Nonviral nanoparticle gene delivery into the CNS for neurological disorders and brain cancer applications

Abstract

Nonviral nanoparticles have emerged as an attractive alternative to viral vectors for gene therapy applications, utilizing a range of lipid-based, polymeric, and inorganic materials. These materials can either encapsulate or be functionalized to bind nucleic acids and protect them from degradation. To effectively elicit changes to gene expression, the nanoparticle carrier needs to undergo a series of steps intracellularly, from interacting with the cellular membrane to facilitate cellular uptake to endosomal escape and nucleic acid release. Adjusting physiochemical properties of the nanoparticles, such as size, charge, and targeting ligands, can improve cellular uptake and ultimately gene delivery. Applications in the central nervous system (CNS; i.e., neurological diseases, brain cancers) face further extracellular barriers for a gene-carrying nanoparticle to surpass, with the most significant being the blood–brain barrier (BBB). Approaches to overcome these extracellular challenges to deliver nanoparticles into the CNS include systemic, intracerebroventricular, intrathecal, and intranasal administration. This review describes and compares different biomaterials for nonviral nanoparticle-mediated gene therapy to the CNS and explores challenges and recent preclinical and clinical developments in overcoming barriers to nanoparticle-mediated delivery to the brain.

This article is categorized under:

Therapeutic Approaches and Drug Discovery > Nanomedicine for Neurological Disease

Therapeutic Approaches and Drug Discovery > Emerging Technologies

Nanotechnology Approaches to Biology > Nanoscale Systems in Biology

1 |. INTRODUCTION

Gene therapies aim to treat diseases by replacing or modifying defective proteins, altering protein expression, or introducing therapeutic proteins, either transiently or long-term. As the molecular mechanisms underlying many diseases continue to be uncovered, novel targets amenable to gene therapy techniques are revealed. Moving beyond transient expression of genetically encoded functional proteins or transient knockdown of defective proteins, gene editing techniques such as Zinc finger nucleases (ZFNs), transcription activator-like effector nuclease (TALENs), and most excitingly CRISPR/Cas are all examples of new biotechnological methods to edit genes for long-term functionality (Khalil, 2020; Khan, 2019). In some disease profiles, it may be beneficial to express more of a certain protein, such as CAL (cystic fibrosis transmembrane conductance regulator-associated ligand) to treat Parkinson’s disease (W. Y. Luo et al., 2019). Directly injecting the protein of interest is challenging due to issues such as protein instability, toxicity, or poor biodistribution profiles (Goswami et al., 2019; Sung & Kim, 2019). Alternatively, reduction of disease-causing proteins may be favorable, as is the case of silencing CHCHD10, a mitochondrial protein linked to amyotrophic lateral sclerosis (Burstein et al., 2018). Delivery of genetic material encoding the protein of interest is an attractive alternative when proteins cannot be delivered directly and/or when the target of action is intracellular. Gene delivery approaches also have great promise for next-generation vaccines, as have been highlighted by the response to the COVID-19 pandemic. Additionally, gene delivery technology can also be used for epigenetic engineering to make lasting changes to gene expression without making a permanent change to a host’s genome sequence.

The central challenge for gene therapy has been safe and effective delivery. Naked nucleic acids are susceptible to degradation by nucleases found extracellularly and in the cytosol (Wadhwa et al., 2020). Cellular uptake of nucleic acids is poor, and they are immunostimulatory (Tzeng & Green, 2018). Scientists, engineers, and clinicians have worked on multiple approaches to gene therapy spanning from physical techniques, such as electroporation (Sokołowska & Błachnio-Zabielska, 2019), to the use of viral vectors, such as adeno-associated virus (AAV; Darrow, 2019), to chemical approaches, such as nanoparticles (S. Zhang, Cheng, et al., 2021). The majority of gene therapy clinical trials, approximately 70%, have utilized viral methods of delivery (S. Ghosh et al., 2020). Viruses, which can be considered as biological nanoparticles, serve as attractive vehicles for intracellular delivery because they have evolved the mechanisms necessary to transfect human cells efficiently (Shirley et al., 2020). Viral gene therapy vectors include AAV, adenovirus, retrovirus, lentivirus, and herpes simplex virus-1. Among these, AAV has been the most used vector in FDA approved gene therapies (Bulcha et al., 2021). These therapies include Glybera, a treatment for lipoprotein lipase deficiency, Luxturna, a treatment for Leber congenital amaurosis, and Zolgensma, a treatment for spinal muscular atrophy (Keeler & Flotte, 2019).

However, there are major drawbacks associated with viral gene therapies including immunogenicity, limited cargo size, and manufacturing expense (Sainz-Ramos et al., 2021). Viral vectors may be recognized by the immune system through pattern recognition receptors, or a patient may already have neutralizing antibodies for a viral antigen, triggering an immune response in both situations and limiting the ability for redosing (Shirley et al., 2020). The packaging capacity of some viral vectors can also be as small as five kilobases, making it difficult to deliver larger genes of interest (S. Ghosh et al., 2020). Manufacturing can also be costly; for example, Glybera was approved by the EMA in 2012 and priced at over $1.2 million per patient, but has since been removed from the market due to its costliness and low use (Shahryari et al., 2019). Nonviral nanoparticles address many of the concerns of viral vectors and will be discussed in depth in this review. While there are many applications for gene delivery nanoparticles, spanning uses from rare monogenic disorders to common infectious diseases, the focus of this review will be on applications to the central nervous system (CNS) for treating neurological disorders and brain cancer.

2 |. NONVIRAL NANOMATERIALS IN GENE THERAPY: LIPID-BASED, POLYMER-BASED, AND INORGANIC NANOPARTICLES

Nonviral particles are attractive gene delivery materials, due to lower immunogenicity, higher genetic payload, ease of bulk manufacturing, and lower costs, as compared to their viral counterparts (Gantenbein et al., 2020; Zu & Gao, 2021). There are three broad categories of nonviral nanoparticles used for gene delivery: (i) lipid-based, (ii) polymer-based, and (iii) inorganic nanoparticles (Ramamoorth & Narvekar, 2015).

2.1 |. Lipid-based nanoparticles

Lipid-based nanoparticles (LBNPs) are composed of lipids and nucleic acids that interact to form a particle for intracellular delivery. LBNPs include liposomes, solid lipid nanoparticles (SLNP), and lipid nanoparticles (LNPs; Figure 1a). Liposomes have been studied extensively as vehicles for drug and gene delivery and are made of phospholipids which create a bilayer capable of encapsulating hydrophobic or hydrophilic cargo within the bilayer or aqueous core, respectively (García-Pinel et al., 2019; Tenchov et al., 2021). Tenchov et al. provide a comprehensive review of currently available liposomal formulations for drug delivery (Tenchov et al., 2021). SLNPs and LNPs differ from liposomes in that their lipids do not create a bilayer to form an aqueous core. SLNPs are made of lipids in the solid state, and are much more stable compared to liposomes (Tenchov et al., 2021). LNPs can contain a mixture of lipids in the solid and liquid states including cationic lipids, ionizable lipids, neutral lipids, and cholesterol as common components. Cationic lipids have positive head groups, and lipids such as 1,2-di-O-octadecenyl-3-trimethylammonium-propane (DOTMA) and 1,2-dioleoyl-sn-glycero-3-phosphoethanolamine (DOPE) have been commercialized as reagents for nucleic acid delivery (Balazs & Godbey, 2010; Malone, 1989). The positively charged, hydrophobic head of cationic lipids can bind to negatively charged nucleic acids (Blakney et al., 2019; García-Pinel et al., 2019). Ionizable lipids, such as C12–200, have reduced charge at neural pH to improve the safety profile and are sensitive to changes in pH, making them positively charged in the endosome to lead to endosomal escape (Hajj & Whitehead, 2017; M. Kim, Jeong, et al., 2021). LNPs may also utilize cholesterol, which helps to stabilize particles and provide structural integrity (J. Kim, Eygeris, et al., 2021). Lipids functionalized with polyethylene glycol (PEG) also help provide stability, and are commonly used to increase particle circulation time (Knop et al., 2010; Figure 2a).

FIGURE 1.

(a) General structures of lipid-based nanoparticles: Liposomes, solid-lipid nanoparticles, and lipid nanoparticles. (b) General structures of inorganic nanoparticles: Iron oxide nanoparticles, gold nanoparticles, carbon dots, silica nanoparticles, and spherical nucleic acids

FIGURE 2.

