The authors state that they have no conflicts of interest.
The study of FOP, a disabling genetic disorder of progressive heterotopic ossification, is hampered by the lack of readily available connective tissue progenitor cells. We isolated such cells from discarded primary teeth of patients with FOP and controls and discovered dysregulation of BMP signaling and rapid osteoblast differentiation in FOP cells compared with control cells.
Introduction: Fibrodysplasia ossificans progressiva (FOP), the most disabling condition of progressive heterotopic ossification in humans, is caused by a recurrent heterozygous missense mutation in activin receptor IA (ACVR1), a bone morphogenetic protein (BMP) type I receptor, in all classically affected individuals. A comprehensive understanding of FOP has been limited, in part, by a lack of readily available connective tissue progenitor cells in which to study the molecular pathology of this disorder.
Materials and Methods: We derived connective tissue progenitor cells from discarded primary teeth (SHED cells) of patients with FOP and controls and examined BMP signaling and osteogenic differentiation in these cells.
Results: SHED cells transmitted BMP signals through both the SMAD and p38 mitogen-activated protein kinase (MAPK) pathways and responded to BMP4 treatment by inducing BMP responsive genes. FOP cells showed ligand-independent BMP signaling and ligand-dependent hyper-responsiveness to BMP stimulation. Furthermore, FOP cells showed more rapid differentiation to an osteogenic phenotype than control cells.
Conclusions: This is the first study of BMP signaling and osteogenic differentiation in connective tissue progenitor cells from patients with FOP. Our data strongly support both basal and ligand-stimulated dysregulation of BMP signaling consistent with in silico studies of the mutant ACVR1 receptor in this condition. This study substantially extends our understanding of dysregulated BMP signaling in a progenitor cell population relevant to the pathogenesis of this catastrophic disorder of progressive ectopic ossification.
Fibrodysplasia ossificans progressiva (FOP), an autosomal dominant genetic disorder of progressive heterotopic ossification (HO), is the most disabling disorder of extraskeletal osteogenesis in humans.(1–3) FOP results in the formation of an ectopic skeleton through a process of soft tissue replacement by bone.(1–3) Presently, there is no cure for FOP, and only limited palliative treatment options are available.(2)
Physical and surgical trauma can induce heterotopic ossification in patients with FOP, which has limited our ability to obtain human biopsy tissue for study. Consequently, most of the molecular and biochemical understanding of FOP has been derived from studies using lymphoblastoid cell lines (LCLs) established from peripheral blood samples from patients with FOP and unaffected family members.(4–8) Whereas these cell lines have been an invaluable resource and have provided critical information on the pathophysiology of FOP, they have inherent limitations. The cells are immortalized (transformed) by Epstein-Barr virus (EBV), grow in suspension culture, and do not express bone-specific genes. Nevertheless, extensive studies using these cell lines have revealed dysregulation of the bone morphogenetic protein (BMP) signaling pathway in FOP.(4–7)
Recently, the mutant FOP gene was mapped by genetic linkage to chromosome 2q23–24, where all affected individuals with classic FOP were found to have an identical heterozygous missense mutation (617G>A; R206H) in the glycine-serine (GS) activation domain of the gene encoding activin receptor IA (ACVR1), a BMP type I receptor.(8) Protein homology modeling of the mutated GS domain predicted constitutive activation of ACVR1 as the underlying cause of the ectopic chondrogenesis, osteogenesis, and joint fusions that occur in patients with FOP.(8,9) More recent in silico modeling of the mutant receptor has refined the model and predicted a complex picture of BMP dysregulation in FOP that includes changes in both basal and ligand-stimulated BMP signaling consistent with our in vitro BMP signaling studies.(10)
To more completely understand the molecular pathogenesis of FOP and the role of BMP signaling in heterotopic ossification, we isolated connective tissue progenitor cells from the tooth pulp of discarded primary teeth of FOP children and unaffected controls. These cells are also called SHED cells: stem cells from human exfoliated deciduous teeth. SHED cells are highly proliferative, adherent cells that have the capacity to differentiate into several different cell types including neurons, adipocytes, chondroblasts, osteoblasts, and odontoblasts and have recently been proposed as a human adult stem cell source for hard tissue engineering.(11–14)
In this study, we investigated BMP signaling and osteoblast differentiation in SHED cells from patients with FOP. We found that FOP SHED cells exhibit ligand-independent leakiness of BMP signaling and ligand-dependent hyper-responsiveness to BMP. FOP cells express higher levels of alkaline phosphatase (ALP) and Runx2 mRNA and more rapidly undergo osteogenic differentiation. These results support previous findings in FOP cells and substantially extend the observation of BMP signal dysregulation in a progenitor cell population relevant to the pathogenesis of this catastrophic musculoskeletal disorder.
