The SPP1 gene encoding osteopontin (OPN) is located on the long arm of human chromosome 4 (4q22.1) and spans ~5.4 kilobases (kb), comprising 7 exons and 6 introns (11). The coding sequence is mainly distributed across exons 2 to 7, producing 5 major alternative splicing transcripts (12). The gene is evolutionarily conserved across vertebrates, with mammalian SPP1 exhibiting >60% nucleotide sequence identity in the coding region and >70% amino acid identity in cross-species comparisons (11). The 5'-end promoter region contains abundant cis-acting elements, allowing dynamic expression of SPP1 under stimuli such as inflammatory signals, hypoxia and carcinogenesis.

OPN is a glycosylated protein encoded by the SPP1 gene, which is widely present in the ECM and is a key component of bone matrix. The molecular weight of OPN protein ranges from 44 to 75 kDa. Alternative splicing of the SPP1 gene gives rise to three primary isoforms of OPN, each with distinct structural and functional characteristics (13). OPN-a, the full-length isoform, is the canonical secreted form (sSPP1) and the predominant isoform within the TME (14). OPN-b lacks exon 5 (OPN-b Δex5); although it is also secreted, it exhibits enhanced integrin binding and is associated with aggressive tumor phenotypes (15). OPN-c lacks exon 4 (OPN-c Δex4) and is often retained intracellularly (iSPP1) or localizes to the nucleus (nSPP1) (16). Functionally, secreted OPN-a and OPN-b mediate paracrine signaling in the TME, while intracellular/nuclear OPN-c modulates cell cycle progression and transcriptional regulation (14,17).

Structurally, SPP1 contains three functionally defined domains that support its diverse biological activities. The integrin-binding domain is acidic and rich in aspartic acid residues. It serves as the main site of phosphorylation, enhancing integrin binding affinity and promoting cell adhesion and migration (18,19). The central domain contains a conserved Arg-Gly-Asp (RGD) motif, which is responsible for binding to various integrins (for example, αvβ3, αvβ5, α5β1) (20). This interaction triggers downstream signaling cascades such as FAK/PI3K-AKT and MAPK/ERK pathways to regulate tumor cell proliferation, survival and epithelial-mesenchymal transition (EMT). The C-terminal domain contains a non-RGD integrin-binding site (SVVYGLR) and a thrombin cleavage site (21). Under the action of thrombin, proteolysis produces N-terminal and C-terminal fragments with distinct functions: The N-terminal fragment retains RGD-mediated integrin binding, while the C-terminal fragment preferentially interacts with the CD44 receptor, promoting stemness, chemoresistance and immune evasion in the TME (22). The functional diversity of OPN is largely attributed to its extensive and variable post-translational modifications (PTMs), which finely regulate its protein-protein interactions, stability and receptor binding specificity (Fig. 1).

SPP1 undergoes extensive PTMs that critically regulate its function:

Phosphorylation: OPN is one of the most highly phosphorylated proteins in vivo, with numerous serine and threonine phosphorylation sites, often occurring in acidic clusters. This modification is essential for high-affinity integrin binding and downstream signaling activation. Casein kinase II (CK2) and other kinases phosphorylate specific residues, creating acidic patches that coordinate calcium ions for integrin recognition (23,24). This extensive negative charge is crucial for mediating high-affinity binding to various integrins, especially αvβ3, through coordination with calcium ions and basic residues on the integrin surface. Phosphorylation is a prerequisite for integrin participation in activating downstream pro-survival and migration signaling pathways such as PI3K/AKT and FAK (25).

Glycosylation: OPN undergoes both N-glycosylation and O-glycosylation. These carbohydrate modifications significantly enhance protein stability, increase its resistance to proteolytic degradation, and affect its secretion (26). Most importantly, glycosylation directly regulates the interaction between OPN and its other major receptor, CD44. Specific glycan structures can either promote or sterically hinder binding to CD44 variants, adding an additional layer of regulation that determines whether OPN signals primarily through integrin-mediated or CD44-mediated pathways in a context-dependent manner (27,28).

Proteolytic cleavage: Thrombin cleavage at Arg156-Ser157 generates N-terminal (containing RGD) and C-terminal fragments. The N-terminal fragment retains integrin-binding activity, while the C-terminal fragment preferentially binds CD44 to promote stemness, chemoresistance and immune evasion (29).

In summary, the multi-domain structure of OPN combined with its complex PTMs enables interactions with multiple cell surface receptors, forming the structural basis for its functional pleiotropy.