(a) Chemical structures of lipids used in fabrication of lipid-based nanoparticles for gene delivery. (b) Chemical structures of polymers used in fabrication of polymeric gene delivery nanoparticles

Among nonviral nanoparticles, LBNPs have been widely explored due to their low cytotoxicity and potential biode-gradability, making them attractive delivery vehicles from a safety perspective (Hadinoto et al., 2013; Rezigue, 2020). Despite the fact that LBNPs are already in the clinic (e.g., ONPATTRO as a LNP for the treatment of hereditary transthyretin-mediated amyloidosis), recent COVID-19 vaccines have expedited approval and usage of LNPs (Urits et al., 2020). Many more clinical trials are in progress for a range of diseases, from cancers (e.g., melanoma, ovarian, gastrointestinal) to other viral infections (e.g., Zika virus, Rabies; Hou et al., 2021). The primary disadvantages of LNPs are related to the efficiency of systemic gene delivery to non-liver targets and particle stability. Modifying lipid composition or the LNP corona can boost delivery efficiency or tissue targeting, while simultaneously improving parameters such as stability or reduced LNP clearance (Francia et al., 2020). Previous work has demonstrated that including cholesterol in the lipid composition can improve gene delivery efficiency and that coating the LNP surface with hydrophilic molecules can further promote biocompatibility (Das & Das, 2019; Guevara et al., 2020).

2.2 |. Polymer-based nanoparticles

Polymer-based nanoparticles offer their own set of advantages, as compared to lipid nanoparticles. Polymeric nanoparticles can offer increased stability, controlled release, and controlled degradation and elimination of the polymer (Moku, 2021; Rai et al., 2019). Polymers also have advantages in being able to finely tune chemical structure to modulate function as well as facile large-scale manufacture (Gagliardi et al., 2021). Commonly investigated synthetic polymers include poly(lactic-co-glycolic acid) (PLGA; Jo et al., 2020), poly(L-lysine) (PLL; Urello et al., 2020), poly(amidoamine) (PAMAM; Sun et al., 2019), poly(methyl methacrylate) (PMMA; Jaiswal et al., 2019), polyethyleneimine (PEI; Boussif et al., 1995; Ullah et al., 2020), poly(beta-amino ester) (PBAE; Karlsson et al., 2020), and poly(alpha-amino ester)/CART (Blake et al., 2020; T. J. Thomas et al., 2019; Figure 2b). Many of these polymers can be synthesized to have different polymeric architectures, including linear, branched, and dendritic structures.

PLGA is a copolymer consisting of lactic acid and glycolic acid, and adjusting the ratios of these two polymers can shift the degradation rate of the formed nanoparticle (Rezvantalab et al., 2018). Unmodified PLGA nanoparticles have a negative surface (Liang et al., 2011), but cationic modifications can improve membrane binding (Wusiman et al., 2019). PLL consists of repeating amino acid L-lysine and forms positively charged nanoparticles by self-assembly with anionic nucleic acids due to the protonation of charged amine groups on the polymer (Zheng et al., 2021). A significant drawback for PLL nanoparticles is poor endosomal escape and potential cytotoxicity associated with positive charges, though studies have been conducted to discern the relationship between concentration and molecular weight with cytotoxicity (Alinejad-Mofrad et al., 2019). PEI is used in both linear and branched forms and its balance of primary, secondary, and tertiary amines can facilitate nucleic acid binding and cellular uptake (charged primary amines) as well as effective endosomal escape (secondary/tertiary amines capable of endosomal buffering; Y. Li & Ju, 2017). PEI displays high gene delivery efficiency, but experiences drawbacks due to its high positive charge density that can lead to cytotoxicity (Y. Zhang et al., 2018). PBAE polymers are synthesized using diacrylate monomers and amine-containing small molecules (Perni & Prokopovich, 2017) and they break down quickly in physiological conditions via hydrolysis of their ester bonds (Shenoy et al., 2005). PBAEs offer both high gene delivery efficiency as well as lower cytotoxicity compared to PEI and PLL (Sunshine et al., 2012), making them an attractive alternative. PLL and PAMAM are widely explored as dendrimers, highly branched polymers that create a spherical conformation (Rana, 2021). Bare PAMAM nanoparticles also form positively charged dendritic structures useful for delivery and the polymer end-groups can be further modified to display a different charge or biomolecule (J. Zhang, Li, et al., 2021). PAMAM branching structure can modulate delivery properties as well as antibacterial properties (Kheraldine et al., 2021). Unlike the previously discussed polyesters and peptides, PMMA is a nonbiodegradable material and is often utilized as a nanoparticle core (Alves Batista et al., 2020; Jia et al., 2019).

All of these polymeric nanoparticles can have tunable particle surface charge and particle size, and modifications to these physicochemical parameters can also help increase gene delivery (Salameh et al., 2020). Cationic polymers are generally used to form “polyplex” nanoparticles through electrostatic self-assembly with anionic nucleic acids. Nanoparticle positive surface charge has been demonstrated to increase cellular uptake with the negatively charged cell surface and thus increase gene delivery efficiency (Ooi et al., 2020). Notably, the charge-altering releasable transporter (CART) system enables effective intracellular delivery and biodegradability and has positive charges that gradually become neutral, allowing for dissociation of the negatively charged nucleic acid cargo from the polymer (McKinlay et al., 2017). Particle size can also affect cellular uptake and influence transfection efficiency. For PBAE NPs, 70–150 nm diameter particles have been shown to have optimal gene delivery in many cases (Iqbal et al., 2020). Similarly, small PLGA nanoparticles (<100 nm) exhibited higher gene delivery than larger particles in many cell lines (Liang et al., 2011; Prabha et al., 2016). For enhancing blood circulation time, particle size is also a critical design parameter to tune to avoid renal filtration and to minimize immune clearance (Hoshyar et al., 2016). Researchers have also investigated other physical and chemical properties of polymeric nanoparticles and their influence on cellular uptake and delivery efficiency. For example, polymeric ellipsoidal nanoparticle conformations were found to have a 30%–40% reduction in cellular uptake compared to spherical nanoparticles (Ben-Akiva et al., 2020). In a chemical modification approach, introducing guanidine into the structure of PBAE nanoparticles yielded cellular uptake more than 10 times that of PBAE particles without guanidine (Tang et al., 2020), and crosslinking PBAE nanoparticles was found to increase particle stability and improve systemic delivery properties (Karlsson et al., 2021).

The primary disadvantage of polymer-based nanoparticles is insufficient gene transfer efficiency for the desired target and application (Dizaj et al., 2014; G. Lin, Zhang, & Huang, 2015). Often this is due to poor targeting and quick clearance of the nanoparticles from the blood before the nucleic acid has been intracellularly delivered (Alexis et al., 2008). Opsonization can cause phagocytosis of nanoparticles prior to cellular uptake and subsequent gene delivery (Behzadi et al., 2017). However, modifications to the nanoparticle surface can be made to delay nanoparticle clearance (Mitchell et al., 2021). One such modification involves the conjugation of PEG to the nanoparticle (Shi et al., 2021). PEG has been widely explored as a stealth coating on nanoparticles, helping to increase circulation time and reduce nonspecific cellular uptake of nanoparticles, especially when used to construct a dense coating (M. Li et al., 2021). The presence of PEG serves as a barrier, preventing proteins from adhering to the nanoparticle, thus reducing opsonization (Shiraishi & Yokoyama, 2019). For a non-PEG approach, zwitterionic materials can also be used as nanoparticle surface coatings to reduce clearance (Gagliardi et al., 2021; Zhou et al., 2020). Another disadvantage of polymeric materials in nanoparticles includes nonbiodegradability, such as in the case of PEI or PAMAM (Rai et al., 2019; Wen et al., 2009), which can lead to higher cytotoxicity. And while manufacturing costs for polymeric nanoparticles may be less costly, there are still challenges associated with scaling up the production of these nanoparticles, including batch-to-batch consistency or polydispersity; however, recent fluidic processing techniques have been developed to improve manufacturing consistency (Abstiens & Goepferich, 2019).

2.3 |. Inorganic nanoparticles

A third class of nanoparticles used for gene therapy is inorganic particles. These include iron-oxide magnetic nanoparticles, gold nanoparticles, carbon dots (CDs), silica nanoparticles, and spherical nucleic acid nanoparticles (SNA NPs; Figure 1b). These nanoparticles are typically relatively lower cost, easy to produce, and can have well-defined tunable shapes and sizes (Roma-Rodrigues et al., 2020).

Iron-oxide nanoparticles can be approximately 5–200 nm in diameter, depending on the fabrication method, and are hydrophobic until coated with molecules such as peptides or polymers to improve biocompatibility (Ali et al., 2016). These magnetic particles can be targeted via external magnetic fields (Lu et al., 2002), and are frequently used for diagnostic purposes and drug delivery (Mahdavi et al., 2013). Iron-oxide nanoparticles generally accumulate in the liver and spleen (Edge et al., 2016), although more research into the distribution and elimination of these particles is needed to further understand how to tune delivery and minimize toxicity. With respect to CNS disorders, iron-oxide nanoparticles have primarily been explored for diagnostic purposes in diseases such as Alzheimer’s, Parkinson’s, and amyotrophic lateral sclerosis (S. Luo et al., 2020).

