Cancer imposes an ever-growing global health challenge, with over 19 million new diagnoses in 2020 and projections exceeding 27.5 million by 2040 [1,2]. Conventional chemotherapy faces major limitations due to non-specific biodistribution, systemic toxicity, and the emergence of resistant clones [3]. In contrast, immunotherapy has transformed oncology by recruiting and amplifying the host's own defenses against cancer [4]. Within this paradigm, bacteria-based therapies have garnered exceptional interest: they combine direct oncolysis with potent immune activation, creating a synergistic assault on tumors [[5], [6], [7]]. The clinical success of Bacillus Calmette–Guérin (BCG) in non–muscle-invasive bladder cancer underscores the promise of live microbial agents [8,9]. Advances in genetic engineering have further expanded this arsenal: talimogene laherparepvec (T-VEC) is now an FDA-approved oncolytic virus for melanoma, and the attenuated Salmonella strain SGN1 is progressing through phase II trials in osteosarcoma patients [[10], [11], [12], [13]].

Bacterial mediated tumor therapy provides a unique solution to the limitations of traditional therapies, such as poor targeting, tumor heterogeneity, and immunosuppressive microenvironment. And with the development of time, from the traditional use of bacterial targeted therapy and drug delivery alone, to now the combination of bacterial attenuation and targeted ability modification, and immune checkpoint inhibitors for joint treatment to enhance the therapeutic effect on tumors. Salmonella, in particular, offers unmatched advantages: as a facultative anaerobe, it selectively colonizes the hypoxic cores of solid tumors, penetrates deep tissue barriers, and is readily programmable via molecular tools [[14], [15], [16], [17], [18], [19], [20], [21], [22], [23], [24], [25], [26], [27], [28], [29], [30]]. Its surface lipopolysaccharides (LPS) and flagellin act as pathogen-associated molecular patterns (PAMPs), igniting both innate and adaptive immunity, reshaping the immunosuppressive tumor microenvironment, and amplifying cytokines such as IL-1β and TNF-α to drive tumor cell death [19]. Despite these strengths, safety concerns persist. Even attenuated strains like S. typhimurium VNP20009—engineered with purI and msbB deletions—elicited adverse events in clinical trials, including thrombocytopenia, anemia, and bacteremia, highlighting the insufficiency of genetic attenuation alone to prevent systemic toxicity [13,20,21,31]. To overcome these challenges, surface functionalization has emerged as a promising strategy. Ligand-decorated Salmonella have demonstrated 2–4-fold greater tumor accumulation and reduced off-target exposure [32]. However, a modular, stimulus-responsive coating capable of dynamically adapting to the physiological environment remains lacking.

Self-assembling peptides, renowned for their biocompatibility and programmable responsiveness to pH, temperature, ionic strength, and redox potential, have shown remarkable utility as “smart” nanocarriers [[33], [34], [35], [36], [37]]. Disulfide-rich peptide hydrogels can encapsulate proteins, cells, and small molecules, releasing them upon.

thiol–disulfide exchange in reductive environments [[37], [38], [39], [40]]. Building on these insights, we developed a glutathione (GSH)-responsive, LPS-targeting self-assembling peptide designed to cloak Salmonella via electrostatic interactions, offering immune shielding without compromising tumor tropism. In this work, we demonstrate that the designed peptide binds LPS on the bacterial surface and self-assemble to form a nanofibrils-rich coating—validated by transmission electron microscopy and confocal imaging. This peptide coating effectively suppresses early inflammatory cytokine release (IL-1β, TNF-α), thereby reducing collateral tissue damage in vivo (Fig. 1). Upon exposure to the reductive conditions and elevated GSH levels within the tumor microenvironment arisen from both intracellular release and extracellular accumulation—the disulfide-linked peptide coating undergoes selective cleavage, thereby disassembling on-site and reactivating the antitumor functions of the coated Salmonella. In vivo, this strategy enhances bacterial colonization, neutrophil infiltration, and tumor regression, all while minimizing off-target toxicity. Our findings establish a self-assembling peptide-based, intelligent coating platform that significantly improves the safety and efficacy of bacteria-mediated cancer therapy, offering promising potential for clinical translation.