Safety considerations for nanoparticle gene delivery in pediatric brain tumors

Abstract

Current standard of care for many CNS tumors involves surgical resection followed by chemotherapy and/or radiation. Some pediatric brain tumor types are infiltrative and diffuse in nature, which reduces the role for surgery. Furthermore, children are extremely vulnerable to neurological sequelae from surgery and radiation therapy, thus alternative approaches are in critical need. As molecular targets underlying various cancers become more clearly defined, there is an increasing push for targeted gene therapies. Viral vectors and nonviral nanoparticles have been thoroughly investigated for gene delivery and show promise as vectors for gene therapy for pediatric brain cancer. Here, we review inorganic and organic materials in development for nanoparticle gene delivery to the brain with a particular focus on safety.

Malignant CNS tumors comprise the majority of solid tumors in children [1]. Although the 10-year survival rate for children with CNS tumors is 68.2%, this survival rate varies drastically across over 100 subtypes of CNS tumors [2]. For example, the 10-year survival rate for children with pilocytic astrocytoma is 96.2%, but just 26.2% for atypical teratoid rhabdoid tumors (AT/RT) [2]. This heterogeneity in tumor subtype and prognosis for pediatric brain cancers highlights the need for safer and more effective treatment. For many CNS tumors, the standard treatment is surgery followed by radiation and/or chemotherapy. However, these treatments often have a host of adverse side effects that are especially problematic for neurological development in children [3]. Certain types of tumors, such as diffuse intrinsic pontine gliomas (DIPGs) which make up to 75–80% of brainstem tumors, are not amenable to surgical resection [4] highlighting the need for safe targeted therapies.

Nanomedicines, including nanoparticle-mediated gene therapy and oncolytic viruses, are promising advanced therapeutics for pediatric brain tumors. An increasing body of literature has begun to show the benefits of using nanoparticles to deliver gene therapy in several diseases, including cancer. However, as this new technology continues to develop, safety considerations are paramount to ensure that these preclinical findings can translate to patient care. In this review, we describe key nanomaterials, summarized in Table 1, and routes of administration for therapeutic gene delivery and discuss advantages and disadvantages of each with respect to safety in pediatric brain cancer.

Table 1. . Summary of safety advantages and disadvantages for vectors discussed.

	Vector	Safety advantages	Safety disadvantages	Ref.
Viral gene delivery	Retrovirus	Safe to use for ex vivo transfection	Integration into host genome may lead to insertional mutagenesis	[5,6]
	Adenovirus	No integration into host genome	Immunogenic	[6,7]
	Herpes simplex virus	No integration into host genome	Immunogenic	[7]
	Recombinant adeno-associated virus	Genes required for viral replication have been removed	Long-term safety currently unknown	[8]
	Oncolytic virus	Viruses can be engineered to limit pathogenicity and infectivity	High rate of adverse events, long-term safety currently unknown	[9]
Nonviral gene delivery: inorganic materials	Gold nanoparticles	Biocompatible	Potentially cytotoxic	[10]
	Iron-oxide nanoparticles	Polymer coatings improve biocompatibility	Potentially cytotoxic, clearance mechanisms unclear	[11–13]
Nonviral gene delivery: organic materials	Liposomes	Clinically approved vector	Potentially neurotoxic in vivo when administered to brain locally	[14,15]
	Lipid nanoparticles	PEGylation decreases immunogenicity in brain parenchyma	High levels of nonspecific binding, potentially cytotoxic	[16,17]
	Dendrimers	Biocompatibility improved through surface modifications	Neurotoxicity in glioma cell lines	[18,19]
	PLGA	Biodegradable	Potential local immune response	[20]
	PEI	Biodegradable with linker modifications	Potentially cytotoxic	[21,22]
	PBAE	Biodegradable	Degradation products need to be studied further	[23]

PBAE: Poly(beta-amino) ester; PEI: Polyethylenimine; PLGA: Poly(lactic-co-glycolic) acid.

