Navigating the intricate in-vivo journey of lipid nanoparticles tailored for the targeted delivery of RNA therapeutics: a quality-by-design approach

Abstract

RNA therapeutics, such as mRNA, siRNA, and CRISPR–Cas9, present exciting avenues for treating diverse diseases. However, their potential is commonly hindered by vulnerability to degradation and poor cellular uptake, requiring effective delivery systems. Lipid nanoparticles (LNPs) have emerged as a leading choice for in vivo RNA delivery, offering protection against degradation, enhanced cellular uptake, and facilitation of endosomal escape. However, LNPs encounter numerous challenges for targeted RNA delivery in vivo, demanding advanced particle engineering, surface functionalization with targeting ligands, and a profound comprehension of the biological milieu in which they function. This review explores the structural and physicochemical characteristics of LNPs, in-vivo fate, and customization for RNA therapeutics. We highlight the quality-by-design (QbD) approach for targeted delivery beyond the liver, focusing on biodistribution, immunogenicity, and toxicity. In addition, we explored the current challenges and strategies associated with LNPs for in-vivo RNA delivery, such as ensuring repeated-dose efficacy, safety, and tissue-specific gene delivery. Furthermore, we provide insights into the current clinical applications in various classes of diseases and finally prospects of LNPs in RNA therapeutics.

Graphical Abstract

Introduction

RNA therapies have emerged as superior alternatives for the rapid and efficient treatment of acute and life-threatening diseases compared with other approaches. With the potential to modulate gene expression or generate therapeutic proteins, RNA therapeutics are applicable for addressing diseases characterized by specific genetic targets. They can target intracellular proteins that are otherwise challenging to reach. Moreover, RNA therapeutics can be personalized and tailored to address patient-specific genetic abnormalities, resulting in fewer side effects. They are widely considered for numerous conditions, including infectious diseases, cancers, immune disorders, heart diseases, and neurological disorders [1]. RNA therapeutics encode proteins or facilitate DNA or RNA editing and are generally categorized into two groups: non-coding oligonucleotide RNAs (e.g. siRNAs, RNA aptamers, miRNAs, sgRNAs, and ribozymes) and macromolecular coding RNAs or mRNAs synthesized through in vitro transcription (IVT) [2]. RNA therapeutics are advancing in clinical development and application worldwide (Table 1). However, the clinical translation of RNA therapies faces significant obstacles, including instability, vulnerability to RNase degradation in the bloodstream and tissues, renal clearance, uptake by the reticuloendothelial system (RES), limited cellular uptake due to size and negative charge, endosomal trapping, off-target effects, and immunogenicity [3]. Scientists have developed strategies to overcome these challenges by appropriately selecting untranslated regions (UTRs), poly-A tail addition, capping, and nucleoside modifications [4]. In addition, diverse non-viral vectors have been devised for RNA delivery, with LNPs emerging as the primary carriers for RNA therapeutics [5,6,7]. LNPs efficiently encapsulate RNA molecules and deliver them to target cells. They offer substantial potential for nucleic acid delivery by protecting against nucleases, avoiding the mononuclear phagocyte system, exhibiting low immunogenicity and toxicity, facilitating cellular uptake, preventing endolysosomal degradation, minimizing non-specific interactions, and enabling easy synthesis and engineering [6]. Additionally, LNP formulation can be tuned to improve stability, transfection efficacy, and extrahepatic targeting [8, 9].

Table 1 Approved RNA therapeutics

In recent years, LNPs have been successfully employed for clinical applications and trials of various RNA therapeutics, such as siRNA, mRNA, and CRISPR–Cas9, to treat diseases like cancer, infections, and genetic disorders [10]. Notably, in 2020, LNP-encapsulated mRNA vaccines of BNT162b2 and mRNA-1273, produced by Pfizer-BioNTech and Moderna respectively, were developed to provoke immune responses against the SARS-CoV-2 spike protein [11]. Furthermore, Onpattro (patisiran), the first FDA-approved RNAi drug in 2018, is an LNP-based siRNA drug targeting the transthyretin gene for treating hereditary transthyretin-mediated amyloidosis. Leqvio (inclisiran) is another LNP-based siRNA drug that targets the PCSK9 gene to lower cholesterol levels and reduce the risk of cardiovascular events, receiving approval from the European Commission in 2020. NTLA-2001, an LNP-based CRISPR–Cas9 therapeutic, was introduced as the inaugural systemic in vivo application for hATTR amyloidosis treatment, advancing into clinical trials in 2020. Despite the numerous advantages of LNP-RNA therapeutics, they face challenges, including immune system interactions with LNP components and the difficulty of targeting various tissues in clinical settings due to the composition of RNA-LNPs.

This updated review focuses on LNP design and application for the growing siRNA and mRNA therapeutics. Notably, it explores the in-vivo fate of RNA-LNPs, their functionalization for targeting gene expression across various diseases via different administration routes with safety and immunogenicity concerns. The review begins with the composition and fabrication methods of RNA-LNPs, considering their structural and physicochemical attributes such as morphology, size, charge, encapsulation efficiency (EE), and stability, optimized via the quality-by-design (QbD) approach. Extensive discussions cover the in-vivo fate of RNA-LNPs, including interactions with blood components, protein corona formation, biodistribution, pharmacokinetics, systemic circulation, blood clearance, intracellular trafficking, and endosomal escape mechanisms. Furthermore, it highlights current challenges and strategies in using LNPs for in-vivo RNA delivery, such as repeated-dose efficacy, safety, immunogenicity, and LNP functionalization for tissue-specific gene delivery. Lastly, the review summarizes the current clinical applications and prospects of RNA-LNPs, encompassing protein replacement therapy, genome editing, and their use in treating various diseases such as hereditary conditions, cancers, and infectious diseases.

RNA-LNP: compositions and fabrication methods

RNA molecules

RNA therapeutics, composed of adenine, cytosine, guanine, and uracil bases linked to a phosphate backbone, function as ncRNAs such as siRNAs or miRNAs, which interfere with target mRNAs through Watson–Crick base-pairing. Additionally, they can act as coding RNAs (mRNAs) [12]. siRNAs have a well-defined structure, consisting of short (usually 20–24 base pairs) RNA duplexes with two-base overhangs in the 3′ region [13]. Additionally, IVT mRNA can be used for protein replacement therapy or immunization [14], avoiding irreversible genome changes and genetic risks associated with DNA-based therapeutics [15]. CRISPR-based genome editing can modify target RNA sequences to treat specific disorders [16]. RNA aptamers can block protein activity, similar to small-molecule inhibitors and antibodies [17]. Both coding and non-coding RNAs demonstrate considerable promise in clinical contexts across diverse medical disciplines. mRNA has attracted significant interest because of its pivotal role in mRNA-based vaccines, gene therapy, and gene editing techniques such as CRISPR–Cas9. Moreover, therapeutic mRNA provides tailored treatment avenues for genetic disorders and cancer. siRNAs have seen notable market success in gene silencing and regulation across various diseases like genetic disorders, metabolic conditions, and viral infections. Among these are drugs such as patisiran (Onpattro), givosiran (Givlaari), inclisiran (Leqvio), lumasiran (Oxlumo), and glecaprevir/pibrentasvir (Mavyret). ncRNAs function as disease biomarkers for diagnosis, prognosis, and treatment response prediction. They have also emerged as therapeutic targets in oncology, cardiology, and neurology. Collectively, both coding and ncRNAs are propelling advancements in personalized and precision medicine paradigms, reshaping clinical practices.

