2.1 Theoretical basis of gene therapy

Gene therapy research assesses three main approaches: (i) introduce exogenous genes into diseased cells to produce normal gene expression products to further supplement missing or loss-of-function proteins (i.e., by upregulating gene expression); (ii) downregulating gene expression by using small interfering RNA (siRNA), antisense oligonucleotides (ASOs), short hairpin RNA (shRNA), or microRNA (miRNA); and (iii) editing mutated genes using zinc finger nucleases, transcription activator-like effector nucleases, or CRISPR/Cas9 technology, resulting in gain or loss-of-function.^14-17

Gene therapies can be classified as germline or somatic, depending on the cells targeted. Germline gene therapies (GGTs) target germ cells; therefore, GGTs can not only treat genetic diseases in the first generation of offspring but also pass on new genes to future generations, so genetic diseases can be mitigated. A form of GGT called “mitochondrial replacement therapy” (MRT) has been developed and acts by replacing mutated mitochondrial DNA (mtDNA) in mutant carrier oocytes with donated, mutation-free counterparts, allowing women with mtDNA mutations to conceive healthy children. The first successful human case involving MRT was reported in Mexico.¹⁸ CRISPR/Cas9 technologies can be used to edit germline genes; however, there are many technical problems with CRISPR/Cas9 gene-edited embryos, such as mosaicism, off-target effects, and chromosomal structural abnormalities.¹⁹ There are also ethical issues surrounding GGTs, since they can have unpredictable consequences.

Somatic gene therapies, which comprise the focus of this study, can be divided into two categories based on the route of gene transfer: The ex vivo route and the in vivo route.¹ In the ex vivo route, the gene of interest is introduced into cells extracted from the patient. The cells are then delivered back into the human body after in vitro proliferation of the cells, screening, drug treatment, or a series of other operations.¹⁴ This approach requires a gene delivery vector, the DNA or RNA that makes up the gene itself, and technologically advanced facilities for processing the cells. For the in vivo route, a genetic material vector is injected directly into patients.²⁰ This approach is similar to other types of drug delivery systems. Compared with the ex vivo route, the in vivo approach is more dependent on gene delivery vehicles.

2.2 CAR gene therapy

DNA-based gene therapies involve the delivery of DNA fragments and CAR T-cell therapy. CAR T-cell therapy involves the in vitro genetic modification of autologous T-cells in patients to encode a chimeric receptor that binds to a specific tumor antigen. These cells are then reinfused into the patient.^{21, 22} CAR T-cell therapies have been successful in treating hematologic malignancies.²³ Autologous T-cells are genetically modified to express CARs targeting the B-cell antigen CD19, which has led to remarkable clinical responses in patients with B-cell-related malignancies. However, CAR T-cell therapies have several major limitations, including life-threatening CAR T cell-related toxicity, limited efficacy against solid tumors, resistance to B-cell malignancies, antigen escape, limited persistence, poor transport, tumor infiltration, and an immunosuppressive microenvironment.²² Transport can be improved by local injections^{24, 25} or chemokine receptors on CAR T-cells.^{26, 27} Immunosuppressive microenvironments can be tackled using immune checkpoint inhibition. Toxicity can be reduced while optimizing efficacy by altering the structure of CARs, modifying CAR-transduced T cells, or implementing “off-switch” or suicide gene strategies.^{28, 29} Mesothelin is another factor that has been investigated in studies on CAR T-cell therapies. Mesothelin is a tumor differentiation antigen that is overexpressed in a wide range of solid tumors.³⁰ Several mesothelin-targeted CAR T-cell therapies have been developed to treat solid tumors,^{31, 32} but the use of CAR T-cell therapies for solid tumors remains challenging because of the biocomplexity of the solid tumor microenvironment.

2.3 RNA-based gene therapy

RNA-based therapies comprise two categories. The first includes RNA molecules or analogs used directly as therapeutic agents and the second includes RNA-targeted small-molecule drugs. Various RNA delivery vectors have been investigated for the widespread use of RNA therapies. Among RNA delivery systems, biomaterials for nonviral RNA delivery have exhibited excellent performance.³³

ASOs are single-stranded nucleic acid polymers comprising 18−30 bases that can combine with specific sequences of target mRNAs through complementary base pairing to upregulate or downregulate protein expression.^{34, 35} Antisense RNAs comprise a class of small RNAs without coding functions of their own, but can suppress the function of target RNAs and regulate the expression of the corresponding gene by complementarily binding to the target RNA, especially to specific regions of the mRNA, through hydrogen bonds between paired bases.

RNA interference (RNAi) is a sequence-specific way to knock down the expression of target genes through double-stranded RNA (dsRNA) mediated mRNA degradation or inhibition of mRNA translation.³⁶ RNAi involves posttranscriptional gene silencing, an effect that may be mediated by miRNA and siRNA. In this process, mRNAs with homologous sequences to the dsRNA are degraded, suppressing the expression of the gene of interest. miRNAs are endogenous, highly conserved, and small noncoding RNAs comprising approximately 22 nucleotides. miRNAs regulate the expression of coding genes by incompletely binding to their 3′-untranslated regions and are further considered key regulators of target gene expression.³⁷ Long dsRNA is cleaved by an intracellular Dicer into short dsRNA comprising 20−25 nucleotides. These molecules are siRNA.³⁸ Compared with conventional gene therapies, siRNAs can effectively silence diseased genes and knock down their function. Recent US FDA approval of the first and second siRNA drugs has marked the beginning of the era of RNAi therapies. These drugs include Patisiran, which is an siRNA formulated as a lipid complex for delivery to hepatocytes,³⁹ and Givosiran, which is an siRNA that is bound to the N-acetyl-D-galactosamine (GalNAc) ligand for the asialoglycoprotein receptor-mediated targeted delivery to hepatocytes,⁴⁰ respectively.⁴¹

RNA activation (RNAa) is mediated by small dsRNA, targets gene regulatory sequences such as promoters and various noncoding regions, and involves transcriptional and epigenetic alterations. Small dsRNAs that act as activators of gene expression have been defined as small activating RNAs (saRNAs) or antigene RNAs.⁴² RNAa requires additional steps, such as saRNA crossing the nuclear membrane, with the mechanism being more complex than that of RNAi,⁴³ thus RNAa occurs later than RNAi. Furthermore, the effects of saRNA-induced gene activation last much longer than those of downregulated expression triggered by RNAi. saRNA-induced gene activation by RNAa has also been implemented as a novel strategy to treat cancer. For instance, p21-saRNA-322 impeded the growth of colorectal cancer by activating the p21 gene.⁴⁴

