentiation of bone progenitors.(43) The demonstration using PTN of human osteoblast migration followed by attachment, as shown by the chemotactic assay, confirm and extend the findings of Imai et al.(7) in their studies using a variety of mammalian cell lines. In situ hybridization and immunolocalization studies confirmed primed human osteoprogenitors express PTN. Thus, PTN may act to feedback on the activity of these cells in a paracrine fashion. However, although expression was observed, it must be noted secretion of mature, functional, and processed protein was not determined. In the current studies, PTN was found to stimulate colony formation, including alkaline phosphatase-positive CFU-F number. Specifically the effects of PTN on CFU-F formation were observed in the later phase of cell culture, irrespective of whether preparations were grown in basal or osteogenic conditions. These effects of PTN on colony formation indicate an effect on late osteoprogenitor populations. In support of this, PTN promoted mineralization and bone formation in primed osteoprogenitor preparations as assessed by in vitro static culture, subcutaneous, and diffusion chamber in vivo assays. Gundle et al.(24) demonstrated bone formation in human bone marrow cell preparations derived from young donors (14–27 years of age) after extended culture, for over 6 weeks, in osteogenic conditions before implantation in combination with hydroxyapatite carrier. In the present study, we used primary human osteoprogenitors derived from aged individuals primed for 21 days in osteogenic conditions before implantation, showing the ability of PTN to drive primed mesenchymal populations along the osteogenic lineage.
However, the cultures of human primary bone marrow stromal fibroblastic cells used in this study include nonadherent hemopoietic cells that are present until their extensive removal, by vigorous washing, on day 6. While the weight of evidence would suggest otherwise, given the aforementioned as well as the in situ hybridization and PTN localization data, it cannot be excluded that the effects of PTN on stromal progenitors may be mediated, at least in small part, indirectly, by effects on hematopoietic cells and hematopoietic progenitors. To date, the effects of PTN on colony formation of hematopoietic progenitor cells, including colony-forming unit-granulocyte, erythroid, macrophage, and megakaryocyte cells, are unknown.
Bone formation involves the directed differentiation of mesenchymal cells into osteogenic cells, a process subject to regulation by a variety of hormones and factors.(17, 18) Bone morphogenetic proteins (BMPs), originally identified as proteins that could induce new cartilage and bone formation in non-bony tissues, are key osteoinductive factors pivotal in the recruitment, commitment, and differentiation of bone progenitors.(43) Native BMPs are 100- to 1000-fold more effective in inducing bone than individual recombinant BMPs.(44) However, despite the availability of recombinant BMPs, which BMP or which combination of BMPs will be most efficacious and cost-effective for bone induction and regeneration in clinical practice remains unclear, and the identification of an anabolic agent for skeletal repair remains, to date, elusive. With this background, the concentrations of PTN required to induce differentiation and mineralization were extremely low, approximately 100-fold lower than concentrations of rhBMPs required to achieve similar effects. In the current study, the effects of PTN on human osteoblast differentiation and mineralization on three-dimensional PLGA scaffolds, were observed using concentrations of PTN as low as 10 pg/ml. Given the relatively high levels of PTN, as much as 3.5 mg/kg of PTN, present in bone matrix,(15) the current findings add further support to the proposed role of PTN in maintaining osteoblastic activity of recruited populations whether at damaged sites or at areas of altered mechanical loading.