Pediatric brain tumors

Young children have a higher incidence of embryonal origin tumors such as medulloblastoma (MB) and AT/RT, which portends a worse prognosis than older individuals with brain tumors [24]. In 2016, due to the heterogeneity of disease within tumor groups, the WHO developed a classification scheme for identifying tumors through molecular parameters [25].

Specifically, MB is an embryonal tumor of the posterior fossa and is considered the most common malignant brain tumor in children. The 2016 WHO classification subdivides MB into four subgroups: wingless (WNT), sonic hedgehog (SHH), group 3 and group 4 [26]. Importantly, these tumors have differing prognosis as well as molecular pathways, which require more idiosyncratic treatments [25]. For example, while patients with WNT tumors – characterized by WNT signaling with a CTNNB1 mutation – have a very good prognosis with >90% 5-year overall survival (OS), patients with group 3 tumors, which feature MYC amplification and/or PVT1-MYC fusion, have a 32% 5-year OS. Current standard of care for MB involves maximal safe surgical resection along with cisplatin-based chemotherapy. For children greater than 3-year old, craniospinal radiation markedly increases OS [27]. Unfortunately, there are several neurocognitive and neuroendocrine sequelae that occur from radiation treatment, with children less than the age of three often precluded from this therapy due to high morbidity.

While AT/RT only account for about 2% of all pediatric brain tumors, their prevalence rises to 20% of CNS tumors in patients younger than 3-year old [28]. Similar to MB, AT/RT features significant molecular heterogeneity with three molecular subtypes that have different prognoses: TYR, SHH and MYC. Treatment options for AT/RT are limited with significant side effects and prognosis is overall dismal [29]. One of the barriers to treatment involves the young age of presentation, which limits the use of radiation due to neurological sequelae.

In older children, high-grade gliomas such as anaplastic astrocytoma, DIPG and glioblastoma (GBM) become more prevalent [30]. Of note, DIPG presents with multiple cranial neuropathies with median OS at 9 months [31]. The current mainstay of treatment for DIPG is radiation therapy, though recent studies have exploited multiple molecular pathways, including mutations of H3K27M, PDGFRA, PI3K and ACVR1 [24].

Compared with treatment of adult brain cancer, one of the greatest challenges to treating pediatric brain tumors is the limited therapeutic options. This limitation is in part due to ongoing neurodevelopment in children, which makes them more vulnerable to the adverse effects of standard treatment options such as radiation, which can cause nearly untenable morbidities despite improved survival [32]. Therefore, there is an urgent need for new therapeutic approaches as safe and effective alternatives to the current methods of care. Precise biological molecular therapy, especially gene therapy, has strong potential for unlocking new avenues for pediatric CNS tumor treatments. Recent studies using small molecule drugs to target genetic pathways have shown promise in treating several pediatric CNS tumor subtypes, elucidating genetic targets that can be modulated more specifically and with fewer potential side effects using targeted gene therapy [33]. Potential gene therapy targets include PDGFRA downregulation to treat DIPG [34], AURKA downregulation or INI1 tumor suppressor gene upregulation to treat AT/RT [35] and BRAF downregulation to treat pediatric low-grade gliomas [36]. Despite these promising recent advances, very few studies on gene therapy for pediatric brain cancers have been published in part due to the lack of animal models that can accurately recapitulate the disease phenotype of a diverse array of pediatric CNS malignancies (see Future perspective). In this review, we present the current research on gene delivery to the CNS to provide valuable insights to the nascent field of gene therapy to treat pediatric brain cancer.

Viral gene delivery

Nanoparticulate systems for gene therapy are divided into two main classes: viral and nonviral. Viruses have evolved for millions of years to be highly efficient at transducing mammalian cells, making them an attractive vector for delivering a cancer-killing gene or providing an oncolytic function. However, safety concerns over vector-mediated immunogenicity and toxicity, as well as limited cargo carrying capacity and difficulty in large scale manufacturing remain significant obstacles to their use to treat pediatric brain tumors [7].