MATERIALS AND METHODS
Amplification Grade DNase I, Superscript reverse transcriptase, Trizol, DMEM culture medium, and GlutaMAX supplement were obtained from Invitrogen (Carlsbad, CA, USA). Human recombinant BMP4 was obtained from R&D Systems (Minneapolis, MN, USA); stock solutions (100 ng/μl) were prepared in 4 mM HCl and 0.5% BSA as recommended by the manufacturer and stored at −70°C. Purified human Noggin protein was a generous gift of Regeneron Pharmaceuticals (Tarrytown, NY, USA). SB203580 (p38 MAPK inhibitor) and PD98059 (ERK inhibitor) were obtained from Upstate Biotechnology, and SP600125 (JNK inhibitor) was from CalBiochem/EMD Biosciences (San Diego, CA, USA). β-cyclodextrin, β-glycerophosphate, ascorbic acid sodium salt, dexamethasone, type II collagenase, and Sigma Phosphatase Substrate (p-nitrophenyl phosphate [p-NPP]) were from Sigma (St Louis, MO, USA).
SHED cell isolation and culture
Exfoliated teeth were obtained from pediatric patients according to Institutional Review Board–approved protocols at the University of Pennsylvania and Thomas Jefferson University. The characteristics of the patient population providing samples for this study are presented in Table 1. Cells were isolated as previously reported(11) with minor protocol modifications. The dental pulp was digested with 2 mg/ml type II collagenase for 1 h (37°C) in serum-free DMEM and filtered through a 100-μm cell strainer (BD Falcon, Franklin Lakes, NJ, USA). Cells in the filtrate were recovered by centrifugation (400g, 10 min) and plated in DMEM with 10% FCS, GlutaMAX supplement, and antibiotics. SHED cell strains were established from patients with FOP and unaffected age- and sex-matched controls. The presence of mutations in codon 206 (R206H) in FOP cells was confirmed by DNA sequence analysis.
Table Table 1.. Patient Population Used in These Studies
For experimental treatments, cells were plated in 6-well plates (5 × 104cells/well) in DMEM/10% FCS and grown for 4–6 days (80–90% confluence). Cells were washed with PBS, incubated for 1 h in serum-free medium, and treated with 100 ng/ml BMP4 in serum-free medium for 1.5 h. For potassium depletion (No K) studies, cells were washed with HEPES buffer (25 mM HEPES [pH 7.5], 140 mM NaCl) and incubated in potassium-free medium (No K buffer; 25 mM HEPES [pH 7.5], 0.5% BSA, 5 mM glucose, 140 mM NaCl, 1 mM MgSO4, 1 mM CaCl2) for 1 h before addition of BMP. To disrupt lipid rafts, cells were treated with 4 mM β-cyclodextrin in serum-free medium for 1 h before ligand treatment. For transfection experiments, SHED cells were seeded into 6-well plates and transfected with a Smad6 expression construct (1 μg/well) for 48 h(7) using TransIT-LT1 transfection reagent (Mirus, Madison, WI, USA) following the recommended protocol. Results presented are those obtained with cells derived from a minimum of two control and two patients with FOP.
RNA isolation and real-time PCR
Cells were washed with PBS, and RNA was isolated using Trizol (Invitrogen, Carlsbad, CA, USA), following standard protocols.(7) After phase separation, 30 μg glycogen (Roche Applied Science, Indianapolis, IN, USA) was added to the aqueous phase to facilitate precipitation of RNA. To prepare cDNA, 5 μg RNA was digested with DNase I (Invitrogen) for 10 min, and cDNA was synthesized with Superscript II reverse transcriptase (Invitrogen) and random primers.
