SEARCH

SEARCH BY CITATION

Keywords:

  • iPS cells;
  • Differentiation;
  • Osteoblasts;
  • Mesenchymal lineages

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
  9. References
  10. Supporting Information

Reprogramming somatic cells into an ESC-like state, or induced pluripotent stem (iPS) cells, has emerged as a promising new venue for customized cell therapies. In this study, we performed directed differentiation to assess the ability of murine iPS cells to differentiate into bone, cartilage, and fat in vitro and to maintain an osteoblast phenotype on a scaffold in vitro and in vivo. Embryoid bodies derived from murine iPS cells were cultured in differentiation medium for 8–12 weeks. Differentiation was assessed by lineage-specific morphology, gene expression, histological stain, and immunostaining to detect matrix deposition. After 12 weeks of expansion, iPS-derived osteoblasts were seeded in a gelfoam matrix followed by subcutaneous implantation in syngenic imprinting control region (ICR) mice. Implants were harvested at 12 weeks, histological analyses of cell and mineral and matrix content were performed. Differentiation of iPS cells into mesenchymal lineages of bone, cartilage, and fat was confirmed by morphology and expression of lineage-specific genes. Isolated implants of iPS cell-derived osteoblasts expressed matrices characteristic of bone, including osteocalcin and bone sialoprotein. Implants were also stained with alizarin red and von Kossa, demonstrating mineralization and persistence of an osteoblast phenotype. Recruitment of vasculature and microvascularization of the implant was also detected. Taken together, these data demonstrate functional osteoblast differentiation from iPS cells both in vitro and in vivo and reveal a source of cells, which merit evaluation for their potential uses in orthopedic medicine and understanding of molecular mechanisms of orthopedic disease. STEM CELLS 2011;29:206–216


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
  9. References
  10. Supporting Information

Mouse and human fibroblasts can be reprogrammed into an ESC-like state by transduction with a combination of transcription factors (Oct 3/4, Sox2, Klf4, and c-Myc or alternatively Oct3/4, Sox2, Nanog, and Lin28) [1–5]. The induced pluripotent stem (iPS) cells resulting from this manipulation function in a manner indistinguishable from mouse ESCs. For example, iPS cells are capable of differentiating into cell types characteristic of the three germ layers in vitro and in vivo, they express many of the markers associated with pluripotent cells, and they have an epigenetic status similar to that of ESCs [1–5]. Although the differentiation potential of iPS cell lines compared with that of ESC lines varies between similar to less potent depending on the lines tested and the directed differentiation analyses performed [6], the accessibility and therapeutic potential of patient's own iPS cells provides a powerful tool for cell-based regenerative medicine, including bone reconstructive surgery and other orthopedic procedures.

Although bone exhibits a remarkable regenerative capacity, aging, disease, or injury often results in significant bone loss, preventing natural replacement of this tissue in the organism. Although bone autograft provides the best clinical outcome for bone replacement therapy [7], it is associated with severe pain at the site of removal and high morbidity [8]. In turn, allogeneic transplants carry not only the risk of immunological rejection but also the transmission of disease from donor to recipient [9]. Autologous mesenchymal stem cells (MSCs) derived from bone marrow offer a promising source of cells for musculoskeletal regeneration because of their potential to differentiate into bone, cartilage, and fat, as well as their potent paracrine anti-inflammatory properties [10, 11]. However, the use of MSCs in orthopedic reconstructive therapy may be restricted by their proliferative potential, which significantly decreases with age [12]. Because of their self-renewal capacity, patient-specific iPS cells would address this limitation by providing an unlimited source of MSCs. Recently, Marion et al. [13] demonstrated that iPS cells generated from both young and aged individuals have elongated telomeres, and their telomers acquire the characteristics of ESCs. Agarwal et al. [14] showed the importance of telomerase in the maintenance of iPS self-renewal. The MSC potential of iPS was described by Lian et al. [11]. The iPS cell-derived mesenchymal cells were able to attenuate the injury associated with hindlimb ischemia in a rodent model and to contribute to tissue regeneration to a greater degree than bone marrow-derived MSCs (BM-MSCs); however, anti-inflammatory properties of these iPS cell-derived MSCs remained undefined in this study. Additional studies have also shown that iPS cells can be differentiated into skeletal muscle, adipocytes, and vascular lineages in vitro [15–19]. Thus, similar to BM-MSCs, iPS cell-derived MSCs may be used as an autologous graft in situations of fracture nonunion, osteoarthritis, and intraoral defects to repair bone and cartilage and potentially reduce inflammation. Significant progress has recently been made with regard to iPS cells in musculoskeletal regenerative medicine; however, the in vivo regenerative potential of these iPS cell-derived lineages has not been addressed.

