ECM-targeted cancer interventions: nanobiotechnological paradigms for imaging and therapy

Introduction

The Extracellular Matrix (ECM) is a complex, three-dimensional network of macromolecules, including collagens, elastin, proteoglycans, and glycoproteins such as fibronectin and laminins that provides essential structural and biochemical support to cells (Hynes, 2009). Far from being a static scaffold, the ECM is a dynamic signaling hub that regulates tissue homeostasis by sequestering and controlling the bioavailability of growth factors, cytokines, and chemokines (Humphries et al., 2006). This regulation occurs via bidirectional communication: cells sense and respond to biochemical and mechanical cues from the ECM (outside-in signaling), while simultaneously secreting and remodeling ECM components to alter their microenvironment (inside-out signaling). This vital reciprocal signaling, primarily through transmembrane receptors such as integrins, supports tissue homeostasis through the translation of microenvironmental data to cellular consequences. For example, effective ECM-mediated adhesion prevents detachment-induced cell death (anoikis) and preserves tissue integrity. Simultaneously, ECM-related signals also guide stem cell differentiation into the correct lineages, ensuring proper tissue composition. Additionally, the ECM acts as a spatial and temporal regulator for proliferation by adjusting the availability of growth factors and, upon injury, orchestrates repair by controlling cell migration and function (Jarvelainen et al., 2009). After all, this dialectic process is the reason cellular function synchronizes and reacts effectively to the structural and physiological requirements of the tissue and maintains long-term homeostasis.

During carcinogenesis, homeostatic balance is disrupted through sustained pro-tumorigenic signals from malignant and stromal cells. This leads to pathological ECM remodeling that no longer supports tissue integrity but instead actively drives cancer progression. Within the tumor microenvironment (TME), the ECM is co-opted by cancer cells and associated stromal cells, transforming it from a regulatory structure into an active partner in tumorigenesis (Lu et al., 2012). This remodeling, often manifesting as desmoplasia, involves two interconnected processes: the excessive production and stiffening of matrix components (e.g., collagen I) by cancer-associated fibroblasts (CAFs), and the dysregulated degradation of existing matrix by proteolytic enzymes such as MMPs. The result is a rigid, disorganized, and pro-tumorigenic matrix. This altered ECM facilitates epithelial-to-mesenchymal transition (EMT), guides metastatic invasion along aligned collagen fibers (Provenzano et al., 2008), and confers resistance to apoptosis and conventional therapies (Netti et al., 2021). The transformation of the ECM from a regulated structure to a disorganized scaffold is not merely a symptom but a fundamental catalyst of malignant progression (Figure 1).

Although the ECM plays a key role in cancer development and in some cases, can create a desmoplastic protective barrier which prevents therapeutic compounds and immune cells from penetrating, it has not been explored as extensively as therapeutic approaches designed against cancer cells. Most treatments like chemotherapy and targeted inhibitors aim to eliminate or adjust the cancer cell population. Targeting the ECM is little bit difficult because of its complex structure and the variations it shows across different types of tumors. And also there’s the concern about causing damage to the healthy ECM that is vital for normal organs. Previous attempts to create broad-spectrum inhibitors like broad-spectrum MMP inhibitors, have mostly fall flat in clinical trials because they had serious side effects. This clearly highlights the need for treatments which are more selective and localized (Gialeli et al., 2011).

By the review, we seek to address a significant translational gap by bridging fundamental ECM biology with recent advancements in nanobiotechnology. We examined current knowledge regarding the interplay between cancer cells and their surrounding stroma, focusing on how the specific physicochemical properties of the pathologically remodelled ECM contribute to cancer progression. We have explored innovative nanobiotechnological approaches, including nanoparticles, nanogels, and nanodevices that are designed to overcome the structural and drug delivery challenges posed by the tumor microenvironment. Specifically, the dense ECM around tumors acts as a physical shield that blocks drug penetration, raises fluid pressure within the tumor to hinder drug movement, and traps therapeutic agents, stopping them from reaching cancer cells in adequate amounts. The central focus of this study was to analyze how nanoscale systems can be engineered for targeted drug delivery, ECM reprogramming, and interference with pro-tumorigenic mechanotransduction pathways. Through this analysis, we provided a comprehensive perspective on how nanobiotechnological interventions targeting the ECM could transform the landscape of cancer therapy.

The extracellular matrix: architecture and function in normal physiology

The complex web of macromolecules, ECM, is essential for maintaining the tissue homeostasis. ECM performs two functions and it gives tissues their necessary structural characteristics and act as a signaling reservoir that regulates cellular activity (Frantz et al., 2010).

Physiological and pathological ECM remodeling in cancer progression

ECM is the physical and biochemical support for tissue function. It consists of structural proteins like collagen for strength (Gelse et al., 2003), adhesive proteins like fibronectin for cell attachment and migration (Hynes, 2009), and gel-like molecules like hyaluronan that hydrate tissues and regulate molecular transport (Theocharis et al., 2016). Importantly, the ECM is an active participant in cellular communication. Being a feedback center, it processes data and stores and mediates the release of signaling molecules, such as growth factors for behavior of cells and converts physical forces into intracellular biochemical signals through receptors such as integrins, ultimately regulating gene expression (Critchley, 2000). This sophisticated bidirectional signaling between cells and their matrix is fundamental to maintaining healthy tissue function and the distinct properties of each organ (Chaudhuri et al., 2020).

In cancer, this precisely regulated system is disrupted, with pathological remodeling transforming the ECM into an active promoter of tumor growth. Disruption of the homeostatic ECM and its transformation into a turbulent, pro-tumorigenic environment is a fundamental transformation process in tumorigenesis. Desmoplasia (ECM remodeling or pathological remodeling) represents a proactive, rather than an inactive byproduct, of cancer progression and plays a critical role in tumor cell fate (Lu et al., 2012). Pathological alteration alters the ECM architecture and signaling processes, allowing ECM to be converted from a regulator of homeostasis to an active powerhouse for tumor growth and anti-tumor therapy resistance, thus generating the very obstacle this review’s nanobiotechnological methods attempt to address.

Mechanisms of pathological remodeling in cancer progression

Pathological ECM remodeling is a dynamic, tumor-coordinated process driven by a corrupted dialogue between cancer cells and activated stromal cells, particularly cancer-associated fibroblasts (Kalluri, 2016). This process is defined by two interlinked, dysregulated activities that create a self-reinforcing cycle fueling progression:

• Aberrant Biosynthesis and Matrix Stiffening: Cancer cells secrete factors that activate resident fibroblasts into a pro-tumorigenic CAF phenotype. These CAFs become hyper-prolific factories, depositing excessive amounts of stiff fibrillar collagen (e.g., Type I) and fibronectin. This overproduction dramatically increases the mechanical stiffness of the tumor microenvironment (Egeblad et al., 2010). The resulting stiffness is not inert; it actively triggers pro-survival and invasive signaling pathways in cancer cells via mechanotransduction, creating a feed-forward loop that encourages further CAF activation.
• Dysregulated Proteolytic Degradation and Bioactive Signaling: Alongside the above, cancer cells and CAFs overexpress and secrete a suite of matrix-degrading enzymes, MMPs and ADAM (A Disintegrin and Metalloproteinase) family proteins (Gialeli et al., 2011). This leads to the uncontrolled breakdown of the existing ECM architecture. Crucially, this degradation plays an active role: it degarades structural barriers to facilitate local invasion and, more importantly, liberates sequestered growth factors and releases cryptic bioactive fragments from matrix proteins called matrikines. These matrikines exhibit distinct, pro-tumorigenic signaling properties that stimulate angiogenesis, chronic inflammation, and further cancer cell proliferation.

Functional consequences in tumorigenesis

The combined effect of these mechanisms transforms the ECM’s properties to collectively drive the tumor progression. The increased stiffness and altered fiber alignment provide contact guidance cues that facilitate cancer cell invasion and metastasis (Provenzano et al., 2008). This mechanically altered matrix also promotes proliferation and confers resistance to therapy through sustained integrin-mediated mechanotransduction. Cancer cells in this process sense the increased stiffness through integrin receptors, which induce a persistent intracellular signaling cascade that inhibits apoptosis and drives cell cycle progression. Furthermore, the resulting dense and disorganized architecture imposes a significant physical barrier. This leads to poor vascularization, hypoxia, and elevated interstitial fluid pressure and these harsh conditions not only foster tumor survival but also creates a major barrier to the penetration of chemotherapeutic drugs and immune cells, directly contributing to treatment failure (Netti et al., 2021) (Figure 1).

Why ECM remains under explored compared to tumor cell–centric therapies?

In spite of evidence proven beyond doubt for the ECM’s key and active role in tumourigenesis (Lu et al., 2012), it is a relatively undertargeted focal point when considered against therapeutic approaches directly targeting cancer cell genetics or cell proliferation. The intrinsic disparity relates to a coincidence of inherent biological complexity and challenging technical barriers faced by conventional drug discovery pipelines to date. The necessity for specialized approaches to the overcoming of these challenges has driven the present emphasis on advanced delivery systems (Avis et al., 2024).

Biological and structural hurdles

One of the biggest challenges is the universal presence and functional redundancy of major ECM components, a phenomenon in which multiple matrix proteins (e.g., collagens, fibronectin, laminins) share overlapping structural and signaling roles, making it difficult to disrupt pathological ECM without compromising essential homeostatic functions in healthy tissues (Hynes, 2009). Major matrix proteins, including collagens, are required for structural integrity and homeostatic function of essential, healthy organs. It has been highly challenging to develop agents that specifically interfere with the pathological ECM within the TME without eliciting dangerous, debilitating off-target systemic toxicity in normal tissues. In addition, the ECM is highly heterogeneous in its composition and stiffness with great variability both between tumour types (e.g., pancreatic vs. breast cancer desmoplasia) and within one tumour (Netti et al., 2021). The highly cross-linked, dense, and frequently pressurized structure of the rigid ECM itself provides a physical barrier that significantly restricts the diffusion and effective penetration of both large biologics and a wide variety of small-molecule drugs and representing a major delivery bottleneck (Provenzano et al., 2008).

Therapeutic and translational limitations

Conventional drug discovery has largely prioritized targeting highly “druggable” molecules, such as specific cell surface receptors (e.g., HER2, EGFR) or intracellular oncogenic pathways (e.g., kinase signaling). In contrast, therapeutic ECM modulation presents a fundamentally distinct challenge that necessitates strategies to selectively degrade, normalize, or reprogram a complex, extracellular, and physically obstructive structure. This distinction is crucial; whereas cellular targets are spatially confined and often accessible to small-molecule inhibitors, the tumor matrix constitutes an expansive, composite physical barrier. This intrinsic difficulty underlies the historical translational setbacks in this field Early clinical attempts, such as with broad-spectrum MMP inhibitors, aimed to block excessive matrix degradation but failed due to severe off-target toxicity (e.g., musculoskeletal syndrome) and a lack of efficacy (Gialeli et al., 2011). These failures highlight a core translational problem: the lack of specificity. Without the ability to localize matrix-modulating agents exclusively to the pathological tumor ECM, systemic administration inevitably disrupts the essential scaffolding of healthy tissues.

