ALK2 R206H mutation linked to fibrodysplasia ossificans progressiva confers constitutive activity to the BMP type I receptor and sensitizes mesenchymal cells to BMP-induced osteoblast differentiation and bone formation

Abstract

Fibrodysplasia ossificans progressiva (FOP) is a rare disabling disease characterized by heterotopic ossification for which there is currently no treatment available. FOP has been linked recently to a heterozygous R206H mutation in the bone morphogenetic protein (BMP) type I receptor activin receptor–like kinase 2 (ALK2). Expression of the mutant ALK2-R206H receptor (FOP-ALK2) results in increased phosphorylation of the downstream Smad1 effector proteins and elevated basal BMP-dependent transcriptional reporter activity, indicating that FOP-ALK2 is constitutively active. FOP-ALK2-induced transcriptional activity could be blocked by overexpressing either of the inhibitory Smads, Smad6 or -7, or by treatment with the pharmacological BMP type I receptor inhibitor dorsomorphin. However, in contrast to wild-type ALK2, FOP-ALK2 is not inhibited by the negative regulator FKBP12. Mesenchymal cells expressing the FOP-ALK2 receptor are more sensitive to undergoing BMP-induced osteoblast differentiation and mineralization. In vivo bone formation was assessed by loading human mesenchymal stem cells (hMSCs) expressing the ALK2-R206H receptor onto calcium phosphate scaffolds and implantation in nude mice. Compared with control cells FOP-ALK2-expressing cells induced increased bone formation. Taken together, the R206H mutation in ALK2 confers constitutive activity to the mutant receptor, sensitizes mesenchymal cells to BMP-induced osteoblast differentiation, and stimulates new bone formation. We have generated an animal model that can be used as a stepping stone for preclinical studies aimed at inhibiting the heterotopic ossification characteristic of FOP. © 2010 American Society for Bone and Mineral Research

Introduction

Fibrodysplasia ossificans progressiva (FOP) is a rare disabling autosomal dominant disorder with an incidence of 1 in 2 million individuals characterized by congenital malformations of the great toes and progressive heterotopic ossification in muscles, tendons, ligaments, and other connective tissues.1, 2 Heterotopic ossification in FOP patients is episodic and occurs spontaneously or in response to trauma, leading to progressive cervical spine fusion, scoliosis, submandibular swelling, hearing loss, and death resulting from pulmonary complications of thoracic insufficiency syndrome.1 At present, no treatment for FOP is available.

A single R206H point mutation in a bone morphogenetic protein (BMP) type I receptor, termed activin receptor-like kinase 2 (ALK2), has been linked as the causative mutation in FOP patients with classic clinical features.3 More recently, atypical FOP patients with heterotopic ossification combined with unusual additional clinical symptoms were all found to have mutations in the protein-coding region of ALK2, some of them in-frame missense mutations different from R206H.2, 4

BMPs are members of the transforming growth factor β (TGF-β) superfamily that were isolated and purified to homogeneity initially from bone through their ability to stimulate ectopic bone formation.5, 6 Partial protein sequence analysis led to the identification of the BMP primary structure by molecular cloning.7, 8 BMPs are expressed not only in skeletal tissues but also in soft tissues, and BMP-4, -6, and -9 have been detected in serum.9, 10

BMPs regulate the fate of various mesenchymal cell types. BMPs inhibit myogenic differentiation but promote the differentiation of mesenchymal cells into osteoblasts,11–13 and BMPs can either stimulate or inhibit adipocyte differentiation.14, 15 The effects BMPs have depend on the isoform, concentration, and cellular context. Certain BMPs have been found to be misexpressed in human diseases. For example, lymphocytes of FOP patients were shown to express elevated levels of BMP-4 that may contribute to ectopic osteogenesis.16 Consistent with this notion, transgenic mice overexpressing BMP-4 were found to suffer from heterotopic ossification reminiscent of FOP.17

