Delivery of gene editing therapeutics

Abstract

For the past decades, gene editing demonstrated the potential of attenuating each of the root causes of genetic, infectious, immune, cancerous, and degenerative disorders. More recently, Clustered Regularly Interspaced Short Palindromic Repeats-CRISPR-associated protein 9 (CRISPR-Cas9) editing proved effective for editing genomic, cancerous, or microbial DNA to limit disease onset or spread. However, the strategies to deliver CRISPR-Cas9 cargos and elicit protective immune responses requires safe delivery to disease targeted cells and tissues. While viral vector-based systems and viral particles demonstrate high efficiency and stable transgene expression, each are limited in their packaging capacities and secondary untoward immune responses. In contrast, the nonviral vector lipid nanoparticles were successfully used for as vaccine and therapeutic deliverables. Herein, we highlight each available gene delivery systems for the treatment and prevention of a broad range of infectious, inflammatory, genetic, and degenerative diseases.

Graphical Abstract

Lipid nanoparticles are assembled by microfluidic mixing of lipids and cholesterol then deployed for nucleic acid cargo cell and tissue delivery.

Introduction

Genome editing can revolutionize how biological processes are harnessed for therapeutic gain. Engineered nucleases facilitate gene-editing serving to improve the diagnosis, prevention, and elimination of infectious, inflammatory, degenerative, cancerous, and genetic diseases. In the last decades, significant progress was made in creating mega nucleases, transcription activator-like effector nucleases (TALENs), and zinc-finger nucleases (ZFNs) ¹.^. Notably, Clustered Regularly Interspaced Short Palindromic Repeats-CRISPR-associated protein 9 (CRISPR-Cas9) used as a genome editing tool has revolutionized genetic manipulations to improve the diagnosis, treatment, and prevention of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), methicillin-resistant staphylococcus aureus (MRSA), human immunodeficiency virus (HIV) excision, stem cell and cancer therapeutics. The promise began in 2007 when Mario R. Capecchi was awarded the Nobel prize in physiology or medicine for this groundbreaking discovery of DNA recombination ². The specificity of gene editing was improved by introducing double-stranded breaks (DSBs) at the targeted site in dsDNA ³ and by using the host cells’ own DNA-repairing mechanisms ⁴. Two major DNA repair pathways in eukaryotic cells are, non-homologous end joining (NHEJ, which is error-prone) and homology-directed repair (HDR, which affords accurate gene editing). While the use of CRISPR-Cas9 as a gene-editing tool has come a long way the major limitation in translating it use to the clinic centers on targeted and off target delivery. Approaches available include the use of viral vectors, virus-like particles (VLPs), nonviral polymeric nanoparticles, and lipid nanoparticles (LNPs). Amongst these LNPs elicit excellent cellular uptake (Figure 1). LNPs are also safer than viral vectors with limited immunogenicity. Additionally, the lipid composition of LNPs can be optimized to delivery efficiency as its external surface can be modified for on-target delivery. Such modifications limit off-target toxicities ⁵.

Figure 1.

Schematic for the LNP delivery of RNA-based CRISPR-Cas9 (sgRNA and Cas9 encoding mRNA). LNPs enter the cells by receptor-mediated endocytosis or hydrophobic interactions. The acidic pH in the endosome anionic interaction with the LNP ionizable lipids and RNAs (Cas9 mRNA and guide RNA) followed by maturation in endosomal environment. These are released into the cytosol following endosomal escape. The Cas9 mRNA is translated to proteins, which forms complexes with the guide RNA (gRNA). The CRISPR-Cas9 complexes are trans-ported to the nucleus. CRISPR-Cas9 activities occur inside the nucleus resulting in the desired gene-editing on targeted locus. In attempts to address therapeutic safety, the review provides a comprehensive review of CRISPR-Cas9-based gene editing on how this can optimize the broad applicability of LNPs including targeted cells and tissue delivery (The figure was created internally by the authors).

Genome Editing Tools

ZFNs, TALENs and CRISPR-Cas9

TALENs and ZFNs are gene editing tools. ZFNs, was the first programmable nucleases, consisting of a non-specific nuclease, FokI (a type II S restriction enzyme from Flavobacterium okeanokoites), fused to a zinc-finger DNA-binding domain ⁶. The zinc finger proteins are formed by binding tandem arrays of Cys₂His₂ residues (two invariant pairs of cysteine and histidine amino acid residues) with zinc (II) ions, wherein each zinc finger (ZF) can recognize a 3–4 base pair (bp) sequence ^7. The ZFNs bind in the major groove of the DNA double helix and can be modified to identify different DNA sequences ^6,8. FokI nucleases cleave the DNA into dimer form; therefore, the ZFNs are designed in pairs to address each target ^8,9. The specificity of the ZFNs is primarily determined by the number of fingers in zinc finger proteins and the extent of interaction of the nuclease domain ¹⁰. Three- and four-finger ZFNs have been successfully used for the genomic cleavage, targeting 18–36 bp sequences per ZF pair ⁸. ZFNs have been bioengineered with linkers that allow for increasing the configurations to selectively identify diverse targets ¹¹. ZFNs have been used for various therapeutic applications, including the generation of HIV-resistant genotypes, by disrupting the CCR5 co-receptor in CD4⁺ T-cells ¹². However, the generation of on-target specificity using ZF motifs is costly ⁷. This has led to the development of TALE-based nucleases. TALENs were developed by fusion of transcription activator-like effectors (TALEs), a protein from Xanthomonas bacteria) with the catalytic domain of the FokI restriction endonuclease ¹³. Structurally, TALEs consist of an N-terminal, a C-terminal, and a highly conserved 30–35 amino acid residue-long central repeat region. The DNA binding specificity of the TALE proteins is associated with a variable pair of residues located at positions 12 and 13 within each repeat of the central tandem ¹³. By rearranging or substitution of each residue of each pair, the specificity for a nucleotide can be modified to achieve optimal selectivity. For example, when the residue pair is NG (where N is asparagine followed by glycine), the TALEs specifically bind to thymine. NI (where N is asparagine followed by isoleucine) recognizes adenine, whereas HD (where H is histidine followed by aspartic acid) recognizes cytosine. However, TALENs are best in identifying methylation-dependent gene modifications ¹⁴. Analogous to ZFNs, TALENs are also used in pairs. Here FokI cleaves the DNA as a dimer. Apart from FokI restriction endonuclease, the TALEs have been fused with other endonucleases (PvuIII) and meganucleases (I-TevI, I-OnuI). Most do not require dimerization which reduces the number of inter-biomolecular interactions. As these have sequence specificity they limit target diversity ¹⁵ and improve specificity ¹⁶. TALENs and ZFNs show different mutation patterns. TALEN-induced mutations result in a higher frequency of deletions, whereas ZFNs lead to equal insertions and deletions ¹⁷. A notable feature of TALENs is its ability to generate Universal Chimeric Antigen Receptor 19 T-cells (UCART19) through the editing of CD52 and T-cell receptor (TCR). This has been used in targeting B-cell acute lymphoblastic leukemia ¹⁸ and demonstrating the translational potential of gene-editing therapy in clinical settings. In this instance, two infants with no feasibility of manufacturing autologous CAR19 T-cells were treated successfully with TALENs for B-cell acute lymphoblastic leukemia ¹⁸. Although TALENSs are difficult to deliver this study established proof-of-concept for their use ¹⁹. Both are hybrid proteins that have been developed to target and cleave the respective dsDNA. The major disadvantage of ZFNs and TALEN lies in their design, as each new target requires considerable time and labor. In contrast, the CRISPR-Cas9 system is an improvement over the former as it demonstrates design feasibility, reduced cost, and improved specificity ²⁰. Designing a protein is relatively challenging for ZFNs, TALEN-mediated targeted DNA binding compared to the construction of a related CRISPR-Cas9 complex. The latter requires changes only in the single-guide RNA (sgRNA) sequence. As such, the ease in its development makes its usage feasible. Moreover, a recent study which compared the specificity of the three most used gene-editing tools ZFNs, TALENs, and CRISPR-Cas9 in a genome-wide sequencing technique (GUIDE-seq/genome-wide unbiased identification of double-stranded breaks enabled by sequencing) to evaluate any off-target activity revealed that SpCas9 (Cas9 from Streptococcus pyogenes) was more specific, efficient, and safe in gene editing than ZFNs and TALENS. This evaluation was made during in vivo targeting of the human papillomavirus 16 (HPV16) ²¹. The specificity of SpCas9 is due to its sgRNAs ^21,22. CRISPR-Cas9 serves as a “best” gene-editing tool based on simplicity.

CRISPR-Cas gene editing tool

History

The first description of CRISPR was made by iap gene (a 29 bp direct repeat sequence with 32 bp spacers and isozyme of alkaline phosphatase) studies ²³. Similar interrupted direct repeats (36 bp direct repeats interspersed by 35–41 bp spacers) are in Mycobacterium tuberculosis ²⁴, archaeal organisms (Haloferx and Haloarcula species), and prokaryotes. All are short regularly-spaced repeats (SRSRs) ²⁵. The term “CRISPR” was introduced in 2002 ^26,27. Shortly thereafter, the cas genes, present adjacent to the CRISPR locus were discovered. These have known helicase/exonuclease activities with a functional relationship to the CRISPR locus ^26,28. A critical insight came when the nucleotide spacers were found to have homology with plasmid and viral genes ^29,30 and leading to the notion that CRISPR cascades alter adaptive immunity ³¹. Integrating the new bacteriophage DNA into the CRISPR array of Streptococcus thermophilus where it recognized a wave of attacking bacteriophage supports the idea that CRISPR affects immunity ³². This supported CRISPR’s role for “genetic vaccination.” The findings spurred studies of spacer sequences derived from invading phages were transcribed into CRISPR RNA (crRNAs), which guided the Cas proteins to its targets ³³. It was later discovered that Cas9 can affect cleavage with crRNAs. The discovery of a small RNA, termed trans-activating CRISPR RNA (trcrRNA), which forms a duplex with crRNA and guides Cas9 to its target ³⁴.

A major breakthrough was in the discovery of a programmable single-guide RNA (sgRNA). This combined with trcrRNA:crRNAs highlights the CRISPR-Cas system as a means for RNA-programmable genome editing ^4,35. The first successful application of the CRISPR-Cas9 was harnessed by the dairy industry. It was used to immunize cultured bacteria against bacteriophages. Later, CRISPR-Cas9 systems were used in mammalian cells ^36,37. Since then, CRISPR-Cas was applied to human, yeast, nematodes, mice, monkeys, rabbits, and other eukaryote cells beyond microbial pathogens ^22,36–39. CRISPR has also been used in clinical diagnostic and therapeutic trials (Figure 2).

Figure 2.

CRISPR scissors are applied as nanotheranostics for treatment of emerging infections. Molecular scissors are delivered in vivo by lipid nanoparticles to edit the genome at the desired site. CRISPR nanodiagnostics to diagnose SARS-CoV-2 using lateral flow strips after collecting a nasopharyngeal swab. CRISPR nanotherapeutics can be applied for the inhibition of SARS-CoV-2 and bacterial infections (The figure was created internally by the authors).

Mechanisms of Action

CRISPR systems were classified based on the structural and organizational variations of the Cas genes. Class 1 contains multi-protein effector complexes while class 2 is a single protein. These two classes are further subdivided into six types (type I-VI) with at least 33 subtypes, at which more are continuously being catalogued ⁴⁰. The class 2 system requires only SpCas9. The SpCas9 protein has two nuclease domains, HNH (named for the characteristic asparagine (N) bookended by single histidine (H) residues) and RuvC-like (named for an endonuclease similar to an E.coli protein) domain, wherein each cleaves the complementary and non-complementary DNA strands ⁴¹. For CRISPR-Cas9, the target DNA is followed upstream by a short sequence termed “protospacer adjacent motif” (PAM). The motif is a short 2–6 bp sequence required for Cas9 nuclease activity ⁴. Approximately ten different CRISPR-Cas proteins were remodeled for gene editing. SpCas9 requires a simple PAM sequence “NGG” (where N is any nucleobase followed by two guanine residues), which presents itself in high frequency at the target DNA. This is to afford the CRISPR-Cas9 system simplicity for its use in diverse applications ⁴¹. Mechanistically, the CRISPR-Cas9 system contains two major activity-based components. This includes a DNA-binding domain, that mediates sequence-specific DNA recognition and binding (sgRNA) and an effector domain that enables DNA cleavage. It may also regulate transcription near the binding site (Cas9 endonuclease) 20 nucleotide sequence complementary to the target DNA site and a tracrRNA sequence critical for a unique dsDNA cleavage. The sgRNA associates with the target DNA site by base complementarity, and Cas9 cleaves the dsDNA to generate a double-stranded break (DSB) leads to activation of a native DNA repair mechanism. These repair pathways are exploited for genomic modifications to create either genetic “knock-outs” using NHEJ or genetic “knock-ins” (both for accurate gene editing) by the homology-directed repair (HDR) pathway ⁴¹ (Figure 3). Both repair mechanisms occur in many cell types. HDR is triggered only when the homologous sister chromatid is available. In other conditions, where the cells are non-dividing, NHEJ is activated to repair the DSBs ⁴². The HDR pathway can copy a designed DNA template with the desired sequence inserted into the target cleavage site. NHEJ inhibits gene expression and the gene knock-out process that results in premature stop-codon insertions or frameshift mutations ⁴³. However, when the purpose is to introduce specific mutations and targeted “knock-ins”, DSBs must be repaired using the HDR pathway. The functional efficiency of HDR in comparison to that of NHEJ is relatively low in mammalian cells ⁴⁴. Methods to improve the HDR efficiency include use of overlapping homology arms, modified Cas9, and intrinsic genetic or chemical modifications, combinatorial approaches were found to improve HDR efficiency. This was done by simultaneously using single-stranded DNA oligonucleotides (ssODN), double sgRNAs, and cell cycle synchronization ⁴⁶. Importantly, after determining the roles that CRISPR-Cas nucleases play in bacterial defense systems its potential for editing genes in mammalian species were made.

Figure 3.

Mechanism of Cas9 endonuclease to improve efficiency. (A) WT Cas9 endonuclease cleaves the dsDNA after the sgRNA binds to the target DNA, leading to double stranded breaks (DSBs). DNA repair mechanisms introduce random insertions/ deletions (indels) or accurate gene-editing (when homologous DNA is present) that introduce the desired changes, (B) dCas9 (generated by D10A and H48A mutations in WT Cas9), lacks dsDNA cleavage activity and binds specifically with the target DNA that can be used to regulate the transcription of the desired gene. When dCas9 is fused with a transcriptional activator (represented by green sphere), the sgRNA (orange) guides the dCas9 to activate the promoter, thereby resulting in gene expression (CRISPRa). Likewise, attachment of transcriptional repressor (represented by red sphere) results in suppressing the gene expression (CRISPRi), (C) To improve the editing efficiency, deaminase is attached to the nCas9 or dCas9. (CBEs); converts cytosine to uracil followed by DNA repair and replication resulting in conversion of uracil to thymine, thereby encouraging the conversion of C·G to T·A. Similarly, ABEs results in A·T to G·C conversions via deaminating adenine to Hypoxanthine/Inosine. WT=Wild Type, nCas9=Cas9 nickase, dCas9=dead Cas9, CRISPRa=CRISPR activation, CRISPRi=CRISPR interference, ABEs=Adenine Base Editors, CBEs=Cytosine Base Editors, H=Hypoxanthine (The figure was created internally by the authors).

