Engineered Cancer-Killing Viruses Move From Lab Curiosity to Approved Therapy

Cancer still claims roughly 10 million lives a year worldwide, with nearly 20 million new cases diagnosed annually, and those numbers are projected to rise as populations age and grow. Even as surgery, radiotherapy, chemotherapy, targeted drugs and immunotherapy have improved outcomes for many patients, durable disease control remains out of reach for a large share of people with advanced or metastatic disease. Tumor heterogeneity, therapeutic resistance, immune evasion and treatment-related toxicity continue to cap the long-term effectiveness of existing strategies, and a comprehensive new review in Clinical and Experimental Medicine examines one of the most ambitious answers to that problem: oncolytic virotherapy, the use of engineered viruses that selectively infect, replicate within and destroy malignant cells.

The core idea is deceptively simple. Unlike conventional therapies that target proliferation or specific oncogenic pathways, oncolytic viruses kill cancer cells directly while largely sparing normal tissue, and in doing so they convert a dying tumor into an inflammatory site that trains the immune system. When infected cells lyse, they release tumor-associated antigens together with pathogen- and damage-associated molecular patterns, which activate dendritic cells, promote antigen presentation and ignite both innate and adaptive antitumor responses. Modern oncolytic viruses are therefore described as dual-action therapeutics, combining direct oncolysis with what researchers call an in situ cancer vaccination strategy.

Selectivity rests on several converging mechanisms. Many tumors overexpress receptors that viruses exploit for entry: adenovirus serotype 5 uses the coxsackievirus and adenovirus receptor, while alternative serotypes and chimeric vectors such as Ad5/35 prefer CD46 or desmoglein-2; measles virus targets CD46 and SLAM; herpes simplex virus enters through HVEM and nectin receptors; and coxsackievirus A11 preferentially infects thoracic malignancies through ICAM-1. Entry alone is not enough, however. Cancer cells frequently carry oncogenic abnormalities, in Ras, EGFR, Wnt/beta-catenin, PI3K/AKT signaling or defects in p53 and RB pathways, that create an intracellular environment favorable to viral replication while weakening antiviral defenses. Reovirus is the classic example: in tumors with activated Ras signaling, protein kinase R activity is suppressed and type I interferon responses are blunted, allowing selective viral propagation while healthy tissue retains protection.

The tumor microenvironment itself helps. Hypoxia, abnormal vasculature, nutrient deprivation and impaired immune surveillance within solid tumors favor viral gene expression and spread, in part through hypoxia-inducible factor-1 signaling. Genetic engineering then layers on further control: receptor retargeting redirects viral entry toward tumor-enriched molecules; transcriptional targeting places essential genes under tumor-selective promoters such as hTERT or survivin; replication detargeting deletes virulence genes or adds microRNA response elements to suppress replication in healthy tissue; and carrier-mediated delivery uses mesenchymal stromal cells to shield viruses from immune clearance and home them to tumors. The review cautions that receptor heterogeneity, and the fact that many exploited receptors are overexpressed rather than exclusively expressed in cancer, still limit the therapeutic window.

How the viruses kill matters as much as how many cells they kill. Rather than a single mechanism, different platforms trigger multiple regulated cell death pathways: caspase-dependent apoptosis, inflammatory necroptosis driven by the RIPK1/RIPK3/MLKL axis, gasdermin-mediated pyroptosis, autophagy-dependent cell death and, more tentatively, ferroptosis, which Newcastle disease virus has been shown to induce in glioma models. These pathways rarely operate in isolation; adenoviral infection can induce apoptosis, autophagy, necroptosis and pyroptosis concurrently, and experimental studies show viral replication and lysis continue even when individual pathways are inhibited, a redundancy that helps the viruses cope with heterogeneous tumors. Inflammatory death modes release more antigens and danger signals than silent cytotoxicity, generating stronger immune priming.

That immunological quality is formalized in the concept of immunogenic cell death. Its canonical hallmarks are surface exposure of calreticulin, extracellular ATP, release of HMGB1 and production of type I interferons. Calreticulin acts as an eat-me signal recognized by CD91/LRP1 on dendritic cells; ATP recruits dendritic cells through P2Y2 and P2X7 and activates the NLRP3 inflammasome; HMGB1 binds TLR4 and RAGE to enhance cross-presentation; and type I interferons, triggered via cGAS-STING, RIG-I and related sensors, activate NK cells and support CD8-positive T-cell expansion. The result can convert immunologically cold tumors, which respond poorly to checkpoint blockade, into hot, T-cell-infiltrated lesions, and can generate systemic responses reaching uninjected metastases, the so-called abscopal effect.

