Cancer immunotherapy has fundamentally transformed cancer care for a wide range of tumor types, shifting the treatment paradigm towards the hope of achieving long-term control of tumors that are naïve to conventional therapies such as surgery, chemotherapy, and radiotherapy (1, 2). The current landscape of immunotherapeutics encompasses immune checkpoint inhibitors (ICIs), adoptive cell therapies, oncolytic viruses (OVs), and cancer vaccines. Despite these advancements, the clinical benefits of immunotherapy have been observed in only a small subset of cancer patients. The efficacy of immunotherapy is strongly influenced by the presence of a high infiltration of lymphocytes and a reduced abundance of immunosuppressive cells (3). T cells are the primary anti-tumor effector lymphocytes and the focus of the majority of immunotherapeutic strategies. The density and distribution of T cell infiltration serve as a critical indicator of the response to immunotherapy interventions. For instance, the effectiveness of ICIs is better in hot tumors with a high lymphocyte penetration than in immunologically inert or cold tumors that lack infiltrating lymphocytes (4, 5).

The pattern of T cell infiltration within tumors is heavily shaped by the tumor extracellular matrix (ECM), which is notably denser and stiffer than a normal ECM. Immunologically cold tumors, which lack significant immune cell infiltration, are commonly characterized by this dense and stiff ECM, a condition known as desmoplasia (6), which is a result of tumor-induced ECM remodeling (7). The altered expression of matrix proteins, including collagens, laminins, and fibronectin, as well as glycans such as glycosaminoglycans and proteoglycans, changes the mechanical properties of the tumor ECM, leading to increased ECM density, stiffness, and interstitial fluid pressure (IFP) (8). This aberrant tumor ECM creates physical barriers that hinder the penetration of immune cells, drugs, and chimeric antigen receptor T (CAR-T) cells, thereby adversely affecting the efficacy of cancer treatment (9, 10).

Oncolytic virus immunotherapy (OVI) harnesses viruses to selectively infect and lyse tumor cells (11). The ensuing virus-mediated immunogenic cell death (ICD) ( Figure 1A ) releases tumor antigens (TAs) and damage-associated molecular patterns (DAMPs) that are captured by antigen-presenting dendritic cells (APCs) inside the tumor and presented to naïve T cells in the lymph nodes, priming them to attack tumor cells (12–17). The primed and activated CD4+ and CD8+ T cells are trafficked into the tumor to exert their effector functions, thus making immunologically cold tumors hot (18–21). The dying tumor cells release TAs and DAMPs, triggering activation and infiltration of T cells leading to an immune activation cycle referred to as the cancer immunity cycle (CIC) ( Figure 1B ). OVI-mediated conversion of cold tumors to hot provides an opportunity for the development of combination therapies that involve novel immunotherapeutic approaches such as immune checkpoint blockade (ICB), adoptive T-cell therapy, and cancer vaccines (22–27). Accumulating evidence implicates the limited intratumoral penetration and poor distribution of oncolytic viruses as two key factors to the poor response to OVI (28, 29). Aberrant expressions of matrix proteins (collagens, laminins, fibronectin, and elastin) and glycans (glycosaminoglycans (GAGs) and proteoglycans (PGs)) by cancer-associated fibroblasts (CAFs) alter the mechanical properties of the tumor ECM, leading to an increase in ECM desmoplasia and an elevated intratumoral IFP (8), creating physical barriers that impede the penetration and distribution of OVs. Overcoming these barriers is a crucial step in unlocking the full potential of OVI in modulating the tumor immune microenvironment (TIME). This review discusses the latest advancements in strategies to enhance the intratumoral penetration of OVs, thereby facilitating the infiltration of tumor-infiltrating lymphocytes (TILs) into cold tumors via the CIC process. We also explore clinical trials and challenges associated with translating these strategies into clinical practice to improve patient outcomes.

In solid tumors, the tumor parenchyma is usually surrounded by an aberrantly organized ECM made of overexpressed components that contain variable proportions of fibrous proteins (collagen, fibronectin, laminin, and elastin), as well as, GAGs (hyaluronic acid (HA)), and PGs (chondroitin sulfate and heparan sulfate) (8, 30). Elevated collagen synthesis, fibrillogenesis, and crosslinking in desmoplastic tumors such as those of pancreatic adenocarcinoma are associated with an increase in ECM density, alignment, and stiffness (31).