Chemically modified RNA backbones involve structural adjustments to RNA molecules through modifications of the phosphate backbone, ribose ring, nucleotides, and 3′- and 5′-terminals (Fig. 1A). These modifications can improve the stability, efficacy, and safety of RNA therapeutics for clinical use [18]. PS modification, which replaces one oxygen atom in the phosphate group with sulfur, improves cellular uptake by increasing hydrophobicity, nuclease resistance, and serum protein binding [19]. Modifications at the ribose 2′ position, such as 2′-F, 2′-MOE, 2′-O-Me, and 2′-LNA, enhance binding affinity and stabilize siRNAs against RNases while preventing immune activation [20, 21]. The 2′-F and 2′-O-Me modifications imitate the biophysical features of 2′-OH and can stabilize siRNAs against RNases while preventing siRNAs from activating innate immune receptors (TLR, MDA-5, and RIG-I); accordingly, all therapeutic siRNAs in clinical trials have 2′-F or 2′-O-Me modifications [22]. 5′- or 3′-RNA conjugates, like GalNAc, target the liver by binding to the ASGPR [23]. More complex modifications, such as PMO [24] and PNA [25], alter linking moieties while maintaining nucleobase pairing. Chemical modifications can be applied to sgRNAs for gene editing [26]. For instance, modifications such as 2′-O-M-3′PS (MS), 2′-O-M, or 2′-OM-3′thioPACE (MSP) at the 5′ and 3′ ends enhance genome editing efficiency [27]. The 2′-OM-3′-phosphonoacetate (MP) modification reduces off-target cleavage while maintaining on-target performance [28]. The 5′-H group modification increases activity and avoids innate immune responses [29]. Additionally, incorporating bridging NAs (2′,4′-BNA^NC[N-Me]) and LNA at specific sites in crRNA reduces off-target DNA cleavage by Cas9 [30]. Likewise, chemically modified nucleotides or conjugates are employed to improve the stability and pharmacokinetics of RNA aptamers [31]. Most clinical trial aptamers are modified with 5′-end PEG for increasing residence time [32], 3′end-capping strategy with inverted thymidine [33], 2′-substitutions on the sugar ring, and PS modifications can improve nuclease resistance and target binding affinity [34]. IVT mRNA, a ssRNA with a 5′ cap, ORF, flanking 5′ and 3′ UTRs, and a 3′ poly (A) tail, can include chemical modifications like Ψ, m¹Ψ, m⁵C, 5hmC, m⁵U, and s²U to minimize the innate immune responses [35, 36].

Fig. 1

Obstacles in the development of RNA therapeutics: A chemical modifications including alterations at the 5′ and 3′ ends (5′-capping, 3′-tail modifications, and 5′ and 3′-end conjugations), nucleotide modifications, ribose sugar substitutions at the 2′ position, and alterations in the phosphate backbone. B Nanocarrier delivery systems including lipid nanoparticle (LNP), lipoplex, polyplex, lipopolyplex, polymersome, polymeric micelle, exosome, and DNA nanostructures

Chemical modifications can enhance nuclease resistance and reduce immune activation. However, developing a safe and effective carrier to protect RNA from harmful physiological conditions and overcome gene delivery barriers is essential (Fig. 1B). Advances in nanotechnology provide potential solutions for intracellular delivery across biological barriers. Nanovectors, such as LNPs, offer key benefits by allowing optimization of their chemical properties, including size, shape, composition, and surface chemistry. The customization of surface properties is also allowed, including the functionalization with targeting ligands, potentially improving delivery specificity and efficiency.

Lipid compositions

LNPs are regarded as leading non-viral vectors for the in vivo delivery of siRNA, mRNA, and gene editing constructs [1] due to their superior encapsulation, biocompatibility, stability, cellular uptake efficiency, and ease of large-scale production [37]. Typically, LNPs consist of ICLs, helper lipids, cholesterol, and PEG-lipids, each playing a crucial role in their functionality (Fig. 2). The molar ratios of these lipid components typically range as follows: 40–50% ICLs, 10–12% helper lipids, 38–45% cholesterol, and 1–2% PEG-lipids. The relationship between lipids and RNA is expressed by the N:P ratio, which denotes the proportion of nitrogen atoms in ICLs to phosphate atoms in RNAs. It is also expressed as a mass ratio, reflecting the ICL weight relative to RNA [38]. Below, please find additional details regarding the lipid components:

Fig. 2

(Adopted from [38])

Structure of lipid nanoparticle (LNP) encapsulating nucleic acids. A Compact nanostructured core. The LNPs exhibit a compact, electron-rich core where the siRNA is tangled with positively charged lipids forming inverted micelle structures, enclosed by cholesterol and helper lipids; B the multilamellar assembly features an external bilayer and a dense inner core; C ‘Bleb’ configuration. The cationic lipid predominantly interacts with the mRNA, creating a solid core, while the helper lipid is mainly partitioned into a ‘bleb’ bilayer

Component 1: ICLs, a critical component, consist of headgroups containing tertiary amines, linkers, and lipid tails. The chemical structure, particularly the amine head group, significantly influences mRNA encapsulation and delivery efficiency [39]. Flexibility in charge is essential during LNP preparation, maintaining a neutral surface charge at physiological pH to prevent immune responses while becoming positively charged in endosomal environments for membrane fusion and effective cytoplasmic release [40]. Optimal ICLs for in vivo activity exhibit a pKa within the range of 6.2–6.5, ensuring neutrality at pH 7.4 and positivity at acidic pH (< 6.0) [41]. Common ICLs are illustrated in Fig. 3 A. DLinKDMA, initially identified as a lead compound with a dimethyl amino headgroup, was modified to increase the spacer length to the dioxolane linker, resulting in DLinKC2DMA, which improved the efficacy of the ICL. Further optimization led to DLinMC3DMA, a gold standard ICL used in LNP formulations for NAs, including Onpattro and Leqvio [42, 43]. The linker bridging the amine headgroup to the lipid tail is pivotal, influencing transfection efficiency, stability, and biodegradability [44]. Linkers are diverse, including ethers, esters, amides, and carbamates, and their functional group, spacing, and length affect transfection efficiency [45]. Maier et al. revealed that esterification at carbons 9 and 10 in the linalyl moiety conferred a potency akin to MC3-lipid, with an accelerated clearance [44]. Shirazi et al. engineered biodegradable multivalent cationic lipids with a disulfide linkage between the head and tail groups, enhancing cell viability in vitro [46]. Lipid tail characteristics, such as saturation and unsaturation, length, level of substitution, and structure, influence transfection efficiency [47]. The lipid tail structure affects particle internalization, with certain lipids providing optimal cellular uptake and capability for inducing endosomal membrane-disruptive H_II phases [48]. When examining various degrees of lipid tail unsaturation, linoleic chains, as found in DLinDMA, offer optimal particle internalization and endosomal release [48]. Recent advancements have led to a new class of ICLs for efficient mRNA delivery, including SM-102, L319, ALC-0315, and Lipid 5 (Fig. 3A) [49, 50]. SM-102 and ALC-0315 are the ICLs used in the Moderna Spikevax and BioNTech Pfizer Comirnaty COVID-19 vaccines [49, 50]. Through combinatorial library design, novel ICLs have been identified like A6, 113-O12B, 16-I, 4A3-Cits, and BAmP-TK-12 for targeted delivery to specific cells and improved mRNA delivery efficiency (Fig. 3A) [51, 52]. For example, Anderson et al. developed A6, an ionizable lipid with alkyne bonds, to target human erythropoietin mRNA delivery to hepatocytes [53]. LNPs composed of 113-O12B for lymph node targeting achieve mRNA delivery to 30% of APCs [54]. Moreover, ROS-sensitive LNPs containing BAmP-TK-12 exhibited targeted mRNA delivery and protein expression in cancer cells [52].

