This article is mentioned in:

Introduction

Brain tumors, especially those originating in the central nervous system (CNS) such as gliomas, rank among the most aggressive and lethal forms of cancer. Gliomas are typically derived from neural stem cells, progenitor cells or dedifferentiated mature neural cells, and are characterized by a poor 5-year survival rate of <30% (1,2). The poor survival rate associated with brain tumors can be ascribed to the difficulty of an early diagnosis (3) and the lack of standard treatment regimens (4).

A major obstacle in brain tumor therapy is the presence of physiological barriers, most notably, the blood-brain barrier (BBB) and the blood-brain tumor barrier, that severely limit the delivery of therapeutic agents to intracranial tumors (5,6). These barriers restrict the permeability of most chemotherapeutic agents and are compounded by active efflux mechanisms and insufficient tumor specificity (7,8). As a result, conventional treatment strategies often fail to deliver adequate drug concentrations to the tumor site, diminishing therapeutic efficacy and increasing the risk of systemic toxicity (8). To address these challenges, increasing attention has been directed toward the development of drug delivery systems capable of traversing the BBB and achieving site-specific accumulation within tumor tissues. In this context, nanomedicine has emerged as a particularly promising approach. Among its various applications, nanoparticle (NP)-based delivery systems stand out for their ability to enhance the pharmacokinetic and pharmacodynamic profiles of therapeutic agents (9,10). Rapid progress in nanotechnology has led to the design and optimization of various nanocarriers, such as liposomes, solid lipid NPs (SLNs), dendrimers, polymeric micelles and inorganic NPs, that are engineered with specific physicochemical properties, including particle size, surface charge and ligand functionalization, to enhance biological compatibility and targeting efficiency (9). These tailored systems have demonstrated the ability to across the BBB, accumulate selectively in tumor tissues and facilitate efficient cellular uptake; thereby, enabling localized and sustained therapeutic effects. In neuro-oncology, NP-based platforms have shown strong potential for selective tumor accumulation and localized drug release, reducing off-target effects and enhancing treatment efficacy (10).

Unlike a number of existing reviews that separately discuss drug (8,9,11) or gene (12,13) delivery systems, the present review systematically integrates both therapeutic modalities. This comprehensive synthesis provides a holistic view of how different nanocarriers, ranging from liposomes to dendrimers and inorganic systems, are being engineered to deliver not only chemotherapeutics but also genetic payloads [such as small interfering RNA (siRNA), short hairpin RNA (shRNA) and micro RNA (miRNA)] across the BBB. Furthermore, the present review also adopts a material-based classification (lipid-based, dendritic, polymeric and inorganic nanocarriers) rather than a payload-based or mechanism-based classification. This structure allows readers to directly compare and contrast the physicochemical properties, targeting strategies and translational potential of each nanocarrier platform. To address the gap between bench and bedside, a new section titled 'Clinical challenges in translating NP-based therapies for brain tumors' is also introduced, where the limitations in immunogenicity, scalability, model relevance and human BBB heterogeneity (critical considerations that are often overlooked in reviews focused solely on laboratory advancements) are discussed (14-18). To support conceptual understanding, new schematic diagrams illustrating the BBB structure, transport mechanisms and NP-cell interactions are presented, which distinguish the present review from more text-heavy reviews. We consider that these additions and the integrative approach significantly enhance the value of the present review and differentiate it from prior literature in the field.

Structural features and transport mechanisms of the BBB

The BBB is a dynamic and highly selective interface that separates the systemic circulation from the CNS. The BBB plays a pivotal role in maintaining neural homeostasis by strictly regulating the entry of ions, nutrients and xenobiotics into the brain parenchyma (5). However, this protective function also presents a challenge for the delivery of therapeutics to brain tumors, including glioblastoma multiforme (GBM), which requires sustained and targeted drug accumulation in the tumor microenvironment (6).

Structural and functional composition of the BBB

The BBB is composed primarily of a monolayer of non-fenestrated endothelial cells that are tightly sealed by continuous tight junctions (TJs). These endothelial cells function within the context of the neurovascular unit, which includes pericytes embedded within the basement membrane, astrocyte end-foot processes that ensheath >90% of the vascular surface and the surrounding extracellular matrix components, such as collagen and laminin (Fig. 1A) (19). Together, these elements ensure integrity, selectivity and homeostatic regulation of the brain microenvironment.

Figure 1 Structural and transport features of the BBB. (A) Overview of the brain capillaries and the neurovascular unit, including endothelial cells, pericytes and astrocyte end-feet. (B) Tight junctions formed by claudins, occludin and ZO-1 limit paracellular diffusion. (C) The main BBB transport pathways include transcellular diffusion, paracellular transport, receptor-mediated and adsorptive transcytosis, efflux by ABC transporters and influx via solute carrier proteins. This figure was created in BioRender (paid version; https://www.biorender.com/). BBB, blood-brain barrier; ZO-1, zonula occludens-1; Ang-2, angiopoietin-2.

BBB endothelial cells possess a unique set of functional characteristics that distinguish them from cells in the peripheral vasculature. A key feature of BBB endothelial cells is their exceptionally high transendothelial electrical resistance, often exceeding 1,000-2,000 Ω cm2 in vivo, which effectively limits paracellular diffusion of water-soluble and charged molecules (20,21). Moreover, these cells exhibit markedly low rates of pinocytosis and transcytosis; thereby, restricting non-specific vesicular transport and preserving the selectivity of the barrier (22). Notably, the BBB possesses an active drug efflux system, predominantly mediated by adenosine triphosphate-binding cassette transporters, which effectively transport a wide variety of endogenous metabolites and exogenous compounds, such as a number of chemotherapeutic agents, back into the bloodstream (23).

