Tendon stem/progenitor cells in heterotopic ossification: functional regulation, molecular mechanisms and targeted therapeutic strategies

Abstract

Heterotopic ossification (HO) is a pathological process in which ectopic bone tissue forms within soft tissues such as tendons, resulting in pain, limited range of motion, and functional disability that reduces the quality of life. In recent years, tendon stem/progenitor cells (TSPCs) have been considered a key cellular source of HO due to their inherent multilineage differentiation potential. This review aims to systematically elaborate on the advances in the functional regulation and mechanisms of TSPCs in HO. Existing evidence (including lineage tracing and clinical studies) indicates that under the stimulation of trauma, inflammation, and other factors, the chondrogenic and osteogenic differentiation of TSPCs are abnormally activated, shifting their fate from tendon repair to ectopic bone formation. This process is centrally regulated by osteochondrogenesis-related signaling pathways such as BMP, as well as inflammatory signaling pathways such as NF-κB. External factors such as mechanical and inflammatory microenvironment also play important stimulatory roles. Based on these mechanisms, drugs and biomaterials targeting the function of TSPCs have shown potential for the prevention and treatment of HO. Future research needs to explore the heterogeneity of TSPCs and the spatiotemporal dynamics of their regulatory networks, providing theoretical support for the clinical translation of therapeutic strategies.

The translational potential of this article

Given the critical role of TSPCs in driving heterotopic ossification through aberrant differentiation, targeting these cells represents a highly promising translational strategy. This review details how external stimuli change the local microenvironment and disrupt the signaling networks within TSPCs, shifting their fate from tissue repair to abnormal bone formation. Highlighting the application of pharmacological agents and biomaterials to precisely modulate the behavior of TSPCs supports a crucial shift toward targeted preventative therapies

Graphical abstract

1. Introduction

Heterotopic ossification (HO) refers to a pathological process characterized by the abnormal formation of lamellar bone in soft tissues such as muscles, tendons, and ligaments. As a severe clinical complication, it often occurs secondary to severe trauma, central nervous system injury, extensive burns, or joint replacement surgery, leading to functional impairments including joint pain, stiffness, and severely restricted range of motion, which greatly affect patients’ quality of life and rehabilitation prognosis [1]. Although the exact global incidence of HO varies, its prevalence can be as high as 20% to 40% in specific high-risk populations, such as patients with spinal cord injury [2], and some reports indicate that its incidence reaches 14% to 62% in patients after hip surgery [3], highlighting the universality and clinical importance of this condition. Currently, the clinical management of HO faces great challenges because its pathogenesis has not been fully elucidated, resulting in a lack of specific preventive and radical treatment methods. The traditional view holds that the initiation of HO originates from acute or chronic inflammatory responses triggered by local tissue injury. This inflammatory microenvironment disrupts tissue homeostasis and may recruit or abnormally activate the inherent chondrogenic and osteogenic precursor cells in the tissue, leading them to differentiate into the osteoblastic lineage through intramembranous or endochondral ossification, and ultimately form ectopic bone in soft tissues [4]. Among various types of HO, those occurring near the attachment points of tendons and ligaments are particularly common [5], such as around the elbow and hip joints, which strongly suggests that tendon tissue itself may play a key role as a cellular source in this pathological process.

For a long time, exploring the cellular origin of HO has been a scientific hotspot in the field. Although mesenchymal stem cells (MSCs) are widely regarded as potential osteogenic precursor cells [6], the role of tissue-specific stem/progenitor cells has attracted increasing attention. Tendon tissue was once thought to be mainly composed of terminally differentiated tenocytes with limited regenerative capacity [7], so its importance as a cellular source of HO was not fully recognized in the past. A revolutionary breakthrough in this research field was achieved in 2007, when Bi et al. first successfully isolated and identified a population of cells, tendon stem/progenitor cells (TSPCs), from human and mouse tendon tissues, with self-renewal capacity, clonogenicity, and multilineage differentiation potential, including tenogenic, osteogenic, chondrogenic, and adipogenic lineages [8]. This landmark discovery not only completely updated the scientific community's understanding of tendon biology, tendon homeostasis maintenance, and injury repair mechanisms, but also provided a new cell biological perspective for the exploration of a series of tendon-related pathological phenomena, such as HO and calcific tendinopathy. TSPCs are a key cell subset that maintains tendon tissue homeostasis and participates in tendon injury repair: under physiological conditions, their main function is to respond to minor tissue damage, proliferate and differentiate into tenocytes to participate in tissue renewal and repair [9]; however, under specific pathological conditions, such as severe trauma, persistent inflammation, or abnormal mechanical stimulation, their delicate cell fate regulatory mechanisms may be disrupted, and they incorrectly switch from the original tenogenic differentiation pathway to the chondrogenic/osteogenic differentiation pathway, thus becoming a potential cellular source and initiating factor of HO. In recent years, several studies have further revealed the heterogeneity of TSPC subpopulations. Specific cell subpopulations, such as CD26⁺ TSPCs [10] or those expressing Gli1 [11] and Tppp3 [12], have been proven to be the main precursor cell sources and play a crucial role in HO models. At the same time, with the application of new technologies such as single-cell sequencing and lineage tracing, researchers can more accurately identify and distinguish functional subpopulations of TSPCs with different differentiation tendencies, which provides a higher-resolution perspective for understanding the cellular origin of HO [13].

Based on the research background mentioned above, we will focus on the core role of TSPCs in HO and systematically sort out the latest research progress in their functional regulation and molecular mechanisms. First, we will summarize the lineage tracing and clinical evidence of TSPCs as the cellular source of HO, establishing their key cellular basis. Furthermore, we will focus on elaborating the complex network composed of core intrinsic signaling pathways and extrinsic microenvironmental factors that regulate TSPCs osteochondrogenesis. On this basis, the review will also discuss potential preventive and therapeutic strategies targeting TSPCs function developed based on different mechanisms, including drug intervention and biomaterial application. Through a systematic review of existing studies, this review aims to provide a new perspective for in-depth understanding of the pathogenesis of HO and point out the direction for the future development of specific preventive and therapeutic strategies targeting TSPCs (Fig. 1).

Fig. 1.

Schematic overview of the functional regulation, molecular mechanisms and targeted therapeutic strategies of TSPCs in HO.

2. Evidence and fate transition of TSPCs in HO

2.1. Evidence for the involvement of TSPCs in HO

Identifying the exact cellular origin of HO is important for understanding its pathological mechanisms and developing specific targeted therapies. In early studies, although histological observations suggested the abnormal differentiation of mesenchymal precursor cells within soft tissues, there was a lack of effective methods to directly trace the fate of these cells in vivo. In recent years, with the help of dual-fluorescent reporter systems and conditional knockout Cre-LoxP genetically engineered mouse models, the core role of TSPCs as the cellular source of HO has been strongly studied.

Various TSPCs subpopulations distribute in different regions of the tendon and exerting functions with common characteristics but distinct specificities. A review published in 2021 summarized numerous identified TSPCs subpopulations with specific molecular markers [14]. Here, we update the newly discovered TSPCs subpopulations over the past five years, along with their tissue locations, biological functions, and associated diseases (Table 1). Several typical TSPCs subpopulations are involved in the regulation of HO. Dey et al. found that Scx⁺ tendon-derived progenitor cells mainly mediate non-traumatic HO of ligaments and joints, while muscle-resident Mx1⁺ mesenchymal cells mediate HO induced by muscle injury in Acvr1^R206H knock-in FOP mice [22]. Also, Feng et al. used lineage tracing to define that Ctsk-Cre can label a subset of Scx⁺ TSPCs, and inhibiting SUFU in Ctsk-Cre-expressing cells to upregulate the Hedgehog signaling pathway can induce HO in tendons and ligaments [23]. The results of Chen et al. showed that Gli1⁺ TSPCs at the tendon site of mice proliferate rapidly after trauma, differentiate into tenocytes and osteochondrocytes, and then abnormally form cartilage and bone at 21 days and 63 days post-injury, respectively [11]. Moreover, single-cell RNA sequencing (scRNA-seq) with lineage tracing technology identified that CD26⁺ TSPCs participates in tendon healing and HO through mediating the Tenascin-C (TNC)-Hippo signaling pathway [10]. Collectively, three distinct subpopulations have been functionally validated as direct cellular contributors to HO: CD26⁺ TSPCs (driving HO via TNC-Hippo-YAP signaling), Gli1⁺ TSPCs (mediating HO through GNAS/PKA/Hedgehog axis) [11], and Tppp3⁺ synovial/tendon sheath progenitor cells (contributing to post-traumatic heterotopic bone formation) [12]. The functional distinction among these subpopulation markers not only refines our understanding of cell heterogeneity in HO but also provides a roadmap for cell-specific precision therapy.

Table 1.

TSPCs subpopulations and their characteristics.

