These data were presented in part at the 27th Annual Meeting of the American Society for Bone and Mineral Research, Nashville, TN, September 23–27, 2005.
The authors state that they have no conflicts of interest.
FOP is a disabling disorder in which skeletal muscle is progressively replaced with bone. Lymphocytes, our model system for examining BMP signaling, cannot signal through the canonical Smad pathway unless exogenous Smad1 is supplied, providing a unique cell type in which the BMP–p38 MAPK pathway can be examined. FOP lymphocytes exhibit defects in the BMP–p38 MAPK pathway, suggesting that altered BMP signaling underlies ectopic bone formation in this disease.
Introduction: Fibrodysplasia ossificans progressiva (FOP) is a rare genetic disorder characterized by progressive heterotopic ossification of connective tissues. Whereas the primary genetic defect in this condition is unknown, BMP4 mRNA and protein and BMP receptor type IA (BMPRIA) protein are overexpressed in cultured lymphocytes from FOP patients, supporting that altered BMP signaling is involved in this disease. In this study, we examined downstream signaling targets to study the BMP–Smad and BMP–p38 mitogen-activated protein kinase (MAPK) pathways in FOP.
Materials and Methods: Protein phosphorylation was assayed by immunoblots, and p38 MAPK activity was measured by kinase assays. To examine BMP target genes, the mRNA expression of ID1, ID3, and MSX2 was determined by quantitative real-time PCR. Statistical analysis was performed using Student's t-test or ANOVA.
Results: FOP lymphocytes exhibited increased levels of p38 phosphorylation and p38 MAPK activity in response to BMP4 stimulation. Furthermore, in response to BMP4, FOP cells overexpressed the downstream signaling targets ID1 by 5-fold and ID3 by 3-fold compared with controls. ID1 and ID3 mRNA induction was specifically blocked with a p38 MAPK inhibitor, but not extracellular signal-related kinase (ERK) or c-Jun N-terminal kinase (JNK) inhibitors. MSX2, a known Smad pathway target gene, is not upregulated in control or FOP cells in response to BMP, suggesting that lymphocytes do not use this limb of the BMP pathway. However, introduction of Smad1 into lymphocytes made the cells competent to regulate MSX2 mRNA after BMP4 treatment.
Conclusions: Lymphocytes are a cell system that signals primarily through the BMP–p38 MAPK pathway rather than the BMP–Smad pathway in response to BMP4. The p38 MAPK pathway is dysregulated in FOP lymphocytes, which may play a role in the pathogenesis of FOP.
Fibrodysplasia ossificans progressiva (FOP) is a severe, disabling genetic disorder characterized by congenital malformations of the great toes and progressive heterotopic ossification.(1,2) Bone formation in FOP begins with large painful swellings that are associated with lymphocytic infiltration into skeletal muscle, followed by the death of muscle cells and development of highly vascular fibroproliferative tissue that matures into bone.(3,4) The bone formed in FOP is normal but is temporally and spatially inappropriate. Physical or surgical trauma to FOP patients often results in the induction of heterotopic bone, restricting the ability to obtain tissue samples.(5–9) Consequently, most studies have been limited to the analysis of peripheral blood lymphocytes that can be safely obtained from patients.