As with other members of the TGF-β family, BMPs elicit their cellular effects via heteromeric complex formation of single-pass transmembrane type I and II serine/threonine receptors.18 The type I receptor, also termed ALK, is a substrate for and acts downstream of the type II receptor and is the predominant determinant of the signaling response induced by the BMP-receptor complex. Type II receptor-induced serine and threonine phosphorylation occurs in the glycine/serine residue-rich domain (GS domain) of the type I receptors.19 Specific artificial mutations in this domain, such as Q207D in ALK2, render the type I receptors constitutively active; that is, they can signal in the absence of ligand stimulation and type II receptors.20 The negative regulator FK506-binding protein 12 kDa (FKBP12) can bind to the type I receptors via the leucine-proline interaction motif in the GS domain21, 22 and inhibits signaling by shielding the serine and threonine residues from type II receptor–mediated phosphorylation.23, 24 Three BMP type II receptors, that is, BMPR-II, ActRII, and ActRIIB, and four BMP type I receptors, that is, ALK1, ALK2, ALK3, and ALK6, have been identified. Each of the type I receptors partners with different affinity to specific BMP isoforms. Whereas BMP-6 and -7 bind with higher affinity to ALK2 than to ALK3 and -6, BMP-2 and -4 bind preferentially to ALK3 and -6.25 BMP-9 binds strongly to ALK1 and with lower affinity to ALK2.26 Each BMP type I receptor has a distinct expression pattern; ALK2 is broadly expressed, including on mesenchymal cells.27

The activated BMP type I receptor phosphorylates specific receptor-regulated (R)–Smad proteins, Smad1, Smad5, or Smad8, which assemble into heteromeric complexes with the common partner (co-)Smad4. Heteromeric Smad complexes translocate into the nucleus, where they either stimulate or repress the transcription of specific target genes.18 The BMP signaling cascade is intricately regulated by extracellular antagonists such as noggin, which block BMP binding to its receptors28 and the inhibitory Smads antagonizing the intracellular signal cascade.29

To determine whether the R206H mutation in ALK2 affects its function, we investigated the effect of ectopic expression of this mutant receptor on basal and BMP-induced signaling responses, in particular with respect to osteoblast differentiation, mineralization, and ectopic bone formation.

Materials and Methods

Materials

Human recombinant BMPs were obtained from R&D Systems (Minneapolis, MN, USA). Dorsomorphin was purchased from Enzo Life Sciences (Farmingdale, NY). Expression plasmids for ALK2, CA-ALK2, Smad6, Smad7, and FKBP12 have been described previously.22, 30 The scaffolds consisted of 2- to 3-mm porous biphasic calcium phosphate particles that were sintered at 1150 °C as described previously.31

Construction of plasmids

A polymerase chain reaction (PCR) was done with primers carrying the FOP R206H mutation. The BbsI-BglII fragment in the wild-type ALK2 construct (pcDNA) was replaced with the PCR fragment cut with BbsI-BglII. The replaced part was sequenced-verified. For the adenoviral constructs, the ALK2 cDNAs were cloned in the pENTR1a and then recombined in the pAd/CMV/V5/DEST using clonase (Invitrogen, Carlsbad, CA). The lentiviral constructs were made by cloning a NheI-SalI fragment from the pcDNA3-ALK2 constructs in the NdeI-XhoI sites of the pLenti-CMV-IRES-GFP vector.

Cell culture

COS, 293, and C2C12 cells were cultured in DMEM with high glucose (Gibco, Carlsbad, CA) containing 10% fetal bovine serum (FBS, Gibco). Bovine aortic endothelial cells (BAECs) were cultured in DMEM with low glucose (Gibco) containing 10% FBS (Gibco). Human mesenchymal stem cells were cultured in α modified essential medium (α-MEM, Gibco) containing 10% FBS (Gibco), 0.2 mM L-ascorbic acid-2-phosphate (Sigma, St. Louis, MO) and 1 ng/mL of basic fibroblast growth factor (Peprotech, Rocky Hill, NJ).

Transcriptional reporter assay

Cells were seeded in 24-well plates and transiently transfected for 4 hours with the different ALK2-expressing plasmids, a β-galactosidase-expression plasmid, and the BRE-luciferase reporter construct32 using Lipofectamine reagent (Invitrogen) according to the manufacturer's protocol. After 2 days, the cells were serum-starved for 8 hours and subsequently stimulated for 16 hours with BMP. Cells were washed and lysed, and activity of luciferase and β-galactosidase, which served as a control to correct for transfection efficiency, were determined. Each transfection was carried out in triplicate, and representative experiments are shown.