Fig. 3

Chemical compositions of lipid nanoparticles (LNPs): A ionizable cationic lipids (ICLs); B helper lipids; C cholesterol and its derivatives; D PEG-lipids

Component 2: Helper lipids, typical phospholipids like DSPC and DOPE, enhance LNP stability, cellular uptake, and endosomal release [27]. DSPC forms stable bilayer structures, which contribute to LNP stability and rigidity, while DOPE, with fusogenic characteristics, facilitates improved fusion of LNPs with endosomal membranes, thereby aiding RNA release (Fig. 3B) [55].

It was determined that ~ 30–40 mol% helper lipid is required to efficiently entrap siRNA within LNPs, providing additional insight into the role of these helper lipids [56]. The helper lipids serve as a mechanism for spacing out ionizable lipids to achieve a membrane surface charge of approximately + 1 per nm² (siRNA has a surface charge of approximately − 1 per nm²). For LNP-pDNA formulations, certain unsaturated phosphatidylcholines (i.e., SOPC and DOPC) improved the LNP activity over DSPC in vitro [57]. DOPE-containing LNP-pDNA systems showed the best activity in murine serum, suggesting a potential role of helper lipids in modifying the LNP surface affinity to distinct apolipoprotein subtypes. LoPresti et al. explored helper lipids with different properties (neutral, cationic, and anionic) and demonstrated that changes in the chemistry and concentration of these lipids influence nanoparticle surface charge, affecting their distribution and efficacy in specific tissues [58].

Component 3: Cholesterol, a naturally occurring lipid, stabilizes LNPs by intercalating into bilayer structures and protecting against serum protein interactions [59]. The role of cholesterol in RNA delivery systems is not fully understood. However, particles lacking cholesterol may extract cholesterol from lipoproteins in blood circulation, potentially leading to destabilization. Cholesterol enhances intracellular delivery by promoting fusion with cellular and endosomal membranes, especially when combined with unsaturated lipids [60]. The size of LNPs is also impacted by cholesterol content, exhibiting a reduction in particle size as cholesterol levels increase within the range of 10–60 mol% [61]. Cholesterol is an exchangeable molecule that can accumulate within the LNP during circulation, significantly reducing surface-bound protein and enhancing circulation half-lives [62]. It is also crucial for nucleic acid encapsulation. Specifically, at least 40 mol% cholesterol is required to achieve nearly complete siRNA encapsulation [63].

Various cholesterol types in LNP formulations have been investigated for their effects on cell transfection efficiency (Fig. 3C). For instance, the incorporation of 20α-OH cholesterol can redirect mRNA-LNP systems from hepatocytes to hepatic endothelial cells and Kupffer cells by altering the LNP surface affinity to distinct apolipoprotein subtypes or scavenger receptors [64]. Zeng et al. discovered that using β-sitosterol instead of cholesterol in LNP formulations can improve the mRNA transfection on immune cells and enhance T-cell proliferation and IL-2 production [65]. Cholesterol derivatives with C24 alkyl tails in LNPs lead to diverse morphologies, with modifications like methyl and ethyl groups increasing multilamellar structures and double bonds causing lipid partitioning, both enhancing gene delivery efficiency [66].

Component 4: PEG-lipids serve dual functions by controlling particle size during formation and improving stability by establishing a hydrophilic spatial hindrance. PEG-lipids prevent particle aggregation, enhance in vivo circulation lifetime, and reduce rapid clearance by RES [37]. However, PEG-lipids can diminish cellular uptake, necessitating a balance for effective intracellular delivery [67]. Very short PEG (< 1 kDa) lacks efficient circulation time, while cellular uptake is diminished by very long PEG chains (> 5 kDa) or high mole fractions (> 15 mol%). Typically, PEG molecules with a molecular weight = 2 kDa at a ratio lower than 5%, preferentially 0.5 mol%, are employed for LNP formulation [68]. Lokugamage et al. found that formulations without PEG-lipids resulted in unstable, polydisperse LNPs over 200 nm in diameter [69]. Increasing the molar ratio of PEG-lipids results in significantly smaller LNPs, regardless of other lipid components. PEG-lipids influence nucleic acid encapsulation efficiency, in vivo distribution, transfection efficiency, and immune response. These characteristics are linked to the molar ratio of PEG-lipids and the structure and length of the PEG chain and lipid tail [63]. Diffusible PEG-lipids contribute to particle stability and facilitate intracellular transport [70, 71]. PEG-DMPE and PEG-DMG with 14-carbon acyl chain rapidly dissociate from LNPs during circulation, whereas PEG-DSPE and PEG-DSG with 18-carbon acyl chain remain associated, extending their half-life at high concentrations [72]. Moderna’s COVID-19 vaccine contains 1.5% PEG2000-DMG, while BioNTech-Pfizer’s vaccine includes 1.6% ALC-0159 (Fig. 3D).

RNA-LNP fabrication

LNP fabrication involves the orchestrated interplay of distinct lipid derivatives crucial for optimizing RNA delivery. The self-assembly of LNPs relies on the electrostatic attraction between lipids and RNA in solution. Anionic mRNA initiates aggregation with cationic lipids, forming NPs via van der Waals forces. In an acidic buffer (e.g., pH = 4), ICLs in ethanol interact with negatively charged RNA, forming inverted micelles. Subsequent ethanol diffusion into the aqueous phase precipitates lipids, resulting in hydrophobic interactions [73]. The detailed mechanism suggests the migration of unbound ICLs into LNP cores, potentially forming multilamellar structures like onion or inverted hexagonal (H_II) structures [66]. A prompt and homogeneous blending of ethanol-solubilized lipids with an aqueous buffer containing dissolved RNA is essential for consistent particle generation.

Various conventional bulk mixing methods for NP synthesis, including thin-film hydration, solvent diffusion, pore extrusion, ultrasonication, melting emulsification, solvent injection, and bulk nanoprecipitation, have been employed; however, conventional approaches may yield inconsistent results (Fig. 4A). These methods for LNP synthesis have several disadvantages such as poor reproducibility due to variations in the synthesis process; requirement of a high amount of surfactant, which can affect the stability and functionality of LNPs; limited control over critical parameters like particle size, PDI, and EE; limited shelf life of LNPs due to chemical degradation of mRNA and impurities in the solution. Some traditional methods require harsh conditions that can negatively impact the integrity of the encapsulated mRNA or other therapeutic agents. These disadvantages highlight the need for improved and more controlled LNP preparation techniques to ensure consistent and effective delivery of therapeutic agents [74, 75]. Although post-treatment methods like membrane extrusion and ultrasonication have been proposed for LNP homogenization, issues remain. Payload stability and carrier deformation during these processes are still unresolved [76].

