Depletion of Mast Cells and Macrophages Impairs Heterotopic Ossification in an Acvr1^R206H Mouse Model of Fibrodysplasia Ossificans Progressiva

Abstract

Heterotopic ossification (HO) is a clinical condition that often reduces mobility and diminishes quality of life for affected individuals. The most severe form of progressive HO occurs in those with fibrodysplasia ossificans progressiva (FOP; OMIM #135100), a genetic disorder caused by a recurrent heterozygous gain-of-function mutation (R206H) in the bone morphogenetic protein (BMP) type I receptor ACVR1/ALK2. In individuals with FOP, episodes of HO frequently follow injury. The first sign of active disease is commonly an inflammatory “flare-up” that precedes connective tissue degradation, progenitor cell recruitment, and endochondral HO. We used a conditional-on global knock-in mouse model expressing Acvr1^R206H (referred to as Acvr1^cR206H/+) to investigate the cellular and molecular inflammatory response in FOP lesions following injury. We found that the Acvr1 R206H mutation caused increased BMP signaling in posttraumatic FOP lesions and early divergence from the normal skeletal muscle repair program with elevated and prolonged immune cell infiltration. The proinflammatory cytokine response of TNFα, IL-1β, and IL-6 was elevated and prolonged in Acvr1^cR206H/+ lesions and in Acvr1^cR206H/+ mast cells. Importantly, depletion of mast cells and macrophages significantly impaired injury-induced HO in Acvr1^cR206H/+ mice, reducing injury-induced HO volume by ~50% with depletion of each cell population independently, and ~75% with combined depletion of both cell populations. Together, our data show that the immune system contributes to the initiation and development of HO in FOP. Further, the expression of Acvr1^R206H in immune cells alters cytokine expression and cellular response to injury and unveils novel therapeutic targets for treatment of FOP and nongenetic forms of HO.

Introduction

Heterotopic ossification (HO) often diminishes movement and quality of life. Patients with the rare genetic disorder fibrodysplasia ossificans progressiva (FOP; OMIM #135100) develop severe and debilitating HO.⁽¹⁾ Episodes of HO commonly follow distinct “flare-up” events that are characterized by swelling, stiffness, pain, and warmth.⁽¹⁾ Flare-ups occur in soft connective tissues following overt trauma, such as tissue injury, surgery, or intramuscular immunizations.⁽²⁾ FOP lesions progress through stages of soft tissue response that have been defined histologically⁽²⁾: early inflammatory connective tissue destruction is followed by fibroproliferative cell expansion, and HO through an endochondral process.

FOP is caused by heterozygous mutations in Activin receptor A type I (ACVR1), a BMP type I receptor.^(3,4) More than 95% of FOP patients possess the same R206H mutation in ACVR1, which confers gain-of-function BMP pathway signaling that is both ligand-independent and ligand-responsive.^(4,5) ACVR1 regulates early chondrogenic fate and Acvr1^R206H enhances chondrogenesis in chondro-osseous progenitor cells that contribute to HO in vivo.⁽⁶⁾

Understanding the cellular and molecular events that regulate disease progression in FOP will help to identify targets and appropriate timing for therapeutic intervention. Tissue inflammation has been associated with the onset of HO in FOP, and biopsies of patient lesions reveal robust immune cell infiltration even in the absence of an overt inflammatory triggering event.⁽²⁾

Despite intense interest in FOP and in HO without known genetic predisposition,^(7–15) a detailed understanding of HO initiation and progression remains limited. Skeletal muscle is a commonly affected site of HO formation.^(2,16) The normal cellular and molecular response to skeletal muscle injury and subsequent tissue repair has been well characterized in mouse models of tissue injury and FOP,^(11,17) and the earliest events, which involve a cellular and molecular inflammatory response, mimic the early events leading to HO in FOP patients.^(2,17) Although skeletal muscle normally repairs and regenerates following injury, in FOP the terminal response to injury is formation of HO. This suggests that the disease-causing ACVR1 R206H mutation and dysregulated BMP pathway signaling diverts the appropriate injury response and repair mechanisms away from muscle regeneration and toward bone formation.

The BMP signaling pathway plays a seminal role in inflammatory responses,^(18–21) and the immune system, most notably the lymphoid and myeloid lineage cells that invade rapidly in response to injury, have been implicated in triggering FOP disease progression.^(13,22) The dysregulated BMP pathway signaling in FOP caused by the ACVR1 R206H mutation^(5,6) may amplify the early immune response to injury and establish a permissive tissue microenvironment leading to HO.

In the present study, we conducted in vitro and in vivo experiments to investigate immune cell contributions to HO development in FOP. To investigate the inflammatory responses through lesion progression, we conducted a detailed analysis of the skeletal muscle injury response using a knock-in Acvr1^cR206H/+ FOP mouse model that faithfully reproduces FOP clinical phenotypes.^(12,23) We investigated BMP pathway signaling and inflammatory cytokine expression in primary mast cells and macrophages, two immune cell populations abundantly present in early FOP lesions, and examined the contributions of mast cells and macrophages to HO in vivo in mast cell-depleted and macrophage-depleted Acvr1^cR206H/+ mice. Together, our data demonstrate a significant contribution of the immune system to the initiation and progression of HO in FOP, show that expression of Acvr1^R206H in immune cells alters cytokine expression and cell response to injury, and identify novel therapeutic targets for treatment of FOP and other forms of HO.

Materials and Methods

Animal Care and Use

A conditional-on knock-in mouse Acvr1^[R206H]FlEx [ref. 23] was used to generate doxycycline-inducible global allele expression Acvr1^{[R206H]FlEx/+};Gt(ROSA)26Sor^{tm1(rtTA*M2)Jae};Tg(tetO-Cre)1Jaw mice (which we refer to as Acvr1^cR206H/+), as described.⁽¹²⁾ Acvr1^+/+ controls were littermates that did not contain an Acvr1^cR206H allele (as verified via PCR genotyping). For mast cell–deficient Acvr1^cR206H/+ mice, Acvr1^cR206H/+ mice were mated with heterozygous White Sash;B6.Cg-Kit^W-sh/HNihrJaeBsmGlliJ (The Jackson Laboratory, Bar Harbor, ME, USA; Stock #012861) to generate compound heterozygous Acvr1^cR206H/+;B6.Cg-Kit^W-sh/HNihrJaeBsmGlliJ mice; offspring crosses generated homozygous mast cell–deficient mice (hereafter referred to as Acvr1^cR206H/+;c-Kit^W-Sh/W-sh). All animal procedures were reviewed and approved by the Institutional Animal Care and Use Committee at the University of Pennsylvania.