Integrins are a class of heterodimeric transmembrane proteins composed of α and β subunits, which are recognized as the main cell adhesion transmembrane receptors. They play multiple roles in transducing biochemical and mechanical signals between cells and the environment as ECM-cytoskeleton linkers in both health and disease (30). SPP1 binds to several integrins via its RGD motif, including: αvβ3, αvβ5, αvβ1, αvβ6 and αIIbβ3 (CD61) (31-33). In cancer, αvβ3 and αvβ5 are the most extensively studied integrin receptors. They are primarily responsible for regulating the adhesion, migration and invasion abilities of cancer cells (34). After binding to αvβ3, SPP1 can activate the downstream FAK and PI3K/AKT signaling pathways, promote cytoskeletal rearrangement, and enhance the motility of cancer cells. In addition, αvβ3 is also involved in SPP1-induced angiogenesis by upregulating the expression of pro-angiogenic factors such as VEGF to promote the formation of tumor neovascularization (35) (Fig. 1).

A landmark study first confirmed that αIIbβ3 (CD61) is the key receptor through which SPP1 regulates stromal cell fate in pancreatic cancer (PC) (36). The study found that SPP1 secreted by epithelial PC cells specifically binds to the CD61 receptor on the surface of mesenchymal PC cells (MPCs), activates the downstream BMP2-GREM1 signaling axis, forms an intercellular positive feedback loop, and maintains the mesenchymal phenotype of MPCs. Knockout of CD61 completely mimics the phenotype of SPP1 knockout, leading to the transformation of MPCs into epithelial-like cells and significantly inhibiting the metastatic ability of PC. This discovery reveals a novel mechanism by which SPP1 regulates PC cell plasticity.

CD44 is a transmembrane glycoprotein with multiple isoforms generated by alternative splicing. CD44v6 and CD44v9 are particularly associated with cancer stem cells (CSCs). SPP1 binds CD44 via a region C-terminal to the RGD motif, overlapping with the SVVYGLR domain (37,38). This interaction activates JAK2/STAT3, PI3K/AKT and MAPK pathways, promoting stemness, chemoresistance and immune evasion (39). Unlike integrins, the binding of SPP1 to CD44 does not depend on the RGD sequence but is achieved through specific regions at its N-terminal and C-terminal ends. CD44 exists in multiple alternative splicing variants, among which CD44v6 plays a specific role in gastrointestinal cancers (40,41), whereas CD44v9 marks CSCs in cholangiocarcinoma and drives stemness, tumor progression, metastasis and chemoresistance by regulating Wnt/β-catenin, EMT and cellular redox status (42) (Fig. 1).

Existing studies have confirmed that SPP1 secreted by CAFs binds to the CD44 receptor on the surface of PC cells, activates the downstream FAK/PI3K/AKT/ERK signaling pathway, upregulates the expression of stemness transcription factors such as NANOG, KLF4, BMI-1 and ABCG2, and significantly enhances the self-renewal ability and tumor sphere formation ability of pancreatic CSCs (5). CD44v6 is not only involved in SPP1-induced maintenance of tumor stemness but also closely related to invasion, metastasis and chemoresistance of PC (43). In addition, SPP1 binding to CD44 on CAFs activates β-catenin/HIF-1α-related transcription, leading to increased production of CXCL12 by CAFs (44). This SPP1-CD44-β-catenin/HIF-1α-CXCL12 axis establishes a feedback loop that promotes cancer progression, creates an immunosuppressive microenvironment, and contributes to both primary and acquired resistance to immunotherapy (44).

Beyond integrins and CD44, SPP1 interacts with several additional receptors and co-receptors that modulate its signaling and functional outcomes. Heparan sulfate proteoglycans serve as co-receptors that modulate SPP1-mediated signaling, while calcium-binding proteins interact with SPP1 through its calcium-binding properties to influence ECM interactions (45).

Notably, CD61 (integrinβ3) has recently been identified as a critical receptor for SPP1 in PC, where it maintains mesenchymal cell fate through the SPP1-CD61-BMP2-GREM1 signaling axis (36). These alternative interactions expand the functional repertoire of SPP1 beyond its canonical integrin and CD44 pathways, highlighting its context-dependent roles in the TME.