Real-time PCR reactions contained forward and reverse primers (Table 2), cDNA (1:5 dilution), and SYBR Green PCR Master Mix (Applied Biosystems, Foster City, CA, USA) as previously described.(7) Each sample was analyzed in triplicate (Applied Biosystems 7000 Sequence Detection System), and target gene mRNA levels were quantified from standard curves and normalized to GAPDH.
Table Table 2.. Primers Used for Quantitative Real-Time PCR
Protein isolation and immunoblotting
Cells were pelleted by centrifugation (1500 rpm, 10 min, 4°C) and resuspended in 250 μl 1x lysis buffer (20 mM Tris [pH 7.5], 150 mM NaCl, 1 mM EDTA, 1 mM EGTA, 1% Triton, 2.5 mM sodium pyrophosphate, 1 mM β-glycerophosphate, and 1 mM Na3VO4, with 1 μg/ml leupeptin and 1 mM PMSF added immediately before use). Cells were sonicated on ice (3 × 10 s), and debris were removed by centrifugation (6000 rpm, 10 min, 4°C). Protein concentrations were determined by BCA Protein Assay (Pierce Biotechnology, Rockford, IL, USA) using BSA as standard. Proteins were electrophoresed through 12% SDS-polyacrylamide gels and transferred to PVDF membranes. Membranes were blocked in PBST (PBS with 0.1% Tween 20) containing 5% milk and incubated with primary antibodies, diluted 1:1000 against Smad1 and Phospho-Smad1 (Cell Signaling Technology, Danvers, MA, USA) in PBST and 1.5% BSA (overnight, 4°C). Bound primary antibodies were detected with a horseradish peroxidase (HRP)-conjugated anti-rabbit or anti-mouse secondary antibody and ECL Plus Western Blotting Detection Reagent (GE-Healthcare, Piscataway, NJ, USA).
Cell differentiation assays
ALP activity was detected histochemically with BCIP/NBT-plus substrate (Moss Substrates, Pasadena, MD, USA) at 37°C.
To quantify ALP enzyme activity, cells were lysed in 0.1 M Tris (pH 7.5) and 0.1% Triton X-100. Cell debris were removed by centrifugation (5000g, 5 min, 4°C) and 50-μl aliquots of supernatant were assayed in 0.5 ml of 1 mg/ml Sigma Phosphatase Substrate in 0.1M Tris (pH 9.5), 1 mM MgCl2; p-nitrophenol (p-NP) formation was detected by spectrophotometry at 405 nm. Protein was determined with a BCA (bicinchoninic acid) protein assay (Pierce, Rockford, IL, USA), using BSA as standard. Enzyme activity is expressed as μmoles p-NP per minute per microgram protein.(15)
To induce mineralization, SHED cells were plated (5 × 104 cells/well, 24-well plates) and after 24 h were treated with osteogenic medium (OM; DMEM/10% FCS, supplemented with 50 μg/ml ascorbate, 10 mM β-glycerophosphate, and 10 nM dexamethasone) for 14–21 days.(11,12) The medium was changed twice weekly. To detect calcium mineralization, cells were stained with 1% Alizarin red (LabChem, Pittsburgh, PA, USA) for 30 min. To quantify mineralization, Alizarin red–stained cells were solubilized in 0.5 N HCL and 5% SDS for 30 min and detected at 405 nm using a Bio-Tek Synergy HT microplate reader.(16)
All results were obtained from a minimum of three determinations from two to five independent experiments. For statistical analysis, we used Student's t-test to compare two samples or ANOVA with Bonferroni's correction when analyzing multiple sets of data. Values were considered to be significantly different at p < 0.05. All statistics were performed with Graphpad software (www.graphpad.com).
Isolation of connective tissue progenitor cells from patients with FOP
Heterotopic ossification is induced in patients with FOP by physical or surgical trauma, severely hampering our ability to obtain primary cells for in vitro studies. However, discarded primary teeth are obtained without trauma and have enabled us to establish adherent connective tissue progenitor cell strains (SHED cells). SHED cells from patients with FOP and unaffected age- and sex-matched individuals (Table 1) were similar in morphology and growth characteristics (Fig. 1A).