Species	Tissue locations	Markers	Biological functions	Associated diseases	References
Rat	Achilles tendon	CD26	Enhance tenogenesis, promote ECM deposition	Tendon HO, tendinopathy	[15]
Rabbit	Patellar tendon	Sox11	Promote cell proliferation, osteogenesis and angiogenesis	Osteonecrosis of the femoral head	[16]
Mouse	Tail tendon, Achilles tendon	Cd9, Cd271	Secrete NGF, mediate neonatal-to-adult tendon transition, drive immature to mature tendon conversion, regulate tendon maturation	Tendon injury and repair dysfunction	[17]
Mouse	Achilles tendon	Prg4	Participate in tendon and ligament homeostasis, inhibit abnormal cell differentiation	Tendon and ligament HO	[18]
Mouse	Achilles tendon	CD55, CD248	High clonogenic potential and tenogenesis capacity, form functional tendon-like tissue	Tendon and ligament injuries, tendinopathy, postoperative tendon rupture	[19]
Mouse	Rotator cuff tendon	CD248	Regulate stemness, participate in tendon-bone tissue remodeling, mediate intercellular communication and ECM organization	Osteoporosis, rotator cuff tears, impaired tendon-bone healing after rotator cuff repair surgery	[20]
Human	Semitendinosus tendon	FGF7, CYGB	Activate FGF7 signaling pathway, promote tenogenesis and tendon regeneration via intercellular communication	Tendon injury, impaired healing and regeneration disorder	[21]
Human	Semitendinosus tendon	ICAM1, ITGB8	Mediate cell adhesion and intercellular interaction, synergize with FGF7 to enhance tenogenesis.	Tendon injury, defect and poor regeneration	[21]
Mouse, Human	Achilles tendon, spinal ligament	RSPO2	Suppresse chondrogenesis, inhibit ectopic endochondral ossification, maintain tendon and ligament homeostasis	OPLL, tendon and ligament HO	[18]
Mouse, Human	Peritendon of Achilles tendon, Spinal ligament enthesis tissue, Tail tendon sheath	CD26	Migrate to tendon midsubstance, differentiate into tenocytes, chondrocytes and osteoblasts	Tendon injury with inadequate healing, tendon trauma HO	[10]
Mouse, Human	Achilles tendon sheath, Posterior longitudinal ligament	Gli1	Mediate Hedgehog signaling, maintain cell properties, regulate osteogenesis	Tendon traumatic HO, FOP, OPLL	[11]

In the future, the application of spatial transcriptomics may construct a four-dimensional spatiotemporal regulatory map of HO occurrence and development, accurately locate the key node cell subsets and microenvironmental niches that drive the abnormal differentiation of TSPCs, and integrated multi-omics analysis is expected to open new dimensions for the research of TSPCs in HO. However, due to the anatomical differences and local microenvironmental disparities between human and mouse tendon tissues, animal models cannot fully simulate the complex pathological environment of clinical traumatic HO, which limits the extrapolation of research conclusions. For instance, human traumatic HO typically develops over weeks to months with sustained low-grade inflammation and pronounced fibrotic remodeling, whereas mouse models often exhibit a more acute and self-limited inflammatory response, and lack the genetic susceptibility or comorbid conditions (e.g., diabetes, aging) commonly seen in patients [24]. Moreover, due to ethical and technical constraints, it is impossible to implant fluorescent labels or genetic switches into the human body, resulting in the inability of lineage tracing technology to be applied to human samples. HO also frequently occurs in skeletal muscle. Lineage tracing shows that Scx-lineage cells contribute to ectopic bone formation in muscle [25]. However, distinct muscle-resident progenitors may have different functions: Mx1⁺ interstitial cells mediate injury-dependent intramuscular HO, while Scx⁺ progenitors drive non-traumatic HO in ligaments and joints [22]. Moreover, a distinct muscle interstitial progenitor population demonstrates potent BMP dependent osteogenesis and represents a predominant cellular origin of HO. [26]. These findings provide comparative insights into cellular heterogeneity and microenvironmental regulation across soft tissues. Therefore, it is necessary to develop new in vitro models, such as human iPSC-induced tendon organoids, to better supplement the evidence for the involvement of TSPCs in HO.

2.2. Fate transition of TSPCs drives HO

The core pathological mechanism of HO lies in the aberrant fate transition of TSPCs (Fig. 2). Under normal physiological conditions, TSPCs differentiate specifically into tenocytes under the regulation of key tendon transcription factors such as Scleraxis (SCX), Mohawk (MKX), and Early growth response 1 (EGR1). They synthesize a parallel fibrous matrix dominated by type I collagen, thereby maintaining tissue homeostasis and supporting repair functions [27,28]. However, when soft tissues are exposed to severe trauma, persistent inflammatory stimulation, or extreme mechanical stress, the epigenetic modifications and transcriptional regulatory networks of TSPCs undergo remodeling. The tenogenic differentiation program is suppressed, while chondrogenic and osteogenic pathways are abnormally activated, ultimately leading to HO formation.

Fig. 2.

Distinct fate transitions of TSPCs. Under the regulation of Scx/Mkx/Egr1 and BMP-12/13/14, TSPCs differentiate into tenocytes. Under the regulation of Sox9, TGF-β3/BMP-2/4/7, TSPCs differentiate into chondrocytes. Under the regulation of Runx2/Osterix (Osx), BMP-2/7, and Wnt/β-catenin, TSPCs differentiate into osteocytes.

HO occurs through two distinct mechanisms: intramembranous ossification and endochondral ossification. Intramembranous ossification, which is mainly involved in craniofacial bone development, is driven by osteoblasts directly differentiated from mesenchymal tissue without a cartilaginous intermediate stage. A classic example of ectopic intramembranous ossification is progressive osseous heteroplasia (POH), caused by loss-of-function mutations in the Gnas gene [29]. In contrast, endochondral ossification begins with the differentiation and hypertrophy of chondrocytes, which are later replaced by osteoblasts. This mechanism is often seen in the natural development of long bones and fracture healing, and underlies most cases of acquired HO. In tendon HO, when local injury repair fails, tendon tissue often becomes refractory to healing and accompanied by chronic inflammation, characterized by cellular hyperplasia, excessive proteoglycan deposition, and collagen matrix degradation. Tendon samples from patients with tendinopathy show obvious chondroid metaplasia, in which tenocytes acquire chondrocyte-like morphological features and upregulate chondrocyte marker genes. This suggests that abnormal differentiation may underlie failed tendon healing and subsequent HO [30]. As resident cells in tendon tissue, TSPCs possess unique clonogenicity, self-renewal capacity, and multilineage differentiation potential. They can be recruited and activated in the early inflammatory phase, then differentiate locally into chondrogenic and osteogenic cells, giving rise to the HO phenotype [31]. Various growth factors and complex cellular signaling cascades likely play critical roles in this fate transition of TSPCs, which will be elaborated in later sections of this review.

Nevertheless, the molecular switch mechanism governing TSPCs fate transition remains incompletely understood. Although the downregulation of tendon-related transcription factors and the activation of osteogenic/chondrogenic pathways have been confirmed, the key epigenetic modifying enzymes or factors that mediate this transcriptional switch have not been systematically characterized, as well as their target gene regulatory networks. In a recent study, the essential roles of the Hippo pathway transcriptional coactivators YAP and TEAD in tendon development were found by integrating transcriptomic and epigenetic data from mouse tendon tissue from the neonatal to early adult stages [32]. scATAC-seq revealed increased chromatin accessibility of osteogenic genes in mesenchymal progenitor cells during HO formation [33], but did not further identify key epigenetic regulators or mechanisms. Future research may focus on dynamic chromatin changes and epigenetic modifiers during TSPCs osteochondrogenesis, and promote the development of epigenetic editing tools and targeted drugs for the treatment of HO.

3. Regulatory signaling pathways in TSPCs osteochondrogenesis

The differentiation fate of TSPCs is not random but is coordinately regulated by a series of highly conserved and sophisticated intracellular signaling pathways. Under physiological conditions, these signaling networks maintain the homeostasis of TSPCs, directing them toward tenogenic differentiation to participate in tissue repair and renewal. However, under pathological conditions such as trauma and inflammation, local microenvironmental disorders can lead to the abnormal activation or inhibition of key signaling pathways, thereby disrupting the original differentiation balance and driving TSPCs to erroneously switch to the chondrogenic and osteogenic lineages, serving as the cellular basis for HO. These pathways include well-known signaling pathways such as the BMP, Wnt/β-catenin, Hedgehog, and Hippo signaling pathways, as well as other important regulatory pathways including Notch, MAPK, PI3K-Akt, and NF-κB (Fig. 3). They do not function separately but form a complex and dynamic regulatory system through multi-level crosstalk, collectively determining the fate of TSPCs. This section will thoroughly review the molecular composition, interactions, and regulatory mechanisms of these signaling pathways during TSPCs osteochondrogenesis and HO development.

Fig. 3.

Aberrant changes in multiple signaling pathways lead to overactivated TSPCs osteochondrogenesis in HO. Under various external stimuli and experimental interventions, a series of intracellular signaling events in TSPCs lead to robust expression of chondrogenic genes (e.g., Sox9, Acan, Col2a1, etc.) and osteogenic genes (e.g., Runx2, Osx, Alp, Ocn, etc.). The green sections highlight inhibitors of several signaling pathways, all of which have been validated during TSPCs osteochondrogenesis.