Several obstacles still have to be overcome before iPS cells can be studied as a potential therapy for orthopedic medicine. Importantly, the persistence of differentiated phenotypes in vivo must be demonstrated. Given the potential of pluripotent stem cells to be expanded and differentiate into multiple lineages similar to ESCs, we hypothesized that clonal iPS cells capable of generating mesenchymal tissues could differentiate to a functional osteoblast lineage, which when cultured on a scaffold would give rise to ectopic mineralized tissue nodules in vivo. Toward this hypothesis, we performed directed differentiation of iPS cells to mesenchymal cells and subsequently to the osteoblast lineage in vitro. These osteoblasts were seeded on gelatin scaffolds and studied in vitro and in vivo using syngenic mice. The osteogenic scaffolds were evaluated for stability of the osteoblast phenotype, proliferation, and osteogenic matrix production.

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
  9. References
  10. Supporting Information

Generation of iPS Cells

Mouse iPS cells were generated by transducing freshly isolated ICR mouse dermal fibroblasts (Harlan Laboratories, Indianapolis, IN) with four retroviruses encoding murine Oct3/4, Klf4, Sox-2, and c-Myc as originally reported by Yamanaka [2]. Mouse Oct3/4, Klf4, and Sox2 pMxs retroviral vectors were obtained from Addgene (Cambridge, MA), whereas the c-Myc construct was generated by cloning mouse c-Myc cDNA into a MSCV-ires-GFP vector (pMIG, Addgene). Viruses were prepared by transient transfection of Phoenix-E cells together with the pCL-Eco packaging vector. Two days post-transduction, the transduced fibroblasts were plated onto a mouse embryonic fibroblast feeder cells treated with mitomycin C (Sigma-Aldrich, St. Louis, MO) and cultured in ESC medium containing Knockout Dulbecco's modified Eagles medium (KO-DMEM) supplemented with 2 mM GlutaMAX-I, 0.1 mM minimum essential medium non essential amino acids solution, 0.1 mM β-mercaptoethanol (Invitrogen, Carlsbad, CA), 15% ESC tested fetal bovine serum (FBS; Tissue Culture Biologicals, Seal Beach, CA), and 1,000 U/ml LIF (ESGRO, Millipore, Billerica, MA) and in the presence of 0.5 mM valproic acid [20] (Chemicals purchased from Sigma-Aldrich unless otherwise stated) until the ESC-like looking colonies appeared in about 2 weeks.

Teratoma Formation

Mouse iPS cells (5 × 105 cells) were injected subcutaneously into the flank of 6-week-old Foxn1nu nude mice (The Jackson Laboratory, Bar Harbor, ME). The animals were monitored for 1 month for tumor formation.

Differentiation of iPS Cells into Mesenchymal Tissues

Based on a modified protocol for embryoid body (EB) formation [21–23], mouse iPS cells were detached from a feeder layer and formed on a Petri dish in ESC medium without LIF. After 2 days, EBs were cultured in ESC medium in the presence of 10−7 M all-trans retinoic acid (ATRA) for 2 days. After 2 days, the EBs were collected, cells dissociated via brief trypsinization, and analyzed by flow cytometry or replated for 1 day on gelatin-coated plates to promote commitment toward the mesenchymal osteoblast, chondrocyte, and adipogenic lineages (3 days total) [21–24]. This medium was replaced with lineage differentiation medium for 3–4 weeks. In the case of adipocyte and osteoblast differentiation, EB outgrowths were used in for differentiation studies, whereas the remaining three-dimensional structures were washed from the culture when the differentiation medium was applied. To promote adipocyte differentiation, cells were cultured in DMEM containing 10% fetal bovine serum supplemented with 500 μM isobutylmethylxanthine, 1 μM dexamethasone, and 100 U/ml Humulin (human recombinant insulin). Thereafter, the cells were refed every 48 hours with this media supplemented with 1 μM troglitazone for 4 weeks when distinct lipid droplets were observed by microscopy. Chondrocyte differentiation was performed by trypsinizing EB to a single cell suspension, diluting cells to a final concentration of 2.5 × 105 cells/ml and forming micromass pellets by centrifugation. Micromass was cultured as nonadherent spheres in 15-ml conical tubes for 4 weeks. Media was changed every 2 days. Differentiation media consisted of high glucose (4.5 g/l), DMEM as base medium, and supplemented with 10% FBS, 1 mM dexamethasone, 17 mM ascorbate-2-phosphate, 35 mM L-proline,1 mM sodium pyruvate, 1× insulin-transferrin-selenium, and 10 ng/ml TGF-β3 (R&D Systems, Minneapolis, MN) [25]. iPS cell-derived EBs were differentiated to the osteoblast lineage by culturing cells to high density (90% confluence) followed by incubating with differentiation medium consisting of DMEM low glucose (Invitrogen), 10% FBS, 5% Pen/Strep, 1 mM dexamethasone, 0.1 M ascorbic acid, 1 M glycerol 2 phosphate. Media was changed every 2 days and 14–17 days later and the differentiation documented by Alizarian red and von Kossa staining. Differentiation was analyzed after 1, 4, and 8 weeks.