Retroviruses are single-stranded RNA viruses that integrate into the host genome and provide long-term expression. The integrative nature of retroviruses can translate to pathogenic outcomes resulting from insertional mutagenesis in the target cells [6]. In comparison, adenoviruses are double-stranded DNA viruses that do not integrate into the host genome [7]; but the viruses are still pathogenic and immunogenic and therefore progress in developing adenoviral-based therapeutics has been slow [6]. Viral vectors derived from herpes simplex virus (HSV) have a very large genetic carrying capacity relative to retroviral and adenoviral vectors, and do not integrate into the host genome. Despite the advantage of large capacity, progress with these HSV viral vectors has also been slow due to immunogenicity of the vectors [7].

Recombinant adeno-associated viruses (rAAVs) have become increasingly popular vectors and are a derivative of wild-type AAV that no longer contain sequences encoding viral proteins. This removal of significant genetic information allows much of the approximately 5-kb rAAV capacity to be replaced with cassettes encoding proteins with therapeutic potential [8]. The removal of genes required for viral replication in particular has also enhanced the safety profile of rAAV vectors. The first AAV gene replacement therapeutic, Luxturna (voretigene neparvovec-rzyl), was approved by the US FDA in 2017 for treatment of patients with RPE65-mediated inherited retinal dystrophy [37]. The second, Zolgensma (onasemnogene abeparvovec-xioi), was approved in 2019 for treatment of pediatric patients with spinal muscular atrophy [38]. These landmark approvals will undoubtedly pave the way for further development of rAAV-based gene therapeutics, though Phase IV trials are still in progress for both Luxturna and Zolgensma as the long-term effects of viral-based gene delivery are still unknown. Despite advances in viral gene delivery, and in particular therapies utilizing rAAV, no treatments are currently clinically available for brain tumors, and those that have reached Phase I trials have not reported data [39].

In addition to their use as gene-delivery vehicles, several oncolytic viruses have also been investigated for direct killing of tumor cells. An early study by Kramm et al. used intrathecal administration of HSV to deliver the HSVtk gene to a disseminated rat gliosarcoma model; the authors demonstrated long-term survival in 90% of treated animals due to the combined effects of suicide gene therapy via ganciclovir activation as well as oncolytic activity of the vector [40]. Studebaker and coworkers published a series of studies demonstrating the efficacy of oncolytic measles virus [41,42] and the oncolytic herpes virus [43] in treating orthotopic xenografts of MB or AT/RT in mice. Several clinical trials using HSV, adenovirus, and modified measles virus to treat MB, AT/RT and/or glioma are currently underway [44]. It is important to note that most of the preclinical studies have been done on immunodeficient animals, making it difficult to assess potential adverse immune responses. Long-term safety of viral treatment is another major concern as children in a clinical trial using retroviral gene therapy to correct X-linked severe combined immunodeficiency developed leukemia nearly 4 years after treatment, despite not exhibiting any adverse symptoms initially [45].

Non-viral nanoparticle delivery

Given the inherent risks associated with viral vectors, alternative materials have been explored for gene delivery. Nanoparticles, which are 1–1000 nm in size, have been developed using a variety of biomaterials that can be broadly categorized into inorganic and organic subsets. Of the organic materials investigated thus far, lipid-based nanoparticles have advanced the furthest with the first therapeutic now commercially available [46]. Recent advancements in polymeric systems, and in particular bioreducible nanoparticles, show particular promise as nonviral gene vectors. Nonviral gene-delivery vectors are generally designed to mimic the sizes of viral particles, approximately as spheres that are 20–100 nm in diameter.

Inorganic materials

Gold nanoparticles (AuNPs) have been explored for applications in drug delivery, gene delivery and imaging. In the context of gene delivery, AuNPs have been functionalized to deliver siRNA, miRNA and plasmid DNA in vitro, as reviewed by Ding et al. [10]. AuNPs delivering siRNA against oncogenes have been shown to decrease tumor burden and improve survival in mice with GBM multiforme [47]. While these delivery systems are advantageous because of their biocompatibility and tunable characteristics, potential cytotoxicity, immunogenicity and long-term biodistribution need to be explored further [10].