Fig. 4

Schematic representation of various lipid nanoparticle (LNP) fabrication techniques: A bulk mixing methods including thin-film hydration, solvent injection, bulk nanoprecipitation, and melting emulsification; B microfluidic methods utilizing distinct mixer designs such as T-junction, HFF, staggered herringbone, ring-type, Tesla, and baffle mixer

Microfluidics offers a more stable and reproducible method for producing LNP formulations. This technique provides precise tunability, short mixing time, low sample consumption, monodisperse particle size distribution, reproducibility, and the ability to scale up production. Key aspects of microfluidics include high controllability of TFR and FRR, as well as mixing conditions, scalability, real-time monitoring, and customized microchannel geometries [77]. Various mixer designs, such as T-junction, HFF, SHM, bifurcating or toroidal, baffle, and Tesla mixers, have been explored to enhance the efficiency of the mixing process and achieve rapid LNP formation [77] (Fig. 4B). The T-junction method, a basic microfluidic device, facilitates rapid mixing at high TFR by combining an aqueous medium with a lipid-organic solvent (ethanol), resulting in supersaturated lipid self-assembly into LNPs without a need for size reduction. HFF involves mixing lipid organic solutions with aqueous solutions in a microfluidic channel, allowing precise control over LNP size and EE. SHM induces chaotic mixing, offering a controlled size and shape by adjusting the TFR and FRR. Bifurcating toroidal mixers achieve large-scale production of uniform LNPs. Baffle mixers introduce successive turns for rapid mixing, allowing fine-tuning of the LNP size. The Tesla mixer efficiently achieves chaotic mixing, enabling a wide range of flow rates for smaller NP production [78, 79]. Each microfluidic design tackles distinct challenges, offering flexible choices for controlled and reproducible LNP synthesis. However, microfluidic mixing methods may pose concerns about environmental waste and expenses in case of employing single-use disposable cartridges, variability between cartridges, and the compatibility of cartridge materials with organic solvents, requiring careful design and application. Zöller et al. aimed to design charge-converting LNPs using a microfluidic mixing technique for their preparation and coating. LNPs consisting of DSPC, cholesterol, MPEG-2000-DSPE, and cationic surfactants were prepared at different flow rate ratios. Utilizing a second microfluidic chip, LNPs were coated with polyoxyethylene (9) nonylphenol monophosphate ester (PNPP) [80]. Roces et al. highlighted that microfluidics offers significant advantages for LNP manufacturing, notably in terms of scalability, reproducibility, and expedited preparation. Critical factors affecting the characteristics of LNPs were assessed, including FRR and TFR, both of which were determined to have a significant impact on particle size. Additionally, the selection of amino lipids, buffers, and nucleic acid payloads also played a crucial role in determining LNP properties. These findings serve as a valuable reference for the development and scaling of LNP systems utilizing microfluidics [61]. Ma et al. developed a scalable method for synthesizing LNPs using inertial microfluidic mixers that are scaled isometrically in three dimensions. A theoretical predictive model was employed to ensure a consistent mixing time across various chips, thereby achieving uniformity in particle size and distribution. siRNA-LNPs produced at various TFRs present a feasible approach for the scalable production of LNPs, potentially promoting the advancement of nucleic acid drug development into clinical applications [81].

Physicochemical attributes of RNA-LNPs

The physicochemical attributes of LNPs are crucial in determining their stability and functionality, directly influencing in vivo fate and gene delivery efficiency [55]. LNP structure and morphology are influenced by various factors, including lipid components and their ratios, preparation methods, and the type of NA.

Morphology

The morphology of LNPs affects cellular uptake, biodistribution, and toxicity. Common methods for assessing LNP morphology include TEM, SEM, and AFM. Cryo-TEM and scattering techniques, such as XRD, SAXS, and SANS, offer valuable insights into the inner structure, shape, and dimensions of LNP forms that may be challenging to identify through other means [82]. LNPs containing a high amount of ICL (more than 40 mol%), DSPC, and cholesterol can form inverted micelles around RNA therapeutics. Conversely, those with a lower amount of ICL (20 mol%) and a higher DSPC content (30 mol%) result in solid cores primarily composed of inverted micelles, with ICL, cholesterol, and DSPC accumulating in the bilayer area [83] (Fig. 2B). The use of DSPC and cholesterol in LNP formulations is strategically based on the interaction of DSPC with ICL through its phosphate group and its complexation with RNA through its choline group. Cholesterol occupies the space within the curvature created by the interaction of ICL with RNA-based therapeutics [45, 84]. PEG-lipids cover the outer layer of LNPs because of their hydrophilic nature and steric size [68]. Unlike siRNAs, mRNA leads to bleb formation, influencing morphology and EE [85] (Fig. 2C). Moreover, substituting DOPE for DSPC resulted in the absence of the bleb structure, developing a condensed core comprising inverted micellar structures [45].

Size and polydispersity

The size of LNPs significantly affects their stability, EE, and biological performance. Techniques such as NTA, DLS, and TEM are commonly used to measure LNP size. Smaller LNPs typically exhibit better stability, lower aggregation, higher EE, and longer circulation time than larger LNPs. However, very small LNPs may reduce gene expression efficiency and increase clearance. An ideal size for cellular internalization is approximately 100 nm, falling within the acceptable size range of 20–200 nm [40]. Factors like lipid type and ratio, NA concentration, preparation method, and storage conditions affect LNP size. Increasing the ICL content from 50 to 80 mol% while decreasing the cholesterol content from 37.5 to 7.5 mol% results in LNPs with sizes increasing from 50 to 100 nm [83]. In addition to mean size, the PDI) indicates size distribution homogeneity and is evaluated using techniques like DLS and NTA. High PDI can lead to inconsistent transfection efficiency, reduced stability, increased aggregation, and heightened immune recognition.

Zeta potential

Zeta potential, which influences LNP circulation and tissue uptake, can be modified by lipid type and ratio, particularly those with ionizable amine groups or PEG chains. Techniques like zeta potential analysis and electrophoretic light scattering measure the LNP charge. While positively charged LNPs have a higher affinity for negatively charged NAs, facilitating their encapsulation and protection, they may also pose higher toxicity risks due to increased internalization, interaction with the cell membrane, lower stability, higher aggregation, and stronger immune response compared with neutral or negatively charged LNPs. An elevated surface charge, whether + 30 mV or − 30 mV, usually leads to rapid elimination from the bloodstream and increased sequestration by the RES. In contrast, LNPs with a neutral charge, between − 10 and + 10 mV, tend to circulate in the blood with a reduced uptake by the RES.

Encapsulation efficiency (EE)

EE, a critical determinant of gene delivery effectiveness and cost-effectiveness for LNPs, reflects the degree to which NAs are safeguarded from degradation and successfully delivered to target cells. Various established techniques, including ultracentrifugation, agarose gel electrophoresis retardation assay, UV–vis spectrophotometry, fluorescence spectrophotometry, and ethidium bromide intercalation assay, are commonly used to quantify EE. Factors such as the type and ratio of lipids, concentration and molecular weight of NAs, preparation method, and storage conditions influence EE%. For example, changes in ICL content notably impact EE, as evidenced by a 30% increase in the molar ratio of ICL, leading to a proportional reduction in cholesterol amount and subsequent decline in siRNA EE [45]. Inadequate molar ratios, particularly cholesterol, diminish the interaction of ICLs with NAs, resulting in decreased EE. In addition, high proportions of PEG-lipid (2.5 or 5 mol%) reduce EE [83].