Another RNA-based therapy involves mRNA, which can encode proteins that exert therapeutic activities. In eukaryotes, RNA polymerase synthesizes precursor mRNA in vivo when it converts genes into primary mRNA transcripts. During mRNA splicing, introns are removed from the transcript, whereas exons are spliced to produce mature mRNA. The advantages of mRNA are as follows: (1) mRNA can theoretically express any protein and thus treating an extensive range of diseases. (2) DNA can be transcribed only into mRNA after entering the nucleus. However, mRNA need not enter the nucleus and can initiate protein translation in the cytoplasm. Therefore, it is more efficient than DNA. (3) mRNA does not affect genetic information by inserting itself into the genome as DNA and viral vectors do. mRNA-encoded proteins are only expressed momentarily and are quickly degraded afterwards without the risk of gene integration. (4) Compared with proteins and viruses, mRNA can quickly translate all the proteins in cells. Moreover, industrial production is simple and inexpensive. In addition, mRNA chains are long-chain macromolecules with negative charges, while the surfaces of cell membranes also have negative charges. It is difficult for mRNA molecules to cross the cell membrane and enter the cell because of electrostatic repulsion. Moreover, because mRNA molecules are single-stranded, they are extremely fragile and can rapidly be degraded by various enzymes in the body. The information encoded by mRNA further covers the sequences of ribosomal proteins and must be delivered to cells to encode proteins. Therefore, there are two barriers to delivering mRNA into cells: The enzymatic degradation of mRNA while being delivered and its electrostatic repulsion with the membrane barrier. Therefore, special modifications or encapsulated delivery systems are necessary to achieve the intracellular expression of mRNA and change its intracellular biodistribution, cellular targeting of mRNA, and uptake mechanisms to facilitate mRNA delivery.⁴⁵

2.4 CRISPR/Cas9-mediated gene therapy

The CRISPR system recognizes and degrades foreign genetic materials, such as RNA and DNA, for bacterial defense by integrating short exogenous sequences into the bacterial genome and transcribing them into CRISPR RNAs (crRNAs).⁴⁶ The CRISPR/Cas9 system, from which the type II CRISPR system was modified, comprises the Cas9 protein and single guide RNA (sgRNA). Under the guidance of sgRNA, Cas9 cleaves the target DNA by recognizing adjacent protospacer motifs. After the target DNA breaks, repair mechanisms are initiated, including the cellular nonhomologous end-joining repair mechanism, which can rejoin the genomic DNA at the break and introduce insertion or deletion mutations. Another repair mechanism is homologous recombination repair (HDR), which requires the involvement of HDR templates to insert or replace gene fragments at the break.⁴⁷

CRISPR/Cas9 has become a research hotspot and has been used to edit genomes to treat various diseases. CRISPR/Cas9 can be delivered in three ways: (1) the delivery of plasmid DNA encoding Cas9 and sgRNA; (2) the delivery of Cas9 mRNA and sgRNA, during which mRNA is converted to Cas9 nuclease through translation in the cytoplasm; (3) and the delivery of ribonucleoproteins (RNPs), which comprise a complex of Cas9 protein and sgRNA. The latter bypasses the transcription and translation process and provides the fastest form of gene editing.⁴² There are advantages and disadvantages of each delivery method, mainly in terms of efficiency, off-targeting, and immunotoxicity (Table 1).

3.1 Physical methods

Physical stimulation has been shown to enhance gene transfer into cells because it allows genes to move near the cell membrane. It also allows for the temporary microdisruption of the cell membrane. Various physical methods have been investigated to enhance gene expression in vivo using needle injections, ballistic pressure injections (gene gun), electric fields (electroporation), hydrodynamic pressure (water perforation), magnetic fields (magnetic transfection method), and ultrasounds (ultrasonic perforation).⁵³ In addition, most biomaterial-based carriers are based on cationic polymers and lipids, which allow genes to maintain physiological pH. These complexes not only protect genes from nuclease attacks, but also promote cellular internalization through electrostatic interactions with cell membranes.

Microinjections comprise a nonendocytic driving technique for the precise delivery of molecules and cells into droplets or for the delivery of genes, molecules, proteins, or viruses into individual cells.^{54, 55} This technique allows for the rapid delivery of genes into the nuclei. The advantages of microinjections include precise doses and timing, high transduction efficiency, and low cytotoxicity. However, manual microinjections are labor intensive and time consuming. They also comprise an empirical technique, meaning that the cell survival rate after microinjections is heavily dependent on the operator, which limits the application of this technique for numerous cells in samples.⁵⁶ Sun et al.^{57, 58} proposed an autonomous microrobot system comprising a pipette holder and a syringe to achieve a high injection success rate.⁵⁹ Despite advances in automated microinjections, current techniques have limitations in terms of accuracy or injection speed. In contrast to active microinjection techniques with injection cycles greater than 0.1 s, Azarmanesh et al.⁵⁵ developed a passive microinjection technique that relies on pressure-driven fluid flow and pulsating flow patterns within a high-throughput droplet microfluidic system to generate continuous droplets while rapidly managing microinjection droplets. This technique reduces the injection period to approximately 3 ms and delivers the droplets with minimal errors.⁶⁰ Passive microinjection technologies can work effectively and in a cost-efficient manner. They can also deliver cells and particles into droplets accurately and noninvasively.⁵⁵ Chen et al.⁶¹ administered microinjections of oocytes and early embryos at different stages of the development of their experimental mice by combining a suction head pipette and piezo-assisted micromanipulator, which significantly improved the cell survival rate after microinjections.

