The Notch signaling pathway in regulating bone and cartilage homeostasis: novel insights into the pathogenesis and therapeutics of osteoarthritis

Abstract

Osteoarthritis (OA) is a chronic degenerative joint disorder characterized by cartilage degradation, bone hyperplasia, and synovitis. It is a primary cause of joint pain and dysfunction globally. The increasing incidence of OA, driven by an aging population, profoundly impairs patients’ quality of life while imposing a growing economic burden on societies and families. Thus, in-depth investigations into the pathogenesis of OA, along with the exploration of novel therapeutic targets and strategies, hold significant clinical implications and social value. The Notch signaling pathway, a highly conserved intercellular communication pathway, plays a pivotal regulatory role in cell proliferation, differentiation, apoptosis, and organogenesis. In recent years, a growing body of research has revealed that the Notch signaling pathway is crucial for maintaining bone and cartilage homeostasis, with its aberrant activation or inhibition being closely linked to the initiation and progression of OA. Therefore, this narrative review performed an extensive PubMed database search using keywords like “Notch”, “osteoarthritis”, “bone”, “cartilage”, “synovitis”, “osteoblasts”, “osteoclasts”, and “chondrocytes”, and reviewed all pertinent literature. It specifically focuses on the role of Notch signaling in the differentiation and function of osteoblasts, osteoclasts, and chondrocytes, shedding light on its mechanism in cartilage damage, subchondral bone dysfunction, and synovitis. It also explores evidence for targeted Notch pathway therapies in OA, aiming to illuminate the molecular mechanisms underlying OA pathogenesis and offer new theoretical insights and therapeutic targets for OA prevention and treatment. Additionally, this narrative review seeks to decipher the mechanisms underlying the context-dependent duality of Notch signaling in bone and cartilage, and provides a critical appraisal of the challenges confronting current targeted therapies.

Introduction

Osteoarthritis (OA) is the most prevalent joint disorder globally, imposing substantial mental and physical burdens on the elderly population. With nearly 250 million individuals affected, it ranks as the leading cause of disability among older adults. OA is a whole-joint disease characterized by cartilage damage, subchondral bone sclerosis, osteophyte formation, and structural alterations in periarticular muscles, which collectively lead to pain and joint dysfunction [1–4]. Its core pathological features involve the degeneration of cartilage and subchondral bone, primarily manifesting as cartilage peeling, fragmentation, and wear, accompanied by periosteal bone proliferation and joint instability. Articular cartilage, predominantly hyaline cartilage, consists of a dense extracellular matrix (ECM) embedded with chondrocytes and water. It lines the bone surfaces, providing a low-friction interface that absorbs external pressure and stimuli, enabling pain-free joint movement [5]. In OA, progressive articular cartilage degeneration results in the loss of joint mobility and function, ultimately impairing patients’ quality of life due to pain and lifestyle limitations. Subchondral sclerosis is widely regarded as a definitive hallmark of OA [6]. However, emerging evidence from certain studies indicates that the subchondral bone experiences distinct microstructural alterations at various stages of OA. Notably, subchondral sclerosis may only manifest during the advanced stages of OA [7]. In the initial phases of OA, an upsurge in bone remodeling activity is observed, accompanied by subchondral bone loss [8, 9]. As the primary structural components of joints, bone and cartilage maintain their homeostasis through intricate signaling networks. Among these, the Notch signaling pathway has garnered significant research attention owing to its pervasive involvement in cellular processes such as proliferation, differentiation, and apoptosis [10]. Initially identified in Drosophila, the Notch pathway has since been confirmed to play critical roles in mammalian skeletal development, including chondrogenesis, ossification, and joint cavity formation [11]. In recent years, its role in OA has been increasingly elucidated. Aberrant regulation of Notch signaling not only affects chondrocyte survival and phenotypic maintenance but also participates in subchondral bone remodeling by modulating the osteoblast-osteoclast balance, thereby accelerating OA progression [12, 13]. Therefore, this narrative review systematically summarizes the regulatory roles of the Notch signaling pathway in bone and cartilage homeostasis, as well as its pathological significance in OA, aiming to provide novel insights for the prevention and treatment of this disease. Furthermore, several fundamental questions remain unresolved in the field. These include the dynamic regulatory mechanisms of Notch signaling across different OA pathological stages, the functional heterogeneity of cell-specific receptors, and the therapeutic potential of “temporal therapy” (e.g., early activation and late inhibition). Exploring these specific controversies and gaps is, therefore, the central aim of this review.

Overview of the Notch signaling pathway

Components of the Notch signaling pathway

The Notch signaling pathway is primarily composed of Notch receptors, Notch ligands, and downstream effector molecules [14]. Notch receptors are single-pass transmembrane proteins, with four subtypes identified in mammals, namely Notch1 to Notch4 [15, 16]. Each receptor consists of three main segments: the Notch extracellular domain (NECD), transmembrane domain (TMD), and Notch intracellular domain (NICD) [17]. The NECD of all Notch receptors contains 29–36 tandem epidermal growth factor (EGF)-like repeats, with Notch1–4 harboring 36, 36, 34, and 29 repeats, respectively [18]. Downstream of the EGF repeats lies a unique negative regulatory region (NRR), which comprises three cysteine-rich Lin12-Notch repeats (LNRs) and a heterodimerization domain (HD). The C-terminus of the NECD contains the three LNRs and forms an NRR with HD, which prevents receptor activation in the absence of ligands [19]. The TMD of Notch contains cleavage sites recognized by γ-secretase complexes, which are critical for signal activation. Within the NICD, the RBPJ (recombinant signal binding protein-J) associated module (RAM) domain is a high-affinity binding motif of 12–20 amino acids, centered around a conserved WxP sequence. A long, unstructured linker containing nuclear localization sequences (NLS) connects the RAM domain to seven ankyrin repeats (ANK). Both the RAM domain and ANK repeats are known to mediate interactions with CBF1/suppressor of hairless/Lag1 (CSL, also called RBPJ) transcription factors [20]. Flanking the ANK domain are two NLSs and an evolutionary divergent transactivation domain (TAD) [21]. Notably, Notch1 and Notch2 possess functional TADs, whereas Notch3 and Notch4 lack this domain [22]. At the C-terminus of the NICD resides a proline (P)-, glutamate (E)-, serine (S)-, and threonine (T)-rich motif (PEST domain), which serves as a substrate for ubiquitin ligases, targeting NICD for proteasomal degradation [14] (Fig. 1A).

Fig. 1.

Structure and cleavage sites of the Notch signaling pathway. A Structural features of Notch receptors (Notch1, Notch2, Notch3, and Notch4). B The structure of Notch ligands (Jag1, Jag2, DLL1, DLL3, and DLL4). C Cleavage sites of the Notch receptor and their corresponding cleavage components. (Created with https://BioRender.com)

Notch ligands, also referred to as DSL proteins, are categorized into two families: Delta-like (DLL1, DLL3, and DLL4) and serrate-like ligands Jagged (Jag1 and Jag2). DLL3 and DLL4 are considered “classical” Notch ligands, with confirmed interactions with the NECD [23]. Additionally, other soluble and transmembrane proteins can bind to Notch receptors, potentially modulating Notch signaling activity. Most Notch ligands are type I transmembrane proteins, characterized by three key structural motifs: an N-terminal DSL (Delta/Serrate/LAG-2) motif, specialized tandem 5–16 EGF repeats that collectively adopt an extended conformation spanning approximately 30–65 nm [24]. The Jagged family contains 16 EGF-like domains, whereas the Delta-like family has 5–9, though the exact number remains to be fully clarified. Compared to Delta-like proteins, Jagged proteins also possess an additional cysteine-rich domain (CRD) positioned downstream of the EGF-like repeats, which may further regulate ligand-Notch binding efficiency or specificity (Fig. 1B).