The difficulty of targeting the ECM is made significantly worse by the fact that the tumor stroma’s composition and structure are not the same in all patients or even within a single tumor. Not all patients have “ECM-rich” or desmoplastic tumors, and the structural and biochemical nature of the stroma differs intrinsically across various oncological indications. The absence of reliable biomarkers for patient stratification means clinical trials often include individuals unlikely to benefit from ECM-targeting agents, diluting therapeutic signals and contributing to perceived failure. Therefore, the critical translational gap is twofold: (1) the inability to achieve tumor-selective delivery of ECM-modulating drugs, and (2) the lack of predictive biomarkers to identify patients with susceptible, matrix-dependent tumors. These interconnected barriers have stalled the translation of compelling ECM biology into clinical practice. This impasse underscores the urgent need for and justifies the current focus on innovative nano-targeted delivery systems (Hosseini et al., 2024). Nanobiotechnology offers a potential solution by engineering carriers that can protect therapeutic payloads, accumulate specifically in the tumor via enhanced permeability, and release agents locally to remodel the pathological matrix while sparing healthy tissue.

Nanobiotechnology for ECM imaging and sensing

Quantum dots for ECM protein tracking

Quantum dots (QDs) are fluorescent nanoprobes that enable dynamic visualization of tumor ECM dynamics through superior photostability and size-tunable emission, allowing long-term tracking of matrix remodeling in cancer progression (Chand et al., 2025). Dkhar and colleagues developed multifunctional QD platforms for real-time imaging of malignant cells and therapeutic responses murine breast cancer models (4T1 tumors in BALB/c mice) (Dkhar et al., 2024). Multiplexed QD barcodes were used to simultaneously track collagen IV and laminin reorganization during metastatic progression in orthotopic breast cancer models (MDA-MB-231 tumors in NOD/SCID mice) (Chen et al., 2018). Key malignant ECM dynamics include fibronectin unfurling, which exposes cryptic binding sites for cell migration (Vogel, 2018), and collagen fiber reorientation that generates contact guidance tracks for invasion (Conklin et al., 2021).

The versatile surface chemistry of QDs enables multiple conjugation strategies for targeted ECM monitoring. Gao et al. (2025) developed quantum correlation imaging with QD biosensors to detect fibronectin and tenascin-C on 200 nm vesicles in patient samples. Park et al. (2023a) employed single-molecule tracking of QD-labeled RGD peptides to reveal integrin conformational dynamics during cancer cell migration. Iyer et al. (2010) used scFv-conjugated QDs to image cancer-associated fibroblasts, while Liu W. et al. (2024) extended this to multiplexed protease imaging. These strategies position QDs as versatile tools for ECM monitoring and theranostic applications.

Conjugation strategies and their cancer relevance

This makes it possible for multiple conjugation strategies to allow QDs to precisely target and monitor ECM proteins in the tumor microenvironment.

Structural studies based on protein conjugation

The QDs can also be conjugated to structural proteins like the hexameric core protein Hcp1 that provides stable constructions keeping photoluminescence for a long time (Schreiber et al., 2012). These conjugates can allow the clustering of the nanostructures, including for “photonic wires,” in aqueous solutions. The strategy can be generalized to the TME to design sensor arrays that capture spatial variations in ECM density and organization, which are hallmarks of desmoplasia. These methods produce highly accurate readouts of matrix structure relevant to tumor staging and monitoring therapeutic response to stroma-modulating agents. QD-based platforms have been modified to enable enhanced precision sensing and imaging of cancer cells, particularly in the imaging of malignant tumor cells to screen them for tumorigenesis and therapeutic response (Dkhar et al., 2024). Further, multiplexed imaging has contributed toward the simultaneous tracking of multiple ECM components to acquire a more holistic understanding of the matrix remodeling dynamics.

Single molecule tracking

The conjugation of QDs to peptides or antibodies will allow the tracking of individual ECM proteins (e.g., collagen, fibronectin) or their receptors (e.g., integrins) in live cells (Rossier et al., 2012). For therapeutic implications, that facilitates real-time tracking of pathological interactions at the molecular level, which should help in predicting the mechanism of drug action with respect to nanotherapeutics or enzyme inhibitors. Recent advances in quantum imaging technologies have substantially enriched these capabilities. A new imaging technology combining QD-based biosensing with quantum correlation imaging has emerged to enable precise analysis of surface protein markers with unprecedented resolution (Gao et al. 2004). This method enables identification and quantification of multiple biomarkers simultaneously, including on vesicles as small as 200 nm, allowing for targeted molecular profiling beyond the traditional microscopic scope. Also, a single-molecule tracking strategy for tracking integrin conformational dynamics is used in monitoring cancer cell migration on remodeled ECM substrates (Park et al., 2023a).

Aptamer-mediated labeling

Aptamers are short, synthetic oligonucleotides with a high binding specificity that have been shown to guide QDs to target membrane proteins which are required for ECM adhesion (Chen L. Q. et al., 2012; Bagalkot et al., 2007). It is an approach that allows the tracking of endocytic pathways, membrane diffusion, and vesicle transport. In cancer investigation, it supports the mapping of the internalization of ECM-encapsulated growth factors and integrin clusters. By tracking these pathways, researchers can investigate metastatic potential and the mechanisms of resistance to counter-regulate cell-matrix adhesion. Recent advances in this area have also produced novel electrochemiluminescence biosensors based on QD-aptamer conjugates showing remarkable sensitivity and specificity in the detection of tumor biomarkers over complex biological matrices (Ran et al., 2025). Furthermore, aptamer-functionalized QDs have been used for real-time imaging of matrix metalloproteinase activity in murine breast cancer models (4T1 tumors in BALB/c mice), providing insights into enzyme dynamics during invasion (Li et al., 2014). It has also been generalized to track exosome-mediated ECM communication, demonstrating how tumor-derived vesicles reprogram the stromal microenvironment for metastasis (Yang et al., 2020).

Single-chain antibody fragment conjugation

The conjugation of single-chain antibody fragments (scFv) to QDs enables precise visualization of protein targets with fine-tuned biospecificity based on controlled nanoparticle coupling (Iyer et al., 2010). This method is particularly powerful in visualizing the low-density yet clinically important ECM remodeling enzymes (i.e., different MMP isoforms) or signaling matrikines located in the dense tumor stroma. The specific localization and activity of these targets may be relevant for the generation of more directed inhibitors. It was recently reported that scFv-conjugated QDs could help in targeted imaging of cancer-associated fibroblasts with their secreted ECM proteins and they can discriminate between desmoplastic tumor area and adjacent stromal tissue (Chen et al., 2019). Moreover, multiplexed imaging with scFv-strained fragments with highly selective motifs has facilitated simultaneous scanning of multiple ECM components and their associated proteases, allowing elucidation of spatial relationships which are complex and may be correlated with invasive potential (Liu F. et al., 2024). This has also been used to guide intraoperative images, where scFv-QD probes directed at tumor-specific ECM signatures appear promising in searching for microscopic residual disease in surgical resection (Gao et al., 2021).

Translational challenges and considerations

While advantageous to use for these applications, there are still issues for applications in QD. Cytotoxic effects and the difficulty of the functionalization systems still need to be minimized for in vivo use (Nekouian et al., 2014). Recent attempts have been made in the field of biocompatibility of heavy-metal-free QDs and surface coatings to improve the biocompatibility of QD while maintaining optical performance (Li et al., 2014). Furthermore, the requirements to simplify the complexity of functionalization protocols should be pursued to allow reproducibility and scalability in clinical manufacturing contexts. Translating basic imaging techniques, such as single-molecule tracking of integrin-ECM dynamics into reliable clinical diagnostic algorithms will require the development of standardized assays capable of distinguishing benign stromal reactions from pathological desmoplasia (Thomas et al., 2024). Improvements in multiplexed imaging and machine learning-based assessment of QD imaging data enable the identification of quantitative ECM remodeling features, such as fiber orientation, density, and alignment patterns, which can serve as predictive biomarkers for treatment response and metastatic risk (Patel et al., 2025).

Gold nanoparticles for plasmonic sensing of ECM biomarkers

Gold nanoparticles (AuNPs) are increasingly employed as a primary platform for nanoplasmonic sensing largely owing to their distinctive localized surface plasmon resonance (LSPR). This optical effect arises from the oscillations of conduction electrons in the surface region, and has made AuNPs very sensitive to small changes of their local dielectric environment influenced by biomolecular binding events. This feature is directly exploited to assess the dynamic and pathological changes in the tumor ECM.

Synthesis, functionalization, and cancer-relevant sensing

AuNPs are prepared according to the precision parameters required, including size and morphology, such as citrate reduction or seed-mediated growth, which determines the overall optical features and the performance of sensing (Karnwal et al., 2024). Biorecognition components are used to tailor AuNPs for ECM studies, such as antibodies against collagen IV to sense basement membrane disruption, peptides directed against the EDA fibronectin splice variant uniquely upregulated in tumor stroma, or DNA aptamers specifically targeting MMP-9 for protease activity monitoring (García et al., 2022; Chen et al., 2023). The functionalized AuNPs promote a detectable LSPR wavelength shift upon target binding, allowing quantification of biomarkers. New multiplexed AuNP platforms, including silica-coated gold nanorods loaded with multiple antibodies, can now have simultaneous sensing of collagen IV, tenascin-C, and MMP-14 in patients’ plasma, facilitating early-stage breast cancer detection (Pramanik et al., 2023). Furthermore, branched AuNPs (nanostars) with improved electromagnetic field confinement can detect ultrasensitive low-abundance matrikine fragments in serum from pancreatic cancer patients (Špringer et al., 2014).

Role in cancer detection and progression monitoring

AuNP-based LSPR sensors detect serum or biopsy ECM biomarkers, such as collagen cross-linking byproducts and MMP activity markers, reflecting active desmoplasia and stromal remodeling. García and team (2022) showed that sensors that target a splice variant commonly upregulated in tumor ECM called fibronectin extra domain A (EDA) were able to distinguish benign and malignant stromal responses in murine models of breast cancer (4T1 tumors in BALB/c mice). These nanosensors track ECM evolution during tumor progression and facilitate estimation of metastatic potential and therapy-induced fibrosis. In a longitudinal study, it was found that AuNP-LSPR sensors were used to monitor MMP-9 activity in breast cancer (4T1 tumors in BALB/c mice) during tumor progression over time, establishing a direct relation between increased protease activity and invasive opportunities (Tai et al., 2022). For real-time matrikine detection, Tian et al. (2019) developed plasmonic gold nanopillar arrays to monitor and quantify ultralow concentrations of laminin-derived matrikines in murine models of pancreatic cancer (PANC-1 xenografts in nude mice). This local electromagnetic field enhancement achieved attomolar sensitivity of these sensors. Furthermore, Drifka et al. (2016) used LSPR to detect hydroxyproline content in collagen, a post-translational modification associated with desmoplastic intensity and treatment resistance in murine models of pancreatic cancer (PANC-1 xenografts in BALB/c nude mice).

Advanced imaging and theranostic potential

AuNPs, in addition to in vitro sensing, may be used as contrast agents for imaging of tumor stroma. Hyperspectral imaging enables real-time monitoring of functionalized AuNPs within living tissues, allowing visualization of their transport through the intricate ECM architecture (Spicer et al., 2018). Their biocompatibility enables theranostic systems to facilitate simultaneous identification of ECM-based biomarkers and photothermal therapy by utilizing LSPR absorption to ablate stromal and cancer cells without endangering normal tissues (Ali et al., 2019). ECM-targeted AuNPs functionalized with collagen-binding peptides accumulate in desmoplastic areas, allowing simultaneous imaging of collagen density and photothermal ablation of cancer-associated fibroblasts modulating the stroma in order to remodel stroma for increased therapeutic penetration (Fernandes, 2024). Nonetheless, dense ECM has physical constraints, with steric hindrance from collagen networks, increased interstitial fluid pressure opposing convective transport, and binding site barriers that trap nanoparticles superficially. In response to these problems, Park et al. (2023b) demonstrated in orthotopic models of 4T1 breast cancer that photothermal CAF ablation decreases collagen density and enhances immune infiltration. Wang et al. (2024) demonstrated that ECM-targeted AuNPs functionalized with collagen-binding peptides significantly enhanced the intratumoral penetration of co-administered doxorubicin in PANC-1 pancreatic cancer xenografts, achieving deeper drug distribution compared to non-targeted nanoparticles.