3.1. BMP signaling pathway

BMP is an important member of the transforming growth factor-β (TGF-β) superfamily and plays a central role in skeletal development, homeostasis, and injury repair [34]. BMP signals through a cell-surface serine/threonine kinase receptor complex composed of type I (ALK2/ACVR1, ALK3/BMPRIA, ALK6/BMPRIB) and type II subunits, activating canonical SMAD and/or non-SMAD pathways (p38 MAPK, ERK, JNK, PI3K-Akt) to regulate target gene expression [35]. In TSPCs, BMP signaling exhibits context-dependent effects. Physiological levels of BMP-12, BMP-13, and BMP-14 promote tenogenesis and support tendon repair [[36], [37], [38]]. Conversely, pathologically overactivated BMP-2, BMP-4, BMP-7, and other subtypes drive TSPCs toward aberrant chondrogenic and osteogenic differentiation, resulting in HO [39]. TSPCs isolated from chronic tendinopathy show significantly higher BMP/Smad activity than those from healthy tendons, and BMP-2 treatment further upregulates osteochondrogenic genes, which can be blocked by the BMP inhibitor Noggin [40]. The BMP inhibitor LDN193189 attenuates the osteogenic potential of CD26⁺ TSPCs while enhancing their tenogenic capacity, preventing HO after tendon injury [15]. In FOP, gain-of-function mutations in the BMP type I receptor gene Acvr1/Alk2 render TSPCs abnormally sensitive to ligands [41], and lineage-tracing studies confirm that Scx⁺ TSPCs are the main cellular source mediating non-traumatic HO [22]. Inhibiting primary cilia-mediated BMP signaling suppresses abnormal osteochondrogenesis in mouse and human primary FOP cells [42]. Future studies should clarify subtype-specific mechanisms of different BMPs in distinct TSPC subpopulations and explore crosstalk between BMP and other regulatory pathways.

3.2. Wnt/β-catenin signaling pathway

The canonical Wnt/β-catenin pathway plays a central regulatory role in skeletal development and metabolism [[43], [44], [45]]. In TSPCs, activation of this pathway suppresses expression of tenogenic genes (Scx, Mkx, Tnmd), indicating an inverse relationship with tenogenic differentiation [45]. Conversely, aberrantly activated Wnt/β-catenin signaling promotes TSPCs osteochondrogenesis and is essential for pathological bone formation in HO [46]. In vitro administration of baicalin promotes osteogenic marker (Alp, Runx2, Ocn) expression in TSPCs via Wnt/β-catenin in a dose-dependent manner, while oral treatment enhances tendon-bone healing in rats [47]. LncRNA Snhg12 targets miR-199a-5p to relieve inhibition of Fzd4, upregulating Wnt/β-catenin and promoting HO formation [48]. The Nr4a1/Wnt/β-catenin axis positively regulates osteogenic differentiation of mouse TSPCs, an effect inhibited by metformin [49]. Macrophage migration inhibitory factor (MIF) activates Wnt/β-catenin through a ROS/HIF-1α positive feedback loop under inflammatory and hypoxic conditions, enhancing osteochondrogenetic gene transcription [50]. Given that Wnt/β-catenin is critical for systemic skeletal homeostasis, and its dysregulation is associated with osteoporosis and bone defects [51], exploring its specific regulatory roles in TSPC osteochondrogenesis and HO formation is of great significance.

3.3. Hedgehog signaling pathway

The Hedgehog signaling pathway is an evolutionarily highly conserved core developmental pathway, first identified in Drosophila embryonic development screens and named for the hedgehog-like protrusions on larval cuticles caused by its mutation [52]. Hedgehog signaling is highly dependent on primary cilia, a specialized organelle that act as a hub for the dynamic assembly of pathway components and signal transduction, thereby exerting its regulatory functions in embryonic development, tissue homeostasis, and cell differentiation [53]. In the resting state, the receptor Patched1 (PTCH1) localizes and accumulates on the ciliary membrane, where it inhibits the activity of another transmembrane protein, Smoothened (SMO), and prevents its entry into the cilia. Meanwhile, cytoplasmic GLI family transcription factors (GLI2 and GLI3) bind to suppressor proteins such as SUFU and undergo partial proteasomal degradation to generate truncated transcriptional repressors (GLI2/3R), which enter the nucleus to inhibit target gene transcription. Upon binding of Hedgehog ligands to PTCH1, the inhibition of SMO is relieved. Activated Smo initiates downstream signaling cascades, dissociating SUFU from GLI transcription factors, preventing processing of GLI into repressor forms, and promoting their conversion to full-length transcriptional activators (GLI2/3A). These activators, together with GLI1, a direct target and potent activator of the pathway, translocate into the nucleus to initiate transcription of downstream target genes [54,55].

In mammals, the Hedgehog pathway includes three homologous ligands: Sonic hedgehog (SHH), Indian hedgehog (IHH), and Desert hedgehog (DHH). DHH exhibits a relatively restricted expression pattern and functions mainly in germ cell development in gonads [56], peripheral nerve myelination [57], and airway/pancreatic islet tissue regeneration [58]. IHH plays an indispensable regulatory role in skeletal development, particularly in endochondral ossification at the growth plates of long bones [59]. As the most extensively studied and functionally diverse member of this family, SHH contributes to the initial patterning of embryonic skeleton as well as postnatal bone homeostasis and repair [60,61]. Chen et al. reported that transfection of Ihh and Shh into rabbit BMSCs significantly induced chondrogenesis while suppressing chondrocyte hypertrophy and senescence [62], further highlighting their critical functions in regulating bone and cartilage development and homeostasis.

Multiple studies have confirmed that the Hedgehog signaling pathway regulates the differentiation of TSPCs. Studies by Thomopoulos lab [[63], [64], [65]] demonstrated that Gli1⁺ cells at tendon entheses (including TSPCs) critically promote tendon development and injury repair. Since GLI1 is a key transcription factor that participates in and upregulates Hedgehog signaling, this cell population is controlled by the Hedgehog pathway, indicating its activating role in tenogenic differentiation of TSPCs. Zong et al. also found that the IHH pathway is important for TSPCs-mediated rotator cuff tendon repair in rats [66]. However, aberrantly activated Hedgehog signaling under pathological stimuli redirects TSPCs differentiation toward chondrogenesis and osteogenesis, leading to HO at tendon sites. It was reported that knockout of Hedgehog pathway suppressor SUFU in Ctsk⁺ Scx⁺ TSPCs results in activation of GLI1/2 transcription factors, activating Hedgehog signaling pathway and leading to spontaneous progressive HO in tendons, while pharmacological inhibition with JQ1 downregulated Hedgehog signaling and alleviated periarticular and tendon HO in mice [23]. Enhanced Hedgehog pathway activity was also detected in a rat Achilles tendon transection model, while treatment with GANT58 (a GLI antagonist) and SAG (a SMO agonist) inhibited and promoted TSPCs osteochondrogenesis and local HO volume, respectively [67]. It is worth noting that Gli1⁺ tendon sheath cells derive from Scx⁺ TSPCs and rapidly proliferate and undergo osteochondrogenic differentiation following Achilles tendon transection in mice, a process regulated by the GNAS/PKA/Hedgehog signaling axis [11]. Moreover, ChIP-seq analysis of TSPCs overexpressing tendon-related transcription factor MKX showed elevated expression of tenogenic genes and reduced expression of chondrogenic or osteogenic genes [68]. Mkx^KO mice developed spontaneous HO as early as one month after birth, accompanied by early activation of the Hedgehog pathway [69]. Future research should further investigate whether key molecules governing tendon development and homeostasis modulate TSPCs osteochondrogenesis through the Hedgehog pathway, providing new directions for uncovering HO mechanisms and developing therapeutic strategies.

While it is well established that the Hedgehog pathway is highly dependent on primary cilia, whether primary cilia contribute to the pathogenesis of HO by mediating this pathway has long remained elusive. Accordingly, our group has performed in-depth investigations in this area in recent years (Fig. 4). We first identified primary cilia as a key regulatory hub for FOP, where the Acvr1^R206H mutant activates BMP and Hedgehog signaling via ciliary compartments, leading to the hyperactivated osteochondrogenesis [42]. We further revealed that primary cilia serve as the core platform for aberrant Hedgehog and PI3K-Akt pathway activation in traumatic HO, mediating a positive feedback loop that drives TSPCs osteochondrogenesis and ectopic bone formation [70]. These findings establish primary cilia as a shared pathogenic node across both genetic and acquired HO, filling a key research gap and providing a novel therapeutic target for HO. Our subsequent work will focus on dissecting ciliary signaling crosstalk, screening targeted therapeutics, and exploring early HO diagnostic biomarkers.

Fig. 4.

The role of primary cilia and mediated Hedgehog signaling in HO pathogenesis. Schematic model of primary cilia-mediated BMP and Hedgehog signaling in FOP (upper). Schematic model of primary cilia-mediated Hedgehog and PI3K-Akt signaling in traumatic HO (lower).