The activation process of the Notch signaling pathway

The activation of the Notch signaling pathway can be categorized into the canonical and non-canonical pathways. Maturation and activation of canonical Notch receptors involve four sequential proteolytic events, designated S1 to S4 [14] (Fig. 1C). Initially, Notch proteins are synthesized as single-chain precursors and transported to the endoplasmic reticulum (ER), where the EGF-like domains of Notch receptors undergo glycosylation [25]. The glycosylated single-chain Notch precursor is then transported to the Golgi apparatus, where the S1 cleavage site—within an unstructured loop protruding from the HD subdomain—is cleaved by furin-like convertases at the RQRR motif, specifically between residues R1654 and E1655 [26]. This S1 cleavage converts the Notch polypeptide into a mature heterodimeric receptor. Subsequently, the mature Notch receptor is transported to the cell surface, where it binds to transmembrane Notch ligands (e.g., Delta-like or Jagged) expressed on adjacent cells. This interaction triggers cleavage at the S2 site (predominantly at residue A1710-V1711) by a disintegrin and metalloproteinase (ADAM) 10/17, generating a transient intermediate termed “Notch extracellular truncation (NeXT)” that comprises the TMD and NICD. Following S2 cleavage, the NRR is exposed, and the bound NECD-ligand complex is internalized by the signal-sending cell through endocytosis, which is essential for ligand-dependent Notch activation [14, 27]. Next, the S3 and S4 sites on the Notch protein are cleaved by a γ-secretase complex that includes presenilin 1/2 (PS1/PS2), nicastrin, anterior pharynx defective-1 (Aph-1), and presenilin enhancer-2 (Pen-2). This complex searches for protein degradation substrates within the TMD. It achieves this by targeting the β-chain, which is recognized by the γ-secretase induced at the carboxyl-terminus of Notch's transmembrane helical region [28, 29]. γ-secretase mediates cleavage at S3/S4, initiating at Val (V1744) to release NICD. This liberation of NICD allows it to translocate into the nucleus to regulate transcription [30]. In the absence of NICD, Rbpjκ (also known as CSL) associates with corepressors to inhibit transcription. Upon nuclear translocation, NICD forms a ternary complex with Rbpjκ and MAML via its RAM domain, recruiting transcriptional co-activators to displace corepressors [31]. CSL then recognizes and binds to specific DNA sequences, thereby regulating the transcription of downstream target genes, such as members of the Hes/Hey family and modulating cell differentiation [32–34]. These interactions convert the original “corepressor complex” into a “coactivator complex”, promoting transcriptional activation of Notch target genes (Fig. 2).

Fig. 2.

Typical and atypical conduction routes within the Notch Signaling Pathway. Activation of the canonical Notch pathway begins with ligand binding, followed by sequential proteolysis by ADAM10/17 and γ-secretase. This cleavage releases the NECD for endocytosis and the NICD, which enters the nucleus to form transcriptional activation complexes with CSL. Alternatively, the non-canonical (CSL-independent) pathway can regulate Wnt/β-catenin, JAK/STAT, PI3K/AKT, and NF-κB pathways to exert its non-canonical biological functions. (Created in https://BioRender.com)

The non-canonical Notch signaling pathway is CSL-independent [16]. Currently, research in this area remains relatively limited. One identified mechanism involves the ANK domain of Notch receptors binding to the intracellular zinc finger protein Deltex, which inhibits the activity of the transcription factor E47 and thereby regulates the expression of related genes [35]. Additionally, Notch can modulate other signaling pathways at the post-translational level, such as the Wnt/β-catenin pathway, JAK/STAT pathway, PI3K/AKT pathway, and NF-κB pathway. These interactions form a complex signaling regulatory network, enabling Notch to exert its non-canonical biological functions [36, 37] (Fig. 2).

As an evolutionarily conserved signaling pathway, Notch exerts spatiotemporally specific effects that govern diverse cellular processes, including proliferation, differentiation, and apoptosis. Its physiological roles have been investigated across multiple systems involved in organogenesis and tissue renewal [38].

The regulatory role of the Notch signaling pathway in bone homeostasis

Bone homeostasis constitutes the core mechanism for preserving bone health and function, with its delicate equilibrium relying on the dynamic interplay between two key cell populations, osteoblasts and osteoclasts. The Notch signaling pathway is indispensable for bone regeneration induced by bone morphogenetic proteins (BMPs), and its role in maintaining bone homeostasis primarily revolves around regulating the differentiation, function, and survival of osteoblasts and osteoclasts themselves [39].

Notch signaling and osteoblasts

Notch signaling exerts a bidirectional regulatory effect on osteoblast differentiation, which appears to depend on cell type, differentiation stage, coculture microenvironment, timing of Notch activation, and crosstalk with other signaling pathways [40, 41] (Fig. 3A).

Fig. 3.

Mechanisms by which the Notch signaling pathway regulates osteoblasts (A), osteoclasts (B), and chondrocytes (C). A The Notch signal regulates MSC differentiation into osteoblasts by BMP/Smad, Wnt/β-catenin, and NF-κB pathways, balancing osteoblast proliferation and differentiation. B Notch signal regulates the differentiation of osteoclasts and myeloid progenitor cells by modulating NFATc1/RANKL signaling, influenced by upstream factors (TNF-α, NSUN2, and Tspan-5/10). C Notch signal controls MSC chondrogenic differentiation and chondrocyte dedifferentiation. Its dysregulation drives cartilage degeneration and endochondral ossification via IL-6/p-STAT3. (Created with https://BioRender.com)

The Notch signaling pathway exerts an inhibitory influence on osteoblast differentiation and bone formation processes. On one hand, targeted deletion of Notch2 in osteoblasts results in a decline in the levels of NICD. This is accompanied by a reduced expression of its target genes, Hey2 and Hes7, which in turn enhances osteoblast function and leads to a substantial increase in trabecular bone mass in the proximal femur and distal tibia [42]. Concurrently, Notch2 deficiency promotes glycolysis and bone formation in the long bones of postnatal mice. In contrast, activation of either Jag1 or the intracellular domain of Notch2 can suppress glucose metabolism and osteogenic differentiation of bone marrow mesenchymal stem cells (BMSCs). This suppression is manifested by decreased glycolysis and mitochondrial complex I gene expression, ultimately leading to reduced mitochondrial respiration, superoxide production, and AMP-activated protein kinase (AMPK) activity [43]. On the other hand, Hes1, a target gene of the RBPjκ-dependent Notch signaling pathway, plays a dual role in BMSCs biology. It promotes the proliferation of BMSCs while simultaneously inhibiting their differentiation into osteoblasts [44]. It is noteworthy that the Notch pathway can induce osteoblast differentiation not only through direct alterations in its downstream factors but also potentially via interactions with the β-catenin and Wnt signaling pathways [45]. For instance, activation of Notch signaling inhibits osteoblast differentiation by overexpressing its target gene Hes1. Conversely, ED-71 can counteract this effect by inhibiting Notch signaling and activating the Wnt/GSK-3β/β-catenin signaling pathway, thereby reversing the inhibitory process on osteoblast differentiation [46]. Additionally, NICD can block the Wnt signaling pathway by reducing Wnt3a transactivation and Wnt/β-catenin expression, without affecting the BMP signaling pathway. This selective inhibition impairs the biological effects of BMP-2 and Wnt, ultimately hindering osteoblast generation [47]. Silencing of Hes-1 (si-Hes) can partially mitigate the inhibitory effects of NICD. Similarly, overexpression of the NICD inhibits osteoblast differentiation and mineralization by downregulating the Wnt/β-catenin signaling pathway, leading to decreased expression of osteocalcin, type I collagen (COL1A1), alkaline phosphatase (ALP) transcripts, and Delta2Delta FosB protein [48]. Other mechanisms include the formation of a complex between Foxo1, NICD, and Mastermind, which inhibits osteoblast formation [49]. Spalt-Like Transcription Factor 4 (SALL4) interacts with Notch2, blocking the nuclear translocation and target gene expression of Notch2. This inactivation of Notch2 signaling ultimately promotes osteoblast differentiation [50]. Additionally, Notch2 colocalizes with GSK3 in the nucleus via NICD, inhibiting the Wnt/β-catenin pathway to hinder osteoblast differentiation [51]. In inflammation-related contexts, tumor necrosis factor (TNF) transgenic mice (rheumatoid arthritis model) exhibit increased expression of non-canonical NF-κB proteins p52 and RELB. These proteins interact with NICD at the Hes1 promoter to upregulate Notch target gene expression. Sustained Notch activation further blocks BMSCs’ commitment to the osteoblast lineage, reducing osteoblast differentiation, whereas Notch inhibitors can prevent bone loss [52]. Furthermore, galectin-3 (Gal-3) binds to Notch1 in a sugar-dependent manner, accelerating Notch1 cleavage and activating Notch signaling. This, in turn, downregulates the expression of osteogenic differentiation markers (Runx2, SP7, ALP, COL1A1, IBSP, and BGLAP) and inhibits osteoblast differentiation [53].