At the molecular level, the restrictive nature of the BBB is governed by the TJ complexes located at the apical lateral membrane between adjacent endothelial cells. These complexes are composed of integral membrane proteins, such as claudins (particularly claudin-5), occludin and junctional adhesion molecules. These proteins are anchored intracellularly by cytoplasmic scaffold proteins that include zonula occludens-1 (ZO-1), ZO-2 and cingulin (Fig. 1B) (24). These structures not only maintain the structural cohesion of the endothelial monolayer but also dynamically regulate paracellular permeability in response to physiological or pathological stimuli.

Collectively, these structural and functional features make the BBB an extraordinarily effective yet formidable obstacle to therapeutic delivery, especially in the treatment of brain tumors. Therefore, any strategy that enhances drug penetration into the brain must be designed with a comprehensive understanding of this intricate barrier system.

Main transport pathways across the BBB

Despite its restrictive nature, the BBB allows specific translocation of molecules through several tightly regulated pathways (Fig. 1C) (19). Among the various mechanisms, the paracellular route refers to the movement of hydrophilic molecules between adjacent endothelial cells. However, this pathway is effectively sealed in the BBB due to the presence of TJs, which consist of transmembrane proteins (such as claudins, occludin and junctional adhesion molecules) connected to the cytoskeleton via adaptor proteins, such as ZO-1 and ZO-2 (25). Unless TJ modulators or novel nanotechnologies are used, most drugs fail to achieve effective brain exposure via paracellular diffusion (26). However, even when TJs are transiently loosened through pharmacological or physical means to enhance permeability, the risks of drug-induced toxicity and barrier disruption persist and limit the clinical applicability of the paracellular pathway (27,28). Therefore, the paracellular aqueous pathway contributes minimally to drug penetration and remains a formidable hurdle in brain tumor therapy.

Unlike the paracellular pathway, the transcellular lipophilic route allows small, lipid-soluble compounds to diffuse directly through endothelial cell membranes. This route is exploited by endogenous molecules, such as oxygen, carbon dioxide and certain hormones, as well as small-molecule drugs, such as some anesthetics. However, successful diffusion requires favorable physicochemical characteristics; typically, low molecular weight, low hydrogen-bonding potential and a relatively high logP value for the lipid-water partition coefficient (29-31). Numerous anticancer drugs and alkylating agents, including doxorubicin (DOX) or temozolomide (32) and monoclonal antibodies (33), do not meet these criteria, rendering passive diffusion across the BBB inefficient. Moreover, even lipophilic drugs are often subject to rapid efflux via transporters, such as P-glycoprotein, further limiting their accumulation in the brain (34).

Receptor-mediated transcytosis (RMT) is one of the most promising routes for NP-based delivery across the BBB due to its high specificity, affinity and potential for receptor recycling. This mechanism involves ligand binding to specific receptors on the luminal surface of endothelial cells, followed by clathrin-mediated endocytosis and exocytosis into the brain parenchyma (11,35). Numerous preclinical studies have shown that NPs conjugated with ligands, such as transferrin, lactoferrin, angiopep-2 or monoclonal antibodies targeting the transferrin receptor or the insulin receptor, achieve significantly enhanced brain penetration (36-38). These ligands not only facilitate BBB transcytosis but also enable subsequent tumor targeting, making RMT an attractive platform for theranostic applications in glioblastoma and other CNS malignancies.

Adsorptive-mediated transcytosis (AMT) relies on electrostatic interactions between the negatively-charged luminal surface of endothelial cells and positively-charged ligands or carriers (39). This non-specific mechanism is particularly useful to enhance the transport of cationic proteins (40), peptides (41) and NPs (42) across the BBB. Cell-penetrating peptides, such as trans-activator of transcription (TAT), penetratin and rabies virus glycoprotein 29, have been widely used to facilitate AMT-mediated delivery of both small molecules and macromolecules (43-45). Although AMT is less selective than RMT, it has the advantage of delivering diverse payloads independent of specific receptor expression, which expands its applicability across different brain tumor subtypes. However, the use of strong cationic charges must be balanced against the risk of cytotoxicity and non-specific distribution.

Solute carrier transporters constitute a broad family of influx proteins at the BBB that enable the uptake of essential nutrients, metabolites and ions into the brain (46). These transporters include glucose transporter 1, which facilitates glucose entry (47), L-type amino acid transporter 1, which is responsible for transporting large neutral amino acids (48), and monocarboxylate transporters, which mediate the influx of lactate and other monocarboxylates (49). Notably, these transporters can be harnessed for brain-targeted drug delivery by chemically modifying drugs or nanocarriers to resemble endogenous substrates. For example, glucose-conjugated or amino acid-decorated NPs have shown promise in facilitating carrier-mediated transport across the BBB (50). Such approaches offer a non-invasive, receptor-independent route to enhance the delivery of small-molecule therapeutics or genetic materials into brain tumor tissues, making solute carrier-mediated influx an attractive mechanism for developing effective brain-targeted therapies.

An in-depth understanding of these transport routes is critical for designing effective nanocarrier platforms for brain tumor therapy. Functionalization with targeting ligands (such as angiopep-2 and transferrin) enables RMT-based entry, while surface charge modulation enhances AMT. Dual-targeting strategies that combine BBB penetration with tumor specificity are also emerging as effective approaches for glioma therapy.

NP-based drug delivery in brain tumors

Advancements in nanotechnology have opened new avenues to overcome the notable challenges posed by the BBB and the complex microenvironment of brain tumors. In particular, NP-based delivery systems offer versatile platforms capable of transporting a wide array of therapeutic agents across the BBB, including small-molecule drugs, nucleic acids and gene-editing constructs, with enhanced precision, stability and efficacy. Over the past decade, a diverse spectrum of nanocarriers has been engineered to meet the distinct physicochemical and biological requirements of CNS drug delivery. These include lipid-based carriers, such as liposomes and SLNs (51,52), polymeric NPs (50) constructed from biodegradable materials, such as poly(lactic-co-glycolic acid) (PLGA) (53) or polyethylene glycol (PEG) (50), dendrimers with highly-branched architectures (54) and inorganic systems, such as gold NPs (AuNPs) (55), graphene oxide (GO) (56) and magnetic iron oxide NPs (IONPs) (57). These platforms possess unique advantages in drug loading capacity, targeting capabilities, biocompatibility and multifunctionality, making them suitable not only for conventional chemotherapy but also for gene silencing, messenger RNA replacement and immunomodulation. Table I (50-69) summarizes representative NP-based delivery systems applied to brain tumor therapy, highlighting their physicochemical characteristics (such as particle size and zeta potential), payloads, targeting strategies and therapeutic outcomes. Furthermore, given the formidable barrier posed by the BBB, a variety of surface modification approaches have been developed to enhance nanocarrier penetration and targeting efficiency. These approaches, including ligand conjugation, cell-penetrating peptides and stimuli-responsive coatings, are summarized in Table II (7,70-86).