Stability

Stability studies are vital for translating LNP-RNA formulations from research to practical application, ensuring the determination of shelf life and optimal storage conditions. Unstable LNPs may undergo issues like aggregation, degradation, leakage, or phase transition, jeopardizing their gene delivery efficacy and safety. Stability testing should be conducted early in preclinical development. Common techniques for stability assessment include DLS, zeta potential analysis, TEM, UV–vis spectrophotometry, fluorescence spectrophotometry, and in vitro/in vivo gene expression assays, aligned with the ICH guidelines [50, 86]. The stability of LNP-RNA formulations depends on factors such as pH, temperature, light, humidity, ionic strength, lipid type and ratio, and NA properties [39, 50]. To augment the stability of LNP-RNA constructs, techniques such as the incorporation of buffers and surfactants, application of stringent processing controls, and using cryoprotectants during lyophilization are employed [86]. For example, both Moderna and BioNTech/Pfizer use sucrose as a cryoprotectant in their mRNA COVID-19 vaccines, incorporating specific buffering systems. Moderna employs Tris HCl buffer to maintain the final pH between 7 and 8, whereas BioNTech/Pfizer uses phosphate buffer. Choosing the appropriate buffering agent and osmolyte is of paramount importance because of the potential pH changes during freezing and storage at ultra-low temperatures [50, 87].

Overall, controlling the quality attributes in LNPs, such as average size, size distribution, zeta potential, structural integrity, loading capacity, and stability, is crucial for the successful delivery and therapeutic application of RNA. An in-depth examination of the physicochemical characteristics of LNPs supports the QbD approach by identifying critical attributes, facilitating risk evaluation, directing experimental frameworks, and ensuring effective control measures. These elements collectively contribute to the development of high-quality pharmaceutical products [88], which will be elaborated in the subsequent section.

RNA-LNP QbD

QbD is a systematic approach used in pharmaceutical development to ensure consistent adherence to predefined quality standards, while also aiming to improve efficacy and safety [89]. The QbD process begins with formulating the QTPP, which outlines the desired attributes of the final product, encompassing efficacy, safety, and other pertinent characteristics. Subsequently, CQAs are identified, representing the specific traits essential for guaranteeing product quality and performance. To attain these CQAs, a comprehensive understanding of CPPs and CMAs throughout drug development and manufacturing is imperative. CMAs pertain to the inherent characteristics of raw materials used in the manufacturing process, such as the properties of the drug and excipients. Conversely, CPPs encompass the pivotal variables within the manufacturing process that must be regulated to ensure the desired product quality. Risk assessment methodologies are frequently employed to prioritize and evaluate the impact of diverse CMAs and CPPs on CQAs, facilitating the identification of the most critical factors necessitating control or optimization to achieve the desired product quality. Once these critical factors are identified, a DoE approach is used to systematically explore how each factor influences product quality. The DoE conducts a series of experiments testing different combinations of CMAs and CPPs to establish a mathematical model of the manufacturing process. This model aids in understanding how alterations in input parameters affect the output (product quality), facilitating the development of robust manufacturing processes [90]. Statistical experimental designs, such as full or fractional factorial designs, RSM, and Box–Behnken designs are employed for DoE [91]. Factorial Designs involve studying the effects of two or more factors simultaneously. Full factorial designs consider all possible combinations of factors at different levels and provide comprehensive information about the interaction between factors. A full factorial design can be used to investigate the effects of multiple formulation variables, such as lipid type, lipid concentration, and solvent type on the particle size and encapsulation efficiency. This approach helps in identifying the most significant factors and their interactions [92]. Fractional factorial designs consider only a subset of the full factorial design that includes only a fraction of the total possible combinations. This design is useful when the number of factors is large, and a full factorial design would be impractical. For example, it can be used to quickly identify which formulation parameters (e.g., lipid type, surfactant type, solvent type) have the most significant impact on particle size and stability [93, 94]. RSM is used to explore the relationships between several explanatory variables and one or more response variables. It helps in identifying the optimal conditions for the desired response. For example, researchers might use RSM to determine the optimal lipid-to-drug ratio and surfactant concentration to maximize encapsulation efficiency and minimize particle size [95]. Box-Behnken is a type of RSM that does not include combinations where all factors are at their extreme values simultaneously. This design is efficient and requires fewer runs than a full factorial design [96, 97]. For instance, it can help in understanding how the interactions among lipid type, stirring speed, and temperature affect the stability and size distribution of the NPs [98, 99]. Factors affecting responses were validated statistically, using ANOVA, and different models (linear, two-factor interaction (2FI), cubic, quadratic) were applied. Three-dimensional surface and contour plots were used to illustrate significant effects [100].

In the QbD approach for LNP formulation, several key concepts are used to ensure the QTPP (efficacy, quality, and safety) of the final product. CQAs, representing specific traits of LNPs crucial for ensuring their QTPP, encompass factors like particle size, polydispersity, zeta potential, EE, and stability. On the other hand, CMAs, denoting inherent characteristics of the raw materials used in LNP manufacturing, may comprise lipid type and concentration, RNA integrity and concentration, stabilizer, and the presence of impurities or contaminants. CPPs are pivotal variables within the manufacturing process that directly shape the formation, characteristics, and performance of LNPs. Examples of CPPs in LNP formulations may involve parameters such as mixing speed and duration, temperature, pH, lipid-to-NA (N/P) ratio, solvent type and concentration, and the methodology employed for NP characterization (Fig. 5). By effectively identifying and managing these CQAs, CMAs, and CPPs throughout the development and manufacturing stages of LNP formulations, researchers can optimize the quality, efficacy, and safety of the final product. Kauffman et al. optimized erythropoietin (EPO)-mRNA-loaded LNPs. Six input variables were considered: C12-200:mRNA weight ratio, phospholipid type, C12-200, cholesterol, and PEG molar contents. Output variables included %EE, particle size, PDI, and EPO serum concentration. Initial screening determined the optimal phospholipid type. Further screening refined the C12-200:mRNA ratio, C12-200%, phospholipid %, PEG %, and cholesterol %. Final optimization identified the best composition: 10:1 C12-200:mRNA weight ratio, 35% C12-200, 16% DOPE, 46.5% cholesterol, and 2.5% C14-PEG2000. The optimal LNPs had a particle size of 102 nm, PDI of 0.158, 43% encapsulation efficiency, pKa of 6.9, and zeta potential of − 5.0 mV. The in vivo efficacy achieved with the injection of 15 µg EPO mRNA was measured at 7065 ± 513 ng/µL [93]. In another study, Blakney et al. optimized SAM LNPs using a full factorial design. They examined five variables: lipid type (C12-200, DDA, DOTAP, cephalin), lipid-to-RNA ratio (1:1, 4:1, 18:1, 90:1), lipid concentration (high, medium, low), particle concentration (high, medium, low), and cationic to zwitterionic lipid ratio. The key outcome was luciferase expression after 10 days. Cephalin lipids at an 18:1 lipid-to-RNA ratio, with low lipid and medium particle concentrations, resulted in a sevenfold increase in luciferase expression compared to the original formulation [92]. Ly et al. optimized LNPs for SAM expression and cellular activation using a Box–Behnken design. The study tested formulations varying in phospholipid content, ionizable lipid type, ionizable lipid content, and pH. The N/P ratio, DOPE and DMG-PEG-2000 contents, flow rate, and temperature were fixed. Output variables included particle size, PDI, EE%, charge, and protein expression. RSM modeling utilized second-order OLS regression, with a Box–Cox transformation to improve accuracy. Optimal conditions were determined using the desirability function and BFGS optimization [98]. Bastogne et al. optimized cationic nano-lipid formulations for siRNA transfection using DoE based on a D-optimal mixture design, testing 36 formulations. The study varied the proportion of DOTAP, PEG surfactant concentration, lecithin proportion, and LNP size. Key outputs included LNP stability, siRNA transfection rate (minimum 30%), and PDI. Polynomial models were employed to link input and output variables, and evaluated using Bayesian estimation, posterior predictive checks, and leave-one-out cross-validation [101]. Young et al. optimized LNP compositions for mRNA delivery to the placenta using a factorial design study. They created 18 unique LNPs (A1–A18) via definitive screening design (DSD). Factors included ionizable lipid type (C12-200, DLin-MC3-DMA), phospholipid type (DSPC, DOPE), and various molar percentages of lipids and cholesterol. Key metrics measured were hydrodynamic diameter (72.2–171.5 nm), polydispersity index (0.120–0.317), mRNA encapsulation efficiency (35.6–83.2%), transfection efficiency, and apparent pKa (5.3–7.1). The optimal formulation contained 35% C12-200, 10% DOPE, 1.5% PEG, and 53.5% cholesterol, with a hydrodynamic diameter of 130.2 nm, PDI of 0.064, EE of 56.5%, and pKa of 6.6 [102].