Gene guns are devices that deliver foreign molecules into cells. By wrapping DNA in heavy metal particles, such as gold, tungsten, and titanium, and using mechanical forces to launch these particles into the cell, the target genes can be integrated into the target cells. For more than two decades, gene guns have been used to transfer DNA to various biological species. However, cell damage caused by the air pressure (20−60 bar) used by the gene gun system and particle impingements in target tissues are limitations of gene gun use.²⁰

Electroporation changes the permeability of a cell membrane by using short, high-voltage pulsed electric fields to treat molecules and direct the physical delivery of genes to the target tissue.⁶² In recent years, electroporation has been applied to gene therapies that involve CRISPR/Cas9 gene editing systems to treat various diseases.^{63, 64} Cells must use electroporated colorimetric dishes or microfluidic devices to be concentrated in small volumes. The cell density typically used in these methods is approximately 0.5−2 × 10⁶ cells/mL.^65-67 This prevents large-scale implementation for industrial production. Considering the electrodes, electroporation can be used only close to the skin⁶⁸ or to surgically transfer genes to deeper tissues, such as a beating heart.²⁰ Electroporated cells are low in vitality,³⁸ and their traumatic and tissue contusions can lead to cell death.⁶⁹

Sonoporation, which uses ultrasounds and microbubbles, comprises the ultrasonic transfection method. This method uses high-intensity ultrasounds to induce instantaneous holes in cell membranes to allow DNA entry. Ultrasound-mediated gene transfer, or acoustic perforation, is a minimally invasive, nonviral, and clinically transformable gene therapy. Compared with other physical gene delivery methods, such as electroporation, ultrasound-mediated gene transfer provides better security and is less invasive. Ultrasound-targeted microbubble destruction (UTMD) refers to the targeted delivery of drugs or genes using ultrasound and microbubbles. Several studies have shown that UTMD may enhance local gene delivery. Microbubbles have a significant cumulative effect on ultrasonic membrane penetration. The disturbance of the cell membrane and blood vessel walls by ultrasonic oxidation of microbubbles can increase permeability and gap delivery.⁷⁰ Intracellular reactive oxygen species produced by ultrasound may contribute to cell membrane penetration without affecting cell viability.⁷¹ The main advantage of ultrasound-mediated microbubble-assisted transfection is the ability to deliver and express nucleic acid only in sonar regions rather than in nontargeted organs.⁷² Ultrasound-mediated microbubble cavitation was used by Zhang et al.⁷³ to facilitate AAV-mediated cochlear gene transfection across the round-window membrane.

Photoporation is the process of delivering nucleic acids mediated by membrane permeabilization induced by high-intensity and short-duration laser pulses on individual cells.^{74, 75} Hasanzadeh Kafshgari et al.⁷⁶ proposed antibody-functionalized gold nanostar-mediated photoporation to efficiently and selectively transfect specific cells with siRNA, mRNA, or Cas9-RNPs for their targeted delivery to human retinal pigment epithelial (RPE) cells (Figure 3A). In addition, the concept of hydrodynamic transport-based water perforation was established in 1999 to achieve effective plasmid DNA delivery to the liver by rapid tail vein injections of large amounts of DNA solution into mice.^{77, 78} Kizer et al. developed an intracellular delivery platform called a “hydroporator,” which enables the rapid intracellular delivery of high-throughput carrier-free nanomaterials and DNA origami nanostructures⁷⁹ using only fluid inertia while maintaining biological stability in living cells (Figure 3B).⁸⁰

Another physical gene transfer method is magnetic transfection, or magnetofection. This method involves linking a viral or nonviral gene delivery vector with magnetic nanoparticles and magnetically directing the vector to the target cells for fast and efficient nucleic acid delivery (Figure 3C).⁸¹ The main advantages of magnetic transfection include its ability to facilitate fast and efficient transfection at low carrier doses. This also allows remote control of vector targeting in vivo. Plank et al.⁸² combined vectors and magnetic particles through appropriate bonds and salt-induced colloidal aggregation. If these particles were mixed with DNA, liposomes, or polymeric complexes such as polyethyleneimine (PEI)-DNA in a salty buffer, they would bind or copolymerize with these compounds. Under a magnetic field, cells were incubated with a vector-magnetic particle mixture that attracted the particles into the cells.⁸² The magnetic particles used were mainly superparamagnetic iron oxide nanoparticles, which were coated with highly biocompatible polymers, such as PEI,⁸³ glucan,⁸⁴ and poly(amidoamine) (PAMAM) dendritic macromolecules.⁸⁵ The results of magnetic transfection comprise the rapid deposition of a full-dose vehicle on target cells, which further leads to a high percentage of cells being rapidly transfected within minutes.⁸⁶

3.2 Viral vectors

Studies have shown that viral vector-mediated gene delivery is the most efficient method of gene transfer. Viral vectors are therefore the most commonly used gene therapy vectors. The production of nonpathogenic viruses for gene therapy has increased in recent years. These viruses include retroviruses, lentiviruses, adenoviruses, and AAVs.⁸⁷

Retroviruses are RNA viruses composed of two RNA strands packaged with proteins. In addition, they comprise a simple genome structure and an expression cassette that can be replaced by therapeutic genes. When retroviruses enter host cells, the viral RNA is reverse-transcribed into a double-stranded DNA molecule, enters the nuclei of the cells, and randomly integrates into the host cell genome as a “provirus.” The provirus is then transcribed into RNA, after which the packaging protein is synthesized. The transcribed RNA genome is then released outside the cell to undergo the cell cycle.⁸⁸

Lentiviral vectors are gene therapy vectors developed based on the human immunodeficiency type I virus (HIV-1). Lentiviruses comprise a family of retroviruses, but unlike general retroviruses, lentiviruses have a wider host range for infection of both dividing and nondividing cells. Research on lentiviral vectors is moving rapidly and aims to efficiently integrate foreign genes into the genomes of host cells for long-lasting expression.⁸⁹

Adenoviruses are large molecules with genomes of approximately 36 kb. They are double-stranded, nonenveloped DNA viruses. Adenoviruses enter cells through receptor-mediated endocytosis, after which their genomes are transferred to the nucleus, but remain outside the chromosomes without integrating into the host cell genome. The host cell range of adenoviruses is broad and can infect terminally differentiated cells that divide and do not divide, such as neurons.⁹⁰

AAVs are single-stranded, linear, DNA-deficient viruses without envelopes. AAVs comprise an icosahedral envelope protein with a diameter of approximately 22 nm. Furthermore, they comprise a single-stranded linear DNA genome of approximately 4.7 kb. AAVs cannot replicate independently; their replication depends on secondary viruses, such as adenoviruses. AAV vectors comprise a new class of safety vectors that have been widely studied and are not pathogenic to humans.^{91, 92}

3.3 Nonviral vector delivery systems

Nonviral vectors must be able to mimic viral functions, such as expressing nucleic acids, targeted cell attachment and internalization, endosomal escape, and nuclear transfer.⁹³ Most gene vectors enter cells primarily through the endocytic pathway. After being internalized through the endocytic pathway, genes are carried to endosomal compartments, where they are then enzymatically degraded after endosomal-lysosomal fusion. An effective gene vector should promote the rapid escape of genes from endosomes. Numerous studies have assessed materials and methods to promote endosome escape and constructed highly efficient and low toxicity nonviral gene vectors.

