Targeted delivery of gene-silencing molecules to brain tumors using specially arranged immune-targeting nanoparticles

Glioblastoma (GBM) is the most malignant primary brain tumor, with an annual incidence of 3.19 to 4.17 cases per 100,000 people worldwide [1]. Despite maximal safe resection, radiotherapy, and temozolomide chemotherapy, median survival rarely exceeds 15 months [1,2,3]. The blood–brain barrier (BBB) and the highly heterogeneous tumor microenvironment severely limit the efficacy of molecularly targeted drugs and immunotherapies [4]. Vascular endothelial growth factor (VEGF), a key driver of angiogenesis, is overexpressed in approximately 90% of GBM cases [5]. Although the anti-VEGF antibody bevacizumab is clinically approved for recurrent GBM, it provides only transient radiographic responses without survival benefit, highlighting the need for more effective VEGF-targeted strategies [6].

RNA interference was first discovered by Fire et al. in 1998, who demonstrated that double-stranded RNA could potently and specifically silence gene expression in Caenorhabditis elegans, establishing the basis for sequence-specific post-transcriptional gene silencing [7]. RNA-based therapeutics, particularly small interfering RNA (siRNA), enable precise silencing of oncogenic or angiogenic targets at the mRNA level and are increasingly being investigated as potential tools for GBM therapy [8,9,10].

Among non-viral nucleic acid delivery systems, lipid nanoparticles (LNPs) represent the most clinically validated platform, as evidenced by their successful application in mRNA vaccine development [11,12,13]. However, conventional LNPs lack tumor-targeting specificity, and existing antibody conjugation methods often suffer from random orientation, reduced antigen binding, and manufacturing complexity [14,15]. To overcome these limitations, we previously established the synthesis of the Fc-binding peptide (FcBP)-lipid for targeted delivery of messenger RNA-loaded anti-transferrin antibody-modified LNPs [16].

Here, we developed a modular, GBM-specific targeting platform using the FcBP-lipid for non-covalent post-insertion of antibodies for preformed siRNA-loaded LNP delivery. This approach preserves antibody orientation and recognizing function, enables rapid and scalable manufacturing, and is adaptable to diverse antibodies and RNA cargos. As a proof of concept, we targeted programmed death-ligand 1 (PD-L1), which is expressed in 35–62% of gliomas [17,18], not for immune checkpoint inhibition but as a tumor-selective entry route for VEGF-siRNA delivery. In subcutaneous GBM models, we demonstrate that PD-L1–FcBP-LNPs achieved >60% tumor growth inhibition without systemic toxicity, demonstrating both potent efficacy and clinical translation potential for precision oncology applications.

Recent advances in RNA therapeutics have increased interest in siRNA for cancer treatment, including GBM [8,9]. However, efficient and tumor-specific delivery of siRNA to solid tumors, such as GBM, remains a major challenge because of poor tumor penetration, rapid systemic clearance, and off-target effects [21,22,23]. LNPs are among the most clinically validated non-viral delivery systems for nucleic acids [24]; however, their inherent lack of tumor-targeting capability limits their therapeutic efficacy without additional surface functionalization [25].

Conventional antibody-mediated LNP targeting strategies often rely on covalent conjugation chemistries, such as maleimide–thiol coupling, which can result in heterogeneous antibody orientation, diminished antigen-binding activity, and added manufacturing complexity [26]. To overcome these limitations, we developed a modular, site-specific targeting platform using FcBP–lipid conjugates. This post-insertion approach enables non-covalent anchoring of antibodies through their Fc region, preserving Fab-mediated antigen recognition and maintaining antibody functionality [16]. In this study, PD-L1 was selected as the surface antigen for antibody-mediated LNP modification, and VEGF was chosen as the siRNA target. The aim was not to evaluate the intrinsic anti-tumor effects of PD-L1 inhibition but rather to exploit PD-L1 expression as a tumor-selective delivery route for siRNA therapeutics. Therefore, we established a subcutaneous tumor model in nude mice with GL261 cells to avoid T-cell-mediated immune responses. PD-L1 was selected as the targeting receptor because 35–62% of glioma patients exhibit PD-L1 expression ≥5% [18]. VEGF-siRNA was chosen because VEGF is overexpressed in approximately 90% of GBM cases [5] and plays a critical role in tumor angiogenesis.