Iron-oxide nanoparticles have also been studied for drug delivery, gene delivery, imaging and radiotherapy. These particle systems are often formulated in conjunction with a polymer coating to improve biocompatibility and have been used to deliver plasmid DNA in vitro [11]. Assessment of iron-oxide nanoparticle safety has been complicated. Soenen et al. describe a variety of safety parameters to test for iron-oxide nanoparticle formulations in vitro, leading to the conclusion that cytotoxicity can arise from a variety of processes and can vary widely between formulations [12]. A critical challenge with nondegradable inorganic nanoparticles is their clearance or potential accumulation. While nanoparticles smaller than approximately 6 nm can be filtered through glomerular filtration in the kidney and cleared from the body in urine, larger nanoparticles, especially those with significant surface charge, can accumulate in tissues over long time periods [13].

Organic materials

Liposomes are soft particles that are formed by a lipid bilayer resulting in an aqueous core. These particles can be used to entrap hydrophobic or hydrophilic cargo and show increased cell permeability with integration of cholesterol and other phospholipids in the lipid bilayer [14]. The first liposomal-based therapeutic was approved in 1995 for patients with ovarian cancer and AIDS-related Kaposi’s sarcoma. Since that time, several therapeutics have been approved for treatment of cancer and fungal infections, and for use in vaccine delivery [14]. Several liposomal-based treatments are also in clinical trials for delivery of siRNA or plasmid DNA, as reviewed by Bulbake et al. [14].

Lipid nanoparticles (LNPs) are made up of cationic lipids with positively charged headgroups. These particles are efficient in packing negatively charged DNA and interacting with cell membranes, but also exhibit high levels of nonspecific binding and cytotoxicity. Additional components, such as cholesterol or PEG, can be incorporated into cationic LNPs; helper lipids such as cholesterol have been shown to improve nanoparticle stability and intracellular trafficking while PEGylation is well known strategy for immune evasion [17]. The structure and composition of each of the cationic headgroups, linker between headgroup and tail, and tail groups can affect activity of the cationic LNPs. Structure–activity relationships of these modifications are reviewed by Buck et al. [17]. LNPs administered via intravenous injection induce a strong immune response upon reaching the brain parenchyma in mice, though this immune response can be reduced via PEGylation of the particles [16]. An important advancement for LNPs was the approval of Onpattro in 2018. Onpattro is a LNP siRNA therapeutic for the treatment of polyneuropathies from transthyretin-mediated amyloidosis [46], which highlights the use of LNPs for gene delivery and clinical translation of LNP-based systems.

Dendrimers are highly branched structures synthesized by stepwise addition of structural layers resulting in monodisperse polymers with uniform molecular weight. In the last several decades, dendrimers have been studied in the context of drug and gene delivery. They are advantageous due to their scalability and surface functionality [48]. Although successful gene transfection with plasmid DNA has been shown in vitro [49], translation in vivo has proven challenging and dendrimers today are primarily used as in vitro transfection reagents (SuperFect and PolyFect). Wang et al. have also described the neurotoxic nature of dendrimers in glioma cell lines suggesting they may not be suitable for brain tumor treatment [19].

Poly(lactic-co-glycolic) acid nanoparticles (PLGA NPs) are made of a biocompatible copolymer of polylactic acid and polyglycolic acid. They have been used to deliver small molecules, proteins and peptides, and nucleic acids. PLGA NPs degrade via hydrolysis and have tunable release characteristics [20]. Most progress has been in drug delivery of chemotherapeutic agents, and PLGA has been used in the targeted delivery of drugs across the blood–brain barrier (BBB) such as paclitaxel in an orthotopic U87-Luc glioma mouse model [50], shikonin in rat brains [51] and rotigotine in mouse brains [52]. Delivery of siRNA with PLGA NPs highlighted the material’s slow degradation properties and durable gene knockdown in vivo [53].