Recent studies have uncovered context-specific roles of these pathways. In the TME, SPP1⁺ macrophages secrete SPP1, which binds to CD44 on CD8⁺ T cells and activates PI3K/AKT signaling. This aberrant activation induces a unique effector-exhaustion intermediate state (T cell stress response), where T cells retain some effector molecules (for example, IFN-γ) but upregulate exhaustion markers (PD-1 and TIM-3), leading to impaired proliferation and cytotoxicity, thereby driving colon cancer liver metastasis (46). In prostate cancer, PTEN deletion constitutively activates PI3K/AKT/mTOR signaling. Activated AKT phosphorylates and degrades the translation inhibitor PDCD4 (relieving eIF4A inhibition) and phosphorylates 4E-BP1 (releasing eIF4E). These parallel pathways increase eIF4F complex availability and cooperate with MNK/eIF4E to selectively enhance translation of mRNAs with complex 5'-untranslated region structures, promoting SPP1 secretion (47). Collectively, these studies reveal a bidirectional immunosuppressive feedback loop: Tumor cell-intrinsic PI3K/AKT activation drives SPP1 secretion via translational regulation; secreted SPP1 activates PI3K/AKT in immune cells (T cells, MDSCs) to induce immunosuppression; and the resulting immunosuppressive microenvironment sustains tumor cell survival and pathway activation. This loop, validated in colon and prostate cancers, may represent a universal immune evasion mechanism in solid tumors. Targeting this axis, for example, combining AKT inhibitors with SPP1-neutralizing antibodies, could overcome immunotherapy resistance. As depicted in Fig. 1, the PI3K/AKT/mTOR pathway is activated downstream of integrin and CD44 engagement and represents a central node in SPP1-mediated survival signaling.

Beyond PI3K/AKT, SPP1 activates MAPK/ERK signaling to promote c-Myc and Cyclin D1 expression, accelerating cell cycle progression, and upregulates MMP-2/MMP-9 to degrade ECM and enhance invasion (48). SPP1-integrin binding also induces FAK phosphorylation, which not only activates downstream PI3K/AKT and MAPK/ERK but also directly regulates cytoskeletal rearrangement and cell motility (49). Furthermore, the JAK1/JAK3-STAT1 pathway is an upstream regulator of SPP1 expression in macrophages: STAT1 activation drives SPP1 transcription, forming pathogenic SPP1⁺ macrophage subsets. JAK inhibitors (for example, tofacitinib) significantly suppress this pathway, downregulate SPP1, and reduce pathogenic macrophages, showing superior efficacy to TNF inhibitors, while SPP1 paracrine signaling amplifies JAK inflammatory signals in a positive feedback loop (50). Additionally, in tumor-associated macrophages (TAMs), SPP1 promotes angiogenesis through activation of the ERK/p38 signaling pathway, further highlighting the context-dependent utilization of p38 MAPK downstream of SPP1 (5). Moreover, In the TME, SPP1-stimulated macrophages (SPP1⁺ TAMs) promote angiogenesis via NF-κB activation (5,51). This NF-κB activation downstream of SPP1 is a critical node linking inflammation to tumor progression (Table I).

In short, SPP1 engagement with integrin or CD44 receptors triggers multiple intracellular signaling cascades critical for tumor progression. The PI3K/AKT/mTOR pathway, activated via FAK, supports cell survival, proliferation, metabolic reprogramming and stemness maintenance (36,52); MAPK/ERK signaling drives proliferation, differentiation and migration; NF-κB activation promotes inflammation, survival and invasion through upregulation of BCL2, MMP9 and CXCL8 (35,53); and CD44-dependent JAK/STAT3 signaling enhances stemness, immune evasion and angiogenesis, while integrin clustering-induced FAK/Src activation facilitates adhesion and invasion (54-56) (Table I). SPP1 splice variants expressed by breast tumors regulate monocyte activation via MCP-1 and TGF-β1, promoting an immunosuppressive myeloid phenotype (57,58). This SPP1-MCP-1/TGF-β1 axis contributes to the establishment of an immunosuppressive TME by skewing monocyte differentiation toward a pro-tumorigenic state. As schematically illustrated in Fig. 1, SPP1 functions as a signaling hub that integrates receptor-specific inputs into coordinated oncogenic outputs.

Collectively, these findings position SPP1 as a central signaling coordinator rather than a mere ligand: By engaging distinct receptor complexes, integrins (αvβ3, αvβ5 and αIIbβ3) and CD44 variants (CD44v6 and CD44v9), SPP1 converges on and amplifies core oncogenic cascades (PI3K/AKT, MAPK/ERK, JAK/STAT3 and NF-κB) in a context-dependent manner, thereby coordinating proliferation, survival, EMT, stemness and immune evasion as an integrated program (Fig. 1, Table I).