Table I Summary of nanoparticle-based delivery systems for brain tumor therapy.

Table I

Summary of nanoparticle-based delivery systems for brain tumor therapy.

A, Lipid-based
First author, year	Nanocarrier type	Particle size (nm)	Surface charge (zeta potential)	Drug loading efficiency (%)	Encapsulation efficiency (%)	Therapeutic payload	Application mode	Targeting strategy	(Refs.)
Cen et al, 2023	Liposomes	120-170	Negative	5.5	90	DOX	Drug	Mitochondria-targeting	(58)
Zong et al, 2014	Liposomes	100-120	Negative		90	DOX	Drug	TAT/T7 dual-targeting	(63)
Estella-Hermoso de Mendoza et al, 2011	SLNs	105-140	Negative	15.31±3.29	85	Edelfosine	Drug	Passive accumulation	(52)
Pandian et al, 2021	SLNs	113	Negative	22.14±1.2	88±2.05	Rutin	Drug	EGFR targeting	(54)
Campani et al, 2020	LPHNs	120-160	Positive		72-100	miR 603	Gene	Passive accumulation	(51)
Saw et al, 2018	Liposomes	~112	Neutral		90	siRNA	Gene	EDB-targeting	(62)
B, Dendrimers
First author, year	Nanocarrier type	Particle size (nm)	Surface charge (zeta potential)	Drug loading efficiency (%)	Encapsulation efficiency (%)	Therapeutic payload	Application mode	Targeting strategy	(Refs.)
He et al, 2011	PAMAM	~14	Neutral			DOX	Drug	Tf/WGA dual-targeting	(59)
Liu et al, 2019	PAMAM	7-34	Positive	4.3	58.9	DOX	Drug	LRP1/EGFR-mediated	(61)
Khoury et al, 2020	PAMAM	4.8±0.9	Neutral	12		JHU29	Drug	TAMs targeting	(60)
Liaw et al, 2021	PAMAM	4.7±0.6	Neutral	18		Triptolide	Drug	TAMs targeting	(65)
C, Polymeric
First author, year	Nanocarrier type	Particle size (nm)	Surface charge (zeta potential)	Drug loading efficiency (%)	Encapsulation efficiency (%)	Therapeutic payload	Application mode	Targeting strategy	(Refs.)
Wohlfart et al, 2011	PLGA nanoparticles	72.5±2.2	Neutral	8.2±0.6	90.4±2.3	PTX	Drug	Passive accumulation	(56)
Jiang et al, 2014	PEG-PTMC	71		6.52	94.8	PTX	Drug	GLUT targeting	(50)
Zhang et al, 2016	PEG-PLA	39.95	Neutral			DOX, AR-825	Drug	αvβ3 integrin-targeting	(69)
Lu et al, 2006	PEG-PLA	70-130	Negative	0.04		6-coumarin	Drug	Passive accumulation	(66)
Negron et al, 2019	PEG-PEI	50	Positive		100	DNA	Gene	Passive accumulation	(67)
D, Inorganic
First author, year	Nanocarrier type	Particle size (nm)	Surface charge (zeta potential)	Drug loading efficiency (%)	Encapsulation efficiency (%)	Therapeutic payload	Application mode	Targeting strategy	(Refs.)
Feng et al, 2017	AuNPs	80	Negative	16%		DOX	Drug	EGFR targeting	(64)
Grafals-Ruiz et al, 2020	AuNPs	20	Negative	AuNP:OMIs (mol/mol)=1:50		RNAi	Gene	ApoE/RVG dual-targeting	(53)
Zhao et al, 2020	GO	100-400	Negative			DOX	Drug	LRP targeting	(57)
Wang et al, 2013	Silica-Graphene	50-250	Negative	1.27±0.11 μg DOX/μg GSP		DOX	Drug	Passive accumulation	(68)
Norouzi et al, 2020	IONPs	4.76±0.7	Positive	3.45±0.01%		Salinomycin	Drug	External magnetic fields	(55)

[i] DOX, doxorubicin; EGFR, epidermal growth factor receptor; EDB, extra domain B; LRP, lipoprotein receptor-related protein; TAM, tumor-associated macrophage; PTX, paclitaxel; GO, graphite oxide; IONP, iron oxide nanoparticles; PEG, polyethylene glycol; PEI, polyetherimide; GLUT, glucose transporter; SLN, solid lipid NPs; LPHN, lipid-polymer hybrid NPs; PAMAM, poly(amidoamine); PLGA, poly(lactic-co-glycolic acid); PTMC, Poly(trimethylene carbonate); PLA, polylactic acid ; GSPI, mesoporous silica-coated graphene nanosheet; OMI, oligonucleotide miRNA inhibitors; AuNP, gold nanoparticles.

Table II Current modification strategies for nanocarriers to cross the BBB.

Table II

Current modification strategies for nanocarriers to cross the BBB.