Soft polymeric nanocomplexes formed by cationic polymers and anionic nucleic acids, often termed polyplexes, have also been explored for gene delivery. Polyethylenimine (PEI) is an amine-rich cationic polymer that has been shown to strongly bind DNA and buffer pH within the endosomal compartment to promote endosomal escape. These particles can efficiently condense DNA and have been shown to successfully transfect a variety of cell types, including primary neurons, as well as newborn mouse brains in vivo [54]. Recent efforts show that PEI NPs are easily scalable and can be generated via flash nanocomplexation for more uniform particles [55]. Although PEI was originally hypothesized to be quite safe due to its extensive previous use in cosmetics and other products, cytotoxicity has been more recently reported in in vitro and in vivo applications [22].

Poly(beta-amino ester) (PBAE) nanoparticles are both cationic like PEI and hydrolytically degradable like PLGA. They can be modified by varying the constituent amine or acrylate groups used during polymer synthesis as well as the terminal end-capping group, giving rise to a large library of polymers to use to form nanoparticles with nucleic acids. The biodegradable nature of PBAEs makes them an attractive option for drug and gene delivery [23]. Karlsson et al. demonstrated that systemic delivery of reporter gene siRNA with bioreducible PBAEs was effective in vivo at knock down in orthotopic tumors derived from human GBM1A implanted intracranially in mice. There was no toxic or adverse effect observed by tissue histology or on liver function [56] and these data highlight the potential safety profile of PBAEs. From a treatment perspective, bioreducible PBAEs encapsulating two miRNAs, miR-148a and miR-296-5p, and administered intracranially were found to inhibit brain cancer stem cells and lead to long-term survivors in orthotopic human GBM xenografts in mice [57]. These nanoparticles were further used to deliver miRNA constructs inhibiting miR-486-5p, which was found to be overexpressed in adult GBM and enhanced resistance to ionizing radiation treatment [58]. PBAE nanoparticles downregulating miR-486-5p in combination with radiation significantly reduced tumor volume in orthotopic murine models of patient-derived GBM tumors compared with nanoparticle or radiation alone without causing significant histological changes in the brain [59]. These studies demonstrate that gene delivery nanoparticles can be used to enhance the cytotoxic effects of conventional treatment such as ionizing radiation. Mangraviti et al. showed increased survival in rats with implanted gliosarcoma locally treated with nonbioreducible PBAE-delivered HSVtk DNA and systemic ganciclovir delivery [60]. Choi et al. showed how this work could be extended to a pediatric context in mice implanted with AT/RT or MB (Figure 1) [61]. Biodegradable nanoparticles for intracellular gene delivery are an exciting emerging approach for the safe and effective treatment of brain tumors, including pediatric cases.

Figure 1. . Biodegradable poly(beta-amino ester) nanoparticles enabled suicide gene therapy in two pediatric brain cancer models.

(A) PBAEs self assembled into nanoparticles with plasmid DNA encoding the HSVtk gene. Upon nanoparticle CED into orthotopic brain tumors in mice, systemically administered prodrug ganciclovir was activated by transfected cells, enabling tumor cell killing. (B) The HSVtk gene therapy significantly improved survival in both the BT-12 AT/RT as well as D425 medulloblastoma models.

AT/RT: Atypical teratoid rhabdoid tumor; CED: Convection-enhanced delivery; GCV: Ganciclovir; NP: Nanoparticle; PBAE: Poly(beta-amino ester).

Reproduced with permission from [61], © Elsevier Inc. (2019).

Methods of delivery for therapy

Delivery of therapeutics to the brain is limited by the BBB, a tight endothelial cell layer that prevents the diffusion of nearly all substances except some small lipophilic molecules. Due to limited transport across the BBB, many therapeutics do not reach the brain and thus need to be administered directly [62]. There are increasing methods of augmenting systemic delivery, including techniques like ultrasound destabilization which induces transient permeability in the BBB by disrupting tight junctions [63]. Delivery of nanoparticles to the brain parenchyma in mice was improved in a pressure-dependent fashion using focused ultrasound and microbubbles [64]. Further utility of ultrasound was described by Airan et al. demonstrating that with delivery across the BBB, payload release can be modulated via focused ultrasound allowing for better localization of therapeutic delivery [65].