Our results demonstrate that FcBP–lipid can be efficiently incorporated into preformed VEGF-siRNA LNPs without altering key physicochemical parameters, including particle size, PDI, zeta potential, and encapsulation efficiency, which is consistent with previous reports using EKGG- and RRGW-based FcBP conjugates [16]. Decoration with anti-PD-L1 antibodies significantly enhanced cellular uptake (Figure 1) and VEGF gene silencing in PD-L1-positive GL261 glioma cells compared with isotype controls, confirming target specificity. In vitro, anti-PD-L1–FcBP–LNPs suppressed VEGF mRNA expression by >70% and inhibited cell proliferation (Figure 2). In vivo, repeated intratumoral administration in a subcutaneous GL261 model markedly suppressed tumor growth (Figure 3), reduced bioluminescent signals (Figure 4), and decreased tumor volume by 65% compared to controls (Figure 5) without affecting body weight. This tumor-suppressive effect was statistically significant, suggesting that FcBP–LNPs achieved functional VEGF knockdown with minimal off-target toxicity.

The reduction in bioluminescent signals over time (Figure 4A,B) provides additional evidence of tumor suppression and delivery specificity. Serial IVIS imaging on days 3, 6, and 9 revealed a progressive decrease in luminescent intensity in tumors treated with PD-L1–FcBP–LNPs, indicating sustained inhibition of tumor progression. In contrast, tumors treated with control LNPs or isotype antibody-modified LNPs showed no significant signal reduction, highlighting the importance of antigen-specific targeting. Moreover, the >60% reduction in tumor volume (Figure 5) strongly correlates with the observed VEGF protein suppression (Figure 6), supporting a mechanistic link between VEGF knockdown and tumor regression. These findings are consistent with previous reports demonstrating that VEGF-driven angiogenesis plays a central role in GBM growth. Most previous studies on PD-L1 in GBM have focused on immune checkpoint inhibition by blocking PD-L1/PD-1 signaling, rather than exploiting PD-L1 as a delivery target [18,27,28]. Our findings position PD-L1 as a promising entry point for LNP-based therapeutics. Anti-VEGF therapy with bevacizumab has not improved the overall survival of GBM patients [6]. In contrast, our results indicate that siRNA-mediated VEGF silencing, especially with tumor-targeted delivery, can exert potent anti-tumor effects. A plausible explanation is that antibody-based VEGF inhibition may trigger compensatory angiogenic pathways, whereas gene-level knockdown suppresses VEGF production more directly. This activity also raises the possibility of combining VEGF-siRNA with inhibitors targeting alternative angiogenesis pathways to prevent vascular escape mechanisms [5].

In addition, anti-PD-L1 antibody-modified VEGF-siRNA LNPs may confer dual functionality by acting as immune checkpoint inhibitors in a normal host immune system, enabling synergistic anti-tumor responses. The modular FcBP-based approach allows for rapid customization with various antibodies and nucleic acid cargos, supporting the development of personalized nanomedicine. The RRGW peptide used herein binds the Fc region with high affinity (K_D ≈ 0.5 nM) [15], ensuring stable antibody anchoring even in the presence of endogenous IgG, which is a critical feature for clinical translation.

Despite these promising findings, this study has limitations. The subcutaneous tumor model does not replicate the BBB or the complex GBM microenvironment. Although IVIS imaging confirmed tumor-specific accumulation, validation in orthotopic glioma models will be essential. In this regard, recent work by Tang et al. [29] has demonstrated BBB-permeable PD-L1-targeted siRNA–LNPs, supporting the feasibility of antibody-guided delivery to brain tumors. Incorporating non-invasive BBB disruption strategies, such as focused ultrasound and microbubbles, may further enhance brain penetration [4,20]. Additionally, intratumoral administration, while effective in this study, is an inherently localized therapy with limited systemic applicability. However, localized gene therapy approaches have been investigated for GBM [30], suggesting that nucleic-acid-based local delivery is a feasible regulatory pathway. Further limitations include the lack of long-term safety data, immune-related toxicity assessment, and evaluation of repeated dosing, which will require further preclinical studies.

From a clinical perspective, the FcBP–LNP platform has several important implications. GBM remains highly refractory to existing treatments, and the proposed system provides a flexible framework for selective siRNA delivery, potentially reducing off-target effects and improving therapeutic efficacy. Beyond siRNA, this platform may accommodate other nucleic acid cargos, including mRNA, miRNA, or CRISPR/Cas9 components [31]. For example, IL-12 mRNA delivery to enhance immune activation or gene editing targeting GBM oncogenes may significantly broaden the therapeutic scope of this platform [32].

In conclusion, we have demonstrated a versatile and reproducible FcBP-mediated LNP system for VEGF-siRNA delivery via anti-PD-L1 antibody modification. This strategy preserves antibody activity and maintains nanoparticle stability, thereby achieving potent and specific tumor suppression through localized administration. Our findings lay the groundwork for further development of FcBP–LNP platforms as clinically translatable nanomedicines for GBM treatment.