Malignant tumours are among the most common causes of disease-related deaths worldwide and pose serious threats to human quality of life [1]. The idea of manipulating the immune system to combat neoplastic disease has been explored for the past 50 years and has recently yielded unprecedented clinical success [2,3]. Among immunotherapies, cytotoxic lymphocytes, including cytotoxic T lymphocytes (CTLs) and natural killer (NK) cells, are a central focus for engaging the immune system in the fight against tumours [4]. Cytotoxic lymphocytes are able to directly lyse malignant cells through multiple mechanisms, particularly the granule-mediated cytotoxic pathway involving the release of lytic granules containing perforin (PFN) and granzymes [5,6]. Granzymes are a family of serine proteases known for their ability to induce cell apoptosis through cleaving a variety of target proteins, among which GrB has been most extensively characterized [7]. PFN is a pore-forming protein that forms ring-shaped lesions that facilitate the uptake of granzymes [8]. Once target cells are recognized, GrB is secreted and enters the cytoplasm of the cells with the help of PFN, where it induces cell apoptosis through either direct cleavage of caspases or the activation of the mitochondrial pathway of apoptosis [9,10].

Although the prospect of using cytotoxic lymphocyte-based immunotherapy in treating haematological malignancies is impressive, the current clinical efficacy in solid tumours has been much less rewarding due to the poor penetration and targeting ability of immune cells into solid tumours [11,12]. Furthermore, the proliferation of infiltrated cytotoxic T cells within tumours can still be inhibited or even inactivated by an immunosuppressive tumour microenvironment [13]. In light of these challenges, a more effective approach would be to directly deliver PFN and GrB into the tumour microenvironment to induce cell apoptosis. Studies related to the targeted delivery of GrB to tumour cells have been conducted [14]. Research on the tumour-targeted delivery of GrB was first reported by the Rosenblum team, who employed the ScFvMEL/GrB fusion protein to target the Gp240 antigen of melanoma to inhibit tumour proliferation [15]. That study provided a proof of concept that GrB-based antitumour agents could inhibit tumour growth in vivo. Although these fusion proteins can retain the targeting ability of GrB, their antitumour effects are still suboptimal owing to their poor tumour-homing ability and susceptibility to lysosomal degradation [16]. Recent studies have demonstrated the ability of these nanomaterials to successfully transport GrB into tumour cells, bypassing capture by the reticuloendothelial system (RES) [17]. For example, Gu et al. designed a fused nanoplatform derived from exosomes and liposomes in which GrB and siRNA targeting Sb9 were encapsulated. This biomimetic nanoplatform enabled the effective transfer of GrB to tumour tissues and displayed satisfactory therapeutic efficacy [18]. Previously, we successfully developed a GrB delivery platform containing a GrB-TAT core and an HA/PMPC shell for active tumour targeting and effective tumour growth inhibition [19].

Despite these favourable therapeutic outcomes, the application of GrB as a single antitumour modality faces challenges, including unsatisfactory cellular uptake and a high therapeutic dose range [20]. In fact, GrB-induced apoptosis highly relies on the coexistence of PFN because GrB cannot be internalized into target cells without the membrane pore formation mediated by PFN [21]. It is therefore hypothesized that the simultaneous delivery of GrB and PFN would exert a more profound tumour suppressive effect than delivery of GrB alone. We previously observed a synergistic effect of GrB and PFN on inhibiting cell proliferation compared to that of GrB alone at the cellular level [19]. However, the simultaneous delivery of exogenous PFN and GrB into tumours awaits further exploration.

Based on the above discussion, in the present study, a nanocapsule-based codelivery system for GrB and PFN was developed. As shown in Scheme 1, GrB and PFN were first conjugated through MMP-2-responsive peptides. Next, methacrylic acid (MMA) and zwitterionic 2-methacryloyloxyethyl phosphorylcholine (MPC) were added to the GrB-PFN conjugate solution as monomers, and a biodegradable crosslinker was added to grow a thin polymer network around each protein conjugate through in situ polymerization, resulting in the formation of dual protein nanocapsules, denoted nGPMs. Notably, the crosslinker employed contains an MMP-2-responsive peptide sequence that is bioresponsive and biodegradable in the tumour microenvironment (TME) [22]. The polymer shells can stabilize the two protein cargoes and retain their activity [23]. Compared with other positively charged nanoparticles, such nanocapsules have longer circulation times and greater accumulation within tumour sites because of the low fouling surface area of MPC [24]. Once nanocapsules reach the TME, the peptide bonds of the peptide crosslinkers between GrB and PFN are cleaved by MMP-2 proteases overexpressed in the TME [25], resulting in the release of GrB and PFN and their subsequent entry into tumour cells to execute cellular apoptosis.