Most cases of FOP are spontaneous, although several multigenerational families show an autosomal dominant pattern of inheritance.(10) The rarity of these families has made traditional positional cloning and linkage analysis difficult.(11,12) Because standard genetic approaches to study FOP are limited, a candidate gene screening approach has been used. In the search for candidate genes, we found that, whereas BMPs are not mutated in FOP patients, the BMP signaling pathway is dysregulated in the FOP lymphoblastoid cell system.(13–17)
BMPs are members of the TGF-β family of proteins that act as potent osteogenic morphogens capable of inducing ectopic bone formation in animal models.(18–22) Recombinant BMPs induce mesenchymal cells to differentiate into bone through cellular events that parallel the stages of bone formation in FOP lesions.(3,23,24) BMPs interact with specific BMP receptors (BMPRs) that phosphorylate downstream effector molecules. Several combinations of type I (BMPRIA, BMPRIB, or ACVR1) and type II (BMPRII or ACVR2) BMPRs mediate BMP signaling pathways.(25–27) Constitutively activated type II receptors phosphorylate type I receptors after ligand binding.(28,29) The activated type I BMP receptors subsequently phosphorylate downstream targets of at least two signaling cascades: the Smad pathway and the p38 mitogen-activated protein kinase (MAPK) pathway.(30,31) Both pathways regulate nuclear transcriptional activity through cofactor association in a cell and developmental-specific context.(32)
The role of BMP signaling in lymphocytes has not been well studied. However, BMP signals control the thymic microenvironment, thereby affecting lymphocyte maturation and thymopoiesis.(33) BMPs are also key regulators of hematopoietic development, and thus, B-cell development during embryogenesis.(34) Additionally, BMPs are associated with cell cycle growth arrest and apoptosis in B cells.(35)
Previously we showed that increased BMPRIA phosphorylation in FOP cells is correlated with increased p38 MAPK phosphorylation but not with phosphorylation of Smads.(17) In this study, we show that the BMP–Smad pathway cannot function in either control or FOP lymphoblastoid cells (LCLs) unless exogenous BMP-specific Smad1 is supplied. We also show that p38 MAPK is the predominant BMP signaling pathway in LCLs and that this pathway is dysregulated in cells from FOP patients.
MATERIALS AND METHODS
Peripheral blood samples were obtained with informed consent from FOP patients and unaffected individuals in accordance with institutional guidelines and Institutional Review Board approval. Lymphoblastoid cell lines (LCLs) were established from peripheral blood mononuclear cells by transformation with Epstein-Barr virus as described previously(14) and grown in Roswell Park Memorial Institute (RPMI) 1640 medium (Invitrogen) supplemented with 15% FBS. For all experiments, cells were grown overnight in serum free RPRI medium. Treatment with BMP4 (150 ng/ml; R&D Systems) for indicated time-points, Noggin (400 ng/ml; a gift of Regeneron Pharmaceuticals) for 30 minutes, or BMPRIA antibody (10 μg/ml; R&D Systems) for 30 minutes was in serum-free medium. For experiments using BMPRIA antibody, cells were also treated with preimmune serum and other control antibodies for 30 minutes in serum-free medium. To block the MAPKs, cells were pretreated for 1 h with SB203580 (1 μM; Calbiochem) for p38 inhibition, PD98059 (10 μM; Calbiochem) for extracellular signal-related kinase (ERK) inhibition, and 420119 (10 μM; Calbiochem) for c-Jun N-terminal kinase (JNK) inhibition and then treated with BMP4 for 90 minutes.
RNA isolation and cDNA synthesis
Total RNA was isolated from LCLs with Trizol Reagent (Invitrogen) following the recommended protocol. Reverse transcription was performed using 5 μg of total RNA, 50 ng random primers (Roche), and Superscript reverse transcriptase II (Invitrogen), following the recommended protocol.
Real-time PCR was performed on an ABI Prism 7000 Cycler. Primers were designed using Primer Express (ABI) for GAPDH: forward, 5′-AGATCATCAGCAATGCCTCCTG-3′ and reverse, 5′-ATGGCATGGACTGTGGTCATG-3′; ID1: forward, 5′-GGTGGAGATTCTCCAGCACG-3′ and reverse, 5′-TCCAACTGAAGGTCCCTGATG-3′; ID3: forward, 5′-TTCCCATCCAGACAGCCG-3′ and reverse, 5′-GCGTTCTGGAGGTGTCAGGA-3′; MSX2: forward, 5′-CCACCCCCTCTAACGGCTAG-3′ and reverse, 5′-AAATTTCAGCTATGTGGTGTGGC-3′. All PCR reactions were performed in triplicate with 50 nM of GAPDH, ID1, or ID3 primers or 300 nM of MSX2 primers and SYBR Green Master Mix (ABI) at 50°C for 2 minutes, 95°C for 10 minutes to denature, followed by 40 cycles of 95°C for 15 s and 60°C for 1 minute. Controls (without template) and standard curves were run for each set of primers. Values are expressed as the mean fold induction ± SE and are results obtained from at least three independent experiments. Statistical analysis was performed using Student's t-test, and the values are considered to be significantly different at p ≤ 0.05. ANOVA analysis was also performed when applicable and resulted in the same statistical significance.