CAFs are activated stromal cells arising from diverse cell types (tissue-resident fibroblasts, mesenchymal stem cells, endothelial cells, pericytes, epithelial cells, and bone marrow-derived stem cells) that are recruited during the genesis of tumor stroma (32, 33). They are responsible for the secretion of aberrant components of the tumor ECM (34). CAFs are a heterogeneous group of cells with different secretory and functional characteristics and are operationally classified into three subtypes including myofibroblast (myCAF), inflammatory (iCAF), and antigen-presenting CAF (apCAF) (35). The phenotypic conversion of stromal cells to CAFs is a result of signals from both tumor cells and recruited normal cells. For instance, tumor cell-derived transforming growth factor-β (TGF-β) induces the conversion of normal gastric resident fibroblasts to CAFs (36). Similarly, tumor-derived exosomes (TDEs) induce phenotypic conversion of vascular endothelial cells to CAFs (37). The myCAFs are the most common CAF sub-type and their abundance correlates with the physical properties of the tumor ECM. High intratumoral collagen deposition by myCAFs is responsible for the increased tumor ECM density (38), while myCAFs-exerted active forces on tumor cells and tumor ECM orchestrate the increase in ECM stiffness (39–41).

The tumor microenvironment (TME) exerts several factors that augment collagen secretion, fibrillogenesis, and crosslinking to further increase ECM desmoplasia. For instance, the hypoxic TME triggers an increase in the expression of lysyl oxidase (LOX) enzymes resulting in increased collagen crosslinking and stiffness (42, 43). Likewise, several cytokines, chemokines, and growth factors that are abundant in the TME exert pro-fibrotic effects to increase the secretion of ECM components by CAFs, tumor cells, and tumor-associated macrophages (TAMs). For example, high levels of TGF-β increase collagen secretion and remodeling by myCAFs and induce the M2 polarization of TAMs (44). The M2-polarized TAMs further promote ECM desmoplasia through the production of reactive nitrogen species and inducible nitric oxide synthase which exerts profibrotic effects on collagen-secreting cells (45). In addition, several pro-inflammatory cytokines that are abundant in the TME such as IL-1β and TNF-α promote tumor ECM desmoplasia by promoting the activation and secretory functions of ECM-secreting cells such as stellate cells in pancreatic ductal adenocarcinoma (PDAC) (46, 47).

In addition to collagen secretion, myCAFs secrete large amounts of HA, which contributes to altered physical properties of the tumor ECM (48, 49) and leads to an elevated intratumoral IFP due to the high hydrophilicity from its negative charge (50–52). Furthermore, the elevated HA in the tumor ECM is associated with the generation of abnormal tumor blood vessels with decreased patency and permeability (53). A combination of elevated IFP and hyperdense and stiffened ECM results in the collapse of intratumoral blood and lymphatic vessels (52). These features are associated with increased resistance to treatment by physically impeding the intratumoral penetration and distribution of drugs, therapeutic nanoparticles, OVs, antibodies, and immune cells (8–10) ( Figure 2 ). Further barriers to penetration of the tumor parenchyma result from high tumor cell density within tumor nests and their tight adherence due to overexpression of cell-to-cell adhesion molecules (56, 57).