Basally leaky and conditionally hyper-responsive BMP signaling in FOP cells
SHED cells grew well in culture until passage 8 (∼30 population doublings) and then began to senesce. Additionally, the ability of SHED cells to respond to BMP declined with increasing time in culture. Consequently, all experiments described in this report used SHED cells at passage 6 or earlier. In the absence of exogenous ligand, ID1 expression was similar in control and FOP cells. However, MSX2 expression, which is Smad-dependent,(7,17) was significantly greater in FOP cells compared with controls (Fig. 1B).
Treatment of SHED cells with BMP4 induced mRNA expression of ID1 and MSX2, two BMP early-response genes. In response to BMP4, the levels of both ID and MSX2 mRNA were consistently higher in FOP cells than in control cells (Fig. 1B). To verify that induction of these genes is ligand dependent, cells were pretreated with the BMP antagonist Noggin(18) before ligand exposure. Noggin treatment reduced BMP induction of ID1 to near basal levels in both control and FOP cells and blocked induction of MSX2 in a similar manner in control cells (Fig. 1B), confirming that gene induction is mediated by BMP. However, in FOP cells Noggin did not completely abolish MSX2 expression, consistent with a component of ligand-independent leaky BMP signaling through the Smad pathway in these cells.
Dysregulation of Smad and p38MAPK signaling in FOP cells
BMP signal transduction is mediated through the canonical Smad pathway and the p38 MAPK cascade.(18–22) To assess BMP signaling through Smad activation in SHED cells, phosphorylation of Smad1 (a BMP pathway activating Smad) was examined in response to BMP. Increased Smad1 phosphorylation was detected by immunoblotting when cells were treated with BMP4, with similar induction profiles over time in both control and FOP cells (Fig. 2A). Quantitation by densitometry showed Smad1 phosphorylation was 1.5-fold higher in FOP cells compared with control cells (p < 0.05) at the 90-min time-point.
Transfection of an expression construct for Smad6 (an inhibitory Smad) into control and FOP cells reduced BMP4-induced ID1 and MSX2 mRNA to near basal levels, although higher levels persisted in FOP cells compared with control cells (Fig. 2B). These results showed that BMP4 treatment activated the Smad signaling pathway in SHED cells, that there was greater basal activity of this pathway in FOP cells compared with control cells, and that the pathway was partially resistant to modulation by inhibitory Smads.
To study the contribution of the p38 MAPK pathway to activation of BMP target genes, SHED cells were incubated with SB203580, a specific inhibitor of p38 MAPK,(7) before treatment with BMP4. Inhibition of p38 MAPK reduced BMP4 stimulation of ID1 mRNA in both control and FOP cells, but had no effect on MSX2 expression, consistent with previous reports that BMP induction of MSX2 was Smad dependent(7,17) (Fig. 3). Treatment of SHED cells with inhibitors of extracellular signal-regulated kinase (ERK) and inhibitors of c-jun N-terminal kinase (JNK) had no effect on ID1 expression (data not shown). Hence, BMP4 induction of ID1 is mediated through both the Smad and p38 MAPK pathways, whereas BMP4 induction of MSX2 is activated exclusively through the Smad pathway in SHED cells.
Clathrin-mediated internalization of BMP receptors
To examine the effect of receptor endocytosis on BMP signaling, SHED cells were pretreated either with potassium-free medium (i.e., potassium depletion) to inhibit clathrin polymerization(23) or with β-cyclodextrin to disrupt lipid rafts and block internalization through the caveolin pathway.(24) We found that potassium depletion blocked BMP4 induction of ID1 mRNA, reducing ID1 mRNA expression to near basal levels in both FOP and control cells (Fig. 4). In contrast, treatment of cells with β-cyclodextrin did not affect BMP-induced ID1 mRNA expression (Fig. 4) in either FOP or control cells. These results suggest that BMP4 signaling in FOP and control cells requires receptor internalization through clathrin-coated pits and that the caveolin pathway does not substantially affect BMP signaling in SHED cells.
Enhanced osteogenic potential of FOP SHED cells
SHED cells have the capacity to differentiate into several connective tissue cell types.(12) To compare the osteogenic potential of FOP and control cells in the absence of exogenous osteo-inductive factors, several osteoblast markers were quantified by real-time PCR. Relative to control cells, FOP cells showed elevated basal expression of Runx2 and ALP mRNA but not osteocalcin (OC) or osteopontin (OP) mRNA (Fig. 5A). The expression of ALP enzyme activity was similar when cells were grown either in the absence or presence of serum (Figs. 5B and 5C), suggesting that elevated ALP activity in FOP cells is not a consequence of a specific response of these cells to a serum-derived factor.
