Heterotopic ossification in mice overexpressing Bmp2 in Tie2+ lineages

Abstract

Bone morphogenetic protein (Bmp) signaling is critical for organismal development and homeostasis. To elucidate Bmp2 function in the vascular/hematopoietic lineages we generated a new transgenic mouse line in which ectopic Bmp2 expression is controlled by the Tie2 promoter. Tie2^CRE/+;Bmp2^tg/tg mice develop aortic valve dysfunction postnatally, accompanied by pre-calcific lesion formation in valve leaflets. Remarkably, Tie2^CRE/+;Bmp2^tg/tg mice develop extensive soft tissue bone formation typical of acquired forms of heterotopic ossification (HO) and genetic bone disorders, such as Fibrodysplasia Ossificans Progressiva (FOP). Ectopic ossification in Tie2^CRE/+;Bmp2^tg/tg transgenic animals is accompanied by increased bone marrow hematopoietic, fibroblast and osteoblast precursors and circulating pro-inflammatory cells. Transplanting wild-type bone marrow hematopoietic stem cells into lethally irradiated Tie2^CRE/+;Bmp2^tg/tg mice significantly delays HO onset but does not prevent it. Moreover, transplanting Bmp2-transgenic bone marrow into wild-type recipients does not result in HO, but hematopoietic progenitors contribute to inflammation and ectopic bone marrow colonization rather than to endochondral ossification. Conversely, aberrant Bmp2 signaling activity is associated with fibroblast accumulation, skeletal muscle fiber damage, and expansion of a Tie2+ fibro-adipogenic precursor cell population, suggesting that ectopic bone derives from a skeletal muscle resident osteoprogenitor cell origin. Thus, Tie2^CRE/+;Bmp2^tg/tg mice recapitulate HO pathophysiology, and might represent a useful model to investigate therapies seeking to mitigate disorders associated with aberrant extra-skeletal bone formation.

Fig. 1. Constitutive endothelial Bmp2 overexpression results in aortic valve dysfunction and pre-calcification.

A Circulating Bmp2 levels detected by ELISA in WT and Tie2 ^CRE/+;Bmp2 ^tg/tg adult mice serum. B Quantification of pulmonary acceleration time (PAT, left panel), and PAT-ejection time ratio (PAT/PET, right panel) measured by ultrasound on 16-week-old WT and Tie2 ^CRE/+;Bmp2 ^tg/tg mice. C Quantification of the aortic valve velocity (AoV Mean and Peak Vel), and pressure gradient (AoV Mean and Peak Grad) measured by ultrasound. D Representative images of acquired data of the aortic velocity peaks detected by ultrasound in WT (≈1000 mm/s) and tg animals (≈1200 mm/s). E Top panels: Masson trichromic staining on consecutive sections of aortic valve from 18-week-old WT and Tie2 ^CRE/+;Bmp2 ^tg/tg mice. Chondrocyte island (arrow) in aortic annulus at the base of the leaflet. Bottom panels: Localization of lipid droplets (arrowheads) identified by Oil Red O staining. Unpaired t test, two tails, mean ± SD *P < 0.05; ***P < 0.001; ****P < 0.0001; ns, non-significant. Scale bar 200 µm.

Fig. 2. Constitutive Tie2-driven Bmp2 expression causes HO lesions in mice.