3.4. Hippo signaling pathway

The Hippo signaling pathway can regulate cell proliferation, apoptosis, and tissue homeostasis by sensing extracellular microenvironmental cues such as cell density and mechanical stress [71]. This pathway relies on a cascade of kinase reactions [72]. Under resting conditions, activated large tumor suppressors LATS1/2 phosphorylate specific serine residues of the core downstream effectors: Yes-associated protein (YAP), a transcriptional coactivator, and its paralog transcriptional coactivator with PDZ-binding motif (TAZ). Phosphorylated YAP/TAZ then bind to 14-3-3 proteins and are sequestered in the cytoplasm, inhibiting their transcriptional coactivator function. Conversely, when upstream inhibitory signals are relieved, the Hippo kinase cascade is inactivated, allowing YAP/TAZ to be dephosphorylated and rapidly translocate into the nucleus. Once inside the nucleus, they bind tightly to target transcription factors including the TEAD family to form transcriptional complexes, thereby initiating the expression of downstream target genes [73].

The Hippo-YAP/TAZ signaling pathway plays an essential role in the development, metabolism, and homeostasis of the skeletal system [74]. During early chondrogenesis, the nuclear transcriptional activity of YAP/TAZ is critical for the proliferation and expansion of progenitor cells. As chondrocytes differentiate toward maturation, activation of the Hippo-YAP/TAZ pathway releases the inhibition of cartilage-specific transcription factors, ensuring normal synthesis and secretion of the cartilage matrix [75]. Under pathological conditions, external stimuli induce abnormal activation and sustained nuclear translocation of YAP/TAZ, triggering hypertrophic transformation of chondrocytes and accelerating cartilage matrix degradation [76]. In the regulation of bone development and metabolism, YAP/TAZ mainly act as positive regulators. Physical mechanical signals in the bone microenvironment strongly suppress the core Hippo kinase cascade through mechanotransduction networks, directly driving extensive nuclear accumulation of YAP/TAZ. These coactivators then cooperate with key osteogenic transcription factors, like RUNX2, to robustly activate the transcription of osteogenic target genes, thereby promoting osteoblast proliferation, differentiation, and bone matrix synthesis and mineralization [77].

In tendon biology, conditional knockout of YAP1 (a key regulator of early tendon growth and maturation) in postnatal Scx-lineage cells alters collagen expression and tissue architecture in tendons [32]. Recent studies concentrate on the influence of the Hippo pathway on TSPCs osteochondrogenesis. Using scRNA-seq, a CD26⁺ TSPCs subpopulation was identified in tendon tissue that contributes to tendon healing and HO [10]. RNA-seq and enrichment analysis revealed that the TNC-Hippo-YAP pathway is significantly upregulated during pathological osteochondrogenesis, while targeting TNC (via gene knockout or neutralizing antibodies) or treatment with the Hippo antagonist XMU-MP-1 (which reduces YAP phosphorylation and inhibits its degradation) both suppressed ectopic bone formation [10]. In addition, studies have shown that mechanical stimulation mediated by primary cilia regulates the chondrogenic and osteogenic differentiation potential of nestin⁺ BMSCs through the F-actin-Hippo-YAP pathway, thereby promoting healing at the bone-tendon interface [78]. Given the similarities between TSPCs and BMSCs in cellular properties and differentiation potential, abnormally hyperactivated Hippo signaling may promote TSPCs osteochondrogenesis via regulation of primary cilia on their surface, ultimately leading to HO. This hypothesis requires further experimental validation in the future.

3.5. Other important signaling pathways

In addition to the core pathways mentioned above, other signaling pathways also participate in the complex regulatory network governing the fate determination of TSPCs. They can function either independently or through extensive crosstalk with core pathways, collectively shaping the pathological microenvironment that drives HO formation and development.

The Notch signaling pathway is an evolutionarily highly conserved mechanism for intercellular communication, mainly mediating juxtacrine signaling between adjacent cells. The pathway is triggered by specific binding of transmembrane ligands (e.g., Delta-like or Jagged family proteins) to Notch receptors (Notch1-4) on neighboring cell surfaces, inducing proteolytic cleavage of the receptor and subsequent intracellular signal transmission. It plays essential roles in regulating cell proliferation, differentiation, and apoptosis [79]. Notch signaling is critical for BMP-induced bone regeneration, and its function in bone homeostasis mainly involves regulating the differentiation, function, and survival of osteoblasts and osteoclasts [80]. It also acts as a key regulator of cartilage growth, development, and joint homeostasis, influencing the entire life cycle of chondrocytes, including dedifferentiation, endochondral ossification, maturation, and cartilage repair [81]. Roberts et al. reported that the expansion of Sox9⁺ chondrocytes derived from Scx⁺ TSPCs is accompanied by downregulation of the Notch2-Dll1 pathway, leading to ectopic endochondral ossification at the tendon-bone junction [82]. Researchers also found that Triptolide significantly reduces the expression of proinflammatory cytokines and inhibits mouse TSPCs osteochondrogenesis by suppressing Notch signaling [83], highlighting the potential therapeutic value of targeting Notch in traumatic HO. Under various pathophysiological conditions, the Notch intracellular domain (NICD) can interact with signaling molecules of the BMP or Wnt/β-catenin pathways to coordinately regulate target gene transcription, thereby amplifying chondro-osteogenic signals [84]. However, this mechanism has not been verified in TSPCs-mediated HO and may represent a direction for future research.

The mitogen-activated protein kinase (MAPK) cascade serves as a central hub integrating diverse extracellular stimuli to regulate cell proliferation and differentiation. It mainly comprises three classical branches: ERK, p38, and JNK, all of which have been implicated in cancer, inflammatory diseases, and neurodegenerative disorders [85]. During normal skeletal development and endochondral ossification, the different MAPK branches exert spatiotemporally specific and synergistic effects. The ERK1/2 pathway acts as a key positive driver in the early stage of osteoblast differentiation, it directly phosphorylates specific serine or threonine residues of the osteogenic transcription factor RUNX2, enhancing its activity and preventing proteasomal degradation, thereby promoting early bone matrix synthesis [86]. Meanwhile, the p38 MAPK and JNK pathways not only contribute to osteogenic differentiation and bone metabolism [87,88] but also play indispensable roles in chondrocyte proliferation, hypertrophic maturation, and cartilage matrix remodeling [89]. In recent years, the role of the MAPK pathway in mediating aberrant fate decisions of TSPCs and HO has gradually been uncovered. Chen et al. confirmed through multi-omics analysis that collagen triple helix repeat-containing 1 (CTHRC1) activates the MAPK pathway by binding to EGFR, promoting the proliferation, migration, and tenogenic differentiation of TSPCs, thereby enhancing tendon repair and alleviating tendinopathy severity [90]. Wang et al. found that secreted protein acidic and rich in cysteine (SPARC) is upregulated in the Achilles tendons of rats with HO, accompanied by elevated phosphorylation of MAPK-related proteins (ERK, JNK, p38), which promotes osteogenic differentiation of mesenchymal stem cells [91]. In contrast, some studies have reported that MAPK/ERK1/2 activity is positively correlated with TSPCs proliferation and migration but negatively associated with their osteogenesis, suggesting it may be inhibited by the phosphorylation status of the BMP pathway [92]. Given the universal and complex roles of MAPK signaling in cellular function, it is crucial to dissect the activation patterns of individual MAPK branches at different stages of HO progression and their effects on TSPCs differentiation.

The PI3K-Akt signaling pathway senses growth factors, hormones, or cytokines via cell-membrane receptors such as receptor tyrosine kinases (RTKs), and subsequently activates downstream effector molecules to participate in a wide range of cellular processes, including cell growth, proliferation, metabolism, migration, and secretion. Its dysregulation is closely associated with numerous human diseases, including cancer, immune disorders, neurological diseases, diabetes, local tissue hyperplasia, and cardiovascular diseases [93]. Genetic and pharmacological studies have confirmed that the differentiation and maturation of cartilage and bone depend on the PI3K-Akt signaling pathway [94,95]. The role of the PI3K-Akt pathway in various forms of HO has been extensively documented. Valer et al. found that the PI3Kα inhibitor BYL719 reduces the activity of downstream SMAD, AKT, and mTOR/S6K, impairs the osteogenic response of mesenchymal stem cells to BMP, and alleviates ectopic bone formation in FOP mice [96]. After spinal cord injury, CXCL2 secreted by M1 macrophages binds to the CXCR4 receptor on BMSCs, activating the downstream PI3K-Akt pathway and the transcription of osteogenic genes such as Ocn and Runx2, thereby inducing neurogenic HO [97]. Regarding tendon HO driven by TSPCs, it was reported that neurotrophin-3 (NT-3) accelerates TSPCs osteochondrogenesis by upregulating the PI3K-Akt signaling pathway and then promotes tendon mineralization and HO formation in rats [98]. Insulin-like growth factor 1 (IGF-1) binds to its receptor and activates the PI3K-Akt-mTOR pathway, inducing TSPCs differentiation into chondrocytes and osteoblasts both in vitro and in vivo, while inhibition of either PI3K or mTOR suppresses IGF-1-stimulated ectopic bone formation [99]. Interestingly, Chen et al. reported that local blood infiltration following tendon injury exacerbates chronic HO, induces apoptosis of TSPCs in vitro, and enhances their chondro-osteogenic differentiation and the PI3K inhibitor LY294002 downregulates the expression of chondro-osteogenic genes in TSPCs and significantly alleviates blood-induced tendon HO in rats [100]. Collectively, the PI3K-Akt signaling pathway plays a crucial role in regulating TSPCs osteochondrogenesis and the induction of HO, while targeted inhibition of this pathway represents a promising therapeutic strategy.