Biomaterials include three categories: metallic, inorganic, and organic materials. Organic materials mainly include polymers, which can be categorized as natural polymers, such as chitosan, hyaluronic acid (HA), and cyclodextrin (CD), or synthetic polymers, such as polycationic peptides, PEI, polyesters, polyethylene glycol (PEG), and polyamide (PAA), depending on their source. Natural cationic polymers are often nontoxic, less immunogenic, and biodegradable; thus, many have been approved by the FDA as safe biomaterials. Synthetic polymers have the advantage of being modified by combining bioactive moieties and functional groups, but are limited by their excessive cytotoxic effects. In addition, because of their high positive charge, they can bind to anionic biomolecules in sera, which can lead to off-target side effects. However, low-molecular-weight cationic polymers have good cell tolerance, but unfortunately, they are inefficient in cellular delivery.⁹⁴

Biomaterial-based vectors often have well-defined structures, designed functions, no immunogenicity, and the potential for large-scale production.⁹⁵ Another advantage of these carriers is that they protect their cargo from enzymatic degradation and immune recognition.⁹⁶ Therefore, cationic lipids, polymers, or lipid–polymer hybrids have been developed for gene delivery by being compounded with negatively charged nucleic acids.¹⁶

4.1 Liposomes and their derivatives

Liposomes are self-assembled vesicles composed of phospholipids with polar head and nonpolar tail groups and stabilizers, such as cholesterol. The cationic nature of liposomal systems and their unique ability to capture lipophilic and hydrophilic compounds allow for the encapsulation of nucleic acids and various drugs in these vesicles. Once inside cells, liposomes are processed via the endocytic pathway, after which nucleic acids are released from endosomes or vectors into the cytoplasm. Liposome formulations are characterized by particle sizes, charges, number of sheets, lipid composition, and surface modifications with polymers and ligands, which determine their stability. Over the past two decades, lipid cations have been extensively studied and modified to reduce cytotoxicity and enhance gene expression levels. For example, the binding of PEG polymers to liposomal membranes has been demonstrated to be a critical strategy for elongating the circulation times of carriers and the prevention of the reticuloendothelial system (RES)-mediated carrier removal through spatial stabilization.⁹⁷

Micelles are thermodynamically stable micellar agglomerates of self-assembled and ordered molecules in aqueous solutions when the surface activator reaches a certain concentration. Lipid micelles are tiny droplets generated when the hydrophilic head of a phospholipid is exposed to water, whereas the hydrophobic tail clusters to avoid water molecules. Phospholipids with short compact tails often form micelles, whereas those with longer tails form liposomes.⁹⁸ Surfactants are amphiphilic molecules with lipophilic tails that cluster inside micelles due to hydrophobicity and hydrophilic head ends that extend outward because of polarity and protect the hydrophobic groups inside the micelles. Micelles may act as vectors for the core genetic material and they can easily enter tissues while rarely triggering immune responses.⁹⁹

Lipid nanoparticles (LNPs) are lipid vesicles with a homogeneous lipid core.¹⁰⁰ LNPs include ionizable and cationic lipids, cholesterol, phospholipids, and PEG lipids, of which ionizable lipids play an important role in protecting nucleic acids from nuclease-mediated degradation. The encapsulated drugs are distributed within the hydrophobic layer, hydrophilic cavity, or on the surface of the nanoparticles. Moreover, the external surface of LNPs can be coupled to targeting ligands. Thus, the surface chemistry of LNPs is highly programmable.¹⁰¹ Lipid polyplexes are double-layered structures with polymer-encapsulated mRNA as the inner core and encapsulated phospholipids as the outer shell. Lipid polyplexes provide better encapsulation and protection for mRNA than conventional LNPs and allow gradual release of mRNA molecules as the polymer degrades.¹⁰² Lipid polyplexes (Lipo-polyplexes) have a balance between stability and cargo release that improves the efficiency of transfection while reducing cytotoxicity.^{103, 104}

4.2 Peptides

Peptides can be integrated as functional components in nonviral gene delivery systems to overcome biological barriers.⁹⁷ Functional peptides that can be incorporated to improve the low transfection efficiency of nonviral systems and ultimately enhance gene expression can be classified according to their use in overcoming biological barriers. These classifications include (1) cell-penetrating peptides (CPPs), which facilitate cell entry; (2) targeting peptides, which improve specific cell binding; (3) endosomal disruption peptides, which overcome endosomal trap barriers; and (4) nuclear localization signal peptides, which facilitate efficient nuclear entry.

CPPs are short peptides of only 30 amino acids. They can penetrate biological membranes and deliver various bioactive substances into cells. CPPs are positively charged alkaline peptides at physiological pH and have large cargo loads, high cellular permeability, and can cross the membranes of different cell types while exhibiting low cytotoxicity and causing no immune response.^{105, 106} In addition, they have been used as carriers for siRNA, plasmid DNA, small molecules, proteins, and other peptides in vitro and in vivo. Peptides can self-assemble with nucleic acids to form peptide-based nanoparticles (PBNs), thus opening peptides to the field of nanomedicine.¹⁰⁷ Depending on their origin, CPPs can be subdivided into three categories: protein-derived peptides, chimeric peptides, and synthetic peptides.¹⁰⁸ CPPs can also be classified into the following types according to their physicochemical properties: Cationic peptides, such as transactivator (TAT) and poly(alpha-l-lysine) (PLL), which comprise short sequences of amino acids such as arginine, lysine, and histidine; amphiphilic peptides, such as MPG (GLAFLGFLGAAGSTMGAWSQPKKKRKV), which are peptides with polar and nonpolar structural domains; and hydrophobic peptides, which are formed from nonpolar residues of valine, leucine, and tryptophan.¹⁰⁹