Skeletal muscle injury with cardiotoxin

Acvr1^cR206H/+ mice were fed a doxycycline chow diet (625 mg/kg doxycycline chow; Envigo Laboratories, Madison, WI, USA; TD.01306) for 5 days prior to cardiotoxin injection to induce mutant gene expression. Hamstring muscles of mice (at 4 weeks of age) were injured by injecting 50 μL of 20 μM cardiotoxin from Naja mossambica mossambica (Sigma-Aldrich, St. Louis, MO, USA; C9759) into hamstring muscles. Mice were euthanized by CO₂ asphyxiation and whole hindlimbs were collected at days 0, 1, 2, 3, 4, 5, 6, 7, 10, and 14 postinjection. Day 0 samples were collected without cardiotoxin injection.

Histology and immunohistochemistry

Tissue samples were fixed in 4% paraformaldehyde for 24 hours and decalcified using Immunocal (Decal Chemical Corporation, Tallman, NY, USA) for 3 days, embedded in paraffin, and sectioned serially at 5 μm. Deparaffinized sections were stained with Alcian Blue/Orange G/Hematoxylin/Eosin. Mast cells were detected by combined eosinophil-mast (C.E.M.) staining (KTCEM; American MasterTech, Lodi, CA, USA).

For immunohistochemistry, deparaffinized sections were treated for antigen retrieval with 10 mM sodium-citrate buffer (pH 6.0) at 95°C for 20 min (for cytokine and chemokine detection) or with Digest-All 2 Trypsin (Thermo Fisher Scientific, Waltham, MA, USA; 003008) at 37°C for 10 min (for immune cell detection). Endogenous peroxidase activity was quenched with 3% hydrogen peroxide. Sections were blocked using Background Buster (Innovex Biosciences, Richmond, CA, USA; NB306), incubated with primary antibodies overnight at 4°C, then with appropriate host horseradish peroxidase (HRP) secondary antibody, DAB detection (SuperPicture Polymer 879263; Thermo Fisher Scientific), and hematoxylin counterstain. Results were compared to negative controls processed without primary antibody. Primary antibodies used were as follows: phosphorylated-Smad1/5/8 (Cell Signaling, Danvers, MA, USA; 13820; 1:50 dilution), myeloperoxidase (Abcam, Cambridge, MA, USA; ab139748; 1:200 dilution), F4/80 (Abcam; ab111101; 1:500 dilution), CD3 (Abcam; ab16669; 1:50 dilution), TNFα (Abcam; ab34674; 1:100 dilution), IL-6 (Abcam; ab6672; 1:400 dilution), IL-1β (Abcam; ab9722; 1:400 dilution), MCP-1 (Abcam; ab25124; 1:200 dilution), IL-13 (Abcam; ab106732; 1:400 dilution), Activin A (LS Bio, Seattle, WA, USA; LS-C195902; 1:400 dilution). We analyzed sections from three to four independent mice (Acvr1^cR206H/+ or Acvr1^+/+) per day, per genotype, per protein. Representative images from each sample set are shown.

Cell culture

To obtain primary mast cells, whole bone marrow from femurs and tibias of Acvr1^cR206H/+ and Acvr1^+/+ mice was cultured in suspension for 6 weeks in mast cell maturation medium (RPMI 1640, 10% FBS, 1X Penicillin/Streptomycin, 25 mM HEPES, 1X MEM non-essential amino acids, 1X GlutaMAX, 100 mM sodium pyruvate, 50 μM β-mercaptoethanol) with added recombinant murine Stem Cell Factor (Peprotech, Rocky Hill, NJ, USA; 250–03; 15 ng/mL final concentration) and recombinant murine IL-3 (Peprotech; 213–13; 12 ng/mL final concentration).⁽²⁴⁾ Mast cell purity was verified by flow cytometry, gating for c-kit receptor/CD117 and Fc_εRI_α receptor double-positive cells⁽²⁵⁾ (see Supporting Fig. 1). Mast cell degranulation was quantified using a degranulation assay kit (see Supporting Materials and Methods).

To obtain primary macrophages, 1.5 mL aged 4.0% thioglycolate solution (Sigma-Aldrich; T9032) was injected intraperitoneally into Acvr1^cR206H/+ and Acvr1^+/+ mice.⁽²⁶⁾ After 5 days, peritoneal exudate was collected from euthanized mice, centrifuged at 300g for 5 min, and cells plated in macrophage growth medium (RPMI 1640, 5% FBS, 10 mM HEPES, 1X Penicillin/Streptomycin, 1X GlutaMAX, 23.3 mM sodium bicarbonate, 50 μM β-mercaptoethanol).

For BMP pathway signaling assay, 1 × 10⁶ mast cells or 5 × 10⁵ macrophages were cultured in Tyrode’s solution for 12 hours or 2 hours, respectively, and then treated with indicated amounts of rhBMP4 (R&D Systems, Minneapolis, MN, USA; 314-BP-050) for 1 hour. For mast cell inflammatory stimulation, mast cell cultures were treated with 100 μM Substance P (Sigma-Aldrich; S6883) plus 15 ng/mL of rhBMP4 for 1 hour. For macrophage inflammatory stimulation, macrophages were treated with 1 ng/mL lipopolysaccharides (Sigma-Aldrich; L2630) plus 15 ng/mL of rhBMP4 for 1 hour.

