The tumor ECM is far more than a passive wall. In healthy tissue, its composition of collagens, proteoglycans, glycosaminoglycans, laminins, and fibronectin is finely tuned to maintain homeostasis. In cancer, the matrix becomes a dynamic scaffold that promotes immune evasion, metastasis, and therapeutic resistance. Dense ECM obstructs the trafficking of drugs and immune cells, drives hypoxia, and fuels immunosuppressive signaling within the tumor microenvironment. Beyond the ECM, many tumors wrap themselves in a glycocalyx enriched in mucins and hypersialylated glycans, which physically blocks receptor-ligand interactions and engages inhibitory Siglec receptors on immune cells. The review argues that editing these barriers enzymatically could transform immunologically cold tumors into hot ones, but it also cautions that such remodeling is a double-edged sword requiring precise spatial and temporal control.

Among the most striking preclinical advances is the engineering of immune cells to produce heparanase, the only mammalian enzyme capable of cleaving heparan sulfate chains on heparan sulfate proteoglycans. Long-term ex vivo expansion of T cells downregulates heparanase expression, partly through p53-mediated repression, leaving CAR-T cells poorly equipped to burrow through matrix-rich tumors. When researchers co-expressed heparanase in CAR-T cells, the engineered cells showed robust tumor-eliminating activity in the presence of ECM, enhanced intratumoral accumulation, superior tumor control, and prolonged survival in animal models, without abnormal accumulation in the lung or liver. A parallel approach anchored an enzymatically active heparanase construct to the surface of NK cells, which restored penetration into tumor spheroids and improved tumor control in vivo, while the membrane-anchored design may help limit systemic enzymatic dissemination.

Hyaluronidase strategies have traveled furthest toward the clinic, though not without painful lessons. Hyaluronan accumulates densely in many tumors, including HER2-positive breast cancers and EGFR-positive head and neck cancers, where it limits antibody access and disrupts the immune synapses needed for antibody-dependent cellular cytotoxicity. The PEGylated enzyme PEGPH20 reached clinical trials, but the SWOG S1313 study found that combining it with modified FOLFIRINOX increased gastrointestinal and thromboembolic toxicity and was associated with inferior survival, while the HALO 109-301 trial raised response rates in pancreatic cancer without improving overall survival. These failures have pushed the field toward localized delivery: CAR-T cells engineered to secrete Fc-fused PH20 showed greater tumor regression in gastric cancer xenografts, dual-engineered cells co-expressing IL-7 and PH20 improved infiltration and survival in hepatocellular carcinoma models, and bioorthogonal click chemistry has been used to decorate CAR-T cells with hyaluronidase plus a pH-responsive anti-PD-L1 antibody.

Remarkably, hyaluronidase activity may do more than clear a physical path. Degradation fragments of high-molecular-weight hyaluronan activated CD103-positive dendritic cells through Toll-like receptor 4 signaling, enhancing CD8-positive T cell responses. Small extracellular vesicles carrying PH20 increased CD8-positive T cell infiltration and suppressed tumor growth in melanoma and breast cancer models, and when combined with anti-PD-L1 therapy, produced durable tumor suppression even in an autochthonous MMTV-PyMT model. A stimulus-responsive nanocarrier delivering hyaluronidase, interleukin-12, and anti-PD-L1 achieved sequential, tumor-triggered release in hepatocellular carcinoma models, depleting hyaluronan, reducing collagen deposition, and extending survival without inducing metastasis. Intratumoral hyaluronidase also boosted a nanovaccine regimen, markedly improving CD8-positive T cell infiltration and interferon-gamma production in melanoma models.

Collagen, the most abundant ECM component, presents a subtler challenge because its role is context-dependent. In pancreatic ductal adenocarcinoma, cancer cells produce an oncogenic type I collagen homotrimer that drives growth through integrin signaling, distinct from the heterotrimer made by cancer-associated fibroblasts. Because collagenase cannot distinguish collagen by origin, researchers engineered Escherichia coli Nissle 1917 bacteria, which preferentially colonize hypoxic tumor regions, to carry ROS-responsive nanocages releasing collagenase and anti-PD-L1 locally. This depleted oncogenic collagen, attenuated integrin-FAK signaling, increased CD8-positive T cell infiltration, and extended survival in orthotopic pancreatic cancer models. In another approach, CAR-T cells were fitted with collagenase nanogel backpacks decorated with a CXCR4 antagonist peptide, which simultaneously degraded matrix and freed T cells from CXCL12-mediated trapping, achieving complete tumor regression in some mice. Yet the review warns that indiscriminate collagen depletion can backfire: in pancreatic cancer models, depleting myofibroblast-derived collagen accelerated tumor progression and immunosuppression.