Substantial evidence suggests that excessive ECM desmoplasia is a significant factor that impairs the efficacy of OVI by hindering the intratumoral penetration of the viruses. Evidence from in vitro experiments of OV oncolysis has shown a decrease in oncolysis efficiency in three-dimensional (3D) spheroid culture conditions compared to two-dimensional (2D) monolayer culture conditions. For instance, oncolytic herpes simplex virus 1 (HSV-1) infection was highly effective in 2D cultured melanoma cell lines leading to a complete lysis of tumor cells while showing significant impairment in 3D cultures, which is consistent with ECM-mediated impairment of virus penetration and replication (28). Furthermore, the penetration of HSV-1 in human melanoma xenografts was found to be highly size-dependent. This was demonstrated by the similar impedance experienced by nanospheres loaded with quantum dots of comparable size to the virions (150 nm in diameter) and the superior penetration achieved by smaller dextran tracer molecules (20 nm in diameter) (54). Further, the degradation of ECM collagen using collagenase significantly enhanced the intratumoral penetration and distribution of HSV-1 in Mu89 melanoma xenografts (54). This treatment was associated with more robust and sustained tumor regression compared to the administration of HSV-1 alone. Additional evidence highlighting the role of tumor ECM barriers in hindering OV penetration was demonstrated in studies designed to enhance the susceptibility of tumor cells to OV infection. For instance, both rapamycin (Rap) and HA have been utilized to augment the susceptibility of gallbladder carcinoma (GBC) cells to infection by oncolytic myxoma virus (MYXV) (58, 59). However, the combination of MYXV with Rap (MYXV+Rap) significantly enhanced the susceptibility and lysis of GBC cell lines, but this effect was not observed in patient-derived xenografts (PDX) (58). Conversely, the combination of MYXV with HA (MYXV+HA) increased both the susceptibility and oncolytic efficacy in PDX models. This enhancement was attributed to hyaluronan-mediated induction of matrix metalloproteases 2 and 9 (MMP-2 and MMP-9) secretion, which facilitates the degradation of type-IV collagen (58, 60).

Recognition of the tumor ECM’s adverse role in obstructing intratumoral OV penetration has spurred research endeavors aimed at overcoming these barriers to improve the effectiveness of OVI. Advances in our understanding of the structural elements and molecular pathways that regulate tumor desmoplasia have paved the way for the development of strategies targeting specific components of the ECM. These ECM-targeting strategies can be categorically divided into two main types: biochemical modulation and physico-chemical disruption ( Figure 3 ).

Physico-chemical strategies for ECM disruption utilize magnetic, sound, and light energies to physically or chemically disrupt the tumor’s ECM. This section summarizes the applications of these strategies to enhance the intratumoral penetration of OVs, as reported in pre-clinical and clinical studies.

In addition to enhancing the intratumoral penetration and distribution of OVs for an effective oncolysis, ECM-targeted OVs can reprogram TIME to facilitate anti-cancer immune reactions. An important characteristic of OVs is their ability to increase the infiltration of T-cells into tumors, effectively converting immunologically cold tumors into hot (171). This conversion is achieved through OV-mediated tumor cell lysis, which releases TAs and DAMPs. These are then taken up by intratumoral dendritic cells (DCs) for priming of T-cells in tumor-draining lymph nodes. The ability of ECM-targeted OVs to increase T-cell infiltration in cold tumors arises from two primary mechanisms: (i) the enhanced efficiency of tumor cell lysis, leading to the release of TAs and DAMPs, and (ii) the overcoming of barriers imposed by the ECM on T-cell trafficking and infiltration of T-cells into tumors (172).

Several oncolytic adenoviruses engineered to carry a therapeutic transgene for PH20 have shown a high efficacy in improving the infiltration of T cells into cold tumors (68, 71, 173). VCN-01 is an example of an ECM-targeted oncolytic adenovirus that carries a PH20 transgene for intratumoral degradation of hyaluronic acid. In a phase I clinical trial involving patients with advanced or metastatic pancreatic cancer, VCN-01 enhanced the infiltration of CD8+ T cells and the upregulation of indole 2,3-dioxygenase, resulting in clinical response rates of 40 to 45% (173). VCN-01 has been granted orphan drug designation for the treatment of retinoblastoma and PDAC (174). ICOVIR17 is another PH20-expressing oncolytic adenovirus that demonstrated improved tumoral infiltration of CD8+ T-cells in an animal model of glioblastoma (68). ICOVIR17 treatment also increased PD-L1 expression by tumor cells and enhanced the efficacy of PD-1 blockade. Owing to their large genome size and ability to carry multiple therapeutic transgenes (175), oncolytic adenoviruses have been engineered to express both PH20 and immunostimulatory molecules for modulating phenotypes of infiltrating T cells. For instance, intratumoral injection of an oncolytic adenovirus armed with therapeutic transgenes for PH20, IL-12, and sPD1-Fc (KLS3020) resulted in increased tumoral infiltration of CD4+ and CD8+ T cells, intratumoral populations of IFN-γ+ effector CD4+ and CD8+ T cells, and enhanced the Teff/Treg ratios of tumoral TILs in animal models of melanoma and colorectal carcinoma (71). These effects led to the activation of systemic anti-cancer immune reactions, which controlled tumor growth in non-injected lesions. Similar effects were observed with a recombinant pseudorabies virus carrying transgenes for PH20, IL-18, and IFN-γ expressing PRV (rPRV-IL-18-γ-PH20), further demonstrating the efficacy of ECM-targeting OVs in improving T cell infiltration of tumors and inducing robust systemic anti-cancer immune reactions (70).