Lineage Phenotyping

Histochemical stains were performed as follows: Hematoxylin stain (Vector Labs, Burlingame, CA) was applied directly on sections for 30 seconds followed by a 10-second rinse in water. Sections were then counterstained with eosin Y (Thermo Scientific) as per standard techniques. In Alizarian red staining, cells were first fixed in 10% formalin for 5 minutes, washed twice with water, incubated with alizarian red solution for 10 minutes, then rinsed several times in water, and allowed to dry. The von Kossa staining cells were rinsed in water and stained with silver nitrate for 30 minutes under a UV light. Cells were then rinsed in water and allowed to dry. In Alcian blue staining, chondrocyte micromass was fixed in 4% paraformaldehyde overnight. Micromass was then sectioned and slides were incubated with 1% Alcian blue solution for 30 minutes. Slides were then washed with 0.1 N HCl and allowed to dry.

Immunostaining of iPS cells was performed, where cells were fixed in cold methanol (−20°C) for 5 minutes and saturated with phosphate-buffered saline containing 10% bovine serum albumin (Sigma-Aldrich). The cells were then incubated with anti-Nanog (Chemicon, Temecula, CA) antibody overnight at 4°C and secondary Alexa-conjugated fluorochromes 594 anti-rabbit antibody (Invitrogen) for 1 hour at room temperature.

Tissues and micromass chondrocyte cultures were fixed in 4% paraformaldehyde, embedded in paraffin, and sectioned. For immunofluorescence analysis, antigens were retrieved by boiling sections in 10 mM sodium citrate for 10 minutes. The prepared sections were then incubated with primary antibody overnight at 4°C and secondary antibody for 1 hour at room temperature and mounted with hard set mounting medium (Vector Laboratories, Burlingame, CA). The primary antibodies used were against keratin 14, Krt14 [26], cytokeratin Endo-A (TROMA-1), adult skeletal muscle myosin heavy chain (A4.1025; Developmental Studies Hybridoma Bank, Iowa City, IA), and aggrecan (Santa Cruz Biotechnology, Santa Cruz, CA).

Implants were isolated as described, embedded in optimal cutting temperature compound (OCT) and snap frozen in liquid nitrogen. Frozen sections were cut at 25 μM followed by fixation in 4% paraformaldehyde. Sections were incubated with primary antibodies (bone sialoprotein [BSP] #WVID1(9C5) and osteocalcin clone M-15; Santa Cruz Biotechnology, Santa Cruz, CA) for 12 hours, followed by a 3-hour incubation with fluorescently labeled secondary antibody. The secondary antibody conjugates used were Alexa-conjugated fluorochromes 594 or 488 anti-guinea pig, anti-rat, and anti-mouse (Invitrogen). Coverslips were mounted using Vectashied and nuclei identified with diamidino-2-phenylindole (DAPI) (Vector Labs, Burlingame, CA).

Flow cytometry was performed to detect mesenchymal markers CD 90, 73, 105, 106, and 133 and lack of the hematopoietic markers CD45 and c-kit as previously described [27] using a Beckman Coulter Cyan ADP. Laser lines and emission filters included 488 to detect fluorescein isothiocyanate using the 530/40 filter, phycoerythrin using 575/25 filter and 635 nm laser line to detect allophycocyanin using the 665/20 filter. Gating strategy included forward scatter/side scatter (FSC/SSC), doublet discrimination, live dead using DAPI, and single color analysis to detect each marker. Negative gates were set to an unstained control and compensation performed using beads (BD Pharmingen, Franklin Lakes, NJ).

Gene Expression Analysis

Total RNA was extracted from fibroblasts, mouse CJ7 ESCs, iPS cells, EB, and differentiated cells using RNeasy kit (Qiagen) per manufacturers protocol. Reverse transcription of RNA to cDNA followed the manufactures protocol (Invitrogen). Semiquantitative polymerase chain reaction (PCR) reactions were conducted using Taq Gold (Applied Biosystems, Foster City, CA) with 1 μl of cDNA for 30 cycles. Primers used to detect the expression of endogenous nanog, oct3/4, and sox2 were previously published [2]. Glyceraldehyde-3-phosphate dehydrogenase (Gapdh) was detected using the following oligos: 5′-CGTCCCGTAGACAAAATGGT-3′ and 5′-TCTCCATGGTGGTGAAGACA. For quantitative (q) PCR detection of Col1a1 (mm00801666_g1), spp1 (mm01204014_ m1), runx2 (mm00501584), sox9 (mn00448840_m1), acan (mn00545807_m1), flt-1 (mm01210866_ml), and pecam/CD31 (mm01242584_m1; Applied Biosystems), 50 ng of cDNA was analyzed in triplicate under using the Light Cycler 480 System (Roche Diagnostics, Basel, Switzerland). Levels were normalized to Gapdh abundance (Applied Biosystems).

Three-Dimensional Culture and In Vivo Analyses of Osteoblast Phenotype

Seeding of the scaffold was performed by cutting Gelfoam surgical sponges (Pfizer Pharmaceuticals) into one centimeter squares using sterile technique. The sponges were impregnated with bone differentiation medium. Differentiated osteoblasts at 8 weeks were trypsinized to obtain a single cell suspension. A total of 8 × 106 cells were suspended in differentiation medium and sponges added. Cells were allowed to adhere for 12 hours under routine culture conditions. Sponges were then placed in a conical tube containing fresh bone differentiation medium. Medium was replaced every other day until the time of harvest.