NF-κB signaling plays a central role in inflammatory responses, immune reactions, cell proliferation and differentiation, and inhibition of apoptosis [101,102], and is critical at all stages of traumatic HO: it drives the early inflammatory response and participates in regulating subsequent cartilage and bone formation [103]. In TSPCs, the activation of NF-κB promotes aberrant osteochondrogenesis. During the inflammatory phase, NLRP3-dependent macrophage pyroptosis triggers IL-1β and extracellular vesicle release, inducing aberrant NF-κB signaling and promoting TSPCs senescence and osteogenic responses [104]. TSPCs osteochondrogenesis at injury sites depends on NF-κB activation, and targeting NF-κB with palovarotene can reverse the progression of HO [105]. Upstream molecules including TNF-α [106], prostaglandin-endoperoxide synthase 2 (PTGS2) [107], and Sirtuin 1 (SIRT1) [108] all can promote TSPCs osteochondrogenesis by regulating NF-κB signaling, suggesting that the NF-κB pathway serves as a central hub for diverse mechanisms underlying HO.

Antioxidant signaling pathways are also involved in HO formation. The balance between reactive oxygen species (ROS) production and antioxidant defense is increasingly recognized as a key determinant of TSPCs fate in HO. Nuclear factor erythroid 2-related factor 2 (NRF2) is a master transcriptional regulator of antioxidant enzymes, and its activation has been shown to counteract traumatic HO. Carnosic acid, a natural NRF2 activator, reduces intracellular ROS levels in TSPCs by upregulating NRF2 and its downstream targets NQO1, HO-1, CAT, and GSR, while suppressing NADPH oxidase NOX1; this dual action inhibits TSPCs osteochondrogenesis and alleviates HO in the mouse model [109]. Similarly, melatonin protects TSPCs from oxidative stress and preserves their osteogenic potential under H₂O₂-induced damage by increasing superoxide dismutase 2 (SOD2) expression, reducing ROS levels via SIRT1-dependent signaling, and inhibiting SIRT1 with nicotinamide abolishes this protective effect [110]. Other ROS-modulating strategies have also demonstrated efficacy against HO: proanthocyanidin-loaded mesoporous silica nanoparticles scavenge excess ROS and suppress aberrant TSPCs osteochondrogenesis [111]; MIF promotes TDSCs osteogenesis by mediating a ROS/HIF-1α positive feedback loop and activating Wnt/β-catenin signaling [50]; the KLF2/PPARγ axis contributes to traumatic HO by regulating mitochondrial dysfunction and ROS production through SOD2 and CAT [112]; and Hedgehog signaling modulates osteogenesis in TSPCs by regulating antioxidant pathways and ROS generation [67]. Forkhead box O3 (FOXO3), another redox-sensitive transcription factor, has been implicated in tendon healing after rotator cuff injury, where it regulates cellular responses to oxidative stress [113]; however, its specific role in HO has not yet been investigated and warrants future exploration. Collectively, these findings establish antioxidant signaling as a critical regulatory axis in HO pathogenesis and highlight NRF2, SOD2, and potentially FOXO3 as promising therapeutic targets for preventing TSPCs-driven ectopic bone formation.

4. Microenvironmental factors regulating TSPCs function in HO

The differentiation fate of TSPCs is dynamically shaped by the local microenvironment, which evolves in distinct initiation and progression stages following tendon injury. During the initiation stage, acute mechanical disruption and immediate inflammatory bursts serve as primary triggers that aberrantly activate TSPCs osteochondrogenesis. As the pathological process transitions into the progression stage, sustained immune dysregulation coupled with metabolic reprogramming establishes a self-amplifying niche that perpetuates TSPCs toward ectopic bone formation. These extrinsic cues do not act in isolation but sequentially and synergistically intersect with intrinsic signaling networks. This section organizes microenvironmental factors according to their chronological contributions, clarifying how mechanical, immune, metabolic, and aging microenvironment participate in the pathogenesis of TSPCs-mediated HO.

5. Mechanical microenvironment

During the initiation stage, acute mechanical disruption serves as a primary trigger that aberrantly activates TSPCs osteochondrogenesis. As a mechanical transduction hub connecting bones and muscles, tendon resident cells adapt physiologically to regular cyclic tension [114]; however, abnormal biomechanical stimuli such as overloading, stress shielding, or pathological shear force are key initiating factors that drive tendon degeneration and subsequent HO [115]. Phosphorylated focal adhesion kinase (FAK) regulates tendon cell development and their response to external force, significantly influencing tendon size, mechanical properties, cellular composition, and ECM maturity [116], and FAK2 inhibition suppresses both inflammation and osteogenesis in blast limb injury models [117]. Mechanistically, following acute injury, ECM remodeling and sclerosis occur, and TSPCs sense these pathological matrix mechanical changes via integrin receptors, inducing integrin clustering at focal adhesions, FAK phosphorylation, and downstream cytoskeleton rearrangement. This significantly upregulates chondrogenic and osteogenic factors while inhibiting tendon-specific transcription factors, initiating HO. During the progression stage, sustained abnormal mechanical loading reinforces and perpetuates TSPCs toward ectopic bone formation, intersecting with inflammatory and metabolic cues. Temporal transcriptomics and proteomics have identified that discoidin domain receptor 2 (DDR2) modulates ECM collagen arrangement and osteogenic differentiation of mesenchymal lineage cells by mediating the FAK/YAP/TAZ pathway [117]. As a core upstream input for the Hippo pathway, mechanical tension directly drives osteochondrogenic gene expression via YAP/TAZ in stiff or high-tension microenvironments, and also acts as a co-regulator of Wnt/β-catenin and BMP/Smad pathways, integrating physical and chemical signals to synergistically accelerate TSPCs differentiation [118]. Multiple studies have shown that local limb immobilization alleviates mechanically induced ectopic bone formation through various mechanisms (e.g., regulating ECM collagen arrangement and the FAK/YAP/TAZ pathway [33,119], inhibiting TNF-α activity and promoting Mkx transcription [120], clearing neutrophil extracellular traps [121]), highlighting that early clinical immobilization may be a key step in preventing HO.

5.1. Inflammatory and immune microenvironment

During the initiation stage, acute tissue damage triggers an immediate inflammatory storm that serves as the primary trigger for aberrant TSPCs activation. Once injury exceeds a critical threshold, damaged cells and matrix release large amounts of danger-associated molecular patterns (DAMPs), rapidly establishing a proinflammatory milieu rich in factors such as IL-1β and TNF-α [4,122]. In this early phase, neutrophils, monocytes, and mast cells infiltrate the injured site. Recent studies have revealed that NLRP3-dependent macrophage pyroptosis activates pattern recognition receptors, triggering NF-κB and NLRP3 inflammasome cascades, leading to maturation and release of IL-1β. These molecules directly drive osteochondrogenic differentiation of TSPCs while also recruiting and activating more inflammatory cells through a positive feedback loop, forming a vicious cycle that amplifies the initial insult [104,123]. As the pathological process transitions into the progression stage, sustained immune dysregulation establishes a self-amplifying niche that perpetuates TSPCs toward ectopic bone formation. Extensive macrophage recruitment occurs, and their phenotypic polarization plays a decisive role. Liu et al. reported that periostin (POSTN) secreted by M1 proinflammatory macrophages inhibits fatty acid β-oxidation in TSPCs via the PTK7 receptor and initiates their osteogenic differentiation [124]. Although M2 anti-inflammatory macrophages exert antifibrotic effects in later repair stages, their increased infiltration paradoxically promotes endothelial-mesenchymal transition through the MSX2/LEF1/Wnt axis, activating chondrogenic and osteogenic markers and leading to HO [125]. Multiple studies have confirmed that targeting the phenotypic switch from M1 to M2 macrophages alleviates HO severity [108,126]. In addition, adaptive immune T lymphocytes participate in progression-stage regulation. Clinically, peripheral inflammatory T-cell subsets such as IFNγ⁺ CD8⁺ T cells and IL17⁺ CD4⁺ T cells have been identified as potential biomarkers of postoperative HO after elbow trauma [127]. Single-cell and spatial transcriptomic sequencing have shown that infiltrating CD4⁺ and CD8⁺ T cells at injury sites can promote local abnormal osteochondrogenesis [128]. However, no studies have addressed the role of B lymphocytes in HO, nor the mechanisms by which adaptive immune cells directly act on TSPCs, representing a promising direction for future research. In summary, the inflammatory and immune microenvironment of HO is highly complex (Fig. 5) and exhibits marked spatiotemporal heterogeneity. Future mechanistic studies should further investigate how this intricate, stage-dependent microenvironment influences TSPCs osteochondrogenesis.