A Nano PET-computed tomography (CT) images of 16-week-old WT and Tie2 ^CRE/+;Bmp2 ^tg/tg mice showing ectopic bone lesions close to ribs, scapulae, and neck (red arrows, ribs, rb; nk, neck; dorsal vertebrae, dv). B H&E staining on sections of skeletal muscle of the hindlimbs of Tie2 ^CRE/+;Bmp2 ^tg/tg mice showing histological features typical of HO lesions. a. Evidence of inflammation in HO lesion. a’ Damaged skeletal muscle fibers with central nuclei (arrows), mononuclear infiltration (black arrowheads), and fat cells (white arrowheads). b, b’ Area of massive fibroblast accumulation. c Ectopic bone in skeletal muscle (arrows) next to tibia. c’ Chondro-osteogenic areas with chondrocytes (white arrowhead), and osteoblasts (black arrowheads). d, d’ Mature ectopic bone with colonizing bone marrow cells (arrows), chondrocytes (white arrowhead), and osteoblasts (black arrowhead). C Top panels: Alcian blue staining of sections of WT and Tie2 ^CRE/+;Bmp2 ^tg/tg knee joint. a, a’ In WT, chondrocyte tissue (in blue) is located in the epiphysis region, tip of the bone (black arrowheads), and head of the fibula (arrow). b, b’ In Tie2 ^CRE/+;Bmp2 ^tg/tg, chondrocyte tissue is located in the epiphysis region at the tip of the bone (black arrowheads), accumulated in connective tissue (arrow) of the meniscus and head of the fibula (white arrow) and chondrogenic areas (white arrowhead) inside the skeletal muscle. Bottom panels: Alizarin red staining of WT and Tie2 ^CRE/+;Bmp2 ^tg/tg knee joint. c, c’ In WT, osteogenic tissue (in red) is located in the epiphysis region and tip of the bone complementary to chondrogenic areas (white arrowheads). d, d’ Intense staining in the head of Tie2 ^CRE/+;Bmp2 ^tg/tg joints (white arrowheads in (d’)). Extra ossification inside the skeletal muscle and head of the fibula (arrows in (d)), and the meniscus (black arrowhead). Scale bar 200 µm.

Fig. 3. Hematopoietic stem cells contribute to HO in Tie2^CRE/+;Bmp2^tg/tg mice.

A Quantification of bone marrow (BM) cellularity. Left, total BM (combined BM of transgenic (tg) and ectopic bone (EB)) cellularity is increased in Tie2 ^CRE/+;Bmp2 ^tg/tg mice. Right, cellularity of WT BM, tg BM, and ectopic BM. Ectopic BM cellularity is highly variable. B Quantification of granulocyte, monocyte colony forming units (CFU_GM). Left, total BM (tg + EB) CFU-GM was increased in Tie2 ^CRE/+;Bmp2 ^tg/tg mice. Right, ectopic BM can give rise to CFU-GMs. C Quantification of the FACS-isolated Lin^-Sca1⁺cKit⁺ (LSK) population. Left, the LSK population is increased in Tie2 ^CRE/+;Bmp2 ^tg/tg BM. Middle, LSK progenitors in ectopic BM are detectable and below normal. Right, the bulk (93.6% in tg and 93.2% in ectopic BM) of the LSK population in Tie2^Cre/+;Bmp2^tg/tg BM is GFP⁺. D Quantification of the FACS-isolated pro-inflammatory Cd11b⁺Gr1⁺ cells from BM. Left, the Cd11b⁺Gr1⁺ population is increased in total Tie2 ^CRE/+;Bmp2 ^tg/tg BM (tg + EB). Right, varying presence of Cd11b⁺Gr1⁺ cells in ectopic BM. E. FACS-isolated pro-inflammatory Cd11b⁺Gr1⁺ cells from PB. Left, Cd11b⁺Gr1⁺ cells are marginally increased in Tie2 ^CRE/+;Bmp2 ^tg/tg PB with HO. Right, Cd11b⁺Gr1⁺ population at different time points, 13, 19, and 21 weeks. At 21 weeks, when mice present HO, pro-inflammatory cells increase in Tie2 ^CRE/+;Bmp2 ^tg/tg. F Representative eosin staining of CFU-fibroblast (Fb). G Representative alkaline phosphatase staining of CFU-osteoblast (Ob). H Quantification showing increased CFU-Fb and CFU-Ob in Tie2 ^CRE/+;Bmp2 ^tg/tg BM. A, B, C, D, H, for WT and Tie2 ^CRE/+;Bmp2 ^tg/tg tg or tg + EB BM groups, unpaired t test, two tails, mean ± SD *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001, ns non-significant. E Right panel, two-way ANOVA and Sidak’s correction.