For subcutaneous implantation of the scaffold, 12-week-old ICR mice were purchased from Harlan Laboratories and housed in the University of Colorado Denver central vivarium under pathogen-free conditions. All procedures were performed according to the Animal Care and Use Committee guidelines at the University of Colorado Denver. Mice were anesthetized with inhaled isoflurane and hair shaved off the back of the recipient mouse to minimize infection. Using aseptic technique a longitudinal 0.5-cm incision was made in the back of the mouse, the skin separated from the underlying muscle with forceps, and the Gelfoam per cell implant placed in this subcutaneous pouch. The skin was closed with 3-0 nylon suture and tissue glue applied over the suture to seal. One such pocket was made in each mouse (using 15 mice). Animals were singly housed for 7 days following implant then housed in groups of 2–3 for the remaining 12 weeks of the experiment. Undifferentiated iPS cells produce teratomas, therefore, we did not include a control group of undifferentiated cells. The groups were performed with gelfoam controls (minus cells) or gelfoam seeded with osteoblasts 24 hours prior.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
  9. References
  10. Supporting Information

Generation of iPS Cells

We generated iPS cells by transducing primary mouse fibroblasts with retroviral vectors encoding four reprogramming factors (Oct3/4, Sox2, Klf4, and c-Myc) [2]. Similar to the observation originally made by Yamanaka [2], our iPS cells exhibited an ESC-like morphology (Fig. 1A) and reactivated expression of endogenous Oct3/4, Sox2, and Nanog, genes normally expressed in mouse ESCs and silenced in somatic cells, as determined by reverse transcription polymerase chain reaction (RT-PCR; Fig. 1B). The reactivation of Nanog expression in our iPS cell clones was further confirmed by immunofluorescence analysis (Fig. 1C). The generated iPS cells formed teratomas following subcutaneous injection into nude mice. Tissues from all three germ layers were present in these tumors as detected by immunofluorescence analysis (Fig. 1D), thus confirming the pluripotency of our iPS cell lines. We used Krt14 as a marker for ectoderm, heavy chain myosin from skeletal muscles for mesoderm and cytokeratin EndoA for endoderm.

thumbnail image

Figure 1. Generation of mouse iPS cells. (A): Skin fibroblasts from newborn ICR mice were transduced with retroviruses expressing reprogramming factors and cultured under ESC conditions. After 2 weeks in culture, colonies exhibiting an ESC morphology under phase-contrast microscopy emerged. (B): Reverse transcription polymerase chain reaction analysis with primers specific to endogenous mouse Nanog, Oct3/4, and Sox2, as well as GAPDH as a control, was performed with total RNA extracted from lane 1 (primary ICR fibroblasts), lane 2 (ICR fibroblasts transduced with four retroviral vectors and cultured for 5 days), lane 3 (mouse ESCs), and lane 4 (mouse iPS cells). (C): iPS cell colonies were positive for Nanog, as determined by immunofluorescence analysis (red, left panel). The feeder cells used to maintain iPS cells served as a negative control for Nanog immunofluorescence (DAPl, right panel). (D): Teratomas were formed when iPS cells were injected subcutaneously into nude mice. H&E staining are shown in top panels and consecutive sections labeled with various antibodies representing three germ layers, that is, ectoderm (Krt14), mesoderm (MyHC), endoderm (cytokeratin Endo-A) are bottom panels. All images were taken with ×20 objectives. Abbreviations: DAPI, diamidino-2-phenylindole; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; ICR, imprinting control region mouse; iPS, induced pluripotent stem cell; MyHC, myosin heavy chain from skeletal muscles.

Download figure to PowerPoint

Differentiation of iPS Cells into Mesenchymal Cell Phenotypes

To determine the potential of a clonal iPS cell line to differentiate into the mesenchymal lineages of bone, cartilage, and fat, we employed differentiation protocols previously formulated for ESCs. EB differentiated from iPS cells were treated with retinoic acid in suspension culture to induce cell commitment toward mesoderm, plated on gelatin followed by culture in lineage-specific differentiation medium according to Kawaguchi [21, 22] with slight modifications. Fat differentiation was evident after 4 weeks by visualization of lipid droplet accumulation (Fig. 2A). Cartilage differentiation was performed by culturing EBs as nonadherent cell spheres or micromass in chondrogenic medium. Analysis of mRNA extracted from iPS cells after 5 weeks of differentiation in chondrogenic medium showed expression of the chondrogenic differentiation factor, Sox-9, and the chondrocyte matrix protein Acan (Aggrecan; Fig. 2B). Further investigation demonstrated immunohistological staining for aggrecan (Fig. 2C, 2D). Differentiation into a chondrocyte phenotype was demonstrated by staining with H&E and Alcian blue, a cationic stain that highlights an extracellular matrix rich in polyanionic glycosaminoglycans (Fig. 2E, 2F). We also performed flow cytometric analysis on cells obtained from 200 dissociated EB following the 3-day ATRA treatment to analyze cell surface expression of markers indicative of a mesenchymal cell. We found that the cells express varying levels of the mesenchymal markers CD 90, 73, 105, 106, and 133 and lack the hematopoietic markers CD45 and ckit (Fig. 2G). Lung MSCs were used as a positive control (not shown).