Considerations and challenges

Although promising, the application of AuNPs for translational analysis demands vigilance against long-term cytotoxicity and the complicated batch-to-batch reproducibility required for clinical assays. The development of specialized surface chemistries is essential for maintaining colloidal stability in complex biological fluids such as serum or plasma. These include polyethylene glycol (PEG) coatings that create hydrophilic barriers to reduce protein adsorption, zwitterionic surfaces that resist non-specific interactions through balanced charges, and dense oligonucleotide shells that provide steric stabilization (Sperling et al., 2022). These modifications also prevent non-specific binding, which occurs when nanoparticles adsorb unintended proteins (protein corona formation) or adhere to healthy tissues, leading to false signals and off-target effects. For example, PEGylated AuNPs show significantly reduced macrophage uptake and prolonged circulation compared to bare nanoparticles (Liu et al., 2023). Similarly, zwitterionic surface modifications have demonstrated enhanced stability and reduced protein corona formation in human plasma samples (Liu et al., 2023). Achieving this specificity is critical to accurately differentiate between pathological ECM remodeling and normal tissue turnover in clinical diagnostics.

Carbon nanotubes and graphene-based biosensors for ECM profiling

Carbon nanotubes (CNTs) and graphene represent a class of nanocarbon materials with exceptional electrical, optical, and mechanical properties, making them formidable platforms for developing next-generation biosensors capable of probing the tumor ECM.

Carbon nanotubes for deep-tissue imaging and sensing

Single-walled carbon nanotubes (SWCNTs) possess intrinsic photostable near-infrared (NIR) fluorescence, a spectral window where biological tissues exhibit minimal autofluorescence and absorption. This opens the possibility of deep-tissue imaging of tumor stroma (Jain et al., 2015). Designing SWCNTs with ECM-targeting ligands allows for the visualization of the spatial distribution and density of various matrix elements in preclinical models, enabling direct viewing of desmoplasia. Hence, SWCNTs modified in a peptide-targeting capacity targeting collagen-rich portions have been used for real-time imaging of tumor-associated fibrosis in murine models of breast cancer (4T1 tumors in BALB/c mice) and pancreatic cancer (PANC-1 xenografts in nude mice), where ECM density was found to be heterogeneous on the basis of the region where drug penetration was low (Hendriks et al., 2022). SWCNTs conjugated with antibodies against tenascin-C were also used to visualize stromal activation during early tumor progression and in response to anti-fibrotic therapies in pancreatic cancer (KPC mice) (Bhattacharya et al., 2023). Moreover, SWCNTs are the foundation for advanced optical biosensors because of their environment-sensitive photoluminescence. This optical sensitivity arises from changes in the SWCNT band gap and quantum efficiency in response to local dielectric shifts, allowing biochemical interactions to be converted into measurable optical signals (Hendriks et al., 2022). SWCNTs functionalized with quencher-containing peptide substrates, for example, the MMP-2-cleavable peptide sequence GPLGIAGQ conjugated to a fluorophore-quencher pair (e.g., Cy5.5 and BHQ-3), show fluorescence recovery upon cleavage by specific ECM enzymes such as MMP-2, which can provide real-time imaging of protease activity in the tumor microenvironment (Bremer et al., 2001; Kim et al., 2012). Likewise, SWCNTs coated with pH-sensitive polymers like PEG-acid-cleavable linkers or polyacrylic acid detect tumor acidosis via NIR fluorescence shifts, providing real-time readouts of glycolytic metabolism and ECM remodeling (Asadi et al., 2024). Their photostable NIR fluorescence enables long-term, non-invasive monitoring of ECM dynamics. For instance, (GT)₁₅ -DNA-wrapped SWCNTs implanted in murine pancreatic tumors tracked hydrogen peroxide fluctuations during chemotherapy, with signals reversing upon treatment withdrawal (Harvey et al., 2023). Similarly, peptide-functionalized SWCNTs incorporating the MMP-9-cleavable sequence PLGLAG and conjugated to a near-infrared dye, non-invasively monitored MMP protease activity in breast cancer models (4T1 tumors in BALB/c mice) over several days, directly visualizing enzyme activity changes during tumor progression and therapy (Bremer et al., 2001).

Graphene-based biosensors for ECM analysis

Due to its high conductivity and surface area, graphene is an effective transducer material for electrochemical and field-effect transistor (FET) biosensors (Garg et al., 2015; Pena-Bahamonde et al., 2022). In addition to antibodies targeting tenascin-C (Peiris et al., 2023) and laminin-332 (Okada et al., 2021), graphene sensors have been functionalized with DNA aptamers targeting osteopontin for breast cancer detection; peptide ligands for SPARC in desmoplastic tumors; and scFv fragments targeting fibronectin EDA splice variants that specifically belong to tumor stroma. Functionalization is performed using EDC-NHS chemistry or pyrene derivatives (Fu et al., 2023). Graphene field-effect transistor (GFET) arrays can be used to detect multiplexed cancer biomarkers. Yin et al. (2023) generated GFET arrays with femtomolar sensitivity that assessed pancreatic cancer exosomal markers (GPC-1 and CD63) in patient plasma. Li et al. (2014) introduced GFET sensors for measuring desmoplastic activity in breast cancer progression by targeting MMP-9 and collagen IV fragments.

Multimodal theranostic platforms

Apart from biosensing, nanocarbons are versatile theranostic platforms that combine diagnostic imaging technology with therapeutic intervention within the tumor stroma. SWCNTs operate as multimodal contrast agents, and the simultaneous conjugation of photoacoustic imaging (PAI) to obtain high-resolution data on ECM density paired with magnetic resonance imaging (MRI) permits detailed anatomical localization visualization by addition of gadolinium (Liu et al., 2011; Richard et al., 2008). This enables non-invasive assessment of stromal parameters including collagen organization and interstitial pressure during therapy (Gong et al., 2025). These platforms allow simultaneous therapeutic action. Specialized ECM-targeted SWCNTs, engineered through covalent conjugation of collagen-binding peptides or non-covalent wrapping with ECM-targeting polymers, serve as multimodal theranostic platforms by integrating diagnostic imaging and therapeutic intervention. Their near-infrared fluorescence enables non-invasive imaging of ECM distribution and tumor accumulation, while exposure to near-infrared irradiation generates localized hyperthermia. This thermal effect ablates cancer-associated fibroblasts and disrupts dense collagen architecture, thereby improving therapeutic penetration (Neelgund and Oki, 2021; Gong et al., 2025). Longitudinal photoacoustic imaging correlates stromal depletion with enhanced drug delivery (González et al., 2022). Current advanced technologies include blending MMP-responsive fluorescence quenching with siRNA delivery for in vivo pharmacodynamic monitoring and gene silencing within the tumor microenvironment (Harvey et al., 2023). Graphene-based nanoplatforms offer complementary functionalities for ECM-targeted theranostics. Graphene oxide nanosheets offer distinct theranostic advantages, serving as dual-modal imaging agents (fluorescence and photoacoustic) while also functioning as pH-responsive carriers for stroma-modulating drugs such as LOX inhibitor BAPN and TGF-β receptor inhibitor SB431542, which have been shown to reduce desmoplasia and enhance drug penetration in murine models of pancreatic cancer (PANC-1 xenografts in nude mice) (Zhou et al., 2024). By integrating diagnostic imaging and therapeutic intervention within a single nanoplatform, nanocarbon-based theranostics address the critical need for real-time target validation and treatment monitoring, potentially guiding patient selection for stroma-targeted cancer therapies (Chen et al., 2025).

Translational hurdles

The clinical translation of CNT and graphene-based biosensors involves significant challenges concerning long-term biological safety, potential immune recognition, and the scalability of manufacturing with strict quality control (Garg et al., 2015). Rigorous toxicological profiling and the development of standardized functionalization protocols are essential steps before these innovative platforms can fulfill their potential in clinical oncology for precise ECM monitoring and guided intervention.

Real-time visualization platforms for TME

Advances in nanobiotechnology paved the road for the construction of advanced imaging platforms for real-time evaluation of ECM dynamics, contributing to understanding the specific pathophysiology of cancer progression and therapeutic response.

Nanobubbles for ultrasound imaging of vascular leakiness and ECM penetration

Gas-filled nanobubbles (∼275 nm diameter) are sensitive ultrasound contrast agents. Their use goes beyond basic imaging to functionally map the tumor’s abnormal vasculature along with its ECM barrier. Microfluidic models of these nanobubbles have proven to be able to extravasate through the hyperpermeable tumor vasculature to provide a way to visualize and quantify flow and diffusion in the perivascular ECM of the nanobubbles in real-time (Cooley et al., 2023). As a visualization platform, this approach provides critical real-time imaging of nanoparticle extravasation dynamics, penetration depth, and spatial distribution within the tumor stroma, thus offering a valuable preclinical tool for assessing how nanotherapeutic candidates interact with and traverse the dense ECM barrier. This imaging capability directly addresses a critical challenge in desmoplastic tumors by providing real-time visualization of how the dense ECM barrier restricts therapeutic agent penetration, thereby identifying the physical obstacles that limit drug delivery.

Förster resonance energy transfer (FRET) sensors for proteolytic activity

To achieve high-resolution, quantitative measurements of enzymatic activity in a three-dimensional matrix, FRET-based sensors are embedded in synthetic, tissue-mimetic hydrogels. These sensors harbor a donor–acceptor fluorophore pair connected with a protease-cleavable peptide sequence, and upon enzymatic cleavage, fluorophore separation disrupts FRET efficiency, generating a measurable ratiometric signal (increased donor and decreased acceptor emission) that directly relates to proteolytic activity (Lee et al., 2020). These platforms have proven particularly valuable for detecting MMPs, the enzymes responsible for pathological ECM degradation during cancer invasion and metastasis. FRET hydrogels are formulated out of collagen- or hyaluronic acid-based matrices with adjustable stiffness to mirror native tumor ECM characteristics and offer a physiologically appropriate vehicle for investigating stroma-modulating therapeutics not available in a traditional 2D culture. Park et al. (2023c) conducted a study on FRET hydrogels to quantify a dose-dependent inhibition of pericellular proteolysis mediated by marimastat in association with decreased tumor cell invasion using murine models of breast cancer (4T1 tumors in BALB/c mice), and Choi et al. (2017) used that to describe MMP-14-specific inhibitors in murine models of breast cancer (MDA-MB-231 xenografts in nude mice). Longitudinal monitoring is essential because ECM remodeling is a dynamic process that unfolds over time. FRET sensors embedded in microfluidic tumor-on-a-chip devices address this need by enabling continuous, real-time tracking of MMP activity under controlled interstitial flow conditions. Using this platform, Nguyen et al. (2023) demonstrated that mechanical forces within the tumor microenvironment, such as fluid shear stress and interstitial pressure, directly influence therapeutic efficacy, revealing critical insights that cannot be obtained from static assays. This integrated approach bridges the gap between conventional biochemical assays and complex in vivo models, providing a reproducible and standardized platform for the preclinical evaluation of stroma-targeted therapies.