Modification strategies	Related protein	Role in crossing the BBB	(Refs.)
Receptor-mediated transcytosis	TfR	Responsible for intracellular transport of transferrin and is the most used and validated receptor.	(72)
	LDLR	Responsible for the endocytosis of LDLs, such as apolipoprotein B and apolipoprotein E.	(70,73)
	Insulin receptor	Widely expressed receptor in the brain microvessels.	(74)
	Insulin-like growth factor 1 receptor	Expressed in the brain and the cerebral vessels	(75)
	LDLR-related protein 1	Expressed on the surface of BBB cells. Angiopep 2 has been investigated for LRP1-mediated drug delivery into the brain.	(71)
	Nicotinic acetylcholine receptors	Highly expressed on the capillary endothelium of the brain, susceptible to the inhibition by peptide neurotoxins.	(77)
Peptides	GYR	Targets the TfR.	(76)
	Interleukin 13 peptide	IL-13Rα2-mediated transcytosis.	(7)
	Transactivating-transduction	Endocytosis micropinocytosis pore formation.	(81)
	Leptin30	Endocytosis pathway	(82)
	Angiopep-2 peptide	Binds with LRP1 and promotes drug delivery by LRP1-mediated transcytosis.	(79)
	Penetratin	Direct penetration endocytosis.	(86)
	DK17	Membrane potential.	(78)
Carrier-mediated transcytosis	Glucose receptors	Utilizing GLUT transporter protein for transmembrane transport	(80)
	Fatty acid	Targeted lipid metabolism as a substrate for energy production.	(84,85)
	Nucleoside transporters	Binds to receptors important for recycling pathways.	(83)

[i] TfR, transferrin receptor; LDL, low-density lipoprotein; LDLR, LDL receptor; GYR, GYRPVHNIRGHWAPG, herein as GYR peptide; DK17, DRQIKIWFQNRRMKWKK, herein as Dk17 peptide; GLUT, glucose transporter; LRP1, LDL receptor-related protein 1; BBB, blood brain barrier.

The present review also provides a holistic view of how different nanocarriers, ranging from liposomes to dendrimers and inorganic systems, are being engineered to deliver not only chemotherapeutics but also genetic payloads (such as siRNA, shRNA and miRNA) across the BBB. In the following sections, these nanocarrier classes and their applications in delivering both chemical and genetic therapeutics to brain tumors are reviewed, with an emphasis on structural design, delivery mechanisms and preclinical outcomes.

Lipid-based nanocarriers

Lipid-based nanocarriers, including liposomes, SLNs and lipid-polymer hybrid NPs (LPHNs), represent some of the most well-established and clinically relevant platforms for brain-targeted drug and gene delivery. These carriers are biocompatible, structurally versatile and easily modifiable, making them ideal for encapsulating hydrophilic and lipophilic drugs, as well as genetic materials, such as siRNA, miRNA and plasmid DNA (51,62).

Liposomes

Liposomes are spherical vesicles composed of lipid bilayers that can be modified with targeting ligands to enhance BBB penetration and tumor specificity. Human interleukin-13 (IL-13) is a cytokine released by activated T cells and can induce proinflammatory and anti-inflammatory immune responses (87). Reports have indicated that IL-13 receptor α2 (IL-13Rα2) is found at high levels in brain tumors, such as GBM and pilocytic astrocytomas, but not in normal tissues (88,89). Therefore, IL-13Rα2 is a promising target to deliver cytotoxic drugs to various aggressive brain tumors. Madhankumar et al (90) developed IL-13-conjugated liposomes to deliver DOX specifically to brain tumors, which led to less cytotoxicity and greater accumulation and retention of DOX in the glioma cells compared with therapy without these liposomes. Mice receiving intraperitoneal injections of IL-13-conjugated liposomes carrying DOX showed a significant decrease in U251 glioma growth (90). In addition to IL-13, numerous other ligands promote the brain tumor targeting ability of liposomes. Zong et al (63) designed a dual-targeting liposomal system modified with TAT (AYGRKKRRQRRR) and T7 (HAIYPRH) peptides. The specific ligand, T7, targets the BBB and brain tumors, while the non-specific ligand, TAT, enhances passage through the BBB and increases tumor penetration. Intravenous administration of dual-targeted DOX liposomes led to enhanced delivery of DOX to brain glioma tissue in a C6 glioma mouse model. Furthermore, Cen et al (58) developed a DOX-loaded liposomal system modified with the SS31 peptide, which exhibits dual-targeting capabilities by facilitating both BBB penetration and mitochondrial targeting within glioma cells. Liposomal system-DOX liposomes accumulated effectively in tumor tissues, significantly inhibiting tumor growth and extending survival rates without notable toxicity. This dual-targeting approach offers a promising strategy to enhance the therapeutic efficacy of liposomal DOX in glioma treatment.

In gene delivery, cationic liposomes have enabled the systemic administration of siRNAs and miRNAs to brain tumors. For instance, Wei et al (91) performed a layer-by-layer assembly of protamine/chondroitin sulfate/siRNA/cationic liposomes followed by T7 peptide modification to create a targeted siRNA-epidermal growth factor receptor (EGFR) delivery system. This modified NP specifically accumulated anti-EGFR siRNA in brain tumor tissues upon intravenous administration, leading to downregulation of EGFR expression and high survival rates in a U87 mouse glioma model.

Cyclophilin A (CypA) is a tumor marker that is elevated in various cancer types, including liver, brain and lung cancer (92-94). CypA plays a crucial role in epithelial to mesenchymal transition and cancer metastasis (94) and can be used as a target for tumor therapy. Saw et al (62) developed a liposome-based extra domain B targeting nanoplatform for CypA siRNA delivery and glioblastoma therapy. After systemic administration, these new siRNA delivery NPs targeted glioma cells and effectively inhibited glioblastoma tumor growth by silencing CypA, which plays a crucial role in the malignant transformation of brain cancer and the maintenance of glioma cell stemness (62). These findings indicate that this RNA interference NP platform is a potentially effective tool for targeted brain tumor therapy.

The biopharmaceutical challenges related to the therapeutic use of non-coding RNAs could potentially be addressed by lipid NPs (95). The therapeutic efficacy of non-coding RNAs delivered by lipid NPs has been confirmed in different forms of cancer (96).