thumbnail image

Figure 2. Differentiation of the mesenchymal adipocyte and chondrocyte lineages from iPS cells. Murine embryoid body derived from iPS cells were treated with all-trans retinoic acid for 3 days to induce mesoderm followed by the treatment with adiopgenic or chondrogenic differentiation medium for an additional 4 weeks. Lipid droplets were visualized in adipocytes using phase-contrast microscopy ×10 objective (A). Chondrogenesis was performed in micromass cultures and documented by quantitative reverse transcription polymerase chain reaction detection of Sox9 and Aggrecan (B) and by Aggrecan immunostaining ×4 objective (D) with DAPI nuclear stain (C). The phenotype of chondrocytes was documented in paraffin sections by H&E (E) and Alcian blue histochemical stains and photographs taken with the ×20 objective (F). (G). Analysis of mesenchymal marker expression by dissociated iPS-derived embryoid bodies following 3-day treatment with ATRA. Abbreviations: FSC, forward scatter; iPS, induced pluripotent stem cell; SSC, side scatter.

Download figure to PowerPoint

The differentiation of iPS cells into osteoblasts was performed by culturing mesodermal-induced iPS in osteogenic medium for up to 8 weeks. Differentiation of iPS cells to an osteoblast phenotype from day 0 (following retinoic acid treatment) to 8 weeks was shown by histochemical staining using alizarin red to indicate sites undergoing calcium deposition and mineralization (Fig. 3A–3D). Additionally, von Kossa silver stain and alkaline phosphatase stain were performed at 8 weeks on the differentiated cells to demonstrate phosphate deposition and alkaline phosphatase activity (Fig. 3E). Both phosphate and alkaline phosphatase activity were present in the culture. 3T3E1 preosteocyte cells were cultured in basal media or osteogenic differentiation media and used to demonstrate the specificity of histochemical staining (supporting information Fig. 1). Analysis of mRNA extracted from iPS cells after 0, 1, 4, and 8 weeks of differentiation in osteogenic medium showed induction of the osteoblast markers, the transcription factor runx2 (cbfa1), extracellular matrix and structural proteins col1a1 (Collagen type I pro alpha I), spp1 (bone sialoprotein or osteopontin [OPN]), and flt-1 (vascular endothelial growth factor receptor 1 [28]; Fig. 3F). Controls for qRT-PCR primers included a bone marrow as a negative and 3T3 E1 cells cultured using osteogenic differentiation medium (supporting information Fig. 2).

thumbnail image

Figure 3. Functional osteoblasts can be induced from iPS cells in vitro by the presence of dexamethasone and ascorbic acid. (A–D): Murine embryoid body derived from iPS cells was treated with all-trans retinoic acid for 3 days followed by osteoblast differentiation medium. Commitment to the osteoblast lineage was documented by 4 and 8 weeks, as demonstrated by alizarin red stain, identified the increase in calcific deposit by osteoblasts (red), and nuclei were visualized with a hematoxylin counterstain (blue). (E): Von Kossa stain (black) and alkaline phosphatase colorimetric detection (red) were performed to characterize a differentiated osteoblast lineage. All images were taken with a ×4 objective. (F): A temporal increase in expression of runx2, col1a1, and spp1 genes was confirmed by quantitative reverse transcription polymerase chain reaction analysis. Abbreviations: BM, bone marrow; D5, day 5; DIF, differentiation medium; iPS, induced pluripotent stem cells.

Download figure to PowerPoint

These data suggest the possibility of using approaches developed for ESC differentiation to create and expand mesenchymal lineages such as osteoblasts and chondrocytes from our iPS cell lines in vitro.

iPS Cell-Derived Osteoblasts Maintain Their Phenotype in a Scaffold In Vitro and In Vivo

The next step in the evaluation of iPS cell differentiation to a bone-like phenotype was the ability to maintain a differentiated osteoblast phenotype when passaged and seeded into a three-dimensional scaffold both in vitro and in vivo. After 4 weeks, the osteoblast cultures had reached confluence and expressed high levels of runx2, col1a1, and spp1 (Fig. 3). At that time, gelfoam carriers, a biodegradable gelatin matrix, were seeded with 8 × 106 iPS cell-derived osteoblasts after 8 weeks of differentiation culture and maintained in differentiation media. Implanted sponges were analyzed in vitro after 48 hours and 2 weeks of culture to confirm the presence of cells using hematoxylin staining (Fig. 4 A, 4B). After 2 weeks, sponges stained with hematoxylin and alizarin red with a hematoxylin counterstain demonstrated an increase in cell number and secreted matrix (Fig. 4C–4E). Sponges seeded with iPS cell-derived osteoblasts maintained the expression of osteoblasts markers runx2, col1a1, and spp1 after 48 hours with higher levels of col1a1 and spp1 2 weeks of differentiation in osteogenic medium (Fig. 4F).