Gold nanoparticles can be fabricated at a range of sizes and shapes, including 1–150 nm spherical particles for drug and gene delivery (P. Ghosh et al., 2008; J. Peng & Liang, 2019). An advantage of gold nanoparticles is that they can be easily surface-functionalized using thiol chemistry to flexibly couple biomolecules to the particle surface. These have been shown useful for drug loading, to enable PEGylation for increased particle circulation time, and to incorporate targeting ligands to enrich cell-specific delivery in vivo (P. Ghosh et al., 2008). Gold nanoparticles can also enable theranostic and combination therapies for both in vitro and in vivo applications, although they must be functionalized to facilitate their intracellular delivery (Ding et al., 2014). CDs are carbon-based particles with sizes less than 10 nm that have been explored for use in imaging applications as well as for drug and gene delivery (Biswal & Bhatia, 2021). CDs are photoluminescent, which is an attractive property for imaging purposes, and they have demonstrated low cytotoxicity, which is favorable for gene delivery applications (Biswal & Bhatia, 2021; Cao et al., 2018).

Mesoporous silica nanoparticles are made from colloidal silica and are used for drug delivery and diagnostic purposes (Janjua et al., 2021). These particles are typically fabricated to be 20–500 nm, have pores that range in size from 2 to 50 nm (Y. Yang et al., 2020), and have demonstrated safety and bioavailability in humans (Bukara et al., 2016). These nanocarriers have widespread use in drug delivery and are increasingly been investigated for nucleic acid delivery. Organic materials have also been explored for use in combination with inorganic particles. Both gold nanoparticles and carbon dots have been engineered to incorporate PEI to improve functionality for nucleic acid binding and endosomal escape (Liu et al., 2012; M. Thomas & Klibanov, 2003). Gold nanoparticles have also been further developed as particle cores that are coated layer by layer with polyelectrolytes, including both PEI and PBAE polymers, to enable simultaneous delivery of siRNA and DNA (Bishop et al., 2015). PEGylation has also been used with gold nanoparticles to increase particle circulation time (P. Ghosh et al., 2008) and with carbon dots as a passivation agent (Z. Peng et al., 2020). Prior work has demonstrated that modifying iron-oxide or gold nanoparticles with amphipathic molecules can promote biosafety (Roma-Rodrigues et al., 2020). However, this field is still young and much more research on cytotoxicity and biocompatibility is required before these nanoparticles can be translated into clinical application.

A final type of nanoparticle studied for nucleic acid delivery is SNA NPs. SNA NPs have a center, either solid or hollow, with nucleic acids radiating outwards in spherical geometry. SNA NPs are frequently built using an inorganic core, such as a gold nanoparticle core, and have been demonstrated as useful vehicles for gene therapy (Rosi et al., 2006). Recent SNA research has demonstrated that cores can be built from varied biomaterials including iron oxide, quantum dots, silica, and platinum (Kapadia et al., 2018). SNA nanoparticles can be designed to cross the BBB, have been used for delivery to the brain (Jensen et al., 2013), and have been assessed in a clinical trial delivering siRNA to glioblastoma (GBM; Kumthekar, 2020). A phase 0 clinical trial (NCT03020017) of NU-0129, consisting of a gold nanoparticle core and Bcl2L12 siRNA directed against the GBM oncogene Bcl2Like12, was found to lead to no significant treatment-related toxicities and the gold nanoparticles were detected in the brain following systemic administration, reaching brain cancer cells as well as tumor-associated endothelial cells and macrophages (Figure 3; Kumthekar, 2020).

FIGURE 3.

Silver staining of pretreatment and posttreatment tumors from two brain cancer patients (105 and 108) treated with SNA NPs, revealing gold in brain tumor cells and tumor-associated endothelial cells and macrophages. (a, d) silver staining of pretreatment tumors in patients 105 and 108. (b, c, e, f) silver staining of tumor sections post-NU-0129 treatment. In tumors resected after NU-0129 administration, Au was present within endothelial cells (e, orange arrowhead), tumor cells (b, c, f; black arrow head), and macrophages (b, e, red asterisk). Cell shown in panel f demonstrates therapy-related nuclear atypia typical for a post-therapy GBM tumor cell (panel f). Reproduced with permission from Kumthekar (2020)

Although several advances have been made to target inorganic nanoparticle localization to the brain, biodistribution and neurotoxicity remain a challenge (Guo et al., 2021). Delivery of black porous silicon nanoparticles intravenously injected exhibited low biodistribution to the brain, as compared to the intestines, kidneys, liver, or other organs (Tamarov et al., 2021). Oligonucleotide nanoparticles also have difficulty crossing the blood–brain barrier (Mendes et al., 2022). Although some studies have demonstrated accumulation of smaller gold particles in the brain, neurotoxicity remains a concern in such cases (De Matteis, 2017). Inorganic nanoparticles may disrupt the BBB or neuronal functions, making neurotoxicity another critical factor to consider during nanoparticle manufacturing (Teleanu et al., 2019). One toxicological study found a significant decrease in brain weight when gold nanoparticles were repeatedly delivered intravenously to mice (L. Yang et al., 2017). New tools continue to be developed to address these issues of bio-distribution and neurotoxicity. Research groups are investigating the effect of nanoparticle shapes on delivery to cerebral endothelium as well as exploring the effect of particle surface charge of cadmium or iron-based nanoparticles on neuronal targeting (Da Silva-Candal et al., 2019; Dante et al., 2017). It is critical to continue investigating these gaps to improve biodistribution and reduce neurotoxicity for patients.

3 |. MECHANISM OF NANOPARTICLE-MEDIATED INTRACELLULAR DELIVERY

Given a desired genetic modification as a goal, the dominating challenge is how to safely and effectively deliver the needed nucleic acid into the cell (Andresen & Fenton, 2021; Vaughan et al., 2020). Nucleic acid uptake is restricted by transport through the extracellular matrix (ECM) and into the cell (Sokołowska & Błachnio-Zabielska, 2019). Gene delivery can be broken down into four main components: (1) membrane binding and cellular uptake, (2) intracellular trafficking and endosomal escape, (3) particle degradation and nucleic acid release, and in the case of DNA, (4) nuclear import and gene expression (Figure 4; Table 1).

FIGURE 4.

Schematic of the intracellular delivery mechanism and trafficking of DNA during gene delivery

TABLE 1.

Comparison of lipid-based, polymeric, and inorganic nanoparticles and their modifiable physiochemical properties affecting delivery

Formulation	Modifiable physiochemical properties	Advantages	Disadvantages
Polymeric	Surface charge or particle size/shape—increase gene delivery Guanidinylation—increase cellular uptake Polymer crosslinking—increase particle stability, systemic delivery PEGylation—delay nanoparticle clearance, reduce nonspecific cellular uptake of nanoparticles Zwitterionic material coatings—reduce clearance	Increased stability Controlled release Controlled degradation and elimination of polymer Tunable structure Ease of large-scale manufacturing	Poor targeting Quick nanoparticle clearance Nonbiodegradability for some polymers (PEI, PAMAM)
Lipid-based	Lipid composition—improve delivery efficiency Hydrophilic coating—promote biocompatibility PEGylation—boost stability and circulation time	Low cytotoxicity Biodegradability	Particle stability Efficiency of systemic delivery to non-liver targets
Inorganic	Surface functionalization—couple biomolecules to surface (targeting, increase circulation time, nucleic acid binding) PEGylation—increase particle circulation time	Lower cost Ease of production Tunable shapes and sizes	Requires more research on cytotoxicity and biocompatibility

3.1 |. Membrane binding and cellular uptake

Nanoparticles can interact with and bind to the surface of cells through multiple molecular interactions. These include nonspecific interactions based on charge (H. Lee, 2021) and specific ligand-receptor binding interactions. Because the glycocalyx, the outer surface of cells composed of glycolipids and glycoproteins, has an overall negative surface charge, positively charged nanoparticles can electrostatically associate with and bind to the cell surface (Möckl, 2020). Together with being able to bind to negatively charged nucleic acids, this is in large part why many of the polymeric biomaterials used for nucleic acid delivery contain amine groups and positive charges (Figure 2b). If the biomaterial does not natively contain positive charges, these can be added through conjugation. For example, gold nanoparticles can be decorated with positively charged peptides (Patel et al., 2019) to aid in membrane association and enable cell penetration (Gong et al., 2019). Rather than focus on positive charge, nanoparticles can be decorated with specific sequences of peptides to improve targeted binding to specific cellular receptors. For example, researchers have shown that attaching a DSPE-PEG nanoparticle to two ligands targeting EGFR and α_vβ₃ integrins, receptors over-expressed on cancer cells, can enable targeting of tumor metastasis (Covarrubias et al., 2019). Multivalent ligand–receptor interactions can enhance membrane binding, and studies have demonstrated that less stiffness in a polymeric nanoparticle could also increase avidity (Farokhirad et al., 2019). Glycocalyx differences in healthy cells versus tumor cells can also attribute to nanoparticle membrane binding differences for cell targeting (von Palubitzki et al., 2020).