Type of CRISPR Cargos

For CRISPR-Cas9-based genome editing, mainly three forms of cargo choices are available (Figure 4). These cargoes include sgRNA/Cas9 protein complex (commonly referred to as the ribonucleoprotein (RNP) complex); plasmid DNA encoding sgRNA, and Cas9; and RNA system consisting of Cas9 mRNA and sgRNA. The aim of using a cargo is that the functioning Cas9 sgRNA complex must be presented inside cell nucleus for gene editing. The choice of payload often directs which delivery vehicle (described in the delivery section) is to be used for achieving genome editing. This section describes the advantages and disadvantages of each type of cargo and the delivery options available.

Figure 4.

Schematic of plasmid, protein, and mRNA cargoes for CRISPR-Cas9 expression. (A) Plasmid-based expression; two plasmids (encoding Cas9 and sgRNA separately) or a single plasmid (encoding Cas9 and sgRNA together) can be used, (B) Ribonucleoprotein (RNP)-based expression; in vitro transcribed sgRNA can be complexed with Cas9 protein, (C) Cas9 mRNA-based expression; an mRNA encoding Cas9 endonuclease (consisting of 5’ cap, UTRs, and 3’ poly-A tails) along with sgRNA can express CRISPR-Cas9 (The figure was created internally by the authors).

Plasmid DNA

The DNA encoding Cas9 and sgRNA can be packaged in a single or two different plasmids which can be used to express the binary CRISPR-Cas components ⁴⁷ (Figure 4). Plasmid construction is a time-consuming and labor-intensive process. For the synthesis of Cas9 protein, both transcription and translation are needed, which leads to slow-onset editing; however, plasmids express the Cas9 protein for a longer duration of time. Sustained expression, wherein the goal is to introduce stable mutations can often result in increased off-target activity ⁴⁸ and a potential risk of insertional mutagenesis ⁴⁹. Despite its drawbacks identified above, plasmid DNA-based expression is a cost-effective alternative. Another advantage of using plasmid-based expression is the ability to achieve multiplexed gene editing: that is, different sgRNAs can be expressed using a single plasmid, thereby simultaneously ensuring multiple gene site editing and large genomic deletions ⁵⁰. Moreover, the selective transcription of Cas9 mRNA and sgRNA can be driven by a cell type-specific promoter. Care must be taken in choosing the proper promoter to use; it must be biocompatible with the host cells. For Cas9 expression, the promoter for the β-actin gene in chickens can be used (pX330 and pX458) ³⁸. Recently, studies have focused on using cell-specific promoters that includes macrophage (CD68)-specific promoters. These serve to drive Cas9 expression and can be delivered by nanoparticles ⁴⁵.

For multiplexing sgRNAs, each sgRNA requires its own promoter for transcription. By doing so, the plasmids can allow for the expression of 2–7 different sgRNA transcripts. ^50,51. For the plasmid DNA-based expression of CRISPR components, many delivery options are available. Hydrodynamic injections have been used in mice to deliver the plasmid cargoes. However, this technique is not currently feasible for human use due to the large infusion volume required ⁵². Also, hydrodynamic injections do not lend themselves to a guarantee that cargo will reach the nucleus of the targeted cells. An electroporation method to deliver cargoes directly to the nucleus of cells during cell division has also been used for plasmid DNA-based delivery in mouse stem cells ⁴⁸. Different types of lipids like polyethyleneimine (PEI) polyplexes, PAMAM dendrimers, PEI-coated DNA nanoclaws have also been studied to deliver plasmid DNA-based cargo carriers ⁵³, but methods entailing such are not yet suitable for applications in human. The most successful delivery system for plasmid DNA that encodes for CRISPR-Cas9 components has emerged from the use of viral vectors ⁵⁴. Although plasmid DNA-based delivery for effectuating CRISPR-Cas9 holds tremendous promise, considerable additional research is required to understand any associated off-target activity that results from the stable expression of plasmid DNA encoding Cas9 before clinical applications are considered.

Ribonucleoprotein (RNP)

For the generation of RNPs, Cas9 and sgRNAs (custom-made) can be either purchased or synthesized ⁵⁵. In research laboratory settings, sgRNAs [generated by in vitro transcription (IVT)] are incubated with the Cas9 protein (which is expressed by E. coli using Cas9-containing plasmids) to generate the RNP complex. The RNP-based delivery of CRISPR components offers several advantages. Since they are ready-for-use protein complexes, no intracellular transcription or translation is required (Figure 4). Accordingly, without consideration to delivery, the efficiency of RNP-based gene editing is superior to that of the plasmid DNA editing ⁵⁶. The RNP complexes show transient expression due to biological degradation over time, which can help in reducing the likelihood of “off-target” genome editing. However, this same property of RNPs (degradation) renders them as less than ideal for use in cells where stable expression of Cas9 is required. Among the various common delivery methods (physical, viral, and non-viral) available for the trafficking of CRISPR-Cas9, physical delivery, which includes electroporation, are the most efficient for RNP-based expression systems ^55,56. Electroporation has been successfully used in cell culture systems and mice zygotes ⁵⁷. Non-viral carriers for RNP delivery show great promise for high-efficiency gene editing because they allow for the ability to enhance the binding affinity of the Cas9 protein. It was suggested that the Cas9 protein can be modified by attaching negatively-charged tags to improve the interaction of RNP with non-viral-based carriers ⁵⁵. Recently, the use of a vapor nanobubble (VNB) photoporation technique was used that allows RNP delivery into different cell types. These include primary human T- and mesenchymal stromal cells. They exhibited 60 and 30% of gene knock-out levels in mesenchymal stromal cells and primary human T-cells ⁵⁸. Although RNP-mediated expression appears promising, more research is required to better understand their immunogenicity and the stability for in vivo delivery.

mRNA

The third type of cargo consists of mRNA encoding the Cas9 nuclease protein. This protein gets directly translated within the cell cytosol. The sgRNA is delivered as synthetic oligonucleotides (Figure 4). As mRNA is susceptible to cell nucleases, various modifications have been made to the chemical constituents comprising mRNA. This enables sustained expression ³⁷. Exogenous mRNA is highly immunogenic and activates multiple pattern recognition receptors (PRRs) and RNA sensors. This includes toll-like receptors (TLRs) 7 and 8, a retinoic acid-inducible gene I (Rig-I), RNA-dependent protein kinase (PKR), melanoma differentiation-associated protein 5 (MDA5), and 2′-,5′-oligoadenylate synthetase/RNase L axis ^59,60. The cell immune system recognizes triphosphate, diphosphate, uridine-rich sequences, double-stranded regions, and abnormal 3′- and 5′-cap structures of RNA ⁵⁹. To prevent the activation of innate immune responses, chemical modification of the exogenous mRNA was considered by adding eukaryotic-mimetic mRNA caps at the terminals of exogenous mRNA and replacing uridine with various bases, including pseudouridine (Ψ), N1-methylpseudouridine (m1Ψ), 5-methylcytidine (m5C), 5-hydroxymethylcytosine (5hmC), 5-methyluridine (m5U), and 2-thiouridine (s2U), ⁵⁹. Moreover, it was suggested that chemical modifications in the sgRNA oligonucleotides can further enhance efficiency of genome-editing ⁶¹. For example, such modifications can be applied to CD34⁺ hematopoietic stem and progenitor cells and human primary T-cells.

The addition of a 5′ cap with a 3′ poly-A tail to the sgRNA enhances its stability and allows it to exhibit more DSBs in human cells than an unmodified equivalent sgRNA ⁶². The association of a hairpin secondary structure-engineered sgRNAs (hp-sgRNAs) with five variants of Cas9 exhibited greater specificity in editing the genome at the targeted sites ⁶³. Generation of rationally chemically modified sgRNA improved site-specific genome editing and gene disruption activities in human cells. 2′-O-methylation base change at the 2′-position of the sugar molecule in RNA favors an energetics-favorable conformational change in RNA. Creating sgRNA with modifications such as 2′-O-methylated nucleotides at 5′ and/or 3′ termini of RNA having a phosphorothioate (PS) backbone greatly enhanced CRISPR-Cas9-mediated genome-editing with increased stability and resistance to exonuclease ⁶⁴. Chemically-modified sgRNA (such as 2′-O-Me- and PS-modifications of three nucleotides at both ends of sgRNA (5′- and 3′-sgRNA) have enhanced gene-editing activity in mammalian cells compared to using unmodified sgRNA  However, in some cases, the addition of 2 to 3 nucleotides at the 5’ end of However, in some cases, the addition of 2 to 3 nucleotides at the 5’ end of sgRNA reduces the formation of the R-loop and cleavage activity of the RuvC domain. Also, the incorporation of a 20 nucleotide RNA hairpin at the 5’ end of sgRNA showed stable R-loop formation with gene-editing activity inactivation ¹⁹. Modification of the 5′-end nucleotide of sgRNA to 5-bromo-2′-deoxyuridine (BrdU) in CRISPR-associated Marinitoga piezophila Argonaute (MpAgo) showed higher specificity and affinity to its complementary nucleic acids with a 300-fold enhanced ability to discriminate on-target sequences from off-target sequences that have only one nucleotide difference ⁶⁶. An in vitro study indicated that exposure of human blood cells to sgRNA with phosphate removal at the 5’ end showed suppression of an interferon (IFN) response ⁶⁷. In zebrafish embryos, it has been reported that treatment of sgRNAs with a G-quadruplex motif at the 3′ end significantly improved the stability and efficiency of site-specific cleavage of sgRNAs. Such sgRNA with its modified 3’ end did not show any non-specific cleavage in zebrafish ⁶⁸. In practice, sgRNAs are generated via performing their chemical syntheses, IVT and intracellular transcription in the host cells their advantages, such as affording high purity, relative ease of sgRNAease of sgRNA modification, and high yield ^{69, 70}. However, the limited length of RNAs remains a barrier to their chemical syntheses. Chemical ligation of two or more RNA fragments using click chemistry could be a potential approach towards overcoming the relevant limitations entailed with their chemical syntheses ⁷¹.

Various modified sgRNAs that included 2′-O-methyl, 3′-phosphorothioate, 2′-O-methyl-3′phosphorothioate (MS), 2′-O-methyl 3′ thiophosphonoacetate, 4′-O-methyl, 2′-F, locked and unlocked nucleic acids (LNAs and UNAs, respectively), in 3′ and 5′ terminals of sgRNA, increased the editing efficiencies, stability, and on-target sgRNAs performances in human cells. ^72–74. In contrast to the limitations in synthetically preparing sgRNAs, the IVT approach that leads to their development is rapid, inexpensive, easy to control, and capable of generating sgRNAs in high yield ^69,75. However, sgRNAs might demonstrate poor stability, high immunogenicity, and variable nucleotide sequence lengths. These could result in undesirable secondary or tertiary structures, and thus yield broad heterogeneous activities ⁶². Several chemical modifications of RNA nucleotides that decrease the immunogenicity and increase the stability have been studied ²⁷. Among various Cas9 nuclease delivery methods, mRNA-based expression is optimal. First, Cas9 mRNA does not need to bypass the nuclear envelope, which in turn, allows for the use of a reduced dosage for achieving efficient transfection. Second, the delivery method does not carry with itself the risk of being integrated into the genome. Third, it can be produced in a cost-effective manner and chemically modified on a larger scale. Most importantly, it is associated with producing limited off-target effects due to its transient expression ⁷⁶. Owing to the success of COVID-19 mRNA vaccines as well as the existing approaches to gene therapies, current efforts disproportionately favor the use of Cas9 mRNA along with sgRNA oligonucleotides as a cargo for the delivery of CRISPR components against multiple pathogens ^45,77,78. Despite the advances in CRISPR-Cas9-based gene editing, concerns exist, some of which are discussed in the next section.

Limitations of CRISPR-Cas9

Progress for CRISPR-Cas9 was realized in genetic, cancerous, and infectious disease treatments ⁷⁹. Though the approach has seen favorable results in pre-clinical research studies the promise of gene editing in clinical trials has yet to be achieved. With regards to clinical trial usage, concerns related to the specificity, efficiency, and in vivo delivery of CRISPR-based gene-editing therapies must first be resolved for their safe and efficient clinical translation.

Specificity

Nucleotide targeting specificity has been a significant concern for each of the ZFNs, TALENs, and CRISPR-Cas9-based systems. CRISPR-Cas9 enzymes leading to the dsDNA breaks at off-target locations (in addition to the targeted sites) can result in undesirable off-target activity ³⁷. The Cas9 enzyme sometimes cleaves dsDNA that does not exactly complement the designed sgRNA, thereby leading to mutations at non-targeted sites. Long-term expression of Cas9 within the targeted cells, editing temperature, and unsuccessful sgRNA design are some of the factors that can significantly impact off-target activity ²¹. Off-target prediction assays including in silico, biochemical, and cell-based systems were developed to model these unwanted mutations ⁸⁰. Some of the examples of off-target prediction assays include GUIDE-seq (genome-wide unbiased identification of DSBs enabled by sequencing) ⁸¹, LAM-HTGTS (linear amplification-mediated high-throughput genome-wide translocation sequencing) ¹¹ and CIRCLE-seq (circularization for in vitro reporting of cleavage effects by sequencing) ⁸². In parallel, studies are being conducted to address the challenge of producing off-target effects caused by wild type SpCas9.

Highly specific Cas9 variants of wild-type SpCas9 have already been identified ²¹. A variant protein called cas9 nickase (nSpCas9) was generated by mutating either the HNH (D10A)- or RuvC (H48A)-like domain, resulting in conferring it the ability to cleave single-stranded DNA (ssDNA), wherein it has been demonstrated that nicks in the genomes are repaired with high fidelity as compared to DSBs ⁸³. Likewise, introducing mutations in both HNH (D10A)- and RuvC (H48A)-like domains results in the generation of a mutant protein called dead cas9 (dCas9), which binds with the sgRNA and targets DNA with high specificity. dCas9 lacks DNA cleavage activity, and thus can essentially deliver any functional domain to a specific locus in the genome. Such delivery can allow for site-specific transcriptional activation (CRISPRa) or transcriptional inhibition (CRISPRi) of the desired gene ⁸⁴. To improve the efficacy of CRISPRi, dCas9 was fused with a repressive chromatin modifier domain, kruppel-associated box (KRAB) from kox1 protein, recruiting chromatin-modifying complexes to silence the transcription of genes targeted by sgRNA ⁸⁵. Moreover, SpCas9 protein can be easily programmed for multiplexing, which means multiple genomic loci can be targeted by several sgRNAs simultaneously to further increase specificity ⁶⁵.