Fig. 5.

Different immune microenvironments lead to different fate of MSCs. The phenotypic transition of macrophages from M1 to M2 is a key factor driving chondrogenic and osteogenic differentiation, as are IFNγ⁺ CD8⁺ T cells and IL17⁺ CD4⁺ T cells.

5.2. Metabolic microenvironment

During the initiation stage, acute tissue damage triggers immediate metabolic reprogramming that serves as a permissive prerequisite for aberrant TSPCs activation. Comprehensive metabolomic analysis of HO tissues in mice have confirmed alterations in multiple metabolites, including glucose, tricarboxylic acid cycle intermediates, and lipid oxidation-related metabolites, indicating that metabolic reprogramming is a prerequisite for HO development [129]. In this early phase, conflicting glucose metabolic patterns have been reported. Sun et al. performed integrated transcriptomic and metabolomic analyses on induced MSCs derived from FOP patients and found that glucose metabolism shifts from glycolysis to oxidative phosphorylation (OXPHOS) during osteochondrogenesis [130]. In contrast, Kang et al. reported increased expression of hypoxia-inducible factor 1α (HIF-1α) in mesenchymal progenitors during HO, which redirects cellular metabolism toward glycolysis, promoting traumatic HO [131]. Imbalanced lipid metabolism also plays a pivotal role in the initiation stage. Sun et al. demonstrated that fatty acid β-oxidation is significantly reduced in TSPCs during the early stage of HO, a change modulated by extracellular vesicles secreted by necrotic and apoptotic macrophages, ultimately triggering osteogenic differentiation of TSPCs (Fig. 6) [132]. This mechanism reveals crosstalk between the inflammatory-immune and metabolic microenvironments that further primes TSPCs for osteochondrogenesis. As the pathological process transitions into the progression stage, sustained metabolic dysregulation establishes a self-amplifying niche that perpetuates TSPCs toward ectopic bone formation. Beyond fatty acid metabolism, cholesterol metabolism constitutes another major branch of lipid metabolism with critical roles in numerous pathophysiological processes, and its reprogramming has been linked to TSPCs dysfunction and tendinopathy [[133], [134], [135]]. However, the influence of cholesterol metabolism on TSPCs osteochondrogenesis and its role in HO during the progression stage remain largely unexplored. Regulation by the metabolic microenvironment does not act independently. Similar to the complex tumor microenvironment [136], it serves as a key downstream effector of the mechanical and inflammatory-immune microenvironments while also retrogradely modulating cellular mechanosensing, inflammatory cytokine secretion, and immune cell activation during HO progression. It is therefore reasonable to infer that analogous mechanisms operate in the HO microenvironment to influence TSPCs osteochondrogenesis. Overall, the evolutionary dynamics of the metabolic microenvironment after tendon injury, the interactive regulatory networks between different metabolic pathways or microenvironmental compartments, and the molecular mechanisms by which metabolic reprogramming targets and regulates TSPCs osteochondrogenesis warrant further investigation.

Fig. 6.

The key role of lipid metabolism imbalance in the functional regulation of TSPCs osteogenesis and the pathogenesis of HO. Following soft tissue injury, macrophage infiltration and necroptosis coincide with paracrine delivery of EV-derived PAK4. PAK4 directly binds and phosphorylates FABP3 at S122 in TSPCs, suppressing fatty acid oxidation (FAO). This regulatory cascade drives TSPCs osteogenesis and ultimately promotes HO. Reproduced with permission [132]. Copyright 2025, Bone Research.

5.3. Aging microenvironment

Aging profoundly remodels the tendon microenvironment, creating a pro-osteogenic niche that predisposes to HO. Within this aged niche, TSPCs exhibit increased osteochondrogenic differentiation potential, mechanistically linked to elevated expression of BMP-2/4/7 and enhanced BMP/Smad signaling [40]. Consistently, TSPCs isolated from older human donors show higher expression of osteogenic and adipogenic genes, while their chondrogenic potential is also upregulated with age [137]. The aged tendon microenvironment further amplifies HO susceptibility through multiple converging mechanisms. Disruption of the circadian clock gene Bmal1 accelerates premature aging and drives progressive tendon/ligament HO via activation of TGF-β/BMP signaling [138], and loss-of-function mutations in Enpp1 promote age-dependent heterotopic ossification of spinal ligaments and Achilles tendons by upregulating Hedgehog signaling [139]. In addition, disturbed glycolipid metabolism in aged or diabetic conditions activates the CXCL13-CXCR5 autocrine axis in senescent tendon-derived stem cells, creating a positive feedback loop that reinforces aberrant osteogenic differentiation [140]. Collectively, these findings indicate that the aging microenvironment is not a passive backdrop but an active biological driver that remodels both TSPCs intrinsic properties and extrinsic niche signals, thereby lowering the threshold for HO initiation and accelerating its progression. Future studies should aim to dissect the age-related epigenetic and metabolic clocks that control TSPC fate within this specialized niche and to develop geroprotective strategies that restore youthful differentiation plasticity in the aged tendon microenvironment.

6. Potential strategies for therapeutic targeting of TSPCs in HO

Strategies for the prevention and treatment of HO are transforming from traditional approaches (surgery, physical therapy, NSAIDs, etc.) toward precision paradigms. With the gradual elucidation of the central role of TSPCs in HO development, as well as their surface markers, signaling pathways, and functional characteristics, targeted regulation of this key cell population at multiple levels and dimensions has become a hot topic in cutting-edge research in this field. An ideal targeted strategy should exhibit high cellular specificity, precisely aberrant TSPCs osteochondrogenesis, and simultaneously promote their differentiation toward normal soft tissue phenotypes such as tendons and ligaments or maintain their quiescent state, without interfering with normal bone metabolism and systemic homeostasis. Current relevant research is mainly focused on two directions: 1) Intervention of the related signaling pathways or key molecules using small molecule compounds or biological macromolecules, with emerging emphasis on subpopulation-specific targeting (e.g., inhibiting TNC-Hippo signaling in CD26⁺ TSPCs, or blocking Hedgehog pathway in Gli1⁺ TSPCs) to achieve selectivity while sparing other TSPCs subsets required for normal tendon repair; 2) Construction of composite biomaterials to improve the local microenvironment or achieve precise delivery of targeted drugs, potentially coupled with subpopulation-recognizing ligands (e.g., anti-CD26 antibodies or Gli1-responsive elements) for cell-specific therapy. These two approaches are not isolated but exhibit a trend of synergistic integration. Combined interventions are expected to overcome the limitations of monotherapy and provide more efficient solutions for the clinical translation of precise HO prevention and treatment.

6.1. Pharmacological intervention strategies targeting key molecules and pathways

Pharmacological intervention targeting specific receptors on the TSPCs surface and intracellular signaling nodes is one of the most extensively studied strategies and closest to clinical translation (Table 2). Its theoretical basis lies in the presence of multiple molecular switches during TSPCs osteochondrogenesis that can be specifically regulated by small-molecule drugs.

Table 2.

Drug intervention strategies that targeting key molecules and pathways.

Drugs	Mechanism	References
Targeting BMP signaling pathway
LDN-193189	Binding to ALK2/3 receptors on the surface of TSPCs, and blocking Smad1/5/8 phosphorylation and nuclear translocation	[141]
LDN-212854	ALK2 bias inhibitor	[22,142]
Noggin	Directly neutralizing excess BMP ligands in TSPCs microenvironment	[40,143]
ALK3-Fc	BMP ligand trap, efficiently binding BMP-2 and BMP-4	[144,145]
NG-25	TAK1 inhibitor	[146]
Palovarotene	RARγ agonists, inhibiting the phosphorylation of Smad1/5/8, and inhibiting the downstream BMP/ALK2/Smad pathway; Inhibiting the synergistic effect of Smad and NF-κB signaling pathway	[105,147]
Targeting Hedgehog signaling pathway
JQ1	BET bromodomain inhibitor, suppressing the activity of GLI	[23]
GANT58	GLI specific inhibitor	[105]
Targeting Hippo signaling pathway
XMU-MP-1	Targeting TNC, an upstream regulator, to significantly inhibit phosphorylation and degradation of YAP	[10]
Targeting PI3K signaling pathway
LY294002	PI3K inhibitor, downregulating PI3K-Akt signaling pathway	[100]
Rapamycin	mTOR inhibitor, inhibiting PI3K-Akt-mTOR pathway	[148,149]
BYL-719	PI3Kα inhibitor, inhibiting the activity of Smad, Akt and mTOR	[96]
Targeting metabolic reprogramming
2-DG	Blocking glycolysis, and down regulating HIF-1α driven glycolytic metabolic reprogramming	[131]
AAV-siCxcl3	Inhibiting autocrine axis of CXCL13-CXCR5 in senescent TSPCs	[140]