Western blot analysis

Cells were seeded in 6-well plates and allowed to grow to confluence. Cells were washed with PBS and serum-starved for 4 hours [human mesenchymal stem cells (hMSCs)] or overnight (C2C12 cells). Cells were stimulated with different concentrations of BMP-6 for 60 minutes, washed with PBS, and lysed in SDS sample buffer. Samples were boiled for 5 minutes and subjected to SDS-PAGE and Western blotting. Smad phosphorylation was detected with an antibody specifically recognizing phosphorylated Smad1 and Smad5, which has been described previously.33

Alkaline phosphatase assay

C2C12 cells were stimulated in normal culture medium with the addition of BMP, and the hMSCs were grown in normal culture medium with the addition of dexamethasone (10 nM) and BMPs. Histochemical examination of alkaline phosphatase (ALP) activity in C2C12 cells or hMSCs was performed using naphtol AS-MX phosphate (Sigma) and fast blue RR salt (Sigma), as described previously.34 To quantify the data, the histochemically stained cell material was solubilized in 50 mM NaOH in ethanol, and absorbance was measured at 550 nm.

Mineralization assay

Mesenchymal stem cells were seeded at a density of 10 × 10³ cells/cm² and grown in osteogenic media (Lonza, Basel, Switzerland) supplemented with BMP-6. Cells were stained with alizarin red S. Briefly, the cells were washed with PBS and fixed with 3.7% formaldehyde. The cells were again washed with PBS and then incubated with 2% alizarin red S solution (pH 4.2) for 2 minutes and subsequently washed with distilled water.

Adenoviral/lentiviral production and transduction

Adenoviruses were produced in 293T cells according to the Invitrogen protocols. C2C12 cells were transduced overnight with an multiplicity of infection (MOI) of 250, and the next day the medium was refreshed. One and a half days after transduction, the cells were stimulated with BMP-6. For lentiviral production, 293T cells were transfected using the calcium phosphate method. For one T175 (confluency 60% to 70%), 9.2 µg of VSV, 17.4 µg of GAG/POL, 13.2 µL of REV, and 26 µg of the pLenti constructs were used. Virus was harvested 48 and 72 hours after transfection. The virus was filtered, and the titer was determined by a p24 ELISA. For transduction, 1.5 million cells were plated in a p145 and were transduced with lentivirus for 4 hours (475 ng p24 of virus per 1 × 10⁶ cells) in the presence of diethylaminoethyl (DEAE)-dextran (20 µg/mL), after which the medium was refreshed. After 3 days, the cells were replated for experiments.

In vivo implantation

To assess the in vivo bone-formation, hMSCs virally transduced with empty vector, wild-type ALK2, or the FOP mutant ALK2 R206H were loaded onto porous biphasic calcium phosphate scaffolds (2 to 4 mm) and implanted subcutaneously on the backs of nude mice. Then 1 × 10⁶ cells in 300 µL were loaded onto five scaffolds in ultralow attachment plates. Prior to implantation into nude mice, hMSCs were cultured on the scaffolds for 1½ weeks in proliferation medium. Immune-deficient mice were anesthetized using isoflurane. Implantation of the scaffolds was performed as described previously.35 In short, five subcutaneous pockets were made, and each pocket was implanted with three particles. The incisions were closed using a Vicryl 5-0 suture. After 6 weeks, the mice were euthanized using CO₂, and scaffolds were explanted, fixed in 1.5% glutaraldehyde (Merck, Darmstadt, Germany) in 0.14 M cacodylic acid (Fluka, St. Louis, MO) buffer (pH 7.3), dehydrated, and embedded in methyl methacrylate (Sigma) for sectioning. Approximately 10-µm-thick sections were processed on a histologic diamond saw (Leica Saw Microtome Cutting System). To visualize the newly formed ectopic bone, the sections were stained with basic fuchsin and methylene blue. Basic fuchsin stains the newly formed bone pink, and the other cellular tissues stain light blue with methylene blue. The CaPO₄ phosphate scaffold material is not stained by this procedure. At least five sections were made from each scaffold. The sections were analyzed using a light microscope and by making high-resolution digital photographs from randomly selected sections from each scaffold graft. The extent of bone formation (areas with intense pink staining) was determined relative to the total available pore area for new bone growth. A custom-made macro was used to measure bone/pore area ratios. As a positive control for the ability of hMSCs to induce ectopic bone, hMSCs were treated prior to implantation for 4 days with 1 mM dibutyryl adenosine cyclic monophosphate (cAMP, Sigma); after 6 weeks in the mice, the explanted scaffolds showed that hMSCs from this donor were capable, albeit not so efficiently, of forming ectopic bone. All animal experiments were approved by the Animal Ethics Committee of Leiden University and conformed with the Guidelines for the Care and Use of Laboratory Animals (National Institutes of Health Publication No. 85-23, revised 1996).