Flow cytometry

To confirm primary bone marrow–derived mast cell purity, mast cells were incubated with c-Kit-APC (BioLegend, San Diego, CA, USA; 105812; 1:150 dilution) and Fc_εRI_α-PE (BioLegend; 134308; 1:150 dilution) antibodies for 30 min and then run on a BD FACSCalibur instrument (BD Biosciences, Franklin Lakes, NJ, USA). Double-positive cells expressing c-Kit and Fc_εRI_α receptors were gated as mast cells (Supporting Fig. 1).

Immunoblot analysis

Total cell protein was recovered using RIPA lysis buffer (Thermo Fisher Scientific; 89900) containing Halt Protease and Halt Phosphatase Inhibitor Cocktails (Thermo Fisher Scientific; 78440) and quantified using Pierce BCA Protein Assay Kit (Thermo Fisher Scientific; 23225). Proteins were electrophoresed through 4–12% Bolt Bis-Tris polyacrylamide gels (Thermo Fisher Scientific; NW04127BOX) and transferred to nitrocellulose (Bio Rad; 1704271). Membranes were blocked in 5% BSA, incubated with phosphorylated-Smad1/5/8 primary antibody (Cell Signaling Technology; 9511; 1:750 dilution) at 4°C overnight, then with GAPDH primary antibody (Cell Signaling Technology; 5174; 1:5000 dilution) at room temperature for 1 hour. Bound primary antibodies were detected with antirabbit horseradish peroxidase-conjugated secondary antibody (Cell Signaling Technology; 7074; 1:10,000 dilution) at room temperature for 1 hour. Detected proteins were imaged with WesternSure PREMIUM Chemiluminescent Substrate (LI-COR, Lincoln, NE, USA; 926–95000) and quantified (LI-COR C-DiGit Blot Scanner and Image Studio software).

RNA isolation and real-time RT-PCR

RNA was isolated from mast cells and macrophages using TRIzol (Thermo Fisher Scientific; 15596026) and quantified. cDNA was synthesized using High Capacity RNA-to-cDNA reagents (Applied Biosystems, Foster City, CA, USA; 4387406). Real-time quantitative PCR reactions contained forward/reverse primers (0.37 μM). cDNA (1:5 dilution), and Fast SYBR Green PCR Master Mix (Applied Biosystems; 4385612); each sample was analyzed in triplicate. Target gene mRNAs were quantified from standard curves and normalized to Gapdh. Forward and reverse primer sequences are in Supporting Table 1.

Clodronate-liposome-mediated macrophage depletion

To deplete macrophages from mice,^(27,28) clodronate-liposomes (100 μL/10 g bodyweight) were injected intraperitoneally. Clodronate-liposomes were obtained from the lab of Dr. Nico van Rooijen (http://www.clodronateliposomes.com; The Netherlands). Confirmation of macrophage depletion was assessed by flow cytometry (see Supporting Materials and Methods).

Micro–computed tomography analysis

As indicated in Supporting Fig. 9, mice were injected with cardiotoxin at 4 weeks of age, and samples were collected at 17 days postinjury and fixed for 24 hours in 4% paraformaldehyde. HO was detected and quantified in high-resolution, cross-sectional reconstructed images of paraformaldehyde (PFA)-fixed hind limbs using a micro–computed tomography (μCT) VivaCT40 imager (Scanco Medical AG, Brüttisellen, Switzerland) at a source voltage of 55 kV, a source current of 145 μA, and an isotropic voxel size of 19.0 μm. Three-dimensional renderings to quantify HO were reconstructed using Scanco μCT V6.1 software from regions of interest that were free-hand drawn around HO every 5 to 10 reconstructed slices and then interpolated for total volume. Users ensured that the HO region of interest did not include skeletal bone. Thresholding values for HO detection ranged from 240 to 1000 Hounsfield units.

Statistical analysis

Data were analyzed statistically using GraphPad (La Jolla, CA, USA) Prism 7 software (unpaired, two-sided, equal variance Student’s t test; two-way ANOVA with Sidak’s multiple comparisons test; one-way ANOVA with Tukey’s multiple comparisons test); values are expressed as the mean values ±SE. Statistical significance was p < 0.05.

Results

The Acvr1 R206H mutation causes early divergence from the normal skeletal muscle repair program

In response to cardiotoxin-induced skeletal muscle injury, mice expressing Acvr1^R206H form heterotopic endochondral bone over a 2-week time course.⁽¹²⁾ To investigate immune cells in skeletal muscle following injury, we examined the progression of injury response and tissue repair of mutant Acvr1^cR206H/+ and control Acvr1^+/+ mice in detail (Supporting Fig. 2; Fig. 1).

Fig. 1.

Progression of heterotopic ossification formation following injury in Acvr1^cR206H/+ mice. Skeletal muscle tissue sections from Acvr1^cR206H/+ and Acvr1^+/+ mice from indicated days following cardiotoxin injury were detected for cartilage (Alcian Blue), bone (Orange G), cell nuclei/immune cells (Hematoxylin), and skeletal muscle (Eosin). (A) Skeletal muscle histology is similar in mutant and control mice before injury (day 0). (B, C) Skeletal muscle degradation and increasingly abundant immune cells are present by days 1 to 2 in both mutant and control mice. (D) At day 3, immune cells increased further, but to a greater extent in Acvr1^cR206H/+ tissue. (E–G) In control tissue, muscle repair progresses through the fibroproliferative stage and clearance of immune cells (days 4 to 6); however, GAG proteins are detected (Alcian Blue) in Acvr1^cR206H/+ tissue. (H–J) In control tissue, skeletal muscle repair continues and is completed over days 7 to 14. In Acvr1^cR206H/+ lesions, endochondral ossification proceeds with maturation to hypertrophic chondrocytes and mineralized bone. Ectopic bone with adjacent regions of mature cartilage and remaining fibroproliferation at day 14 is shown. Representative images from n = 4 mice per day, per genotype are shown. Scale bar = 100 μm for all images. Dotted lines indicate tissue areas as noted. M = skeletal muscle; I = immune infiltration; FP = fibroproliferation; C = chondrocytes; B = bone.