ECM-targeting OVs that inhibit the synthesis of ECM components can also enhance the intratumoral infiltration of T cells. For instance, an oncolytic herpes simplex virus (HSV-1) expressing murine OX40L (OV-mOX40L) led to increases in tumoral CD4+ and CD8+ TILs in animal models of pancreatic ductal adenocarcinoma (93). The OV-mOX40L treatment also increased the expression of effector CD8+ T cells expression of IFNγ and GZMB, and decreased the expression of inhibitory receptors, such as PD-1 and LAG-3 (93). Similarly, an increase in intratumoral infiltration of CD8+ T-cells was observed with 0X40L expressing OV (LOAd713) in animal models of pancreatic cancer (95). The addition of immunomodulatory genes further enhances the phenotype of infiltrating T cells by increasing the number of effector TILs. For instance, an oncolytic adenovirus carrying a therapeutic gene for the anti-fibrotic hormone relaxin (RLX) and genes for immunomodulatory cytokines IL-12 and GM-CSF (oAd/IL12/GM-RLX) increased the intratumoral infiltration of activated CD4+, IFN-γ+ and CD8+, IFN-γ+ T cells in a Syrian hamster model of pancreatic cancer (77). Combination therapy of oAd/IL12/GM-RLX with αPD1 increased the proportions of intratumoral CD8+, IFN-γ+ and CD8+, PFRN+ T cells, leading to a durable suppression of tumor growth.

In addition to converting cold tumors to hot by increasing their T cell infiltration, ECM-targeted OVs contribute to a global improvement in TIME, addressing immunosuppression and enhancing anti-cancer immune responses. These effects include overcoming the exhaustion of TILs (93), reducing the presence of immunosuppressive cells such as Tregs (71, 93, 176), M2-like TAMs (68, 93, 176), and myeloid-derived suppressor cells (MDSCs) (95, 176), as well as DC maturation and antigen presentation (176). These collective effects position ECM-targeted OVs as an effective tool to augment anti-cancer immune reactions, thereby enhancing the efficacy of immunotherapy in refractory tumors.

Due to their dual capabilities of inducing ICD and enhancing OV penetration, physico-chemical strategies for tumor ECM disruption have the potential to profoundly modulate TIME and enhance anti-cancer immune reactions. For instance, in mouse models of glioblastoma and malignant meningioma, PDT using a KillerRed-expressing engineered oncolytic HSV increased the infiltration of a diverse range of immune cells, including lymphocytes, NK cells, monocytes, macrophages, and neutrophils (166). Similarly, magnetically-steered MNP-encapsulated OVs triggered an increase in tumoral infiltration of activated immune cells, such as neutrophils, cytotoxic T-cells, NK cells, and macrophages, as well as a decrease in tumoral infiltration of Tregs and B cells (154).

Several ECM-targeting strategies have shown promising outcomes in pre-clinical studies, as outlined in Tables 1 , 2 . However, the majority of these strategies have not yet been evaluated in clinical trials. Clinicaltrials.gov lists twelve registered clinical trials involving three ECM-targeted OVs: VCN-01, LOAd703, and DNX-2440. These trials are predominantly early-phase (phase I or phase I/II) and are focused on assessing the safety and optimal dosage of the OVs when used as a single agent or in combination with other treatments, including ICB, CAR-T therapy, and cytotoxic chemotherapy ( Table 3 ). A significant proportion of these trials employ combination therapy, with only three using OVI monotherapy. Four trials have reached completion, three are still actively recruiting participants, and two have been terminated due to resource limitations at the treating centers (NCT03555149) and stock issues (NCT03714334).