Protein extracts were sonicated on ice in cell lysis buffer (Cell Signaling) containing 1 mM phenylmethylsufonyl fluoride (PMSF). SDS sample buffer was added, and the samples were denatured at 100°C for 5 minutes. Proteins were size fractionated on Novex SDS-polyacrylamide gels under reducing conditions before transfer to polyvinylidene difluoride (PVDF) membranes. Membranes were blocked in TBST (20 mM Tris [pH 7.4], 150 mM NaCl, 0.1% Tween 20) with 5% milk for 1 h at room temperature. p38 and phospho-p38 were detected with primary antibodies (1:1000; Cell Signaling) in TBST with 5% BSA overnight at 4°C and anti-rabbit-IgG-horseradish peroxidase (HRP)-conjugated secondary antibody (1:1000; Cell Signaling) for 1 h at room temperature. Blots were developed with ECL + chemiluminescence reagents (Amersham Biosciences). Scanned images were quantified using NIH Image software and normalized to total p38.
An assay for p38 MAPK activity (Cell Signaling) followed the recommended protocol. Briefly, whole cell lysates were prepared as described above. Phospho-p38 was immunoprecipitated with immobilized monoclonal antibody overnight at 4°C. The beads containing immunoprecipitated phospho-p38 were pelleted and washed, and a kinase assay was performed using 200 μM ATP and 1 μg of ATF2 fusion protein as substrate for 30 minutes at 30°C. Reactions were terminated by addition of SDS sample buffer and heating for 5 minutes at 100°C. Proteins were immunblotted as described above except incubation was with a phospho-ATF2 antibody (1:1000) in dilution buffer (TBST with 5% BSA) overnight. The membranes were incubated with HRP-conjugated secondary antibody (1:1000) for 1 h at room temperature and developed using LumiGLO substrate. Scanned images were quantified using NIH Image software.
Control and FOP LCLs were transfected with Smad1 or Smad6 expression constructs(36) using TransIT Jurkat Transfection Reagent (Mirus). Briefly, 4 μg DNA was added to 10 μl TransIT Jurkat Reagent in 200 μl of serum free media and incubated with 5–10 × 106 cells for 48 h at 37°C, followed by RNA isolation and cDNA synthesis for real-time PCR analysis.
BMP4 increases p38 MAPK phosphorylation levels in FOP cells
BMP4 did not induce a detectable increase in Smad1 phosphorylation in FOP and control LCLs, and very low levels of Smad1 mRNA were detected in either cell type.(17) However, BMP4 stimulated an increase in ID1 mRNA levels in these cells, and this induction could be blocked using SB203580, a specific p38 MAPK inhibitor.(17) These results suggest that LCLs do not use the canonical BMP–Smad signaling pathway to regulate ID1 expression, but instead signal through the p38 MAPK pathway in response to BMP4.
To examine p38 MAPK signaling, control and FOP LCLs were treated with BMP4 for 0, 30, and 90 minutes. After treatment, p38 phosphorylation levels were determined by immunoblotting with p38 and phospho-p38 antibodies. p38 MAPK phosphorylation increased in response to BMP4 in both cell types. Phosphorylation levels were 1.6-fold greater at 30 minutes and 1.9-fold greater at 90 minutes in FOP cells compared with control cells (Fig. 1).