SPP1 promotes cancer cell proliferation through multiple mechanisms. Activation of PI3K/AKT signaling enhances cell cycle progression by upregulating Cyclin D1 and downregulating p27 (36). SPP1-integrin engagement activates the Ras-Raf-MEK-ERK cascade, which translocates to the nucleus and phosphorylates transcription factors including ELK1, c-FOS and c-JUN. The MAPK/ERK pathway increases MYC expression, driving metabolic reprogramming and ribosome biogenesis (75). SPP1 also inhibits apoptosis by upregulating BCL2 and BCL-XL via NF-κB and AKT signaling (35). SPP1 also upregulates IAP family members that directly bind and inhibit caspases-3, caspases-7 and caspases-9, further suppressing both intrinsic and extrinsic apoptotic pathways (8). SPP1-induced NF-κB increases c-FLIP expression, which inhibits death receptor-mediated apoptosis by competing with caspase-8 assembly (6) (Table I).

In OC, SPP1-induced STAT3 activation is required for survival under hypoxic and nutrient-deprived conditions (54). Furthermore, SPP1 promotes autophagy under stress conditions, providing metabolic intermediates that support survival. Conversely, excessive SPP1 signaling may suppress autophagic cell death by activating AKT (76).

Together, these mechanisms establish SPP1 as a central coordinator of tumor cell proliferation and survival: It integrates growth factor signaling, stress adaptation (for example, hypoxia and nutrient deprivation), and autocrine/paracrine cues to sustain cell cycle progression, apoptotic resistance and metabolic fitness across diverse cancer types.

SPP1 serves as a potent, multifaceted inducer of EMT, a cellular reprogramming process essential for metastatic dissemination, therapy resistance and stemness acquisition. SPP1 regulates EMT at transcriptional, post-transcriptional and post-translational levels via autocrine and paracrine mechanisms.

SPP1-induced AKT activation stabilizes Snail by phosphorylating GSK-3β, preventing Snail degradation. Snail directly represses CDH1 (E-cadherin) transcription by binding to E-box elements in its promoter. In CRC, Slug directly binds to the SPP1 promoter, creating a positive feedback loop that amplifies EMT (77). SPP1 signaling upregulates EMT transcription factors including Snail, Slug, Twist and ZEB1/2, leading to downregulation of E-cadherin and upregulation of N-cadherin, vimentin and fibronectin (64,78). Additionally, SPP1 activates matrix metalloproteinases (MMP2 and MMP9) that degrade ECM components, facilitating invasion (79). SPP1 directly binds to and stabilizes vimentin, preventing its ubiquitination and degradation, thereby promoting EMT independent of transcriptional changes (80). SPP1-induced COL11A1 expression drives EMT and metastasis via integrin-mediated FAK signaling. SPP1⁺ TAMs and FAP⁺ CAFs synergistically promote EMT and liver metastasis via β-catenin/HIF-1α activation (81). Several novel molecular mechanisms underlying SPP1 function in PC have been recently elucidated. Epithelial PDAC cells secrete SPP1, which activates CD61 on mesenchymal cells, thereby inducing BMP2 and GREM1 expression and establishing a bidirectional regulatory loop that maintains cellular heterogeneity (36). Additionally, LINC01133 has been identified as a long non-coding RNA that binds to Arp3 and reduces SPP1 mRNA degradation, consequently increasing SPP1 expression and promoting EMT in PDAC (82). Furthermore, VSIG4⁺ TAMs epigenetically regulate SPP1 transcription; VSIG4 deficiency reduces lactate production and histone H3K18 lactylation, leading to decreased STAT3-mediated SPP1 transcription and disruption of TAM-neutrophil immunosuppressive interactions (83).

In PC, SPP1 maintains mesenchymal cell fate through the SPP1-CD61-BMP2-GREM1 axis. Epithelial PDAC cells secrete SPP1, which binds to CD61 on mesenchymal PDAC cells, activating NF-κB and inducing BMP2 expression. BMP2 maintains mesenchymal identity, while GREM1 (BMP antagonist) feeds back to epithelial cells to sustain SPP1 secretion (36).

In summary, SPP1 functions as a master coordinator of EMT and metastatic progression rather than a single-pathway inducer: It orchestrates EMT at transcriptional (Snail/Slug/Twist/ZEB), post-transcriptional and post-translational levels, couples EMT to ECM remodeling and stemness acquisition, and maintains mesenchymal cell fate through reciprocal paracrine axes (for example, SPP1-CD61-BMP2-GREM1 in PDAC), thereby integrating invasion, therapy resistance and metastatic competence into a coordinated phenotypic program.