Uninjured mutant tissues have normal skeletal muscle morphology (day 0; Fig. 1A). At the initial stages of wound healing (days 1 to 2 following injury) in control Acvr1^+/+ tissue, abundant immune cells and degradation of damaged tissue occurs and this response was also observed in mutant tissue (Fig. 1B, C). Control tissues subsequently showed progressive clearance of immune cells by day 6 (days 3 to 6; Fig. 1D–G). By contrast, Acvr1^cR206H/+ tissues showed further muscle degradation with increased immune cells at day 3 (indicated by dense hematoxylin staining; Fig. 1D), followed by robust and prolonged presence of fibroblasts and accumulation of Alcian Blue-positive glycosaminoglycan, an extracellular matrix component associated with chondrogenesis⁽²⁹⁾ (Fig. 1E–G).

By day 7 postinjury, Acvr1^+/+ lesions exhibited further immune cell clearance and presence of regenerating muscle (Fig. 1H); however, Acvr1^cR206H/+ lesions showed a persistent presence of fibroblast cells together with ectopic pre-hypertrophic chondrocytes (Fig. 1H). By day 10, immune cells were absent from Acvr1^+/+ tissue, with skeletal muscle repair reaching near completion at day 14 (Fig. 1I, J). By contrast, heterotopic endochondral ossification had significantly progressed from day 6 to day 7 (Fig. 1H) in Acvr1^cR206H/+ tissue, with abundant hypertrophic chondrocytes and bone matrix present at day 10 (Fig. 1I) and more mature heterotopic bone at day 14 (Fig. 1J). These data support that although the initial response to skeletal muscle injury in Acvr1^cR206H/+ mice appears to parallel that of Acvr1^+/+ mice, the Acvr1^cR206H/+ response subsequently diverges toward fibroproliferation, chondrogenesis, and ectopic bone. This progression toward bone formation occurs in the context of a robust and prolonged immune cell response.

BMP signaling is increased in Acvr1^cR206H/+ posttraumatic lesions

We examined the relative levels of BMP pathway signaling in Acvr1^cR206H/+ and Acvr1^+/+ lesions following skeletal muscle injury. Consistent with the gain-of-function R206H mutation in Acvr1,^(5,30) increased phosphorylated-Smad 1, 5, and 8 (pSmad1/5/8) protein was detected in Acvr1^cR206H/+ lesions compared to Acvr1^+/+ at all stages (Supporting Fig. 3A). Elevated pSmad1/5/8 protein was most prominent in the nuclei of fibroproliferative cells and prehypertrophic chondrocytes, consistent with accelerated chondrogenesis resulting from the R206H mutation.⁽⁶⁾ We also detected higher pSmad1/5/8 expression in the bone marrow in Acvr1^cR206H/+ mice (Supporting Fig. 3B), particularly in day 2 sections, confirming that activation of the conditional Acvr1^cR206H allele occurs widely and is active in cell types known to be responsive to BMP signaling. Enhanced BMP pathway signaling in the bone marrow is of particular note since bone marrow is the primary adult hematopoietic tissue and source of hematopoietic stem cells from which mature immune cell populations are derived.⁽³¹⁾

Immune cell numbers are elevated and prolonged in Acvr1^cR206H/+ tissue following injury

Previous studies using patient biopsies and FOP mouse models recognized that multiple immune cell types and inflammatory mediators are present in developing HO lesions, with most attention given to cells of the innate immune system such as neutrophils, macrophages, and mast cells.^{(8,11,12,22,32)} However, these studies were limited to the earliest stages of lesion progression. To examine the immune cell composition in postinjury tissue in detail, we investigated the cellular inflammatory response to skeletal muscle injury during early-stage and intermediate-stage lesion progression.

Neutrophils are recruited within hours after tissue injury in response to damage-associated molecular pattern (DAMP) pathway activation⁽³³⁾ and then release proteases that degrade damaged tissue in preparation for clearance by macrophages.⁽³⁴⁾ Neutrophils were significantly elevated in early Acvr1^cR206H/+ lesions compared to Acvr1^+/+ (Fig. 2A), reaching peak numbers at day 2 and persisting at higher levels at day 4.

Fig. 2.

Immune cell numbers are elevated and prolonged in Acvr1^cR206H/+ lesions. Specific immune cell populations were detected and quantified in injured Acvr1^cR206H/+ and Acvr1^+/+ skeletal muscle over time; early-stage (day 2), intermediate-stage (day 6), and late-stage (day 14) postinjury are shown. (A) Neutrophils were detected by myeloperoxidase IHC. (B) Monocytes/macrophages were detected with F4/80 IHC. (C) Mast cells were detected by C.E. M. stain kit, and indicated by red arrows. (D) T cells were detected with CD3 IHC. Cells were quantified from three fields of view per independent sample; n = 3 for neutrophils, monocytes/macrophages, T cells; n = 4 for mast cells. Representative images are shown. Scale bar = 50 μm. Data shown are mean values ± SE; two-way ANOVA with Sidak’s multiple comparisons test comparing Acvr1^cR206H/+ versus Acvr1^+/+ per day was performed; *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. IHC = immunohistochemistry.

Tissue-resident macrophages rapidly respond to neutrophils, phagocytosing damaged tissue and debris generated by neutrophils, and secreting cytokines and chemokines that recruit progenitor cells to facilitate wound healing.^(35,36) Additionally, circulating HSC-derived monocytes are recruited to inflamed tissues and can differentiate into macrophages.⁽³⁶⁾ F4/80-positive monocytes/macrophages were significantly increased in early-stage and intermediate-stage (days 2 to 6) Acvr1^cR206H/+ lesions compared to Acvr1^+/+ (Fig. 2B). Although numbers of monocytes/macrophages in control tissue returned to preinjury levels by day 6, these cells persisted in Acvr1^cR206H/+ lesions.

Mast cells contribute to diverse biological processes, including wound repair, fibrosis, pain pathophysiology, and allergy,^(37–40) and have been previously reported in biopsy samples from FOP patients⁽⁸⁾ and in FOP mouse models.^(11,12) Consistent with previous studies, we determined that mast cells were significantly greater in Acvr1^cR206H/+ lesions, with elevated levels of mast cells evident by day 4 following injury and persisting at high levels through day 14 and heterotopic bone formation (Fig. 2C).