Efficiency

dsDNA cleavage by CRISPR-Cas9 results in the activation of repair pathways in the edited cells, which in turn, leads to gene editing. The two most common pathways are NHEJ and HDR, wherein the latter results in gene editing with higher accuracy. However, the editing efficiency of the HDR repair pathway is low ⁶². The strategies to improve the HDR repair efficiency by inhibiting the NHEJ pathway was described. This includes regulating HDR-related factors and by synchronizing the cell cycle ²¹. To improve the editing efficiency base-editing tools were developed that allow the irreversible conversion of one targeted base to another without introducing dsDNA cleavage ⁷⁹. Two classes of Base Editors (BEs) have been developed. This includes adenine base editors (ABEs) that convert A·T to G·C ⁸⁶ and cytosine base editors (CBEs). Each result in the transition of C·G to T·A ⁸⁷. BEs are comprised of the Cas9 nickase or dCas9 fused to a deaminase and a guide RNA to direct the Cas9 protein to the target base within the specified window. The use of base editors has resulted in 70 and 100% editing efficiencies in mouse neurons and cultured mammalian cells, respectively ²¹. Base-editing techniques have been widely adapted in prokaryotes, mice, monkeys, and human embryos ²¹. Further advances were made by establishing the prime editing (PE) technique. PE mediates all the possible base-to-base conversions as well as targeted deletions and insertions with increased accuracy ⁸⁸. The prime editing complex consists of prime editing guide RNA (pegRNA) that identifies the target sequence by sgRNA and contains the information to replace the targeted sequence in the reverse transcriptase template sequence, wherein it also has a primer binding site (PBS) that is required to start reverse transcription at the 3’ end of the nicked DNA; and a Cas9 nickase fused to reverse transcriptase ⁸⁹. PEs do not induce detectable off-target effects efficient delivery was optimized.

Overview for Gene Delivery

Delivery of gene-editing tools is a daunting task. Recently, technologies were used in delivering CRISPR-Cas9 components uses non-viral and viral systems. The methods include transfection by lipids, cationic polymers, or lipid nanoparticles (LNPs), electroporation (nucleofection), and transduction. For cell delivery, physical methods such as lipofectamine-mediated transfections and electroporation of plasmid DNAs encoding SgRNAs and Cas9 proteins have been widely used ^19,57. However, for in vivo applications, electroporation-based techniques for their delivery are not suitable. Instead, viral-based vectors, including the adeno-associated virus (AAV) ¹⁴ and lentivirus, have been used in the past few years. These viral vectors have limited future opportunities for vast applications, due to their immunogenicity, limited payload size, replication, chances of active mutation and carcinogenesis, which are significant safety concerns and fully rollout for genetic medicine. Cas9 orthologues like Staphylococcus auricularis Cas9 (SauriCas9; 1061aa), Staphylococcus aureus Cas9 (SaCas9; 1053 aa) instead of SpCas9 (1350 aa) have been used ⁷⁹ to overcome the drawback of the low packaging capacity of AAV. Nevertheless, the in vivo delivery of CRISPR-Cas systems is concerning. In consideration of the barriers to effective and efficient delivery, we devote considerable attention and offer insight into the advantages of delivery methods for successful CRISPR-Cas9-based gene-editing.

Delivery Systems

Viral Vectors

Significant technological advancements have been made in the field of gene-editing with the use of viral-based vectors as delivery vehicles. An early mention of using viruses as vectors was in the early 1980s when the inherent capacity of viruses for transducing tissues and cells with foreign nucleic acid was made for gene delivery ⁹¹. Throughout the years, viral vectors were developed based on their projected efficacy for delivery and expression ⁹². Lentiviruses, adenoviruses, and adeno-associated viruses (AAVs) (Figure 5) are common vector systems. They are comprised of a protein capsid used for cargo encapsulation, the transgene of interest, and the promoter regions ⁹³. Their delivery abilities are based on systemic administration. AAV vectors are used for in vivo applications; whereas lentivirus vectors are used for ex vivo purposes ⁹⁴. Viral vector systems are widely employed due to their high efficacy (John Wiley and Sons LTD, 2021) (Figure 5). However, deleterious effects can result due to immune responses. Pre-existing immunity, limited cargo capacity and potential for insertional mutagenesis reduce efficacy. Alternative non-viral delivery may improve safety, efficiency, and scalability ⁹⁵. Moreover, viral-based vectors require high viral particle quantities for efficient gene editing ⁹⁶.

Figure 5.

Viral vectors used for gene delivery. Comparison of the three most-widely used viral vectors: adenovirus (AVs), adeno-associated virus (AAVs) and lentivirus. Total number of currently active clinical trials that employ viral vectors and its schematic representation (figure concept adopted from John Wiley and Sons LTD. (2021) (The figure was created internally by the authors).

Lentiviruses

Lentiviruses are a genus in the Retroviridae family that are approximately 100 nm in diameter. They have enveloped spheres containing two sense-strand RNAs that are encapsulated by nucleocapsid proteins. Reverse transcriptase is present in these viruses along with integrase and protease proteins. Lentiviruses fall into the category of complex retroviruses. HIV-1, a prominent lentivirus, possesses a 9.7 kb genome flanked by 5’- and 3’-Long Terminal Repeat (LTR) regions that are crucial for the replication cycle. The three essential core protein genes have the following functions: (i) the gag gene is ultimately responsible for originating the matrix proteins and the capsid; (ii) the pol gene encodes for reverse transcriptase and integrase; and (iii) the env gene encodes for glycoproteins associated with cellular tropism of the virus ⁹⁷. Additionally, the tat and rev genes, which are present in lentivirus, are associated with transcription, elongation, and nuclear export of viral RNA ^98,99. Moreover, auxiliary genes such as vif, vpu, vpr and nef, in lentiviruses increase pathogenesis and viral titer ¹⁰⁰. The most prominent advantage of using lentiviruses in gene-editing applications is their ability to integrate into the host genome and exhibit extended transgene expression. Lentivirus can form replication-competent retroviral vectors (RCR), which pose a risk of large deletion or insertional mutagenesis as dysregulation in viral replication will occur due to RCR. The FDA mandates thorough testing for the detection of RCRs in lentivirus products, a process that is both very expensive and time-consuming. PCRs need to be performed on patients’ specimens from time to time for those participating in clinical trials, which increase burdens on both the patients and on the trial ¹⁰¹. Chimeric Antigen Receptor-T (CAR-T) cells developed from lentivirus for acute lymphoblastic leukemia give rise to B-cell aplasia due to sustained CAR T-cell effect. It can result in B-cell aplasia for more than 5 years following treatment due to the long half-life of T-cells. ¹⁰² Late onset of malignancies is also a risk factor for patients receiving lentivirus vector treatments, as proto-oncogenes might be activated near the integration site ¹⁰³. Moreover, the sites of integration are difficult to predict, which adds to the challenges of utilizing lentiviruses as a therapeutic delivery vehicle (Figure 5). Although multiplexable methods, such as CRISPR - enhanced viral integration site sequencing (CReVIS-seq), for mapping the viral-integration sites have been developed, they are limited to needing specific sections of virus-host boundary sequences to work effectively ¹⁰⁴. A single lentivirus vector was demonstrated to express a number of genes ¹⁰⁵, which may not be safe for patients long term. As mentioned earlier, the provirus may also integrate into the host genome, thereby causing a cascade of pathogenic reactions. Moreover, there may also be mobilization of structural viral genes in the target cells. Lentiviral vectors while being developed post additional questions on whether they will be safe enough as an effective in vivo therapy ¹⁰¹. They are modified from HIV-1 and while several viral genes are removed concerns remain for recombination. The lentiviral vector contains the viral LTRs and the packaging signal. While the lentiviral packaging genes are contained in separate plasmids and the emerging pseudo lentiviral particles remain replication deficient whether their longer-term use could have secondary untoward outcomes remains undefined. On balance and to date lentiviral vectors are used to deliver genes into cells as to date they are both a dependable and safe method as a gene delivery system.

Adenoviruses

Adenoviruses (Family: Adenoviridae) are DNA viruses without an envelope. They contain an icosahedral protein capsid that can accommodate a 26- to 45-kb double-stranded, linear DNA genome. About 51 different serotypes of adenoviruses are present in humans, from which Ad5 and Ad2 are most naturally occurring. Nearly 12 different proteins make up the virion, including core proteins, terminal proteins, major and minor capsid proteins, maturation protease and packaging machinery. The icosahedral structure is formed by the major coat protein hexon and the capsomers are formed by penton base and fibers. Three minor coat proteins conserved within the family are proteins I, IIIa, V, and VIII, with each playing a pivotal role in the viral life cycle. The Ad genome is packaged within a icosahedral capsid having a diameter of 95 nm measuring from vertex to vertex ¹⁰⁶. Ad genome is flanked by two short, inverted terminal repeats that act as the origin of replication ¹⁰⁷. Inverted Terminal Repeat (ITR) forms a hairpin-like structure that serves as a self-primer to enhance primase-independent DNA synthesis along with easing its integration into the non-host genome. Proper viral packaging is carried out by a genetic component following the ITR in the Ad genome ⁶⁵. The Ad contains core proteins V, VII, and µ, which are highly basic and play a role in neutralizing the negative charges to allow for the dsDNA to condense within the viral core ¹⁰⁸. AVs while efficient for gene delivery, they do not uniformly target disease affected tissues. Their toxic effects do limit the vector amount that can be administered to patients. Limitations in viral titers and packaging efficiencies also limit their general utility for treating more chronic disease. Nonetheless and on balance, AVs are effective as vectors because of their epichromosomal persistence in host cells, high efficiency of transduction in dormant and dividing cells, and its broad tropism ⁶⁵.

There are three generations of viral vectors: The first-generation vectors are devoid of the E1 and E3 regions and can package a payload maximum of approximately 8.2 kb of cargo carrier gene(s). The second-generation vectors, along with lacking E1 and E3 regions, also lack the E2 and E4 regions. The third-generation vectors are known as ‘gutless’ vectors that contain only the ITRs and the packaging signal. Also known as ‘high capacity’ adenoviral vectors (HCAds) ¹⁰⁹, these third-generation vectors have packaging capacity of approximately 36 kb of genetic material. Therapeutic Ad vectors have been used in vivo since the early 1990s with the first human trials in patients afflicted with peripheral vascular disease and lung tissues of patients with cystic fibrosis ^38,110. However, numerous pathologies have been reported in pre-clinical and clinical trials. In 2002, a patient fatally suffered from multisystem organ failure upon involvement in a clinical trial for Ornithine transcarbamylase deficiency (OTCD), with other participants experiencing an acute rise in hepatic transaminase, hypophosphatemia, and thrombocytopenia ¹¹¹. Delivery of SpCas9 using Ad vector in targeting gene Pten (which is associated with nonalcoholic steatohepatitis (NASH)), was shown to cause liver damage (hepatomegaly and steatohepatitis) and cellular and humoral immunity. ¹¹². Li et al. reported 13% editing was achieved in vivo in red blood cells (RBCs) of the edited hematopoietic/stem progenitor cells that were performed via using the CRISPR-Cas9 in B-thalassemia mice ¹¹³. Intravenous administration of Ad5 for Cas9 system targeting mutated human SERPINA gene [encodes for alpha-1-antitrypsin (AAT)], showed the possibility of the use of gene editing in hepatocytes. ¹¹⁴ Efficient tumor regression was reported by “knocking out” mutated oncogenes such as KRAS or EGFR by CRISPR-Cas system ^115,116. The heterogeneity of different cancers, along with their tendency to metastasize or diffusely spread, may result in the therapy being less effective than if such was otherwise. A challenge in Ad vector delivery is that the common human Ad serotypes have pre-existing immunity in the form of neutralizing antibodies (NAbs) against them. Seroprevalence of 35% and 90% against HAd5 was found in the United States and Cote d’Ivoire (West Africa), respectively ¹¹⁷. Also, adenoviruses have high clearance rates post-systemic administration due to their interaction with myeloid cells.

Adeno-Associated Viruses (AAVs)

AAVs was a serendipitous discovery in 1965, wherein such was first deemed as an impurity in adenovirus preparations ¹¹⁸. AAV is classified as a dependoparvovirus because it requires a ‘helper virus’ in assisting the completion of its life cycle. AAVs are non-enveloped and are smaller than Ad. They contain a single-stranded genome of 4.7 kb that is flanked by 2 ITRs. These ITRs are 145 nucleotides long and form a hairpin structure that can self-prime ¹¹⁹. The three major genes encoding for a functioning AAV are the Cap gene that encodes for the Capsid consisting of 60 outer-coat proteins responsible for protecting the genomic DNA and cell binding; the Aap gene which encodes for an assembly-activating protein that acts as a scaffold for capsid assembly; and the Rep genes that are responsible for viral packaging and replication. The Rep genes Rep 52, Rep 40, Rep 68, and Rep 78 are present in the first Open Reading Frame (ORF). However, they lack the required genes needed to replicate their genome ¹²⁰. The capsid of AAV is a T=1 icosahedral, 60-mer structure containing VP1, VP2 and VP3 in a 1:1:10 ratio, respectively. The capsid sequence dictates the varying transduction efficiencies in different tissues ¹²¹. Since the advent of AAVs for usage in humans [which afforded the in the delivery of the cystic fibrosis transmembrane regulator (CFTR) gene packaged with the recombinant AAV2 capsid (rAAV2-CFTR)], AAVs have been widely used to deliver a gene of interest in vivo ¹²². Because of the recent advances in discerning host-virus interactions, recombinant AAV (rAAV) is becoming the leading platform for delivering genes of interest for treating different human diseases ⁶². AAVs are comprised of an icosahedral protein capsid of ~26 nm in diameter having a single-stranded DNA genome of ~4.7 kb in length that can be either the plus (sense) or minus (anti-sense) strand ¹²³. The small size of the AAV genome limits insert size ⁹³. This aspect especially poses a unique challenge to deliver the CRISPR-based cargos. A dual-vector system using ITR-mediated recombination was exploited to overcome this challenge ¹²⁴. However, the genome-editing efficacy of this dual-vector approach is potentially limited in its extent of overall success, as accurate editing can be achieved if and only if both vectors are taken up by and expressed in the given cells in a manner that affords them a period of their contemporaneous presence. One other important positive aspect of using an AAV-based CRISPR-Cas9 delivery system is to limit unwanted outcomes. This is affirmed because of its long-term expression of its genome editing abilities. Many approaches towards achieving transient expression of the genome editing system have been reported, including a drug-inducible promoter for gene expression, protein activation by light, drug-inducible assembly, reconstitution from a split construct, and protein inactivation by a drug-inducible destabilizing domain ^22,125,126. To date, their clinical efficacy has yet to be demonstrated. Immunogenicity is a major issue to contend with in using AAV viral vector-based delivery systems, especially if requiring the use of a high dose ¹²⁷ (Figure 5). AAV packaging capacity is lower than that of other viral vectors, thereby at times preventing full-length BEs from being packaged into the vector ¹²⁸. Maurya et al. wrote that AAV-mediated therapy is “relatively safe” at a low threshold of vector dose ¹²⁹ with efficacy being mostly dependent on the dose administered ¹³⁰. This could imply the potential endpoint limits for using AAVs in vivo at “safe” doses. With that stated, it is necessary to emphasize that the efficacy of AAVs cannot be interpreted from in vitro studies because the transduction of AAVs in vivo does not positively correlate with that obtained when applied in vitro. In short, even though AAVs have undergone extensive testing in pre-clinical studies, their translation for human use is still lacking ¹²⁰.