Dual-sensitive nanoprobes for mapping the tumor physicochemical landscape

A major innovation is the design of dual-sensitive fluorescent nanoprobes targeted for simultaneous activation by two important components of the tumor microenvironment: local acidosis (low pH) and elevated MMP expression. These nanoprobes are uniformly embedded within stiffness-tunable 3D hydrogels, composed of polyethylene glycol or hyaluronic acid crosslinked with MMP-cleavable peptides (e.g., GPLGIAGQ), allowing them to report on MMP activity and pH changes throughout a matrix that mimics tumor mechanics. It enables the spatial co-mapping of tumor margins (low pH) and local proteolytic activity (MMP-9 expression) in real-time (Rainu and Singh, 2024). This is an important dual-parameter sensing because it indicates the functional coupling of biochemical signaling and biophysical forces driving cancer progression. For example, it can visualize how an acidic, stiff microenvironment promotes increased MMP activity to drive invasion, presenting a robust means of examining the crosstalk of tumor-stroma and of validating multi-mechanism therapeutics.

Glycan-binding fluorophores

Rhobo6 is a recently developed glycan-binding fluorophore that provides wash-free visualization of ECM components in living tissues, circumventing multiple challenges of conventional labeling techniques. Although the current method of imaging ECM glycans is often dependent on fixation followed by permeabilization and multiple washing steps to remove unbound probes, it precludes dynamic studies in living specimens and introduces artifacts from sample processing (Fiore et al., 2024a). Other conventional glycan-binding agents like lectins or antibodies are also limited in tissue penetration owing to their large molecular size and exhibit rapid photobleaching that limits long-term imaging of ECM dynamics. Rhobo6 helps to overcome this issue with its small molecular size, improved penetration into tissue, and the fluorogenic nature of the molecule such that it turns fluorescent only after binding to glycans, which eliminates the need for washing steps and allows continuous observation and quantification of ECM remodeling in living tissues. It is useful by enhancing tissue penetration and reducing photobleaching and thus, an excellent candidate to investigate ECM dynamics within a variety of biological settings, such as tumor progression or metastatic niche formation, as well as stromal remodeling induced by therapy (Fiore et al., 2024b).

Although these developments in nanobiotechnology present potential opportunities in real-time ECM imaging, there are still issues of assuring the biocompatibility and safety of these nanomaterials to clinical use. More studies are required to deal with these issues and maximize the application of these technologies in therapeutic practice. In Figure 2 few nanobiotechnological approaches for ECM imaging and Biosensing were illustrated.

ECM-targeted nanotherapeutic approaches

Nanotherapeutic strategies that are based on ECM-targeting are aimed at improving drug delivery systems by taking advantage of the special behaviors of the ECM (Figure 3). These systems make use of ECM components to enhance biocompatibility, targeting and therapeutic efficacy, especially in cancer and fibrotic diseases. There are essential strategies and examples of ECM-specific drug delivery systems described as follows.

ECM-derived nanoparticles

Nanoparticles derived out of ECM proteins (e.g., collagen, gelatin) and glycosaminoglycans (e.g., hyaluronic acid) are used to deliver drugs and repair tissues (Hogan et al., 2023). In cancer therapy, gene delivery and regenerative medicine, these nanoparticles have demonstrated potential to increase drug targeting and tissue-specific responses (Hogan et al., 2023).

ECM-disrupting nanoconstructs for active tumor targeting

A prime example of a multi-mechanism delivery system is the c(RGDyK)-HAase-IONP nanoconstruct, designed to sequentially overcome barriers in the tumor microenvironment for improved drug delivery.

Active targeting

c(RGDyK) cyclic peptide acts as a high affinity ligand of the αvβ3 integrin receptor, commonly upregulated in tumor endothelial cells and multiple tumor cell types. This allows for active docking and sustained retention of nanoparticles at the tumor site compared to passive accumulation of nanoparticles (Chen and Chen, 2011).

ECM barrier degradation

Hyaluronic acid (HA) is an important glycosaminoglycan in the tumor ECM and causes the increase in the interstitial fluid pressure as well as steric barrier. The HAase enzyme conjugated to the nanoparticle degrades HA locally, which causes stroma loosening, reducing pressure, and expanding diffusion channels (Zhu and Marchant, 2011).

Integrated integrin-mediated targeting coupled with the disruption of ECM through enzymatic degradation facilitates the penetration and accumulation of nanoparticles deep within the tumor parenchyma, together increasing penetration and accumulation of nanoparticles and providing a broader drug distribution and enhanced therapeutic response when compared with those in passive, non-targeted systems that are non-disruptive (Chen and Chen, 2011; Zhu and Marchant, 2011).

ECM-mimetic materials for enhanced cellular interaction and uptake

In contrast to synthetic nanocarriers, ECM-mimetic materials are engineered to replicate native ECM components or display ECM-derived peptide sequences (e.g., RGD, IKVAV, YIGSR), thereby exploiting natural cell-matrix recognition pathways to enhance nanoparticle targeting and internalization (Mitrousis et al., 2018). These biomimetic carriers are constructed from natural polymers such as collagen, hyaluronic acid (HA), or gelatin, and their mechanisms of action include.

Ligand-receptor mediated uptake

Cancer cells overexpressing specific ECM receptors can be actively targeted using ECM-mimetic coatings. For example, hyaluronic acid (HA)-coated nanoparticles bind selectively to CD44 receptors, which are overexpressed on various cancer cells and cancer stem cells, promoting receptor-mediated endocytosis and enhancing intracellular drug delivery (Hwang et al., 2020a). A similar type of integrin-enriched RGD peptide-functionalized nanoparticles target integrins αvβ3 and αvβ5 on tumor endothelial cells and certain cancer cells, facilitating internalization via integrin-mediated endocytosis (Zhao et al., 2017).

Camouflage and biocompatibility

Coating nanoparticles with native ECM components such as collagen or albumin reduces opsonization and recognition by the mononuclear phagocyte system, thus extending circulation half-life (Liu et al., 2023). For example, gelatin-based nanoparticles exhibit prolonged systemic circulation due to their resemblance to denatured collagen, which minimizes immune clearance (Khan et al., 2022).

Stimuli-responsive drug release

ECM-mimetic hydrogels and nanoparticles may be developed to degrade in the tumor microenvironment under the influence of local stimuli. MMP-cleavable peptide crosslinkers added to HA-based hydrogels provide for controlled drug release in environments with high protease activity, for instance (Yin et al., 2020). Likewise, pH-responsive gelatin nanoparticles take advantage of the acidic tumor microenvironment to induce drug release exclusively in malignant tissues (Park et al., 2024).

With the biological cues that cells learn by evolution, ECM-mimetic nanomaterials actively engage cellular recognition pathways, facilitating rapid uptake and receptor-mediated endocytosis. The “Trojan horse” strategy not only enhances targeting precision and mitigates off-target toxicity, but also expands the therapeutic potential through enhanced accumulation of drugs in intracellular space (Mitrousis et al., 2018; Hwang et al., 2020b).

Mechanotherapy and nanocarrier engineering

It is the pathological stiffening of the ECM, in turn resulting from aberrant collagen cross-linking, hyaluronan deposition and the resulting fibroblast activity of cancer, leading to the inhibition of nanoparticle diffusion, compression of vasculature, increased interstitial fluid pressure, and sustains pro-invasive mechanotransduction through YAP/TAZ and Piezo1 signaling (Cassani et al., 2024; Wang J. et al., 2025). Mechanotherapy is designed to normalize this stiffened ECM with four synergistic treatments:

• Biochemical Remodeling: Nanoparticle-supplemented LOX inhibitors (e.g., BAPN) inhibit collagen cross-linking. Specifically, SPNcb (Semiconducting Polymer Nanoparticle conjugated with a LOX inhibitor and a photosensitizer) nanoremodelers decrease matrix stiffness by approximately 50%, as measured by atomic force microscopy and shear wave elastography, and can promote synergism with photodynamic immunotherapy (Zhang C et al., 2023).
• Physical Disruption: Ultrasound-induced microbubble cavitation disrupts collagen networks, resulting in lower solid stress and an increase in extravasation of nanoparticles (Hanson, 2018). Magnetically actuated nanoparticles, for example, Iron oxide nanoparticles (IONPs) coated with PEG and functionalized with RGD peptides, enable spatiotemporally controlled ECM deformation in response to external magnetic fields, facilitating dynamic stromal remodeling and enhanced penetration of co-administered therapeutic agents such as doxorubicin or paclitaxel into the tumor core (Zimina et al., 2024).
• Nanoparticle-Intrinsic Mechanodesign: Carrier geometry and mechanics are used to remove biophysical barriers. Fusiform, high-curvature nanoparticles such as gold nanostars or mesoporous silica nanorods (aspect ratio >3) possess lower hydrodynamic drag that allow increased extravasation against interstitial fluid pressure and promote rotation-dependent “hopping” during interstitial transit. A similar trend emerges for softer, more deformable nanoparticles such as liposomes with lower cross-linking density or poly(ethylene glycol) (PEG) hydrogels, which penetrate deeper into narrow ECM pores by adapting their shape to confined spaces (Jiang et al., 2024).
• Mechanosignaling Pathway Blockade: Nanocarrier–delivered integrin antagonists, FAK inhibitors or YAP/TAZ interference agents disrupt pathological feedback loops. Inhibitors of Piezo1/TRPV4 inhibit calcium-dependent mechanotransduction, thereby restricting metastatic growth (Fu et al., 2025).

Taken together, these approaches move from a passive ECM evasion model to active, mechanistically precise stromal normalization, and highlight mechanotherapy as a central approach in advanced nanomedicine design (Cassani et al., 2024; Chen et al., 2025).

Specific case studies

PEGPH20 (Pegvorhyaluronidase alfa)

It was a clinical-stage therapy developed to degrade the HA-rich scaffold of pancreatic ductal adenocarcinoma (PDAC), which is attributed to dense hyaluronan-driven desmoplasia. PEGPH20 along with chemotherapy (nab-paclitaxel/gemcitabine) promoted better progression-free survival in patients with high-HA tumors, in phase Ib/II studies. Subsequent phase III trials, however, indicated increased occurrences of thromboembolic events without any improvement in overall survival, leading to the halt of PEGPH20’s development while also providing valuable insights into the intricacies of disrupting stroma (Hingorani et al., 2018; Van Cutsem et al., 2020).

MMP-2/9-activated therapeutic nanoparticles

Researchers have developed nanoparticles designed to release drugs when cleaved by MMP-2 or MMP-9, enzymes that are often elevated within the tumor ECM. For instance, polyamidoamine (PAMAM) dendrimers, highly branched, spherical polymers with well-defined surface functionality, were conjugated to doxorubicin via an MMP-9-cleavable peptide linker. These dendrimer-drug conjugates self-assemble into ∼10 nm nanoparticles that remain inert during circulation but undergo rapid drug release upon encountering MMP-9-rich tumor microenvironments. In studies using orthotopic mouse models of breast cancer (4T1 mammary tumors in BALB/c mice), these nanoparticles selectively accumulated within tumor tissue and facilitated targeted doxorubicin release upon activation by tumor-associated MMP-9. This approach resulted in enhanced suppression of tumor growth and reduced systemic toxicity compared to both free doxorubicin and nanoparticles lacking cleavable linkers (Zhu et al., 2019).

Collagen-binding immunotherapy

Instead of breaking down collagen, this approach focuses on delivering therapeutics directly to it. Liu et al. (2022) developed a SIRPαFc fusion protein conjugated with a collagen-binding domain (CBD-TKKTLRT) derived from von Willebrand factor. In studies using A549 non-small cell lung cancer xenografts in nude mice, this collagen-targeting immunoconjugate demonstrated significantly enhanced tumor accumulation and retention compared to unmodified SIRPαFc. The CBD-conjugated protein remained within the tumor ECM for extended periods, resulting in markedly increased infiltration of CD11b+F4/80+ macrophages, particularly the anti-tumor MHC II+ M1 phenotype, within tumor tissues. This targeted approach achieved superior antitumor efficacy compared to conventional antibody alone. This example illustrates how utilizing the ECM can serve as a reservoir for prolonged localized drug effects, enhancing immune cell recruitment and therapeutic response (Chen et al., 2018).