Clinical translation has been real but narrow. Talimogene laherparepvec (T-VEC), an HSV-1 engineered with deletions of ICP34.5 and ICP47 plus a GM-CSF transgene, was approved by the FDA in 2015 and later by the EMA for unresectable cutaneous, subcutaneous and nodal lesions in melanoma. In the pivotal phase III OPTiM trial of 436 patients, T-VEC achieved a durable response rate of 16.3 percent versus 2.1 percent for GM-CSF, an overall response rate of 26.4 percent with 10.8 percent complete responses, and median overall survival of 23.3 versus 18.9 months. Real-world cohorts have reported higher response rates, including 67 percent with 35 percent complete responses among 121 routinely treated melanoma patients, most of them previously treated. Other approvals include the adenovirus H101 (Oncorine) in China for nasopharyngeal and head and neck carcinoma in combination with chemotherapy, and the triple-mutated HSV-1 G47Delta (teserpaturev) in Japan for malignant glioma, conditionally approved in 2021 on single-arm phase II survival data. Rigvir, an ECHO-7 virus approved in Latvia in 2004, has never undergone the large randomized trials required by major regulators.

Combination strategies are where much of the current excitement, and disappointment, lies. T-VEC plus pembrolizumab in a phase Ib cohort produced an overall response rate of 61.9 percent, and T-VEC plus ipilimumab in a randomized phase II study improved response rates to 35.7 versus 16.0 percent. Yet a large phase III trial of T-VEC plus pembrolizumab in checkpoint-naive patients failed to significantly improve progression-free or overall survival, although response and durable response rates favored the combination. A neoadjuvant trial of T-VEC plus nivolumab achieved a 74 percent pathological response rate in resectable melanoma. Evidence outside melanoma rests mostly on smaller cohorts or preclinical models, and early-phase work is extending to armed vectors such as T7011, which co-expresses CD19, CCL5, IL-12 and an anti-PD-1 payload to partner with CAR-T cells.

Formidable obstacles remain. Intratumoral injection, the route used by approved agents, only reaches accessible lesions; systemically delivered viruses are neutralized by antibodies, complement and phagocytic clearance, and dense stroma and abnormal vasculature restrict spread. Tumors can acquire resistance through chronic type I interferon activation and IRF7 upregulation, which simultaneously restricts viral replication and suppresses antitumor immunity, a paradox because the same sensing pathways needed for immunogenicity also clear the virus. Manufacturing replication-competent viruses under Good Manufacturing Practice is costly and biologically variable, regulatory frameworks for living therapeutics are still maturing, and no validated predictive biomarkers exist to guide patient selection. The disappointing phase III result of the vaccinia virus Pexa-Vec in hepatocellular carcinoma underlines that engineering alone does not guarantee clinical benefit.

The review’s authors argue that the field’s future lies less in discovering new viruses than in precision engineering and rational combination: multifunctional platforms armed with cytokines, checkpoint blockers or bispecific T-cell engagers; delivery systems using nanoparticles, extracellular vesicles or cell carriers; multidimensional biomarker signatures integrating genomic, immune and imaging data; and better preclinical models such as organoids and humanized mice. With 171 oncolytic virus trials registered worldwide, 62 of them completed, oncolytic virotherapy has moved from a century-old observation that infections occasionally shrank tumors to a small but established class of approved cancer immunotherapies, one whose next chapter will depend on matching viral design to each tumor’s biology rather than on any single breakthrough virus.

Subject of Research: Oncolytic virotherapy as an engineered cancer immunotherapy platform

Article Title: Oncolytic virotherapy for cancer: biological mechanisms, clinical advances, and future perspectives

Article References: Halawa, E. M., AbdelHamid, S. G., & ElHefnawi, M. (2026). Oncolytic virotherapy for cancer: biological mechanisms, clinical advances, and future perspectives. Clinical and Experimental Medicine, 26(1), Article 358. https://doi.org/10.1007/s10238-026-02311-3

Image Credits: AI Generated

DOI: 10.1007/s10238-026-02311-3

Keywords: oncolytic virotherapy, T-VEC, immunogenic cell death, tumor microenvironment, herpes simplex virus, immune checkpoint inhibitors, viral engineering, melanoma, combination therapy, cancer immunotherapy, viral tropism, clinical trials