Protein-derived CPPs comprise polypeptide regions of naturally occurring proteins that primarily translocate proteins into cells. The TAT sequence (RKKRRQRRR) originates from the TAT transcription factor of HIV. Bahadoran et al.¹¹⁰ developed PAMAM dendrimers coupled to TAT. When TAT/PAMAM nanoparticles were formulated, skin penetration and cellular uptake of plasmid DNA were slightly improved. TAT-conjugated PAMAM achieved transdermal delivery of DNA better than unmodified PAMAM dendrimer.¹¹⁰ Furthermore, CPP- and cationic poly(ethyleneimine)-conjugated gold nanoparticles (AuNPs) were developed by Niu et al.¹¹¹ for transdermal delivery of plasmid DNA to reverse the progression and metastasis of cutaneous melanoma. Synthetic CPPs are synthetic poly-arginine- or poly-lysine-based CPPs or more efficient membrane-penetrating CPPs modified from natural CPPs, such as MPG and Pep-1 (KETWWETWWTEWSQPKKRKV). Pep-1 was the first approved CPP on the market, with the trade name Chariot. Peptides made from lysine and arginine only, called “poly-Lys” and “poly-Arg,” respectively, were the first to be evaluated as artificial CPPs due to their ability to be internalized in living cells. When used as carriers, poly-Lys PLLs are usually bound to other proteins and synthetic compounds, such as PEI, whereas unbound poly-Lys has a lower transfection rate.¹¹² Chimeric CPPs are synthetic peptides formed by the fusion of two or more naturally occurring sequences. Amphipathic CPPs derived through covalently linking hydrophobic structural domains to effectively target the nuclear localization sequences (NLS) of cell membranes are also chimeric peptides. For example, MPG and Pep-1 based on simian virus 40 NLS PKKRKV.¹⁰⁹

4.3 Polymeric nanoparticles

Nanoparticles are less than 100 nm in diameter and vary in shape, size, and materials used to prepare them. Nanomedicines are often intended to enhance therapeutic efficacy by efficiently delivering drugs to the target site and/or reducing toxicity by minimizing their accumulation in healthy body sites. Nanoencapsulation can protect therapeutic agents from degradation in the biological environment and can further provide solubilization.¹¹³ Nanoparticles have positively charged surfaces, whereas cell surfaces are negatively charged; therefore, nanoparticles can enter target cells efficiently. Once inside the cell, nanoparticles undergo lysosomal escape and disassemble within the cytoplasm, releasing nucleic acids that enter the nucleus through nuclear pores and rely on the host enzyme system to express functional proteins for therapeutic purposes. As nanoparticles enter the circulation of the host, they encounter enzymes, plasma proteins, RES, phagocytes, and other components. The protein layer that forms on their surface is called a “protein crown.” Nanoparticles with protein crowns are readily recognized by the immune system, which leads to rapid clearance by the mononuclear phagocyte system (MPS). Nanoparticles with protein crowns also accumulate in the small pulmonary capillaries and cause serious toxic effects.¹¹⁴ The prolonged circulation of nanoparticles in the blood is a prerequisite for achieving the controlled and passively or actively targeted release of the encapsulated gene or drug at the desired site of action. The most widely used method for masking or camouflaging nanoparticles involves the adsorption, grafting, or binding of PEG or other hydrophilic polymers, such as polysaccharides, to the surfaces of the particles, reducing toxic side effects and prolongs the circulation of the particles in the blood. The main factors that affect the circulation of nanoparticles are the particle size, surface charge, and hydrophilicity. The surface modifications of polymeric nanoparticles involve PEGs and polysaccharides.¹¹⁵ In this section, we focus on synthetic and natural polymeric nanoparticles.

4.4 Inorganic nanoparticles

Inorganic nanoparticles typically have smaller particle sizes, narrower size distributions, and surface chemistry that is more suitable for ligand coupling than that of other polymers or LNPs. The most common inorganic nanoparticles are MSNs, which have custom-made homogeneous mesoporous structures, high specific surface areas and pore capacities, selective surface functionality, allow control over morphology, and are biocompatible and biodegradable (Figure 4A). MSNs also have high loading capacities and controlled release properties for therapeutic agents if modified with stimulus-responsive groups, polymers, or proteins.¹⁷⁶ The mesoporous structures of MSNs allow siRNA to be delivered along with other biomolecules, which results in efficient delivery to target tissues and enhanced gene expression. Mora-Raimundo et al.¹⁷⁷ proposed a modified MSN system capable of transporting and delivering the SOST gene, which encodes sclerostin, using siRNA and osteostatin through subcutaneous injections.

Metal nanoparticles, particularly AuNPs, are essential for various biomedical applications¹⁷⁸ as they can be designed in different sizes and shapes. Metal nanoparticles are highly stable, nontoxic, biocompatible, remarkably reactive, and have a large surface area. In addition, they are electrostatically charged and therefore can be functionalized by other biomolecules, such as nucleic acids and drugs.¹⁷⁹ Metal nanoparticles can also protect DNA from degraded by nucleases and transfer various drug molecules, nucleic acids, or vaccines to controlled release target sites to treat intracellular diseases.¹⁸⁰ The functionalization of nanoparticles is significant for reducing their toxic effects, improving their delivery efficiency and accumulation in target tissues, and increasing their stability. More efficient and safer gene carriers based on functionalized metal nanoparticles are being explored. For example, cationic carbo-silane dendrimer-modified AuNPs have been synthesized by Abashkin et al.¹⁸¹ for the delivery of anticancer siRNAs siBCL-xL and siMCL-1 (Figure 4B). Silver NPs are valuable alternatives to AuNPs due to their lower prices and higher reactivity, which increase their range of applications for surface functionalization. Pedziwiatr-Werbicka et al.¹⁸² demonstrated that carbo-silane dendritic surface-modified silver nanoparticles may be used as carriers for anticancer siRNA, such as siBcl-xl, and protect the siRNA from enzymatic degradation to ensure effective cellular uptake.

Magnetic nanoparticles are another form of inorganic nanoparticle. Their magnetism allows enrichment, separation, and directional movement or localization (Figure 4C). Magnetic nanoparticles are safe, have a strong binding capacity, low immunogenicity, can protect DNA, exhibit superparamagnetism, and can bind to large DNA fragments. Iron oxide magnetic nanoparticles, such as magnetite Fe₃O₄ or its oxidized and more stable forms of magnetite γ-Fe₂O₃, are superior to other metal oxide nanoparticles in terms of biocompatibility and stability. They are also the most commonly used magnetic nanoparticles in biomedical applications. The versatility of superparamagnetic magnetic nanoparticles is attributed to their ability to respond to external magnetic fields and function simultaneously and/or differently with bioactive agents.¹⁸³

4.5 Dendrimers

Dendrimers are monodisperse polymers with highly dendritic structures. They comprise an oligomer that is repeatedly and linearly connected by dendrimer units, an inner core, a polymer backbone, and side chains of dendrimer units. Dendritic macromolecules form a dendritic-like structure by repeatedly growing and branching. As the number of polymerized generations increases, the degree of branching expands, and a closed 3D spherical structure is eventually formed. The spatial steric hindrance of the dendritic units as well as the configuration and flexibility of the dendrimers themselves may be regulated by controlling the structures of the dendritic units, the number of generations, and the distance between the polymer backbones. Dendrimers comprise a new type of polymeric nanostructure and are characterized by hyper-branched 3D structures with multiple functional groups on their surfaces, which enhance their activity and make them versatile and biocompatible. Their unique properties, such as nanoscale uniform size, high degree of branching, multivalency, water solubility, available internal cavities, and convenient synthesis methods, make them promising reagents for biological and drug delivery applications.¹⁸⁴