BMP4 increases the kinase activity of p38 MAPK in FOP cells
BMP4 induced higher levels of ID1 mRNA in FOP cells compared with control cells.(17) To determine if this was caused by increased kinase activity of p38 MAPK in response to BMP4, phosphorylation of ATF2 (a direct target of p38 MAPK) was detected by immunoblotting after 0, 30, and 90 minutes of BMP4 stimulation. Consistent with p38 MAPK phosphorylation levels (Fig. 1), the kinase activity of p38 was increased in response to BMP4 in both cell types. Phosphorylation levels were 2.1-fold greater at 30 minutes and 2.0-fold greater at 90 minutes in FOP cells compared with control cells (Fig. 2).
ID1 and ID3 are downstream targets of BMP–p38 MAPK signaling
ID1 and ID3 have been shown to be downstream transcriptional targets of both Smad and p38 MAPK signaling pathways in multiple cell types.(37,38) In response to BMP4, there was a significant increase of ID1 (Fig. 3A) and ID3 (Fig. 3B) mRNA in control and FOP LCLs as measured by real-time PCR. In FOP cells, this induction was ∼5-fold greater for ID1 and 3-fold greater for ID3 compared with control cells. To confirm that induction of ID1 and ID3 mRNA in these cells is through activation of the p38 MAPK pathway, cells were treated with SB203580, a specific p38 MAPK inhibitor. In all cells, treatment with the inhibitor reduced ID1 (Fig. 3A) and ID3 (Fig. 3B) mRNA expression to basal levels. In contrast, treatment with PD98059 and 420119, inhibitors of ERK and JNK, respectively, had no significant effect on ID1 (Fig. 3A) or ID3 (Fig. 3B) mRNA levels. These results confirm that, in lymphocytes, BMP4 induces ID1 and ID3 gene expression through the p38 MAPK cascade.
BMP4-mediated ID1 induction occurs through BMPRIA
The BMPRIA–ACVR2 receptor complex occurs on the surface of LCLs and mediates BMP signal transduction.(17) To further characterize the BMP signaling pathways in LCLs, control and FOP cells were treated with an anti-BMPRIA antibody to suppress BMP signaling at the receptor. Blocking BMPRIA significantly reduced levels of BMP-induced ID1 in both control and FOP cells (Fig. 4A). We also used Noggin, a BMP antagonist, to block BMP4 binding to its receptor. Noggin treatment reduced ID1 mRNA to basal levels in both cell types (Fig. 4B). Taken together, these data show that ID1 induction is BMP4-mediated at least in part through BMPRIA.
Smad1 transfection increases expression of ID1, ID3, and MSX2 mRNA
Smads are not phosphorylated in response to BMP4, and the levels of Smad1 mRNA and protein are low in LCLs.(17) These results are consistent with studies in non-EBV transformed primary human B lymphocytes (data not shown). To determine if lymphoblastoid cells are capable of transducing BMP signals through the Smad pathway, we examined the ability of Smad1 transfected cells to increase expression of MSX2, a gene regulated by BMP-activated Smads but not by BMP-activated p38 MAPK signaling.(39) Smad1 (receptor associated) and Smad6 (inhibitory) expression constructs were transfected into control and FOP LCLs. After transfection, cells were treated with BMP4 for 90 minutes. In the absence of transfected Smad1, MSX2 levels did not change when control and FOP cells were treated with BMP4, supporting our hypothesis that LCLs do not use the Smad pathway in response to BMP4 (Fig. 5). However, MSX2 mRNA was increased in response to BMP4 when Smad1 was transfected into cells, but not when Smad6 was transfected. Interestingly, basal levels of MSX2 were ∼3-fold greater in FOP cells compared with control cells both with and without Smad1 transfection. Treatment with Noggin or BMPRIA antibodies did not alter levels of MSX2 mRNA in non–Smad1-transfected cells (data not shown), suggesting that another, non–BMP-dependent pathway is active in these cells or that a mutation causing basal constitutive activation of a BMP receptor or BMP associated protein is responsible for the upregulation of basal levels of MSX2 in FOP cells. As expected, inhibiting p38 MAPK with SB203580 had no effect on MSX2 expression (Fig. 5), even in the presence of BMP4.