Virus-like Particles (VLPs)

Virus-like particles are a class of delivery vehicles that resemble the structure of true virus particles, wherein the immune system rapidly recognizes them ¹³¹. VLPs address certain drawbacks of viral-based delivery systems: (i) the low likelihood of integration of viral vectors into the genome of transduced cells; and (ii) an increase in off-target editing frequency due to extended expression in transduced cells ¹³². VLPs can have enveloped or non-enveloped structure with one to four proteins in its outer layer. VLPs are self-assembling structures that contain the wild-type genome. In a study, the therapeutic cargo carrier such as base editing machinery, was linked with the gag polyprotein of Friend Murine Leukemia Virus (FMLV) to make FMLV based VLPs ¹³².

VLPs and their Types for Gene Therapy

There have been numerous attempts to develop effective VLP-based gene editing systems. Choi et al. demonstrated that pre-packaging the Cas9 protein in lentiviral particles reduced off-target editing because the exposure and expression was transient ¹³³. The first report of in vivo gene editing using VLPs was by engineered VLPs, “eVLPs”. Nanoblades, which are Cas9-sgRNA ribonucleoproteins produce efficient genome editing in in vitro systems such as bone marrow-derived macrophages and primary human fibroblasts. Fusion of Cas9 from Streptococcus pyogenes to Gag from MLV allowed the incorporation of the Cas9 endonuclease into VLPs internal structure. Nanoblades were more effective in inducing DSBs in vitro compared to lipofection or electroporation. They can also be made into complexes with templates of DNA repair to facilitate homologous recombination-based “knock in” in cells without transfection. The subsequent design of eVLPs with the fourth generation eVLPs afforded a 16-fold greater packaging capacity than their first-generation equivalents. Off-target editing was very low, and they found that it was efficacious in a myriad of cell types. This also yielded effective base editing in brain, liver and other organs of mice. Base editing mediated by Pcsk9 had greater than 60% efficiency following a single intravenous dose. eVLPs have been reported to be of some utility in a myriad of both ex vivo and in vivo applications. It is worth noting that the modular structure of eVLPs provides countless number of opportunities for delivering proteins or therapeutic agents of interest, but a priori knowledge of the applicable cargo carrier size that can efficiently diffuse out from these capsids does beget additional consideration ¹³⁴.

Nonviral Vector

Nonviral delivery vehicles are emerging as promising alternatives for gene delivery due to their improved safety profile, scalability, and versatility. These vectors offer advantages such as lower immunogenicity and reduced risk of insertional mutagenesis compared to viral vectors ¹³⁵. Non-viral vectors can be categorized into different types. These include, but are not limited to, lipid- and polymer-based vectors. Lipid-based vectors that include liposomes or lipid nanoparticles (LNPs) can encapsulate and protect the therapeutic genetic material. They offer efficient cellular uptake and endosomal escape that enable successful gene delivery. Lipid-based vectors can also be optimized to enhance stability, transfection efficiency, and targeting (Figure 6). Alternatively, polymer-based vectors utilize synthetic polymers, such as polyethyleneimine (PEI) or poly(lactic-co-glycolic acid) (PLGA), to form complexes with DNA or RNA. The complexes protect the genetic material from degradation and facilitate cellular internalization. Lipid and polymer-based vectors can be modified to controlled release and create specific delivery by surface modifications. Non-viral vectors hold potential for clinical translational research. Ongoing advancements in vector design, formulation optimization, and delivery serve to enhance efficiency, safety, and therapeutic potentials ^{136 137}.

Figure 6.

Graphic description of currently developed nanocarriers for in vivo gene editing. A description of the components of this description are now provided. A micelle is defined as a self-assembled, colloidal structure formed by the aggregation of amphiphilic molecules in a liquid medium. Amphiphilic molecules possess both hydrophilic and hydrophobic regions within the same molecule. In an aqueous environment, these molecules arrange themselves spontaneously to minimize exposure of their hydrophobic regions to water. The hydrophobic tails of the amphiphilic molecules come together in the center, shielded from water, while the hydrophilic heads face outward, interacting with the surrounding aqueous medium. Micelles have a core-shell structure, with the hydrophobic core and the hydrophilic shell. This unique arrangement allows them to solubilize and transport hydrophobic materials in a biological aqueous environment. They are a spherical vesicle composed of lipid bilayers analogous to cell membranes with an aqueous core. Liposomes are artificially prepared consisting of phospholipids, the primary building blocks of cell membranes. The phospholipids have hydrophilic heads and hydrophobic tails which enable them to form a bilayer structure in an aqueous environment. Cationic lipids have a positive charge due to the presence of quaternary ammonium groups in their structure. When mixed with negatively charged nucleic acids, electrostatic interactions occur, leading to the formation of lipoplexes. Lipoplexes are widely used in gene delivery and gene therapy applications. (The figure was created internally by the authors).

Targeted Gene Delivery

Nanoparticles (NPs) can be made to better protect genetic cargo carriers of in vivo elimination which occurs via normal excretion processes and metabolic processes. Based on the objective of the therapy, NPs can be modified enabling all kinds of routes of administration. Additionally, the efficacy of large-size cargo carrier materials of the NPs encases and affords its delivery to targeted cells and tissues. The chemical basis and any chemical modifications thereof of NPs can alter their biodistribution, ability to afford cellular and tissue targeting, and the mechanism affording cellular absorption such that the desired therapeutic result is achieved ¹³⁸.

Polymeric nanoparticles (PNPs)

New drug delivery systems are becoming increasingly important due to the intricacy of some diseases and the toxicity associated with some medicines ¹³⁹. PNPs can increase the bioavailability and biodistribution of therapeutic payloads ¹⁴⁰ and have proven effective for targeting brain subregions ¹⁴¹ and improve drug solubility, release, distribution, limit immunogenicity, and can deliver multiple drugs simultaneously ^141,142. Toxicities that are governed by their chemistry, size, shape, aggregation, and electromagnetic characteristics ^143–145 can be overcome by design and exposure limitations ¹⁴⁶. Delivery of nucleic acids to cells in the form of plasmid DNA (pDNA) or small interfering RNA (siRNA) can improve outcomes ¹⁴⁷. Viral-based vectors allow high rates of gene transfer but may induce inflammatory and immunological responses. To mitigate these issues, non-viral-based cargo carriers that included cationic liposomes or polymers were developed. On doing so, they form “lipoplexes” or “polyplexes” (Figure 6). These complexes show promise as an alternative strategy to viral-based vector gene therapy ¹⁴⁸. First, a micelle is a self-assembled, colloidal structure formed by the aggregation of amphiphilic molecules in a liquid medium. Amphiphilic molecules possess both hydrophilic and hydrophobic regions within the same molecule. In an aqueous environment, these molecules arrange themselves spontaneously to minimize exposure of their hydrophobic regions to water. The hydrophobic tails of the amphiphilic molecules come together in the center, shielded from water, while the hydrophilic heads face outward, interacting with the surrounding aqueous medium. Second, micelles have a core-shell structure, with the hydrophobic core and the hydrophilic shell. This unique arrangement allows them to solubilize and transport hydrophobic materials in a biological aqueous environment. They are a spherical vesicle composed of lipid bilayers analogous to cell membranes with an aqueous core. Third, liposomes are artificially prepared consisting of phospholipids, the primary building blocks of cell membranes. The phospholipids have hydrophilic heads and hydrophobic tails which enable them to form a bilayer structure in an aqueous environment. A lipoplex is a complex formed by combining lipids (usually cationic lipids) with nucleic acids, such as DNA or RNA. Fourth, cationic lipids have a positive charge due to the presence of quaternary ammonium groups in their structure. When mixed with negatively charged nucleic acids, electrostatic interactions occur, leading to the formation of lipoplexes. Lipoplexes are widely used in gene delivery and gene therapy applications.

Genetic cargo within endosomolytic polymeric carriers is protected by reversible covalent modifications, cellular receptor ligands, and labile siRNA conjugations. Each can be integrated into dynamic poly-conjugate technologies. A significant advancement in the use of RNA interference (RNAi) in therapeutic applications and gene function research would be the affording of efficient in vivo transport of siRNA to the proper target cell. Dynamic poly-conjugate technology has key features that include a membrane-responsive polymer, reversible masking of polymer activity until it reaches the acidic environment of endosomes, for example, targeting hepatocytes in vivo after simple, low-pressure IV injection ¹⁴⁹. RNAi, target degradation by RNase H-mediated cleavage, splicing manipulation, non-coding RNA inhibition, gene activation, and programmed gene editing are all examples of using oligonucleotides to influence gene expression. As a result, these molecules could be used to treat a wide range of diseases, especially as numerous oligonucleotide-based medicines have only recently received regulatory approval. Despite recent technological breakthroughs, effective oligonucleotide delivery, particularly to extra-hepatic regions, remains a key translational challenge ^150,151. Although NP delivery systems can be used to deliver genes, there are still some challenges that need to be addressed to improve their effectiveness.

mRNAs

Researchers worldwide have been investigating different technologies to develop mRNA delivery systems for more than 30 years. LNPs are novel biopharmaceutical drug delivery systems composed of different classes of lipids. LNPs as a drug delivery vehicle were first approved in 2018 for the siRNA drug. LNPs are found to be the breakthrough nonviral-based technology that can stably deliver mRNA to the target organ, especially due to protecting the mRNA from nuclease-mediated degradation ¹⁵². mRNA LNPs have recently come under the spotlight due to serving as remedial platforms following the acceptability of the SARS-CoV-2 mRNA LNP vaccines (BNT162b2 and mRNA-1273), with mRNA vaccine doses now injected into billions of individuals worldwide. These semi-effective mRNA-LNP vaccines have showcased the modularity and the rapid manufacturing capabilities of mRNA-LNP systems ¹⁵³.

LNPs are lipid-based nanostructures, composed of multiple components such as cholesterol, a phospholipid, a PEG-lipid (polyethylene glycol lipid), and an ionizable lipid. Each type of lipid has a distinct role in typical LNPs formulations and formation. For example, ionizable lipids are responsible for mRNA binding and LNP core formation. Cholesterol and phospholipids are considered helper lipids that give stability to the mRNA-lipid complex. PEGylated lipids protect the lipid shell and the LNPs from opsonization ¹⁵⁴ (Figure 7). LNPs can be classified into different subgroups based on lipid composition and the synthetic method used for their development ¹⁵⁵. For instance, liposomes, exosomes, transferosomes, nanostructured lipid carriers, liposome-like structures or LNPs, solid LNPs (SLNPs), and lipid nanoemulsions are some of the well-developed lipid-based NPs to date ¹⁵⁶. LNPs excel in efficacy over that of viral-based vectors due to their decreased immunogenicity, the use of versatile payloads (from several nucleotides to several million in length), tailorable functional groups, and ease of scale-up which enables commercial production ¹⁵⁷. The types of lipids used in mRNA-LNP formulation and methods for synthesis are discussed below.

Figure 7.

mRNA delivery by LNPs. Key components of mRNA-LNPs; administration routes of mRNA-LNP formulations; advantages of mRNA-LNPs over other delivery systems; mRNA-LNPs used in different diseases (The figure was created internally by the authors).

Lipids used in mRNA LNPs Preparation

Cationic Lipid Nanoparticles (cLNPs)

Cationic lipids have three distinct components that include a positively charged (hydrophilic) amine functional group; a hydrophobic chain; and a covalent linkage between the hydrophobic chain and the hydrophilic moiety-linker group. The positively-charged amine group (ammonium based) can readily associate with the negatively-charged nucleic acid through opposing charges interactions, which makes transfection of LNPs effective ¹⁵⁸. The hydrophobic moiety includes the length of alkyl chain and its degree of saturation. The polar hydrophilic headgroup positively charged made through the protonation of one or multiple amino groups designed with specific spatial properties. This headgroup contain amines, quaternary ammoniums, guanidiniums, amino acids, peptides, and a variety of heterocyclic compounds. ¹⁵⁹. Headgroups with one or multiple positive charges can attract negatively charged nucleic acids. This occurs through electrostatic forces that forms a complex called a “lipoplex” made up of cationic lipids and anionic nucleic acids. This structure is crucial for the cLNPs function as it protects the nucleic acids from degradation and delivers them to target cells. Prior studies have found that the acid dissociation constant (pKa) of the headgroup and the position of the charge on the lipid bilayer interface are key factors for the successful delivery of nucleic acids. ¹⁶⁰. Recently, research was performed that confirmed the acid dissociation constant (pKa) of liposomal nanoparticles and showed that they possess the ability to transport siRNA across biological barriers and ultimately achieve gene silencing ¹⁶¹. The presence of a charge on the lipid can cause toxicity and quick elimination from the bloodstream. To address this issue, anionic lipids with pKa values of around seven or lower have been developed, which have lower toxicity and can efficiently encapsulate nucleic acids for both in vitro and in vivo activities as they exist in an uncharged state ¹⁶². Cationic lipids are the optimal carrier for gene delivery, as they can be easily synthesized by modifying each of the constituent domains of the ionizable lipids ¹⁵⁹. The cationic lipids that are used as part of non-viral-based gene carriers include N-(1-(2,3-dioleyloxy) propyl)-N-,N-,N-trimethyl-ammonium (DOTMA), 1,2-dioleoyloxy-3-(trimethylammonium) propane (DOTAP), 2,3-dioleyloxy-N-(2-(sperminecarboxamido)ethyl)-N,N-dimethyl-1-propana trifluoroacetate (DOSPA), DC-cholesterol (dimethylaminoethane-carbamoyl-cholesterol) and 1,2-dimyristyloxy-propyl-3-dimethyl-hydroxyethylammoniumhydroxyethyl ammonium bromide (DMRIE). Recently, various polymeric carriers, such as chitosan, poly(β-amino esters) (PBAEs), protamine (when used only with polyethyleneimine (PEI), and poly(2-dimethylaminoethyl methacrylate) (PDMAEMA) and other polyamines have been adapted from their use in siRNA/DNA delivery to now afford mRNA delivery ¹⁶³. A linear PEI derivative, jet PEI, which is commonly used for DNA and siRNA transfection, can also be used for mRNA transfection, even though it shows lower efficiency than lipoplexes ¹⁶⁴. LNPs have been reported to reduce the susceptibility of nucleic acids to degradation by nucleases. Modifying the ratio of cationic lipid to nucleic acids, by having more positive charge, improves nucleic acid delivery by making it easier to bind to negatively charged cell surfaces and break down endosomal membranes. Despite their relatively low transfection efficiency when compared to that of using viral-based vectors, LNPs can be used in a wide range of applications due to the rapid and facile preparation of lipoplexes whilst displaying relatively low toxicity ¹⁶⁵. However, cationic LNPs systems tend to degrade when stored for a long time in aqueous solutions, particularly due to their tendency to undergo hydrolysis. ¹⁶⁶. The physical and chemical properties of cationic lipids greatly hinder their cellular uptake and effectiveness in gene therapy. One study observed that the lipoplex size is the primary factor in determining the efficiency of lipofection in vitro ¹⁶⁷. Furthermore, studies have indicated that larger liposomes are rapidly cleared from the bloodstream by the reticuloendothelial system (RES), while smaller liposomes exhibit prolonged circulation. Moreover, liposomes with a positive charge tend to exhibit relatively toxicity levels compared to neutral or negatively charged liposomes. ¹⁶⁸. Aside from their short circulation half-life, unacceptable toxicity, and high net-positive charge, their tendency to aggregate with erythrocytes is the primary limitation to the clinical translation of cLNPs ¹⁶⁹. The concept of using ionizable cationic lipids can be a solution to (i) the low-transfection efficiency of LNPs comprised of uncharged lipids and (ii) their biocompatibility issues.