Studies were next carried out to assess the ability of SHED cells to differentiate into osteoblasts. We initially treated cells with BMP2 or BMP4 in the presence of ascorbate and β-glycerophosphate but were unable to induce differentiation. However, as previously reported for SHED cells,(12) when cells were cultured in medium supplemented with dexamethasone, ascorbate, and β-glycerophosphate (OM), differentiation was observed. We examined expression of ALP and Runx2 mRNA after 2 and 14 days in OM. Runx2 expression remained relatively constant over the time-course examined (data not shown), whereas ALP mRNA expression increased in both control and FOP cells after 14 days (Fig. 6).
The mineralization of SHED cells in response to OM was assayed by Alizarin red staining, indicative of calcium deposition. Increased Alizarin staining in control cells was detected at 3 wk; however, FOP cells showed a more rapid mineralization response with increased Alizarin staining by 2 wk (Fig. 7A). Consistent with these data, OC mRNA, a marker of mineralization, was increased in FOP cells grown in OM for 14 days and in control cells after 21 days (Fig. 7B).
FOP is a challenging condition to study. Because physical and surgical trauma exacerbates FOP by inducing bone formation, it has been difficult to obtain biopsy samples from patients with FOP for detailed biochemical and molecular analysis. In this report, we describe primary osteogenic progenitor cells (SHED cells) that are derived safely and nontraumatically from the dental pulp of discarded primary teeth of patients with FOP. These cells are the first adherent primary connective tissue cell strains from patients with FOP that have been used to study BMP signaling and osteoblast differentiation in this condition. This study substantially extends observations of BMP pathway dysregulation previously reported in lymphoblastoid cell lines (LCLs) from patients with FOP.(4–7)
The importance of studying BMP signaling in connective tissue cells from patients with FOP is further highlighted by recent long-term follow-up studies of hematopoietic and connective tissue cell lineages in a patient with FOP who underwent bone marrow transplantation for treatment of intercurrent aplastic anemia. This study strongly supports that resident connective tissue progenitor cells are more directly relevant to the pathogenesis of heterotopic ossification in FOP than nonadherent circulating cells of hematopoietic origin, as in the LCLs.(25)
The intracellular pathways that mediate BMP signaling have been found to be highly cell type specific. We previously reported that human LCLs signal primarily through the p38 MAPK pathway, but on introduction of a Smad1 expression construct, are fully capable of using the Smad pathway.(7) In this study, we found that SHED cells were responsive to BMP signaling and that both the Smad and p38 MAPK limbs of the BMP signaling pathway were active and dysregulated in FOP cells. Of particular note, we found that FOP SHED cells exhibited ligand-independent basally leaky BMP signaling and ligand-dependent hyper-responsiveness to BMP stimulation. These findings are consistent with recent protein homology modeling of the mutated ACVR1 receptor in FOP that predicts dysregulation of the receptor under both ligand-independent and ligand-stimulated conditions.(8,9)
The particular signaling pathway used by BMPs can have major physiological consequences. We observed differences in the basal and ligand-stimulated response of FOP cells through the Smad and p38 MAPK pathways. At present, very little is known about how cells differentially regulate various components of the BMP pathway in health and disease. For example, BMP4 inhibits proliferation and promotes myocyte differentiation of lung fibroblasts through the JNK pathway and the BMP-specific Smad pathway.(26) Inactivating mutations in BMPRII that lead to primary pulmonary hypertension attenuate Smad phosphorylation but preserve MAPK activation in pulmonary vascular cells.(27)
Endocytosis of cell surface receptors serves crucial roles in cell signaling and in establishing and maintaining morphogenetic gradients.(21) BMP signaling in SHED cells was inhibited by potassium depletion, which blocks the formation of clathrin lattices.(23) In contrast, β-cyclodextrin, which disrupts lipid rafts,(24) had no apparent effect on BMP signaling in either FOP SHED cells or control SHED cells. These results strongly suggest that, after ligand binding, receptor-mediated endocytosis through clathrin-coated pits is necessary for BMP-mediated signal transduction. Our results are consistent with reports showing that internalization of TGF-β receptor complexes into clathrin-coated pits is required for TGF-β–induced Smad activation(21) and that BMP2-mediated Smad signaling requires clathrin-dependent endocytosis.(22)
In addition to hyper-responsiveness to BMP4, we found that FOP SHED cells expressed higher basal levels of the osteogenic markers Runx2 and ALP and also mineralized more rapidly than control cells. Because Runx2 is a crucial regulator of bone formation,(28) increased expression of Runx2 may “prime” FOP SHED cells, accelerating their differentiation into osteoblasts. This could account for clinical findings of the rapid induction of heterotopic ossification observed in some patients with FOP.