However, some studies have yielded contradictory findings. For instance, osteoblast-specific deletion of RUBCN accelerates autophagic degradation of NICD and downregulates Notch signaling, thereby negatively regulating osteoblast differentiation [54]. During the early stage of osteoblast differentiation, activation of the NICD or endogenous Notch via DLL1 significantly promotes calcified nodule formation, stimulates BMP-2-induced osteogenic differentiation, and inhibits adipogenic differentiation [55]. Similarly, in primary human bone marrow mesenchymal stem cells (hBMSCs), those expressing Mastermind1 and NICD or harboring the Jag1 transgene exhibit enhanced mineralization, nodule formation, and ALP activity, with induction of both spontaneous and stimulated osteoblast differentiation [56]. Notch signaling also potentiates BMP9-induced BMP/Smad signaling activity. This not only upregulates the expression of key osteogenic differentiation factors in BMSCs, such as Runt-related transcription factor 2 (Runx2), COL1A1, and inhibitor of differentiation (Id) 1, but also activates the expression of activin receptor-like kinase 2 (ALK2) [57]. Beyond differentiation, Notch signaling influences osteoblast mineralization by enhancing anti-apoptotic effects. Specifically, Notch activation promotes osteoblast differentiation and maturation, enhances mineralization, and significantly increases cell proliferation. Concurrently, it upregulates the anti-apoptotic factor BCL-2 and downregulates the pro-apoptotic factor Caspase 3 [58].

Notch signaling and osteoclasts

The differentiation of osteoclasts is tightly regulated by multiple signaling pathways, among which the Notch signaling pathway is critical [59]. Similar to its role in osteoblasts, the function of Notch signaling in osteoclastogenesis remains controversial, with evidence supporting both inhibitory and stimulatory effects (Fig. 3B).

Notch signaling promotes osteoclast maturation. Tspan-5 and Tspan-10 enhance Notch activation by facilitating ADAM10 maturation [60]. Interestingly, the expression level of Notch2 in osteoclast lineage cells is significantly higher than that of Notch1, suggesting that Notch2 is the primary Notch receptor regulated by Tspan-5/10 [61]. Epigenetic modifications are also involved in this process. Specifically, NSUN2 promotes the degradation of KDM6B mRNA through m5C modification in a PPIA-dependent manner. This reduces KDM6B levels, leading to NUMB hypermethylation, disruption of its interaction with NICD, activation of the Notch signaling pathway, upregulation of receptor activator of NF-κB ligand (RANKL) expression, and ultimately promotion of osteolytic bone metastasis [62]. Inhibition of Notch signaling using Dibenzodiazepine (DBZ) reduces the expression of nuclear factor of activated T cells c1 (NFATc1), a key transcription factor for osteoclast differentiation. Conversely, overexpression of the Notch2 intracellular domain (NICD2) can rescue the inhibition of osteoclast differentiation by promoting proline-rich tyrosine kinase 2 (PYK2) autophosphorylation and microtubule acetylation [63]. The mechanism by which Notch regulates osteoclasts is reflected in the differential effects of distinct receptor-ligand axes. For example, the Notch2/DLL1 axis promotes osteoclastogenesis, whereas the Notch1/Jag1 axis inhibits it. The osteoclastogenesis-promoting effect mediated by the Notch2/DLL1 axis is dominant, and blocking DLL1 can alleviate bone erosion and loss in arthritis [64]. However, the roles of individual Notch ligands and receptor subtypes in osteoclastogenesis vary across different cell types. Another study reported that immobilized Notch ligand DLL1 can inhibit osteoclast differentiation in vitro, while Jag1 has been shown to directly induce osteoclast generation in bone marrow [65, 66]. Besides, the timing of exposure of osteoclast precursors to RANKL and Notch activation influences osteoclast generation. When RANKL is stimulated first and Notch is activated later, immobilizing Jag1 to stimulate Notch signaling can enhance osteoclastogenesis, manifested as an increase in cell size and mineral absorption, while inhibiting Notch reduces osteoclast fusion and absorption [67, 68]. In contrast, when Notch is activated before RANKL stimulation, Jag1 inhibits osteoclastogenesis, which is associated with Notch-mediated suppression of osteoclast precursor proliferation [65, 69]. Moreover, CaM-kinase IV (CaMKIV) interacts with the Notch1 intracellular domain (NICD1) and phosphorylates it, reducing Fbw7-mediated proteasomal degradation to enhance its stability and promote osteoclast differentiation. Notch1IC can also counteract the inhibitory effect of KN-93 on osteoclast differentiation, a function absent in phosphorylation-deficient Notch1IC [70]. Similarly, Notch2 interacts with the NF-κB subunit p65 and is recruited to the NFATc1 promoter, thereby supporting osteoclastogenesis [61]. The expression level of Hes1 increases concomitantly with Notch2 during osteoclast differentiation [71]. Deactivation or downregulation of Hes1 impairs osteoclast generation and reverses the osteoclastogenesis-promoting effect of Notch2, indicating that Hes1 mediates the regulatory role of Notch2 in osteoclast differentiation [72]. Similarly, TNF-α induces the transcription of Jag1, DLL1, DLL3, and Notch2, which in turn upregulate Hes1 and promote osteoclast differentiation under TNF-α stimulation [73, 74]. In osteoclasts with enhanced Notch2 activity, TNF-α increases interleukin-1beta (IL1B) levels in a Hes1-dependent manner, further inducing osteoclastogenesis [75]. Notch3 gain-of-function mutations can lead to osteopenia by increasing the expression of the RANKL-encoding gene Tnfsf11 in osteoblasts and bone cells, while decreasing the level of the osteoprotegerin (OPG)-encoding gene Tnfrsf11b [76]. This suggests that Notch3 dysfunction may indirectly promote osteoclastogenesis and differentiation by regulating RANKL and other molecules secreted by osteoblasts and bone cells. Further research is required to clarify the spatiotemporal and cellular environmental regulation of Notch signaling in osteoclasts under both physiological and pathological conditions.

The regulation of Notch signaling in osteoclast differentiation is “cell state-dependent”, meaning it varies according to the cell’s own developmental stage or functional status. For example, it can inhibit the differentiation of immature cells, enhance the differentiation of adherent bone marrow-derived macrophages, and precisely regulate the process of osteoclast formation [68]. Specifically, in osteoclast precursors, DLL1 can suppress osteoclastogenesis and downregulate the expression of c-Fms on their surface. Simultaneously, Notch1 increases OPG expression in stromal cells and inhibits macrophage-colony stimulating factor (M-CSF) expression, thereby impeding osteoclast development [65]. Jag1 also exerts a regulatory role by weakening the capacity of bone marrow macrophages to differentiate into osteoclasts. Compound knockout of Notch1, 2, and 3 promotes the proliferation of osteoclast precursors and enhances their responsiveness to RANKL, leading to increased osteoclastogenesis [69]. In tertiary myeloid progenitor cells, activation of Notch1 signaling can inhibit osteoclast formation, reduce the number of TRAP⁺ osteoclasts, and downregulate the expression of RANK protein and the Ctsk gene. Conversely, disruption of the Notch regulatory pathway (via deletion of RBP-J) enhances osteoclastogenesis, as evidenced by an increase in osteoclast number and size, along with elevated TRAP expression [77]. Activated Notch1 signaling can downregulate the expression of osteoclast differentiation-related genes such as Fos, Ctsk, and RANK, resulting in a bone sclerosis phenotype [78, 79]. Furthermore, although DLL1-induced Notch signaling does not alter the intrinsic differentiation potential of osteoclast lineage cells, it can arrest osteoblast maturation. This disruption of osteoblast-osteoclast coupling impairs the maturation and function of osteoclasts, significantly inhibiting bone metabolic turnover [80].

The dual regulatory effects and experimental discrepancies of Notch signaling on osteoblasts and osteoclasts arise from its context-dependent nature, which manifests in variations across cell types, differentiation stages, specificity of receptor–ligand interactions, as well as the timing of signal activation and microenvironmental influences. Specifically, Notch1 inhibits differentiation in osteoblast precursors but promotes mineralization in mature osteoblasts. Conversely, while Notch signaling inhibits osteoclastogenesis in osteoclast precursors, it promotes differentiation in adherent bone marrow-derived macrophages. Moreover, the Notch2/DLL1 axis enhances osteoclastogenesis, whereas the Notch1/Jag1 axis exerts an inhibitory effect. The differential predominance of these pathways at various pathological stages of OA leads to apparently contradictory experimental outcomes. Furthermore, RANKL stimulation followed by Notch activation enhances osteoclast function; in contrast, reversing this sequence produces inhibitory effects. Additionally, TNF-α within the inflammatory microenvironment can amplify the pro-osteoclastogenic influence of Notch signaling.