Ionizable Cationic LNPs (iLNPs)

The quaternary ammonium head group of cationic lipids was substituted with a titratable moiety to make ionizable lipids. These ionizable lipids have an electrostatic charge on the head group that is pH-dependent on the head group pK_a of the lipid. The ionizable containing LNPs are among the most advanced non-viral-based delivery platforms for efficient gene therapy, as they can self-assemble into NPs when mixed with polyanionic nucleic acids. Also, these ionizable lipids, with modifiable pK_a values below pH 7 (at a pH less than 6) are positively charged, thereby allowing high encapsulation efficiency (EE%) of negatively charged nucleic acids such as mRNA, siRNA. When injected, the LNPs have a neutral charge in physiological environments with a pH above their pKa, allowing them to evade uptake by the RES, increase circulation and decrease toxicity associated with positive charge ¹⁶². However, once LNPs are taken into the endosome, where the pH is lower than the pKa of ionizable lipid, the amine group of the ionizable lipid loses a proton and binds to the negatively charged endosome lipid head groups. This interaction facilitates endosomal escape ¹⁷⁰. Liu et al. described the safe use of aminoionized LNPs for Cas9 mRNA delivery. In their study, they found approximately 70% of gene editing that caused tumor cell apoptosis. They also observed about 50% of tumor growth inhibition and improved the survival rate by 30% with a single intracerebral injection of CRISPR-LNPs against PLK1. However, when these ionized LNPs were engineered with an antibody targeted delivery, they afforded selective uptake in ovarian tumors, enabled up to 80% gene editing; and increased the survival rate to 80%. The ability to disrupt gene expression opens new avenues for cancer research and treatment ⁶².

iLNPs have been widely used for the systemic RNA therapeutic delivery. Previously, various ionizable lipid materials were reported for LNP formulations such as C12–200 (1-,1’-((2-(4-(2-((2-(Bis(2-hydroxydodecyl)amino)ethyl)(2-hydroxydodecyl) amino)ethyl) piperazin-1-yl)ethyl) azanediyl)bis(dodecan-2-ol), UK-E12 (proprietary), cKK-E12 (3,6-bis(4-(bis(2-hydroxy-dodecyl) amino)butyl)piperazine-2,5-dione) and DLin-MC3-DMA (4-(dimethyl-amino)-butanoic acid, (10Z,13Z)-1-(9Z,12Z)-9,12-octadecadien-1-yl-10,13-nonadecadien-1-yl ester). These LNPs showed efficient gene silencing in the liver at the dose level of 0.002 mg of siRNA per kilogram (mg/kg). The mechanism of cellular uptake was well-investigated, and adsorption of serum ApoE (apolipoprotein E) on the surface of LNPs was a major factor facilitating the intracellular delivery of LNPs into hepatocytes through low-density lipoprotein (LDL) receptors. In recent years, the medicinal chemistry-guided structure-activity relationship (SAR) approach has led to the development of various novel ionizable cationic lipids. In conjunction with sophisticated formulation procedures such as the involvement of microfluidic approaches (which can carefully tune the size of the resulting LNPs), these novel lipids have swiftly increased the likelihood of successful therapeutic interventions. Some recent advances are also seen in the “next generation” lipid and lipid-like molecule developments to make the mRNA-LNP delivery even more efficient. These lipoids are derivatives of ionizable lipids that are comprised of tertiary amines, but are modified by the addition of C12–200 and cKK-E12. These modified lipids demonstrate a gene-silencing efficacy that is akin to that of DLin-MC3-DMA. ^171,172. Hence, it appears that the use of ionizable lipids in LNPs formulations could be an outside-of-the-box approach for CRISPR-Cas9 delivery that affords efficacious gene silencing or expression of target genes towards or against a variety of infectious diseases.

Helper Phospholipids

Along with cationic lipids, a standard mRNA-LNP (lipoplex) also contains helper lipids that are added to the lipid composition to enhance the stability of LNPs. In the early stages of cLNP development (for their use in gene delivery), DOPE (1,2-dioleoyl-sn-glycerol-3-phosphoethanolamine) was used as a helper lipid ¹⁷³. DOPE has a small head group, a phosphoethanolamine, and two unsaturated bulky oleoyl chains. This typical structure of DOPE gives it a pronounced cone shape and helps its assembly in the non-bilayer hexagonal phase ¹⁵⁸, which enhances the fusion of LNPs with cell membranes. This peculiar property is the reason why DOPE is considered a “fusogenic lipid” ¹⁷⁴. However, some studies have indicated that LNPs based on DOPE exhibit reduced colloidal stability. This is attributed to their excessive fusogenic properties, which could potentially result in cytotoxicity. ⁴⁹. For this reason, currently, phosphatidylcholines (PCs) have replaced DOPE as a helper lipid in mRNA-LNP formulations. PCs are a natural component of the cell membrane. Particularly, more saturated PCs such as DSPC(distearoylphosphatidylcholine) and HSPC(hydrogenated soybean PC), which have a gel-to-liquid crystalline phase transition temperature, allow LNPs to demonstrate high stability in the formulation ¹⁷⁵. DSPC and CHOL lipid combinations were used to make stable LNPs for the nucleic acid delivery ¹⁷⁶. Replacing DOPE with DSPC having 40% ionizable lipid served to increase in vitro gene-silencing efficacy ¹⁷⁷. Other studies demonstrated that a higher molar ratio of zwitterionic phospholipids in an LNP composition can improve its delivery ^77,178.

Cholesterol

Cholesterol is a natural cellular component. It is regarded as a membrane “stiffener”, which induces an intermediate state by increasing the fluidity of the lipid hydrocarbons below and decreasing the fluidity above the Tc ¹⁷⁹. Once LNPs are internalized into cells, cholesterol converts the lamellar phase of LNPs into a hexagonal phase by lowering the transition temperature. Therefore, cholesterol helps mRNA release into the cytosol from the internalized LNPs ¹⁸⁰. Cholesterol also acts as lipid bilayer stabilizer in both liposomes and LNPs by filling in gaps between phospholipids ¹⁸¹. Cholesterol helps to stabilize LNPs from serum protein. It also promotes membrane fusion in vivo, and thus enhances intracellular delivery ¹⁷⁶. Increased molar fraction of cholesterol in DC-Chol-DOPE (3β-[N-(N’,N’-dimethylaminoethane) carbamoyl] cholesterol and 1,2-dioleoyl-sn-glycero-3-phosphoethanolamine) DNA lipoplexes demonstrates a higher DNA transfection efficiency ¹⁸². Results from a different study indicated that the circulation half-life and the pharmacokinetics (PK) of LNPs improved when 30% cholesterol is mixed with pure DSPC. The circulation half-life increased from a few seconds to five hours. However, adding a higher mole fraction of cholesterol did not further enhance the circulation time ¹⁸³. Some recent studies are investigating natural cholesterol analogues such as betasitosterol’s, fucosterols, campesterol, and stigmasterol as alternatives to cholesterol. Incorporation of analogues instead of cholesterol appears to improve mRNA transfection efficiency ¹⁵⁴

PEG Lipids

The careful construction of an LNP formulation having a suitable combination of excipients is the most critical aspect in optimizing pharmacokinetic and pharmacodynamic (PD) parameters. However, in doing so, it is important to curtail potential adverse effects from activation of the immune response. That is, an LNP should demonstrate negligible activation of immune cells during in vivo administration. PEG has been widely used as a coating that offers stealth-like properties ¹⁸⁴. The recent development of PEG-lipid was a recognizable milestone for the safe efficacious clinical use of LNPs. The shielding effect by PEG‐lipids protects the LNP surface against opsonins; protects LNPs against uptake by the mononuclear phagocytic system; and prevents their aggregation when in circulation ¹⁸⁵. This multifunctional role of shielding lipids is also beneficial towards its production and storage, as it prevents aggregation and affords LNPs the ability to maintain their nano-sized hydrodynamic diameter. The addition of a PEGylated lipid can improve LNPs colloidal stability in vitro and circulation time in vivo. To improve the bioavailability and circulation half-life of these LNPs, PEGylation helps to shield nanoparticles from opsonization and recognition by the immune system, thereby leading to their prolonged systemic circulation. The shielding effect of PEGylation on LNPs depends on the LNPs’ core structure, lipid composition to which PEG is anchored, nanoparticle elasticity, and PEG molecular properties ^{186 176 187}. The “PEG dilemma” (i.e., the sway between its benefits mentioned above and its drawbacks from the decreased uptake that impede endosomal release) can be exploited by leveraging this property to afford its intentional gradual release into the blood circulation. Immunogenicity and toxicity of dynamic-poly-conjugates are usually unanticipated as is the extent of efficient functional delivery. Patients who have ingested PEG-containing items, but not PEGylated drugs, can develop anti-PEG antibodies. As a result, administering PEGylated medications to those person who have developed anti-PEG antibodies can lead to rapid clearance, decreased therapeutic efficacy, hypersensitivity, and even life-threatening reactions ¹⁵⁰. pH-Sensitive anionic helper lipids can trigger low pH-induced changes in LNPs surface charge and stabilization, both of which can facilitate the endosomal release of genetic materials ¹⁷⁶.

mRNA-LNP Preparation Methodologies

Thin-Film Hydration

Various preparations were implemented to encapsulate genetic material in LNPs. There are several ways to encapsulate electrostatic interactions between the genetic material and cationic lipids. Conventional LNPs methods are based on a lipid film preparation along with subsequent hydration with an aqueous buffer that contains the genetic material, which allows for passive encapsulation of the genetic payload. Thin-film hydration is the most common and simplest method for mRNA-LNP production. These hydrated vesicles with heterogeneous particles (with diameters > 100 nm) with multilamellar vesicles was formed then require undergoing an additional down-sizing procedure, such as their extrusion or sonication. However, this method is difficult to scale-up and lacks extensive reproducibility. In early 2000, ethanol injection and microfluidic techniques were introduced to overcome these large-scale LNP manufacturing challenges ^188–191. Initially, lipids are dissolved in an organic solvent (ethanol and chloroform) and subsequently transferred to a round bottom flask. Next, the solvent is removed using negative vacuum pressure, thereby forming a thin lipid film on the inner surface of the round bottom flask. Next, the mRNA is dissolved in an appropriate aqueous buffer. Hydration of the lipid film with this mRNA-containing buffer results in producing mRNA-loaded LNPs. As electrostatic interaction plays a major encapsulation role. pH adjustment (to pH 4) creates an environment that influences the lipid charge allowing for the interaction with anionic genetic materials. Once lipid bilayer vesicles form with RNA/DNA, ethanol and unencapsulated materials are separated (by dialysis or filtration) to stabilize the LNP. As materials come encapsulated, the pH is adjusted to pH 7 to reduce the LNP surface charge so that it becomes biocompatible. The diameter of the LNPs can be selected by passing such through a polycarbonate filter with a specific diameter cut-off pore size. Three types of LNPs are formed by thin-film hydration: (i) small unilamellar vesicles (SUVs), (ii) giant unilamellar vesicles (GUVs), and (iii) multilamellar vesicles (MLVs) ^192,193 (Figure 8).

Figure 8.

mRNA-LNP production and purification methods. The figure illustrates the several different techniques used for LNP productions. These include thin-film hydration. This method involves creating a thin film of lipids by dissolving them in a solvent and then evaporating the solvent to leave a thin film of lipids on a surface. The film is then hydrated with a solution containing the molecule to be delivered, and the LNPs are formed by the self-assembly of the lipids around the molecule. Solvent evaporation is a method involves dissolving the lipids and the molecule to be delivered in a solvent, and then slowly evaporating the solvent to allow the LNPs to form. Emulsion-based synthesis is a method that involves the formation of an oil-in-water or water-in-oil emulsion, in which the lipids and the molecule to be delivered are suspended in one of the phases and the other phase is an aqueous solution. The LNPs was formed by the self-assembly of the lipids around the molecule as the solvent is evaporated. Forth, LNPs production by T junction mixture is a method of creating LNPs using a T junction mixture. This method involves creating a mixture of lipids and other ingredients, such as mRNA in a T junction reactor. Last, Microfluidics technology for LNPs production; A method involves the use of specialized microfluidic devices such as PNI to create and manipulate mRNA-LNPs in a controlled manner. There are several methods for purifying mRNA-LNPs including that includes Amicon centrifugation, gel filtration chromatography, size-exclusion chromatography, TFF and dialysis methods are the most common methods for purification of mRNA-LNP (The figure was created internally by the authors).

Solvent Injection and Emulsion Evaporation

Herein, lipids are first solubilized in water-miscible solvents (isopropyl alcohol, ethanol, and acetone). Next, the mixture is introduced into an aqueous solution with or without surfactants by utilizing an injection needle. The presence of surfactant in the aqueous phase supports the formation of lipid droplets at the injection site whilst concurrently stabilizing the LNPs by decreasing the surface tension between the solvent and water. The resultant mixture is filtered through filter paper to remove the excess amount of lipid present ¹⁹⁴. This LNP preparation method exhibits several advantages over other preparation methods including fast manufacturing without the use of expensive devices, incorporation of safe and acceptable organic solvent, and their easy handling ¹⁹⁵. An additional method that can be used in mRNA-LNPs formation is the emulsion evaporation method. In this method, initially lipids are dissolved in an organic solvent. Next, an aqueous phase (containing mRNA and surfactant) is added to generate an emulsion. Subsequently, the organic solvent is removed by evaporation under negative pressure. Evaporation of the organic solvent leads to the formation of mRNA-loaded LNPs in the aqueous phase ¹⁹⁶. This approach includes single-emulsification and double-emulsification methods. If the formulation is an “oil-in-water” or “water-in-oil system”, then it will be a single emulsion formulation. On the other hand, a double emulsion is formed when the formulation contains “oil-in-water-in-oil” or “water-in-oil-in-water” ¹⁹⁷ (Figure 8).