More than 200 dendrimers have been synthesized by the scientific community. These include polypropyleneimine (PPI), PAMAM dendrimers, Fréchet-type dendrimers, core–shell tecto dendrimers, chiral dendrimers, liquid crystalline dendrimers, peptide dendrimers, multiantigen peptide dendrimers, sugar dendrimers, hybrid dendrimers, and polyester dendrimers. Among these, the most commonly used dendrimers are PAMAMs, PLLs, PEI, and PPI.^{184, 185} The surface and internal modifications of PAMAM dendrimers improve their physicochemical properties, cell specificity, and transfection efficiency and can further reduce their cytotoxicity to healthy cells. The three most commonly used modification strategies for PAMAMs include (1) surface modifications with functional groups, (2) hybrid vector formation, and (3) the generation of supramolecular self-assemblies.¹⁸⁶ PPI was the first dendrimer introduced and comprises polyalkylamines with primary amine terminal groups and internally composed of several tertiary amine groups. The ability of PPI dendrimers to bind to DNA is enhanced as the number of generations increases, but their toxicity also increases.¹⁸⁷ Many scholars have chemically modified PPI to improve its performance as a gene vector. For example, Liu et al.¹⁸⁸ used facile fluorination to prepare efficient and low cytotoxicity gene vectors based on PPI dendrimers.

4.6 Extracellular vesicles

EVs are vesicle-like bodies with a double-layered membrane structure, ranging from 40 to 1000 nm in diameter, and are detached from the cell membrane or secreted by the cell. EVs are defined by the International Society for EVs as lipid vesicles secreted by cells outside the cell. According to the process of their generation, release pathways, sizes, contents, and functions, EVs can be classified into three types: MVs, exosomes, and apoptotic bodies^{189, 190} (Figure 5A).

EVs have several advantages as gene delivery systems. First, EVs are multifunctional vectors, so can encapsulate and deliver various biological cargoes. EVs can also naturally cross biological barriers and are able to migrate to tissues or regions without a blood supply. Furthermore, after reaching the target tissue, EVs can remain for a long time. Their low clearance rate can be attributed to their biocompatibility. Another important feature of EVs is that they are genetically modifiable. Their surface proteins can be modified for various purposes. For example, cell- or tissue-targeting peptides can be attached to the surfaces of EVs to reduce systemic toxicity. Several studies have also shown that EV-based targeted gene therapies can improve therapeutic efficacy and safety.¹⁹¹ Kamerkar et al. demonstrated that, unlike liposomes and other synthetic drug nanoparticle vectors, exosomes exhibit an enhanced ability to deliver RNAi and suppress tumor growth. Exosomes contain plasma membrane-like phospholipids and transmembrane and membrane-anchored proteins that enhance endocytosis, facilitating the delivery of the internal contents of the vesicles¹⁹² and reducing their clearance in the circulation. CD47, which is a widely expressed integrin-associated transmembrane protein in exosomal proteins, protects cells from being removed by monocytes.¹⁹³ The use of exosomes also minimizes the cytotoxic effects observed when synthetic nanoparticles are used in vivo.¹⁹⁴

Therefore, researchers have been developing exosome-based delivery systems for gene therapy. For example, Liang et al.¹⁹⁵ designed an exosome-based chondrocyte-targeted miRNA delivery system for the repair of cartilage defects (Figure 5B). Furthermore, Yu et al.¹⁹⁶ developed a liquid natural extracellular matrix enriched with exosomal miR-29 to treat pulmonary fibrosis. Exosomes can also serve as vectors for the delivery of CRISPR/Cas9 plasmids. The capacity of natural exosomes to load plasmids is low; therefore, exosomes must be modified to deliver plasmids. Liposomes can package and deliver large plasmids, but display relatively high levels of cytotoxicity due to the unnatural properties of lipids. Exosome–liposome hybrids may be formed by fusion of the lipid bilayer of the exosomal membrane lipid bilayer with the liposome to allow the encapsulation and delivery of large DNA molecules, such as CRISPR/Cas9 expression vectors, and mitigate liposome-related toxicity. Exosome–liposome hybridization can therefore expand the applications of exosomes in drug delivery. Wan et al.¹⁹⁷ developed an exosome-based nanoplatform that allows RNP-based CRISPR/Cas9 genome editing therapies for liver disease. RNPs can be efficiently loaded into human hepatic stellate cell (LX-2) derived exosomes by electroporation, and exosomes loaded with Cas9-RNPs, or exosome^RNP, can be specifically delivered to the liver. Exosome^RNP has been shown to target the p53 upregulated modulator of apoptosis, cyclin E1, and K(lysine) acetyltransferase 5 in gene therapies for mouse models of acute liver injury, chronic liver fibrosis, and hepatocellular carcinoma (HCC), respectively.¹⁹⁷

5.1 Microneedles

Microneedles are micron-sized (less than 1000 μm), conical, radial, or multifaceted puncturing protrusions that can be solid, hollow, coated, or dissolvable. They work by penetrating the stratum corneum of the skin to create hundreds of reversible microchannels without causing skin damage or pain (Figures 6A and B). Microneedles provide safe passages for therapeutic substances, especially large molecules, to bypass barriers and deliver various therapeutic compounds.¹⁹⁸

Wan et al.¹⁵ reported a microneedle patch that mediates the synergistic transdermal delivery of CRISPR/Cas9-based genome editing agents and glucocorticoids for the effective treatment of inflammatory skin diseases. The results showed the microneedle patch system could degrade rapidly in vivo after penetrating the skin to release therapeutic substances. Blood biochemical functional indices have also validated that microneedle patch-mediated therapy has minimal systemic toxicity and is safe for clinical use.¹⁵ Lara et al.¹⁹⁹ further demonstrated that siRNA delivered through microneedle arrays could reduce the expression of the targeted endogenous gene CD44 in human skin xenograft models.