One promising method to bypass the BBB altogether includes using an intra-operative adjunct that allows for direct intratumoral exposure. One of the current standards of care for glioblastoma involves using wafers loaded with carmustine. This platform not only reduces systemic toxicities but also focuses its cytotoxic effect in the tumor microenvironment [66]. In a similar fashion, nanoparticles have the potential to be loaded onto platforms that can be placed into the site of the tumor intra-operatively. Promising preclinical studies using hydrogels for local delivery of siRNA via PBAEs in a breast cancer model [67] and with AuNPs in a colon cancer model [68] highlight more recent advances in sustained local delivery with hope that such methods could also someday be applied to brain tumors.

Additionally, systems such as Ommaya reservoirs also offer a route of CNS penetration to the cerebrospinal fluid, though this may be less effective for intraparenchymal disease. However, this technology is already in clinical use for the delivery of chemotherapeutic agents and can be co-opted to improve the transport of nanoparticles [69]. Moreover, convection-enhanced delivery (CED) has gained traction as a method for intracranial delivery with targeted delivery to tumors, as seen most recently with treatment of DIPG [70]. CED has shown synergistic potential with nanoparticle delivery in animal models, further supporting its use as a delivery mechanism [60], including for use with pediatric tumors [61].

After bypassing the BBB, nanoparticles must also overcome transport barriers posed by the brain extracellular space (ECS) in order to reach tumor cells [71]. Previous studies using the model protein lactoferrin showed that in vivo diffusion through the ECS was reduced by 60% by binding to heparan sulfate proteoglycans [72], highlighting the need to engineer delivery systems to bypass this additional barrier. One approach to facilitate brain ECS transport is to limit the size of nanocarriers; integrative optical imaging studies have shown that the brain ECS contain fluid filled ‘pores’ 38–64 nm in diameter, allowing diffusion of materials under the 64-nm limit [73]. Nance et al. reported that the size limit can be further extended by a dense PEG coating, which enabled nanoparticles as large as 114 nm in diameter to diffuse freely within fresh human and rat brain tissue [74]. To target nanocarriers to tumor cells, peptide ligands or antibodies can be conjugated onto nanoparticle surfaces to facilitate receptor-mediated endocytosis. Common targeting ligands include Angiopep-2 [75], antibodies against neuron-glial antigen 2 [76] and chlorotoxin peptide derived from scorpion venom [77]. These cancer targeting approaches are especially relevant for disease subtypes like DIPG, which extensively infiltrates normal brain tissue and cannot be surgically resected [78].

Conclusion

Safety must always be a primary consideration when assessing new treatment modalities. The pediatric population presents unique challenges due to ongoing development and, potentially, an immature immune system. They are vulnerable to long-standing developmental consequences when treated with radiation at a young age as well as systemic toxicities from treatments. Many emerging types of nanoparticles have shown efficacy for delivery to the brain. A key safety feature needed for these nanoparticles is that they are either small enough to be eliminated from the body or that they break down into smaller nontoxic bioeliminable components that can be easily cleared. The variety of choices now available with nanomaterials, including hydrolytically degradable materials and materials with modular functionality, allows for flexibility to tailor treatment to individual tumors. In particular, PBAEs, which can be both hydrolytically degrading and bioreducible, represent a promising approach for nanomedicine formulations. Although still in preclinical development, they show excellent safety and efficacy, as well as the potential for specific delivery to particular types of tumors. Delivery route is critical to safely and effectively reach the brain with nanoparticle therapeutics and both systemic routes (utilizing targeting agents and/or physical methods of delivery enhancement like focused ultrasound) and local delivery (via CED and pumps) are showing clear progress. Ultimately, there will need to be a combinatorial optimization approach consisting of an engineered nanoparticle designed for efficient intracellular delivery to pediatric brain tumors, molecularly targeted therapeutic nucleic acids of interest, and an optimized delivery route to maximize safety and efficacy in pediatric populations.