Perhaps the most unexpected recruit is nattokinase, a serine protease from Bacillus subtilis best known for its fibrinolytic activity in the Japanese fermented food natto. Repurposed for oncology, intratumoral nattokinase degraded fibronectin, improved tissue perfusion, reduced hypoxia, and indirectly suppressed cancer-associated fibroblast activity. In breast tumor xenografts, nattokinase pretreatment produced a sixfold increase in intratumoral CAR-T cell accumulation and superior therapeutic outcomes. However, the enzyme was administered locally, and its systemic pharmacokinetics, long-term safety, and allergenic potential as a bacterially derived protein remain insufficiently characterized.

The glycocalyx itself has become a target through sialidases and mucinases. Hypersialylation cloaks tumor cells, blocking NKG2D activating receptors and engaging inhibitory Siglecs. Antibody-sialidase conjugates, including trastuzumab fused to Salmonella Typhimurium sialidase (T-Sia 2), enhanced NK cell-mediated killing and delayed tumor growth in trastuzumab-resistant breast cancer models, while an anti-PD-1-sialidase conjugate preferentially desialylated PD-1-positive T cells and improved tumor control in melanoma. Sialidase has also been fused to bispecific T cell engagers and engineered into CAR-T cells secreting Clostridium perfringens neuraminidase, which preserved a naive-like phenotype and improved persistence. Encouragingly, E-602, a first-in-class engineered human sialidase, demonstrated tumor desialylation and early clinical activity in the Phase 1/2 GLIMMER-01 trial when combined with the anti-PD-1 antibody cemiplimab, marking the furthest clinical advance of any glycocalyx-editing strategy.

Mucin-targeting enzymes complete the toolkit. StcE, a mucin-selective metalloprotease from E. coli O157:H7, trims glycocalyx thickness by roughly ten nanometers enough to alter cancer-cell susceptibility to immune attack, and restored NK cell killing of mucin-rich tumors. Because high-dose systemic StcE caused platelet depletion and hemorrhagic toxicity in mice, an attenuated variant was fused to a HER2-targeting nanobody, reducing tumor burden and metastatic spread in mammary tumor models. Researchers also developed a modular leucine-zipper Zip-NK platform for controllable surface display of StcE and sialidase on NK cells, and nanovesicles co-displaying StcE with a CD47-targeting nanobody enhanced macrophage phagocytosis and tumor control while avoiding the toxicities of free enzyme.

The review closes with sobering translational caveats. Excessive matrix degradation may promote metastasis, off-target enzyme activity can damage normal tissues, and matrix reconstitution is rapid, with hyaluronan-rich matrices rebuilding within 24 hours after enzyme removal in vitro. Anti-drug antibodies have already limited repeated dosing of at least one sialidase construct, and microbial enzymes carry inherent immunogenicity risks that PEGylation and protein engineering have not fully solved. Biomarker-guided patient selection, combining substrate abundance with immune-exclusion profiling and spatial imaging, will be essential, since hyaluronan-high status alone failed to predict benefit in pancreatic cancer trials. Most strategies remain preclinical, but with only a handful of microbial enzymes explored so far, computational protein design and screening of the vast microbial world could rapidly expand the barrier-editing toolkit, potentially turning the tumor’s own fortress into its greatest vulnerability.

Subject of Research: Enzyme-mediated remodeling of the tumor extracellular matrix and glycocalyx to enhance immunotherapy in solid tumors

Article Title: Enzyme‐mediated remodeling of the extracellular matrix and glycocalyx to enhance immunotherapy in solid tumors

Article References: Mirmohammadsadegh, N., & Amin, M. (2026). Enzyme‐mediated remodeling of the extracellular matrix and glycocalyx to enhance immunotherapy in solid tumors. Bioengineering & Translational Medicine, Article e70180. https://doi.org/10.1002/btm2.70180

Image Credits: AI Generated

DOI: 10.1002/btm2.70180

Keywords: extracellular matrix, glycocalyx, immunotherapy, CAR-T cells, heparanase, hyaluronidase, collagenase, sialidase, mucinase, tumor microenvironment, drug delivery, cancer