Statistical analysis

All transcriptional response and ALP assay results are expressed as the mean ± SEM, unless otherwise stated. Student's t test was used for statistical analysis, and p < .05 was considered to be statistically significant. All in vivo data were analyzed statistically using SPSS Version 16.0 for Windows (IBM, Armonk, NY). A nonparametric Kruskal-Wallis test was used to evaluate significance between ALK2-R206H-transduced hMSCs cells, empty-vector-transduced hMSCs, and wild-type ALK2–transduced cells. A value of p ≤ .05 was considered statistically significant.

Results

The FOP mutation R206H in ALK2 confers constitutive activity

Point mutations in the BMP type I receptor ALK2 have been linked to FOP. The function of R206H ALK2 (FOP-ALK2), the most recurrent mutation and the one associated with classic symptoms of FOP,2 was chosen for analysis. We sought to determine whether this mutation alters ALK2 activity and in particular if it renders ALK2 constitutively active, as hypothesized by Shore and colleagues.3 We compared the ability of the FOP-ALK2 mutant to wild type ALK2 and the constitutively active (CA)–ALK2 mutant to activate the BMP-Smad-dependent luciferase reporter (BRE-luc) in the absence or presence of BMP-6 (Fig. 1A). We observed that in the absence of BMP-6, cells expressing FOP-ALK2 show significantly elevated reporter activity compared with control pcDNA3 or wild-type ALK2–transfected cells (p < .001), albeit the level was not as high as on transfection of the artificial CA-ALK2 mutant construct. On BMP-6 stimulation, FOP-ALK2- and CA-ALK2-transfected cells reached similar levels that were slightly elevated compared with pcDNA control–transfected cells. We consistently observed that ectopic expression of wild-type ALK2 inhibited the BMP-6-induced reporter activity. Identical results (as shown in Fig. 1A) for COS cells were obtained in multiple other cell types, including C2C12 cells (data not shown). The constitutive activity of the FOP-ALK2 mutant was attenuated by dorsomorphin,36 a small-molecule ATP analogue that inhibits BMP and CA-ALK2-induced Smad-dependent signaling responses (Fig. 1B). In addition, ectopic expression of inhibitory Smads, that is, Smad6 and Smad7, blocked FOP-ALK2 mutant–induced constitutive BRE-luc activity (Fig. 1B).

Figure 1. The R206H FOP mutation in ALK2 confers constitutive activity. (A) COS cells were transfected with pcDNA3 or the indicated ALK2-pcDNA3-based expression plasmids (ie, wild-type ALK2, CA-ALK2, and FOP-ALK2), together with the BRE-luc transcriptional reporter. COS cells were not stimulated or stimulated with 100 ng/mL of BMP-6 for 16 hours before cells were lysed, and luciferase activity was measured. RLU = relative luciferase units. (B) COS cells were transfected with pcDNA3 or the FOP-ALK2 expression plasmid together with the BRE-luc transcriptional reporter with or without a Smad6 or -7 expression plasmid. COS cells were either not stimulated or stimulated with 100 ng/mL of BMP-6 in the absence or presence of 500 nM dorsomorphin for 16 hours before cells were lysed and luciferase activity was measured. RLU = relative luciferase units. (C) C2C12 cells were transduced with adenoviruses expressing HA-tagged wild-type, CA-, and FOP-ALK2. The LacZ adenovirus–expressing β-galactosidase was used as a negative control. Cell extracts were fractionated by SDS-PAGE and blotted. The membranes were incubated with p-Smad1 antibody, which specifically recognizes phosphorylated Smad1/5/8. Cell lysates also were subjected to Western blotting to check equal expression of transduced ALK2 adenoviruses using HA-tag antibody and β-actin antibody as a loading control.