In addition to cells of the innate immune system, T cells have also been reported in FOP lesions,^(11,32) and a BMP4 overexpression mouse model that lacked mature lymphocytes formed reduced HO.⁽⁴¹⁾ At day 3 postinjury, T cells were more than twofold higher in Acvr1^cR206H/+ lesions compared to Acvr1^+/+, similar to myeloid lineage cells, and higher numbers of T cells persisted through the progression to heterotopic bone (Fig. 2D).

In summary, both myeloid and lymphoid immune cell populations are increased in Acvr1^cR206H/+ lesions, indicating that Acvr1^cR206H significantly amplifies the cellular inflammatory response following tissue injury.

The proinflammatory cytokine response is elevated and prolonged in Acvr1^cR206H/+ lesions and in Acvr1^cR206H/+ mast cells

A maximal inflammatory response depends on the synthesis and secretion of proinflammatory cytokines and chemokines that amplify and sustain an ongoing immune response and act as chemoattractant molecules for progenitor cells that participate in tissue repair.^(42–45) To determine whether the elevated and prolonged immune cell response in Acvr1^cR206H/+ lesions (Fig. 3) is accompanied by general or specific differences in proinflammatory cytokines and chemokines, we examined expression in early-stage, intermediate-stage, and late-stage Acvr1^cR206H/+ and Acvr1^+/+ lesions.

Fig. 3.

Proinflammatory cytokine expression is elevated and prolonged in Acvr1^cR206H/+ lesions and primary Acvr1^cR206H/+ mast cells. Specific proinflammatory cytokines were detected by immunostaining of injured tissues from early-stage (day 2), intermediate-stage (day 6), and late-stage (day 14) Acvr1^cR206H/+ and Acvr1^+/+ lesions. (A) TNFα, (B) IL-6, and (C) IL-1β. n = 3 to 4 per day, per genotype; representative images are shown. Scale bar = 100 μm for all images. (D, E) mRNA expression of specific proinflammatory cytokines in mast cells and macrophages was detected by qRT-PCR following mast cell treatment with 100 μM Substance P or macrophage treatment with 1 ng/mL lipopolysaccharides, then 15 ng/mL BMP4 for 1 hour. Data were normalized to Gapdh and are shown as mean values ± SE; Student’s t test compared expression in Acvr1^cR206H/+ versus Acvr1^+/+; *p < 0.05. ns = not significant.

The proinflammatory cytokines TNFα, IL-1β, and IL-6 were detected at higher levels in Acvr1^cR206H/+ lesions compared to Acvr1^+/+ throughout lesion progression (Fig. 3). TNFα and IL-6 were robustly present and detected as early as day 2 postinjury in Acvr1^cR206H/+ lesions (Fig. 3A, B), whereas IL-1β was more slightly increased (Fig. 3C). These cytokines were detected throughout the injured tissues. Lower magnification images can be seen in Supporting Fig. 4.

Immune cells are a major source of secreted cytokines and chemokines during inflammation.⁽⁴²⁾ To examine the effect of enhanced BMP signaling conferred by the gain-of-function ACVR1 R206H mutation on the inflammatory potential of immune cells, we examined primary murine mast cells and macrophages in vitro. We examined BMP pathway signaling competency in Acvr1^+/+ mast cells and macrophages, detecting expression of a broad panel of BMP/TGF-β type I and II receptor mRNAs in both cell types, including Acvr1, the type I receptor mutated in FOP (Supporting Table 2). BMP pathway activity, quantified by pSmad1/5/8 protein levels, was increased in both mast cells and macrophages expressing the Acvr1^cR206H allele compared with Acvr1^+/+ (Supporting Fig. 5A, B) with relative levels of pSmad1/5/8 between Acvr1^cR206H/+ and Acvr1^+/+ similar to that reported previously.⁽⁶⁾

BMP pathway activation has been previously reported to enhance the inflammatory state of multiple immune cell populations.^{(18,21,46,47)} To investigate whether expression of the Acvr1^cR206H allele in mast cells and macrophages confers enhanced inflammatory signaling, we examined proinflammatory cytokine expression. Acvr1^cR206H/+ mast cells exhibited significantly elevated mRNA expression of TNFα and IL-6 and a trend toward elevated IL-1β compared with Acvr1^+/+ mast cells (Fig. 3D). No differences in cytokine expression between Acvr1^cR206H/+ and Acvr1^+/+ macrophages were detected (Fig. 3E).

Degranulation is a directed response to inflammatory stimulation in granulocytes, including mast cells,⁽⁴⁸⁾ releasing numerous inflammatory mediators, including TNFα and other cytokines.⁽⁴⁹⁾ Quantification of degranulation showed elevated activity in Acvr1^cR206H/+ mast cells compared to controls (Supporting Fig. 6).

MCP-1 is a chemokine abundantly produced by white adipose tissue⁽⁵⁰⁾ that functions to recruit monocytes and other immune cells,⁽⁵¹⁾ and was previously proposed as a predictive biomarker for blast injury patients who developed HO.⁽⁵²⁾ In both Acvr1^cR206H/+ and Acvr1^+/+ lesions, MCP-1 was mainly localized to white adipose adjacent to skeletal muscle, with generally low detection in the lesions throughout their progression (Fig. 4A). IL-13 is a proinflammatory cytokine with a well-established role in fibrosis pathology^(53,54) and, like MCP-1, was suggested as a predictive biomarker for blast injury patients who developed HO.⁽⁵⁵⁾ IL-13 expression was relatively low and equivalent in Acvr1^cR206H/+ and Acvr1^+/+ lesions (Fig. 4B). Lower magnification images can be seen in Supporting Fig. 7; negative control sections are in Supporting Fig. 8.

Fig. 4.