There have been reports for and against a potential mitogenic role for pleiotrophin/HB-GAM/OSF1.(45–47) Pleiotrophin was originally purified as a weak mitogen from bovine uterus by Deuel et al.,(48) and the same group went on to show that pleiotrophin was mitogenic for, and promoted, neurite outgrowth.(46) Chauhan et al.(47) showed that overexpression of bovine pleiotrophin complementary DNA (cDNA) in National Institutes of Health 3T3 cells resulted in enhanced cell number, anchorage-dependent growth, and tumor formation in nude mice, indicating a profound influence on cell growth. In contrast, a number of other groups failed to find a mitogenic effect on neurite outgrowth.(9,49,50) For HB-GAM, Hampton et al.(9) showed HB-GAM was not a mitogen for Balb/3T3 cells or human umbilical vein endothelial cells, although when presented as a substrate to chick embryo cerebral cortical derived neurons, neurite extension activity was observed. Studies using Swiss mouse 3T3 and National Institutes of Health 3T3 cells by Kuo et al.(49) showed little, if any, mitogenic activity. Furthermore, Rauvala and Peng(51) have suggested that HB-GAM may even be a proliferation arrest gene and that HB-GAM has a role in cell contact-dependent proliferation arrest. The current studies indicate PTN promotes adhesion, migration, growth, and differentiation of human mesenchymal cell populations; however, only a modest mitogenic effect was observed, as assessed by total colony formation. Further in vitro and in vivo studies using PTN and mesenchymal populations are required to determine if PTN is mitogenic in this system.
The current studies have used adsorbed growth factors, although the exact concentration of PTN adsorbed on the PLGA scaffolds is unknown. However, our data indicate sufficient PTN is present to induce attachment and migration into the scaffold compared with PLGA scaffolds lacking either PTN or serum in which negligible cell attachment was observed. In the current studies, we have used porous PLGA scaffolds generated using novel supercritical fluid technology developed by Howdle et al.(36) These scaffolds have a number advantages, including the ability to generate structures of defined porosity, degradation characteristic, and significantly, the potential for the incorporation of guest particles/factors. The use of super-critical fluid mixing technology to generate defined polymer scaffolds(36, 37) presents new approaches for cell and growth factor delivery. The absence of solvents and thermal processing in the generation of these scaffolds makes this an attractive approach to growth factor incorporation. In an extension of this work, Howdle et al. have demonstrated high protein (ribonuclease) loading into foamed PLA scaffolds that retained full activity on subsequent release from the PLA over 3 months.(36, 37) The current study, along with our recently published studies using transduced BMP-2 secreting osteoprogenitor cells, confirms the conductivity and potential of the porous PLGA scaffolds and indicates the promise such a tissue engineering strategy holds.(3) A number of tissue engineering strategies are in progress to incorporate growth factors, typically BMPs, into delivery systems and tissue engineered scaffolds for skeletal repair, including the use of collagen or hyaluronic acid with PLGA or PLA and encapsulation of BMP-2 in PLGA.(22,52,53) Clearly, the ability to generate scaffolds containing encapsulated PTN and other guest agents will provide a natural developmental step to explore, which may provide unique delivery vehicles for skeletal regeneration.
In conclusion, these studies indicate PTN has the ability to promote adhesion, migration, expansion, and differentiation of human osteoprogenitor cells. Furthermore, our results indicate that PTN acts specifically on late human osteoprogenitor populations rather than early osteoprogenitor subpopulations as evidenced by the time-course studies on PTN action. These findings indicate the capacity to generate osteoinductive surfaces within a porous template for human bone marrow cells that will allow the differentiation and ultimately mineralization of bone marrow cells. The successful generation of three-dimensional biomimetic structures incorporating PTN indicates the potential for de novo bone formation that exploit cell-matrix interactions.
The monoclonal antibody to type I collagen was a generous gift from Dr Larry Fisher (National Institutes of Health, Bethesda, MD, USA). We are also grateful to Prof Heikki Rauvala (Helsinki, Finland) for the donation of PTN antibody. We acknowledge the expert technical assistance of Mrs Anne Linge. We thank Drs Michael Watson and Martin Whitaker (University of Nottingham) for scaffold generation and Gill Campbell (Rowett Institute, Aberdeen) for in situ hybridization advice. We thank DePuy International Ltd. and the Wishbone Trust for PhD studentship funding to XY and RST, respectively. KMS is an EPSRC Advanced Fellow. This work was supported by grants from the Biotechnology and Biological Sciences Research Council (BBSRC), Engineering and Physical Sciences Research Council (EPSRC), Royal Society, and The Nuffield Foundation.