SLNs

SLNs are submicron colloidal carriers composed of biocompatible solid lipids stabilized by surfactants. Compared with conventional liposomes, SLNs offer superior physicochemical stability, lower burst release and improved control over drug release kinetics (97). SLNs are capable of encapsulating both hydrophilic and lipophilic agents and are amenable to various administration routes, including intravenous and oral delivery (52,54).

SLNs have shown notable success in enhancing brain delivery of chemotherapeutics. For example, edelfosine-loaded SLNs administered orally to glioma-bearing mice crossed the BBB efficiently and accumulated within tumor tissues, leading to a significant reduction in tumor volume and prolonged survival compared with non-SLN therapy (52). Beyond synthetic drugs, SLNs have been used to deliver poorly-soluble natural products with anticancer properties. Rutin, a flavonol glycoside with pro-apoptotic and anti-proliferative effects, has low water solubility and poor oral bioavailability (98). Encapsulation within SLNs improves the BBB permeability and bioavailability of rutin, resulting in enhanced brain accumulation and tumor suppression in preclinical models (54,99). Further innovations include surface-modified SLNs with targeting ligands, such as transferrin or apolipoprotein E (ApoE) mimetics, which exploit RMT pathways to improve brain specificity (100). Additionally, SLNs have been used for co-delivery systems, such as the simultaneous encapsulation of paclitaxel (PTX) and P-glycoprotein inhibitors, to overcome drug efflux mechanisms in glioma cells, with less cytotoxicity and enhanced therapeutic outcomes (101).

LPHNs

LPHNs combine the structural stability of polymeric cores with the membrane fluidity and biocompatibility of lipid shells; thereby, bridging the functional advantages of both systems. These hybrid carriers are particularly well-suited for gene delivery due to their high encapsulation efficiency, low cytotoxicity and the ability to promote endosomal escape and intracellular release of nucleic acid cargos (102).

One of the most promising hybrid platforms is the self-assembled lipid-coated calcium phosphate NP (SANP). These carriers typically consist of a calcium phosphate core complexed with genetic material that is coated with PEGylated cationic lipids, such as N-[1-(2,3-dioleoyloxy) propyl]-N,N,N-trimethylammonium and modified lipids, such as ceramide-PEG2000. In glioma models, SANPs have demonstrated high miRNA and shRNA delivery efficiency, with significant accumulation in tumor tissue and potent gene silencing effects in vivo (51,103). As with miRNA therapy, SANPs have been used for co-delivery of nucleic acids and chemotherapeutics, such as bisphosphonates or DOX (104). The lipid shell allows prolonged systemic circulation and BBB traversal, while the calcium phosphate core promotes endosomal rupture via a 'proton sponge' effect, enhancing cytoplasmic release. For instance, a recent study incorporated miRNA-124 and PTX into SANPs to simultaneously silence oncogenic pathways and induce apoptosis, achieving synergistic antitumor efficacy with minimal systemic toxicity (105).

Further optimization strategies include the use of targeted peptides (such as arginyl-glycyl-aspartic peptides and T7) on the NP surface to increase glioma-specific uptake, as well as pH-responsive lipids that enable controlled gene release in the acidic tumor microenvironment (63).

Dendrimers

Dendrimers are highly-branched, monodisperse and nanoscale macromolecules that offer exceptional versatility for drug and gene delivery in brain tumors. Their well-defined architecture, multivalent surface and internal cavities allow for the simultaneous loading of therapeutic agents and targeting ligands (106). Among dendrimers, poly(amidoamine) (PAMAM) dendrimers have been the most extensively studied due to their biocompatibility, tunable size and ease of functionalization. These features enable dendrimers to serve as flexible nanocarriers for drug molecules, genetic materials and multifunctional payloads (107).

In drug delivery, dendrimers function either as nanocapsules to encapsulate hydrophobic agents or as nanoconjugates bearing covalently attached drugs via cleavable linkers. PAMAM dendrimers have shown particular promise in brain tumor models when engineered with dual-targeting moieties. For instance, dendrimers co-modified with angiopep-2 and an EGFR-binding peptide exhibited enhanced BBB permeability and glioma-targeting ability via low-density lipoprotein receptor-related protein 1 (LRP1)- and EGFR-mediated transcytosis. This platform significantly improved the delivery of DOX, resulting in prolonged survival and reduced systemic toxicity in glioma-bearing mice compared with therapy without this platform (61). Similarly, dendrimers functionalized with transferrin and wheat germ agglutinin achieved dual-site targeting, facilitating enhanced uptake into glioma cells and accumulation within tumor tissues; thereby, improving therapeutic outcomes after systemic administration of DOX (59).

Dendrimers also provide a promising scaffold for gene delivery owing to their cationic surface, which facilitates complexation with nucleic acids. PAMAM dendrimers have been used to deliver a range of genetic payloads, including siRNA, shRNA and plasmid DNA, for gene silencing and transcriptional modulation in brain tumors (108). Surface modification with PEG or β-cyclodextrin improves stability and biocompatibility while reducing immunogenicity (109). By tuning the surface charge and size, these dendrimer complexes can traverse the BBB and achieve transgene expression within tumor cells. Notably, dendrimers support intracellular delivery via endosomal escape mechanisms and avoid rapid degradation of nucleic acids in circulation (110).

Beyond direct tumor cell killing, dendrimers have demonstrated potential in modulating the tumor immune microenvironment. Hydroxyl-terminated PAMAM dendrimers have been shown to selectively accumulate in tumor-associated macrophages and activated microglia, particularly in glioma tissues with compromised BBB integrity (60,111-113). A notable example is the targeted delivery of triptolide, a potent immunomodulatory agent, via dendrimer conjugation. This formulation not only enhanced tumor localization but also reprogrammed tumor-associated macrophages from an M2-like immunosuppressive phenotype to an M1-like pro-inflammatory state, resulting in both reduced tumor burden and minimized systemic toxicity in glioblastoma (65).