Variable expression of MCP-1, IL-13, and Activin A in Acvr1^cR206H/+ and Acvr1^+/+ lesions. Sections from early-stage (day 2), intermediate-stage (day 6), and late-stage (day 14) Acvr1^cR206H/+ and Acvr1^+/+ lesions were immunostained for expression of multiple proinflammatory cytokines. Immunohistochemistry was performed to detect (A) MCP-1, (B) IL-13, and (C) Activin A expression. (D) Higher magnification insets demonstrating fibroblast and chondrocyte Activin A expression in Acvr1^cR206H/+ lesions are shown, indicated by dotted rectangles in C. n = 3 to 4; representative images per day, per genotype, are shown. Scale bar = 100 μm for all images in A–C; scale bar = 50 μm for all images in D.

Activin A is a TGF-β superfamily ligand involved in inflammatory stimulation and regulation of several immune cell populations.⁽⁵⁶⁾ Recently, Activin A was reported to aberrantly activate BMP pathway signaling through Acvr1^R206H and contribute to HO progression in FOP.⁽²³⁾ We detected higher Activin A expression in intermediate-stage and late-stage Acvr1^cR206H/+ lesions relative to controls (Fig. 4C), with Activin A expression in fibroblastic regions as well as at ectopic chondrocytes (Fig. 4D).

These data reveal an elevated and prolonged proinflammatory response in Acvr1^cR206H/+ tissues in response to injury, as well as in Acvr1^cR206H/+ mast cells, with higher expression of multiple, but not all, proinflammatory cytokines and chemokines investigated.

Depletion of mast cells and macrophages significantly impairs injury-induced heterotopic ossification development in Acvr1^cR206H/+ mice

The two immune cell populations previously implicated in HO development are mast cells and macrophages.^(8,15,41,57) Given the increased numbers and prolonged presence of mast cells and macrophages in FOP lesions (Fig. 2B, C), as well as their relevance in non-FOP HO disorders,⁽¹³⁾ we investigated the contribution of mast cells and macrophages to injury-induced HO by using immunodeficient Acvr1^cR206H/+ mouse lines.

The homozygous mutant White Sash c-Kit^W-sh/W-sh mouse fully lacks mast cells due to a 3-megabase (mb) inversion in the upstream regulatory region of the c-Kit transcriptional start site.^(58,59) The heterozygous mutant c-Kit^W-sh/+ mouse exhibits partial mast cell deficiency.⁽⁵⁹⁾ Importantly, the c-Kit^W-sh/W-sh mouse exhibits no other immune cell deficiencies.⁽⁵⁹⁾ To deplete mast cells in the context of global expression of the FOP mutation, we generated Acvr1^cR206H/+;c-Kit^W-sh/W-sh and Acvr1^cR206H/+;c-Kit^W-sh/+ mice. To deplete macrophages, Acvr1^cR206H/+ mice were injected with clodronate liposomes (Clo), which selectively induce apoptosis in mature macrophages^(27,28) (Supporting Fig. 9). This pharmacological approach allowed generation of combined mast cell–deficient and macrophage-deficient mice by injecting Acvr1^cR206H/+;c-Kit^W-sh/W-sh mice with clodronate liposomes. Macrophage depletion was confirmed via flow cytometry (Supporting Fig. 10A, B) and mast cell depletion was confirmed via histological staining (Supporting Fig. 11).

Compared to Acvr1^cR206H/+ mice, Acvr1^cR206H/+;c-Kit^W-sh/+ mice consistently had a moderate, but statistically insignificant, decrease in HO volume (Fig. 5A, B). However, fully mast cell–deficient Acvr1^cR206H/+;c-Kit^W-sh/W-sh mice exhibited a ~50% decrease in HO volume relative to Acvr1^cR206H/+ mice. Macrophage-deficient Acvr1^cR206H/+;Clo-treated mice had a similar ~50% decrease in HO. Combined deficiency of mast cells and macrophages (Acvr1^cR206H/+;c-Kit^W-sh/W-sh;Clo) resulted in a further reduction in HO, with a ~75% decrease relative to Acvr1^cR206H/+ mice (Fig. 5A, B).

Fig. 5.

Mast cell and macrophage depletion impairs formation of heterotopic ossification in Acvr1^cR206H/+ mice. Cardiotoxin-injured skeletal muscles of Acvr1^cR206H/+ mice with intact or depleted mast cells and/or macrophages were examined by μCT to detect heterotopic ossification after 17 days. Mast cells were genetically ablated in Acvr1^cR206H/+;c-Kit^W-sh/W-sh mice or partially ablated in Acvr1^cR206H/+;c-Kit^W-sh/+ mice. Macrophages were chemically depleted in Clo-treated mice. Data are compared to Acvr1^cR206H/+ (positive HO control) and Acvr1^+/+ (negative control). (A) Representative μCT 3D volume renderings showing mean HO volume per hind limb are shown. (B) Quantification of HO for each cohort (n = 5 to 20); HO per hind limb was added for total HO per mouse. Data shown are mean values ± SE; one-way ANOVA with Tukey’s multiple comparisons test compared Acvr1^cR206H/+ versus other cohorts; *p < 0.05, **p < 0.01. ns = not significant; Clo = Clodronate-liposomes.

Ectopic bone, cartilage, and fibroproliferation were reduced in Acvr1^cR206H/+;Clo and Acvr1^cR206H/+;c-Kit^W-sh/W-sh histological samples at day 17 (Supporting Fig. 11). A further reduction was observed in samples from the combined deficiency Acvr1^cR206H/+;c-Kit^W-sh/W-sh;Clo cohort (Supporting Fig. 11). PSmad1/5/8 protein was also reduced in the regions of fibroproliferative cells in mast cell and/or macrophage deficient mice (Supporting Fig. 12), suggesting that macrophages and mast cells influence BMP-pSmad1/5/8 signaling in the tissue.

Together, our data show that both mast cells and macrophages contribute to HO development.