Two completed phase I trials evaluated the safety and preliminary efficacy of VCN-01in patients with pancreatic and other advanced solid malignancies. In the first trial, VCN-01 was administered intratumorally to pancreatic cancer patients in combination with either gemcitabine or gemcitabine plus nab-paclitaxel. The treatment was well-tolerated and led to disease stabilization (177). The dose-limiting side effects included asthenia, grade three serum transaminasemia, and a fatal pancreatic fistula in one patient. The efficacy analysis revealed that all treated lesions remained stable, with new lesions appearing in five patients after four months, in one patient after eight months, and in another patient after 31 months. In the second trial, VCN-01 was administered as a single intravenous infusion of VCN-01 to 16 patients with locally advanced or metastatic solid tumors, or as part of two combination schedules with gemcitabine and nab-paclitaxel to 26 patients with pancreatic cancer (84). The treatment demonstrated an acceptable safety profile with an encouraging efficacy. The dose-limiting side effects included one patient with grade 4 raised aspartate aminotransferase (AST), one patient with grade 4 febrile neutropenia, and one patient with fatal thrombocytopenia and enterocolitis. In patients with metastatic PDAC, the combination therapy of VCN-01 and nab-paclitaxel and gemcitabine resulted in an overall treatment response rate of 50%, with 36% of patients showing stable disease for more than 12 months. Post-hoc analyses showed a median progression-free survival of 7.2 months and a median overall survival of 13.4 months. These outcomes are superior to those reported with combination therapy of oncolytic reovirus (Pelareorep) in combination with gemcitabine which resulted in an objective response of 3% and a median overall survival rate of 10.2 months (178).

Another two completed trials are phase I/II studies assessing the safety and efficacy of a TMZ-CD40L-expressing oncolytic adenovirus (LOAd703) in pancreatic cancer and melanoma patients. In one of these trials (179), LOAd703 was administered intratumorally in combination with standard nab-paclitaxel/gemcitabine (nPG) chemotherapy to patients with advanced pancreatic cancer. The combination therapy was found to be tolerable and feasible with observed side effects attributed to LOAd703 being low-grade and short-lived, except for one patient with grade 3 transaminasemia. LOAd703 treatment resulted in an overall response rate (ORR) and disease control rate (DCR) of 44% and 94% respectively, with corresponding immune activation. A follow-up study by the same research group is focused on evaluating a combination therapy of intratumoral LOAd703 and intravenous anti-PD-L1 monoclonal antibody atezolizumab in melanoma patients. The results of this study have not yet been published. The results of these completed phase I/II trials indicate the feasibility and safety of ECM-targeted OVs and suggest potential clinical benefits in the treatment of advanced and treatment-resistant solid tumors such as pancreatic cancer and melanoma.

Despite promising results in pre-clinical studies of ECM-targeted OVs, several limitations may hinder their clinical translation. In most animal studies, intratumoral injection is the most commonly used delivery method to avoid recognition and clearance of the viruses by the host immune system. However, intratumoral administration is not feasible in patients with metastatic tumors due to the presence of multiple lesions, some of which may not be detectable. This assertion is supported by the results of a phase I trial in metastatic PDAC which showed disease progression in non-injected, distant metastatic sites (177). To overcome this limitation, OVs can be encapsulated using suitable tumor-tropic delivery vehicles such as extracellular vesicles, cells, or polymers (180–182). Furthermore, most of the engineered ECM-targeted OVs are directed to a single ECM component, and thus can only be effective in tumors with dysregulation of that particular component. For instance, collagenase-expressing OVs such as KMMP9 and JD0G will only be effective in tumors with aberrant collagen expression and similarly, hyaluronidase-expressing OVs such as VCN-01 and ICOVIR17 will only be effective in tumors with overexpression of hyaluronic acid. However, the composition of the tumor ECM is highly heterogeneous resulting in significant differences in relative quantities of different ECM components within tumors of similar histological types (183, 184). This necessitates the pre-treatment proteomic characterization of the tumor ECM to determine dysregulated ECM components for personalized treatment approaches. Alternatively, ECM-targeted OVs can be engineered to target multiple ECM components that are ubiquitously overexpressed in tumors. Another important consideration for ECM-targeted virotherapy is the possibility of enhancing tumor cell dissemination by breaking the ECM barriers. This is due to the role played by the tumor ECM in tumor cell confinement, proliferation and migration (185). The release of the ECM barrier may thus increase the number of circulating tumor cells to enhance metastatic dissemination of the treated tumors. Moreover, the physicochemical strategies of ECM disruption face several drawbacks including possibility of normal tissue injury due to inability to specifically target tumor cells. Also, the rapid and poorly controlled disruption of the ECM barriers by focused ultrasound, MNPs or PDT may enhance the dissemination of previously confined tumor cells.