Fig. 4. Transplanting WT BM into transgenic Tie2^Cre/+;Bmp2^tg/tg mice delays HO onset.

A Schematic representation of transplant assays. WT (CD1) and Tie2^Cre/+;Bmp2^tg/tg mice were transplanted with WT BM cells. B Circulating Bmp2 levels are maintained in each group 3 months after transplant. C Hematopoietic cell engraftment. FACS-quantification of CD45⁺ cells from PB. Tie2^CRE/+;Bmp2^tg/tg and WT mice show similar HSC engraftment. D FACS-quantification of myeloid (CD11b⁺), B Lymphoid (B220⁺) and T Lymphoid (CD3⁺) cells in the two groups of transplanted animals. E Nano-PET-CT imaging of transgenic Tie2^CRE/+;Bmp2 ^tg/tg mice transplanted with WT HSC. These animals develop HO (red arrows) in ribs, hind limbs, and dorsal vertebrae (ribs, rb; hind limbs, hl; and dorsal vertebrae, dv) at 8, 12, and 14 weeks after transplant, suggesting that ectopic bones formation started before transplant. F Representation of the variability of HO onset in non transplanted Tie2^CRE/+;Bmp2^tg/tg mice and Tie2^CRE/+;Bmp2^tg/tg transplanted with WT HSCs showing significant delay in the onset of HO (8–18 weeks versus 16-28w). Kaplan-Meier curve showing that Tie2^CRE/+;Bmp2^tg/tg mice transplanted with WT HSCs survive 10 weeks longer on average non-transplanted transgenic mice. C, D Two-way ANOVA and Sidak’s correction **P < 0.01; F unpaired t test, two tails, mean ± SD. t test ****P < 0.0001. E Survival curve and survival curve analysis, **P < 0.01.

Fig. 5. Tie2^Cre/+;Bmp2^tg/tg resident skeletal muscle cells are characterized by pSmad1/5 and chondro-osteogenic marker expression, and increased fibro-adipogenic progenitors.

Immunodetection of indicated osteo-chondrogenic marker proteins (red), with GFP (green), IB4 (white) and DAPI (blue), on consecutive WT or Tie2 ^CRE/+;Bmp2 ^tg/tg hindlimb skeletal muscle sections, A Nuclear immunostaining of GFP⁺ pSmad 1/5 (arrows) interspersed in muscle interstitium (a’ and b’). White arrowheads indicate GFP⁺ adipocytes. B Cytoplasmic Sox9 immunostaining in cells (arrows in c’ and d’) surrounded by damaged skeletal muscle fibers identified as central nuclei (open arrowheads in c’). C. Nuclear immunostaining of Osterix (Osx) in cells surrounded by fibers (arrows in e’ and f’). Arrowheads indicate strong Osx immunostaining in ectopic bone emerging areas. Arrowheads in f’ indicate GFP⁺ adipocytes. D Top, FACS quantification of the CD45^-PDGFRα⁺Sca1⁺ fibro-adipogenic (FAP) population in Tie2 ^CRE/+;Bmp2 ^tg/tg skeletal muscle. Bottom, FACS quantification showing that 84% of the CD45^-GFP⁺ cells are PDGFRα⁺Sca1⁺ cells. E Immunostainings of pSmad1/5/8 and PDGFRα (red), GFP (green), and DAPI (blue), on consecutive Tie2 ^CRE/+;Bmp2 ^tg/tg hindlimb skeletal muscle sections showing expression in connective tissue surrounding muscle fibers. Bottom panels are magnifications of areas indicated in top panels. Arrows indicate co-localization of pSmad1/5 and PDGFRα cells. Arrowheads indicate co-localization of PDGFRα and GFP⁺ cells. Cells expressing the three markers are indicated by arrows and arrowheads (middle bottom panel). Scale bars 200 µm. Unpaired t test, two tails, mean ± SD *P < 0.05; ****P < 0.0001.