thumbnail image

Figure 4. Induced pluripotent stem (iPS) cell-derived osteoblasts maintain their phenotype and deposit bone extracellular matrix on a scaffold in vitro. One-centimeter cubes of gelfoam sponge were seeded with iPS cell-derived osteoblasts following 8 weeks of differentiation culture. (A, B): After 48 hours, the gelfoam scaffolds were stained with hematoxylin to confirm the presence of cells. Images were taken with ×10 and ×20 objective. (C, D): After 2 weeks, the scaffolds were stained with alizarin red and hematoxylin counterstain, which indicated an increase in the cells present within the scaffold, as well as matrix deposition by the osteoblasts. Images were taken with a ×10 and ×20 objective. (E): Nodules positive for alizarin red were detectable on gelfoam cubes in vitro after seeding with iPS cell-derived osteoblasts cultured in differentiation medium for 2 weeks. A representative phase-contrast micrograph is presented. Scale bar = 5 mm. (F): Maintenance or increased expression of runx2, col1a1, and spp1 genes was documented using quantitative reverse transcription polymerase chain reaction analysis. Abbreviation: BM, bone marrow.

Download figure to PowerPoint

Following passage osteoblast cells continue to express matrix and react to alizarin red, however, take another week to increase the gene expression of more differentiated osteoblasts as demonstrated by analysis of seeding on sponges presented in Figure 4. Therefore, we waited until 8 weeks when the osteoblasts had expanded in numbers before seeding gelfoam for in vivo studies.

For in vivo analysis, 24 hours after seeding 8 × 106 iPS cell-derived osteoblasts on gelfoam cubes, the sponges were subcutaneously implanted on the dorsal surface of immunocompetent ICR mice. After 12 weeks, no tumor formation was detected and the implants were extracted and analyzed for phenotypic markers of osteogenic cells and mineralization. Implants containing osteoblasts were isolated from 8 of 15 mice. Grossly, the implants were identifiable between the cutaneous and peritoneal muscular layers and appeared to recruit vasculature (Fig. 5A). Paraffin-sectioned implants (6 μm thickness) stained with H&E demonstrate vascularization of the implant discernable by the presence of red blood cells in the lumens of microvessels (Fig. 5B, 5C). Additional implants were flash frozen in OCT and frozen sectioned to 25 μm thickness to maintain tissue integrity and antigenicity for immunofluorescence. To determine the presumptive origin of the microcirculation, we performed qRT-PCR analysis of to detect gene expression indicative of vascular endothelial differentiation. Using the marker pecam1/CD31, which is expressed by both progenitors and differentiated endothelium, we found no evidence of vasculature in the osteoblast cultures or in vitro sponges (Fig. 5D).

thumbnail image

Figure 5. Subcutaneous implantation of osteoblast seeded scaffolds recruited vasculature in vivo. After 12 weeks, implants were identifiable (A) as surrounded by vasculature. Paraffin sections stained with H&E demonstrated the subcutaneous localization of the implant (B) and the vascular supply within the implant, as evidenced by the presence of red blood cells in the lumens of microvessels ([C], arrows) using phase-contrast microscopy. (D) Expression of pecam-CD31 was analyzed to determine whether vascular cells were present in the differentiating iPS osteogenic cultures. Scale bar = 50 μM. Abbreviations: BM, bone marrow; D5, day 5; iPS, induced pluripotent stem cells.

Download figure to PowerPoint

Alizarin red and von Kossa stains localized osteoblasts secreted matrix and calcium in the implants (Fig. 6A–6C). Fluorescent immunostaining demonstrated the presence of osteoblast markers osteocalcin and BSP in serial sections (Fig. 6D–6G, supporting information Fig. 3). Thus, we demonstrate that osteoblasts differentiated from iPS cells maintained their lineage-specific phenotype after passage onto scaffold both in vitro and in vivo. The phenotype of these osteoblast-derived iPS cells appears stable at both 4 and 8 weeks following directed differentiation. In vivo formation of mineralized nodules using a mouse model suggests the persistence of a differentiated phenotype and highlights the potential for these cells to be used in graft development and cell-based therapy for musculoskeletal repair.

thumbnail image

Figure 6. Osteoblast-seeded scaffolds maintain their phenotype and deposit bone extracellular matrix in vivo. After 12 weeks, gelfoam scaffolds were embedded in OCT, frozen, and 25-mM sections were analyzed. Histochemical alizarin red (A, B) and von Kossa (C) staining confirmed the presence of matrix and calcium within the nodules and hematoxylin the presence of osteoblasts. (D–G): Consecutive sections were immunoreactive to osteocalcin and BSP. DAPI was used to label nuclei. Scale bar = 2.5 mm. Abbreviations: BSP, bone sialoprotein; DAPI, diamidino-2-phenylindole; OCT, optimal cutting temperature compound.

