Bacteriophages could help oncolytic viruses penetrate cold tumors

A 2026 review argues that bacteriophages could help overcome two major obstacles facing oncolytic virotherapy: poor delivery into solid tumors and the immunosuppressive conditions that prevent a lasting antitumor immune response. Rather than treating phages as a replacement for oncolytic viruses, the authors describe two experimental strategies. One merges components of the two virus classes into replication-deficient chimeric delivery vehicles; the other transfers selected phage-inspired functions into replication-capable oncolytic viruses. Most supporting evidence remains preclinical, and the review identifies stability, immunogenicity, manufacturing and regulatory challenges that must be resolved before these approaches can be evaluated broadly in patients.

Oncolytic virotherapy uses naturally occurring or engineered viruses to replicate in and destroy tumor cells while largely sparing healthy tissue. Its defining feature is a dual mechanism: viral replication directly lyses malignant cells, while the resulting immunogenic cell death can initiate an adaptive immune response that may persist after the virus is no longer present. Viral rupture releases damage-associated molecular patterns, including ATP, calreticulin, mitochondrial DNA and HMGB1, alongside tumor-associated antigens. These signals recruit and activate antigen-presenting cells, which move to lymph nodes and stimulate tumor-specific CD8+ and CD4+ T cells. The review notes that Talimogene laherparepvec, a modified HSV-1 therapy expressing granulocyte-macrophage colony-stimulating factor, is the only FDA-approved replicating oncolytic virus described in the article. In a neoadjuvant monotherapy study of difficult-to-resect cutaneous basal cell carcinoma, its objective response rate was 55.6%, and 50% of patients reached surgical resection without reconstructive surgery. These results demonstrate clinical feasibility for selected cancers, but they do not establish broad performance in immune-suppressed solid tumors.

Many solid tumors present the biological conditions of a “cold” tumor: inflammation and immune-cell infiltration are limited, antigen presentation is defective, and physical and signaling barriers impede therapeutic entry. A dense desmoplastic extracellular matrix is a central obstacle. Cross-linked collagen, hyaluronic acid and proteoglycans restrict the movement of viruses, drugs and immune cells, while cancer-associated fibroblasts add growth factors and immunosuppressive cytokines. Matrix stiffness can also alter biomechanical signaling through YAP/TAZ, and the matrix may provide a scaffold for migration and metastasis. Oncolytic vaccinia virus engineered to express human hyaluronidase-1 improved viral spread, immune-cell infiltration and chemotherapy penetration in mouse subcutaneous tumor models, illustrating how matrix degradation can help. But the enzyme’s effects are context dependent: HYAL1 has also been associated with tumor progression in some cancers and with brain metastasis in triple-negative breast cancer. A proposed phage-derived alternative is the bacterial enzyme HylP, a hyaluronate lyase that cleaves hyaluronan by beta-elimination rather than hydrolysis and produces unsaturated fragments described in the review as lacking the pro-inflammatory and tumorigenic signaling attributed to fragments generated by mammalian HYAL1.

Low antigenicity creates a second major problem. Tumor cells may reduce or lose major histocompatibility complex classes I and II, weakening activation of cytotoxic and helper T cells. Such defects can arise through mutations, epigenetic changes, unstable messenger RNA, increased protein degradation or disrupted endosomal trafficking. A low tumor mutational burden also means fewer neoantigens available for immune recognition. Strategies include expressing immunostimulatory cytokines, removing viral genes that suppress MHC-I and presenting tumor antigens, but oncolytic virotherapy depends heavily on functional antigen presentation to generate durable adaptive immunity. The tumor microenvironment also contains suppressive cell populations. Regulatory T cells release interleukin-10 and transforming growth factor beta; M2-polarized tumor-associated macrophages support growth, angiogenesis and metastasis; and heterogeneous myeloid-derived suppressor cells can promote angiogenesis and further immune suppression. Experimental approaches to deplete regulatory T cells, repolarize macrophages or inhibit myeloid-derived suppressor cells have produced partial preclinical results, but off-target effects, systemic toxicity and limited monotherapy efficacy constrain translation. T cells that do enter a tumor may also encounter checkpoint signals such as PD-1 and CTLA-4, although systemic checkpoint inhibition can produce autoimmune-like adverse effects and tumors may develop resistance over time.

Bacteriophages are viruses that infect bacteria and do not target human cells. The review highlights several characteristics that may be transferable to cancer therapy. Some phages cross biological and endothelial barriers through receptor-mediated uptake and transcytosis, allowing movement across several cell layers. Filamentous M13 phages can enter epithelial and endothelial cells through several endocytic routes, while their high aspect ratio permits a snake-like movement, or reptation, through pores in dense matrices resembling a tumor extracellular matrix. Phages also carry enzymes that degrade bacterial biofilms and capsules. T4 offers a large genome of approximately 170 kilobases and uses an ATP-powered motor to pack DNA tightly into its capsid. The resulting pressure can place DNA in a dense, liquid-crystalline state that resists nuclease degradation. By contrast, adeno-associated viruses deliver about 4.7 kilobases, and the larger genomes of herpesvirus and vaccinia virus can be constrained by pathogenicity or cytotoxicity. These features suggest that phage components could improve payload capacity, environmental stability and access to dense tumor tissue.