Gene delivery nanoparticles can enter the cell through endocytosis pathways or through physically mediated means (Riyad & Weber, 2021). The primary mechanisms for endocytosis utilized include the clathrin and caveolin uptake pathways (X.-X. Ma, Xu, et al., 2020). Research has also demonstrated that by blocking caveolae-mediated endocytosis, cell internalization levels drop dramatically, but that it is the clathrin-mediated endocytosis that is most correlated with successful gene delivery (J. Kim et al., 2014). Further, small changes to polymer end-group structure or molecular weight can tune cellular uptake mechanisms, and thereby improve gene delivery efficiency (J. Kim et al., 2014). Alternatively, cellular uptake can also be by macropinocytosis, nanopore formation, or other noncanonical pathways (Wu et al., 2021). Nanoparticles around 120–200 nm in size are taken up via clathrin and caveolin endocytosis as well as macropinocytosis, while larger nanoparticles are internalized through phagocytosis (Foroozandeh & Aziz, 2018). Although mechanisms of cellular uptake pathways have been well established, novel techniques for visualizing the cellular uptake process into endosomes and subsequent endosomal release have been developed in the last few years, including high-content and high-throughput Galectin-9 or Galectin-8 imaging based platforms (Z. Ma et al., 2022; Munson et al., 2021; Rui et al., 2022). Physical methods can also be utilized such as microinjection or electroporation to facilitate intracellular delivery (Donahue et al., 2019). Nanoparticle characteristics, such as size, charge, or shape, can affect the efficiency of cellular uptake (Fornaguera et al., 2020; Zu & Gao, 2021). A research study utilizing silica nanoparticles loaded with miRNA demonstrated that particles around 160 nm in diameter internalized the fastest, after an incubation time of 45 min (Haddick et al., 2020); however, this may not hold true for all nanoparticles, as speed of cellular uptake could also depend on the material, endocytic pathway, or other qualities. With gold nanoparticles, smaller sizes (25–50 nm) and spherical shapes were internalized more efficiently than larger sizes (>50 nm) and rod, cube, or prism shapes (Carnovale et al., 2019). Another study focusing on lipid nanoparticles found increased cellular uptake in particles that were 30 nm in diameter compared to larger particles closer to 100 nm or 200 nm in diameter (Nakamura et al., 2020). Multiple studies have found lower cellular uptake of ellipsoidal nanoparticles compared to spherical nanoparticles with a range of cells (Ben-Akiva et al., 2020; Champion et al., 2007).

3.2 |. Intracellular trafficking and endosomal escape

Once a nanoparticle has crossed the cellular membrane, it needs to be transported through the cytoplasm and ultimately releases its nucleic acid cargo to the cytosol. Typically, a nanoparticle moves through the endosomal pathway, and without an engineered mechanism for escape, traffics from the early endosome to late endosome, before fusing into a lysosome. It can also be sorted to further organelles, such as the mitochondria or endoplasmic reticulum (Donahue et al., 2019). For a gene delivery nanoparticle to be successful, it needs to escape into the cytosol to avoid degradation in the endolysosomal pathway. There are several mechanisms by which nonviral nanoparticles escape from the early endosome. Nanoparticles with protonatable amines can buffer acidity in the endosome leading to osmotic rupture, and other particles can be functionalized to lead to biomaterial-mediated membrane destabilization (Durymanov & Reineke, 2018; Smith et al., 2018). For lipid nanoparticles, endosomal escape relies on the presence of ionizable lipids (Maugeri et al., 2019). Once acidified in the late endosome, these nonionizable lipids induce shape changes that are associated with membrane fusion (Schlich et al., 2021). With polymeric nanoparticles, experiments support that polymers with titratable amine groups lead to a proton sponge effect, in which protons and Cl⁻ ions flow into the endosome and cause endosomal swelling, contributing to rupture through osmotic pressure (Degors et al., 2019). This mechanism is controversial as other experimental observations have shown that adding amines to increase buffering capacity does not always increase endosomal escape and subsequent transfection (Funhoff et al., 2004), and other studies have shown that acetylation of amines in PEI, which decreases buffer capacity and changes biophysical properties of the nanoparticles, ultimately yielded nanoparticles with significantly improved transfection efficiency (Forrest et al., 2004). Vermeulen et al. give a comprehensive report on the data both supporting and contradicting the proton sponge effect and give insight into experimental methods that allow researchers to probe this phenomenon and the mechanisms involved in nanoparticle endosomal escape (Vermeulen et al., 2018).

Nanoparticles with insufficient endosomal escape can be engineered to increase delivery to the cytosol and subsequent transfection. For example, PLGA enables nucleic acid encapsulation but relatively poor endosomal escape, but this escape is enhanced when PLGA nanoparticles are decorated with PEI, demonstrating the critical role of titratable amines (Tsai et al., 2021). Studies with a library of PBAEs found that their structure, rich in titratable amines that buffer in the physiological range of pH 5–7 (Sunshine et al., 2012), have up to an order of magnitude improved endosomal escape compared to commercially available transfection agents such as PEI (Rui et al., 2022). Apart from chemical properties, nanoparticle physical properties can also play an important role in endosomal escape. For example, in the endosome, peptides can have a triggered shape change to alpha helix, which can lead to membrane destabilization and endosomal escape (Lundberg et al., 2007). Another study designed rod-shaped nanoparticles and found that they exhibited higher rates of endosomal escape compared to spherical nanoparticles (J. Lee et al., 2021).

3.3 |. Nanoparticle degradation and nucleic acid release

Another consideration for the design of nanoparticles is their biodegradability (Chen et al., 2020; Rai et al., 2019). Biodegradable constituents, such as PBAE or PLGA, can help improve safety and reduce toxicity of particles that they make up (Corbo et al., 2016; Essa et al., 2020; Negron et al., 2020; B. Nguyen & Tolia, 2021). Rate of degradation may be affected by nanoparticle shape, density, porosity, or other factors that can be engineered in nanostructures (Moghaddam et al., 2019). Nanoparticles may degrade hydrolytically in water, through disulfide linkages, or by enzyme action. Covalent disulfide bonds are susceptible to reduction, thus their presence in nanoparticles leads to ease of degradation in the intracellular reducing environment of the cytosol and subsequently cargo release (Beaupre & Weiss, 2021). Alternatively, enzyme cleavage by trypsin, lactate dehydrogenase, or other proteases can take place, as demonstrated by nanocarriers such as poly-(isobutylene-alt-maleic anhydride)-graft-dodecyl (PMA) coated nanoparticles (Zhu et al., 2019). Lipid nanoparticles can utilize similar degradation methods, undergoing degradation through disulfide bonds (Qiu et al., 2021) or ester bonds (Kawase et al., 2021), while maintaining delivery functionality.

Nonbiodegradable materials may be bioeliminable or they may accumulate in certain tissues in vivo. For example, spherical gold nanoparticles ~20 nm in diameter were found to accumulate in the liver and spleen, as opposed to the heart, lung, brain, or kidney (Bailly et al., 2019). With essentially all nanocarriers, and especially those with biological components on the surface such as peptides and protein, an immune response can be triggered, leading to antibody production and quick clearance of the particles, especially following redosing.

Ultimately, for gene delivery to be successful, the nucleic acid must be released from the particle to the cytosol. While this can be driven by degradation of a nanoparticle, it can also be driven by the thermodynamics of the nucleic acid unbinding from the biomaterial (Bishop et al., 2013). With polymeric PBAE materials, these binding properties were found to be adjustable through molecular weight, adding carbons to the polymer structure, or modifying polymer end groups (Bishop et al., 2013). It has been demonstrated that multivalent positive charges contribute to stronger binding and slower release of nucleic acids (Schaffer et al., 2000). For gold nanoparticles, release of nucleic acid typically occurs after endosomal escape through the role of glutathione (GSH) in the cytosol (Ding et al., 2014).

3.4 |. Nuclear import and gene expression

In the case that the genetic cargo of a nanoparticle is DNA, the delivery voyage is not over in the cytosol, and the nucleic acid must next enter the nucleus (Shahryari et al., 2021). One commonly explored approach to facilitate this import utilizes nuclear localization sequences (NLS) to direct a plasmid to the nucleus (Bitoque et al., 2021; Maggio et al., 2020). NLSs are recognized by importins, which shuttle the NLS and associated cargo through the nuclear pore. Although the nuclear pore complex can perform ~1000 transports per second, naked nucleic acid is imported much slower than when associated with an NLS (Munkonge et al., 2003) and DNA nuclear targeting sequences can help transport cytosolic DNA into the nucleus (Guen et al., 2021). For genome editing, transfection of rapidly dividing cells, that break down their nuclear membrane frequently, is much easier than with nondividing cells. Further, editing is mediated by the homology-directed repair mechanism, which operates during DNA replication in cell division (Waldron, 2017).

The most essential part after a nanoparticle has trafficked through the cell is expression of the gene of interest. In the nucleus, RNA polymerase must bind to a promoter sequence in DNA to initiate transcription to mRNA. Afterward, in the cytosol, a ribosome must bind to the mRNA to initiate translation (Clancy, 2008). During plasmid design, there are various engineering considerations that can boost gene expression and durability, such as codon optimization, inclusion of enhancers, and exclusion of CPG islands and bacterial elements. Enhancers promote gene expression, so inclusion of these sequences can increase the amount of the target exogenous protein (Arnold et al., 2020). Conversely, CPG sequences are more abundant in bacterial DNA compared to mammalian DNA and can be recognized as foreign by the immune system for silencing and degradation (Mulia et al., 2021). Removal or reduction of these elements can reduce the elimination of the gene of interest that is being delivered.