Inhibition of ECM remodeling enzymes

LOX catalyzes collagen and elastin crosslinking, increasing ECM stiffness to promote tumor progression and therapy resistance (Yang et al., 2020). To counteract this, LOX-targeting lipid nanovesicles have been developed that disrupt collagen alignment in triple-negative breast cancer, enhancing drug penetration (Vita et al., 2021). Polymeric nanoparticles (e.g., PLGA-based or PEGylated dendrimers) encapsulating the LOX inhibitor BAPN (β-aminopropionitrile) have also shown sustained release and reduced metastasis in murine models of breast cancer (4T1 tumors in BALB/c mice) (Kanapathipillai et al., 2012). A recent “polymeric ECM nanoremodeler” developed by Patrick et al. (2023) consists of a pH-responsive diblock copolymer (PEG-b-PDPA) loaded with the photosensitizer chlorin e6 and the LOX inhibitor BAPN. Upon accumulation in the acidic tumor microenvironment, the nanoparticle disassembles to release both payloads; subsequent light activation generates reactive oxygen species to induce immunogenic cell death while LOX inhibition concurrently reduces ECM stiffness, together promoting T cell infiltration and suppressing desmoplastic tumor growth (Patrick et al., 2023). Separately, MMPs drive ECM degradation during cancer invasion. Although early broad-spectrum MMP inhibitors failed in clinical trials due to dose-limiting toxicity and lack of specificity, emerging nanotechnological approaches now enable targeted delivery and isoform-specific inhibition, reviving the therapeutic potential of targeting these enzymes (Raeeszadeh-Sarmazdeh et al., 2020). Examples include gelatinase-responsive nanoparticles releasing the broad-spectrum MMP inhibitor marimastat specifically within MMP-rich tumors in murine models of breast cancer (4T1 tumors in BALB/c mice) (Liu et al., 2022), gold nanoparticles functionalized with the selective MMP-14 inhibitory peptide CTTHWGFTLC that blocks cancer cell invasion in murine models of breast cancer (4T1 tumors in BALB/c mice) without affecting off-target MMPs (Park et al., 2023d), and MMP-responsive hydrogels crosslinked with MMP-cleavable peptides (GPLGIAGQ) that release anti-PD-L1 antibodies in response to local protease activity in murine models of breast cancer (4T1 tumors in BALB/c mice) (Guo et al., 2024). These targeted strategies represent significant advances over historical MMP inhibitors and are progressing through preclinical evaluation.

Although these ECM-targeted strategies have potential in improving cancer treatments, there are still problems with specificity and reduction of possible side effects of ECM modulation.

Disrupting ECM–Cell interactions

Disrupting ECM-cell interactions is a promising strategy in cancer nanotherapy, as the ECM acts as a mechanical barrier that hinders drug delivery to tumor cells. Various nanoparticle-based approaches have been developed to modulate the ECM, thereby enhancing drug penetration and therapeutic efficacy (Luo and Chen, 2023).

Nanoparticle strategies of ECM disruption

Direct removal of ECM components

Nanoparticles are increasingly employed to deliver anti-fibrotic agents that decrease ECM density, thereby enhancing drug penetration (Chen M. et al., 2023). For example, liposomal nanoparticles transporting collagenase IV degrade fibrillar collagens (type I and III), while those delivering MMP-9 target denatured collagen and basement membrane components in murine breast cancer xenografts (MDA-MB-231 tumors in nude mice) (Zinger et al., 2019). Similarly, hyaluronidase-conjugated silica nanoparticles enzymatically digest hyaluronic acid in murine pancreatic ductal adenocarcinoma models (PANC-1 xenografts in nude mice) (Zhang et al., 2026). Additionally, polymeric micelles (PEG-PLGA) delivering the anti-fibrotic drug pirfenidone downregulate collagen I and fibronectin expression in cancer-associated fibroblasts, reducing fibrosis in desmoplastic melanoma (B16F10 tumors in C57BL/6 mice) and triple-negative breast cancer models (4T1 tumors in BALB/c mice) (Voutouri et al., 2024).

Mechanical and biomimetic approaches

Mechanical stimulation combined with biomimetic nanoparticles offers a dynamic approach to remodel the ECM and facilitate drug delivery (Aghlara-Fotovat et al., 2021). Ultrasound-triggered mechanical stimulation using perfluorocarbon nanodroplets or gas-filled microbubbles induces transient cavitation or acoustic streaming, physically disrupting dense collagen networks and temporarily increasing local porosity of the tumor stroma (Huang et al., 2022). Biomimetic strategies exploit ECM-derived peptides to regulate stromal cell behavior. For example, gold nanoparticles functionalized with collagen-mimetic peptides (GPO)∼10∼ or dendrimers displaying fibronectin-derived fragments (PHSRN) competitively occupy integrin receptors on CAFs, inhibiting pro-fibrotic signaling and excessive ECM production. These nanoparticles have been used to deliver TGF-β inhibitors (SB431542) or FAK inhibitors (defactinib) directly to activated fibroblasts (Smith et al., 2023). Additionally, externally guided magnetically responsive nanoparticles, such as, iron oxide cores coated with PEG and conjugated with RGD peptides, exert directed physical forces to reorganize collagen fibers in murine models of breast cancer (4T1 tumors in BALB/c mice), reducing interstitial fluid pressure and promoting convective transport of therapeutic agents (e.g., doxorubicin or paclitaxel) into the tumor core (Zimina et al., 2024). These biomimetic approaches harness innate cell-matrix recognition to achieve sustained stromal remodeling with minimal off-target effects.

Cancer therapy implication

Improved drug delivery

The dense, desmoplastic ECM characteristic of many solid tumors acts as a physical barrier that limits the penetration and uniform distribution of chemotherapeutic agents. ECM-targeted nanotherapeutics address this challenge by actively remodeling the tumor stroma to enhance drug delivery. For instance, hyaluronidase-conjugated nanoparticles degrade hyaluronic acid-rich matrices in pancreatic ductal adenocarcinoma, reducing interstitial fluid pressure and increasing the accumulation of co-administered chemotherapeutics such as gemcitabine (Zhang et al., 2026). Similarly, collagenase-loaded liposomes have been shown to disrupt perivascular collagen barriers, improving the transvascular transport and intratumoral distribution of doxorubicin in breast cancer models (Vogel, 2018). These ECM-priming strategies create a window of opportunity for enhanced therapeutic penetration, potentially improving patient outcomes (Luo and Chen, 2023).

Overcoming therapy resistance

Tumor microenvironment-mediated resistance represents a major obstacle to effective cancer treatment. ECM stiffness itself promotes cancerous cell survival signaling through integrin-mediated mechanotransduction, activating pathways such as PI3K/AKT that confer resistance to both chemotherapy and targeted agents (Sleeboom et al., 2024). By normalizing ECM architecture, nanoparticle-based strategies can resensitize tumors to conventional therapies. For example, LOX-inhibiting lipid nanovesicles composed of phosphatidylcholine (PC) and cholesterol, encapsulating BAPN, reduce collagen crosslinking in triple-negative breast cancer by irreversibly binding to the active site of LOX, thereby decreasing tissue stiffness and restoring sensitivity to paclitaxel (Vita et al., 2021). Additionally, MMP-responsive nanoparticles developed with peptide linkers (e.g., GPLGIAGQ) cleavable by tumor-associated MMPs, release TGF-β inhibitors specifically within the tumor microenvironment, thereby reversing CAF-mediated chemoprotection and enhancing chemosensitivity in pancreatic cancer models (KPC mice) by disrupting the protective stromal niche that shields cancer cells from chemotherapy (Ji et al., 2023).

Radiotherapy modulation

The efficacy of radiotherapy is profoundly influenced by the tumor microenvironment. Tumor hypoxia, exacerbated by dense ECM compressing blood vessels, limits radiation-induced DNA damage and promotes radioresistance (Ran et al., 2023). ECM-targeted interventions can alleviate hypoxia by normalizing the abnormal tumor vasculature. For instance, hyaluronan degradation using PEGPH20, when combined with radiation, has demonstrated improved tumor oxygenation and enhanced radiation response in preclinical models (pancreatic cancer, KPC mice) (Jacobetz et al., 2022). Furthermore, gold nanoparticles functionalized with ECM-binding peptides such as collagen-binding peptide (CBP, sequence TKKTLRT) or fibronectin-binding peptide (FBP, sequence WQPPRARI) can be retained within the tumor stroma to serve as radiosensitizers, amplifying local radiation delivery specifically within desmoplastic regions (Singh et al., 2024).

Photothermal and photodynamic therapy enhancement

Photothermal therapy (PTT) employs nanoparticles to convert near-infrared light into localized heat, ablating cancer cells within the tumor (Cheng et al., 2021). However, heterogeneous heat distribution due to dense ECM often limits PTT efficacy. ECM-depleting strategies can overcome this limitation; for example, pretreatment with collagenase-loaded nanoparticles prior to PTT has been shown to improve heat penetration and achieve more uniform tumor ablation in desmoplastic melanoma models (B16F10 tumors in C57BL/6 mice) (Huang et al., 2022). Similarly, photodynamic therapy (PDT) generates reactive oxygen species to kill cancer cells, but its effectiveness is hampered by hypoxia. ECM remodeling nanoparticles that concurrently deliver LOX inhibitors and photosensitizers have demonstrated dual benefits: LOX inhibition reduces stromal density to improve oxygenation, while PDT induces immunogenic cell death, together promoting anti-tumor immune responses (Patrick et al., 2023).

Nanoparticle-mediated ECM oxygenation for radiosensitization

Direct hypoxia amelioration through catalytic oxygen generation by manganese dioxide (MnO₂) nanoparticles is realized through the catalytic conversion of endogenous tumor H₂O₂ to O₂, which locally promotes pO₂ and degrades stromal barriers. MnO₂ nanosheets combined with radiotherapy produced total tumor regression in 60% of treated mice in murine breast cancer models (4T1 tumors in BALB/c mice) using external beam radiation (6 Gy) (Zhu et al., 2016). When combined with collagen-binding catalase nanoparticles, sustained oxygen production (∼40% increase in pO₂ over 48 h) was achieved across multiple radiation fractions, significantly improving fractionated radiotherapy outcomes in lung cancer xenografts (A549 tumors in nude mice) (Song et al., 2016). Additional enzymatic ECM degradation strategies including hyaluronidase-conjugated gold nanoparticles (Singh et al., 2024) and collagenase-loaded liposomes (Zinger et al., 2019), further alleviate vascular compression to enhance tumor perfusion and oxygenation. LOX inhibitors, such as BAPN delivered via lipid nanovesicles, decrease collagen alignment and improve vascular function, enhancing radiation response in triple-negative breast cancer models (Vita et al., 2021). These ECM-targeted strategies address the hypoxia barrier through vascular decompression, stiffness normalization, and direct oxygen generation, providing multiple approaches to enhance radiotherapy efficacy in desmoplastic tumors.

ECM-targeted nanotherapeutics for immunomodulation

The dense and immunosuppressive tumor ECM imposes both physical and immunological barriers that limit immunotherapy efficacy. ECM-targeted nanotherapeutics address these barriers through multiple interconnected mechanisms, enhancing the delivery, activation, and function of immunotherapeutic agents.