Overall, the relationship between Notch signaling and osteoblasts or osteoclasts is complex and paradoxical. Notch signaling can interfere with key processes such as osteoblast differentiation, glucose metabolism, and bone formation, while also promoting osteoblast differentiation, mineralization, and enhancing anti-apoptotic capacity. This bidirectional and context-dependent regulatory effect warrants in-depth exploration.

The regulatory effect of the Notch signaling pathway on cartilage homeostasis

Current studies have demonstrated that Notch signaling is a key regulator of cartilage growth, development, and joint homeostasis, influencing the entire life cycle of chondrocytes, though its precise role remains controversial [81, 82]. Researchers detected Notch signaling activity in fibrocartilage stem cells (FCSCs) from normal human temporomandibular joints and further confirmed that activated Notch signaling in FCSCs can mediate osteogenesis and chondrogenic differentiation [83]. Similarly, during in vitro induction of FCSCs differentiation into chondrocytes, members of the Notch/Delta/Serrate family and Dlk2 (a Notch ligand associated with adipogenesis) are expressed in normal ATDC5 cells [83, 84]. These findings indicate that Notch signaling is present in normal cartilage tissue and participates in cartilage growth and development. In addition, reduced cartilage-specific Notch signaling can temporarily downregulate matrix metalloproteinase 13 (MMP13) expression in mouse joints and delay cartilage degeneration [12]. But, in a mouse model with specific knockout of core Notch signaling genes, no signs of joint site impairment were observed at embryonic day 12.5 (E12.5) in the absence of Notch. Nevertheless, after 8 months of cartilage growth, degeneration was evident, including articular cartilage and meniscal degeneration, articular cartilage mineralization, and subchondral bone structural changes [85]. This suggests that Notch signaling plays a complex role in cartilage homeostasis; both its sustained activation and permanent downregulation can lead to cartilage degeneration and joint dysfunction. This phenomenon may arise from the stage-specific roles of Notch signaling in cartilage tissue. During embryonic development, when cartilage is in the early formation stage, other signaling pathways (e.g., Wnt, BMP) may compensate for Notch function to maintain the basic developmental program of cartilage anlagen, explaining why short-term Notch deficiency does not result in abnormal joint phenotypes. By 8 months, however, cartilage enters a homeostasis-maintenance phase, where Notch signaling becomes irreplaceable in regulating chondrocyte proliferation-differentiation balance, matrix metabolism, and subchondral bone remodeling. Long-term loss of Notch signaling disrupts this homeostasis, impairing chondrocyte function and triggering OA-like pathological changes.

Notch signaling influences chondrocyte dedifferentiation, endochondral ossification, differentiation, and cartilage repair. Reports indicate that inhibiting Notch signaling can delay chondrocyte dedifferentiation and maintain CoL2A1 expression [86]. During chondrocyte differentiation, Notch1, Notch2, Rbpj, and Hes1 show high expression, whereas Notch3, Notch4, and other members of the Hes/Hey family are rarely expressed [12]. Endogenous RBPjκ is critical for maintaining Notch function. RBPjκ-dependent Notch signaling enhances the final stage of endochondral ossification by inducing the target gene Hes1, which transactivates MMP13 and vascular endothelial growth factor a (Vegfa). Notch signaling regulates cartilage through two mechanisms: RBPjκ-dependent and RBPjκ-independent pathways [87]. The RBPjκ-dependent pathway governs the final stages of chondrocyte maturation, endochondral ossification, and OA progression, while the RBPjκ-independent pathway is crucial for chondrocyte proliferation and survival [12]. Among them, the RBPjκ-independent Notch pathway acts as an important non-cell-autonomous regulator of perichondral bone formation and is a key cartilage-derived signal required to coordinate chondrocyte and osteoblast differentiation during endochondral bone development [88]. Enhanced Notch signaling inhibits chondrogenic differentiation, truncates hypertrophic cartilage regions, and downregulates Sex-determining region Y-box 9 (SOX9) and Runx2 expression, disrupting chondrocyte differentiation and bone formation and leading to skeletal deformities [89, 90]. Conversely, loss of Notch signaling function results in expanded hypertrophic cartilage, reduced vertebral bone mineralization, and upregulated SOX9 expression, also causing abnormal chondrocyte differentiation [91]. Whether enhanced or diminished, Notch signaling can further perturb normal chondrocyte lineage development by altering the expression of downstream targets such as Hes7, leading to abnormal changes in cartilage structure and function [92]. Notch signaling also promotes the differentiation of MSCs into chondrocytes. For instance, thrombospondin 2 (TSP2) promotes chondrogenic differentiation of human adipose-derived mesenchymal stem cells (hADMSCs) via the Jag1/Notch3 signaling pathway. Additionally, BMP9 targets and activates the Notch1/Jag1 signaling pathway to promote chondrogenic differentiation of adipose-derived mesenchymal stem cells (ADMSCs), thereby facilitating cartilage repair [93–95]. Besides, Notch signaling accelerates cartilage ossification and promotes cartilage tissue healing [96]. Alternatively, it has been proposed that Notch signaling regulates osteochondral repair by selecting specific subpopulations of osteogenic progenitor cells.

The mechanism by which Notch influences cartilage degeneration is reflected in the correlation between its persistent activity state and its effects. Transient or physiological Notch signals promote cartilage anabolic metabolism by inducing SOX9 and maintaining a balance between anabolism and catabolism [97]. However, sustained or high levels of Notch activity can inhibit chondrogenic genes like SOX9, CoL2A1, and Acan, while inducing catabolic factors such as MMP13, thereby triggering pathological responses [98]. Additionally, sustained Notch activation can upregulate IL-6 and p-STAT3. As a potential direct target gene of Notch, IL-6 is involved in mediating the induction of MMP13 and CoL3A1 by Notch [99]. Notch and its effector Hes protein can also promote cartilage degeneration through various pathways by regulating the phosphorylation and activity of STAT3 [100].

In summary, Notch signaling is critically involved in multiple processes, including chondrogenesis, chondrocyte differentiation, dedifferentiation, cartilage healing, and degeneration (Fig. 3C). The regulation of cartilage homeostasis by Notch signaling is highly context-dependent, with experimental variations primarily arising from differences in signal intensity, duration, cellular localization, and pathway specificity. Transient physiological Notch activation promotes anabolic cartilage metabolism, whereas sustained high-level activation induces catabolic effects. Notch signaling is typically highly active during early chondrogenesis and gradually declines as chondrocytes differentiate. However, when chondrocytes approach hypertrophy, Notch again acts as an enhancer of hypertrophy by modulating SOX9 expression [101]. Beyond SOX9, other Notch targets, such as HES, HEY, NF-κB, Rbpjk, as well as Pax6 (identified as a Notch1 target in embryonic stem cells), may also be influenced by Notch signaling at specific stages or under certain conditions during cartilage development. These controversies highlight that therapeutic targeting of Notch signaling in chondrocytes requires precise control over intervention intensity and timing to avoid adverse outcomes resulting from non-selective pathway inhibition or activation.

The role of the Notch signaling pathway in the pathogenesis of OA

The pathological process of OA involves abnormalities in multiple tissues, including cartilage, subchondral bone, and synovium [102]. The Notch signaling pathway acts as a precise “regulatory switch”, capable of finely modulating pathological changes in cartilage, subchondral bone, and synovial tissue through a series of complex signal transduction mechanisms, thereby playing an indispensable role in OA progression [13, 101] (Fig. 4).

Fig. 4.

The Notch signaling pathway influences the pathological processes of OA-related cartilage, subchondral bone, and synovium by regulating relevant cytokines. In cartilage, it regulates chondrocyte apoptosis/hypertrophy, ECM degradation, inflammation, senescence, and microenvironment. In subchondral bone, Notch signaling affects subchondral bone by regulating bone cell dysfunction and subchondral bone remodeling imbalance. Additionally, it may also participate in this process by regulating the function of SF. (Created in https://BioRender.com)

The role of Notch signal in cartilage of OA

Numerous studies indicate that activating Notch signaling in articular chondrocytes speeds up OA progression. It’s activated in cartilage from total knee arthroplasty patients, with higher expression in severely damaged regions [103, 104]. Other research confirms that Notch1, Jag1, Hes5, and NICD are highly expressed in human OA cartilage, positively linked to cartilage degeneration [105]. Moreover, Notch markers like Notch1, Jag1, Hes1, and Hes5 are also present in the fibrocartilage of a temporomandibular arthritis mouse model [106]. This implies that Notch signals could be potential OA severity biomarkers. In brief, Notch impacts OA via multiple pathways, and the following sections delve into its mechanism of action.