Epidemiological studies of multiple human cancers consistently show that SPP1 expression is dysregulated in tumor tissues. This repeated overexpression emphasizes the potential of SPP1 as a pan-cancer molecule and a promising diagnostic biomarker. Compared with normal tissues, SPP1 is significantly overexpressed in multiple solid tumors, including HCC, GC, PDAC, CRC, esophageal squamous cell carcinoma (ESCC), OC and BC (84-89). Correspondingly, elevated SPP1 protein levels have also been detected in the serum of patients with these malignancies. Clinically, high SPP1 expression is a well-recognized predictor of poor prognosis in cancers such as PC, CRC, clear cell renal cell carcinoma and LC (90-93). In addition, SPP1 overexpression is closely associated with metastatic progression in BC and melanoma (94-96). Overall, these epidemiological and clinical evidence confirm that SPP1 is a molecule that is consistently upregulated in cancer. Its strong correlation with aggressive disease and poor patient outcomes makes it a clinically relevant target for diagnosis, prognostic stratification and therapeutic development.

SPP1 is widely overexpressed across multiple human malignancies and exhibits distinct regulatory mechanisms in different tumor types. This chapter systematically reviews the research progress of SPP1 in PC, GC, CRC, BC, LC and HCC, with particular emphasis on its roles in tumor initiation and progression, metastasis, TME remodeling, immune regulation and therapeutic targeting.

SPP1 is a key regulator of TAM recruitment, polarization and function, contributing to the formation of an immunosuppressive and pro-TME. It promotes the recruitment and polarization of macrophages toward a pro-tumor M2-like phenotype through signaling pathways such as CD44/PI3K/AKT and CD44/JAK2/STAT3. This leads to increased secretion of factors that support tumor proliferation, angiogenesis, and ECM remodeling, including VEGFA and IL-6 (54,55,114,115). Hypoxia further upregulates SPP1 in macrophages and enhances their pro-tumor functions (116). In CRC, SPP1⁺ TAMs are enriched at the invasive front via the HIF-1α/STAT3 pathway, and their density correlates with lymph node metastasis (117). Similarly, single-cell RNA sequencing (scRNA-seq) of PDAC organoids showed that SPP1⁺ macrophages are associated with tumor angiogenesis and proliferation (118) (Fig. 2).

SPP1⁺ TAMs engage in critical crosstalk with other stromal cells. They interact with CAFs and neutrophils to promote ECM remodeling, angiogenesis and immune exclusion (83,119,120). In PDAC, spatially co-localized SPP1⁺ APOE⁺ TAMs and CTHRC1⁺ GREM1⁺ myofibroblastic CAFs (myCAFs) synergistically form an immunosuppressive microenvironment through active ECM deposition and induction of EMT (121). Analysis of 14 independent CRC cohorts found a positive correlation between FAP⁺ fibroblasts and SPP1⁺ TAMs, which is regulated by chemerin, TGF-β and interleukin-1. This interaction stimulates the formation of immune-excluded, desmoplastic structures that limit T cell infiltration and reduce the efficacy of anti-PD-L1 therapy (81). In HCC, DAB2⁺ and SPP1⁺ TAMs enhance the function of FAP⁺ CAFs through signals such as TGF-β, PDGF and ADM, participating in ECM organization and angiogenesis (122) (Fig. 2). Importantly, the CXCL9:SPP1 macrophage polarity index, defined by the expression ratio of CXCL9 and SPP1 (rather than conventional M1/M2 markers), has a strong prognostic association with survival across multiple cancers and predicts anti-PD-1 response (123). TAM-derived secreted SPP1 (predominantly OPN-a) mediates these paracrine effects, highlighting the isoform- and source-specificity of SPP1 function in the TME.

SPP1 directly impairs antitumor T cell function. Macrophage-derived SPP1 interacts with CD44 on T cells, inducing exhaustion markers (PD-1, TIM-3 and LAG-3) and suppressing cytotoxic activity via the SPP1-CD44 and CD155-CD226 pathways, thereby limiting tumor elimination (115,124,125). High SPP1 expression in the TME is associated with reduced T cell infiltration and proliferation, while inhibition of SPP1 can restore T cell activation and enhance immunotherapy efficacy (126). In metastatic GC, crosstalk between SPP1⁺ TAMs and exhausted CD8⁺ T cells is the key immunosuppressive mechanism (113). Furthermore, spatial transcriptomics in primary PDAC tissues identified SPP1 as a promoter of the immunosuppressive microenvironment at the tumor front, characterized by accumulation of FOXP3⁺ regulatory T cells (Tregs) (127). SPP1 can also be directly expressed by T cells. SPP1⁺ exhausted T (Tex) cells are associated with NKT cell activity and poor prognosis in OC and may be induced by immune checkpoint inhibition (107). Additionally, CAF-derived SPP1 activates CD44 on tumor cells to enhance stemness features in PC cells, as evidenced by increased CD44⁺, ALDH⁺ and AF⁺ populations and elevated self-renewal and drug-resistance markers (37). Thus, SPP1⁺ TAM-CAF interactions create a desmoplastic barrier that physically excludes T cells from tumor nests. In CRC, SPP1⁺ TAMs and FAP⁺ fibroblasts form immune-excluded structures that limit T cell infiltration and reduce the benefit of anti-PD-L1 therapy (15,81,98). Spatial transcriptomics in PDAC reveals that SPP1 promotes accumulation of FOXP3⁺ Tregs at tumor fronts (64). As depicted in Fig. 2, this SPP1-mediated T cell exhaustion represents a key node in the immunosuppressive network, where SPP1⁺ TAMs directly impair cytotoxic T cell function through SPP1-CD44 interactions.