4 |. CENTRAL NERVOUS SYSTEM ANATOMY AND COMMON NEUROLOGICAL DISEASE PROFILES

4.1 |. Central nervous system anatomy: A brief review

The central nervous system is comprised of the brain and the spinal cord, and its primary function is to process incoming sensory information and send outgoing responses (Ludwig et al., 2021). The major components of the brain include the cerebrum, cerebellum, diencephalon, and brainstem. The cerebrum includes the cerebral cortex and serves many functions, such as movement, speech, thinking, and others (Parkins, 1997). The cerebellum coordinates muscle movement and balance. The diencephalon is found at the center of the cerebrum, and includes the thalamus, hypothalamus, and pituitary gland. Its general function is to regulate bodily functions (e.g., hormones) and maintain homeostasis (Lechan & Toni, 2000). Finally, the brain stem connects the rest of the brain to the spinal cord, and it contains the midbrain, pons, and medulla (J. Zhang, 2019). Tissues in the central nervous system contain nerve cells, or neurons, which allow for the relay of information. An action potential initiates at the cell body of the neuron, and travels down the axon to the axon terminal. This electrical signal releases neurotransmitters into the synapse space between neurons, causing downstream biological responses (Sheffler et al., 2021).

4.2 |. Barriers of entry into the central nervous system

There are two notable membrane barriers to the CNS, the blood-cerebrospinal fluid barrier (BCB) and the blood–brain barrier (BBB). These two barriers exist to maintain homeostasis within the brain (Dotiwala et al., 2021). The brain and spinal cord reside in ~140 ml of cerebrospinal fluid (CSF; Pardridge, 2020a), which is separated from blood via the epithelial cells of the choroid plexus (Tumani et al., 2018). Delivery of drug therapeutics into the CSF needs to cross this barrier, which is relatively leaky compared to the blood–brain barrier, due to the presence of gap junctions and pinocytosis vesicles. The BBB is the barrier between blood vessels and the interior of the brain mass, consists of endothelial cells, and is supported by astrocytes, pericytes, and microglia. These peripheral cells help maintain the BBB structure and stability (Bors & Erdő, 2019). The presence of tight junctions and adherens junctions in the endothelial cell layer severely limits diffusion and pinocytosis through the membrane barrier, making the BBB a significant challenge for gene delivery to the CNS (Bors & Erdő, 2019).

4.3 |. Common central nervous system diseases: Neurological disorders and cancers

As quality of life improves globally and the general population ages, the rate of CNS disorders has also increased (Borumandnia et al., 2021). Neurological disorders are the 5th leading cause of death in the world, with strokes causing over two-thirds of these deaths (Awan et al., 2019). The World Health Organization reports that 15 million people experience a stroke every year (WHO EMRO, 2021). After stroke, the most burdensome neurological disorders include dementias (including Alzheimer’s disease) and migraines (Feigin & Vos, 2019).

Other neurological disorders include epilepsy, Huntington’s disease, spinal muscular atrophy, and Parkinson’s disease (Guekht et al., 2021). Epilepsy is characterized by sporadic seizures, and has been found to be linked to a variety of genes (Perucca et al., 2020), many of which can be targeted with gene therapies in the hippocampus (Riban et al., 2009). Huntington’s disease, caused by a mutation in the huntingtin (HTT) gene, leads to the deterioration of neurons and results in disabilities in movement or thinking (Migliore et al., 2019). Similarly, spinal muscular atrophy is a defect in the SMN1 gene, causing muscle weakness (Keinath et al., 2021). Parkinson’s disease is another neurological disorder linked to specific genes (LRRK2, GBA, and SNCA), manifesting in the form of tremors (Stoker & Barker, 2020), and regions of interest for gene therapies are the substantia nigra in the midbrain, and the caudate nucleus and putamen in the basal ganglia of the cerebrum (Axelsen & Woldbye, 2018). These neurological disorders are due to underlying genetic defects and/or could be modulated by the expression or inhibition of target proteins and therefore are amenable to gene therapy treatments.

A neurological disease that is especially relevant to gene therapy interventions is brain cancer. Brain cancers occur when cells in the brain grow and divide unchecked and is one of the most lethal types of cancer, with a low 2-year survival rate (Aldape et al., 2019). A predictive estimate suggests that up to 85,000 individuals will be newly diagnosed with a brain or CNS cancer in 2021 (Miller et al., 2021). Additionally, brain cancers are the second most commonly occurring cancer in children, after leukemia (Siegel et al., 2020). Brain cancers have also been found to be related to many neurological disorders, including but not limited to stroke, Alzheimer’s disease, Parkinson’s, and multiple sclerosis (Tandel et al., 2019). Tumors in the brain can be primary or recurrent masses, which originate in the brain, or metastatic, originating elsewhere in the body and relocating to the brain. Of primary tumors, most (~70%) were nonmalignant with the rest (~30%) being malignant (Ostrom et al., 2020). Glioblastoma comprises ~49% of all malignant brain tumors, making it the most common malignant tumor. Five-year survival of individuals diagnosed with glioblastoma is bleak, at less than 5% (Baid et al., 2020) and the median survival time is approximately 15 months (Ladomersky et al., 2019; Tamimi & Juweid, 2017). Meningioma makes up 55% of nonmalignant tumors and is the most common nonmalignant tumor (Ostrom et al., 2020). Meningioma has a 5-year survival rate of ~85% and median survival time of over 4 years (Holleczek et al., 2019). Many other brain cancer types exist as well, such as astrocytomas, oligodendroglioma, choroid plexus tumors, medulloblastomas, and hemangiomas (Ostrom et al., 2019), and all are promising candidate diseases for a safe and effective cancer gene therapy.

Existing treatments for malignant brain tumors such as glioblastoma include surgical resection, radiation therapy, and chemotherapy (Tan et al., 2020). However, despite treatments, survival rate remains low, and risks or side effects are abundant. Surgery is an expensive and invasive procedure requiring extensive recovery time, while radiation therapy and chemotherapy is associated with negative side effects, including nausea or pain, which subsequently decrease an individual’s quality of life (Prieto-Callejero et al., 2020). Notably, radiation therapy is a very limited option in pediatric cases, due to neurological sequelae (Armstrong & Sun, 2020). Consequently, nonviral gene therapy presents an attractive, low-cost alternative to current standards of care, especially as new biological mechanisms can be targeted. Beyond delivery of cytotoxic genes, some examples of additional nonviral gene therapy approaches with mechanisms that can suppress brain tumors include nanoparticles delivering miR-486–5p antagomirs or cancer inhibiting miRNAs (Lopez-Bertoni et al., 2018, 2020), delivering factors such as bone morphogenetic proteins to inhibit cancer cell stem cells (Piccirillo et al., 2006), or targeting microcephaly genes (Iegiani et al., 2021).

5 |. ROUTES OF NANOPARTICLE DELIVERY INTO THE CENTRAL NERVOUS SYSTEM

5.1 |. Systemic delivery

Nonviral gene therapy nanoparticles can be delivered into the CNS systemically or locally. Systemic delivery is beneficial in the case that a disease or tumor is not limited to a single or easily accessible location; however, the nanoparticle must then traverse the BBB and may experience quick clearance by the immune system (Mitchell et al., 2021). Delivering nanoparticles, or any other biologic, through the BBB has proven to be the most prominent hurdle in delivery to the CNS (Mitchell et al., 2021). Transport through the BBB is limited to passive diffusion of very small molecules (molecular mass under 400–500 Da, with diffusion increasing as molecular mass decreases within that range) and ions, making transport across the blood–brain barrier very difficult. Lipophilicity also affects BBB crossing, and molecules with log P between 1.5 and 2.5 diffuse most readily. Molecules with high polar surface areas or high electrostatic charges will not readily cross the BBB. These key parameters influencing the ability to cross the BBB can be summed up by Lipinski’s “rule of five” (Mikitsh & Chacko, 2014). Consequently, several challenges exist with crossing the blood–brain barrier to deliver therapeutics to the brain (Tumani et al., 2018). It has been reported that close to 100% of biologic drugs do not cross the blood–brain barrier, and no clinical trials have successfully demonstrated a biologic crossing of this barrier (Choudhari et al., 2021; Pardridge, 2020b). Challenges specific to nonviral nanoparticles crossing the BBB following systemic delivery include low passage efficiency, particle aggregation, or short circulation time (Ahlawat et al., 2020). Some efforts to improve nanoparticle transport include PEGylation, conjugation to transferrin, or conjugation to other targeting ligands (Perez-Martınez et al., 2012). Huang et al. describe systemic delivery of PAMAM-PEG complexes conjugated to Transferrin which led to increased gene expression in the brain compared to unconjugated complexes (Huang et al., 2007). Other researchers, such as Zhang et al., have targeted complexes using an antibody against Transferrin to facilitate BBB crossing (Y. Zhang et al., 2008). Another group employed nanoparticles comprised of oligo(ethylene glycol) with STAT3i siRNA and used iRGD, a tumor penetrating protein, to help the nanoparticle pass the BBB and target glioblastoma tumor cells after systemic injection (Gregory et al., 2020).