Chondrocyte apoptosis and degradation of matrix

As a key Notch ligand, Jag1 is upregulated in OA chondrocytes, and its degradation delays cartilage degeneration [12, 107]. It is worth mentioning that numerous microRNAs(miRNAs) and circRNAs are involved in regulating chondrocyte apoptosis and ECM degradation. For instance, miR-140-5p inhibits Notch signaling in cartilage progenitor cells (CPCs), downregulates the expression of Jag1, cleaved Notch1, and Hey1, enhances cell viability, and inhibits apoptosis, exerting a protective effect against OA [108]. Certainly, DAPT (a specific Notch inhibitor) and miR-140-5p mimetics also protect against OA by inhibiting Jag1/Notch signaling, which suppresses IL-1β-induced OA-like changes in CPCs, reducing cell viability, migration, and chondrogenesis, and increasing cell apoptosis [109]. However, the transcription factor Yin Yang 1 (YY1) exacerbates OA by transcriptionally repressing miR-140-5p and enhancing the Jag1/Notch signaling pathway, thereby interfering with CPCs’ fate reprogramming [108]. Furthermore, circ-0104873 acts as a sponge for miR-875-5p, upregulating Notch3 to activate the Notch signaling pathway and promote OA progression [110]. IL-1β treatment upregulates circFOXK2 in chondrocytes, which further competes with miR-4640-5p for binding, enhances Notch2 expression, and forms the circFOXK2/miR-4640-5p/Notch axis [111]. This axis promotes the transcription of inflammatory factors such as IL-33, IL-17F, and IL-6, thereby inducing chondrocyte apoptosis, accelerating ECM degradation, inhibiting cell proliferation, and driving OA progression. Except for miRNA and circRNA, the E3 ubiquitin ligase ITCH binds to Jag1 via the WW-PPXY motif and degrades Jag1 through K48-linked ubiquitination, thereby inhibiting the Notch1 signaling pathway. Specifically, this process promotes chondrocyte proliferation, inhibits apoptosis, and reduces ECM degradation. Conversely, overexpression of Jag1 reverses this protective effect and exacerbates OA-induced articular cartilage damage [107]. Moreover, S-phase kinase-associated protein 2 (SKP2) triggers Krüppel-like factor 11 (KLF11) ubiquitin-mediated degradation and inhibits H3K27me3 to transcriptionally activate JMJD3, thereby activating the Notch1 pathway and promoting OA chondrocyte apoptosis and ECM degradation [112]. Similarly, Hes1 promotes ECM degradation and chondrocyte apoptosis by inducing cartilage catabolic metabolic factors such as ADAMTS5, IL-6, and ADAMTS8. ADAMTS8 and Notch form a positive feedback regulatory loop that inhibits the production of type I collagen and aggrecan, exacerbating the OA phenotype. Thus, deletion of Hes1 in articular cartilage after skeletal maturation inhibits OA development [113, 114].

Chondrocyte hypertrophy

In the advanced stages of OA progression, chondrocytes undergo dedifferentiation. At times, chondrocytes exhibit an increase in size and adopt a thickened morphology [115]. Sustained OA-related stimuli lead to the activation of Notch signaling, which further upregulates the expression of markers associated with chondrocyte hypertrophy and differentiation, including Col X, MMP-13, and Runx2. Conversely, it downregulates the expression of SOX9, aggrecan, and Col II. Subsequent cartilage hypertrophy triggers degradation of the cellular microenvironment, thereby exacerbating OA pathology. In contrast, si-Notch signaling reduces chondrocyte hypertrophy and differentiation, as well as the terminal differentiation driven by extracellular matrix remodeling [116, 117]. Other studies have further confirmed that activating Notch signaling not only induces the production of OA dedifferentiation markers such as Col I, MMP13, and eNOS but also decreases the synthesis of Col II and aggrecan. Notably, treatment with DAPT has been shown to effectively alleviate articular cartilage damage [118]. Interestingly, Notch-mediated regulation of MMP13 is dependent on Runx2; interference with Runx2 reduces the sensitivity of MMP13 expression to Notch signaling regulation [117]. However, when Runx2 is overexpressed, its effect may reach a saturation point or be constrained by other synergistic factors, thereby precluding further amplification of Notch's regulatory role.

Chondrocyte senescence

Research has shown that elevated expression of the Notch pathway marker NICD exacerbates chondrocyte senescence by upregulating MMP13 levels in chondrocytes [12], while myosin light chain 3 (MYL3) deficiency enhances CME by promoting the interaction between MYO6 and clathrin, thereby inducing Notch receptor internalization and NICD nuclear translocation, leading to activation of Notch signaling, ultimately upregulating the expression of age-related genes (SASP), triggering chondrocyte aging and exacerbating OA [119]. On the contrary, translation factor eukaryotic initiation factor 5 A (eIF5A) inhibits Notch signaling by suppressing the expression of histone acetyltransferase cyclic adenosine monophosphate response element binding protein (CREB)-binding protein (CREBBP), downregulating the expression of NICD, MMP13, and aging markers (P16 and P21), while increasing COL2 expression, thereby delaying OA [120].

Inflammatory response

Under the persistent stimulation of OA-associated inflammatory responses, chondrocytes sustain damage. Notch2 contributes to OA pathogenesis by enhancing the activity of inflammation-related pathways in epiphyseal chondrocytes and disrupting the transcriptomic profiles and cellular cluster interactions of joint or synovial fibroblasts [121]. Notch2 interacts significantly with TNF-α. Overexpression of Notch2 induces the expression of Notch target genes, including members of the Hes and Hey families. However, in the presence of TNF-α, TNF-α reduces the expression of the RBPJκ nuclear protein complex, causing RBPJκ to dissociate from DNA-binding sites. This leads to downregulated expression of target genes such as Hes1 and Hey, while inducing IL-6 expression. Besides, Notch2 can potentiate TNF-α-induced NF-κB activation, potentially by decreasing the level of unphosphorylated p65 [61, 122]. In short, Notch2 may influence the subpopulation of phagocytic chondrocytes by activating OA-related inflammatory pathways, upregulating phagocytic pathways, downregulating OA-associated genes (e.g., Gdf5 and Fgfr3), and amplifying Rac2-mediated responses to TNF-α. These processes ultimately enhance the chondrocyte inflammatory response to TNF-α [123–125].

Cartilage microenvironment

Notch signaling can aggravate OA phenotype to a certain extent by disrupting the chondrocyte microenvironment. During the progression of temporomandibular joint arthritis, excessive angiogenesis occurs in the articular cartilage, leading to the overproduction of hypoxia-inducible factor-1alpha (HIF-1α), vascular endothelial growth factor (VEGF), and activation of Notch signaling [126]. Abnormal levels of HIF-1α can induce the expression of Notch receptors through the HIF-1α-VEGF-Notch axis, thereby promoting VEGF-mediated angiogenesis. This process causes blood vessels to invade cartilage from the subchondral bone, disrupting the hypoxic microenvironment that is critical for maintaining cartilage homeostasis. Notably, inhibiting the HIF-1α-VEGF-Notch axis has been shown to reduce vascular invasion and exert a protective effect on cartilage [127].

Pain response

In OA mice, the Notch signaling pathway is activated in dorsal root ganglia (DRG) innervating the knee joint, where it synergizes with toll-like receptor 4 (TLR4) to promote the production of the chemokine C–C motif ligand 2 (CCL2). Inhibition of Notch signaling via the γ-secretase inhibitor DAPT or soluble Jag1 reduces CCL2 expression and alleviates knee joint hyperalgesia. Conversely, immobilization of Jag1 activates Notch signaling and potentiates these effects [128]. These findings suggest that the Notch signaling pathway could serve as a potential target for alleviating OA-associated pain [129].

To sum up, activation of the Notch signaling pathway and its family members accelerates OA progression through multiple mechanisms, regulating chondrocyte apoptosis and matrix degradation, exacerbating chondrocyte hypertrophy and senescence, participating in inflammatory responses, disrupting the cartilage microenvironment, and modulating pain signaling. These mechanisms play key roles in the initiation and progression of OA, offering multiple potential targets for OA therapeutic intervention.