Stromal-targeted delivery of immunomodulators

Nanoparticles functionalized with ligands against overexpressed ECM components (collagen, fibronectin, tenascin-C) or stromal cell markers (fibroblast activation protein) enable precise accumulation within the tumor stroma. This allows localized delivery of TLR agonists, cytokines, or STING agonists directly into the immunosuppressive microenvironment, converting immunologically “cold” tumors into “hot” ones while limiting systemic toxicity (Kar et al., 2025). For instance, STING agonist-loaded nanoparticles activate innate immune pathways, stimulating type I interferon responses and augmenting dendritic cell cross-presentation to cytotoxic T lymphocytes (Luo et al., 2017). TLR agonist-encapsulated ECM-penetrating nanogels reprogram tumor-associated macrophages from pro-tumor M2 to anti-tumor M1 phenotypes, simultaneously alleviating immune suppression and enhancing CD8⁺ T cell infiltration (Voutouri et al., 2024).

Reprogramming the stromal niche via ECM-mediated pathways

Beyond simply delivering immunomodulators to the stroma, ECM-targeted nanotherapeutics can actively interfere with pro-tumorigenic signaling pathways orchestrated by the ECM. The chemokine CXCL12, which is abundantly bound to ECM components within the tumor stroma, recruits immunosuppressive cells via its receptor CXCR4. Nanoparticle-based blockers of the CXCL12/CXCR4 axis, such as polymeric nanoparticles loaded with AMD3100 (a CXCR4 antagonist) disrupt the recruitment and retention of immunosuppressive cells (e.g., myeloid-derived suppressor cells) within the tumor microenvironment. A more sophisticated strategy involves single ECM-targeted nanoplatforms designed for co-delivery of CSF-1R antagonists and STING agonists. CSF-1R signaling promotes M2 macrophage polarization, while STING activation drives pro-inflammatory responses. By simultaneously blocking CSF-1R and activating STING pathways within the same stromal niche, these nanoparticles prevent immunosuppressive M2 polarization while actively promoting pro-inflammatory M1 phenotypes, thereby enhancing antigen presentation and facilitating cytotoxic T lymphocyte infiltration into previously resistant tumors (Nam et al., 2019).

Photothermal immunomodulation and ECM remodeling

PTT not only ablates tumor cells but also actively remodels the ECM to enhance immune infiltration. Gold nanorods irradiated with an 808 nm laser (1 W/cm², 5–10 min) generate localized hyperthermia (45 °C–55 °C) that denatures collagen and disrupts fibrillar architecture. In breast cancer models, PTT with silica-gold nanoshells (808 nm, 2 W/cm², 5 min) resulted in a 40% reduction in tumor collagen content and a threefold increase in nanoparticle penetration depth (Marangon et al., 2017; Zhang et al., 2022). Photothermal nanomaterials also regulate cancer-associated fibroblasts (CAFs). Lipid-coated gold nanostars functionalized with FAP-α antibodies and irradiated with 808 nm laser (1.5 W/cm², 5 min) selectively ablate CAFs while releasing TGF-β inhibitors, downregulating collagen I and fibronectin production (Li et al., 2022; Chen X. et al., 2023). Carbon-based photothermal agents such as SWCNTs and reduced graphene oxide (rGO) nanosheets similarly remodel desmoplastic pancreatic tumors (PANC-1 xenografts in nude mice) and promote T-cell infiltration while increasing checkpoint blockade efficacy (Kermanizadeh et al., 2021).

Bioinspired lipoprotein nanocarriers for photothermal ECM remodeling

Bioinspired lipoprotein nanocarriers are engineered to mimic endogenous high-density lipoproteins (HDLs), natural lipid-protein complexes responsible for cholesterol transport. These nanocarriers typically consist of a phospholipid monolayer shell surrounding a hydrophobic core (e.g., gold nanocrystals or copper sulfide), with surface-embedded apolipoprotein-mimetic peptides that enable recognition by scavenger receptors (e.g., SR-B1) overexpressed on tumor cells and stromal cells. This biomimetic design exploits natural transport pathways, including receptor-mediated transcytosis and enhanced permeability and retention (EPR) effects, for efficient tumor accumulation (Tang et al., 2022). After localization, photothermal lipoproteins such as gold nanocrystal-loaded reconstituted high-density lipoproteins (rHDL) or copper sulfide-encapsulated liposomes produce localized hyperthermia upon near-infrared exposure, leading to collagen denaturation and hyaluronic acid depolymerization. The resulting decreased matrix density and interstitial fluid pressure create transient pores that facilitate penetration of co-administered immunotherapies (Guo et al., 2024). Thermal disruption combined with enzymatic degradation results in synergistic and sustained stromal normalization, promoting immune cell infiltration and reducing recurrence (Liu F. et al., 2024). However, continuing challenges, including limited laser penetration in deep-seated tumors, potential off-target thermal injury to surrounding healthy tissues, and manufacturing complexity, hinder clinical translation. Current research focuses on ultrasound-activated systems and enhanced tumor-specific targeting to overcome these limitations (Liu W. et al., 2024).

Immunotherapy sensitization

The ECM acts as a physical and immunological barrier to immune cell infiltration. Dense collagen networks exclude cytotoxic T lymphocytes from the tumor core, while matricellular proteins such as tenascin-C suppress effector function. Nanoparticle-mediated ECM modulation can overcome these barriers. For instance, losartan-loaded lipid nanoparticles have been shown to downregulate TGF-β1 expression in cancer-associated fibroblasts, reducing collagen I synthesis and hyaluronan production. This stromal remodeling increased CD8⁺ T cell infiltration and enhanced anti-PD-1 efficacy in previously resistant hepatocellular carcinoma models (Hepa1-6 tumors in C57BL/6 mice) (Zhao et al., 2023). Additionally, MMP-responsive hydrogels, composed of PEG or hyaluronic acid (HA) crosslinked with MMP-cleavable peptides, enable spatiotemporally controlled release of immune checkpoint inhibitors such as anti-PD-L1 antibodies specifically within protease-rich tumor microenvironments, thereby reducing systemic immune-related adverse events (Guo et al., 2024).

ECM biomarkers for patient stratification

The ECM has emerged as an important determinant of immune checkpoint inhibitor (ICI) response, with ECM-derived biomarkers offering significant potential for patient selection and treatment optimization. ECM remodeling directly modulates immune cell infiltration and activity within tumors, thereby influencing responsiveness to immunotherapy (Fejza et al., 2023). Collagen degradation products have shown particular promise as predictive biomarkers. Elevated serum levels of PRO-C3 (a marker of type III collagen formation) and reC1M (a fragment of type I collagen degraded by MMPs) correlate with poor overall survival in patients receiving ICI therapy. Paradoxically, these same markers may identify patients most likely to benefit from ECM-targeted immunotherapy combinations, as high collagen turnover indicates an active desmoplastic stroma amenable to modulation (Bistrup et al., 2024). Early serum proteome changes during treatment also hold predictive value. Goh et al. (2025) identified that increases in pro-inflammatory cytokines (IL-6, TNF-α) and decreases in immunosuppressive factors (TGF-β) within the first 2–4 weeks of therapy predict response to anti-angiogenic agent (bevacizumab) plus ICI combinations in advanced nasopharyngeal cancer. Given the complexity of the tumor microenvironment, a multicomponent approach integrating multiple biomarker classes, including ECM biomarkers (PRO-C3, reC1M, fibronectin EDA), immune cell phenotypes (CD8⁺ T cell infiltration, TAM polarization), and tumor genomic characteristics (mutation burden, microsatellite instability) is essential for precision ICI response prediction (Roelofsen et al., 2022). In advanced renal cell carcinoma, elevated levels of PRO-C19 and TGFβ-LAP correlate with shorter survival but paradoxically predict better responses to nivolumab-ipilimumab combination therapy (Gupta et al., 2022). ECM-guided stratification in lung cancer has enabled patient grouping based on ECM grades, corresponding with distinct clinical features and mutation profiles, allowing optimization of therapeutic strategies (Dansık et al., 2025). Collagen quality, particularly type V collagen pro-peptide levels, has proven more predictive than collagen quantity, with elevated levels associated with increased mortality across multiple solid tumors (Nissen et al., 2023).

Translational and clinical perspectives

The ECM is closely linked with tumor cells and the surrounding stroma. Because of its complexity, this network is a promising yet still underused target for therapy. Its dynamic and context-dependent remodeling by tumor progression influences nearly all aspects of malignancy, from sustained proliferation and immune evasion to metastatic dissemination and therapy resistance. New approaches to normalizing the tumor ECM rather than merely inhibiting malignant cells are emerging as a valuable complementary element of traditional oncologic therapies. These matrix-focused modalities range from inhibition of pro-fibrotic enzymes, disruption of pathological ECM-cell signaling, and stromal cell phenotypic reprogramming to various mechanistic strategies. However, the successful transfer of these promising preclinical techniques to durable clinical outcomes is constrained by some persistent issues. Of these, the most prominent are the heterogeneity of ECM composition across and within tumor types, the difficulties in targeted delivery without disturbing normal tissue homeostasis and the low predictive value of current preclinical models. A close look at critical clinical case studies (both successful and halted) reveals the translational challenge and highlights the imperative requirement for adaptive, biomarker-targeted trial designs in future studies.

Therapeutic strategies

Targeting ECM Components: Instead of non-targeted matrix destruction, this strategy utilizes two mechanistically different modalities of action against pro-tumorigenic ECM macromolecules. Anti-collagen strategies target collagen biosynthesis, mainly for type I and type III fibrillar collagens, by targeting essential enzymes for post-translational modification (e.g., prolyl hydroxylases), or by inhibiting integrin-mediated signaling cascades dependent on collagen architecture to govern tumor cell migration and survival. Anti-hyaluronan strategies involve either pharmacological inhibition of hyaluronan synthase isoforms (HAS2, HAS3) to inhibit its synthesis or enzymatic degradation of pre-existing hyaluronan using nanoparticle-delivered hyaluronidase. While systemically administered PEGPH20 failed to improve overall survival when combined with chemotherapy in the Phase III HALO-301 pancreatic cancer trial (Hakim et al., 2019), preclinical studies continue to explore its potential as a radiosensitizer in hyaluronan-high tumors. PEGPH20 degrades hyaluronan, reducing interstitial fluid pressure and decompressing collapsed tumor blood vessels. This improves tumor perfusion and oxygenation, thereby overcoming hypoxia-mediated radioresistance and enhancing radiation-induced DNA damage (Saida et al., 2022). The aim is not simply to “break down” the matrix, but to selectively disrupt tumor-supportive ECM functions while maintaining structural homeostasis within normal tissues (Desai et al., 2025).

Inhibition of mechanotransduction pathways: This approach focuses on the biomechanical characteristics of the tumor stroma rather than its composition. The core mechanistic focus is the LOX-mediated collagen cross-linking axis. Members of the LOX and LOX-like (LOXL) family catalyze the oxidative deamination of lysine and hydroxylysine residues in collagen and elastin to produce reactive aldehydes that spontaneously form covalent intra- and intermolecular cross-links. This cross-linking process stiffens the ECM, enhances integrin-mediated mechanosignaling, and promotes tumor invasion. Therapeutic intervention occurs via small-molecule LOX inhibitors (e.g., the pan-LOX inhibitor PXS-5505), LOX-neutralizing antibodies, or nanoparticle-mediated LOXL2 silencing. One important observation made in recent works is that such inhibition of LOX not only decreases collagen levels but also fundamentally alters collagen fiber architecture, reducing fiber width, length, and straightness, which reduces tissue stiffness and disrupts mechanotransduction pathways. Moreover, LOX inhibition is stage-dependent: early-acting intervention remodels the matrix, while late-stage disease with pre-existing cross-linked collagen is largely refractory to such treatments (Desai et al., 2025; Chen W. et al., 2025).