Phage display provides another potentially useful feature. Target peptides can be presented in dense, repetitive arrays on major coat proteins such as M13 pVIII, with roughly 2,700 copies per phage, or on minor proteins such as pIII, which can accommodate larger peptide fusions despite having only about five copies. T4 capsids can also display hundreds of selected coat proteins. The RGD-4C ligand, for example, recognizes alpha-v-beta-3 and alpha-v-beta-5 integrins associated with cancer cells and tumor vasculature. Repetitive tumor-peptide display may promote antigen uptake and cross-presentation by antigen-presenting cells, potentially redirecting part of the antiviral response toward tumor targets. This strategy still depends on functional antigen-presentation machinery, and repeated exposure to common phages can produce neutralizing antibodies. Phage immunomodulation is similarly bidirectional. Unmethylated CpG motifs in phage DNA can stimulate Toll-like receptor 9 and act as an adjuvant, and the newly proposed Bacteriophage Risk Index is intended to quantify such immunogenic potential from motif frequency, spacing and sequence context. Other phage capsid proteins have shown anti-inflammatory activity in experimental systems, including suppression of nuclear factor-kappa B signaling and reduced expression of interleukin-1 beta, interleukin-6 and, in some cases, tumor necrosis factor alpha.

The most established hybrid described in the review is AAVP, an adeno-associated virus-phage vector created by inserting AAV cis elements and transgenes into a filamentous M13-derived capsid displaying RGD4C. Early work reported ligand-dependent tumor transduction and liver de-targeting, while later experiments investigated endosomal escape, proteasomal degradation and extracellular-matrix penetration. AAVP derivatives have been used to induce tumor-cell apoptosis, alter tumor metabolism and hypoxia, deliver CRISPR-Cas9 components, and increase susceptibility to drugs. Most of these findings came from cell cultures or murine models. Their principal drawback is that the hybrid is replication deficient, so it does not reproduce the sustained oncolysis and immunogenic cell death of a lytic oncolytic virus. Its persistence in the tumor may also be shortened by anti-phage antibodies. The authors therefore see AAVP primarily as a complementary platform for delivering payloads or conditioning antitumor immunity, potentially before or alongside a replicating oncolytic virus, rather than as a substitute for oncolysis.

A modular alternative aims to retain viral replication while importing individual phage-derived functions. Directly transferring bacterial genes into mammalian viruses is difficult because the two systems differ in codon usage, transcription, polyadenylation, ribosomes, introns and exons, and adding many modifications could impair replication or stability. The review instead favors specialized oncolytic viruses carrying a limited number of relevant modifications, with phage mechanisms recreated through engineering where direct gene transfer is incompatible. One example is PeptiCRAd, in which selected tumor peptides with positively charged polylysine tails are electrostatically attached to negatively charged oncolytic adenovirus or vaccinia capsids. A reported 15-minute incubation allows the peptide layer to be exchanged without rebuilding the virus. The approach can prime tumor-reactive T cells rapidly and may redirect attention from viral antigens toward tumor antigens, but the coating is not genetically amplified as the virus replicates and can dissociate over time. Other modular approaches include biomimetic membranes, which can extend circulation and improve tumor distribution, and mesenchymal-stem-cell-derived vesicle coatings, which may shield viral particles from neutralizing antibodies and help them enter tumor cells. Both require complex manufacturing and can produce variable yields or disrupt viral infection.

The review identifies a specific molecular target for further development: a mammalian-encoded form of a prokaryotic hyaluronate lyase. Creating one would require solutions to transcriptional, post-translational and immunogenicity problems, with codon optimization, protein redesign and AI-assisted engineering proposed as possible tools. The broader experimental priority is not a universal oncolytic virus that corrects every feature of a tumor microenvironment, but a set of specialized products matched to tumor type, route of administration and treatment stage. Dense matrices in pancreatic cancer, breast cancer and glioblastoma may call for matrix-degrading activity or peptides directed at cancer-associated fibroblasts. Intratumoral treatment may benefit from phage-derived tail-fiber proteins that anchor the virus to tumor receptors, whereas intravenous treatment must confront neutralizing antibodies, liver sequestration by Kupffer cells and limited circulation time. Early treatment may prioritize direct killing, antigen release and immune priming; later rounds may require longer persistence and access to disseminated or residual disease.

Despite the appeal of combining phage delivery systems with oncolytic virotherapy, the evidence base remains dominated by laboratory experiments, murine tumor models and small early clinical studies. Phage hybrids can be unstable, anti-phage immunity can reduce vector half-life, and manufacturing a replication-capable construct that preserves both stability and infectivity is technically difficult. Regulation, production scale, long-term safety and the variability of human tumors are additional obstacles. The review’s central proposal is therefore a research direction rather than an established treatment: specialized replication-capable oncolytic viruses could be engineered to borrow phage traits such as barrier penetration, payload delivery, selective peptide display and controlled immune stimulation, while phage-inspired coatings or companion vectors could help modify the tumor microenvironment. Clinical testing will need to establish whether these combinations actually improve viral infiltration, immune activation and treatment outcomes in people, rather than merely overcoming selected barriers in experimental models.

Subject of Research: Review of bacteriophage-derived strategies for improving oncolytic virotherapy in immunosuppressive solid tumors

Article Title: Unlocking oncolytic virotherapy potential: leveraging bacteriophages to conquer immunosuppression

Article References: Boulfrad, B., & Larbi, A. (2026). Unlocking oncolytic virotherapy potential: leveraging bacteriophages to conquer immunosuppression. Frontiers in Microbiology, 17, Article 1957551. https://doi.org/10.3389/fmicb.2026.1957551

Image Credits: AI Generated

DOI: 10.3389/fmicb.2026.1957551

Keywords: oncolytic virotherapy, bacteriophages, cold tumors, tumor microenvironment, viral therapy, cancer immunotherapy, viral delivery, immune suppression, extracellular matrix, phage display, Unlocking, oncolytic