The mRNA levels of ID1 (Fig. 6A) and ID3 (Fig. 6B), genes regulated by both Smad and p38 MAPK pathways,(37,38) increased significantly in LCLs in response to BMP4 and were further stimulated by Smad1 transfection in both cell types, although to a greater extent in FOP cells. Smad1 transfection alone increases ID1 and ID3 mRNA to levels comparable with those seen with BMP-treated untransfected cells, suggesting that promiscuous signaling is occurring in cells with increased levels of Smad1. In Smad1-transfected cells, there was only a partial decrease in the expression of ID1 and ID3 in the presence of a p38 MAPK inhibitor, confirming that ID1 and ID3 are regulated through both Smad and p38 MAPK pathways.
In this study, we showed that, in response to BMP4, lymphoblastoid cells signal through the p38 MAPK pathway and that this limb of the BMP pathway is dysregulated in cells from FOP patients. We also showed that the lack of BMP-stimulated Smad signaling in control and FOP LCLs can be reversed by transfecting the cells with a Smad1 expression vector and that this permits increased expression MSX2, a gene specifically regulated by BMP-activated R-Smads. Human LCLs are therefore a tool for studying BMP-induced p38 MAPK signaling pathways independently of Smad signaling. These results support and greatly extend our earlier preliminary findings.(14,17,40)
BMP4 increases p38 MAPK phosphorylation and activity in control and FOP cells. Whereas basal p38 phosphorylation was similar in control and FOP cells, BMP4 treatment induced p38 phosphorylation to significantly higher levels in FOP cells, suggesting increased activation of p38 MAPK in these cells. Previous studies showed that LCLs from FOP patients express increased levels of BMPRIA on their surface and exhibit defects in ligand stimulated internalization/degradation of this receptor(17); thus, the excessive p38 MAPK phosphorylation in FOP cells may be in part a consequence of overabundant and/or overactive BMPRIA at the cell surface.
The BMPRIA–ACVR2 receptor complex is an important BMP signaling receptor on the surface of LCLs.(17) Inhibition of BMP signaling with the BMP antagonist Noggin reduced ID1 mRNA to basal levels. ID1 mRNA induction was also blocked when cells were treated with an anti-BMPRIA antibody, although to a lesser extent than with Noggin. This may be a consequence of the inability of the antibody to block all of the receptor complexes. Our data support that BMP4-stimulated ID1 induction is BMPRIA-mediated; however, we cannot exclude that BMPs induce some ID1 expression through an alternate receptor complex at the cell surface.
ID1 and ID3 are members of a family of proteins involved in the control of differentiation and cell cycle progression, which act by sequestering ubiquitously expressed basic helix-loop-helix (bHLH) transcription factors.(41,42) Overexpression of ID proteins is associated with a block(41) or stimulation(43–46) of differentiation or an alteration of cell fate(47–49) in different cell contexts. Whereas ID1 and ID3 mRNA levels increase in the presence of BMP4 in both control and FOP cells, the induction is ∼5-fold greater for ID1 and 3-fold greater for ID3 in FOP cells compared with controls. Stimulation of ID1 and ID3 expression is blocked by a p38 MAPK inhibitor, but not ERK or JNK inhibitors. Taken together with the p38 phosphorylation results, these data suggest a dysregulation of BMP signaling in FOP lymphocytes that is reflected in increased p38 MAPK signaling.
Recent studies of Id1 and Id3 heterozygous knockout mice showed that these genes are positive factors promoting bone formation.(50) The mutant mice revealed suppression of proliferation and mineralization in osteoblasts and decreased BMP-induced bone formation in vivo. Other studies of Id1 and Id3 knockout mice showed the importance of these genes in angiogenesis,(51) which is crucial for new bone formation. Taken together, these data suggest that the BMP4-induced dysregulation of ID1 and ID3 in FOP cells may be involved in heterotopic bone formation.
