>90% of glioblastoma (GBM) recurrences occur within a 2–3 cm radius of the resection margin. [1] The probability of this GBM recurrence post-standard treatment is nearly absolute, with an average survival of 6 months for GBM patients with obvious residual lesions post-surgery and even lower in the event of multifocal recurrence. [2] It is unrealistic to eradicate the tumor without residual cells using current imaging techniques. Despite the remarkable success of adjuvant radiotherapy and chemotherapy, residual microscopic lesions in the brain still pose considerable challenges for clinicians to overcome. It can be argued that immunotherapy, which endogenously triggers tumor-specific destruction, may present a potential opportunity among a range of novel therapies.

The programmed death (PD) pathway serves as an endogenous negative feedback loop of T cell activity that is commonly exploited by human tumors, such as GBM, to suppress the antitumor efficacy of infiltrating CD8+ cytotoxic T lymphocytes (CTLs). However, central nervous system (CNS) diseases have been historically limited by the poor blood-brain barrier (BBB) penetration of drugs. [3] The molecular weight of nivolumab was calculated to be 146 kDa, far exceeding the 400 Da threshold for permeation of BBB. [4] This means that PD-1/PD-L1-mediated inhibition generally occurs extrinsically to the tumor, indicating that effector T cells activated by tumor-associated antigens in peripheral lymphoid tissue can only access the tumor microenvironment (TME) via a pre-coated PD-1 antibody mode. [5] As tumors progress, particularly in the case of recurrent GBM, any lymphocytes activated against tumor antigens are assumed to have migrated to the tumor site, making antibody-based drugs intrinsically unable to reach the lymphocytes expressing PD-1 in TME.

We found that immunosuppressive monocytic myeloid-derived suppressor cells (M-MDSCs) were more likely to be recruited by adhesion factors on GBM-associated endothelial cells upregulated post-chemoradiotherapy. These cells are continuously generated during tumor progression, entering TME and expressing PD-L1 at a high level, allowing GBM to exhaust CTLs and evade attack from the immune system, thereby facilitating GBM relapse (Fig. 1A). By utilizing the aforementioned mechanism, we propose the construction of a drug delivery system for long-term targeting peripheral M-MDSCs, leveraging M-MDSCs to transport the drugs crossing BBB (Fig. 1B). We cloned a sequence of a secretory PD-L1 trap protein into the pcDNA3.1 vector and subsequently conjugated it with cationic protamine to form a drug core with slightly negative charges, which was then encapsulated with pre-prepared Ly-6C-targeted cationic liposomes, and finally modified with a layer of red blood cell membrane (RBCM) on the surface to obtain the nanomedicine αLy-6C-LAMP (Fig. 1B).