Bone formation is normally confined to skeletal formation during embryonic development and during skeletal repair in response to injury. However, alterations in the usual processes that regulate when and where bone formation occurs can cause disease. The pathologic condition in which bone forms outside the normal skeleton within soft tissues of the body is described as heterotopic ossification.1 In most cases, this extraskeletal bone formation is not a primary pathologic event but occurs secondarily, frequently following severe soft tissue trauma. Heterotopic ossification is a relatively common complication of CNS injury (eg, traumatic brain injuries, spinal cord lesions, tumors, and encephalitis), total hip arthroplasty, deep tissue burns, multiple forms of tissue damage (including war-related injuries that lead to amputations), and end-stage cardiac valve disease. Formation of heterotopic bone can result in severely debilitating consequences, such as restricted joint range of motion, severe pain, and limitations to prosthetic use. There is a clear clinical need to understand the osteoinductive mechanisms that lead to heterotopic ossification in order to develop therapeutic strategies. However, relatively little is understood about the mechanisms through which this process occurs.

At least three components2 appear to be required for the de novo formation of ectopic bone: An induction event promotes molecular and cellular signals conducive to bone formation, progenitor cells are recruited to differentiate to form bone and/or cartilage, and a permissive tissue environment supports progression to bone tissue. These events are analogous to those which occur during skeletal fracture repair,3 a process that has been the target of much investigation yet remains incompletely understood.

Except in inherited forms of heterotopic ossification, which are genetically programmed to form heterotopic bone, ossification does not form spontaneously in soft tissues. These rare human genetic diseases provide opportunities to identify cellular pathways and regulatory mechanisms that are directly involved in the formation of bone in postnatal tissues.4 Inactivating mutations in the GNAS gene, which encodes G protein regulatory subunits, have been identified as a cause of progressive osseous heteroplasia (POH). Fibrodysplasia ossifican progressiva (FOP) is caused by activating mutations in ACVR1, which encodes a bone morphogenetic protein (BMP) type 1 receptor. While the specific contributions that the ubiquitously expressed GNAS gene makes directly to osteogenic induction are not easily apparent, BMPs have been well established as inducers of bone formation.5 BMPs are members of the transforming growth factor β (TGF-β) family of signaling proteins that regulates a wide range of cellular activity, including differentiation, proliferation, apoptosis, migration, positional information, and stem cell renewal, in addition to key functions in embryonic development and skeletal formation.6

It is generally accepted that changes in the tissue microenvironment that are conducive to support the induction and progression of bone formation are required; for example, inflammation, angiogenesis, hypoxia, extracellular matrix, and mechanical forces all have been implicated in directing osteogenesis.3, 6–8 However, the cellular and molecular mechanisms that are activated through these tissue changes remain largely undetermined. In this issue of the Journal of Bone and Mineral Research, Leblanc and colleagues9 move us a step closer to being able to elucidate factors within the cellular environment that regulate cell differentiation to form bone tissue through their study investigating potential osteoprogenitor cells and requirements for specific tissue microenvironments that contribute to the pathophysiology of BMP-induced heterotopic ossification. Their approach used a series of in vitro and in vivo assays and implicates BMP-9 acting through the ALK1 BMP type 1 receptor as an inducer of muscle-resident stromal cells (mrSCs) to undergo osteogenesis and to form heterotopic bone tissue. The most potentially significant findings of this study were that BMP-9 could induce in vivo heterotopic ossification only in the context of tissue injury and that soluble factors from injured skeletal muscles could induce expression of osteogenic markers in mrSCs.

The authors report, as expected, that BMP-2 has osteoinductive capacity in the presence or absence of tissue injury and induces robust heterotopic bone. By contrast, BMP-9 induced heterotopic ossification in their mouse model only in the context of damaged tissue. This observation indicates that BMP-2 and BMP-9 are not functionally equivalent and suggests that specific and required conditions and/or factors are induced by injury and act in concert with BMP-9 to stimulate osteogenesis. This is supported by an apparent discrepancy in response to BMP-2 and BMP-9 in in vitro and in vivo experiments: whereas BMP-9 more robustly induces osteogenic markers in mrSCs in vitro than BMP-2, BMP-2 shows a stronger in vivo induction of heterotopic ossification, indicating that specific interactions with in vivo factors modulate BMP-9 osteoinduction. (It is worth noting that the experimental system applies BMP 3 days after injury and therefore examines only the effects of BMP subsequent to the earliest cellular changes that facilitate heterotopic ossification.)