The use of solid, coated, lysis, or hollow microneedles for gene therapy with nucleic acids as cargo needs to be improved. Microneedles alone may not facilitate the delivery of genetic material better than intradermal injections, but the combination of microneedles and electroporation is more effective than injections alone. The combination of microneedles with other delivery techniques may therefore also be effective. Compared with other physical delivery vehicles that have disadvantages such as skin damage, high manufacturing costs, and not being preferred by patients, microneedles perform well in these aspects. Moreover, various bioactive ingredients can be added to the formulations of dissolvable microneedles to enhance their effects. Microneedles are also painless when used and, therefore, better for DNA-based vaccinations than other forms of immunization that require multiple injections and cause pain for days or even weeks.¹⁹⁸

5.2 Microspheres

Microspheres are flowable spherical particles that can be loaded with specific substances. They are also biocompatible and biodegradable. Their particle sizes range from 1 to 250 microns. There is a wide variety of microsphere formulations that can be classified into three categories according to their structure: pore-forming, double-layered, and magnetic microspheres. Commonly used methods for microsphere preparation include evaporation of emulsification solvent, phase separation, spray drying, electro-spraying, and microfluidic methods.^{200, 201} Different manufacturing processes can produce microspheres with different structural characteristics to allow microspheres to be used in different medical applications.²⁰²

Materials for the preparation of microspheres include glass (such as that made of silicate, borate, or phosphate), ceramics, and polymers. Polymer-based porous microspheres have been extensively investigated for drug release and as delivery vehicles for other biological components. Polymer-based microspheres are mainly classified into natural polymeric microspheres (e.g., proteins, collagen, chitosan, and alginate) and synthetic polymeric microspheres (e.g., PLGA, PLA, PGA, and PCL) that can be hydrolytically degraded in vivo.²⁰¹ Diez et al.²⁰³ prepared different PLGA microspheres encapsulating DNA using two formulation schemes. The results showed that the formulations, molecular weights, and compositions of the polymers used to prepare the particles are important determinants of the sizes and encapsulation and release behaviors of these delivery systems (Figure 6C). These microspheres have other advantages, as they can deliver plasmid DNA at a rapid and controlled rate.²⁰³

Microspheres are excellent vehicles for the establishment of oral nucleic acid delivery systems. They can protect the encapsulated material during transport through the gastrointestinal tract and facilitate efficient uptake and intracellular transport to the desired target site, while being safe and well tolerated. Attarwala et al.²⁰⁴ developed a multicompartment system based on an oral system of nanoparticles within microspheres for in vivo gene and siRNA delivery to treat colitis in mice (Figure 6D). The results exhibited effective transgene expression or gene silencing, as well as downstream anti-inflammatory effects.²⁰⁴ In addition, Guo et al. designed microsphere systems that could slowly release loaded substances locally and inject them into intervertebral discs to exert sustained therapeutic effects. This microsphere-based delivery system can alleviate low back pain, inhibit tissue degeneration, and promote disc regeneration or repair by targeting the delivery of cells, drugs, bioactive components, and genetic modifiers to the interior of the disc tissue.²⁰⁵

There are also limitations to the use of microspheres. When designing and developing new microsphere formulations, it is difficult to reproduce or predict the full impact of the human environment on injectable microspheres, including changes in pH, temperatures, and enzymes.²⁰⁶ Synthetic polymeric microspheres have been proven to be safe and biocompatible, but migrate from the injection site, which leads to potential risks of embolism and further organ damage.²⁰⁰

5.3 Hydrogels

Hydrogels may be formed through the polymerization of small molecules with numerous hydrophilic groups. Due to their water-rich composite structure and the cross-linking of long polymer chains, hydrogels exhibit elastic, adhesive, and mechanical properties that make them suitable as biomaterials.^{207, 208} Hydrogels can be classified into three main categories according to their size: macro/bulk hydrogels, microgels, and nanogels (Figure 7A).²⁰⁹ DNA hydrogels at the micro- or nanoscale are more responsive to stimuli and have enhanced delivery efficiencies. Based on the origins of the monomers, hydrogels can also be divided into two categories: naturally occurring hydrogels, such as chitosan, HA, gelatin, collagen, fibrin, and alginate, and synthetic hydrogels, which are based on hydrophilic synthetic polymers such as PAA, polyvinyl alcohol (PVA), PAM, and PEG.^{210, 211} DNA hydrogels comprise a hydrophilic polymer network of cross-linked DNA strands. Unlike conventional vectors, DNA hydrogels can improve cell transfection efficiency, protect biomolecular payloads from nuclease or proteases degradation, and reduce nonspecific distribution.²¹² DNA hydrogels are promising carrier materials that are biocompatible, simple to prepare, have tunable mechanical properties, and allow for a controllable phase transition. In addition to enabling the in situ encapsulation of drugs, DNA hydrogels allow for molecular recognition within target regions and the integration of multiple components for synergistic therapy.²¹³ In 1996, Nagahara and Matsuda²¹⁴ designed the first DNA-based hydrogel by cross-linking single-stranded DNA grafted onto polyacrylamide chains. Considering that DNA strands are programmable, complementary, and chemically modifiable, they can be manipulated to form various DNA building blocks with unique geometries and form highly predictable and structured DNA networks. In addition, 3D scaffolds in DNA hydrogels provide mechanical rigidity and numerous attachment sites, enhancing the function of hydrogels as stable immobilization matrices for bound nanoparticles or molecular components.²¹² Due to their biocompatibility, porosity, sequence programmability, and tunable versatility, DNA hydrogels have been extensively studied in bioanalyses and biomedicine.²¹²

Recently, Song et al.²¹⁵ designed a thermosensitive hydrogel loaded with a gene complex and implanted it into a postoperative cavity to inhibit the immune escape of residual tumor cells after surgery. A novel nonviral vector, G5-BGG, was synthesized and formed a gene complex with an shRNA plasmid. A PLGA–PEG–PLGA hydrogel loaded with G5-BGG/shRNA871 and combined with temozolomide caused the downregulation of CD47 protein expression, increased macrophage infiltration into residual tumors, and significantly prolonged the survival of mice. This study demonstrates the potential use of hydrogel vectors in gene therapies for glioblastoma.²¹⁵ Hydrogels possess good flexibility and biocompatibility, although their mechanical properties are weak. Additionally, conventional hydrogels lose their original properties and suffer irreversible damage when they are damaged or fatigued by use, thus severely limiting the expansion of hydrogel applications. Therefore, there is a requirement for hydrogels with excellent deformability. As a functional hydrogel that can repair itself after damage, self-healing hydrogel can be realized through external stimulation (light, heat, pH adjustments, and self-healing agents) or the interaction of functional groups within the hydrogel (dynamic covalent bonding, noncovalent bonding interaction).²¹⁶ The inflammatory response to self-healing hydrogels is relatively weak.²¹⁷ Chen et al.²¹⁸ constructed hydrogels of aldehyde HA (HA-CHO) and poly(amidoamine) PAMAM/siRNA complexes. This novel injectable self-healing hydrogel was found to have significantly alleviated disc inflammation while slowing disc degeneration by efficiently and stably silencing IFN gene expression in nucleus pulposus cells.²¹⁸