5.2 |. Local delivery

Local delivery methods offer the advantage of bypassing the BBB and thus results in higher rates of delivery and less systemic toxicity, but these methods can often be more invasive (Luly et al., 2020). Existing methods for local delivery of nanoparticles into the CNS include intrathecal (IT), intracerebroventricular (ICV), and intranasal injection (Calias et al., 2014; Cerqueira et al., 2020) These methods involve injecting the nanoparticle into the spinal fluid, cerebral cavities, or nose, respectively. A study in 2021 compared the efficacy of each delivery route for an AAV gene therapy, finding the highest delivery through intracerebroventricular injection and the lowest delivery through intranasal injection. Overall, all three delivery routes demonstrated a decrease in the neurological disease being tested (Belur et al., 2021). Rapid clearance of gene delivery vehicles represents the most significant challenge with delivery to the cerebrospinal fluid (Householder et al., 2019). In the case of brain cancer, over small length scales (a few mm) intracranial/intratumoral injection has also shown promise for gene delivery nanoparticle injection and convection-enhanced delivery (Figure 5; Mangraviti et al., 2015) as well as PEGylation (J. Kim et al., 2020; Mangraviti et al., 2015) can improve transport properties further. This study also employed convection-enhanced delivery (CED) to improve delivery of the nanoparticles. CED involves implantation of a catheter connected to a pump which increases fluid pressure and allows therapeutics to bypass the BBB and travel further than if they relied on passive diffusion alone (Stine & Munson, 2019). There have been several drawbacks noted with CED, including backflow, air bubbles, and unsuccessful flow patterns based on brain structures (Mehta et al., 2017). CED entered the clinic over 20 years ago, with clinical trials delivering standard chemotherapies as well as conjugated toxins and liposomal formulations to GBM, though results have generally only demonstrated modest effects thus far (Jahangiri et al., 2017).

FIGURE 5.

Local brain delivery of PBAE/GFP nanoparticles via CED leads to effective tumor transfection in vivo. Coronal section of a 9 L tumor-bearing rat brain at 7 days post-PBAE/GFP infusion showing the tumor region (a, scale bar = 2 mm). Fluorescence microscopy images show GFP+ transfected cells in the tumor area (b, scale bar = 2 mm). Enlarged area shows a wide distribution of GFP+ cells within the entire tumor area including the periphery (c, scale bar = 500 μm). The colocalization of GFP and Cy5 shows that the nanoparticles penetrate into the cells and successfully transfect them (d–f, scale bar: 50 μm). Red, Cy5; green, GFP; blue, DAPI. Reproduced with permission from Mangraviti et al. (2015)

5.2.1 |. Intracerebroventricular injection

The ICV route of delivery involves injecting the gene therapy treatment into the cerebral ventricles through the use of an Ommaya reservoir (Duma et al., 2019; H. J. Kim, Cho, et al., 2021). This device consists of a dome and catheter, and is first implanted under the scalp to provide a route for injection into the ventricles (Magill et al., 2020). This method demonstrates effective biologic delivery rates, likely because the proximity to brain tissue counteracts the quick clearance time. The downsides are that this method is an invasive and time-consuming process (Taylor et al., 2021). Much work has also been done around intracranial injections using convection-enhanced delivery (CED), which promotes fluid flow through a pressure gradient, thus increasing distribution in the brain (Stine & Munson, 2019). ICV is a commonly used method for testing nanoparticles for gene delivery in vivo, with examples for drug delivery (Monge-Fuentes et al., 2021) and nucleic acid delivery (Akita et al., 2015; Shyam et al., 2015; Tanaka et al., 2018), and studies have been performed to assess the safety of ICV as a delivery route in humans (Cohen-Pfeffer et al., 2017; Slavc et al., 2018).

5.2.2. |. Intrathecal injection

Alternatively, IT injection takes place in the lower spinal cord into CSF, which can then circulate up through the brain (Fowler et al., 2020). Because this method presents a less invasive procedure than ICV injection, it also results in a quicker process and shortened recovery. IT injection has been used to explore nanoparticle administration in vivo, including studies looking at administration of fluorescent nanoparticles (Householder et al., 2019), nanoparticles delivering small molecules (Ramírez-García et al., 2019), and nucleic acids (Nabhan et al., 2016).

5.2.3 |. Intranasal administration

In intranasal injection, the nanoparticle is delivered into the olfactory epithelium, which allows direct access into the CNS, without the need to bypass the BBB (Erdő et al., 2018; Trevino et al., 2020). The axons of the nerve cells in the olfactory epithelium extend directly into the CNS via the olfactory bulb (Bryche et al., 2020). This method is the least invasive out of the discussed routes, however, it also generally yields the lowest delivery rates. For ex vivo nanoparticle-mediated gene delivery to mesenchymal stem cells for use as delivery vehicles, it has been demonstrated that the engineered cells can subsequently be delivered intranasally and migrate through the brain, including homing to brain tumors to secrete transfected protein and extend survival (Mangraviti et al., 2016). Additional intranasal delivery techniques have been explored, including a submucosal gel depot of a liposomal AntagoNAT formulation, which demonstrated delivery approximately 40% of ICV delivery of the same liposomal formulation and a safer route of administration (Padmakumar, Jones, Pawar, et al., 2021). A similar method was also developed using the same intranasal gel depot route of administration but instead using a polymeric shell for sustained delivery of an AntagoNAT (Padmakumar, Jones, Khorkova, et al., 2021; Table 2).

TABLE 2.

Comparison of nanoparticle routes of administration into the brain

Routes of administration	Advantages	Disadvantages	Applications
Systemic	Favorable for diseases or tumors not limited to a single or easily accessible location	Quick clearance, BBB transport	Systemic glioblastoma targeting (Gregory et al., 2020)
Intracerebroventricular	Effective biologic delivery rates, bypasses BBB	Invasive, time-consuming	Dopamine delivery for Parkinson’s disease, siRNA (targeting BACE1 and APP) delivery for Alzheimer’s disease, mRNA delivery to Astrocytes and brain neuronal cells (Monge-Fuentes et al., 2021; Shyam et al., 2015; Tanaka et al., 2018)
Intrathecal	Moderate invasiveness, shortened recovery time, bypasses BBB	Nanoparticle clearance prior to delivery	Chronic pain prevention using NPs targeting neurokinin 1 receptor, delivery of frataxin mRNA for Friedreich’s ataxia (Nabhan et al., 2016; Ramírez-García et al., 2019)
Intranasal	Noninvasiveness, bypasses BBB	Lowest delivery rate	Delivery of BMP4 to target brain tumors, delivery of AntagoNAT for upregulation of brain-derived neurotrophic factor (BDNF; Mangraviti et al., 2016; Padmakumar, Jones, Khorkova, et al., 2021)

6 |. ENGINEERING NONVIRAL NANOPARTICLES FOR APPLICATIONS IN THE CENTRAL NERVOUS SYSTEM

6 1 |. Optimizing design to improve nanoparticle delivery into the central nervous system

6.1.1 |. Targeting ligands to facilitate BBB crossing

A variety of modification approaches have been explored to improve nonviral nanoparticle design, improving delivery into the CNS. In a study conducted by Topal et al., SLNPs were functionalized with ApoE lipoproteins through a biotinylation process, and loaded with donepezil, for the treatment of Alzheimer’s disease (Topal et al., 2021). These APoE functionalized SLNPs were 147.5 ± 0.8 nm in diameter, with a zeta potential of −9.6 ± 0.5 mV. The lipid used in the particle formulation was Dynasan 116, a triglyceride lipid (Lu et al., 2018). ApoE lipoproteins function to transfer lipids to cells in the body and cross the blood–brain barrier through endocytosis pathways (Hartl et al., 2020). The authors found that when ApoE ligands were introduced to lipid nanoparticles, the nanoparticles were also taken up through a model of the BBB. The researchers demonstrated a significant increase in ApoE-nanoparticle uptake across the BBB in rat brain endothelial cells and other related BBB cell types (Topal et al., 2021). The BBB model utilized was created in a well plate and validated through trans-endothelial electrical resistance measurements. These measurements used electrical resistance across the membrane as a proxy for “tightness” of the barrier (Vigh et al., 2021). While these experiments were performed in a simplified rat BBB model, development of next-generation in vitro BBB models could provide improved alternatives to evaluate BBB passage (Linville & Searson, 2021). In another recent study, Zhang et al. created a protein-nanoparticle micelle that associates with ApoE in vivo that localizes to brain tissue (Z.-A. Zhang, Xin, et al., 2021). The PEG–PLA nanoparticles contain amyloid β-protein, which associates with ApoE to form the protein corona. These nanoparticles were synthesized with an average diameter of 103 nm, and a zeta potential of 7.2 mV. The positive zeta potential enabled electrostatic interactions with the anionic ApoE, forming a coated nanoparticle with an average diameter of 172 nm and 0.7 mV zeta potential. After demonstrating effective BBB passage in an in vitro transwell BBB model, these nanoparticles were intravenously administered to mice in vivo, where they successfully bypassed the BBB and led to reduction in glioma tumor size (Z.-A. Zhang, Xin, et al., 2021).