The mechanisms leading to the differential responses to BMP-2 and BMP-9 will require further investigation. One possibility is that the abilities of BMP-2 and BMP-9 to induce heterotopic ossification are inherent in properties of the specific BMPs to bind and/or activate specific receptors that are expressed by the relevant target cells. Another consideration is whether the bioavailability of BMP-2 and BMP-9 may be responsible for the observed responses. For example, ligand bioavailability could be influenced by differences in rates of extracellular diffusion, binding to components of extracellular matrix, activation by cofactors or processing/activation, and/or susceptibility to agonists. Tissue injury may provide changes in cellular context that increase the ability of BMP-9 to activate signaling.

The authors' intriguing observations leave many questions remaining, however, including whether BMP-9 has a unique role in the induction of heterotopic ossification. The authors suggest that specific signaling by the BMP-9 ligand and the ALK1 receptor is important in the pathophysiology of heterotopic bone. However, while BMP-9 is shown to activate in vivo heterotopic ossification and expression of osteogenic markers in mrSCs, other BMPs possess these properties as well. In addition to BMP-9, expression levels of BMP-2, BMP-4, and BMP-6 all were increased in response to skeletal muscle injury.

Much of the information regarding BMP-9/ALK1 signaling is derived from in vitro studies in which BMP receptors are overexpressed, and the cells then are treated with high concentrations of specific BMP proteins. Such experimental systems are likely assaying the response of high levels of receptor and ligand homodimers, a situation that may not be relevant physiologically. Heterodimers of BMP ligands and BMP type 1 receptor heterodimers have been shown to be the functional forms of BMP receptor complexes, at least in some contexts.10 At endogenous physiologic levels of BMPs and BMP receptors, ALK1 may be responsive to BMP heterodimers or homodimers other than BMP-9.

Leblanc and colleagues further show that expression of BMP type 1 receptors in addition to ALK1 are increased in response to tissue injury. While BMP-9-induced heterotopic ossification was shown to be mediated through the ALK1 type 1 receptor, the study does not address whether ALK1 acts independently of other BMP type 1 receptors or other receptors contribute to heterotopic ossification, perhaps through type 1 BMP receptor heterodimer formation with ALK1. Although experimental data support that ALK1 binds BMP-9 and activates the BMP-Smad1/5/8 pathway in response to BMP-9,11 there is additional evidence that BMP-9 binds to and signals through receptors other than ALK1, including ALK2,12, 13 the BMP type 1 receptor that is mutated in patients with FOP.4 It is interesting to speculate that in these experiments, BMP-9 in the context of injury is inducing heterotopic ossification by activating ALK1-ALK2 heterodimers or that a sequential series of BMP signaling events mediated independently by ALK1 and ALK2 is required for induction of heterotopic ossification.

A final and important aspect to consider is the identity of osteoprogenitor cells that participate in heterotopic bone formation. The authors note two populations of skeletal muscle progenitor cells that have been identified: satellite cells (myogenic stem cells) and muscle-resident stromal cells (mrSCs; mesenchymal cells). Multipotent mrSCs (Sca1+, CD31, Lin) were shown to express ALK1 and respond to BMP-9 in vitro, and S+CL cells increased in number in injured tissues, suggesting that these cells are a source of osteoprogenitors for heterotopic bone formation. While the data support this conclusion, contributions by other cells that are resident in or migrate to muscle in response to injury cannot be excluded. A number of studies have identified and characterized additional cell populations as osteoprogenitors.14–16 The relationships of these cell populations to each other and to hereditary and nonhereditary forms of heterotopic ossification remain to be resolved. Of relevance is the recent report17 of endothelial cell transdifferentiation to osteoprogenitor cells in response to mutant ACVR1/ALK2. Interestingly, ALK1 was identified initially as an endothelial-specific type 1 receptor.11 It also cannot be overlooked that heterotopic bone is a tissue that frequently forms through endochondral ossification and is comprised of multiple cell types. While the brunt of investigations of heterotopic bone formation has focused on osteoblast formation and differentiation, recruitment and differentiation of additional cell types are critical to the process as well.


  1. Top of page
  2. Disclosures
  3. References

The author states that she has no conflicts of interest.


  1. Top of page
  2. Disclosures
  3. References