Future perspective

Our increasing understanding of the molecular basis of pediatric brain tumors [79] and innovations in materials development for gene delivery have the potential to revolutionize the treatment of these devastating childhood diseases. Looking ahead, one of the major hurdles in the advancement of these multidisciplinary studies is the lack of animal models that accurately recapitulate the disease phenotype of pediatric brain tumors. The majority of the preclinical studies presented in this review were done in rodent models of localized intracranial tumors established using adult GBM cells or pediatric patient-derived cell lines. Unlike adult GBM, however, several pediatric CNS malignancies including MB and AT/RT present with tumor dissemination into the cerebrospinal fluid, which creates additional drug-delivery challenges in addition to being a grave negative prognostic factor [80]. Current models of disseminated pediatric tumors are established by injecting human pediatric cancer cells into the lateral ventricle [41,42]. While this approach recapitulates the genotype and tumor growth patterns of human pediatric brain tumors, the xenografts must be established in immunocompromised animals, which crucially lacks interactions with the immune system. Recent studies harnessing the power of the immune system have shown promise in discovering new treatment paradigms for adult GBM [81,82]. Innovations in gene and immunotherapy applied to clinically relevant animal models will be a crucial next step for research into therapeutics for pediatric brain cancers.

Executive summary.

Pediatric brain tumors

• Malignant CNS tumors comprise the majority of solid tumors in children. Standard treatment for CNS tumors is surgery followed by radiation and/or chemotherapy.

• Diffuse and infiltrative tumor types are not amenable to surgical resection and use of radiation treatment in pediatric cases is limited due to development of neurological sequelae. Safe, targeted therapies are needed.

Viral gene delivery

• Viral vectors have shown success in gene delivery and to date two recombinant adeno-associated viruses have been approved for clinical use, though recombinant adeno-associated viruses have not yet been successfully applied to therapeutics in the brain.

• There are currently multiple clinical trials investigating oncolytic viruses for the treatment of medulloblastoma, atypical teratoid rhabdoid tumor and/or glioma.

Non-viral nanoparticle delivery

• Inorganic nanoparticles, such as gold and iron-oxide nanoparticles, have been explored for gene delivery, though most utility for these materials has been for imaging purposes. Clearance of inorganic materials remains a concern.

• Liposomes and lipid nanoparticles have had clinical success as delivery vectors of small molecules and siRNA, respectively, though immunogenicity of lipid nanoparticles remains an important consideration for further development.

• While dendrimers have been very effective tool for gene delivery in vitro, neurotoxicity in glioma lines suggest they may not be suitable for use in the brain.

• Polymeric nanoparticles have been used to deliver a variety of nucleic acids in vitro and in vivo and are a promising vector for safer gene delivery owing to the bioreducible nature of several polymer classes.

Methods of delivery for therapy

• Delivery of therapeutics to the brain is limited by the blood–brain barrier requiring many therapeutics to be administered directly.

• Focused ultrasound allows for efficient delivery of nanoparticles to the brain and can be utilized for local delivery.

• Hydrogels containing nanoparticles, placed intra-operatively, have shown sustained therapeutic delivery in breast and colon cancer models, though this method has not yet been applied to the brain.

• Reservoirs, already commonly used to administer chemotherapeutic agents, could be co-opted for delivery of nanoparticles taking advantage of convection-enhanced delivery techniques.

• Several approaches to aid transport through the brain extracellular space have been proposed including limiting nanocarrier size and utilizing targeting peptide ligands or antibodies.

Future perspective

• Lack of animal models for pediatric brain tumors remains a critical roadblock to therapeutic development.

• Combinatorial optimization of engineered nanocarriers in conjunction with targeted delivery mechanisms is needed to maximize safety and efficacy.