Fig. 5

Fundamental procedures for applying the quality-by-design (QbD) approach in developing lipid nanoparticles (LNPs) which includes quality target product profile (QTPP), critical quality attributes (CQAs), critical process parameters (CPPs) and critical material attributes (CMAs). In the QbD approach by using statistical, analytical, and risk-management methodologies optimized products can be designed and developed

Overall, by identifying the desired QTPP and managing CQAs, CMAs, and CPPs throughout the development process, along with implementing a risk control strategy for manufacturing and continuous process improvement, researchers can enhance the quality, efficacy, and safety of the final product [103].

In-vivo fate of RNA-LNPs

The in-vivo fate of LNPs and RNA cargo depends on several factors, including LNP composition, size, surface charge, PEGylation, injection route, target tissue, and immune response. Concerning LNP composition, the choice of ICLs, helper lipids, cholesterol, and PEG-lipids affects stability, fusogenicity, endosomal escape, cellular uptake, and transfection efficiency. Particle size affects circulation time, biodistribution, aggregation, and clearance. Larger particles enhance gene delivery efficiency, whereas smaller particles provide better control over body disposition. Surface charge, determined by ionizable lipids and their pKa values, influences interactions with serum proteins, cellular membranes, endosomes, and immune cells. Optimal gene delivery requires a balance between positive and negative charges. PEGylation, achieved through PEG-lipids, forms a hydrophilic steric barrier, preventing aggregation and improving circulation lifetimes, but potentially inhibiting cellular uptake. The chemistry and density of PEGylation on the surface of LNPs can also affect membrane permeability, biodistribution, and immune response. The injection route influences the initial exposure of LNP to the blood circulation and organs, yielding varied biodistribution patterns and gene expression levels. Target tissue, the intended organ or cell type for mRNA delivery, impacts LNP biodistribution by influencing interactions with tissue-specific receptors, transporters, enzymes, and immune cells. The immune response, involving innate and adaptive systems, influences LNP biodistribution through opsonization, phagocytosis, cytokine production, and Ab formation. LNPs with high surface charge or PEGylation can activate complement and undergo macrophage uptake, whereas those with low surface charge or PEGylation may evade immune recognition, extending circulation time [104]. Several rational design strategies can enhance the fate of RNA therapeutics loaded into LNPs. For example, adding electrolytes, such as sodium chloride, can mitigate the repulsion between cationic liposomes and lipoplex intermediates during complex formation. This contributes to stabilizing mRNA encapsulation, reducing particle size, preventing aggregation in saline mixtures, enhancing mRNA stability in serum, and improving gene expression administration [105]. Therefore, advancements in the rational design and manufacturing of LNPs offer promising avenues for enhancing in vivo performance in delivering RNA therapeutics.

Preclinical evaluation tools

Achieving successful translation of LNPs requires careful control of their properties, which influence colloidal stability, decomposition, biodistribution, and cellular interactions. These properties can change when LNPs are exposed to biological fluids. Therefore, conducting in-vitro assessments of LNPs in media that simulate biological conditions is essential for accurately predicting their physiological behavior [106]. In-vivo characterization studies of LNPs encompass a variety of methods to evaluate their performance, distribution, effectiveness, and safety. For biodistribution and pharmacokinetic analysis, which entails examining how LNPs distribute throughout the body after administration, various techniques are employed, such as radiolabeling, luminescence or fluorescence bioimaging, microscopy, and organ perfusion techniques. Radiolabeling is the least cumbersome approach for quantifying LNP biodistribution [107]. Measuring fluorescence encounters challenges from biological autofluorescence but using probes like Cy7 allows for tracking LNPs in live mice organs via confocal microscopy [108]. Liver perfusion is particularly effective for evaluating LNPs’ local distribution, offering simplicity compared to other tissue perfusion methods for organs like the brain, lungs, kidneys, muscles, and tumors [109]. Preclinical biodistribution data for approved RNA therapeutics can be accessed in the FDA’s Application Review Files or the European Public Assessment Reports. Reporter genes, such as Fluc, offer a high-sensitivity method for analyzing in-vivo gene expression. QWBA and MAR map the dispersion of radiolabeled compounds in tissue sections, providing cellular distribution data. They rely on specific radiolabeling procedures. For example, preclinical studies for Onpattro, Oxlumo, Leqvio, Givlaari, and COVID-19 vaccines utilized QWBA in rats. LC–MS/MS, a gold standard for bioanalytical chemistry, measures siRNA levels (e.g., Onpattro and Givlaari [110, 111]) and evaluates the pharmacokinetics and biodistribution of synthetic lipid components (e.g., Onpattro and the Pfizer/BioNTech COVID-19 vaccine [112, 113]). However, MS-based quantifications face challenges with high molecular weight and anionic RNA therapeutics. It enables the visualization of single-molecule RNA and multiple target mRNAs simultaneously, while RT-qPCR and ddPCR offer improved sensitivity, precision, and reproducibility for RNA biodistribution studies. In vivo BLI provides insights into mRNA-LNP biodistribution and translation kinetics using luminescent and fluorescent reporter proteins or labeled LNPs (Fig. 6). Therefore, the choice of evaluation tools depends on the specific requirements and challenges associated with RNA therapeutics, each offering unique advantages and limitations [114].