Targeting the BMP signaling pathway is the most widely investigated direction. Rui et al. confirmed that TSPCs express higher levels of BMPRIA (ALK3), BMPRIB (ALK6), and BMPRII receptors than BMSCs [150]. Lui et al. also found that TSPCs derived from injured tendons show significantly enhanced sensitivity to Smad signaling upon BMP-2 stimulation, accompanied by an aggravated abnormal chondro-osteogenic phenotype [151]. Moreover, lineage tracing revealed that specific overexpression of ALK2 (ACVR1) in Scx + TSPCs is sufficient to drive HO formation at the Achilles tendon enthesis [25]. Accordingly, the BMP receptor kinase inhibitor LDN-193189 competitively binds to ALK2/3 receptors on TSPCs, effectively blocking Smad1/5/8 phosphorylation and nuclear translocation. Peterson et al. demonstrated that it inhibits BMP-2-induced TSPCs osteochondrogenesis and reduces the severity of traumatic HO [141]. Meanwhile, the ALK2-biased inhibitor LDN212854 exhibits favorable therapeutic effects on both genetic FOP-related HO and traumatic HO [22,142]. In recent years, other potential drugs targeting the BMP pathway have emerged. The BMP antagonist Noggin directly neutralizes excess BMP ligands in TSPCs microenvironment and specifically blocks TSPCs osteochondrogenesis in tendon injury models [40,143]. The BMP ligand trap ALK3-Fc binds BMP-2 and BMP-4 with high affinity, markedly suppressing abnormal osteogenic differentiation of early stem/progenitor cells in vitro and ectopic bone volume in soft tissues in vivo with fewer adverse effects [144,145]. As a downstream effector of the non-canonical BMP pathway, TGF-β activated kinase-1 (TAK1) inhibitor NG-25 also exerts therapeutic effects on TSPCs abnormal differentiation and HO both in vitro and in vivo [146]. Notably, Palovarotene, an orally bioavailable retinoic acid receptor γ (RARγ) agonist, can inhibit Smad1/5/8 phosphorylation and downstream BMP/ALK2/Smad-dependent chondrogenesis and osteocyte differentiation, thereby reducing endochondral ossification, which was approved by the FDA in 2023 for the treatment of FOP [147]. One study reported that Palovarotene attenuates TSPCs-mediated HO by suppressing the synergistic effect of Smad and NF-κB signaling following inflammatory microenvironment stimulation, providing potential evidence for expanding its clinical indication to traumatic HO [105]. However, most other candidate drugs discussed above (e.g., LDN-193189 [141], LDN-212854 [22,142], Noggin [40,143], ALK3-Fc [144,145], NG-25 [146], GANT58 [105], XMU-MP-1 [10], LY294002 [100], and 2-DG [131] et al.) remain at preclinical stages, with efficacy demonstrated only in rodent models of HO, and none have yet entered clinical trials for traumatic HO. Thus, translating these promising agents into human therapies requires rigorous validation of safety, pharmacokinetics, and target specificity in patient-relevant settings.

The Hedgehog signaling pathway is a key driver of TSPCs osteochondrogenesis [152]. Feng et al. reported that JQ1, a small-molecule inhibitor targeting BRD4, alleviates HO by downregulating Hedgehog activity through suppressing transcription of Gli1 and Gli2 [23]. The specific GLI inhibitor GANT58 also inhibits osteochondrogenic differentiation and ROS production in TSPCs in vitro, reducing tendon ectopic bone volume after trauma [67]. Given that Vismodegib and Sonidegib, as SMO receptor inhibitors, have been widely used in the treatment of basal cell carcinoma of the skin, whether they can modulate TSPCs osteochondrogenesis to treat HO urgently requires validation. However, caution is needed as Hedgehog signaling plays a dual role in tendon biology. Fang et al. reported that Hedgehog signaling is essential for rotator cuff enthesis formation [66], so complete inhibition may impair normal tendon enthesis healing, suggesting that precise regulation is required for Hedgehog pathway intervention in TSPCs. For the Hippo signaling pathway, XMU-MP-1, which targets its upstream regulator TNC, significantly suppresses YAP phosphorylation and degradation and mitigates TSPCs-mediated ectopic bone formation [10]. Nevertheless, its dose safety and systemic side effects remain insufficiently validated, leaving considerable room for further research. In addition, inhibitors targeting the PI3K-Akt-mTOR pathway, such as LY294002 [100] and rapamycin [148,149], show potent inhibitory effects on abnormal TSPCs osteochondrogenesis. Beyond these canonical pathways, Kong et al. found that HIF-1α-driven glycolytic reprogramming promotes HO by upregulating glucose transporters and glycolytic enzymes (e.g., HK2, ENO1), and blockade of glycolysis with 2-deoxy-D-glucose (2-DG) markedly inhibits osteogenic differentiation of mesenchymal progenitors [131]. Chen et al. reported that glycolipid metabolic disorders activate the autocrine CXCL13-CXCR5 axis in senescent TSPCs, which in turn positively reinforces their abnormal osteogenic differentiation, while genetic silencing of CXCL13 effectively alleviates HO in diabetic mice [140]. These findings indicate that interventions targeting metabolic reprogramming in TSPCs present unique advantages and represent a promising therapeutic strategy.

6.2. Composite biomaterials and local smart delivery systems

For the prevention and treatment of aberrant osteochondrogenic differentiation in TSPCs, composite biomaterials and local smart delivery systems offer unique advantages in regulating TSPCs fate by constructing biomimetic microenvironments that enable spatiotemporally controlled release of drugs or bioactive factors (Table 3).

Table 3.

Composite biomaterials and local intelligent delivery system.

Biomaterials	Mechanism	References
PA loaded MSNs	Inhibiting STAT1 phosphorylation, promoting M2 macrophage polarization, and indirectly downregulating mTOR pathway	[153]
Capsaicin-loaded ROS-responsive hydrogels	Inhibiting PI3K-Akt-mTOR axis to reduce proinflammatory factor secretion	[154]
GDSH	Inhibiting MAPK, AMPK, Smad, Hippo and PI3K/Akt signaling pathways	[155]
rKL@MPs-ASF	Alleviating oxidative stress in TNF-α-induced inflammatory microenvironment	[156]
3D-CBD@LDN/Exos	Inhibiting BMP type I receptors ALK2 and ALK3, and preventing Smad1/5/8 phosphorylation and nuclear translocation	[15]
CHP-PLGA-RAPA nanoparticles	Inhibiting mTOR signaling	[157]
Freeze-dried porous hydrogel loaded with chiral lysine	Intervening IL-1β/NETs/AIM2 axis and regulating the cross talk between neutrophils and macrophages	[158]
Hydrogel with parallel-grooved topographical cues fabricated in 3D sandwich model	Enhancing PI3K-Akt signaling pathway	[159]
D-ZnO or L-ZnO nanoparticles	Clearing H2S and block Ca2+/ERK pathway	[160]
Local administration of FGF9 in Matrigel	Activating FGFR3-BMPR1a pathway to promote local lymphangiogenesis and regulates post-traumatic inflammatory microenvironment	[161]
GsMTx4-loaded GelMA	Inhibiting mechanical sensitive ion channel Piezo1, and activating Apelin signaling pathway	[162]
LT-NPs coupling Licochalcone A	Inducing HSP70 to block mtDNA leakage and reprogramming macrophage phenotype from M1 to M2 by targeting cGAS-STING axis	[163]
PPP-loaded MSNs integrated with PCL@HA-ADH@PA/Fe	Inhibiting PI3K-Akt pathway, activating TSPCs mitophagy, restoring immune homeostasis, and suppressing osteogenesis while enhancing osteoclastogenesis	[164]

Targeted hydrogel delivery represents a powerful approach to inhibit aberrant osteochondrogenic differentiation of TSPCs and mitigate HO (Fig. 7). Capsaicin-loaded ROS-responsive hydrogels upregulate tenogenic markers and collagen synthesis and in downregulates osteogenic markers and ALP activity in TSPCs under IL-1β-induced inflammatory conditions, while inhibiting the PI3K-AKT-mTOR axis to reduce proinflammatory cytokine secretion, promoting ordered collagen regeneration and suppressing HO in Achilles tendon defect models [154]. Moreover, GsMTx4-loaded GelMA hydrogels combined with short-term immobilization can inhibit the mechanosensitive ion channel Piezo1 and activate the Apelin signaling pathway, promoting tendon regeneration while suppressing the transformation of TSPCs into osteoblasts and chondrocytes, then alleviates HO development [162]. Furthermore, a composite delivery system of PPP-loaded MSNs (IGF-1R inhibitor) integrated with ferric ion-crosslinked hydrogel patches exerted anti-HO effects via sustained release of active components to regulate the HO microenvironment multidimensionally, inhibit the PI3K-Akt pathway and activate TSPCs mitophagy, with favorable efficacy and biosafety in rat models [164]. Dang et al. also synthesized a glucose-responsive dual-drug sequential-release hydrogel composed of oxidized hyaluronic acid–dopamine and phenylboronic acid-functionalized carboxymethyl chitosan. Loaded with irisin and connective tissue growth factor via dendritic mesoporous silica nanospheres, this system exerts anti-inflammatory effects and promotes normal tenogenic differentiation of TSPCs while inhibiting abnormal osteochondrogenic differentiation through multiple signaling pathways in a diabetic rat model of chronic tendinopathy [155].