The role of Notch signal in the subchondral bone of OA

Osteocyte dysfunction represents a key pathological feature in the progression of OA. Osteocytes within the subchondral bone of OA undergo phenotypic transformation, leading to altered bone mineralization [130, 131]. Studies have shown that Notch1 antisense transgenic (Notch1 AS) mice exhibit activation of the hedgehog signaling pathway via induction of Gli-1 and Gli-2, resulting in osteophyte formation and increased subchondral bone plate density [132]. In the inflammatory microenvironment of OA, M1 macrophages can inhibit Notch signaling in osteocytes by multiple pathways. On one hand, their secreted pro-inflammatory cytokines (e.g., IL-1, IL-6, TNF-α) or extracellular vesicles containing miR-34 and miR-146 directly suppress Notch signaling [133–136]. On the other hand, they indirectly activate pathways such as Wnt, MAPK, and mTOR, thereby inhibiting Notch signaling [137–140]. The expression level of Hes1 reflects the activity of the Notch signaling pathway [141, 142]. Inhibiting Notch signaling downregulates its target gene Hes1, hindering bone cell maturation, manifested as a decrease in dentin matrix protein 1 (DMP1) and an increase in E11, leading to abnormal bone mineralization. Immature bone cells can also promote osteoclast differentiation and exacerbate subchondral bone abnormalities by increasing RANKL and reducing OPG [143]. Activation of Notch signaling reverses these changes, restoring osteocyte maturation and normal mineralization, indicating that Notch signaling contributes to the development of subchondral bone abnormalities in OA by regulating osteocyte maturation and mineralization.

Imbalanced subchondral bone remodeling is another critical pathological feature of OA, and the Notch signaling pathway plays a key regulatory role in this process [144]. When occlusal load is reduced, the condylar subchondral bone exhibits OA-like characteristics, including decreased bone mass, narrowed trabeculae, and widened trabecular spacing, accompanied by reduced expression of Notch1, Jag1, DLL1, and other signaling molecules. Following partial restoration of occlusal load, subchondral bone trabeculae gradually widen, trabecular spacing narrows, and the expression of Notch-related signaling molecules increases [145]. These findings suggest that the Notch signaling pathway is involved in the response and remodeling of subchondral bone to mechanical stimulation.

In conclusion, the Notch signaling pathway contributes to OA pathogenesis by regulating osteocyte maturation and mineralization, as well as mediating responses to mechanical stimuli (Fig. 4).

The role of Notch signal in the synovium of OA

Synovitis critically drives OA progression, with Notch signaling potentially mediating this process via synovial fibroblasts (SF) regulation [146]. Prior studies revealed differential expression patterns of Notch homologs in OA and rheumatoid arthritis (RA) synovium compared to healthy tissue [147, 148]. TNF-α can induce the expression of IL-6, MMP11, Notch1, Notch4, and Jag2 in rheumatoid arthritis synovial fibroblasts (RASF) [149]. Notably, a Notch1 fragment is present in the nucleus of synovial cells, where it participates in TNF-α-induced proliferation of RA synovial cells [150]. Interestingly, expression of Notch1, Notch4, and Jag2 is also detected in the synovium during mouse embryonic development [150]. This suggests that synovial tissue formation during early embryonic development involves a series of precisely regulated molecular events, and the expression of Notch-related molecules may constitute part of these conserved developmental mechanisms. Notch3 is also significantly upregulated in SF. Its deficiency or blockade alleviates inflammation and prevents inflammatory joint damage. Notch3 has been identified as a key receptor for SF differentiation and pathological amplification [151]. Under hypoxic conditions, the NICD2, NICD3, and HIF-1α are highly expressed in the synovial tissue of patients. HIF-1α can directly regulate the expression of Notch1 and Notch3 genes, promoting RASF invasion and angiogenesis. Specifically, Notch1 regulates SF migration and epithelial-mesenchymal transition, while Notch3 modulates anti-apoptotic and autophagic processes [152, 153]. Besides, inhibition of the STAT3/NF-κB/Notch1 signaling axis impairs SF migration and reduces the progression of inflammatory arthritis [154].

In short, TNF-α activates Notch signaling in SF, which synergizes with the NF-κB pathway to mediate TNF-α-induced IL-6 production in SF [155]. Whether Notch directly binds the IL-6 promoter requires further validation. Notch signaling plays a key role in OA pathogenesis (Fig. 4). Future studies may clarify its activation mechanism in OA-SF and its effects on inflammatory cytokine secretion and ECM degradation.

Potential strategies for targeting the Notch signaling pathway in OA treatment

The following explores potential strategies targeting the Notch signaling pathway, focusing on key factors such as γ-secretase inhibitor (GSI) and specific neutralizing antibodies (Table 1).

Table 1.

Notch-targeted therapeutic strategies

Category	Interv	Mechanism of action/outcomes	Clinical status	Ref
γ-secretase inhibitor	DAPT	Reducing knee hyperalgesia	Preclinical research	[128]
		Suppressing cartilage degradation		[12, 156]
		Delaying chondrocyte dedifferentiation and re-differentiation		[86]
		Improving the OA-like degeneration of cartilage progenitor cells, promoting cell viability, and inhibiting cell apoptosis		[109]
		Inhibiting angiogenesis of articular cartilage		[126]
		Promoting G1 phase arrest		[157]
	MK-0752	Decreasing cellular invasion and increasing apoptosis	Preclinical research	[158, 159]
	LY411575	Suppressing osteoclast differentiation	Preclinical research	[160]
	PF-03084014	Antitumor activity, increased apoptosis, and reduced cancer stem-like cells	Phase I clinical trial	[161, 162]
	RO4929097	Antitumor activity	Phase II clinical trial	[163]
	DBZ	Reducing the formation of new blood vessels	Preclinical research	[164]
	Nirogacestat	Inhibiting tumor cell growth	Clinical application	[165]
	LY3039478	Antitumor activity, inhibiting cancer metastasis	Phase I clinical trial	[166]
Pan-Notch inhibitor	CB-103	Antitumor activity, anti-proliferation	Phase II clinical trial	[167]
Specific neutralizing antibody	Anti-Notch1 mAb	Inhibiting osteoclastogenesis	Preclinical research	[64]
	Anti-Notch2 mAb	Enhancing osteoclastogenesis	Preclinical research	[64]
	OMP-59R5	Inhibiting tumor growth	Preclinical research	[168]
	Anti-Jag1 mAb	Reducing the number of cancer stem cells	Preclinical research	[169, 170]
	Anti-DLL1 mAb	Improving inflammation, reducing the number of osteoclasts, and inhibiting trabecular bone loss	Preclinical research	[64]
	OMP-21M18	Inhibiting tumor angiogenesis	Phase I clinical trial	[171]
Others	miR-140-5p	Promoting cartilage matrix synthesis	Preclinical research	[109]
	miR-30a	Promoting chondrocyte differentiation	Preclinical research	[172]
	miR-145	Reducing chondrocyte apoptosis	Preclinical research	[173]
	miR-9	Promoting differentiation and regeneration of chondrocytes	Preclinical research	[174]
	miR-146a	Inhibiting cartilage degeneration, senescence, and inflammatory response	Preclinical research	[175]
	GRb1	Reducing chondrocyte apoptosis and promoting cartilage regeneration	Preclinical research	[176]
	Curcumin	Inhibiting chondrocyte hypertrophy	Preclinical research	[177]

γ-secretase inhibitor

Activation of the Notch signaling pathway relies on γ-secretase-mediated cleavage of the Notch receptor at the S3 site, which releases the active NICD. GSI represent a widely studied class of drugs targeting the Notch pathway [178]. Thus, GSI administration can block Notch signaling activation, thereby exerting therapeutic effects on bone diseases [66, 179, 180]. Currently, DAPT is the primary GSI used in OA research. By inhibiting cleavage and activation of Notch receptors, DAPT has been shown in OA animal models to reduce the secretion of chondrocyte hypertrophy markers and inflammatory factors, while significantly increasing the synthesis of ECM components such as type II collagen and aggrecan. These effects promote chondrocyte ECM synthesis, delay cartilage degeneration, and improve subchondral bone remodeling [12, 86]. Meanwhile, in animal experiments, DAPT treatment in OA model rats markedly alleviated articular cartilage degeneration, downregulated the expression of MMP-13, Notch1, Jag1, and Hes5, inhibited osteophyte formation, and delayed the progression of temporomandibular joint osteoarthritis [109, 156]. Additionally, DAPT intervention can suppress the expression of HIF-1, VEGF, and Notch, thereby inhibiting angiogenesis in articular cartilage and preventing osteosarcoma growth by promoting G1 phase arrest [126, 157].