Fig. 6

Summary of common labeling and identification techniques for LNPs, RNAs, and protein products. Lipid nanoparticles (LNPs) are labeled with fluorescent and radioisotope tags for imaging purposes (1 and 2). LNP components are analyzed using mass spectroscopy (MS) (3). Specific nucleic acid sequences are detected using reverse transcription-quantitative PCR (RT-qPCR) (4). Reporter gene expressions, such as enhanced green fluorescent protein (eGFP) and luciferase (LUC), are visualized through fluorescent and bioluminescent imaging (5). Immunohistochemistry (IHC) techniques identify proteins produced by mRNA-LNP (6) or suppressed by siRNA-LNP. Fluorescence in situ hybridization (FISH) probes bind to specific nucleic acid sequences within cells, enabling their visualization (7)

Blood components forming protein corona

Upon entry into the body, naked genes engage with various biological components, depending on the specific biological fluids at the administration site, such as lung surfactants during inhalation, interstitial fluid with local injection, and blood plasma post-intravenous administration [115]. The efficacy of in vivo gene delivery significantly hinges on interactions with blood components, encompassing blood cells, ions, and plasma proteins. Cationic non-viral vectors, like lipoplexes, may aggregate in ionic physiological media that proteins like serum albumin can prevent. Similarly, erythrocytes in phosphate-buffered saline exhibit comparable interactions, which can be mitigated by adding proteins [116]. Plasma proteins binding to LNPs form a “biomolecular corona,” comprised of diverse biomolecules such as apolipoproteins, complement elements, immunoglobulins, and coagulation factors [117]. The composition of this corona, determined by LNP characteristics and the presence of NAs, critically alters their physiological attributes, impacting distribution, longevity in the bloodstream, cellular internalization, immune recognition, stability, and pharmacokinetics [118, 119]. LNP surface is predominantly coated with biomolecules like ApoE, vitronectin, albumin, and β2-glycoprotein I, which form stable associations influencing further interactions. The protein corona comprises a firmly attached “hard” corona and a more transient “soft” corona, dynamically interacting with surrounding proteins. Equipped with inherent targeting abilities, the protein corona enhances receptor-specific interactions and uptake via endocytosis [120]. Factors such as biofluid type, concentration, temperature, and origin crucially shape the corona, dictating LNP clearance, tissue targeting, and therapeutic effectiveness. ICLs, altering their charge based on pH, utilize ApoE to enhance liver-targeted gene delivery, where the level of ApoE influences delivery efficiency, while lipoproteins generally suppress gene expression. In liposomes containing DOTAP and cholesterol, fibronectin plays a pivotal role in gene expression in the lung, as does albumin corona in inflammatory liver and tumor conditions [121]. Hence, designing LNPs necessitates careful modulation of protein interactions for natural targeting or minimizing blood cell interactions for passive or active targeting.

Liver accumulation and extrahepatic distribution

The distribution of non-viral vectors in target tissues is heterogeneous and not completely understood, influenced by variations in blood flow and the diversity of receptor expression. The vascular barrier poses a significant challenge, hindering gene delivery to cells beyond the blood vessels. However, some tissues, such as the liver, spleen, and tumors, feature endothelial fenestrations that aid in delivering genes to cells outside the bloodstream. Notably, the size of liver sinusoidal endothelial fenestrae ranges from 100 to 160 nm, with variations across different species [122]. For effective hepatocyte delivery, carriers must be smaller than the endothelial fenestrations. NPs pass through murine tumor vessel walls more easily than human tumors, which show variable permeability depending on the cancer type [123]. Researchers should adjust NP size for optimal vascular permeability during pre- and post-administration phases [124]. When dealing with tight vascular structures like the blood–brain barrier, a simple particle size control is insufficient. An effective strategy may involve leveraging transcytosis by endothelial cells and targeting specific receptors such as the transferrin receptor [125]. Alternatively, physical stimuli such as ultrasound can disrupt the vessel wall, facilitating gene transfer into the tissue [126].

LNPs naturally exhibit a strong affinity for the liver, making them ideal carriers for hepatic delivery [127]. The presence of ApoE in the protein corona of LNPs facilitates their interaction with cell receptors such as LDLR, VLDLR, and LRP, predominantly in the liver. The high affinity of ApoE for LDLR on hepatocytes enhances LNP uptake via receptor-mediated endocytosis. Cheng et al. observed that the PEG-lipid component of LNPs can detach, allowing ApoE to adsorb onto the LNP surface [128]. Beyond liver targeting, different components of the protein corona can direct LNPs to extrahepatic tissues or organs [129, 130]. HDL on the surface of LNP interacts with the SR-BI receptors on lymphatic endothelial cells [131]. Vitronectin facilitates targeted delivery of LNPs to tumor cells through interactions with α_vβ₃ integrins [132]. LNPs with a vitronectin-rich corona show reduced delivery to HepG2 cells with LDL receptors compared to those with an apolipoprotein-rich corona [133]. Coronas rich in albumin extend circulation time and facilitate lymphatic drainage [134]. Gp60 and SPARC receptors are the key targets for delivering LNPs with an albumin-rich corona [135]. A vitronectin-rich corona may inherently target the lungs and tumors, while the affinity of b2-GPI for phosphatidylserine allows for spleen targeting by LNPs [132]. Figure 7 illustrates LNP natural targeting via protein corona formation, highlighting how the corona composition serves as a strategy for natural LNP targeting.

Fig. 7

Natural targeting of various cells by LNP-RNA therapeutics influenced by PEG-lipid shedding and protein corona composition. ApoE, high-density lipoprotein (HDL), vitronectin, and albumin can naturally target LNPs to the liver parenchyma, lymphatic vessel, and tumor and breast cancer cells, respectively

The protein corona composition on LNPs, enriched with specific proteins, is primarily determined by the controlled attributes of the LNPs. Notably, differences in the protein corona between liver and lung-targeted LNPs are unrelated to surface charge, suggesting that the introduction of amide bonds influences the corona composition, thereby facilitating lung-specific mRNA delivery [136]. Dahlman et al. have demonstrated that altering the length of the hydrophobic tail in lipids derived from 2,5-piperazinedione enables RNA delivery independent of ApoE or LDLR, highlighting the significance of lipid chemical structure in directing LNP navigation [137]. Another study found that modifying LNP helper lipids can change tissue tropism: DSPC-containing LNPs accumulate in the spleen, while DOPE-containing LNPs target the liver due to different affinities for ApoE [118]. The presence of proteins in the LNP corona does not guarantee targeted receptor interaction. Factors such as protein availability, conformation, and orientation are crucial, as not all proteins function as targeting agents; so, understanding how adsorbed proteins on LNPs affect cellular recognition is vital for targeted delivery [138]. Chan et al. used the corona “fingerprint” of gold NPs, suggesting that cellular uptake prediction is more accurate when based on the corona rather than particle size, charge, or clustering. However, in LNPs, only a minority of corona proteins enhance cellular association, highlighting the complexity of developing targeted delivery systems for gene therapies beyond liver applications [139]. Approaches like grafting targeting moieties onto LNP surfaces or adjusting surface chemistry enhance tissue- or cell-specific delivery [139].

Systemic circulation and blood clearance

The fate of LNPs within the body is determined by their stability, circulation, interaction with the immune system, and clearance from the bloodstream, all of which are crucial for the success of gene therapy [140]. Some LNP formulations are ineffective in transporting genetic material due to their instability in biological fluids, highlighting the need to evaluate their stability and integrity under these conditions. For example, cholesterol-free LNPs containing pH-sensitive cationic lipids, egg sphingomyelin, and PEG-lipid demonstrate instability in blood, resulting in premature siRNA release [141]. The adsorption of proteins onto LNPs triggers coagulation processes and opsonin attachment, leading to their recognition and elimination by the MPS [142]. These events reduce NP delivery to target organs and decrease therapeutic effectiveness. Electrostatic attraction between positively charged LNPs and negatively charged proteins causes particle agglomeration, which shortens their circulation time and reduces cellular internalization. Additionally, the protein corona can compromise particle integrity, leading to premature cargo release and affecting biodistribution and circulation time [143]. Serum proteins can destabilize LNPs, necessitating thorough testing for agglomeration, premature RNA release, and dye leakage in biological fluids.