Newer LNPs Production Methods

The major drawbacks of the above-mentioned “conventional” methods for preparing LNPs include their lack of scalability and reproducibility. Large-scale preparation of LNPs that have narrow polydisperse particle size(s) and consistent morphology is challenging due to variations in mass-transport processes and varying production conditions. In recent years, microfluidic mixing techniques have been extensively explored for their scalability in GMP synthesis with low polydispersity index (PDI) which helps to develop homogenous LNPs formulations, that ultimately results in higher targeting efficiency and enhanced biodistribution. It is a versatile technique that has been used to encapsulate a variety of macro-biomolecules such as mRNA, plasmid DNA, and gold nanoparticles into LNPs. Furthermore, siRNA encapsulation efficiencies approaching 100% have also been achieved, which has afforded potent in vivo gene silencing ¹⁹⁸ (Figure 9). This technology is mainly based on an ethanol-dilution method. For such, the lipid components are dissolved in ethanol to make a lipid solution, with subsequently adding DNA or RNA to a suitable buffer, such as citrate, acetate, and maleic acid buffer. These two solutions are then passed through a microfluidics chip at a specific flow rate ratio (FRR) and a total flow rate (TFR). By the electrostatic interaction between the negatively charged RNA and positively charged cationic lipids, RNA-loaded LNPs are formed. The dilution of ethanol with an appropriate buffer is a mechanism that helps to form the LNPs ⁶⁵. The flow rate controls the particle size, whereby small monodisperse particles are produced with faster mixing of the aqueous and ethanol solutions. During mixing, the increase in polarity leads to rapid nucleation of nanoparticles and conversion of the intermediate disk-like planar fragments to vesicles, that prevents growth increases in particle size. This microfluidics-based approach provides advantages over other methods. For instance, lipid-based nanoparticle RNA delivery from a laboratory setting to practical applications allows for accurately tuned particle size with high reproducibility ^{237, 238}. Some of the recent advances were also made on T-Junction for production of LNPs. T-Junction mixing enables controlled and reproducible particles which involves rapid and turbulent mixing of two input streams at a T-junction controlled by adjusting the flow rate. Higher flow rates yield smaller particle sizes, while lower flow rates may result in higher PDIs. Flow rate adjustment allows for fine-tuning of particle characteristics ¹⁹⁹

Figure 9.

LNP formulation schemes using microfluidics ^200,201. All the lipids are dissolved in an organic solvent, whereas the nucleic acid (mRNA or siRNA) is dissolved in aqueous media. Both solutions pass through a chip at a specific flow rate ratio and total flow rate to achieve nucleic acid encapsulated LNPs. In the last decade mixing tools and fabrications of microfluidic devices have allowed for a considerable improvement in the manufacturing of LNPs. Various mixing and microfluidic devices are described in Table 1 (The figure was created internally by the authors).

Characterization

Gene delivery employs amphiphilic lipids, characterized by a hydrophobic tail and a hydrophilic head ²⁰⁶. Typically, an amide linkage facilitates the interaction between the hydrophobic tail and the hydrophilic head, resulting in the formation of amphiphilic lipoids. This interaction is crucial for incorporating protein components within the membrane, as these proteins harbor hydrophobic segments. These hydrophobic segments engage in hydrophobic interactions with the phospholipid tails, contributing to the stability of the proteins. The resulting structure was confirmed by ¹H and ¹³C NMR spectra ⁶². Physicochemical characteristics of the LNPs, such as size, shape, charge, polydispersity, and surface composition, each play a vital role in the cellular internalization and resultant efficacy of the particles. Interestingly, the size of the LNPs can be optimized to target specific organs ²⁰⁷. Choice of administration route also depends on LNP size in some cases ^208,209. Undoubtedly, size significantly influences different parameters. Thus, it is important to know the LNPs physicochemical properties. One common method for particle size, size distribution, and polydispersity measurement is dynamic light scattering (DLS). This characterization method is based on the Brownian motion of particles dispersed in a liquid. It can analyze nanoparticle diameter size distribution in the range of 0.5 to 1000 nm. DLS is an indirect method which calculates the particle shape pattern, which might not always accurately reflect the true particle shape ²¹⁰. Some important factors to be noted in DLS measurements include sample dilution needed to avoid multiple scattering and particle-particle interaction. This could result in the destabilization of the formulations. Also, the hydrodynamic radius of the particle measures is critical in the evaluation of DLS measurements. This assumes the particles are in the form of spheres. It need be noted that signals from large particles can mask the signals from smaller particles ²¹¹. Another method of size distribution measurement, that also happens to be based on Brownian motion of particles, is nanoparticle tracking analysis (NTA). Hydrodynamic diameter is calculated via using the Stokes-Einstein equation. NTA gives more accurate size distribution calculation than DLS. Still, it has similar limitations like DLS such as experiencing the effects of sample dilution ²¹². Size exclusion chromatography (SEC) can also be used in size distribution measurement. SEC is mainly used to purify the sample by removing free drugs from the drug-loaded NPs. The separation of particles primarily depends on the pore size of the solid or gel matrix ²¹³. LNP morphology can be identified by the microscope technique, such as negative-stain TEM (transmission electron microscopy) and cryo-TEM (cryo-transmission electron microscopy). Cryo-TEM gives detailed information on particle shape, size, and the membrane structure ²¹⁴. Another important physicochemical parameter is surface charge (i.e., ζ-potential) of the lipid nanoparticles, as such impacts cellular internalization ²¹⁵. The surface charge of LNPs usually depends on the phospholipid head groups, which incorporates positive-charged amines or negatively-charged carboxylate groups. The ζ-potential of an LNPs formulation indicates particle stability, as it regulates the strength of particle-particle interaction ⁶⁴. For stable suspensions, a ζ potential of less than −30mV or more than 30mV would provide sufficient inter-particle repulsion ²¹⁶. One study suggested that the ratio of DOPE and DOTAP was a critical factor affecting the ζ-potential ⁶⁵. Another study investigators observed that the ζ-potential of LNPs has a linear relationship with the molar ratio of ionic lipids in the lipid composition of an LNP formulation ⁶⁴. LNPs are physically stable (in terms of maintaining homogenous distribution and preventing self-aggregation) and maintain EE% during storage time. Here, EE% represents the percentage of a prepared drug that is encapsulated within the LNPs. EE% is determined after the LNPs are prepared and purified, ensuring that any free mRNA or mRNA adsorbed on the LNP surface is removed. EE% remains constant during storage serving to represent the initial amount of mRNA loaded into the LNPs. Förster resonance energy transfer (FRET) serves as a fluorescence-based technique permitting assessment of the quantify of LNP fusion ²¹⁷. This serves to expand prior works which examined the stability of the Cas9:sgRNA/80-O16B, 80-O16B, Cas9:sgRNA/80-N16B, 80-N16B NPs by measuring the change in average hydrodynamic diameter (<D_h>) over time; wherein a lack in D_h change suggested that LNPs maintained high stability and displayed negligible aggregation after two days of storage at room temperature ⁶². In another study, a novel polyethylene glycol phospholipid-modified cationic lipid nanoparticle (PLNP)-based delivery system was developed then analyzed for transfection stability. For this, the investigators compared the stability of representative PLNP/DNA-a to that of a control analogue (lipofectamine 2000-encapsulated Cas9-sgPLK-1a DNA (Lipo2000/DNA-a), a commercially available reagent) in different culture media with varying serum concentrations. They found that the size of the Lipo2000/DNA-a particles increased with an increase in serum concentration. The results indicated that PLNP/DNA-a was more stable than Lipo2000/DNA-a in an environment having a high serum concentration ⁶⁵. Similarly, the stability of the PLNP/DNA was evaluated in vitro at different pH conditions. Their studies revealed that lower pH values increased the diameter, thereby highlighting the instability of the particle’s structure. More specifically, these results indicated that PLNP/DNA particles could decompose under acidic conditions (e.g., in the endolysosomal system due its intravesicular pH of ~5) ⁶⁵. The stability of 5A2-DOT-10 LNP encapsulated with Cas9GFPP RNP complexes was determined throughout storing such at 4°C for two months. The PDI (polydispersity index) and size of the nanoparticles were tested at different times of storage. Analyzing the particle size diameter at different timepoints revealed that the size of the LNPs did not change and remained uniform (PDI < 0.2). Similarly, gene-editing efficiency was evaluated in HeLa-GFP cells after adding 24 nM sgRNA nanoparticles, with gene-editing quantification carried out after 3 days. The gene-editing efficiency of the 5A2-DOT-10 remained constant even after its storage of 60 days. ²¹⁸. How much mRNA is loaded during LNP preparation and remains encapsulated during storage is crucial in mRNA-LNP development and characterization. The RiboGreen^® assay was conducted to determine the EE% of sgRNA and Cas9mRNA in various lipids viz: L6-, L8-, L10-, and MC3-LNP obtained from the lipid library. The EE% was found to be high (> 90%) for the various nanoliposomes except for L1-cLNP (~65%) ²⁰⁶. The CRISPR-Cas9 LNPs were formulated for a single administration for persistent in vivo gene editing. In another study, 61.5% EE% of CRISPR-gold was found for complex Cas9 RNPs, which met the necessary efficiency for developing Cas9 delivery vehicles ⁶⁵. Zhen and co-workers used an encapsulation ratio (%) of [(A₂-A₁)/A₂] ˣ100 % to calculate the CRISPR-Cas9 crude content in the supernatant. The EE% of the nanoliposome complex was 96.87% ± 0.13 ²¹⁹. The EE% of the complex was further confirmed using gel retardation electrophoresis. The EE% of the complex was determined to be 95% by performing a grayscale analysis ²¹⁹. Stability should also be measured in terms of gene disruption efficiency. The cLNP formulations encapsulating GFP sgRNA were evaluated for in vitro gene disruption efficiency by analyzing GFP-associated fluorescence signal loss (via monitoring reduction of GFP signal) in HEK 293 cells expressing GFP (HEK293/GFP) for cancer therapy ²⁰⁶. In all, LNPs should be rationally designed such that they retain the mRNA payload during storage as well as efficiently release. The CRISPR-Cas9 gRNA-HPV16 E6/E7 nanoliposome complex release profile was evaluated after fitting the composite object release model as per the release profile in vitro. The R² value was found to be higher in the Higuchi equation, thereby indicating that the nanoliposome complex conformed to the nanoliposome characteristics. Also, the complex release rate was 37.7% at 3 hours. This meets the required liposome formulation guidelines as per the relevant pharmacopeia ²¹⁹. mRNA expression following its release after being taken up by its targeted cells can be measured using flow cytometry. The expression level of GFP and its congener BFP (blue fluorescent protein) in the CRISPR-gold treated BFP-HEK was quantified using flow cytometry. The analysis of BFP-HEK cells was performed after seven days of Cas9 treatment. The level of GFP and BFP expression was quantified from the visualized fluorescence intensities from localized cell specimens using fluorescence microscopy ⁶⁵.

Routes of LNPs Administration

In vivo biodistribution (BD) and PK of LNPs are highly affected by the selected route of administration ²²⁰. Moreover, its influence on the strength of the immune response as well as any potential side effect(s) makes it an important point of consideration. LNPs serve as unique delivery vehicles for transfection of genetic material that offer the ability to target specific cells and effectively deliver the therapeutic cargo to relevant intracellular sites. In the past few years, various delivery routes have been explored for mRNA-LNP delivery.

Parenteral

The first demonstration of parenteral delivery of LNPs surfaced in 1960 with the use of a parenteral fat emulsion of submicron size for parenteral nutrition (Intralipid^®221. Intravenous (IV) administration is preferably chosen when it is desired for mRNA-LNPs to be delivered into the liver. The major concerns with this delivery route were the extent of systemic toxicity and immunogenic response ²²². Therefore, IV delivery of mRNA-LNP is uncommon. Other parenteral routes include intramuscular (IM), intradermal (ID), and subcutaneous (SC) routes of administration (Figure 10). IM route remains the most common form of delivery and is easy to carry out. IM delivery of mRNA using LNPs has shown successful experimental outcomes in many preclinical and clinical studies ¹⁷¹. For instance, a robust immune response was observed by IM delivery of mRNA LNPs vaccines against H10N8 and H7N9 influenza viruses. This was seen in a phase 1 clinical trial ²²³. Intradermal (ID) injection involves the administration of LNPs directly into the dermis, directly below the epidermis. It was observed in some studies that ID administration requires one-fifth of a standard IM dose to achieve an equivalent response. ^224,225 In another study, mRNA-LNP vaccines administered via the ID route effectively induced a Th1 immune response and efficiently primed cytotoxic T-cells. Subcutaneous (SC) delivery for mRNA-LNP administration permits the LNPs to be administered to the connective tissue. Noteworthy, SC administration of mRNA-LNPs encoding fibroblast growth factor 21 (FGF21) was performed. This was for use of a therapeutic agent in type 2 diabetes (T2D) and non-alcoholic steatohepatitis (NASH). A therapeutic level of FGF21 following such administration was observed ²²⁶.

Figure 10.

The route of administration of the delivery vehicle affects the BD and PK properties of the desired cargo (The figure was created internally by the authors).

Local Administration

Several in vivo studies involving local administration of CRISPR-Cas9 with LNPs have shown promising results, such as their direct delivery in the inner ear ²²⁷, muscles ⁶⁵, brain ^22,228, and tumors ²²⁹. Localized ear delivery of Cas9:guide RNA complexes with cationic lipid mediated LNPs ameliorated hearing loss in a model of human deafness. They have designed (and validated in primary fibroblasts) in vitro genome-editing agents that preferentially disrupt the dominant deafness-associated allele in the Tmc1 (transmembrane channel-like 1) Beethoven (Bth) mouse model. Then, they subsequently injected Cas9:guide RNA lipid complexes that targeted the Bth allele into the cochlea of neonatal Bth/+ mice. They found substantially-reduced progressive hearing loss, higher hair cell survival rates, and lower auditory brainstem response (ABR) thresholds in injected ears compared to ears not injected with such or ears injected with complexes that target an unrelated gene ²²⁷. In another study by Lee et al, they demonstrated localized muscle delivery of vehicles composed of gold nanoparticles conjugated to DNA and complexed with cationic endosomal-disruptive polymers. This multi-complex NP was used to transport the Cas9 ribonucleoprotein and donor DNA following local injection such that efficient correction to the DNA mutation that causes Duchenne muscular dystrophy transpired with minimal off-target DNA damage ⁶⁵. Wang et al. reported tumor-localized delivery of the Cas9 protein and sgRNA plasmid using gold nanoclusters (GNs) encapsulated in the core of the nanocapsules. They modified the GNs with HIV-1-transactivator of transcription peptide, the cargo (Cas9/sgRNA) that can be delivered into cell nuclei that is subsequently utilized to treat the melanoma by designing sgRNA targeting Polo-like kinase-1 (Plk1) of the tumor. This nanocapsule (polyethylene glycol-lipid/GNs/Cas9 protein/sgPlk1 plasmid, LGCP) led to a > 70% down-regulation of Plk1 protein expression in A375 cells in vitro and suppressed melanoma progression by 75% in mouse models ²²⁹.