The multivalency of dendrimers allows them to be co-loaded with chemotherapeutic agents and nucleic acids, facilitating combination therapies in a single nanoplatform. For example, hybrid systems have been developed to simultaneously deliver DOX and siRNA, thereby achieving synergistic effects through cytotoxicity and gene silencing (114). Such designs enable coordinated regulation of tumor proliferation and resistance mechanisms, providing a rational strategy for multifaceted glioma therapy.

Polymeric NPs

Polymeric NPs have garnered notable interest for both drug and gene delivery in brain tumor therapy due to their excellent biocompatibility, structural flexibility and tunable physicochemical properties (115). Typically synthesized from US Food and Drug Administration-approved polymers, such as PLGA, PEG, polycaprolactone and polyethylenimine (PEI), these systems allow for fine control of particle size, surface charge, release kinetics and functional modification (50,56,67).

PLGA-based NPs are widely used to deliver chemotherapeutic agents across the BBB. For example, PTX-loaded PLGA NPs coated with poloxamer 188 significantly increased BBB penetration and PTX accumulation in glioblastoma tissues, outperforming free drug administration in a rat GBM model (56). Co-delivery systems based on PLGA and chitosan have also been developed to encapsulate carmustine and O6-benzylguanine. This combination synergistically depleted O6-methylguanine-DNA methyltransferase, sensitizing tumors to alkylating agents and improving survival in F98 glioma-bearing rats (116). Targeted polymeric micelles further enhance brain accumulation. For instance, 2-deoxy-D-glucose-modified poly(ethylene glycol)-co-poly(trimethylene carbonate) PTX NPs were developed as a potential dual-target drug delivery system that can enhance BBB penetration through glucose transporter (GLUT)-mediated cross-cellular action and improve drug accumulation in gliomas through GLUT-mediated endocytosis. These glucose-decorated NPs increased drug accumulation in the brain and led to improved survival rates in an RG2 mouse glioma model, compared with plain NPs and PTX (50). Similarly, transferrin receptor-targeted micelles carrying PTX showed prolonged tumor retention and increased therapeutic efficacy in U87 MG intracranial gliomas (117). Other advanced polymeric formulations include DOX-loaded polybutylcyanoacrylate NPs and methoxy-PE G-poly(ε-caprolactone) micelles, which have demonstrated superior tumor accumulation, reduced systemic toxicity and enhanced survival in C6 glioma models (69,118).

Polymeric nanocarriers also support efficient gene transfer across the BBB. One example is cationic albumin-conjugated PEG NPs loaded with plasmid-encoding tumor necrosis factor-related apoptosis-inducing ligand. These particles have been shown to cross the BBB via AMT and induce selective apoptosis in glioma cells following systemic administration, with strong antitumor effects in vivo (66). DNA-loaded NPs formulated with an inclusion of densely PEGylated PEI (that is, brain-penetrating NPs) were recently reported to provide widespread transgene expression in the rat brain while exhibiting good in vivo safety profiles (119). Additionally, Negron et al (67) developed polymer-based DNA-loaded NPs with small particle diameters (~50 nm) and non-adhesive surface PEG coatings, which efficiently penetrated both brain tumor tissue and healthy brain parenchyma. These brain-penetrating NPs facilitated widespread transgene expression in heathy rodent striatum and aggressive brain tumor tissue through intracranial administration. These systems efficiently delivered DNA and siRNA while minimizing neuroinflammation and cytotoxicity.

In summary, polymeric NPs offer an adaptable platform to deliver a broad spectrum of therapeutic agents, including cytotoxic drugs, plasmid DNA and siRNA, across the BBB. Their ability to incorporate targeting ligands, modulate release kinetics and achieve sustained transgene expression makes polymeric NPs invaluable in both standalone and combination therapy strategies for malignant brain tumors.

Inorganic NPs

Inorganic NPs offer distinct physicochemical advantages, such as precise size control, surface reactivity, magnetic or optical responsiveness and structural rigidity, which makes them highly valuable in brain tumor therapy (120). Unlike organic materials, inorganic carriers can integrate diagnostic and therapeutic functions within a single platform (that is, theranostics) (121). This section focuses on three prominent classes: AuNPs, graphene-based nanocarriers and IONPs, detailing their respective applications in both drug and gene delivery.

AuNPs

AuNPs have been used to enhance drug stability, targeting and controlled release. For example, self-assembled gold nanospheres (~80 nm) have been designed to remain stable in circulation but disassemble in response to acidic and reductive conditions in the glioma microenvironment. This responsiveness improves intratumoral accumulation while minimizing off-target toxicity through renal clearance of dissociated particles (64). In targeted chemotherapy, antibody-conjugated gold-silica nanoshells directed against IL-13Rα2, which is upregulated in glioblastoma, enabled precise tumor cell recognition and thermal ablation. Upon near-infrared irradiation, these nanoshells induced cell death in U373 and U87 glioma models (122). Moreover, Feng et al (64) described a practical method for assembling AuNPs into ~80 nm nanospheres for use as a drug delivery platform. This platform not only enhanced drug retention in brain tumors but was able to dynamically switch to the single formulation for excretion. These self-assembled AuNPs can target brain tumor cells and respond to tumor microenvironmental parameters, including high vascular permeability and acidic and reductive conditions, suggesting that gold nanoassemblies may be an effective targeting strategy for brain tumor treatment (123).

AuNPs are also efficient vehicles for genetic payloads due to their high surface area and ease of nucleic acid conjugation. For instance, β-cyclodextrin-modified dendrimer-entrapped AuNPs have been used to co-deliver siRNAs targeting Bcl-2 and vascular endothelial growth factor. These carriers achieved robust gene silencing in glioma cells with minimal cytotoxicity and good serum stability (124). Another notable approach involves spherical nucleic acids, in which oligonucleotides are densely arranged on AuNP cores. When encapsulated in ApoE-modified liposomes, these spherical nucleic acids accumulate in glioma tissues and facilitate efficient miRNA modulation, opening avenues for post-transcriptional gene regulation (53).