Discussion

Inflammatory triggering events, such as intramuscular vaccinations, tissue trauma, or viral infections, induce new episodes of HO in FOP patients with high frequency,^(2,60,61) and the presence of immune cells at early stages of HO lesion formation has been noted.^(2,22) Our study reveals that multiple immune cell populations that participate in a normal tissue injury response are increased in Acvr1^cR206H/+ lesions. Additionally, these cells persist at high levels during progression to heterotopic bone instead of returning to preinjury levels as occurs during wound repair. This heightened immune cell response is accompanied by increased proinflammatory factors, and, at least in mast cells, Acvr1^cR206H alters expression of a subset of proinflammatory cytokines. Combined depletion of mast cells and macrophages in vivo further demonstrated key roles for both immune cells in promoting HO.

We observed increased numbers of immune cell populations of neutrophils, monocytes/macrophages, and mast cells in Acvr1^cR206H/+ lesions; however, we also note that additional immune cell subtypes that were not examined may have specific responses to the mutation. The monocyte/macrophage marker F4/80 used to identify tissue resident–derived and monocyte-derived macrophages is expressed on both proinflammatory, early-responding M1 and anti-inflammatory, late-responding M2 macrophages⁽⁶²⁾; additional markers, such as CD163 or CD203, could be used to distinguish the two populations.⁽⁶²⁾ In a notexin skeletal muscle injury model, the timing of macrophage depletion determined the extent of muscle fiber regeneration, highlighting the importance of immune cell subtype responses to injury.⁽⁶³⁾ CD3 is a pan-T cell marker that we used to identify all T cell subtypes⁽⁶⁴⁾; additional markers, such as CD4 and CD8, can identify T cell subtypes that may differentially contribute in the response to injury. Our study identified distinct classes of immune cell contributors to HO formation in FOP; however, future experiments using more discriminating and quantitative methods, such as fluorescence-activated cell sorting (FACS), will be important in further investigations of the contribution of immune cell subtypes at various stages during the response to injury.

The initial response of Acvr1^cR206H/+ skeletal muscle to injury is qualitatively similar to Acvr1^+/+ control tissue and the usual skeletal muscle repair process.^(65–68) Prior to injury (at day 0), no statistical differences between mutant and wild-type in the numbers of immune cell types that we examined were detected, although we cannot exclude chronic effects of the mutation prior to injury by additional criteria. Appearance of immune cells and expression of proinflammatory cytokines, such as TNFα, IL-6, and IL-1β, are observed immediately after injury in both mutant and control. By 72 hours after injury, the mutant tissue begins to diverge from normal skeletal muscle repair events with immune cells and proinflammatory cytokines persisting over time and at higher levels. We also observed increased Activin A protein expression in intermediate-stage and late-stage Acvr1^cR206H/+ lesions, particularly in the fibroproliferative cell regions, ectopic cartilage, and ectopic bone (Fig. 4C, D; Supporting Fig. 7). Activin A is expressed in multiple immune cells, including macrophages, mast cells, T cells, dendritic cells, and B cells,^(56,69) and has been shown to induce HO in FOP.^(23,70) Whether the increased Activin A protein observed in Acvr1^cR206H/+ lesions is produced directly by immune cells and whether Activin A and/or other cytokines produced by immune cells promote HO formation, either singly or in combination, requires further detailed investigations.

We determined that depletion of macrophages or mast cells in Acvr1^cR206H/+ mice reduced HO volume by about half, while depletion of both macrophages and mast cells further reduced HO volume to ~25% of positive controls. Mast cells and macrophages synthesize and secrete many of the same inflammatory cytokines, including IL-6 and TNF-α^(18,49); depleting one cell population may partly reduce inflammation, whereas depleting both macrophages and mast cells combine for a greater effect. We acknowledge that our clodronate-liposome macrophage depletion method and dose regimen may deplete both M1 and M2 macrophage populations, potentially masking differential contributions of each subtype in the response to injury and subsequent HO development. Future experiments that adjust the timing of clodronate-liposome administration to deplete M1 or M2 subtypes independently will help elucidate specific roles of each subtype to HO formation.

The depletion of mast cells and/or macrophages may also alter inflammatory signaling networks among remaining immune cells, given the significant crosstalk of mast cells and macrophages with other immune cells.^(71,72) Depletion of these two cell populations, which express the proinflammatory cytokines IL-6 and TNF-α, may reduce subsequent proinflammatory cytokine expression in remaining immune cells at the site of soft connective tissue injury, possibly contributing to the reduced HO seen in mast cell–depleted and/or macrophage-depleted Acvr1^cR206H/+ mice.

During chronic inflammatory myopathies, expression of TNFα, IL-6, and IL-1β persists and can lead to increased skeletal muscle catabolism and reduced regenerative capacity,⁽⁷³⁾ as well as increased proliferation of recruited fibroblasts.⁽⁷⁴⁾ The elevated and prolonged expression in Acvr1^cR206H/+ lesions of the same proinflammatory cytokines present in chronic inflammatory myopathies suggests that the response to injury in Acvr1^cR206H/+ tissue mimics this condition. However, not all cytokines are upregulated in response to the R206H mutation. IL-13, a cytokine associated with fibrosis, is only mildly expressed in Acvr1^cR206H/+ and Acvr1^+/+ lesions, suggesting that although robust fibroproliferation is a noted component of FOP lesion progression, it is distinct from classic fibrosis.

Proinflammatory cytokines, including TNFα, IL-6, and IL-1β, are produced by infiltrating immune cells to mediate the initial tissue response to injury.^{(34,37,73,75–77)} Under normal conditions, as expression of these proinflammatory cytokines decreases, anti-inflammatory cytokines, such as IL-4, IL-10, IL-13, and TGF-β, are upregulated to promote tissue regeneration.^(66,68) In Acvr1^cR206H/+ tissue, the response to injury involves persistent chronic inflammatory microenvironment that accompanies more tissue damage, potentiating the inflammatory response and possibly stimulating progenitor cells to become ectopic cartilage and bone.⁽⁶⁾

It will also be of interest to investigate whether altered expression or activity of anti-inflammatory cytokines occurs in Acvr1^cR206H/+ tissue and contributes to HO. IL-4 is an anti-inflammatory cytokine that is involved in transition of proinflammatory M1 macrophages to anti-inflammatory/wound healing M2 macrophages,⁽⁷²⁾ and also acts as a myoblast recruitment factor during skeletal muscle growth.⁽⁷⁸⁾ Our preliminary data suggest that IL-4 expression was mildly decreased in Acvr1^cR206H/+ lesions at day 6, compared to Acvr1^+/+ lesions (Convente and colleagues, unpublished data), but this response requires additional investigation.