Inhalation

Direct delivery of mRNA to the lungs can be achieved through nebulization. Nebulized mRNA LNPs show distinct delivery outcomes compared to systemically administered LNPs. The biological environment in the lungs during nebulization differs in terms of biomolecules and cell types, presenting unique barriers for LNPs compared to bloodstream administration ²³⁰. Moreover, aerosol droplets experience shear stress, which causes particle fragmentation and decreases particle stability ²³¹. Some studies have demonstrated that lipid composition optimization can address this problem. For example, a higher molar concentration of PEGylated lipids was found to enhance the efficacy of mRNA LNPs administered by nebulization. They found that optimized mRNA-encoded LNPs protected mice from the H1N1 subtype of the influenza A virus more efficiently than the optimized mRNA LNPs administered systemically ²³⁰.

Intranasal

The conventional mRNA LNPs administration methods that include intravenous, IM, ID, and subcutaneous routes are associated with injection-related effects (pain, trouble with repeated administration, risk of infection, and low patient compliance). The intranasal route for mRNA LNPs delivery is one of the more ideal noninvasive routes of administration, as administration can be repeatedly performed with low infection risk and high patient compliance. In addition, higher permeability and vascularization of nasal mucosa make the nasal tissues more accessible than other tissues ^232,233. It is evidenced that mRNA LNPs can be given by nasal administration. Due to their colloidal nature, LNPs can be readily transported and distributed via their drainage throughout the lymphatics within the nasal cavity. Brain delivery is also benefited from this route of administration since LNPs can be delivered directly to the olfactory lobe, thereby bypassing the blood-brain barrier (BBB) to allow for higher bioavailability ^234,235. Several studies have demonstrated more promising results for mRNA LNP delivery to the brain via utilizing the intranasal route of delivery ²³⁶.

In utero

Most congenital disorders that occur due to protein deficiencies are treated with postnatal protein or enzyme supplements. We believe the high morbidity and mortality rates associated with these irreversible disorders can be minimized with the use of prenatal in utero mRNA delivery. Although viral-based vectors have been explored for this purpose, their higher immunogenicity and low genomic integration limit their application ¹²⁰. Several studies have been performed to optimize the efficacy of LNPs by screening from a library of LNP formulations for effective in utero mRNA delivery to the fetal liver, lung, and spleen. A wide range of mRNA including GFP (green fluorescence protein) and EPO (erythropoietin) mRNA has been found to be efficiently delivered by using a library-optimized LNP formulation ²³⁷. Library design is based on the advantages seen through recent developments of mRNA-LNP vaccines. These were produced based on a developed platform and the rapid manufacturing. The screened libraries led to an optimal mRNA-LNP vaccine for efficacy, stability, and toxicity. Each in composite highlights the critical aspects uncovered in the design of mRNA-LNP platforms that include the means to optimize those platforms.

Ocular Administration

Gene therapy via utilizing ocular administration is advantageous due to several reasons, such as affording a relatively low immune response and being readily effectuated. The use of a viral-based vector can be an option for ocular mRNA delivery; however, the major concern lies in its immunogenic response and potential for long-term toxicity ²³⁸. Non-viral-based vectors, especially LNPs, are found to be an attractive alternative to viral-based vectors due to their low immunogenicity and high safety profile ⁵.

Administration Routes

As mRNA-LNPs can be delivered through many different routes, their biodistribution, expression kinetics, and therapeutic efficacy is highly influenced by such ²²⁰. The specific application or specific desired outcome largely governs the route choice for the delivery of mRNA LNPs. For example, in the case of genetic disorders related to protein deficiency, mRNA LNPs delivery via utilizing IV can be considered a viable option for its administration route. The reasoning behind this is that when mRNA LNPs are administered intravenously, they primarily accumulate in the liver. As the liver is an organ that inherently produces an array of proteins, their accumulation in such lends itself the potential to affording improved outcomes. That is, IV administration can be used to produce proteins that are deficient in congenital, metabolic, or hematologic disorders ²³⁹. However, the major drawback of their IV administration is their tendency to activate the immune response and undergo high systemic clearance ²⁴⁰. Also, it can lead to unwanted accumulation of LNPs in different organs and induce off-target side effects. As such, the introduction of LNPs via performing IM, ID and SC injection are preferable routes of administration due to predefined tissue-specific delivery. The results from several studies have indicated that local administration of mRNA LNPs can effectively afford the expression of tissue-specific protein supplements, such as those associated with the eye, brain, and heart ^238,241,242. For treating prenatal congenital diseases, the delivery of LNPs via in utero administration can be the option of choice for achieving protein expression in the fetal liver, lung, and spleen ^120,237. Again, repeated drug delivery by invasive methods is not feasible in these instances. In such cases, inhalation and the intranasal route of administration can make mRNA delivery more patient-friendly. Additionally, the intranasal route can give better results when the mRNA LNPs are intended for brain delivery ²³⁵.

Targeted LNPs Delivery

Targeted mRNA LNP delivery can be a promising means to increase therapeutic efficacy and mitigate potential off-target effects by nontargeted tissue. One approach to make the LNPs target-specific is called selective organ targeting (SORT) ⁴⁹. To create SORT the investigators added a fifth lipid along with the four standard lipids for mRNA LNP formulations, which are ionizable cationic lipids, helper phospholipids, cholesterol, and PEGylated lipids. They modified the lipid composition based on the notion that the internal and external charge of LNPs affects their biodistribution ²⁴³. They observed that adding a permanently charged 1,2-dioleoyl-3-trimethylammonium-propane lipid (DOTAP) at an increasing molar percentage (0 to 100%) shifted luciferase expression from the liver to the spleen to the lung. Incorporation of a negatively charged SORT molecule, 2-dioleoylglycerolcero-3-phosphate at 10 to 40% achieved spleen-specific luciferase expression. They also observed that by using several other classes of permanent and ionizable lipids, they attained similar outcomes, thereby bolstering the utility of their approach. As an extension of their efforts, the same group explored the use of helper phospholipids. For doing so, they synthesized ionizable phospholipids and optimized them in a manner that afforded tissue-specific mRNA delivery as well as improvements in endosomal escape of mRNA ²¹. Currently. cancer is the primary cause of death. The major concern with conventional treatments (traditional chemotherapy) is the effect of off-target drug delivery. Intratumor LNP injection may afford promising results. However, it is more invasive. To overcome this barrier, a noninvasive anchored secondary scFv approach was developed. This method involves a targeting antibody that is added to the LNP membrane ²⁴⁴. The pK_a value of the lipids, which affects the ionization status of the LNP membrane, may also influence the scFv-based tissue-specific delivery by altering the electrostatic interactions with specific tissue. In one study, researchers found that LNPs having lipids with lower pK_a values on the membrane, such as 6.45 or 6.5, tend to distribute throughout the hepatocytes of the liver, whereas LNPs that have membrane lipids with higher pK_a values, such as 6.8 or 7.1, are more likely to be routed to the liver sinusoidal endothelial cells (LSECs) ²⁴⁵. Another prospect for achieving targeted tissue-specific delivery would be the specific modified mRNA translation systems (SMRTs). This is an approach that entails a self-controlled expression of mRNA inside a specific type of tissue irrespective of its cellular uptake by other tissue types. It is an on-off system containing two mRNA together in a single circuit (in series), which controls the mRNA expression in select tissue types and prevents expression in non-targeted tissue ²⁴⁶. Conjugating CD4⁺ antibodies to LNPs was found to lead to specific targeted and efficient mRNA translation in CD4⁺ T-cell populations demonstrating the efficacy of mRNA-based immunotherapeutic delivery. The study demonstrated that in vitro targeting of human T-cells with anti-CD4⁺-conjugated luciferase mRNA, resulted in robust binding (measured by radioactivity and flow cytometry) and efficient luciferase expression in CD4⁺ T-cells. They also assessed the functional potential of the conjugated mRNA-LNPs to facilitate genome editing using the Cre-LoxP mice reporter system and demonstrated successful genetic recombination (~60% efficiency in spleen), irrespective of the activation status (resting or activated) of CD4⁺ T-cells. Furthermore, a timepoint analysis showed a gradual decrease in the expression of mRNA in the spleen compared to the stable expression in lymph nodes. They suggested that a plausible reason for such could be the natively longer residence time of T-cells in lymph nodes than in the spleen. In sum, the study presented a CD4⁺ T-cell-targeted LNP platform for in vivo delivery of mRNA therapeutics ²⁴⁷. Most recent study by Verve Therapeutics’s team developed an optimized GalNAc-LNPs for targeted CRISPR base editing therapy in patients with low-density lipoprotein receptor (LDLR) deficiency by optimized N-acetylgalactosamine (GalNAc) for asialoglycoprotein receptor, ASGPR)-based ligand enhanced liver editing while reducing off-target effects. Their GalNAc-LNPs demonstrated high editing efficiency in LDLR deficient non-human primates and achieved long-lasting protein reduction in wild-type monkeys (https://www.nature.com/articles/s41467-023-37465-1)

PK and BD

Systemic delivery of gene therapy systems often results in rapid degradation and/or prompt clearance of the material. The genetic material can be protected from degradation by encapsulating it within a stable non-viral-based vector (e.g., in LNPs). One of the primary reasons attributed to the success of this approach is that LNPs can achieve better biodistribution and pharmacokinetic profiles of gene delivery in vivo ²⁴⁸. Replacing the hydrogen or carbon atoms of nucleic acids with radioactive isotopes ³H or ¹⁴C has little effect on the structure or PK behavior of the nucleic acid ²⁴⁹. However, van de Water et al. ²⁴⁸ found that the biodistribution and pharmacokinetics are different when they used ³H-versus ¹¹¹In-labeled unformatted siRNA. Hence, it is essential to choose a suitable isotope to monitor the PK and biodistribution behavior of LNPs. Studies used fluorescence imaging, due to its low cost and relative safety, to monitor the distribution and pharmacokinetics of LNPs during gene therapy ^49,103. One report explained that the LNP-gene therapy having positive charge could bind to different types of serum proteins, including heparin, albumin, and lipoprotein ²⁵⁰. The binding strength between the two is influenced by the charge density and surface structure of the lipid/plasmid complex. Also, extensive lung accumulation was observed likely due to entrapment of the LNP-gene formulation in lung capillaries ²⁵¹. The build-up of these in the lungs may be caused by the ionic interaction with the extensive surface area of endothelial cells. However, others found that negatively-charged LNP-gene complexes do not lead to lung accumulation ²⁵². When DOPE and a stearylated octa-arginine (STR-R8) peptide were employed as the inner coat and the outer coat consisting of cationic lipid di-octadecyltrimethylammonium propane-cholesterol-YSK05 distribution can be amended. In this instance optimized amounts of YSK05 (1-methyl-4,4-bis[(9Z,12Z)-octadeca-9,12-dienoxy]piperidine) and GALA (glutamic acid-alanine-leucine-alanine) peptides were used, which produced a high gene expression level in lung (>10⁷ RLU/mg protein) with high lung:liver and lung:spleen ratios. The investigators used the same exact amount and type of lipid and the same pDNA dose in a 1-step/2-step coating. The results showed that transfection activity was higher in lungs in the case of 2-step coating. In all double-coating improved the efficiency of gene expression.