These findings suggest that GSI holds promising application prospects for OA treatment. However, GSI are not exclusively involved in Notch signaling activation but also participate in other physiological processes, potentially leading to side effects. For instance, in rodents, GSI may induce high intestinal toxicity, disrupt bone development, impair normal nervous system function, and cause adverse reactions such as cognitive impairment [181–183]. Identifying negative regulatory regions of Notch or using antibodies that antagonize specific Notch receptors may help alleviate or limit such adverse effects. Ligand-specific antibodies targeting DLL1 and DLL4 have been shown to reduce toxicity in treating a mouse model of graft-versus-host disease [184]. Other GSIs, including LY411575, PF-03084014, RO4929097, DBZ, MK-0752, Nirogacestat, and LY3039478, have been used for targeted Notch inhibition in various diseases, but their inhibitory effects in OA remain unclear [158, 160, 161, 163–166]. Therefore, developing tissue-specific or receptor-targeted Notch inhibitors and exploring the mechanisms of action of different GSIs in OA will provide valuable directions for their clinical application [185].

Specific neutralizing antibody

Beyond GSI, specific antibodies targeting Notch receptors or ligands can block receptor-ligand interactions. These antibodies inhibit pathway activation by directly binding to Notch ligands or receptors, preventing ligand-receptor engagement. Compared to GSI, neutralizing antibodies offer higher target specificity and lower off-target toxicity.

Neutralizing antibodies against Notch receptors include anti-Notch1, anti-Notch2, and anti-Notch3 antibodies. Anti-Notch1 agonist mAb inhibits osteoclastogenesis and ameliorates abnormal subchondral bone remodeling, while anti-Notch2 agonist mAb enhances osteoclastogenesis [64]. Additionally, Tarexumab (OMP-59R5), a blocking antibody targeting both Notch2 and Notch3, inhibits tumor growth [168]. Though its role in OA remains unclear, Notch2/3 is confirmed to exacerbate OA progression. Thus, OMP-59R5 or similar antibodies may alleviate OA symptoms by suppressing Notch2/3 signaling, reducing bone remodeling, and inflammation.

Ligand-targeting neutralizing antibodies include anti-Jag1, anti-DLL1 antibodies, and DLL4 inhibitors. Anti-Jag1 mAb blocks Jag1-Notch binding, inhibiting pathway activation [169]. However, studies show anti-Jag1 antibodies exacerbate joint inflammation, an effect absent in CD8⁺ T cell-deficient models [170]. This suggests Jag1 not only regulates Notch signaling directly but also indirectly suppresses CD8⁺ T cell overactivation via antigen-presenting cells to maintain immune balance. Blocking Jag1 abolishes this negative regulation, leading to CD8⁺ T cell hyperactivation, pro-inflammatory factor release, and aggravated joint inflammation. Anti-DLL1 mAb reduces K/BxN-induced joint inflammation, decreases osteoclast numbers in affected joints, and inhibits ovariectomy-induced trabecular loss [64]. Demcizumab (OMP-21M18), a DLL4-blocking antibody, inhibits tumor angiogenesis and stem cell activity in cancer trials [171]. While untested in OA, it may theoretically alleviate OA symptoms. Moreover, novel small-molecule Notch inhibitors are under development, potentially expanding OA treatment options.

Others

SiRNA, miRNA, and components of traditional Chinese medicine can exert protective effects on OA by regulating the expression of key molecules in the Notch pathway. For example, siRNA targeting Notch1 and Hes1 can specifically silence these target genes, thereby inhibiting chondrocyte dedifferentiation and senescence [116]. MiR-140-5p mimetics can enhance the inhibition of Jag1/Notch signaling and promote cartilage matrix synthesis [109]. Beyond miR-140-5p, miR-30a, miR-145, miR-9, and miR-146a also improve the OA cartilage phenotype through distinct mechanisms [172–175]. In addition, isolated active compounds such as ginsenoside Rb1 (GRb1) and curcumin alleviate chondrocyte apoptosis to a certain extent, promote cartilage regeneration, inhibit cartilage hypertrophy and inflammation, and delay OA progression [176, 177]. Unexpectedly, GRb1 exhibits a more potent cartilage-protective effect than DAPT, which may be associated with its multiple properties and multi-target mechanisms in OA, warranting further investigation.

Current research on targeted Notch therapies exhibits considerable publication bias, with most studies reporting positive outcomes while often overlooking negative results. As previously noted, certain γ-secretase inhibitors exacerbate joint damage in aged OA models by impairing subchondral bone repair. Although anti-Jag1 antibodies effectively block Notch signaling, they can also relieve CD8⁺ T cell suppression, thereby aggravating joint inflammation. Moreover, the intestinal toxicity and neurological side effects associated with non-selective γ-secretase inhibitors have not been adequately evaluated in OA models. In addition, γ-secretase inhibition may disrupt normal immune, intestinal, and vascular functions—side effects that are clinically unacceptable in OA management [186]. These limitations underscore the inherent constraints of current OA animal models, including the inability of single-injury models to replicate multifactorial clinical pathology, and the scarcity of models incorporating high-risk factors such as aging or obesity. Moving forward, it will be essential to develop composite disease models and emphasize head-to-head comparative studies alongside long-term safety assessments to prevent clinical translation from being misled by publication bias.

Targeting the Notch signaling pathway has provided new hope and directions for OA treatment (Fig. 5). However, most of these strategies remain in the research stage and face numerous challenges. One of the central challenges lies in the critical role of the Notch pathway in maintaining cellular homeostasis in various tissues (e.g., skin and blood vessels). This fundamental physiological function means that targeting it can lead to severe off-target side effects and consequently creates a very high bar for drug specificity. Furthermore, the in vivo efficacy and safety of these therapeutic agents remain pressing issues that need to be thoroughly validated. Future studies should further explore the precise mechanisms of the Notch signaling pathway in OA and develop safer, more effective, and specific targeted therapeutic drugs and approaches.

Fig. 5.

Therapeutic approaches to ameliorate OA by targeting and inhibiting the Notch signaling pathway. GSI, specific neutralizing antibodies, small interfering RNA (siRNA), microRNA, and natural plant active ingredients regulate OA pathogenesis via the Notch signaling pathway. (Created in https://BioRender.com)

Controversy and transformation challenges of the Notch signaling pathway in OA

The Notch signaling pathway does not function in isolation within a single joint tissue; instead, it forms a complex regulatory network through intercellular crosstalk among chondrocytes, synovial fibroblasts, and osteoblasts. For instance, IL-6 secreted within OA cartilage can stimulate the expression and activation of Notch3 in synovial fibroblasts via paracrine signaling, thereby establishing a positive feedback loop across joint tissues [146, 187].

The controversy primarily concerns the context-dependent role of Notch signaling in determining cell fate. Firstly, in immature chondrocytes, Notch1 activation maintains stemness and promotes tissue repair [188]. Conversely, in mature chondrocytes, the Notch1-Hes1 axis upregulates ADAM8 expression, which exacerbates the osteoarthritic phenotype and promotes chondrocyte apoptosis [114]. Direct evidence is still lacking to elucidate whether this divergence is driven by epigenetic modifications, non-canonical pathway activation, shifts in intrinsic cellular states, or alterations in extracellular matrix stiffness. Indeed, the Notch pathway is integral to chondrocyte terminal differentiation, maturation, and endochondral ossification [189]. Secondly, studies using Notch gain-of-function mouse models have demonstrated that sustained Notch activation promotes an OA-like phenotype, inducing chondrocyte hypertrophy and cell cycle arrest via BMP/SMAD-mediated upregulation of p57 [190]. This suggests a pathogenic role for persistent Notch signaling. However, subsequent evidence indicates that the functional outcome critically depends on the degree, duration, and timing of activation. Continuous Notch activation is generally pathological, whereas transient activation more closely resembles physiological signaling and can be beneficial. For example, excessive Notch1 activation causes severe early cartilage loss and apoptosis, while its transient activation supports ECM synthesis [98]. Furthermore, and more surprisingly, inhibition of Notch signaling in a Notch1 antisense transgenic model also exacerbates OA, aggravating osteophyte formation [132]. Mechanistically, Notch inhibition activates the Hedgehog signaling pathway, which subsequently drives the key pathogenic processes of osteophyte formation and chondrocyte hypertrophy [132]. This reveals that Notch may exert a protective mechanism by inhibiting the Hedgehog pathway. Meanwhile, discrepant findings have been reported regarding the correlation between Notch signaling marker expression and osteoarthritis severity. Several studies indicate that Notch activity is markedly elevated in late-stage OA cartilage [113, 191]. In contrast, other work, such as an early-stage temporomandibular joint OA model induced by unbalanced occlusal load, demonstrates that the expression of Notch1, Jagged1, and Hes1 is significantly suppressed in both subchondral bone and cartilage, suggesting an inhibition of Notch1 during initial disease phases [192]. These conflicting results are likely attributable to variations in pathological stages, sampling locations (e.g., weight-bearing versus non-weight-bearing areas), or technical detection methods. Furthermore, substantial differences in inflammatory status, mechanical loading, and cellular composition exist between conventional surgical and chemical OA induction models. These methodological disparities may underlie the inconsistent Notch signaling activation patterns observed across studies. Compounding this issue, human cartilage samples are typically obtained from terminal-stage surgical resections, which inherently limit the ability to capture dynamic Notch signaling changes during early OA and introduce a significant sampling bias.