Fig. 7.

Novel composite hydrogels exhibit great potential in the treatment of HO. Hydrogels function via diverse mechanisms, such as: 1) Remodeling membrane receptors (purple). Reproduced with permission [162]. Copyright 2026, Elsevier. 2) Regulating intracellular signaling (green). Reproduced with permission [154]. Copyright 2026, Materials Today Bio.

At the nanodelivery system level, mesoporous silica nanoparticles (MSNs) loaded with the natural small molecule Parishin A promote M2 macrophage polarization by inhibiting STAT1 phosphorylation, indirectly downregulating the mTOR pathway and suppressing TSPCs osteochondrogenesis [153]. Meanwhile, the sustained-release property of MSNs significantly reduces the dosage and injection frequency of Parishin A, achieving dual effects of regulating the tendon immune microenvironment and preventing heterotopic ossification [153]. In addition, a recombinant α-Klotho (rKL)-loaded silk fibroin microsphere-integrated aligned silk fibroin scaffold (rKL@MPs-ASF) protects the tenogenic differentiation potential of TSPCs by alleviating oxidative stress under TNF-α-induced inflammation, reducing inflammatory cell infiltration and HO formation [156]. Xu et al. reported that exosomes derived from CD26⁺ TSPCs pretreated with the BMP inhibitor LDN193189, modified with collagen-binding peptides (CBD) and loaded onto 3D-printed scaffolds, achieve synergistic tendon regeneration and HO prevention via a dual-targeting strategy in aged rat Achilles defect models [15]. On the other hand, collagen hybrid peptide (CHP)-modified PLGA nanoparticles enable targeted delivery of rapamycin to pathological tendon collagen, blocking chondro-osteogenic differentiation of TSPCs by inhibiting mTOR signaling and significantly suppressing HO progression in mouse models [157]. Yuan et al. innovatively developed a novel biomaterial combining chiral lysines with bacterial cellulose, which regulates neutrophil–macrophage crosstalk by interfering with the IL-1β/NETs/AIM2 axis, blocking macrophage pyroptosis and subsequent aberrant chondro-osteogenic differentiation of TSPCs, thereby preventing traumatic HO [158]. In terms of topological regulation, sandwich-like biomimetic scaffolds with parallel groove-structured hydrogel surfaces fabricated by 3D printing maintain aligned growth of TSPCs in a 3D microenvironment, enhance tenogenic differentiation via the PI3K-Akt pathway, and attenuate inflammatory phenotypes, promoting tendon regeneration and reducing HO in vivo [159]. Meanwhile, D- or L-type zinc oxide (ZnO) nanoparticles, as effective hydrogen sulfide scavengers, inhibit endogenous H2S-promoted abnormal osteogenic activity of TSPCs by blocking the Ca²⁺/ERK pathway [160]. Zhang et al. explored a lymphatic-targeting strategy and found that Matrigel loaded with FGF9 promotes local lymphangiogenesis by activating the FGFR3-BMPR1a pathway, regulating the post-traumatic inflammatory microenvironment and inhibiting HO [161]. A mild photothermal cascade nanoplatform can couple licochalcone A delivery with near-infrared photothermal therapy, while mild heat stress (about 42°C) induces HSP70 to block mitochondrial DNA leakage and inhibit the cGAS-STING axis, reprogramming M1 macrophages toward an M2 phenotype, suppressing TSPCs senescence, uncoupling the vicious cycle of inflammation and degeneration, and significantly inhibiting HO and restoring biomechanical function of tendon [163].

In summary, these composite biomaterials and smart delivery systems, through interdisciplinary integration of materials science, biology, and engineering, provide multidimensional therapeutic approaches targeting TSPCs for HO prevention and treatment, ranging from molecular intervention to tissue regeneration.

7. Conclusions and future research directions

HO is a common musculoskeletal disorder that severely impairs patient function, with its major challenge in prevention and treatment lying in the incompletely elucidated pathogenesis. Focusing on TSPCs as the key cell population, this review systematically summarizes the lineage tracing evidence and clinicopathological basis of their involvement in HO, elaborates on how core signaling pathways such as BMP, Wnt/β-catenin, Hedgehog, and Hippo coordinately regulate the osteochondrogenic differentiation fate of TSPCs, and analyzes the molecular mechanisms by which disorders in mechanical, inflammatory-immune, and metabolic microenvironments drive their pathological transformation. On this basis, receptor inhibitors, signaling pathway modulators, emerging biomaterials, and targeted delivery systems targeting aberrant TSPCs osteochondrogenesis show promising prospects for prevention and treatment, with some strategies already entering the clinical translation stage. However, current research still has obvious limitations regarding TSPC subpopulation heterogeneity, multi-factor microenvironmental crosstalk, epigenetic and metabolic reprogramming targets, and the specificity and safety of existing interventions. To address these gaps, future research should focus on four priority areas: 1) Spatiotemporal regulation: Integrate single-cell multi-omics and spatial transcriptomics to construct a dynamic map of TSPCs differentiation trajectories, with particular emphasis on resolving the distinct fates of distinct subpopulations during HO initiation and progression; 2) Epigenetic editing: Identify key epigenetic modifiers (e.g., chromatin remodelers, DNA methyltransferases, histone-modifying enzymes) that govern the transcriptional switch from tenogenic to osteochondrogenic programs, and develop targeted epigenetic editing tools to reverse aberrant TSPCs fate; 3) Human organoids: Establish humanized models such as tendon organoids from healthy donors and HO patients to screen subpopulation-selective inhibitors and recapitulate human disease features that animal models cannot capture; 4) Subtype-targeted therapy: Move beyond broad pathway inhibition toward cell-subset-restricted interventions by targeting subpopulation-specific surface markers or their unique signaling dependencies, thereby achieving precision prevention of HO while preserving normal tendon repair functions.

Author contributions

Conceptualization: XHZ, BWL; Methodology: XLZ, YG; Investigation: XLZ, BWL; Visualization: XLZ, MLC, HJ; Project administration: RG, HT; Supervision: TLL, BWL; Writing – original draft: XLZ, YG; Writing – review & editing: XHZ, BWL.

Data and materials availability

The data that support the findings of this study are available from corresponding author upon reasonable request.

Declaration of AI and AI-assisted technologies in the writing process

During the preparation of this work, the author did not used AI and AI-assisted technologies.

Funding

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Declaration of competing interest

All authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this manuscript.

Specifically, no author has any financial or personal relationships with other people or organizations that could inappropriately influence or bias their work, including but not limited to employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or other funding, patents or patent applications, and personal or professional relationships that might create a conflict of interest.

Abbreviations list

Abbreviation	Full term
2-DG	2-deoxy-D-glucose
CBD	collagen-binding peptide
CHP	collagen hybrid peptide
CTHRC1	collagen triple helix repeat-containing 1
DAMPs	danger-associated molecular patterns
DDR2	discoidin domain receptor 2
DHH	Desert hedgehog
EGR1	Early growth response 1
FAK	focal adhesion kinase
FOP	fibrodysplasia ossificans progressiva
FOXO3	Forkhead box O3
HIF-1α	hypoxia-inducible factor 1-alpha
HO	heterotopic ossification
IGF-1	insulin-like growth factor 1
IHH	Indian hedgehog
MAPK	mitogen-activated protein kinase
MIF	macrophage migration inhibitory factor
MKX	Mohawk
MSCs	mesenchymal stem cells
MSNs	mesoporous silica nanoparticles
NICD	Notch intracellular domain
NRF2	nuclear factor erythroid 2-related factor 2
NT-3	neurotrophin-3
OXPHOS	oxidative phosphorylation
POH	progressive osseous heteroplasia
POSTN	periostin
PTCH1	Patched 1
PTGS2	prostaglandin-endoperoxide synthase 2
RARγ	retinoic acid receptor gamma
rKL	recombinant α-Klotho
ROS	reactive oxygen species
RTKs	receptor tyrosine kinases
scRNA-seq	single-cell RNA sequencing
SCX	Scleraxis
SHH	Sonic hedgehog
SIRT1	Sirtuin 1
SMO	Smoothened
SOD2	superoxide dismutase 2
SPARC	secreted protein acidic and rich in cysteine
TAK1	TGF-β activated kinase-1
TAZ	transcriptional coactivator with PDZ-binding motif
TDSCs	tendon-derived stem cells
TGF-β	transforming growth factor-beta
TNC	Tenascin-C
TSPCs	tendon stem/progenitor cells
YAP	Yes-associated protein
ZnO	zinc oxide