MSX2 is a member of a small family of homeobox genes related to the Drosophila gene muscle segment homeobox (msh).(52) Unlike the ID genes, MSX2, an obligatory Smad target gene,(39) is not upregulated in LCLs after BMP4 stimulation, further supporting our hypothesis that the Smad pathway is not active in these cells. However, basal levels of MSX2 are ∼3-fold higher in FOP cells than control cells, suggesting that another, non–BMP-dependent pathway may be regulating MSX2 expression. Many signaling cascades are known to regulate MSX2, such as the fibroblast growth factor (FGF) pathway. Mutations in the FGF receptors (FGFRs) and MSX2 have been found to cause craniosynostosis, a defect characterized by premature closure of one or more of the fibrous joints between the bones of the skull before brain growth is complete.(53,54) Control and FOP LCLs were unresponsive to FGF, showing no increase in MSX2 expression (data not shown). Lymphocytes may not possess the receptors necessary to signal through FGF or MSX2 may be regulated through another BMP-independent signaling pathway in this cell type. Another hypothesis is that a mutation causing basal constitutive activation of a BMP receptor or BMP associated protein is responsible for the upregulation of basal levels of MSX2 in FOP cells.
To determine if LCLs are capable of transducing BMP signals through the Smad pathway, we transfected Smad1 (receptor associated) and Smad6 (inhibitory) expression constructs into control and FOP cells. When cells are transfected with Smad1, MSX2 basal expression increases, with further upregulation after BMP4 treatment. Because MSX2 is a Smad-specific target gene, we conclude that in the presence of adequate levels of Smad1, LCLs are capable of signaling through this pathway. Conversely, when inhibitory Smad6 is transfected into cells, no change in MSX2 mRNA levels was observed. To further examine signal transduction, Smad1 transfected cells were treated with BMP4 and a p38 MAPK inhibitor. Whereas p38 MAPK inhibitors did not effect MSX2 expression, the inhibitors partially reduce ID1 and ID3 mRNA levels in Smad1 transfected cells. In untransfected cells, p38 MAPK inhibitors reduce ID1 and ID3 mRNA to basal levels. Taken together, these results show that both the Smad pathway and the p38 MAPK pathway control the transcription of ID1 and ID3 when cells are transfected with Smad1. Therefore, it is likely that in FOP, the BMP-Smad pathway is important in other cell types, such as connective tissue cells, muscle cells, peripheral nerve cells, or other physiological targets of BMP4 pathway overactivity.
The data presented here show that BMP signaling in LCLs occurs through the p38 MAPK pathway but not the Smad pathway. However, when given exogenous Smad1, LCLs have sufficient Smad pathway components to signal through this pathway. To our knowledge, no other cell types signal exclusively through the p38 MAPK pathway in response to BMP4, making this an excellent model system to dissect the BMP–p38 MAPK signaling pathway independently of the BMP–Smad pathway.
We also report that FOP LCLs have high levels of phosphorylated p38 MAPK in response to BMP4, leading to elevated kinase activity and upregulation of downstream transcriptional target genes. The dysregulation of BMP signaling in FOP cells may provide important new insights into the mechanisms of heterotopic bone formation in these patients. The difficulty of finding additional multigenerational FOP families for genetic analysis and the inability to safely obtain FOP connective tissue cell lines shows the importance of studying the BMP signaling pathway in the available LCL system, which may enable us to understand the primary molecular pathogenesis in this disabling disorder.
The authors thank Dr Phoebe Leboy, University of Pennsylvania School of Dental Medicine, for the Smad1 and 6 expression constructs and all of her invaluable contributions to this research and manuscript. The authors thank Dr Aris Economides and Regeneron Pharmaceuticals for the generous gift of Noggin protein. Finally, the authors thank all the members of our laboratory for their helpful discussions. This research was supported in part by the International FOP Association, the Center for Research in FOP and Related Disorders, the Ian Cali Endowment, the Stephen Roach-Whitney Weldon Endowment, the Roemex and Grampian Fellowships, the Isaac and Rose Nassau Professorship of Orthopaedic Molecular Medicine, and the National Institutes of Health (NIH R01-AR41916).