To consolidate our finding that FOP-ALK2 confers constitutive activity, we compared the ability by which ectopic expression of wild-type ALK2, CA-ALK2, FOP-ALK2, or β-galactosidase affected Smad1 phosphorylation. We found that in the absence of BMP ligand, the ectopic expression of FOP-ALK2 stimulated Smad1 phosphorylation, albeit not as efficiently as CA-ALK2 (Fig. 1C). FOP-ALK2 did not significantly potentiate BMP-6-induced Smad1 phosphorylation (data not shown). Thus the BRE-luc and Smad1 phosphorylation responses induced by FOP-ALK2 are fully consistent with each other. Previously, mesenchymal cells isolated from the teeth of FOP patients were reported to have elevated p38 MAP kinase activity.37 However, we observed no increase in phosphorylated p38 levels in cells transduced with the FOP-ALK2 mutant compared with LacZ-transduced cells or in potentiation of the BMP-6-induced p38 activation (data not shown).

The R206H mutation is positioned in close proximity to the FKBP12-binding site on ALK2.21, 22 FKBP12 antagonizes activation of type I receptors of the TGF-β superfamily by sterically hindering access of the type II receptors to the serine and threonine residues in the GS domain of the type I receptors. The CA-ALK2 mutant is resistant to FKBP12-mediated inhibition. Similarly, FKBP12 recruitment to FOP-ALK2 also may be affected and be the cause of the enhanced receptor activity. To investigate this possibility, we determined the ability of FKBP12 to inhibit both basal and BMP-6-induced BRE-luc activity in cells expressing either wild-type ALK2, CA-ALK2, or FOP-ALK2. As expected, in wild-type ALK2–expressing cells, both basal and BMP-6-induced BRE-luc activity was susceptible to FKBP12-mediated inhibition. The FOP-ALK2 mutant, however, was resistant to the inhibitory effect of FKBP12 in a manner similar to CA-ALK2 (Fig. 2).

Figure 2. The FOP-ALK2 mutant receptor is partially resistant to the inhibitory effect of FKBP12. Bovine Aortic Endothelial Cells (BAECs) were transfected with wild-type, CA-, and FOP-ALK2 expression plasmids together with the BRE-luc transcriptional reporter with or without an FKBP12 expression plasmid. BAECs were either not stimulated or stimulated with 100 ng/mL BMP-6 for 16 hours before cells were lysed and luciferase activity was measured. RLU = relative luciferase units.

Ectopic expression of the FOP-ALK2 mutant sensitizes cells to BMP-induced osteoblast differentiation

To analyze the effect of FOP-ALK2 on basal and BMP-6-induced osteoblast differentiation, we ectopically expressed the mutant type I receptor by adenoviral gene transfer in C2C12 myoblasts and compared its effect with that of wild-type ALK2, CA-ALK2, and LacZ. C2C12 cells are thought to correspond to satellite cells and can transdifferentiate in osteoblast-like cells on stimulation with BMPs or by expression of CA-ALK2.12, 30 The osteoblast-like differentiation can be measured easily by determining alkaline phosphatase (ALP) activity; ALP is an early maker for osteoblast differentiation. In contrast to CA-ALK2, FOP-ALK2 (and wild-type ALK2) did not raise basal ALP levels, but BMP-6-induced ALP activity was greatly enhanced by ectopic FOP-ALK2 or CA-ALK2 expression (Fig. 3). The fold induction by BMP-6 is similar for CA-ALK2- and FOP-ALK2-transduced cells. Similar results as for BMP-6 were obtained using BMP-2 and BMP-4, albeit less efficient than as found for BMP-6 (data not shown). The latter can be explained by decreased affinity of BMP-2 and BMP-4 for ALK2 compared with BMP-6.

Figure 3. Ectopic expression of FOP-ALK2 sensitizes cells to undergo BMP-induced osteoblast differentiation. (A) C2C12 cells were transduced with adenoviruses expressing HA-tagged wild-type, CA-, and FOP-ALK2 or LacZ adenovirus (expresses β-galactosidase, used as a negative control) in the absence or presence of either 50 or 100 ng/mL of BMP-6 for 3 days. ALP activity in and/or associated with cells was measured histochemically. (B) ALP activity as shown in panel A was solubilized, and absorbance was m spectrometrically easured at 550 nm. (C) Ectopic expression of HA-tagged wild-type, CA-, and FOP-ALK2 was determined by blotting cell lysates with HA-tag antibody. The blot was incubated with β-actin antibody as a control for equal loading.