Graphene-based nanocarriers

GO sheets, with their large surface area and π-π interaction capability, are excellent carriers for chemotherapeutics. Angiopep-2-modified GO loaded with DOX was shown to enhance uptake by LRP1-upregulated glioma cells, achieving greater cytotoxicity than that with free DOX or unmodified GO formulations (57). In a dual-modality platform, mesoporous silica-coated GO nanosheets (GSPI) were functionalized with targeting peptide and loaded with DOX (GPSID). The resulting GSPID system combined chemo- and photothermal therapy. Upon near-infrared laser exposure, the GSPID system induced significant glioma cell death through thermal enhancement of drug activity (68).

Although limited compared with drug delivery, GO has also been used to deliver nucleic acids. The surface of GO enables electrostatic complexation with negatively-charged RNA or DNA (125). Modified GO-PEI hybrids can deliver shRNA into glioma cells with good transfection efficiency, especially when further engineered for nuclear targeting or stimuli-responsive release. For example, co-delivery of irinotecan and shRNA targeting stomatin-like protein 2 (SLP2) using magnetic GO hybrid nanocarriers improved glioma treatment efficacy by downregulating oncogenic pathways and enhancing chemotherapeutic synergy (126).

IONPs

IONPs exhibit superparamagnetic properties, allowing external magnetic fields to enhance tumor-specific delivery across the BBB. For example, DOX-loaded trimethoxysilylpropyl-ethylenediamine triacetic acid-stabilized IONPs, when co-administered with a cadherin-binding peptide and magnetically-guided, significantly enhanced toxicity in glioma cells by temporarily loosening TJs at the BBB (127). Targeting was further improved using linTT1 peptide-functionalized IONPs, which bind to p32/gC1qR proteins upregulated in the glioma vasculature. These IONPs boosted glioma-specific uptake and promoted co-delivery of other NPs, such as albumin-PTX and silver nanocarriers (128). Another innovative application involved salinomycin-loaded IONPs, modified with PEG and PEI to improve circulation and BBB transit. These NPs selectively induced apoptosis in GBM cells via the generation of reactive oxygen species while sparing healthy tissue, making them a promising candidate for magnetically-targeted chemotherapy (129).

Although still at a preclinical stage, IONPs have been adapted for magnetofection, which is a technique that uses magnetic force to accelerate and direct nucleic acid delivery (130). PEI-coated IONPs complexed with plasmids or siRNA have been shown to transfect glioma cells both in vitro and in vivo under magnetic guidance (131). Co-delivery systems integrating siRNA and small molecules (such as cisplatin) are also under exploration, aiming to synergize gene silencing and DNA damage mechanisms.

Delivery strategies based on the protein corona (PC) in brain tumors

Although NPs as drug carriers hold great promise for brain tumor therapy, numerous challenges remain in clinical translation (132). The PC poses a significant biological challenge for the translation of targeted drug delivery systems from the laboratory to clinical use, greatly impacting the targeting accuracy and biodistribution (133). NPs entering a biological environment experience surface modification due to dynamic physicochemical interactions with proteins and other biomolecules. The surface becomes coated with plasma proteins, lipids and other molecules, forming a PC (134). The PC modifies the size and surface properties of NPs (135), which makes it difficult for NPs to arrive at target sites, including brain tumors (Fig. 2) (136,137). Therefore, PC-mediated targeting by maintaining the function of target plasma proteins on the NP surface may offer a new approach for specific drug delivery. The two most important types of proteins in the PC are opsonins and de-opsonins (138). Mononuclear macrophages can easily identify and clear NPs with surface opsonins, which greatly reduces drug delivery (139). De-opsonins show the opposite effect to that of opsonins as they prevent NP uptake by macrophages and prolong the NP circulation time. Albumin, the most abundant plasma protein, is a typical de-opsonin. Coating NPs with albumin significantly inhibits the adsorption of opsonin, reduces macrophage phagocytosis, decreases NP cytotoxicity and prolongs NP circulation in the blood (140).

Figure 2 General structure of PC and nanoparticles with PC across the BBB in brain tumors. The chemical structure of PC and the dynamics of its configuration depend on the affinity of proteins to the functional groups located on the surface of NPs and the concentration of protein. High affinity coated proteins form 'hard' corona, while low affinity coated proteins form 'soft' corona. The BBB is selective and restrictive to a variety of molecules. Endothelial cells and the basement membrane, together with strong lateral tight junctions, further impede its permeability. This figure was created in BioRender (paid version; https://www.biorender.com/). BBB, blood-brain barrier; NP, nanoparticle; PC, protein corona.

PC-mediated targeting provides a novel approach to precise drug delivery by accurately controlling the interaction modes of functional plasma proteins on the surface of NPs. Zhang et al (141) developed bioinspired liposomes (SP-sLips) by modifying the liposomal surface with a short non-toxic peptide derived from amyloid β-protein (Aβ)1-42. This peptide specifically interacts with the lipid-binding domain of exchangeable apolipoproteins. SP-sLips absorb plasma apolipoproteins A1, E and J, and achieve brain-targeted delivery by exposing the receptor-binding domain of apolipoproteins. DOX-loaded SP-sLips have demonstrated a significant improvement in brain distribution and anti-brain cancer effects compared with DOX-loaded plain liposomes. Another study (142) fabricated PEG-polyethylene-polylactice acid nanomicelles (PMs) decorated by Aβ25-35 with an amide bond (Aβ-CN) to load PTX. Aβ-CN peptide was used to bind the C-terminal domain of ApoE when exposed to physiological conditions and specifically bound to the lipid-binding domain of ApoE in vivo, forming an ApoE-enriched PC surrounding Aβ-CN-PMs. This formation demonstrated enhanced anticancer effects and a favorable tissue distribution profile, with rapid accumulation in brain tumor tissues.