Combination Therapies: This approach aims to block the functional receptors through which tumor and stromal cells physically interact and respond to the ECM. Integrin targeting, specifically αvβ3, αvβ5, and α5β1 isoforms, employs monoclonal antibodies (e.g., cliengitide) or RGD-mimetic peptides to block bidirectional mechanosignaling between ECM ligands and the intracellular cytoskeleton, leading to reduced tumor adhesion, migration, and survival signaling. CD44 targeting disrupts the interaction between hyaluronan and its principal receptor, which is implicated in the acquisition and maintenance of cancer stem cell-like phenotypes, chemoresistance, and peritoneal dissemination. MMP inhibition, being problematic due to off-target toxicity, is being revisited with more selective approaches, including MMP-2/MMP-9-specific inhibitors and nanoparticle-based delivery systems that localize inhibition to the tumor microenvironment. The therapeutic rationale is that blocking the receptors that read ECM signals may be more feasible and less toxic than attempting to normalize the ECM itself. These strategies are not mutually exclusive; rather, they are increasingly combined with ICIs (e.g., anti-PD-1/PD-L1) or chemotherapy (e.g., paclitaxel, gemcitabine), where ECM-receptor blockade sensitizes otherwise resistant tumors by disrupting adhesion-mediated survival pathways and facilitating immune cell infiltration (Mayasin et al., 2024).

Challenges in clinical translation

Despite decades of intensive preclinical research and a compelling biological rationale, no matrix-targeting agent has yet achieved mainstream clinical implementation for cancer therapy (Winkler et al., 2020). This translational gap stems from interconnected biological, pharmacological, and technical challenges.

Complexity and functional duality of the matrisome

The matrisome is defined as the ensemble of ECM proteins and associated factors that represents a dynamic and often contradictory target for therapeutic intervention.

• Tumor-suppressor/tumor-promoter duality: Individual matrisome components exert opposing functions at different stages of cancer progression. For instance, collagen III acts as a tumor suppressor by preserving tissue architecture and restricting early tumor growth, whereas collagen I promotes invasion and metastasis through integrin-dependent mechanotransduction and fiber alignment that guides cancer cell migration (Winkler et al., 2020). Similarly, thrombospondin-1 (TSP-1) exhibits dual roles: it suppresses angiogenesis and tumor growth in early-stage disease but promotes metastasis and immune evasion in advanced tumors via TGF-β activation and recruitment of immunosuppressive cells (Huang et al., 2017). The proteoglycan decorin initially serves as a tumor suppressor by sequestering TGF-β and inhibiting receptor tyrosine kinases; however, its fragmentation in later stages generates bioactive fragments that enhance tumor cell survival and chemoresistance (Neill et al., 2012). This functional duality within the matrisome demands careful consideration of timing and context in therapeutic design, as indiscriminate targeting risks disrupting tumor-suppressive functions while attempting to inhibit pro-tumorigenic roles.
• Immune-modulatory ambiguity: Matrisome degradation generates bioactive fragments called matrikines that exert complex and often opposing effects on immune responses. Hyaluronan-derived matrikines exhibit context-dependent immunomodulation: low molecular weight hyaluronan (LMW-HA, <200 kDa) acts as a damage-associated molecular pattern (DAMP), activating TLR2/4 on immune cells to promote inflammation, whereas high molecular weight hyaluronan (HMW-HA, >1,000 kDa) is immunosuppressive, forming a barrier that excludes T cells and promotes regulatory T cell accumulation (Berdiaki et al., 2023). Collagen-derived matrikines also display duality: endostatin (a fragment of collagen XVIII) suppresses angiogenesis and may enhance anti-tumor immunity by alleviating hypoxia, while tumstatin (a fragment of collagen IV) can promote macrophage polarization toward pro-tumorigenic phenotypes (Ricard-Blum and Salza, 2021). Other collagen-derived matrikines include arresten (collagen IV), canstatin (collagen IV), and restin (collagen XV), each with distinct effects on angiogenesis and immune modulation. Laminin-derived peptides similarly exhibit opposite functions, some activate dendritic cells and T cells, while others such as LG3 (a laminin α-chain fragment) facilitate myeloid-derived suppressor cell (MDSC) recruitment (Okazaki et al., 2002). This immune-modulatory ambiguity within the matrisome poses inherent therapeutic risks: ECM degradation may inadvertently release immunosuppressive fragments that counteract concurrent immunotherapy or, on the other hand, trigger excessive inflammation that promotes tumor-promoting wound-healing responses. Therefore, matrisome-targeted therapies require highly specific, context-aware approaches rather than broad degradation strategies.

The EPR paradox: murine models versus human disease

The enhanced permeability and retention (EPR) effect has been regarded as the cornerstone of nanomedicine for tumor targeting. However, the translational failure of EPR-dependent nanotherapeutics has exposed fundamental discrepancies between preclinical models and human disease, as critically analyzed by Kim et al. (2023).

• Quantitative distinctions: In murine studies, subcutaneous tumors can represent up to 10% of total body weight, for example, CT26 colon carcinoma models in BALB/c mice or 4T1 breast cancer models provided an artificial reservoir that filtered a large portion of the injected nanoparticle dose. The equivalent ratio in a 70 kg human would yield a basketball-sized tumor (∼7 kg), a clinical rarity that fundamentally distorts pharmacokinetic predictions (Kim et al., 2023).
• Vascular and interstitial variances: Human cancers exhibit greater heterogeneity and fewer endothelial fenestrations compared to murine xenografts such as A549 lung cancer models in nude mice. Additionally, the dense human ECM results in elevated interstitial fluid pressure (IFP), which reverses the convective gradient, causing nanoparticles to exit tumors via slow diffusion rather than being retained through efficient vascular extravasation (Kim et al., 2023).
• The size dilemma: While murine studies suggest optimal nanoparticle uptake in the 100–200 nm range, demonstrated in models like B16F10 melanoma in C57BL/6 mice, human tumors with denser ECM and higher IFP favor much smaller particles (<50 nm) that can diffuse through the more restricted interstitial pores, characteristic of desmoplastic human stroma (Kim et al., 2023).

Delivery barriers within the tumor microenvironment

Even the dense cross-linked ECM itself presents a formidable barrier to the delivery of ECM-targeting therapeutics, an access paradox that remains inadequately addressed.

• Interstitial transport limitations: Excessive collagen deposition reduces interstitial pore size. For macromolecular therapeutics such as monoclonal antibodies, antibody-drug conjugates, or nanocarriers, the collagen network forms an almost impenetrable steric barrier (Winkler et al., 2020).
• Vascular compression and hypoperfusion: Solid stress from proliferating cancer cells and expanding ECM compresses intratumoral blood vessels, decreasing functional vascular density. This limits the systemic delivery of nanotherapeutics as well as the infiltration of endogenous immune effector cells (Martin et al., 2019).
• CAF-mediated resilience after ECM perturbation: CAFs respond to mechanical or enzymatic ECM disruption by upregulating pro-fibrotic signaling pathways, including transforming growth factor-beta (TGF-β) and platelet-derived growth factor (PDGF). These factors drive rapid synthesis of replacement matrix proteins such as collagen and fibronectin, leading to reactive fibrosis that quickly reconstitutes, or even aggravates the original stromal barrier. This compensatory mechanism creates a therapeutic paradox: while initial ECM degradation improves drug penetration, the subsequent fibrotic rebound may necessitate repeated or continuous treatment, which paradoxically increases the risk of off-target toxicity due to prolonged exposure of healthy tissues to therapeutic agents (Mayasin et al., 2024)

Toxicity and selectivity

The classical failure of first-generation matrix-targeting agents was mostly related to mechanism-based toxicity due to low target selectivity (Mayasin et al., 2024).

• MMP inhibitor failures: Early MMP inhibitors (marimastat, batimastat) non-selectively targeted multiple MMP family members crucial for normal tissue homeostasis. The resulting musculoskeletal syndrome, painful joint inflammation and fibrosis, was dose-limiting and halted clinical development (Mayasin et al., 2024).
• LOX inhibition challenges: LOX isoforms contribute to wound healing, vascular integrity, and normal tissue remodeling. Tumor-selective LOX inhibition without damaging homeostatic processes requires delivery systems capable of intra-tumoral localization that current technologies do not yet reliably provide (Winkler et al., 2020).
• Systemic off-target effects: Systemically administered ECM-degrading enzymes risk proteolytic damage to healthy tissues, particularly at sites of physiological ECM remodeling such as skin, bone, and healing wounds. Nanocarrier encapsulation mitigates this risk, but it does not eliminate it altogether (Mayasin et al., 2024).

Regulatory and development hurdles

ECM-targeting nanotherapeutics fall into a regulatory gray area, as these multifunctional platforms often combine characteristics of drugs, biologics, and medical devices within a single product. The activity of medical devices involves the components that perform their main function through physical, mechanical, or structural, rather than pharmacological mechanisms, such as nanoparticles that physically disrupt ECM architecture via photothermal heating, magnetic nanoparticles that mechanically remodel collagen fibers under external guidance, and scaffolds, like nanostructured scaffolds, that act as barriers or depots for sustained drug retention in the tumor stroma (Yadav et al., 2025). The medical device part results from physical mechanisms including photothermal ablation (using gold nanorods, Liu et al., 2022), magnetic collagen reorganization (Zimina et al., 2024), or ultrasound-induced cavitation (Huang et al., 2022). Multifunctional nanocarriers (such as lipid-coated silica nanoparticles co-loaded with doxorubicin, collagenase, and imiquimod) require independent safety and pharmacokinetic evaluation for each constituent followed by demonstration of a synergistic benefit, rapidly increasing cost and development schedules (Chen M. H. et al., 2024; Yadav et al., 2025). Manufacturing replicability is still difficult, as small differences in ligand density or enzyme loading have an important influence on their effectiveness (Desai, 2023). Chronic toxicity profiles of non-biodegradable inorganic nanoparticles remain poorly characterized. Gold nanoparticles, for instance, can persist in liver and spleen for months following intravenous administration, with unclear long-term immunological consequences (Cucci et al., 2019). Furthermore, certain rigid, high-aspect-ratio carbon nanotubes have been shown to induce granuloma formation and mesothelioma-like lesions in rodent models, raising concerns similar to those associated with asbestos exposure due to their biopersistence and fibrous geometry (Poland et al. 2008). As such, regulatory authorities are now increasingly demanding extensive biodistribution and chronic toxicity studies costing millions, a process that may prove impractical in academia (FDA, 2023). Furthermore, clinical translation is further hampered by a shortage of validated companion diagnostics for patient stratification, as evidenced by the PEGPH20 phase III failure (Van Cutsem et al., 2020). These complex and multifactorial barriers require cooperation among academia, industry, and the regulatory community to develop specific directions on how to build the ECM-targeted nanotherapeutics.