Subsequently, we compared the effect of lentiviral-mediated ectopic expression of wild-type ALK2, CA-ALK2, or FOP-ALK2 on their ability to induce mineralization of hMSCs. BMP-6 can potentiate this response in osteogenic medium, but to make the effect by BMP-6 on FOP-ALK2-mediated mineralization more apparent, assay conditions were chosen in which BMP-6 alone had no effect. Whereas both FOP-ALK2 and CA-ALK2 were not sufficient to efficiently induce mineralization, BMP-6-induced mineralization was significantly promoted in CA-ALK2- and FOP-ALK2-expressing hMSCs (Fig. 4). Thus expression of FOP-ALK2 (and CA-ALK2) sensitizes mesenchymal cells to BMP-induced biologic responses.

Figure 4. Ectopic expression of FOP-ALK2 sensitizes cells to BMP-induced osteoblast differentiation. Human mesenchymal stem cells were transduced with lentviruses; empty vector or HA-tagged wild-type, CA-, and FOP-ALK2 in the absence or presence of 100 ng/mL of BMP-6 for 15 days. Mineralization was measured histochemically with alizarin red S. (B) Ectopic expression of wild-type, CA-, and FOP-ALK2 forms was measured by blotting cell lysates with HA-tag antibody. Two nonspecific bands are shown to demonstrate equal loading.

Ectopic expression of FOP-ALK2 in hMSCs grown on CaPO4 scaffolds stimulates ectopic bone formation

One important hallmark of FOP is heterotopic ossification. To examine whether expression of FOP-ALK2 is sufficient to induce ectopic bone formation of mesenchymal cells, we assessed the ability of hMSCs virally transduced with empty vector, wild-type ALK2, or FOP-ALK2 to stimulate in vivo bone formation (Fig. 5). The hMSCs were loaded onto porous biphasic calcium phosphate scaffolds (2 to 4 mm) and subcutaneously implanted on the backs of nude mice. Prior to implantation, the hMSCs were cultured on the scaffolds for 1½ weeks in proliferation medium. Implantation of the scaffolds was performed as described previously.35 After 6 weeks, the mice were euthanized, and scaffolds were explanted and analyzed histochemically for ectopic bone formation by staining sections with basic fuchsin and methylene blue. As a positive control for the ability of hMSCs to form ectopic bone, hMSCs were treated for 4 days with 1 mM dibutyryl adenosine cyclic monophosphate (cAMP) prior to implantation; after 6 weeks in the mice, the explanted scaffolds showed that hMSCs from this donor were capable, albeit not so efficiently, of forming ectopic bone (data not shown). We observed that in this assay, FOP-ALK2-expressing hMSCs scored significantly higher in inducing ectopic bone formation than empty-vector-transduced control hMSCs (Fig. 5B). Ectopic bone formation by wild-type ALK2–expressing hMSCs was slightly inhibited. Thus FOP-ALK2 mutant–expressing hMSCs mimic heterotopic ossification of patient cells.

Figure 5. Ectopic expression of FOP-ALK2 in hMSCs grown on CaPO₄ scaffolds stimulates ectopic bone formation. hMSCs were transduced with lentviruses expressing wild-type ALK2, FOP-ALK2, or empty vector (used as control). Cells were loaded onto porous biphasic calcium phosphate scaffolds and implanted in nude mice. After 6 weeks, the mice were euthanized, and scaffolds were explanted and processed for immunohistochemical analysis. Newly formed ectopic bone in the sections was visualized by staining with basic fuchsin and methylene blue. The sections were analyzed using a light microscope, and high-resolution digital photographs were made of randomly selected sections from each scaffold graft. Yellow arrows point to newly formed bone. (B) The extent of bone formation (areas with intense pink staining) was measured relative to the total available pore area for new bone growth. Data were analyzed statistically using a Kurskal-Wallis nonparametric test.

Discussion

In this study, we demonstrate that the R206H mutation in ALK2 (FOP-ALK2) confers constitutive activity to this receptor; that is, it signals in the absence of exogenously added BMP ligands by inducing Smad1 phosphorylation and BMP transcriptional reporter activity. A possible mechanism is the lack of the FKPB12-mediated negative regulation of the FOP-ALK2 mutant versus the wild-type ALK2 receptor. Moreover, we found that ectopic expression of FOP-ALK2 in mesenchymal cells makes those cells more receptive to exogenous BMPs with respect to differentiating into functional mineralizing osteoblasts. Furthermore, expressing this mutant ALK2 in hMSCs confers an ability of these cells to contribute to newly formed bone when seeded onto calcium phosphate scaffolds.

