Prostate cancer (PCa) is the most common malignancy in the male genitourinary system globally, with approximately 1.6 million new cases annually [1]. Despite multimodal therapies, including radical prostatectomy, radiotherapy, androgen deprivation therapy (ADT) and immunotherapy, the prognosis for locally advanced or metastatic PCa remains suboptimal. Once progressing to castration-resistant prostate cancer (CRPC), 60 %–70 % of patients will develop distant metastases within three years, and the five-year survival rate collapses to only 10 %–15 % [2]. Immune checkpoint inhibitors (ICIs) are the primary immunotherapeutic agents, which restore the body's immune response to tumors by blocking inhibitory signaling pathways in T cells. However, various ICIs have failed to significantly extend the overall survival of CRPC patients [3]. The anti-PD-L1 agent pembrolizumab yields an objective response rate of just 5 % in PD-L1-positive metastatic CRPC patients [4]. While combination therapy with the anti-CTLA-4 agent nivolumab and the anti-PD-1 agent ipilimumab is more effective than monotherapy, the objective response rate remains modest at 25 % [5]. This is primarily attributed to PCa being a classic “cold tumor”, characterized by sparse T-cell infiltration and a profoundly immunosuppressive microenvironment. In PCa, immunosuppressive cells like tumor associated macrophages (TAMs), regulatory T cells (Tregs), and myeloid-derived suppressor cells (MDSCs) promote tumor growth and angiogenesis by secreting IL-10, TGF-β, and metabolic by-products such as kynurenine and lactate, thereby silencing T-cell activation and promoting exhaustion [6]. Cancer-associated fibroblasts (CAFs) further reinforce this immunological desert by depositing extracellular matrix barriers that physically impede T-cell trafficking [7].

Beyond intrinsic tumor mechanisms, the commensal microbiota critically modulates antitumor immunity. Microbial metabolites and pathogen-associated molecular patterns (PAMPs) engage Toll-like receptors (TLRs) on antigen-presenting cells (APCs), orchestrating both tumorigenesis and therapeutic responsiveness across urological malignancies [[8], [9], [10], [11]]. Microbial immunotherapy, as one of the pioneering immunotherapeutic approaches, has been revitalized by advances in genetic engineering and nanotechnology. The application of bacteria, including Clostridium, Bifidobacterium, and Salmonella, and their derivatives in cancer therapy has a history exceeding a century [12,13]. Rich in PAMPs like lipopolysaccharides, flagellin, and viral nucleic acids, microbiota can be recognized by TLRs on APCs and activate them. Upon activation, APCs boost CD8+ T cell activation and tumor cell killing by expressing costimulatory molecules, secreting cytokines, and presenting tumor antigens. Escherichia coli Nissle 1917 (EcN) is one of the most commonly used bacteria in the field of cancer immunotherapy due to its probiotic properties, lack of expression of intestinal or cytotoxic toxins, and tumor targeting capabilities. EcN can reverse the “cold” characteristics of various tumors, promote tumor regression, and induce immune memory, thereby preventing tumor recurrence [13,14]. However, bacterial immunotherapy confronts challenges, including the risk of systemic infection, excessive immune responses, potential genotoxicity, and high costs associated with production and transportation, which collectively restrict its clinical application [15]. Achieving a balance between the efficacy and safety of bacterial therapy is crucial for its application in cancer treatment.

Bacterial outer-membrane vesicles (OMVs) offer a safer, scalable alternative. These 20–250 nm nanoparticles inherit bacterial PAMPs yet lack replicative competence, achieving high biocompatibility and stability [16]. In the past decade, the clinical potential of OMVs has been extensively investigated across diverse fields, including vaccine development, drug delivery systems, cancer immunotherapy and diagnostics [17]. Notably, vaccines formulated with OMVs derived from Neisseria meningitidis have garnered regulatory approval for clinical application [18]. In 2017, Kim et al. first reported the use of OMVs in tumor therapy, revealing that OMVs from Escherichia coli could induce IFN-γ production in immune cells, thereby mediating antitumor effects [19]. Building on these findings, subsequent research has focused on engineering OMVs to augment their immunostimulatory properties while mitigating their cytotoxicity. For instance, Li et al. successfully engineered OMVs to express PD-1 molecules on their surface. The interaction between PD-1 on OMVs and PD-L1 on cancer cells effectively alleviates T cell immunosuppression, enabling T cells within the tumor microenvironment (TME) to sustain their antitumor activity [20]. On this basis, Sun et al. conjugated two nanobodies targeting the CD47/SIRPα and PD-1/PD-L1 pathways simultaneously to the surface of OMVs. The modified OMVs were shown to preserve their inherent immunostimulatory properties while concurrently alleviating the immunosuppressive effects on T cells and macrophages within the TME, thereby augmenting their antitumor efficacy [21].

Beyond surface modification, leveraging OMVs as drug delivery vehicles has emerged as a prominent research focus. Multiple studies have indicated that OMVs themselves lack direct cytotoxicity against tumor cells. Therefore, their antitumor efficacy can be enhanced by loading chemotherapeutic drugs or other agents with direct tumor-killing properties into OMVs. Oncolytic virotherapy represents a dual-mechanism therapeutic strategy that combines selective tumor lysis with systemic immune activation, offering a novel solution to overcome immune evasion in PCa. Among these, OH2 is a genetically engineered oncolytic virus based on the type II herpes simplex virus (HSV-2) HG52 strain. It has attenuated neurovirulence by deleting the neurovirulence genes ICP34.5 and ICP47 and carries the human granulocyte-macrophage colony-stimulating factor (GM-CSF) gene to enhance APCs recruitment. The antitumor effects of OH2 have been previously reported in various cancer cell lines, including colorectal cancer, breast cancer, hepatocellular carcinoma, gastric adenocarcinoma, and glioblastoma [22,23]. Relevant clinical trials have been successively launched, demonstrating durable antitumor effects in patients with melanoma and metastatic esophageal cancer, rectal cancer and sarcoma [[24], [25], [26]]. However, there are significant limitations to the administration of oncolytic viruses (OVs). When administered intravenously, approximately 90 % of the viral particles are cleared by neutralizing antibodies in the bloodstream within 24 h. As a result, the treatment with OVs is currently primarily confined to direct intratumoral injection, which makes it difficult to reach lesions located deep within the body. Mehrnoush suggested that extracellular vesicles (EVs) can enhance the targeted delivery of OVs to tumors and protect them from the host immune system's clearance after systemic administration. Moreover, EVs can also deliver OVs to metastatic tumors and trigger a strong antitumor immune response [27]. Ban et al. demonstrated that the encapsulation of oncolytic adenoviruses within engineered OMVs significantly augments viral accumulation within tumor tissues, consequently bolstering the immunotherapeutic efficacy of the oncolytic adenoviruses [11]. In this study, we utilized EcN derived OMVs to encapsulate oncolytic virus OH2 (OH2@EcN-OMV) for the treatment of PCa. Our findings indicated that OH2@EcN-OMVs exhibit precise tumor tissue targeting ability and could lead to direct tumor cell lysis. Additionally, it facilitated the polarization of tumor-associated macrophages (TAMs) towards the M1 phenotype, enhancing their antigen-presenting capabilities and antitumor functions. Consequently, this approach significantly amplifies the immunotherapeutic efficacy of OH2.