SPP1 promotes the formation of an immunosuppressive microenvironment by recruiting and activating MDSCs. It is a key factor for recruiting MDSCs into the TME in prostate cancer (47). In BC models, SPP1 knockdown reduces MDSC infiltration and restores the cytotoxicity of CD8⁺ T cells (15). In monocytic MDSCs, SPP1 expression enhances the activities of arginase 1 and inducible nitric oxide synthase. This consumes local L-arginine and produces nitric oxide, which together inhibit T cell proliferation and function (128). This MDSC recruitment, illustrated in Fig. 2, further amplifies the SPP1-centered immunosuppressive circuit.

SPP1 reprograms quiescent fibroblasts into activated, pro-tumor CAFs. SPP1 binds to integrin αvβ5 on fibroblasts, triggering the FAK-Src-Ras-MAPK cascade, which phosphorylates the transcription factor AP-1 and upregulates CAF marker genes including FAP, α-SMA and PDGFRβ (129,130). Activated CAFs then secrete dense ECM, increase TME stiffness, and activate YAP signaling in tumor cells, promoting proliferation and invasion (131). Conversely, a positive feedback loop exists: SPP1-activated CAFs secrete TGF-β and IL-1, which in turn enhance the pro-inflammatory activity of SPP1⁺ macrophages. Conversely, in the context of PTEN deletion, activation of YAP1 in macrophages can directly induce SPP1 secretion through the β1 integrin-PYK2 pathway, thereby stimulating angiogenesis in glioma (132). Importantly, CAFs themselves can serve as a source of SPP1. In prostate cancer, SPP1⁺ myCAFs promote androgen deprivation therapy resistance through the SPP1-ERK paracrine mechanism (133). In NSCLC, co-culture with SPP1⁺ CAFs increases the metastatic potential of tumor cells (134) (Table II). As shown in Fig. 2, SPP1-induced CAF activation creates a feedforward loop that reinforces ECM deposition and immune exclusion, thereby establishing a physical barrier against T cell infiltration.

SPP1 fundamentally remodels the ECM to promote an invasive and immunosuppressive TME. SPP1⁺ macrophages and CAFs synergistically remodel the ECM by upregulating metalloproteinases (for example, MMP9, MMP12 and MMP7) that degrade structural components, and lysyl oxidase that cross-links collagen, increasing stiffness (81,116,126). This remodeling is closely related to hypoxic conditions and the maintenance of CSCs. Additionally, SPP1 enhances ECM stiffness, forming a dense desmoplastic barrier that physically excludes cytotoxic lymphocytes and contributes to immunotherapy resistance. In PDAC, SPP1⁺ CAFs are associated with a rigid matrix, and targeting SPP1 reduces this stiffness (135). In HCC, the interaction between SPP1⁺ macrophages and CAFs stimulates the formation of the tumor immune barrier, limiting immune cell infiltration into the tumor core (126). This ECM remodeling, summarized in Fig. 2, physically excludes T cells and creates a niche conducive to tumor invasion and therapy resistance.

Collectively, these findings establish SPP1 as a central orchestrator of the immunosuppressive TME: Secreted predominantly by SPP1⁺ TAMs and CAFs, SPP1 coordinates a multi-component regulatory network, driving M2-like macrophage polarization, CD8⁺ T cell exhaustion, MDSC recruitment, CAF activation and ECM remodeling, through reciprocal paracrine crosstalk that converges to form an immune-excluded, therapy-resistant (Fig. 2, Table III). Its multifaceted interactions with T cells, MDSCs, CAFs and the ECM make it a promising therapeutic target for reprogramming the TME to enhance immune infiltration and improve cancer treatment outcomes.

Single-cell transcriptome analyses have revealed that the expression of CXCL9 and SPP1 in TAMs is almost completely mutually exclusive at the single-cell level, forming two macrophage subsets with opposing functions: CXCL9⁺SPP1⁻ antitumor TAMs and CXCL9⁻SPP1⁺ pro-tumor TAMs. Traditional M1 (CD80 and ITGAX) and M2 (CD163 and MRC1) markers cannot distinguish these two subsets and are not significantly associated with patient prognosis, whereas a higher proportion of SPP1-high TAMs correlates with shorter overall survival (OS) (123).