Our data regarding the constitutive activity of FOP-ALK2 with respect to Smad1 phosphorylation and BRE-luc transcriptional reporter activity and the ability to inhibit these responses by dorsomorphin and Smad7 are consistent with recent findings reported by Fukuda and colleagues.4 However, we found that dorsomorphin is not a selective ALK2 kinase inhibitor and also has considerable inhibitory activity on many other kinases (including other ALKs), precluding its (pre)clinical use. We found that in addition to Smad7, as reported by Fukuda and colleagues,4 ectopic expression of Smad6 also was effective in antagonizing FOP-ALK2-induced constitutive BRE-luc transcriptional activity. In contrast to the enhanced BMP-induced BRE-luc activity by FOP-ALK2, wild-type ALK2 consistently downregulated this response. Ectopic expression of ALK2 may interfere with use of the endogenous ALK3 or ALK6 receptor in these cells, which are needed for efficient BMP signaling. Our findings that FOP-ALK2-expressing cells display enhanced sensitivity to differentiate toward the osteoblast lineage in response to BMP stimulation is consistent with the study of Billings and colleagues, which found that connective tissue progenitor cells isolated from the teeth of FOP patients show enhanced BMP signaling and osteogenic differentiation.37 However, in contrast to what was found in these FOP patient–derived cells, we did not observe deregulated p38 MAP kinase signaling by FOP-ALK2.

We observed that by introducing the R206H mutation in ALK2, the sensitivity of the receptor to FBBP12-mediated inhibition was decreased, providing a mechanism by which FOP-ALK2 becomes constitutively active and sensitizes cells expressing the mutant receptor to exogenous ALK2 ligands. Groppe and colleagues have proposed this possibility previously based on in silico modeling of wild-type and FOP-ALK238; protein modeling predicted that the R206H mutation generates a pH-sensitive structural change in the GS-domain that may perturb FKBP12 interaction.

Whereas, on the one hand, we found that the FOP-ALK2 mutant in BRE-luc and Smad1 phosphorylation assays had elevated basal activity, no elevated basal activity of this mutant receptor was observed in the osteoblast differentiation and mineralization assays. However, we demonstrated that the ectopic expression of FOP-ALK2 mutant potentiated BMP-induced osteoblast differentiation and mineralization but did not further increase the BMP-induced BRE-luc and Smad1 phosphorylation. These differential effects of FOP-ALK2 on basal and BMP sensitization can be explained by different thresholds of activation in BRE-luc and Smad1 phosphorylation responses versus osteoblast differentiation and mineralization. BRE-Luc/Smad1 phosphorylation appears to require low levels of receptor activity and reaches a maximum signaling response at lower receptor activation levels than induction of ALP activity and mineralization.

Muscle injury in FOP patients triggers local heterotopic bone formation. This suggests that trauma-induced signals cooperate with FOP-ALK2 in mediating this response. Our finding that ectopic expression of FOP-ALK2 greatly sensitizes the cells to BMP signaling suggests that BMPs themselves could represent such trauma-induced signals. Muscle injury may induce local BMP production; cells from Duchenne muscular dystrophy patients express increased levels of BMP-4.39 BMP-4 was found to be upregulated in lymphocytes of FOP patients.16 Although muscle and blood specimens are impossible or at least difficult to obtain from FOP patients because sample collection may be sufficient to induce ectopic bone formation, it will be of great interest to determine muscle and circulating levels of BMP-4, BMP-6, and BMP-9 in FOP patients or in a transgenic mouse model with a FOP-ALK2 knock-in mutation during spontaneous and trauma-induced heterotopic bone formation.

At present, there are no therapies for FOP. In this study we have developed a novel relevant in vivo animal model for FOP-associated heterotopic bone formation. The FOP-ALK2 mutant–expressing hMSCs but not wild-type ALK2- or nontransfected hMSCs seeded onto calcium phosphate scaffolds induce ectopic bone formation. This assay can be used to test new pharmacologic compounds that target FOP-ALK2 mutant receptors that are causally associated with FOP. Rather than using small-molecule ATP analogues that may inhibit wild-type ALK2 and the closely related ALK1, ALK3, and ALK6, in our laboratory we are attempting to specifically target FOP-ALK2 mutant expression using antisense and siRNA-based oligonucleotides.

Disclosures

All the authors state that they have no conflicts of interest.

Acknowledgements

This work was supported by the Centre for Biomedical Genetics (PtD), the Dutch Organization for Scientific Research (NWO 918.66.606), and IOP Genomics Grant IGE07001.