Download figure to PowerPoint

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
  9. References
  10. Supporting Information

Cell-based therapy for repair of bone and cartilage offers a promising treatment for both degenerative and genetic musculoskeletal diseases. However, engineered grafts are far from being a standard of treatment for bone and cartilage repair. BM-MSC, ESCs, and more recently, iPS cells have been demonstrated as having the potential to differentiate to bone and cartilage, as well as other mesenchymal lineages. Each cell type offers its own challenges. Although BM-MSC may function as an autologous graft, also suited for gene therapy, and demonstrate anti-inflammatory properties, when obtained from aged donors they may demonstrate a limited ability to proliferate [11, 29, 30]. Although ESCs are pluripotent and exhibit high differentiation and proliferation potential, their origin is the topic of ethical debate as they must be derived from fertilized embryos. In addition, the repair of tissues with embryonic stem (ES)-derived cells may result in immune rejection, as these cells are allogeneic. Data on immunological properties of human and murine ESCs and their differentiated derivatives are still controversial, ranging from those claiming unique immune-privileged properties for ESCs [31, 32] to those which refute these conclusions [33, 34]. iPS cells are the most recent addition to this cast and they offer benefits of the aforementioned cell types yet they are the least well characterized. Nevertheless, it has been shown that iPS cells function in a manner similar to ESCs, that is, they are capable of forming multiple cell types in vitro and in vivo, they express most (but not all) of the markers associated with pluripotent cells, they have an epigenetic and telomere status similar to that of ESCs, and they can be used to make fertile mice [13, 35]. The immunological properties of iPS are not well characterized.

Here, we have assessed the ability of murine iPS cells to be directionally differentiated to multipotent mesenchymal cells and subsequently the osteoblast lineage in vitro. The multipotent mesenchymal cell stage was demonstrated by differentiation of iPS cell-derived mesenchymal cells into three characteristic mesenchymal lineages including osteoblast, adipocyte, and chondrocytes. Lian et al. recently demonstrated the ability of human iPS cells to give rise to mesenchymal cells with a cell surface phenotype and differentiation potential similar to those of BM-MSC [11] using clinically compliant culture techniques optimized for ESCs [36] adapted from ES differentiation to MSC using OP9 feeder cells [37]. While we also employed a differentiation strategy optimized for ESCs, however, the two strategies differ significantly. The differentiation strategy we chose to evaluate was based on the premise that ATRA treatment would increase osteoblast and chrondrocyte lineage commitment of iPS-derived EB outgrowths through runx2 induction [21–24]. ATRA is known to increase neural crest markers by ESCs that provides an origin for mesenchymal elements [22]. Runx2 is expressed in neural crest-derived mesenchyme by precursors to bones, teeth and chondrocytes [38]. Runx2 deficiency or downregulation results in a shift toward adipogenesis [39]. We made EB from iPS cells over 2 days using a hanging drop method followed by plating on gelatin-coated dishes in ES medium with ATRA for 3 days. After 3 days, the medium was changed to lineage-specific differentiation medium and outgrowth evaluated for expression of lineage appropriate genes. Conversely, Lian et al. cultured dissociated iPS cells, without an EB and ATRA treatment stage, in medium to support MSC outgrowth containing defined factors such as basic fibroblast growth factor, platelet-derived growth factor AB, and epidermal growth factor [36, 37]. These culture conditions required 1 week to detect a MSC phenotype confirmed by flow cytometric analysis to detect cell surface molecules including CD73, CD105, and CD133 and the lack of hematopoietic markers CD 45 and CD34 [11]. Lineage-specific differentiation for both studies was 3–4 weeks. Following a 3-day treatment with ATRA, we confirmed expression of the mesenchymal markers CD 90 (50%), lower levels of CD 73, 105, 106, and 133 and lack of hematopoietic markers CD45 and ckit.

The differences between iPS cell-derived mesenchymal cells and BM-MSCs provide a compelling argument for the use of iPS cell engineered cells. Lian et al. [11] demonstrated that iPS cell-derived mesenchymal cells could proliferate to 120 population doublings while maintaining a normal karyotype, had a 10-fold greater level of telomerase activity and had more of a protective effect in a rodent model of hindlimb ischemia than BM-MSCs. The difference between cell types was attributed to the increased ability of iPS-derived mesenchymal cells to survive, engraft, and promote de novo vasculogenesis and muscle differentiation. Independent groups have also demonstrated successful directed differentiation of iPS cells into adipocyte, vascular, and skeletal muscle with efficiencies similar to those of ESCs [16–19].