The BMP signaling pathway is known to have multiple distinct roles during skeletal muscle injury and repair, several with relevance to the responses identified in Acvr1^cR206H/+ tissue and FOP. BMP ligands induce inflammatory cytokine production and edema,^(20,21) two features that frequently precede HO initiation in FOP patients.^(1,22) BMP signaling also contributes to inflammatory activation of multiple immune cell populations, including macrophages and T cells.^(18,46) BMP signalling upregulates expression of Substance P,⁽⁵⁷⁾ a neuroinflammatory peptide that stimulates mast cell degranulation, induces TNFα and IL-6 expression, enhances migration of chondrogenic/osteogenic bone marrow stromal cells, and has been observed in biopsies of early FOP patient lesions.^{(15,48,79,80)} Additionally, although BMP signaling is permissive for muscle stem cell proliferation, an early step of skeletal muscle regeneration, persistent BMP signaling reduces myogenic regenerative potential.⁽⁸¹⁾ In contrast to the normal repair program, our data suggest that elevated BMP pathway signaling throughout lesion progression in FOP induces the altered inflammatory events that disrupt the normal skeletal muscle injury response and repair program, ultimately leading to ectopic endochondral bone formation.

The present study highlights the immune system as an appealing target for therapeutic intervention in heterotopic ossification and FOP. Inflammation is the earliest recognized event in HO development. Therefore, inhibiting disease progression at this stage may limit or prevent HO. Additionally, our data identify a potential therapeutic window for anti-inflammatory drugs, as Acvr1^cR206H/+ and Acvr1^+/+ lesions appear comparable in response to injury up to 2 to 3 days postinjury. The use of immunosuppressive therapies has previously shown promise for preventing HO development in FOP, although this case report only followed a single patient.⁽⁸²⁾

Our in vivo data suggest that mast cells and macrophages are key cells for HO progression in FOP. Importantly in the context of developing therapeutics to inhibit HO formation in FOP, multiple methods that block mast cell inflammatory signaling, including targeting the Substance P neuroinflammatory pathway, all inhibited HO formation in multiple HO models.^(15,57,83) Our data showing reduction of HO postinjury in Acvr1^cR206H/+; c-Kit^W-sh/W-sh mast cell-deficient mice (Fig. 5A, B) are consistent with reduced HO formation in previous studies using a mast cell–deficient neuron-specific enolase (Nse)-BMP4;c-Kit^W-sh/W-sh mouse model of HO,⁽¹⁵⁾ and further support an integral role for mast cells in HO formation. Macrophages have been shown to produce osteoinductive signals,⁽⁸⁴⁾ promote osteogenic differentiation of mesenchymal stem cells,⁽⁸⁵⁾ inhibit osteoclastogenesis,⁽⁸⁶⁾ and promote fracture repair,⁽⁸⁷⁾ highlighting their potential relevance to HO development. We found that ablation of only mast cells or only macrophages reduced HO formation significantly but not completely, whereas ablating both resulted in enhanced inhibition, indicating that a single target may be insufficient to completely prevent HO. Notably, our results show that although inflammatory upregulation in FOP is significant, not all inflammatory factors are elevated, suggesting that specific inflammatory pathways can be targeted while maintaining a functional immune system. Thus, an optimal treatment regimen may inhibit HO formation and maintain substantial immune function in patients.

Our observations of an immune response that precedes HO in an FOP mouse model forming spontaneous (non-injury induced) HO suggests that an immune response could be a general feature of HO initiation and/or early progression.⁽¹¹⁾ An inflammatory component is associated with almost all forms of HO,⁽¹³⁾ notably nongenetic HO disorders associated with severe trauma such as invasive arthroplasties,^(88–90) combat blast injuries, and other traumatic wounds.^(91,92) The prevalence of HO in the above scenarios is common, approaching 30% following hip replacement surgery and 65% following combat blast injuries.^(92,93) Strikingly, nongenetic HO disorders and FOP have been determined to share immunological mediators. Serum levels of IL-6 were significantly elevated in combat blast injury patients who developed HO compared to those who did not develop HO.^(52,55) This is notable given the robust increase in IL-6 in Acvr1^cR206H/+ lesions and Acvr1^cR206H/+ mast cells compared to wild-type, suggesting a possible common inflammatory response during HO development in FOP and nongenetic HO disorders. Our experiments using the Acvr1^R206H genetic model of HO provide additional support for a role of IL-6 in HO and identify additional specific immune cells and inflammatory cytokines contributing to HO in the context of this BMP receptor mutation. An important question that remains to be addressed is whether these same immune cells and cytokines are necessary components of nongenetic forms of HO. Other causes of HO, including burn injury, nerve injury, and BMP overexpression,^(94–96) have been examined through mouse injury models and BMP-implant assays and can be used to develop a better understanding of the inflammatory contribution to nongenetic HO disorders and verify possible common inflammatory mediators with genetic HO.

MCP-1 was identified as a biomarker that predicted HO development in combat blast injury patients.⁽⁵²⁾ Although MCP-1 expression was equivalent between Acvr1^cR206H/+ and Acvr1^+/+ lesions, its presence in the FOP setting could contribute to the chronic inflammatory response. Although there may be unique inflammatory contributions to FOP, such as the role indicated for Activin A in disease pathology,⁽²³⁾ common inflammatory mediators across HO disorders establish the immune system as an appealing treatment target that may benefit large numbers of patients.

Our study highlights the immune system as a major contributor to the initiation and development of HO in FOP, and greatly expands the current knowledge on the immunological contributions to FOP disease etiology. Our results provide a strong foundation for future studies on the immunological contributions to FOP. Given the shared immunological features of FOP and nongenetic HO disorders, our findings may provide insight into treatment for more common nonhereditary forms of HO, which could significantly benefit a wide range of patients.