5.4 Scaffolds

Scaffold-based systems are effective methods for improving the efficiency of CRISPR/Cas9 targeting gene editing delivery systems. Ho et al.²¹⁹ proposed an innovative scaffold-mediated CRISPR/Cas9 delivery system by anchoring the CRISPR/Cas9 complex on the surface of an injectable scaffold intended for local delivery. LNP-Cas9-RNPs and chemokine CXCL12α were loaded onto a mesenchymal stem cell membrane-coated nanofiber (MSCM-NF) scaffold that mimicked the bone marrow microenvironment, targeting the interleukin (IL)-1 receptor accessory protein (IL1RAP). Continuous local delivery of the Cas9/IL1RAP sgRNA via the CXCL12α-loaded LNP/MSCM-NF scaffolds provided an effective strategy to attenuate LSC growth to improve the treatment of acute myeloid leukemia.²¹⁹

Scaffold-based sustained-release drug delivery systems not only provide long-lasting effects but also minimize the risk of toxicity during systemic application (Figure 7B). Short durations and inefficient deliveries have been the main limitations of siRNA due to its sensitivity to nucleases and poor intracellular cytoplasmic delivery efficiency.²²⁰ Nelson et al. uniformly doped pH-responsive polymeric micellar nanoparticles in porous, noncytotoxic, biodegradable, and injectable polyester polyurethane (PU) scaffolds to optimize siRNA delivery. siRNAs are depleted in 1 week in rapidly dividing cells,²²¹ and they have also been found to be released through diffusion-based mechanisms across 3 weeks. These findings demonstrate the durable effects and controlled release potential of scaffolds.²²²

Scaffold-based gene therapy offers a promising approach to tissue engineering as it allows for the transfection of cells to enhance the sustained expression of target proteins or silencing of target genes associated with bone and joint diseases.²²³ Costard et al.²²⁴ used MgAl-NO₃ layered double hydroxide loaded with miRNA nanoparticles and doped them into collagen-nanohydroxyapatite scaffolds. This resulted in the successful overexpression of target miRNA in mesenchymal stromal cells (MSCs) and demonstrated an effective miRNA delivery platform for gene therapies in regenerative medicine.²²⁴ Collagen-hydroxyapatite scaffolds can also be used as platforms for the delivery of nanoparticles composed of plasmid DNA. Such scaffolds have been found to facilitate induction of transient gene expression in rat bone marrow MSCs (BM-MSCs) through PEI–plasmid DNA nanoparticles encoding the IL-1 receptor antagonist. These scaffolds improved the mineralization of the BM-MSCs and demonstrated their potential for future therapeutic applications in vivo (Figure 7C).²²⁵

6.1 Genetic vaccines

The recent COVID-19 pandemic caused RNA-based therapies and vaccines for infectious diseases to be brought to the forefront of research worldwide.³⁴ The pandemic further led to the marketing of the first approved mRNA vaccine. The COVID-19 BNT162/Comirnaty vaccine from BioNTech/Pfizer and the COVID-19 mRNA-1273/Spikevax vaccine from Moderna were both approved by the US FDA for emergency use in 2020 and re-approved in 2021 and 2022, respectively.²²⁶ RNA vaccines against hepatitis B, influenza, and respiratory syncytial viruses have also been developed and are undergoing clinical trials.³⁴

LNPs are the most widely used nonviral delivery systems for nucleic acid vaccines.³⁴ The delivery of RNA to target cells via LNPs allows to treat various diseases, including COVID-19.²²⁷ The composition of LNPs in COVID-19 vaccines includes ionizable lipids, cholesterol, helper lipids, PEG2000 lipids, and mRNA.²²⁸ However, the excipient, PEG, may cause adverse effects. PEG has been shown to trigger allergic reactions, including anaphylaxis and contact urticaria, in some vaccine recipients.²²⁹ Therefore, alternative chemicals to PEG should be identified to improve mRNA vaccine vectors or develop new delivery vehicles based on polymers or protein particles, such as protamine. Storage and transport in the form of lyophilized powder is the most commonly used method for mRNA drugs. For mRNA vaccines encapsulated in liposomes, the addition of sucrose can reduce the crystallization temperature and the amount of ice formation during the aqueous phase. Furthermore, the interfacial film formed by the adsorption of sugar molecules at the oil-water interface can also increase the spatial potential resistance and electrostatic repulsion between droplets, which can in turn improve their low-temperature and mRNA stability. The addition of cryoprotectants (such as alginate, sucrose, and mannitol) to the mRNA-LNP solution and storage in liquid nitrogen can ensure the efficiency of an optimal mRNA delivery system in the long term^{230, 231} (Figure 8).

Vaccination is an important way to control and eliminate potentially infectious diseases and cancers. Traditional live attenuated vaccines have been used for a long time, but serious safety concerns regarding toxicity and the risk that pathogens mutate back to an infectious state have limited their use.^{232, 233} Nucleic acid vaccines have numerous advantages. For example, they avoid the problems associated with recombinant protein vaccines, such as improper protein folding or high protein purification costs. They are also free from the infection risks associated with attenuated or inactivated vaccines that produce self-activating infectious organisms. Nucleic acid vaccines can also activate humoral and cellular immune responses, greatly enhances the protective immune response. Furthermore, the efficiency and stability of mRNA transactivation can be improved by various chemical modifications, increasing immunity.²³⁴ Biomaterial-based antigen delivery systems have emerged as innovative strategies to improve the efficacy of genetic vaccines. Antigen delivery systems are divided into two broad categories: one is based on microparticle-based delivery systems, including microparticles and nanoparticles, and the other involves hydrogel- and scaffold-based delivery systems, such as solid implants, scaffolds, and hydrogels.^{235, 236} Depending on the specific properties of the delivery system being used, different functions can be performed. For example, they may ensure stability by maintaining antigen release and/or by providing reservoirs at the injection site, protecting materials from degradation, and delaying the antigen clearance from the injection site.

6.2 Gene therapy for clinical diseases

Gene therapy is widely used in the treatment of clinical diseases, and in this section, we summarize the gene therapy strategies based on biomaterial vectors in recent years (Table 2).