SPP1⁺ TAMs are predominantly distributed in hypoxic regions of tumors and exhibit strong co-localization with GLUT1⁺ cells (a hypoxia marker). In vitro experiments have confirmed that hypoxia directly induces high SPP1 expression in macrophages via the HIF-1α pathway, which is the key regulator of their spatial distribution and functional polarization. Functionally, SPP1⁺ TAMs are highly coordinated with the pro-tumor programs of the entire TME: Their presence is significantly associated with the activation of EMT, angiogenesis, glycolysis and mTOR signaling pathways. As a secreted protein, SPP1 promotes tissue remodeling, angiogenesis and tumor metastasis by acting on integrin and CD44 receptors on the surface of fibroblasts and endothelial cells. Concurrently, SPP1⁺ TAMs inhibit the infiltration of T cells, B cells and dendritic cells, thereby establishing an immunosuppressive microenvironment. SPP1 inhibition combined with ICB synergistically reverses immunosuppression and overcomes immunotherapy resistance (36,38,81). SPP1 inhibition also sensitizes tumors to chemotherapy, targeted therapy and stromal-targeting agents (81). These findings have been validated in more than 10 human cancer types, including head and neck cancer, LC, CRC and liver cancer, indicating that TAM polarity defined by SPP1 is a ubiquitous feature of the TME with significant clinical prognostic value and therapeutic guiding implications (136).

In sum, elevated SPP1 expression in tumor tissues and patient serum correlates with poor prognosis across multiple cancers. Meta-analyses have confirmed that high SPP1 expression is associated with reduced OS and disease-free survival (90).

SPP1 is a key driver of tumor progression and immune suppression, making it an attractive therapeutic target. Small molecule inhibitors offer a promising approach for targeting SPP1 in tumor therapy, acting either on SPP1 itself or its receptors (Table IV).

Another approach is to disrupt the binding of SPP1 to its receptors, such as integrins and CD44. The cyclic RGD peptide cilengitide is a characteristic inhibitor of the key SPP1 receptors αvβ3 and αvβ5 integrins. Cilengitide competitively occupies the ligand binding site, potentially blocking SPP1-integrin interactions (141,148). In addition, blocking this interaction can also inhibit downstream FAK signaling, which is crucial for SPP1-mediated cell migration and invasion. As shown in Fig. 3, inhibition of the key SPP1 receptors αvβ3 and αvβ5 integrins will attenuate downstream FAK signaling, thereby reducing tumor cell migration and invasion, and potentially synergizing with immunotherapy by reversing the immunosuppressive TME (Fig. 3).

In summary, although SPP1 expression can be regulated through various upstream pathways and receptor-blocking drugs show antitumor effects, the research on highly specific, direct SPP1-targeting drugs is still in the early stage, representing a key area for future development.

Nucleic acid aptamers, especially RNA aptamers, are short, structured oligonucleotides that can bind to SPP1 with high affinity and specificity, thereby blocking its pro-cancer functions. This class of therapeutic agents has the advantages of low immunogenicity and the ability to be chemically modified to enhance stability and delivery (149). Their applications are further expanded, such as aptamer-drug conjugates (ApDCs) and aptamers constructed entirely from therapeutic nucleoside analogs (150). These modalities enable precise delivery by discriminating isoforms and cellular sources. OPN-c-specific ADCs deliver cytotoxic payloads to tumor cells, improving therapeutic index (144). SPP1 RNA aptamers offer low immunogenicity and modifiability, with preclinical efficacy in esophageal cancer (55). The diverse therapeutic strategies targeting SPP1, ranging from neutralizing antibodies and RNA aptamers to small molecule inhibitors and combination immunotherapies, each with distinct mechanisms of action, are illustrated in Fig. 3.

Preclinical studies have demonstrated the effectiveness of SPP1-targeting aptamers. In ESCC models, SPP1 RNA aptamers significantly inhibited tumor growth and reduced M2 macrophage infiltration (55). In mouse glioblastoma models, systemic administration of 4-1BB (CD137)-SPP1 bispecific aptamers improved survival by triggering co-stimulatory immune responses. Notably, SPP1 aptamers alone did not significantly improve survival, highlighting the benefits of combinatorial approaches (137). Beyond oncology, SPP1-specific aptamers have shown protective effects by regulating immune responses in non-neoplastic diseases such as chronic liver disease (151).

Overall, these findings highlight the great potential and clinical application prospects of SPP1-targeting aptamers in cancer therapy (Table IV, Fig. 3).