Addressing these limitations will enhance the clinical translation of ECM-targeted virotherapy to enhance the penetration and distribution of OVs in tumors. The resulting improvement of tumor cell lysis and release of DAMPs will enhance immune cell infiltration into tumors, converting cold tumors into hot. This will open new possibilities of combination immunotherapies of OVs and ICIs, cancer vaccines or cellular therapies.

Due to vast heterogeneity in the composition of tumor ECM both between and within tumor subtypes, personalized treatment approaches will enhance the future utilization of ECM-targeted OVs. Personalized treatment approaches are made possible due to the rich arsenal of ECM-targeted OVs that are directed to virtually all ECM components including collagen, hyaluronic acid, and proteoglycans. Central to personalized approach in ECM-targeted OVs is the profiling of individual patient’s ECM components to inform personalized selection of ECM-targeted OVs that patient-specific dysregulated ECM components. The characterization of individual patient’s ECM proteome can be achieved by using several assays such as ELISA, immunohistochemistry (IHC) or mass spectrometry-based approaches (186). Of these assays, IHC is the most attractive option due to its ability to provide both quantitative and spatial information about the abundance and localization of ECM components in the tumor stroma (187). Furthermore, IHC is a routine test in majority of oncology hospitals and thus does not require significant capital investments in terms of human resource and equipment. The implementation of personalized approaches will ensure the right ECM-targeted OV for the right patient, which can lead to substantial improvements in the outcomes of treatment.

Oncolytic virus immunotherapy holds great promise in the fight against cancer, offering specificity and the ability to stimulate long-term anti-cancer immune responses. However, several barriers hinder the effectiveness of OVs in infecting, replicating, and lysing tumor cells, including dense and stiff tumor ECM, abnormal vasculature, and elevated intratumoral IFP. To address these challenges, various strategies have been developed, including targeted degradation of ECM components, inhibition of collagen synthesis, modulation of ECM organization, and the use of mechanical disruption techniques. Despite the lag in clinical translation, Phase I clinical trials of VCN-01, LOAd703, and DNX-2440 have demonstrated an acceptable safety profile and early evidence of clinical efficacy in a range of solid tumors. It is important to note that a re-normalization of the ECM by matrix-targeting OVs not only can enhance the intratumoral distribution of other therapeutics, such as chemotherapy drugs, monoclonal antibodies, and CAR T cells, but it can also convert cold tumors into hot tumors, thereby enhancing the efficacy of immunotherapies. Pre-clinical studies have demonstrated the benefits of combination treatments involving ECM-normalizing OVs with immune checkpoint inhibitors, CAR-T cells, TIL therapy, and cancer vaccines. Lastly, physicochemical strategies for disrupting the tumor stroma, such as UTMD, magnetic steering of MNP-decorated OVs, and photodynamic therapy offer alternative approaches to improving intratumoral OV distribution. These approaches have shown promising results in pre-clinical studies and have the potential to reduce treatment-related adverse events. The realization of the potential of OVs as a tool for transforming immunologically cold tumors into hot tumors will enhance the efficacy of immunotherapy and broaden the scope of patients who can benefit from immunotherapy.