A major challenge in clinical translation involves improving the precision of Notch-targeting therapies. Most currently available Notch inhibitors, including GSI, cannot differentiate between pathological and physiological Notch signaling. Long-term use of these agents may disrupt normal bone homeostasis, for example, by suppressing bone formation and potentially inducing osteoporosis [90, 193]. Although specific neutralizing antibodies theoretically offer greater targeting precision, their suitable patient populations remain unclear due to the absence of comprehensive Notch receptor subtype expression data in OA patients. Furthermore, the majority of Notch inhibitors in preclinical development still lack sufficient specificity, which frequently results in off-target adverse effects in the cardiovascular, immune, and digestive systems, thereby limiting their potential for long-term OA treatment [194, 195]. While the development of Notch signaling pathway inhibitors is steadily advancing, their confirmed applications in OA remain relatively limited. Additionally, osteoarthritis is a multifactorial disease characterized by complex crosstalk between Notch signaling, obesity-associated metabolic disturbances, and mechanical stress-induced damage. Targeting Notch alone shows limited efficacy because it fails to simultaneously block other contributing pathological pathways. Therefore, a critical unresolved issue is how to effectively integrate Notch-targeted therapy with metabolic interventions and mechanical load management to develop synergistic multi-target treatment strategies.

In summary, the Notch signaling pathway exhibits a dual role in osteoarthritis, serving as both a pathogenic driver and a protective mechanism for tissue repair [196]. Future studies should prioritize the development of chondrocyte-specific temporal knockout mouse models to clarify the dynamic regulatory thresholds of Notch signaling across OA pathological stages. It is equally important to construct a cell-specific expression atlas of Notch receptors and ligands within the osteoarthritic joint microenvironment, which includes cartilage, synovium, and subchondral bone, to identify subtype-specific therapeutic targets. Furthermore, advancing joint-localized drug delivery systems would allow precise spatiotemporal modulation of Notch activity, thereby minimizing systemic toxicity while preserving therapeutic efficacy and accelerating the translation of Notch-targeting strategies from bench to bedside. Ultimately, progress in these key areas will be essential to transform Notch signaling from a controversial pathway into a clinically actionable target for OA therapy.

Conclusions and prospects

This paper summarizes the Notch signaling pathway’s role in bone and cartilage homeostasis, particularly its pathological relevance to OA progression. Studies show Notch1/2/3 regulate osteoblast, osteoclast, and chondrocyte fate via intricate mechanisms, with receptor-ligand combinations and activation timing influencing cell differentiation and cartilage degeneration. Aberrant Notch signaling in OA drives cartilage loss, subchondral bone abnormalities, and synovitis. Importantly, the timing of Notch signaling activation is different for joint maintenance and cartilage growth, which may be a significant breakthrough in understanding the mechanism of Notch signaling in OA. Notch-targeted therapies (e.g., γ-secretase inhibitors and neutralizing antibodies) hold promise for OA treatment but face clinical challenges like side effects and dose optimization.

Next research may focus on whether early transient Notch activation (before injury) exerts a preventive effect, and whether the benefits of transient Notch signaling can protect mice from joint damage during pathological processes [197]. A central paradox in the Notch field arises from its context-dependent dual effects, which are driven by variables such as cell type, disease stage, and specific receptor-ligand interactions. To mitigate the off-target risks and variable efficacy of current strategies, it is imperative to prioritize the development of cell-type-specific targeting approaches. In addition, it is necessary to elucidate the cell-type-specific and microenvironment-driven dynamics of Notch signaling in OA, develop safer, more efficacious targeted therapies, and integrate gene editing with single-cell sequencing to decode its regulatory network, thereby advancing precision OA treatment.

Abbreviations

OA

Osteoarthritis

ECM

Extracellular matrix

NECD

Notch extracellular domain

TMD

Transmembrane domain

NICD

Notch intracellular domain

EGF

Epidermal growth factor

NRR

Negative regulatory region

LNRs

Lin12-Notch repeats

HD

Heterodimerization domain

RBPJ

Recombinant signal binding protein-J

RAM

RBPJ associated module

NLS

Nuclear localization sequences

ANK

Ankyrin repeats

CSL

CBF1/suppressor of hairless/Lag1

TAD

Transactivation domain

DLL

Delta-like

Jag

Jagged

DSL

Delta/Serrate/LAG-2

CRD

Cysteine-rich domain

ER

Endoplasmic reticulum

ADAM

A disintegrin and metalloproteinase

NeXT

Notch extracellular truncation

BMPs

Bone morphogenetic proteins

BMSCs

Bone marrow mesenchymal stem cells

AMPK

AMP—activated protein kinase

si-Hes

Silencing of Hes-1

COLlA1

Type I collagen

ALP

Alkaline phosphatase

SALL4

Spalt-Like Transcription Factor 4

TNF

Tumor necrosis factor

Gal-3

Galectin-3

Runx2

Runt-related transcription factor 2

SOX9

Sex-determining region Y-box 9

hBMSCs

Human bone marrow mesenchymal stem cells

Id

Inhibitor of differentiation

ALK2

Activin receptor-like kinase 2

RANKL

Receptor activator of Nfkb ligand

DBZ

Dibenzodiazepine

NFATc1

Nuclear factor of activated T cells c1

NICD2

Notch2 intracellular domain

PYK2

Proline-rich tyrosine kinase 2

NICD1

Notch1 intracellular domain

CaMKIV

CaM-kinase IV

IL1B

Interleukin-1beta

OPG

Osteoprotegerin

M-CSF

Macrophage-colony stimulating factor

FCSCs

Fibrocartilage stem cells

MMP13

Matrix metalloproteinase 13

E12.5

Embryonic day 12.5

Vegfa

Vascular endothelial growth factor a

TSP2

Thrombospondin 2

hADMSCs

Human adipose-derived mesenchymal stem cells

ADMSCs

Adipose-derived mesenchymal stem cells

miRNA

MicroRNAs

CPCs

Cartilage progenitor cells

YY1

Transcription factor Yin Yang 1

SKP2

S-phase kinase-associated protein 2

KLF11

Krüppel-like factor 11

MYL3

Myosin light chain 3

eIF5A

Translation factor eukaryotic initiation factor 5A

CREBBP

Cyclic adenosine monophosphate response element binding protein-binding protein

HIF-1α

Hypoxia-inducible factor-1alpha

VEGF

Vascular endothelial growth factor

DRG

Dorsal root ganglia

TLR4

Toll-like receptor 4

CCL2

C-C motif ligand 2

Notch1 AS

Notch1 antisense transgenic

DMP1

Dentin matrix protein 1

SF

Synovial fibroblasts

RA

Rheumatoid arthritis

RASF

Rheumatoid arthritis synovial fibroblasts

GSI

γ-Secretase inhibitor

siRNA

Small interfering RNA

GRb1

Ginsenoside Rb1

Authors’ contributions

HW and ML designed the article, wrote the original manuscript, and created the figures. XW helped revise the manuscript and provide interpretations of the relevant articles. JH and XZ contributed to the manuscript review, supervised the project, and funded the acquisition. All authors read and approved the final manuscript.

Funding

This work was supported by the National Natural Science Foundation of China (Grant No. 32371184), National Key Research and Development Program of China (2024YFC3607304), and the Basic scientific research project of higher education institutions of Liaoning Province (JYTZD2023131).

Data availability

All data generated or analysed during this study are included in this published article.

Declarations

Consent for publication

All authors have approved the manuscript and agree with its submission to “Cell Communication and Signaling”.

Competing interests

The authors declare no competing interests.