Lipid nanoparticles can be successful in crossing the BBB by incorporating functional biomolecules into the nano-structure. Ray et al. demonstrated that lipid nanoparticles functionalized with aptamers could increase nanoparticle trafficking through the BBB (Ray et al., 2021). Aptamers are short nucleic acids between 40 and 100 nucleotides that can be targeted toward a specific protein (Y. Zhang et al., 2019). The authors added a G-3 aptamer to LNPs composed of an ionizable cationic lipid and cholesterol and that had previously been demonstrated to be endocytosed by the CCR5 protein. The nanoparticles investigated ranged from ~60 to 90 nm in diameter with a zeta potential of −3 to +11 mV. In a simplified BBB model in HeLa cells and TZM-bl cells, over 60% transport was maintained across the BBB normalized to cellular uptake without the barrier model (Ray et al., 2021). In another study, incorporation of neurotransmitter lipidoids into lipid nanoparticles was found to increase the nanoparticle transport across the BBB (F. Ma, Yang, et al., 2020). Using a lipidoid derived from tryptamine combined with a PEGylated lipid (DSPE), Ma et al. were able to show intravenous delivery of an oligonucleotide (Tau-ASO) into a mouse brain in in vivo experiments. The Tau-ASO is a gene silencer, and the studies demonstrated significantly reduced levels of the Tau proteins in mice brains (F. Ma, Yang, et al., 2020). This study demonstrates the potential for delivering genes through the blood–brain barrier with LNPs, while bypassing the invasive procedures of intercranial injection or BBB disruption methods. In an alternative approach, Wang et al. found that the inclusion of borneol stearic acid helped guide LNPs and increase cellular uptake in the brain after intranasal administration (Wang et al., 2019).

Researchers have also designed nanoparticles to cross the BBB via conjugation to cells. Ayer et al. explored polymeric nanoparticle transport across the BBB, facilitated by T-cells, in both in vitro and in vivo models as T-cells can cross the BBB by extravasation through the endothelial layer (Ayer et al., 2021). PEGylated nanoparticles were conjugated onto the T-cell surface and the complexes were able to migrate across a mouse brain endothelial cell layer. Further in vivo work in mice demonstrated that these T-cell nanoparticles were transported across the BBB after systemic injection in the carotid artery (Ayer et al., 2021). Cox et al. investigated polyisoprene nanoparticles decorated with a maleimide group and a drug to facilitate BBB penetration (Cox et al., 2019). These particles were ~160–185 nm in diameter, with positive zeta potentials of approximately 40–50 mV. The maleimide group allowed for covalent binding with proteins to help facilitate BBB trafficking. By binding α2-macroglobulin, a protease that is able to cross the BBB, the nanoparticles were able to cross an in vitro BBB (Cox et al., 2018). Surface modifications of inorganic nanoparticles can also be performed to improve BBB crossing. In one study, gold nanoparticles were enclosed in an exosome expressing a rabies virus glycoprotein peptide to provide a targeting mechanism (Khongkow et al., 2019). Together, the particle size was 105 nm in diameter, with a zeta potential of approximately −10 mV. The exosome-gold nanoparticles were delivered intravenously to mice and were found to localize to the brain mechanism (Khongkow et al., 2019).

6 2 |. Engineering nanoparticles in applications for CNS disease models

6.2.1 |. Engineering nanoparticles for neurological disease use

Several additional studies have demonstrated the utility of nanoparticle modification to benefit application to a specific disease profile. Pinheiro et al. employed rabies virus glycoprotein (RVG29) in DSPE-PEG lipid nanoparticles to demonstrate nanoparticle delivery through the BBB. In this study, the authors delivered quercetin, which has been shown to reduce amyloid-beta peptide aggregation, and thus improve Alzheimer’s disease (Pinheiro et al., 2020). A variety of nanoparticles were examined in this study, ranging ~ −30 to −20 mV in zeta potential. The nanoparticles without the RVG29 were closer to ~150–200 nm in diameter, while nanoparticles conjugated with RVG29 were larger, at about ~200–250 nm in diameter. A similar approach is done by Gan et al. employed RVG29 in polymeric nanoparticles in a Parkinson’s disease profile. Parkinson’s disease is a motor neurological disease, commonly characterized by tremors. Currently, treatment options are continuous intake of L-dopa, which has negative side effects and is inconvenient to the individual. Gan et al. incorporated microRNA-124 and RVG29 into a polymeric nanoparticle, reducing inflammatory responses in a Parkinson’s model (Gan et al., 2019). The polymer was a PLGA-PEG, forming nanoparticles on average 162 nm in diameter. The authors first confirmed that nanoparticles with RVG29 significantly increased nanoparticle transport across the BBB in a simplified in vitro BBB model (Gan et al., 2019). The subsequent experiments demonstrated that delivery of microRNA-124 with the RVG29 polymeric nanoparticles reduced inflammatory cytokines. An in vivo study in mice was also performed, which validated the in vitro study (Gan et al., 2019).

6.2.2 |. Engineering nanoparticles for brain cancer use

Increasingly, there are studies on nonviral gene delivery nanoparticles to treat brain cancer, with a focus on the research on glioblastoma models as model animal systems, given the lethality of this grade IV primary brain tumor. Saha et al. demonstrated that amphetamine-functionalized lipid nanoparticles (consisting of β-amphetaminylated cationic lipid 16-BACL, di-cationic lipid 16-DCL, neutral lipid DOPC, PEG lipid DSPE-PEG[2000], cholesterol, and cargos PD-L1siRNA, paclitaxel, and NIR-dye) could increase BBB transport. In in vivo orthotopic mice glioblastoma models, the authors demonstrated that these positively charged 60–70 nm amphetamine-decorated lipid nanoparticles localized preferentially to the brain over other organs (including the heart, lung, spleen, liver, and kidney; Figure 6; Saha et al., 2020). The authors also demonstrated that delivery of a PDL-1 siRNA into an orthotopic glioblastoma mouse model reduced PDL-1 expression and increased survivability of the mice (Saha et al., 2020). In a polymeric nanoparticle approach, Mangraviti et al. aimed to reduce gliomas through the use of PBAE gene therapy. This study showed that positively charged PBAE- based nanoparticles approximately 134 nm in diameter carrying an HSVtk suicide gene plasmid were able to spread through orthotopic brain tumors (Figure 5) and extend survival in a rat glioma model (Mangraviti et al., 2015). In a different approach, Yang et al. delivered CRISPR/Cas9 plasmids using positively charged 180 nm diameter lipid–polymer hybrid nanoparticles aiming to knockdown MGMT genes, which can impair the efficacy of existing glioblastoma treatment drugs. In this study, the physical method of focused ultrasound (FUS) microbubbles was utilized to increase blood–brain barrier permeability (Q. Yang et al., 2021). FUS has been explored in the clinic for a variety of CNS disorders, including essential tremor and neuropathic pain, though most of these methodologies involved using FUS to create thermal lesions in the brain (Chang & Chang, 2017). More recently, FUS has been explored for its ability to modulate BBB permeability as described in a study by Yang et al. and used in conjunction with a variety of nonviral vectors such as polymers and liposomes (Fisher & Price, 2019; C.-Y. Lin, Hsieh, et al., 2015). Safety of FUS has been assessed in the clinic (Idbaih et al., 2019; Mainprize et al., 2019), though further investigation is necessary.

FIGURE 6.

In vivo biodistribution studies demonstrating localization of lipid nanoparticles to the brain following systemic delivery. Orthotopically established glioblastoma bearing C57BL/6 mice (n = 3) were i.v. administered with the NIR-dye loaded lipid nanoparticle of 14/16/18-BACL on day 10 post tumor implantation. Mice were sacrificed 24 h post i.v. administration, organs digested, and fluorescence intensity was measured using a multimode reader. Normalized mean fluorescence indicates the highest accumulation of fluorescently labeled nanoparticles of 16-BACL (a, *p < 0.05, parametric t-test). The median fluorescent heatmap also supports the notion that the fluorescently labeled nanoparticle of 16-BACL shows the highest accumulation in the brain among the three (b). To understand the targeting ability, both the fluorescently labeled nanoparticle of 16-BACL- and NT-lipid were intravenously administered on day 10 post glioblastoma implantation. Mice were anesthetized 24 h post i.v. injections, mouse head areas shaved, and the intensity of NIR-dye was first observed under a noninvasive imager. (c) (I) Mice treated with the NIR-loaded lipid nanoparticle of 16-BACL and (II) mice treated with the NIR-loaded nontargeting control lipid nanoparticle. Immediately after completion of noninvasive imaging (c), ex vivo images of different organs isolated from mice treated with the NIR-dye loaded lipid nanoparticle of 16-BACL (d) and the NIR-dye loaded nontargeting lipid nanoparticle (e) were recorded. In both parts (d and e), (I) brain; (II) lungs; (III) heart; (IV) liver; (V) spleen; and (VI) kidneys. Reproduced with permission from Saha et al. (2020)

Gene therapy for pediatric CNS cancers, medulloblastoma, and atypical teratoid/rhabdoid tumors, has also been explored. Choi et al. used PBAE nanoparticles and the HSVtk gene to increase survival time in in vivo pediatric brain cancer mouse models. The authors first screened for nanoparticle delivery efficiency in BT-12 (atypical teratoid/rhabdoid tumor) and D425 (medulloblastoma) cells. These cell lines were transplanted orthotopically to create mouse models, to which the authors delivered the PBAE/HSVtk NPs intracranially using linear PBAEs 537 and 447, chosen for their ability to transfect cells in vitro and in vivo, and which give rise to 100–150 nm nanoparticles with positive surface charge (Choi et al., 2020). Given the progress thus far in brain cancer gene therapy, there are many avenues to further investigate to create safe and effective technology to treat brain cancers.