The clinical translation of SPP1-targeted therapy faces challenges, mainly due to the complexity of SPP1 interactions with multiple receptors and its indispensable role in normal physiology. SPP1 participates in key processes such as cell adhesion, tissue integrity and immune regulation, affecting the activation and migration of macrophages and T cells (152). Therefore, systemic inhibition of SPP1 has the potential to disrupt these functions, possibly leading to impaired tissue repair and dysregulated immune responses.

Despite these promising therapeutic strategies, the clinical development of SPP1 pathway inhibitors has encountered significant setbacks, underscoring the need for a more circumspect assessment. A cautionary precedent is cilengitide, a cyclic RGD peptide targeting αvβ3/αvβ5 integrins, key receptors mediating SPP1 signaling. Despite robust preclinical evidence of anti-angiogenic and anti-invasive activity and encouraging phase I/II results (141), the phase III CENTRIC trial (EORTC 26071-22072) in newly diagnosed MGMT-methylated glioblastoma failed to demonstrate any survival benefit (median OS 26.3 months in both cilengitide and control arms), leading to the termination of cilengitide development as an anticancer agent (153). The companion CORE phase II trial in MGMT-unmethylated glioblastoma yielded paradoxical results, with standard-dose cilengitide showing modest survival improvement while dose intensification did not improve outcomes, further highlighting the unpredictability of integrin-targeted therapy in clinical settings.

Several interconnected factors likely account for these failures and pose broader challenges to SPP1-targeted therapy. First, off-target toxicity remains a fundamental concern: SPP1/OPN plays indispensable physiological roles in bone remodeling, wound healing and immune homeostasis. OPN-deficient mice exhibit impaired wound healing with disorganized collagen fibrillogenesis and reduced tissue debridement (154), suggesting that systemic SPP1 inhibition could compromise tissue repair and bone integrity in patients. Second, the failure of integrin-targeted agents such as cilengitide indicates that broad-spectrum blockade of integrin signaling is often compromised by functional redundancy among integrin family members, interference with normal physiological processes, and paradoxical dose-response relationships. These limitations suggest that SPP1-targeted therapy should not aim for global suppression of SPP1 but rather focus on selectively disrupting the pro-tumorigenic signaling axis between SPP1 and key integrins, particularly αvβ3. αvβ3 plays a central role in therapeutic resistance by activating FAK/Src and PI3K/Akt pathways, thereby promoting cell survival, DNA damage repair and EMT, which collectively reduce the sensitivity of tumor cells to radiotherapy, chemotherapy and targeted agents. As one of the major ligands of αvβ3, elevated SPP1 expression is likely to drive acquired resistance through sustained activation of these pathways. Therefore, targeting the SPP1-αvβ3 interaction may offer greater specificity than direct integrin inhibition, while avoiding broad interference with the physiological functions of other integrins (155,156).

To overcome these hurdles, future drug development must prioritize precision-enhancing strategies, including tissue-specific targeting of tumor-associated SPP1 isoforms or TME-activated therapeutics to minimize off-target effects. Advanced modalities, such as bispecific antibodies co-targeting SPP1 and immune checkpoints, lipid nanoparticle-delivered siRNA for SPP1 silencing and SPP1-directed ApDCs, represent next-generation approaches. Rational combination therapy, pairing SPP1 inhibitors with chemotherapy, TGF-β blockers, or immune checkpoint inhibitors (ICIs; for example, anti-PD-1), may overcome drug resistance. Finally, biomarker-driven patient stratification using predictive biomarkers will ensure that SPP1-targeted therapies are administered to populations most likely to benefit.

Evidence demonstrating that SPP1 can exert opposing functions depending on the cellular and tissue context was discussed. For example, the tumor suppressor Chromobox Homolog 7 (CBX7) has been shown to directly inhibit SPP1 gene expression, and loss of CBX7 correlates with SPP1 overexpression and malignant progression in various carcinomas, including thyroid, breast and lung cancers (161,162). Conversely, the CBX7-interacting protein HMGA1b acts as a positive regulator of SPP1 gene expression, counteracting CBX7-mediated repression and highlighting the complex transcriptional regulation of SPP1 (161). This transcriptional duality reflects a broader theme in SPP1 biology: The same molecule can be co-opted for tumor promotion or tumor suppression depending on the regulatory networks active in a given cellular context. Additional studies have shown that OPN can act as a tumor suppressor in certain models; for instance, in prostate cancer, OPN has been reported to induce apoptosis in early-stage lesions, while promoting metastasis in advanced disease (163). These paradoxical observations underscore the importance of considering the specific biological context when interpreting SPP1 functions.