In these studies, we found that treatment of SHED cells with medium containing ascorbate, β-glycerophosphate, and dexamethasone (OM) induced osteogenic differentiation (Fig. 7), as previously reported for SHED cells,(12) human embryonic stem cells (hESs), and human bone marrow stromal cells (hMSCs).(29) Unexpectedly, we found that neither BMP2 nor BMP4 was unable to promote osteogenic differentiation, even though SHED cells clearly respond to BMP treatment by upregulating BMP response genes ID1 and MSX2, with FOP cells showing an enhanced response to ligand treatment compared with control cells. Taken together, our data suggest that endogenous BMP signaling may be necessary but not sufficient for osteogenic differentiation in these cell strains. These findings may reflect an altered differentiation potential of SHED cells or a requirement of the permissive OM to induce the differentiated osteogenic phenotype directly. Perhaps other factors, in addition to ascorbate and β-glycerophosphate, are necessary to promote BMP-mediated osteogenic differentiation. It is well known that flare ups in patients with FOP are observed after physical trauma, and the presence of inflammatory mediators may enhance the ability of BMPs to induce osteogenic differentiation of these cells. Alternatively, SHED cells may respond to BMP exposure differently in vivo, where they are growing in a 3D environment of collagen, heparan sulfate proteoglycans and other connective tissue proteins, cytokines and growth factors, and other resident cell types. To address these questions, future experiments will be directed at a detailed examination of factors involved in the osteogenic differentiation of SHED cells.
Extensive clinical evaluations of patients with FOP strongly suggest that mutations in ACVR1 can have wide-ranging physiological effects in multiple cell and tissue types, but the specific contributions of the various intracellular BMP pathways (Smad, p38 MAPK, JNK, ERK) in many of those cell types is unknown. In addition to forming heterotopic bone, patients with FOP have congenital skeletal malformations, degenerative arthropathies, osteochondromas, and progressive hearing impairment, thus showing the widespread cell and tissue effects of dysregulated ACVR1-mediated BMP signal transduction. The in vitro studies presented here provide a basis for further examination of the various components of the dysregulated BMP signaling pathway (specifically the Smad and p38 MAPK pathways) in cells and tissues from FOP knock-in mice, presently under development. Such in vivo studies will likely show additional levels of complexity in BMP signaling, and other signaling pathways, that will be relevant to an integrated molecular understanding of FOP at the cellular, tissue, and organ levels.
This study provides the first detailed analysis of primary connective tissue progenitor cells obtained from patients with FOP. SHED cells are a readily available source of osteoprogenitor cells that can be safely obtained from these patients and provide a well-defined experimental system to examine the dysregulated signal transduction and osteoblast differentiation in FOP. Future studies with these cells should substantially increase our understanding of FOP pathophysiology and provide novel insights into putative therapeutic strategies for this disabling condition.
The authors thank Dr Songtao Shi (NIH), Dr Pamela Robey (NIH), Dr David Glaser, Dr Robert Mauck, Rita Bhagat, Meiqi Xu, and members of the FOP laboratory for helpful discussions during the development and progress of these studies. We acknowledge Dr Aris Economides, Regeneron Pharmaceuticals, for generously providing purified Noggin protein. This work was supported in part by the Center for Research in FOP and Related Disorders, the International FOP Association, the Ian Cali Endowment, the Weldon Family Endowment, the Isaac & Rose Nassau Professorship of Orthopaedic Molecular Medicine, and a grant from the National Institutes of Health (R01-AR41916).