6.3 |. Translating nonviral gene therapy to clinical approaches

6.3.1 |. Approved and ongoing clinical trials

In many cases, gene delivery studies have progressed to the clinical trial phase, and some have obtained FDA or EMA approval. Nusinersen, a treatment for spinal muscular atrophy, was approved by the FDA in 2016 and the EMA in 2017 (Winkle et al., 2021). Nusinersen is an antisense oligonucleotide targeting an intronic splicing silencer (ISS-N1), delivered intrathecally over multiple doses, to increase production of survival motor neurons (Goodkey et al., 2018; Q. Li, 2020). Complications of repeated intrathecal injection remain a possibility, and as a result, utilizing a nanoparticle gene delivery mechanism could have advantageous implications in the future.

A Phase 0 clinical trial utilized spherical gold nanoparticle cores conjugated with radially oriented siRNAs as spherical nucleic acids for the treatment of glioblastoma (Kumthekar et al., 2021). The densely packed siRNAs were designed to target Bcl2L12, a glioblastoma oncogene (Paul, 2017). The results demonstrated nanoparticle trafficking through the BBB and into glioblastoma cells after an intravenous injection (Figure 3), successfully and safely reducing protein expression of Bcl2L12 (Kumthekar et al., 2021). Other clinical studies using gold nanoparticles to deliver a treatment for glioma are ongoing, indicating numerous possibilities for nonviral nanoparticle delivery to the brain (Mueller, 2021).

While there are several clinical trials are also underway for neurological disorders, many of these trials are utilizing viral gene therapy treatment methods; these include lentiviral or AAV vectors to treat epilepsy, Alzheimer’s disease, among others (Tuszynski, 2021, p. 2; University College, London, 2020). One clinical trial studied BIIB094 for the treatment of Parkinson’s Disease (Biogen, 2022). BIIB094 is an antisense oligonucleotide aiming to reduce the leucine-rich repeat kinase 2 (LRRK2), which is linked to Parkinson’s disease (A. P. T. Nguyen et al., 2020). Another clinical trial is investigating the effects ALN-APP, an RNAi targeting amyloid precursors for the treatment of Alzheimer’s disease (Alnylam Pharmaceuticals, 2022). A pipeline review of gene delivery biotechnology companies (including Moderna, Alnylam, GenerationBio, Editas, CRISPR Therapeutics, Translate Bio, and Beam) found one instance of preclinical work being conducted for neurological diseases. In addition, in late 2021, a startup company (GenEdit), raised $26 M to support their platform of nonviral, non-lipid, polymer nanoparticles for treating multiple diseases, including those of the central nervous system (Cross, 2021). The sparsity of ongoing clinical trials involving nonviral gene therapy methods indicates that this is an emerging field and that there is potential for nonviral nanoparticle optimization in this area (Table 3).

TABLE 3.

Clinical trials on nonviral nanoparticles for gene delivery to the brain

Name	Indication	Therapeutic molecule	Type of nanocarrier	Clinical trial
Nusinersen	Spinal muscular atrophy	Antisense oligonucleotide targeting ISS-N1	Antisense oligonucleotide	NCT04591678
NU-0129	Glioblastoma	siRNAs targeting Bcl2L12, a glioblastoma oncogene	Gold nanoparticles	NCT03020017
BIIB094	Parkinson’s disease	Antisense oligonucleotide targeting LRRK2	Antisense oligonucleotide	NCT03976349
ALN-APP	Alzheimer’s disease; cerebral amyloid angiopathy	RNAi targeting amyloid precursor protein (APP)	RNA interference	NCT05231785

6.3.2 |. Formulation stability and manufacturing challenges

As nonviral nanoparticles are translated for clinical use, pharmacovigilance and formulation stability remain extremely important considerations for patient safety. There are multiple approaches to increase the shelf life, stability, and batch-to-batch consistency of nanoparticles (Najahi-Missaoui et al., 2020). For example, coating materials, such as PEG, have been utilized with lipid nanoparticles to improve stability, while flame spray pyrolysis has been used to prevent dissolution of inorganic nanoparticles (Phakatkar et al., 2021). New techniques, such as flash nanocomplexation, have demonstrated size-controlled development of lipid nanoparticles, thus reducing particle heterogeneity (He et al., 2018). The FDA and EMA have also devised parameters of stability and quality that must be met (Souto et al., 2020). With quality assurance paramount for patient safety and reproducible patient responses, nanomedicine production and characterization methods are continuously improving to meet desired specifications for clinical products.

Because nonviral gene therapy is an emerging approach to treating disease, there are also several unaddressed manufacturing and development challenges for preclinical programs, including meeting production capacity, upscaling manufacturing equipment, and GMP development (Vuksanaj, 2020). For some of these challenges, the FDA has published guidance documents for production of gene therapy, including suggestions to produce many small batches of product, rather than few large batches (Food and Drug Administration (FDA), 2021). Some companies have opted to work with contract development and manufacturing organizations (CMOs) in the short term, rather than developing the capability in-house, due to high costs and long lead-times (Mullin, 2021). Others are evaluating investments in technological innovations to boost production or investments in building manufacturing capabilities to meet gene therapy demands in the long term (Roper & Middleton, 2021). Technological innovations include automated systems and new software to replace manual bench work, to increase production efficiency and capacity (Gene Therapy Industry Report, 2021). Further developments are necessary to streamline the translation of gene therapies, including development of low-cost GMP manufacturing of nucleic acids at scale (The Expanding Frontier of Nucleic Acid Therapeutics, 2022). Thus, while nonviral gene therapy manufacturing currently has challenges, biotechnology companies are implementing new technologies and approaches in order to overcome these issues.

7 |. CONCLUSION

Many advancements have been made in the development of nanomaterials for gene delivery and their ability to overcome extracellular delivery and intracellular delivery bottlenecks. These nanomaterials primarily include lipid-based, polymer-based, and inorganic nanoparticles which can be administered to an individual through intracranial, ICV, IT, or nasal injection. Biodegradable components of nanoparticles are advantageous as they are more easily eliminated from biological systems, and size is a key parameter affecting particle distribution and uptake. Much work has been conducted to modify the nanoparticles for extracellular passage through the blood–brain barrier and intracellular passage to the nucleus. Surface coatings and ligands can be used to help target nanoparticles to specific locations or help improve circulation time following administration. Key structural features for extracellular delivery include size, lipophilicity, polar surfaces, and electrostatic charges, all parameters that researchers seek to address in designing novel nanoparticles systems that aim to cross the BBB.

Key structural features for intracellular delivery include (1) the ability to bind to and/or encapsulate the nucleic acid cargo to facilitate cellular uptake, which can be accomplished with hydrophobic materials and positively charged materials; (2) the ability to promote endosomal escape, which can be accomplished by membrane-disruptive materials and proton-buffering materials; (3) the ability to release the nucleic acid in the cytosol, which can be facilitated by degradable linkages throughout the nanomaterial. Methods for delivery that make use of anatomy (such as ICV injection) or physical methods (CED, image-guided focused ultrasound) can further enhance the effectiveness of gene delivery nanoparticles to treat CNS and brain disorders.

However, there are many challenges remaining for nonviral nanoparticles for gene therapy. First, nonviral vectors still demonstrate much lower gene delivery efficiency compared to viral vectors; boosting this transfection efficiency will be critical in advancing nonviral nanoparticle usage in gene therapy. In addition, there are significant hurdles translating research findings into clinical work, due to unaccounted differences between animal models and human patients, including especially the blood–brain barrier. There may also be unforeseen side effects when these gene therapies are implemented in a human individual. Stability, scale-up, and other manufacturing considerations remain design challenges for nanoparticle systems. Finally, there is the ever-present challenge of reducing gene therapy costs to the patient, given the difficulty of pricing these therapeutics. Despite these challenges, research regarding nanoparticle gene delivery to treat CNS disorders continues to evolve and holds immense promise in addressing complex diseases.

DATA AVAILABILITY STATEMENT

Data sharing is not applicable to this article as no new data were created or analyzed in this study.