Conclusions

Viral vectors are commonly used for CRISPR-Cas9 delivery, but they have drawbacks like insertional mutagenesis, undesired tissue expression, immune reactions, and translational challenges. Non-viral alternatives offer advantages such as biocompatibility, low biodegradability, minimal immunogenicity, tissue targeting, and the ability to deliver various payloads. These non-viral vectors aim to overcome limitations of viral vectors, including low loading efficiency and the need to split CRISPR-Cas9 components, to achieve effective delivery of CRISPR-Cas9 systems ex vivo and in vivo. ^229,253,254. Non-viral vectors offer advantages such as scalability, reproducibility, and cost-effectiveness over viral vectors, making them a preferred drug delivery option. However, addressing challenges like lower editing efficiency and payload issues requires further research. Recent advancements in nanotechnology have led to the exploration of various non-viral delivery systems for genome editing. Lipid-based synthetic nanoparticles/liposomes, polymeric/ polypeptide-based carriers, and natural extracellular vesicles are among the well-studied systems for CRISPR-Cas9 delivery ²⁰⁶. The development of non-viral-based vectors and lipid-based systems for delivering CRISPR cargos holds promise for in vivo gene excision. Factors affecting CRISPR-Cas9 editing efficiency include sgRNA sequence, nuclease exposure duration, delivery efficiency, and DSB repair mechanism. Design factors for nanoformulations also influence in vivo gene editing proficiency. However, developing delivery systems for clinical acceptance remains challenging. Selection of appropriate components for lipid-based LNPs depends on cargo carrier characteristics, lipid interaction, physicochemical properties, and tissue-targeting potential. Variations in gene disruption correlations raise questions about non-viral-based systems. Understanding CRISPR cargo carriers and lipid roles is crucial for LNP development. Effective delivery for therapeutic efficacy against viruses like HIV requires avoiding liver metabolism/clearance and targeting spleen and lymph nodes. Understanding each step of non-viral lipid-based vector gene transfer is necessary. Critical parameters for nanocomplex formulations include particle size, ζ-potential, charge ratio (N/P), endosomal escape, gene-loading efficiency, and release profile. Optimal electrostatic interactions between nucleic acids and lipids are important for encapsulation and binding. Various cationic lipids are widely used as nonviral-based gene carriers due to their high transfection efficiency ^229,255. Optimizing the ratio of cationic and neutral lipids to gene payload is crucial. Lipoplexes have demonstrated great potential for in vitro, ex vivo, and direct delivery of Cas9 plasmid/mRNA/protein and sgRNA. However, effective delivery of cargo carriers remains a challenge for CRISPR-based human engineering. Escape from the endosomal pathway by peptides (paradexin) ²⁵⁶, PEI-mediated osmotic pressure and liposome-mediated membrane fusion ²⁵⁷, use of pH-responsive swelling of lipids/polymers, phototherapy-based approaches ^62,258, NIR irradiation ⁴⁹ and cationic peptide-mediated membrane destabilization. All were explored in attempts to expanding LNP suitability ^206,259. Poor stability of LNPs in biological fluids upon exposure with serum also poses a challenge in the effective delivery of genome editing components ^235,260,261. Structural modifications of LNPs to lipid nano-shells offer an alternative cargo carrier system in such cases. These lipid nano-shells can also be explored for multicomponent systems with the Cas9 protein, plasmid DNA containing target genes, gold nanoparticles ²²⁹ and/or lipofectamine ⁶⁵. The targeting potential of LNPs also remains a critical point of consideration, as unwanted expression or undesirable tissue site targeting may lead to adverse effects. The risks of off-target cleavage by specificity, which leads to longer exposure to the Cas nuclease could result in undesired mutation ^{37,38,262–264}. Mismatch of sgRNA, due to tolerance of the Cas9 nuclease, results in expression at unsought genomic locales. Such undesirable genome editing in adjacent tissues also results in such adverse reaction and low therapeutic spectrum ²⁶⁴. In the recent past, advanced strategies, such as use of double Cas9 nickase ^73,264–266, have been explored to reduce off-target expression-associated side effects. Appropriate selection of suitable lipids, polymers and ligand-mediated surface decoration on nano-formulations also offered an alternative strategy for optimizing the expression at target sites. In a recent study, LNPs were explored for the successful delivery of Cas9 mRNA and sgRNA effectively to the liver in order to modulate expression of transthyretin in mouse with 97% reduction in serum protein level even after several months of treatment ¹⁶³. Recently, SORT NPs for tissue-specific mRNA delivery and CRISPR-Cas9 gene delivery have been explored. In this strategy, re-engineered LNPs by employing supplemental SORT molecule(s) for the targeting of Cas9 mRNA/sg RNA, were devised for selective targeting of lung, liver, and spleen ⁴⁹. Based on the current understanding of cellular barriers, the molecular structure of cell membrane, and cell surface topology, considerable progress has been made in utilizing ligand-based LNPs for effective targeting that results in limited adverse events linked to reduced off-target expression. RGD-mediated integrin receptor targeting via lipid nano-shells comprised of liposomal templated hydrogels can also improve the genome-editing efficacy of CRISPR-Cas9 components, including sgRNA targeting PLK1 ⁶⁵. Environment-based selective targeting also provides an option by using pH-sensitive lipids that respond to the pH of tumor tissue and endosomes for triggering and open stimulation of delivery systems ^267,268. pH-sensitive liposomes were also designed by mimicking the viral fusion with endosomal membrane using a protein at pH 5–6, which allowed for the delivery of their genetic material to the cytosol before reaching the lysosomes ^269,270. PE, a lipid-based membrane fusion promoter, is used in such nano-formulations ²⁷¹. Exploring various cellular markers, and unique receptor epitopes and understanding of physical conditions surrounding the target cells would be ideal. However, phenotypic heterogeneity among targeted cells remains a greater challenge for effective delivery through these approaches. Therefore, LNP-mediated direct delivery of Cas9 protein for rapid action and high efficiency should be further explored in the future to address the problem of off-targeted adverse events. Multi-layered NPs offer an exciting option for delivering CRISPR cargo carriers with considerable in vivo expression efficiency ²⁷². Multi-components such as a cationic core, organic shell and multifunctional ligands on the external surface of these artificial multi-layered LNPs exhibit excellent site-specific delivery, and thus provide improved genome-editing efficiency with low adverse effects ²⁷². For designing therapeutically active CRISPR-Cas9 delivery systems, proper consideration of the route of administration is also warranted to control the in vivo fate of these cargos. The success of IV-administered nano-formulations of CRISPR-Cas9/sgRNA in mouse models appears promising. However, detailed studies on their deposition in organs like liver, kidney, and lungs, as well as on-site expression efficiency, are necessary for effective nano-carrier development. Intra-tumoral administration has demonstrated better safety and site-specific targeting, resulting in successful gene knockout and greater localized effects. ^273,274. Ocular delivery or retinal targeting through intravitreal administration routes has also been reported to “knock-in” therapeutic genes using Cas9/PAMAM polymeric NPs ²¹⁹. Localized infusion of lipid-coated polymeric NP-mediated Cas13a/crRNA into the bladder cavity also showed excellent targeting potential with low systemic adverse effects ²⁰⁶. In view of these studies, safe gene editing at disease-specific targeted organs seems possible due to overcoming the associated challenges of formulation design and selection of compatible components in diseases like retinal degeneration, urinary tract disorders, and a variety of cancers. Furthermore, being of bacterial origin, the risk of SpCas9 in inducing immunogenic reactions in the recipient host is very high due to prolonged expression of bacterial Cas9 protein on MHC-I molecule, thereby resulting in the generation of antibodies and elicitation of T-lymphocyte-mediated immune response ^206,275. This immune cascade can result in adverse events, even leading to treatment failure. Designing LNPs faces a challenge due to the self-immunogenic potential of Cas9. While the lipids used in nano delivery systems are biocompatible and non-immunogenic, preventing premature exposure of the encapsulated gene in the system before reaching the target site is crucial to avoid unwanted immune responses. Ensuring effective loading and biochemical modification is essential for designing successful LNPs. Monitoring and evaluating humoral and cellular immune responses during clinical trials, along with considering ease of administration, are important factors that govern efficacy. Interactions with blood components and the inhibitory effects of serum present significant challenges for effective LNP delivery ²⁷⁶. To achieve successful clinical translation of LNPs, designing customized nano-carriers is crucial. This requires a deeper understanding of nano-bio interactions, safety considerations, loading capacity, compatibility with multifaceted strategies, spatiotemporal release, surface stabilization, electrostatic properties, and route of administration. Modifiable LNP components such as size, shape, charge, and surface properties should be tailored for accurate application in specific patient populations.

Table 1.

Microfluidics for LNP delivery ^{119,202–205}

Type of mixing device	Staggered herringbone micromixer	Toroidal mixer	T-junction	Hydrodynamic flow focusing	Baffle mixer
Schematic Design
Advantages	-Controlled mixing -High EE% (>95%), -Uniform particles (PDI < 0.1) -Scalable between small, large batches on parallelization.	-LNPs encapsulating siRNA by mixing preformed vesicles with siRNA in ethanol- water solutions and by mixing lipids in ethanol aqueous components containing siRNA	-Controlled rapid mixing -High EE% -Uniform particle size, -Broad solvent compatibility	-Can accelerate lipid convection to the mixing interface by reducing the migration distance of the organic stream -Precise control of LNP size -Low PDI203	-High EE% (>90%) - LNP size can be fine-tuned
Flow rate capacity	1–20 mL/min	1 mL/h to ≥ 20 mL/h	40–60 mL/min	0.025–2000 mL/min	0.1 mL/min
Particle size (diameter)	< 100 nm	< 100 nm	< 40 nm	< 150 nm	< 100 nm
Scalability	Multiple staggered herringbone micromixers in parallel station is required to achieve scalability under GMP conditions	One chip required to achieve scal-ability under GMP conditions	A general method of choice for the commercial production of large batches of LNPs	Fully scalable manufacturing is possible	Fully scalable manufacturing is possible

Table 2.

Prototypical clinical trials of LNP-based nucleic acid delivery systems.

LNP-based nucleic acid delivery systems for cancer therapy
Clinical Trials		Phase		Disease	Interventions	Route	Sponsor
NCT00882180		Phase I		Solid tumors	siRNA against VSP/VEGF-A	Intravenous infusion	Alnylam Pharma
NCT01262235		Phase I/II		Gastrointestinal Neuroendocrine Tumors	siRNA against PLK1	Intravenous infusion	Arbutus Biopharma
NCT02191878		Phase I/II		Hepatocellular carcinoma	siRNA against PLK1	Intravenous infusion	Arbutus Biopharma
NCT03323398		Phase I/II		Relapsed/Refrac tory Solid Tumor Malignancies or Lymphoma	mRNA-2416 encoding OX40L, alone (Phase I) or in combination with i.v. PD- L1 inhibitor, Durvalumab (Phase 2)	Intratumoral	Moderna TX, Inc.
NCT03323398		Phase I		Advanced/metas tatic solid tumors or lymphoma	mRNA encoding human OX40L	Intratumoral	Moderna TX, Inc.,
NCT03313778		Phase I		Solid tumors	mRNA encoding for personalized neoantigens	Intramuscul ar	ModernaT X, Inc.,
NCT02340156		Phase II		Recurrent glioblastoma	Plasmid vector encoding wild- type p53 DNA (+temozolomi de)	Intravenous infusion	SynerGene
NCT01505153		Phase I		Advanced and/or metastatic Cancer	Plasmid vector encoding shRNAs against STMN1	Intratumoral	Strike Bio
NCT02716012		Phase I		Advanced liver cancer	small activating RNA to activate the CEPBA gene	Intravenous infusion	Mina Alpha
NCT02410733		Phase I		Melanoma	Four mRNAs encoding melanoma- associated antigens (MAAs)	Intravenous infusion	BioNTech
NCT02316457		Phase I		Triple negative breast cancer	mRNAs encoding tumour- associated antigens (TAAs) and/or personalized neoantigens	Intravenous infusion	BioNTech
NCT01808638		Phase Ib/Iia		Advanced or metastatic pancreatic cancer	siRNA against PKN3 (+ gemcitabin e)	Intravenous infusion	Silence Therapeuti cs GmbH
NCT03739931		Phase I		Metastatic disease	mRNA-2752- OX40L, IL-23, and IL-36γ mRNA with LNPs and Durvalumab	Intratumoral	Moderna TX, Inc
NCT01591356		Phase I		Advanced cancers	siRNA against EphA2	Intravenous infusion	M.D. Anderson Cancer Center
LNP-based nucleic acid therapies for systemic diseases
Clinical Trials	Phase		Disease		Interventions	Route of immunizati on	Sponsor (s)
NCT04601051	Phase I		Familial Amyloid Polyneuropathy, Transthyretin- Related Familial Amyloid Cardio- myopathy, Wild- Type Transthyretin Cardiac Amyloidosis		NTLA-2001- CRISPR-Cas9 system delivers by LNPs	intravenous	Intellia
NCT05398029	Phase1b		CRISPR base-editing medicine to PCSK9 gene turn off		VERVE-101	intravenous	Verve
			hepatic protein production for heart disease
NCT03767270	Phase I/2		Ornithine Transcarbamylase Deficiency		MRT5201-human ornithine transcarbamyl-ase mRNA with lipid- based nanoparticles	intravenous	Translate Bio, Inc.
NCT04159103	Phase I/2		Propionic acidaemia		mRNA-3927 encoding Propionyl- CoA carboxylase	intravenous	Moderna TX, Inc.
NCT03375047	Phase I/2		Cystic fibrosis		MRT5005	Inhalation	Translate Bio, Inc.

LNPs are a successful delivery vehicle that allow for improved treatment outcomes for vaccines and therapies. This includes treatment and prevention for a broad range of genetic, infectious, and cancerous diseases. Clinical trials for cytotoxic chemotherapy agents, antibiotics, and nucleic acid therapeutics are outlined in this table. LNP-based COVID-19 mRNA vaccines from Pfizer/BioNTech and Moderna were successfully developed with international impact.

Statement of significance.

CRISPR-Cas9 gene editing for disease treatment and prevention is an emerging field that can change the outcome of many chronic debilitating disorders.

Abbreviations

AAT

Alpha-1-antitrypsin

AAVs

Adeno-associated viruses

AAVs

Adeno-associated-based Viruses

ABEs

Adenine base editors

ABR

Auditory brainstem response

Ad

Adenoviruses

AIDS

Acquired immunodeficiency syndrome

AMD

Age-related macular degeneration

ASSET

Anchored secondary scFv enabling targeting

AVs

Adenovirus-based Vectors

BBB

Blood-brain barrier

Bes

Base Editors

BFP

Blue fluorescent protein

BrdU

5-bromo-2′-deoxyuridine

Bth

Beethoven

CAR-T

Chimeric Antigen Receptor-T

CBEs

Cytosine base editors

CFTR

Cystic fibrosis transmembrane regulator

cLNPs

Cationic Lipid Nanoparticles

CReVIS-seq

CRISPR -enhanced viral integration site sequencing

cryo-TEM

cryo-transmission electron microscopy

CTL

Cytotoxic T lymphocytes

DLS

Dynamic light scattering

DSB

Double-stranded break

EE%

Encapsulation efficiency

EPO

Erythropoietin

FGF21

Fibroblast growth factor 21

FMLV

Friend Murine Leukemia Virus

FokI

A type II S restriction enzyme from Flavobacterium okeanokoites

FRET

Förster resonance energy transfer

FRR

Flow rate ratio

GALA

Glutamic acid-alanine-leucine-alanine

GFP

Green fluorescence protein

GNs

Gold nanoclusters

GUIDE-seq

Genome-wide unbiased identification of DSBs enabled by sequencing

GUVs

Giant unilamellar vesicles

HCAds

High capacity adenoviral vectors

HDR

Homology-directed repair

HIV

Human immunodeficiency virus

HPV16

Human papillomavirus 16

ID

Intradermal

IFN

Interferon

iLNPs

Ionizable Cationic LNPs

IM

Intramuscular

iNPs

Inorganic nanoparticles

ITR

Inverted Terminal Repeat

IV

Intravenous

IVT

in vitro transcription

LAM-HTGTS

Linear amplification-mediated high-throughput genome-wide translocation sequencing

Lipo2000/DNA-a

Lipofectamine 2000-encapsulated Cas9-sgPLK-1a DNA

LNAs

Locked nucleic acids

LNPs

Lipid nanoparticles

LSECs

Liver sinusoidal endothelial cells

LTR

5’- and 3’-Long Terminal Repeat

mAbs

Monoclonal antibodies

MHC-I

Major histocompatibility complex-I

MLVs

Multilamellar vesicles

MpAgo

Marinitoga piezophila argonaute

MRSA

Methicillin-resistant staphylococcus aureus

Nabs

Neutralizing antibodies

NASH

Nonalcoholic steatohepatitis

NASH

Non-alcoholic steatohepatitis

NHEJ

non-homologous end joining

NPs

Nanoparticles

NTA

Nanoparticle tracking analysis

ORF

Open Reading Frame

OTCD

Ornithine transcarbamylase deficiency

PAM

Protospacer adjacent motif

PDI

Polydispersity index

pDNA

Plasmid DNA

PE

Prime editing

pegRNA

Prime editing guide RNA

PEI

Polyethyleneimine

PK

Pharmacokinetic

PKR

RNA-dependent protein kinase

Plk1

Polo-like kinase-1

PLNP

Phospholipid-modified cationic lipid nanoparticle

RRs

Pattern recognition receptors

PS

Phosphonothioate

rAAV

Recombinant AAV

rAAV2-CFTR

Recombinant AAV2 capsid RBCs Red blood cells

RCR

Replication-competent retroviral vectors

Rig-I

Retinoic acid-inducible gene I

RNAi

RNA interference

RNP

Ribonucleoprotein

RT

Reverse transcriptase

SaCas9

Staphylococcus aureus Cas9

SARS-CoV-2

Severe acute respiratory syndrome coronavirus 2

SauriCas9

Staphylococcus auricularis Cas9

SC

Subcutaneous

SEC

Size exclusion chromatography

sgRNA

Single-guide RNA

siRNA

Small interfering RNA

SLNPs

Solid LNPs

SMRTs

Specific modified mRNA translation systems

SNAs

Spherical nucleic acids

SORT

Selective organ targeting

SRSRs

Short regularly- spaced repeats

ssODN

Single-stranded DNA oligonucleotides

STR-R8

Stearylated octa-arginine

SUVs

Small unilamellar vesicles

T2D

Type 2 diabetes

TALENs

Transcription activator-like effector nucleases

TCR

T-cell receptors

TEM

Transmission electron microscopy

TFR

Total flow rate

TLRs

Toll-like receptors

Tm

Transition temperature

Tmc1

Transmembrane channel-like 1

trcrRNA

Trans-activating CRISPR RNA

UCART19

Universal Chimeric Antigen Receptor 19 T-cells

UNAs

Unlocked nucleic acids

VLPs

Virus-like Particles

VNB

Vapor nanobubble

ZFNs

Zinc-finger nucleases