Case studies and clinical insights

The translational landscape of ECM-targeted cancer therapies has evolved considerably, moving beyond the broad-spectrum MMP inhibitor failures of the past toward highly selective, mechanistically rationalized approaches, now advancing through preclinical and early-phase clinical development (Prinomastat Study, 2024). The niclosamide-based ECM remodeling platform Penetrium™, formulated as polymeric micelles composed of poly(ethylene glycol)-block-poly(lactic acid) (PEG-PLA) encapsulating niclosamide with surface functionalization using fibroblast activation protein (FAP)-targeting peptides, has displayed impressive preclinical efficacy. Studies demonstrated 48.3% tumor reduction in combination with anti-PD-1 therapy, 85.8% metastasis suppression with paclitaxel, and cessation of lung metastasis when combined with bevacizumab in murine models of TNBC (4T1 tumors in BALB/c mice) and metastatic lung cancer (LLC models in C57BL/6 mice). These findings were further validated in spontaneous canine mammary tumors, which closely mirror human disease pathophysiology (Hyundai Bioscience, 2025a; Hyundai Bioscience, 2025b; Hyundai ADM Bio, 2025). Simultaneously, integrin-targeting strategies have gained momentum: an α5β1-function-blocking antibody that inhibits fibronectin fibrillogenesis has been shown to enhance CD8⁺ T cell infiltration and complement PD-L1 blockade. Additionally, CRB-601, an integrin αvβ8-blocking antibody that inhibits latent TGFβ activation in the ECM, has undergone first-in-human phase 1/2 testing with excellent safety and no dose-limiting toxicities in advanced solid tumors (Khan et al., 2025; CRB-601 Study, 2025). The field of selective enzyme inhibition is being revisited, as demonstrated by GS-5745, an MMP-9 specific antibody evaluated in biomarker-driven phase 1 trials targeting bevacizumab-resistant glioblastoma (GS-5745 Study, 2018). In particular, local sustained delivery systems such as the GBP hydrogel have shown promise. This system consists of a gelatin-based matrix crosslinked with MMP-cleavable peptides and sequentially loaded with pirfenidone followed by paclitaxel, enabling controlled ECM priming prior to chemotherapy release. The approach achieved superior tumor suppression in oral squamous cell carcinoma (OSCC) patient-derived xenografts (PDX) established in NSG mice as well as pancreatic PDX models (derived from human tumor specimens implanted in nude mice) following a single intratumoral injection (Sequential drug release system, 2025).

In addition to direct anti-tumor benefits, ECM constructs such as HA35 (35 kDa hyaluronan fragment) are now being used as therapeutically novel fragments for radiotherapy-related pain relief in colorectal cancer, demonstrating the expanding clinical application of ECM-derived molecules (HA35 Study, 2024). Taken together, these varied case studies demonstrate the maturation of the field from non-specific matrix degradation toward precise stromal targeting. Integrin-mediated immune modulation is exemplified by the α5β1-function-blocking antibody that enhances CD8⁺ T cell infiltration and complements PD-L1 blockade (Khan et al., 2025). Localized matrix remodeling is achieved through GBP hydrogel systems for sequential drug release in OSCC and pancreatic cancer models (Sequential drug release system, 2025). Finally, supportive care applications are illustrated by HA35, a 35 kDa hyaluronan fragment used for radiotherapy-related pain relief in colorectal cancer (HA35 Study, 2024). Collectively, these approaches underscore a paradigm shift informed by mechanistic biomarkers, rational combination design, and lessons from previous failures (Winkler et al., 2020; Mayasin et al., 2024).

Preclinical models for ECM-targeted nanomedicine

The development of preclinical models targeting the ECM for nanomedicine is vital in the process of cancer therapy development since they allow to overcome the barrier between labs and clinics. The models concentrate on the TME and especially on ECM, which is a major factor in drug delivery and the success of drugs. The subsequent points present the main features of such preclinical models (Table 1). The use of 3-dimensional tumor spheroids and patient-derived organoids, integrated with decellularized or synthetic ECM hydrogels, makes it feasible to conduct high-throughput screening of nanoparticle penetration, and cytotoxicity to diffusion-limited, matrix-rich microenvironments, in a way that remains faithful to the patient-specific stromal architecture (Millard et al., 2017; Sensi et al., 2023). Tumor microenvironment-based on-chip platforms allow dynamic interstitial flow, vascular perfusion, immune cell recruitment control, and provide real-time visualization of extravasation of nanoparticles and ECM penetration under physiologically relevant mechanical forces (Giannitelli et al., 2024; Rahman, 2024). Biomimetic hydrogels serve as tunable in vitro platforms to investigate how specific ECM properties influence nanoparticle behavior and tumor invasion. For example, hydrogels embedded with aligned collagen fibers mimic the oriented matrix architecture found at tumor-stromal boundaries, enabling studies of contact-guided cancer cell migration. Hydrogels with controlled cross-linking density (e.g., using varying concentrations of genipin or glutaraldehyde) allow systematic evaluation of matrix stiffness effects on nanoparticle diffusion. Additionally, hydrogels functionalized with variable ligand densities (such as RGD peptides) enable quantitative assessment of integrin-mediated cell adhesion and mechanotransduction. These engineered platforms provide critical insights into the biophysical barriers that regulate therapeutic delivery and metastatic progression (Ahn et al., 2025; Cassani et al., 2024). The chick embryo chorioallantoic membrane model offers an inexpensive, vascularized platform for intermediate-throughput screening of nanoparticle accumulation in ECM-rich micrometastases. This model has demonstrated superior predictive value for chemoresistance compared to conventional monolayer cultures (Guller et al., 2021). Genetically modified mouse models like the KPC pancreatic cancer model are capable of the exact comparison to human desmoplasia, elevated interstitial fluid pressure, and co-evolution of immune-stroma, making these models a good approach for validation of ECM-degrading NPs as well as to the development of LOX inhibitors, but because of their low levels of throughput and high costs, their utility is limited in the early stage (Cassani et al., 2024; Yadav et al., 2025). Humanized ECM-containing patient-derived xenografts from co-implantation with human mesenchymal stem cells or cancer-associated fibroblasts represent the current gold standard for the successful validation of those constructs at the late-stage of preclinical validation after mimicking patient-specific matrix composition and CAF phenotypes (Ahn et al., 2025; Cirigliano and Fine, 2024). Non-mammalian animal models, particularly Galleria mellonella larvae and zebrafish xenografts, are deemed more representative and inexpensive medium-throughput platforms for toxicity screening and real-time imaging for NPs and ECM interaction (Yadav et al., 2025; Cassani et al., 2024). Although there is no one best model that fully models the complexity of human tumor ECM, these platforms should start by conducting high-throughput spheroid screening and proceed towards microphysiological mechanistic studies to GEMM/PDX validation (Cassani et al., 2024; Giannitelli et al., 2024; Yadav et al., 2025), which is the most viable approach to advance ECM-targeted nanotherapeutics to the clinic.

TABLE 1. Preclinical model for ECM targeting.

Ongoing clinical investigations

Clinical studies using ECM-targeted nanomedicine have advanced from proof-of-concept research to biomarker-guided trials. The case studies section showcased completed trials with discontinued prinomastat Phase III, completed GS-5745 glioblastoma trial, and CRB-601 Phase 1 data. Current focus is on next-generation platforms to tackle shortcomings from previous approaches (Table 2). Penetrium™, a tumor-agnostic small molecule that reprograms stromal cells to reduce pathological matrix synthesis, is currently in open-label Phase 1 trials for patients with terminal breast and lung cancer (Hyundai ADM Bio, 2026). This is the first sponsor-supported ECM normalization clinical trial conducted outside pancreatic cancer. Elastin-like polypeptide (ELP) nanocarriers, derived from human tropoelastin, exhibit temperature-dependent self-assembly and can passively target ECM-rich inflammatory microenvironments due to their small size and prolonged circulation. These carriers demonstrated sustained symptomatic relief in a Phase 1 osteoarthritis trial and are currently under Phase 2 evaluation. Their repurposing for oncology is being explored based on prolonged intratumoral retention observed in preclinical pancreatic and breast cancer models, where ELP-based delivery enhanced the accumulation of co-administered chemotherapeutics (Dunshee et al., 2021). HA35, a 35 kDa low molecular weight hyaluronan fragment derived from enzymatic digestion of high molecular weight hyaluronan, preserves binding affinity for CD44 and TRPV1 receptors. It functions as an analgesic through TRPV1 modulation, providing relief from radiotherapy-associated pain. A Phase 2 trial is currently recruiting patients with colorectal cancer to evaluate its efficacy in this supportive care setting (HA35 Study, 2024). [¹⁷⁷Lu]Lu-PSMA-617 radioligand therapy, currently approved for prostate cancer, is now being studied in ECM-rich, PSMA-low tumors. It exerts dual cytotoxicity on PSMA-expressing tumor cells and surrounding ECM-embedded stromal cells, exerting its effects not on the matrix per se but on ECM-embedded elements of the cells (Novartis, 2023). These current trials include predictive biomarkers, resistance-based rational combinations, and patient selection based on ECM profiles, essential advancements not included in conventional trials.

TABLE 2. Clinical translation of ECM-Targeted or ECM-Inspired platforms.

Future directions and emerging paradigms

The combination of nanotherapeutics with ECM-targeting strategies signifies a new era in cancer therapy, where future progress will depend not on single advances but on the intelligent fusion of nanobiotechnology with immunology, biomaterials science, and personalized medicine. Several emerging trends will shape the next-generation of ECM-based nanotherapeutics.

AI and computational models for ECM signature prediction

Artificial intelligence is poised to revolutionize ECM biomarker discovery. Deep learning architectures such as DeepECMp, which utilize feed-forward neural networks, achieve 83.6% accuracy in detecting ECM proteins, significantly outperforming traditional approaches (Ghafoor and Nguyen, 2021). Similarly, ECM-LSE employs deep latent space encoding to identify functionally relevant ECM features that may inform patient stratification and therapeutic targeting (Al-Saggaf et al., 2021). These computational tools hold promise for accelerating the translation of ECM signatures into clinical practice.

Next-generation combination strategies

Beyond current immunomodulatory approaches, future nanotherapeutics will increasingly integrate ECM-targeting with emerging modalities. For example, the convergence of ECM modulation with personalized neoantigen vaccines or adoptive cell therapies may further enhance anti-tumor immunity. Additionally, advances in materials science will enable the design of smarter, multi-responsive nanocarriers capable of sensing and adapting to dynamic changes in the tumor ECM.

Clinical translation and adoption

Successful implementation of ECM-targeted nanomedicines will require close collaboration between academia, industry, and regulatory agencies to establish standardized guidelines for safety assessment and quality control. The development of companion diagnostics capable of real-time ECM profiling will be essential for patient selection and treatment monitoring. Furthermore, educational initiatives to familiarize clinicians with matrix-centric oncology concepts will facilitate broader clinical adoption.

In summary, the future of ECM-targeted nanomedicine lies in leveraging advanced computational models, rational combination designs, and precision biomarker-based strategies to achieve durable therapeutic outcomes.

Conclusion

The ECM represents a dynamic and underexploited frontier in oncology. Despite its well-established role in tumour growth, metastasis, and therapeutic resistance, current cancer treatments often overlook the ECM as a direct therapeutic target. The integration of nanobiotechnology into ECM research has opened transformative prospects for both diagnostics and therapy, offering precision tools such as ECM-targeted nanoparticles, enzyme-responsive drug carriers, and nanosensors for real-time biomarker detection. However, several knowledge gaps persist: the heterogeneity of ECM composition across tumour types, limited understanding of spatiotemporal ECM remodeling, and challenges in achieving site-specific delivery without systemic toxicity. Addressing these limitations demands deeper molecular profiling of ECM signatures and development of multifunctional nanoplatforms capable of both sensing and intervention. Looking ahead, future research should focus on the convergence of ECM-targeted nanotherapeutics with immunotherapy and precision medicine, enabling synergistic modulation of tumour–immune dynamics. Personalized ECM maps, aided by AI-driven nanodiagnostics, could guide adaptive treatment strategies tailored to each tumour’s matrix profile. Ultimately, by bridging these translational gaps, ECM-centered nanomedicine could redefine the next-generation of cancer therapeutics, transforming the tumour microenvironment from a barrier into a therapeutic aid.

Generative AI statement

The author(s) declared that generative AI was used in the creation of this manuscript. Images are created using the Unique commands through Google Gemini Pro.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Footnotes

Edited by: Jaison Jeevanandam, National Institute of Mental Health, Czechia

Reviewed by: Dharmendra Kumar Yadav, Gachon University, Republic of Korea

Sutapa Biswas Majee, NSHM Knowledge Campus, Kolkata, India