We demonstrate for the first time that iPS cells could be terminally differentiated into functional osteoblasts that would maintain their phenotype on a three-dimensional gelatin scaffold in vitro and in vivo. As iPS cells are somatic cells manipulated to become pluripotent stem cells, it is vital that their potential to terminally differentiate and maintain a lineage-specific differentiated phenotype, including bone, is documented in vivo otherwise tumorigenesis may become an issue. The osteoblasts lineage was defined by increased expression of the characteristic genes runx2, collagen typeIaI, and BSP/OPN. As anticipated runx2 expression increased with ATRA treatment, whereas the expression of flt-1, col1a1, and spp1 genes increased with differentiation and was maintained over time in vitro [17, 24]. Flt-1 is a marker of both osteoblasts and osteoclasts [28]. When the osteoblasts were passaged and seeded on the gelatin scaffold, the early gene runx2 initially increased and was followed after 2 weeks by increased expression of matrix genes collagen I and BSP in vitro. The osteoblasts production of collagen may enhance further differentiation and secretion of more mature osteogenic matrix [40]. Researchers have shown that to engineer cell-based bone grafts, in vitro commitment or differentiation of the cells used is absolutely necessary [41, 42]. Duan et al. [43] showed that enamel matrix derivatives increased runx2 expression and may be useful in periodontal tissue regeneration. Osteogenic matrix production was confirmed by alizarin red stain and von Kossa to detect mineralization. Isolated implant structures demonstrated the presence of a microcirculation, which is necessary to support complete osteoblastic cell differentiation and functional bone tissue generation. The cells present in the implant formed a matrix, which consisted in part of BSP and osteocalcin. Additionally, remodeling occurred in the implant. Following implantation the osteoblasts were initially spread apart on the sponge-like scaffold. Histological analysis of the tissue revealed that the iPS-derived osteoblasts were densely packed with detectable microvasculature and circulation. As the iPS-derived osteoblast cultures lacked the endothelial marker pecam/CD31, the circulation was likely host derived. The iPS cell-derived osteoblasts did not form tumors in syngenic or nonimmunocompromised mice, which indicates a stable phenotype and no reversion to an embryonic stage of differentiation. These results are in direct contrast to the implantation of undifferentiated iPS or mixtures of undifferentiated iPS and differentiated iPS-derived cells which under all circumstances when subcutaneously implated formed teratomas.

This study presents evidence that iPS cells can serve as a potential source of osteoblasts. These iPS cell-derived osteoblasts increase their expression of bone-specific genes and osteogenic matrix when seeded on a gelatin scaffold in vitro and in vivo, demonstrating their potential use in cell-based therapy. However, future studies are necessary to extend and confirm these conclusions using additional criteria. First, in addition to in vitro commitment or differentiation of the cells used to engineer a bone graft, the environmental conditions in vivo must promote and support osteogenesis [41, 42]. Additional factors may be impregnated into the scaffold, which would enhance bona fide bone formation. Second, the differentiation of ESCs to osteoblasts is affected or enhanced in culture with various scaffolds, including nanofiber [44, 45]. Ideal scaffold properties that support terminal differentiation in vivo must be defined. Finally, the issue of reproducibility of engineered tissues using human cells still exists. To date, most studies have been performed using human ESCs and BM-MSC. With multiple cell types available, the most ideal selection markers for cells used to seed grafts capable of inducing stable bone must be identified [46]. Additionally, human iPS cells must be evaluated for the same potential. The use of integrating virus in human cells poses problems and risks, however, Somers et al. [47] recently described the generation of human iPS from various disease populations using a polycistronic construct which can be removed form the host genome.

In summary, we performed directed differentiation of iPS cells to mesenchymal cells and subsequently the osteoblast lineage in vitro. These osteoblasts were seeded on gelatin scaffolds where they demonstrated stability of the osteoblast phenotype, proliferation, and osteogenic matrix production both in vitro and in vivo. The maintenance of a stable osteoblast phenotype by iPS cell-derived osteoblasts in vivo spotlights these cells as a viable source for further study, in combination with ES and MSCs, for clinical cell-based therapy to treat musculoskeletal diseases. Additionally, the expansion and differentiation potential of patient-specific iPS cells also provides the scientific community with a valuable tool to study the cell-based mechanisms of bone and cartilage diseases. Taken together, a better understanding of these processes promise to offer new therapeutic avenues in the long term.

Acknowledgements

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
  9. References
  10. Supporting Information

This work was supported by R03HL096382-01 and 1R01 HL091105-01 to S.M.M.; NIH grants AR052263 and AR50252 to D.R.R.; and a research grant from the Dystrophic Epidermolysis Bullosa Research Association (DebRA) International from DebRA Austria to G.B. and D.R.R. Additional support was provided by DK NIDDK R01-078966.

References

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
  9. References
  10. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
  9. References
  10. Supporting Information

Additional supporting information available online.

FilenameFormatSizeDescription
STEM_566_sm_suppinfofigure1.eps14406KSupplemental figure 1. The 3T3 E1 murine osteogenic precursor line was used as a control for alizarin red and von Kossa histochemical stains. 3T3E1 cells were cultured in DMEM with 10% FBS or osteogenic differentiation medium. The specificity of staining was confirmed by using the respective negative (A,B) and positive (C,D) controls. The images were taken with 10x objective.
STEM_566_sm_suppinfofigure2.tif1004KSupplemental figure 2. Controls for semiquantitative PCR included RNA isolated from the 3T3 E1 osteogenic precursor line cultured in differentiation medium (diff bone) and whole bone marrow. The specificity of runx2, col1a1 and spp1 gene analysis was confirmed by semiquantiative PCR using the respective positive and negative controls.
STEM_566_sm_suppinfofigure3.tif1476KSupplemental figure 3. A contextual perspective of phenotypic analysis of isolated osteoblast seeded scaffolds. After 12 weeks, gelfoam scaffolds were embedded in OCT, frozen and 25μM sections analyzed. Consecutive sections were immunoreactive to osteocalcin and bone sialoprotein (B-C). DAPI was used to label nuclei (A). Scale bar = 2.5mm.

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.