Clinical challenges in translating NP-based therapies for brain tumors

Despite the notable preclinical progress achieved using nanocarriers for drug and gene delivery in brain tumors, translating these systems into clinical practice remains a formidable task. Numerous NP platforms, ranging from liposomes and dendrimers to inorganic and hybrid systems, have demonstrated efficient BBB penetration, tumor targeting and therapeutic efficacy in animal models. However, only a handful have advanced to clinical trials (14,17) and none have become standard-of-care for glioma treatment. Several key clinical challenges hinder this transition: i) Safety, immunogenicity and long-term biocompatibility. Nanocarriers often exhibit unpredictable interactions with the immune system. For example, cationic surfaces can activate complement cascades or induce cytokine release, leading to hypersensitivity or off-target effects (18). Even biocompatible materials, such as PEG, may elicit anti-PEG antibodies after repeat administration (16), undermining therapeutic efficacy. Moreover, long-term accumulation of inorganic NPs (such as gold or iron oxide) in non-target tissues raises concerns about potential chronic toxicity. ii) Physiological complexity of the human BBB and tumor microenvironment. While most nanocarrier studies rely on rodent models, the human BBB is markedly more selective and the glioblastoma vasculature is heterogeneous and dynamically remodeled during disease progression (15). As a result, targeting strategies that succeed in preclinical models may underperform or fail in clinical settings. Additionally, the tumor microenvironment, including hypoxia, immune suppression and abnormal interstitial pressure, can impair nanocarrier penetration and distribution within brain tumors (143). iii) Reproducibility, scale-up and manufacturing constraints. The structural complexity of nanocarriers presents a major hurdle for consistent large-scale production. Batch-to-batch variability in size, surface charge and drug loading efficiency can affect pharmacokinetics and therapeutic outcomes. Moreover, regulatory agencies require detailed physicochemical characterization, stability profiling and Good Manufacturing Practice compliance, all of which remain difficult to achieve for sophisticated multifunctional systems. iv) Lack of predictive models for human translation. Currently, to the best of our knowledge, there is a lack of robust in vitro or in silico models that accurately predict nanocarrier behavior in the human brain. This limits the ability to assess clinical feasibility early in the developmental pipeline, resulting in high attrition rates during translation.

To bridge the gap between bench and bedside, future efforts must focus not only on optimizing the design and targeting efficiency of nanocarriers, but also on improving their biological stability, minimizing immunogenicity and enhancing reproducibility on a clinical scale. Among the most critical biological factors influencing in vivo behavior is the formation of a PC, a dynamic biomolecular layer that adsorbs onto NPs upon contact with biological fluids.

Conclusions and new perspectives

Innovative local chemotherapy using gene therapy-based NPs has been proposed as a treatment for patients with brain tumors. Zhang et al (130) modified the porous structure of IONPs by attaching carboxyl groups to enable the codelivery of siRNA targeting glutathione peroxidase 4 (si-GPX4) and cisplatin with high drug loading efficiencies. This nanoformulation exerted notable effects on glioblastoma cells, but the effects on normal human astrocytes were limited. During intracellular degradation, IONPs significantly increased iron (Fe2+ and Fe3+) levels, while cisplatin destroyed nuclear DNA and mitochondrial DNA, resulting in apoptosis. Additionally, the co-released si-GPX4 enhanced the therapeutic efficacy by promoting ferroptosis, suggesting that si-GPX4 may be a safe and effective inducer of ferroptosis and apoptosis for combined brain tumor therapy. Chuang et al (126) developed a co-delivery strategy for the chemotherapeutic drug, irinotecan (CPT-11), and SLP2 shRNA using dual targeting NPs by combining magnetic targeting with ligand-mediated active targeting for the treatment of brain tumors. The magnetic GO (mGO) nanocarrier was conjugated with chitosan and urocanic acid (mGOCU) for shRNA complexation and endosomal escape. After PEGylation by interaction of the hydrophobic tails of 1,2-distearoyl-sn-glycero-3-phosphoethanolamine-N-[maleimide(polyethylene glycol)-2000] with the mGO surface followed by binding gastrin releasing peptide to PEG (mGOCUG), CPT-11 was loaded via π-π interaction and the SLP2 shRNA was complexed to form mGOCUG/CPT-11/shRNA. Intravenous administration of mGOCUG/CPT-11/shRNA exhibited excellent antitumor efficacy in orthotopic brain tumors, which presents a versatile mGO-based drug and gene delivery system for effective combination therapy in brain tumors.

Notably, efforts to incorporate multiple organic and inorganic nanomaterials as potential therapeutic agents have also been made. A theranostic liposome incorporating superparamagnetic IONPs, quantum dots and cilengitide has been developed for optical imaging and as a therapeutic drug delivery system (144). Wu et al (145) have also created an aqueous, surfactant-free ferrofluid composed of superparamagnetic IONPs coated with silicate mesolayers and carbon shells. It was found that this NP decreased the viability of brain tumors more than that of the respective primary analogues of the NP. Conjugated polymer NPs have been designed and synthesized by integrating a metal oxide magnetic core (Fe3O4 and NiFe2O4 NPs; 5 nm) into their matrix during the nanoprecipitation process (146). This incorporation may be used for simultaneous magnetic resonance and fluorescence imaging in brain tumors.

The effectiveness of NP-based delivery systems indicates the potential to develop more effective and personalized combination therapies for patients with brain tumors. However, despite these advancements, drug carriers with controllable size, integrated targeting function, appropriate structure and the ability to cross biological barriers must be developed and further optimized in the future.

Availability of data and materials

Not applicable.

Authors' contributions

JXS reviewed the literature and wrote the first draft. ZCL produced the figures. FG and XJ collected part of the data. YYM supervised and revised the manuscript. JXS, ZCL, FG, XJ and YYM were involved in writing the paper. All authors read and approved the final version of the manuscript. Data authentication is not applicable.

Ethics approval and consent to participate

Not applicable.

Patient consent for publication

Not applicable.

Competing interests

The authors declare that they have no competing interests.

Acknowledgements

Not applicable.

Funding

This work was supported by the Medical Science and Technology Project of Zhejiang Province (grant nos. 2023KY465 and 2021KY529).

References