Furthermore, LNPs possess similar characteristics to viruses, including a lipid envelope and a comparable size and curvature. Viruses also exhibit a biomolecular corona that affects infectivity and DC activation [144], indicating an evolutionary role in evading the immune system. Consequently, a significant challenge in LNP delivery is evading the immune system [145]. NPs often face challenges in (pre)clinical trials due to strong immune responses after injection, with subsequent administrations potentially triggering an adaptive immune reaction.

Various strategies, including modifying LNP composition by adding cholesterol with high-phase transition temperature phospholipids, can reduce LNP clearance by altering protein adsorption on the NP surface [146]. PEGylation technology induces a “stealth effect” on NPs by hindering protein adsorption and opsonization. This enhances LNP circulation stability, reduces macrophage phagocytosis, and prolongs blood residency. Proteins like clusterin in the PEGylated NP corona further enhance the stealth effect. The efficacy of PEG-lipid depends on its quantity, molecular size, and surface density. Specifically, increasing the PEG-lipid content with longer chains (C18) on LNPs boosts circulation duration, while shorter chains (C14) do not have the same effect. At least 1.5 mol% PEG-lipid is necessary to alter LNP pharmacokinetics. Additionally, high PEG content reduces protein interactions, while the chain length influences the protein corona composition [147, 148]. Utilizing PEG and related modifications requires meticulous assessment due to potential limitations. The “PEG dilemma” refers to the reduced cellular internalization and endosomal release associated with PEGylation. Solutions to this issue involve using pH-sensitive, enzyme-cleavable, and detachable PEG-lipids. Research on de-PEGylation kinetics for siRNA delivery indicates that PEGs with short, saturated tails are essential for effective de-PEGylation and gene silencing, exemplified by Onpattro. Such PEGs can swap with proteins like ApoE post-administration, improving hepatocyte targeting and demonstrating the strategic use of corona formation for circulation control and tissue targeting [149].

Cellular uptake and intracellular trafficking

Non-viral vectors are typically internalized through endocytosis, utilizing various pathways such as clathrin-mediated, caveolae-mediated, and macropinocytosis [150]. Endocytosis efficiency is influenced by receptor types and densities, with receptors like LDL and ASGPR known for their high uptake capacities [151]. Once internalized, particles are transported to endosomes, undergo degradation, and are eventually directed to lysosomes upon acidification. Achieving endosomal escape is crucial for effective gene delivery, as some pathways can circumvent lysosomal degradation. Intracellular transfer mechanisms, independent of endocytosis, include membrane fusion, as observed in viruses like the Sendai virus and HVJ envelope vectors [152]. Fluorescence imaging is a suitable method for tracking the subcellular distribution of non-viral vectors, particularly in monolayer-cultured cells, enabling quantitative analysis and visualization [153]. However, challenges arise in three-dimensional cultures like spheroids due to low fluorescence detection, and assessing subcellular distribution in living organisms using tissue sections remains challenging [154].

Endosomal escape efficiency is pivotal for successful gene delivery. LNPs are typically trapped in endosomal compartments, with only a small fraction managing to escape [155]. Although the precise mechanism is not fully understood, positively charged lipids may enhance electrostatic interactions and fusion phenomena with negatively charged endosomal membranes, releasing mRNA molecules into the cytoplasm [156]. Jiang et al. [157] introduced a protein probe, ddRLuc, composed of a signal sequence fused to a human IgG1 Fc fragment. This probe binds cells via the Fc receptor. Once expressed in cells, the fusion protein (ddRLuc-Fc) undergoes glycosylation in the endoplasmic reticulum and is subsequently secreted. Activation of ddRLuc-Fc depends on deglycosylation mediated by the cytosolic enzyme, NGLY1, following cellular entry and endosomal release [158, 159]. Common strategies to enhance endosomal escape use pH-sensitive polymers, fusogenic lipids or peptides, and the design of more potent and safer ICLs. Endosomal escape can be optimized by adjusting the pKa of ionizable lipids [160]. Additionally, the properties of lipidic tails can influence the endosomal escape of LNPs [85]. For instance, lipids with branched tails demonstrate improved endosomal escape compared to those with linear tails due to stronger protonation at endosomal pH [161]. Moreover, the type of helper lipids and the ratio of lipids affect endosomal escape [162]. Among helper lipids, DOPE demonstrates high gene delivery efficiency in vitro by inducing fusion with the endosomal membrane. However, fusion with erythrocytes occurs in vivo, reducing efficiency [163]. PEI facilitates endosomal escape through the proton sponge effect, inducing disruption of endosomes due to chloride ion influx [164]. Additionally, fusogenic peptides like GALA aid in pH-sensitive endosomal release. Furthermore, inorganic salts such as calcium phosphate promote endosomal escape through acidification and dissolution [165]. Additionally, certain techniques can achieve efficient gene delivery without requiring endosomal escape by creating temporary pores in the cellular membrane using external forces. These techniques include electroporation, microbubbles, ultrasound application, and the hydrodynamic strategy involving rapid injection of a substantial volume of NA solution [166,167,168]. Following entry into the cytosol, LNP diffusion is hindered by cytoskeletal structures [169, 170]. In dividing cells, exogenous NA localizes in the nucleus during cell division when the nuclear membrane disintegrates. Conversely, non-dividing cells experience limited mass transport due to the intact nuclear membrane. Nuclear pores allow the diffusion of molecules up to 60–100 kDa, requiring NLS for larger materials [171]. The nuclear transport of RNA therapeutics is crucial for influencing intranuclear gene activity. Tammam et al. demonstrated that chitosan conjugated with an NLS peptide effectively transported 150 nm into the nucleus [172].

An ideal nanovector must tightly load mRNA before entering the cytoplasm for optimal protection and delivery, and then rapidly release the mRNA once in the cytoplasm. Various methods, such as FRET, gel electrophoresis, flow cytometry, and qPCR, can investigate the mechanisms and kinetics of mRNA dissociation from LNPs. These methods provide insights into the location, extent, and rate of mRNA release from LNPs and correlate with gene expression levels.

To achieve controlled unpacking of RNA-LNPs and to regulate in vivo behavior, stimuli-responsive nanovectors have been developed to respond to intracellular triggers such as pH, redox, ATP, and enzymes and external triggers like magnetic fields, light, and ultrasound [173]. These nanovectors can stably encapsulate RNA in physiological environments and non-target sites but release RNA cargo upon structural changes triggered by specific stimuli in target cells [174, 175] (Fig. 8). pH-responsive LNPs remain stable at physiological pH but become positively charged in acidic environments, enhancing gene release and endosomal escape. This adaptability aids in targeted delivery and reduces cytotoxicity [176]. For example, using a pH-catalyzed hydrolysis strategy, Tanaka et al. designed a self-degradable ionizable lipid (ssPalmO-Phe) containing disulfide bonds and phenyl esters. This strategy promotes LNP degradation and mRNA release through intra-particle enrichment of reactants, enhancing transfection efficiency [158]. In addition, DOP-DEDA, a novel charge-reversible lipid derivative, shows different charges at varying pH levels, ensuring stability at physiologic pH without PEG-lipids. DOP-DEDA LNPs effectively suppress PLK1 mRNA and protein expression and enhance siRNA delivery with apoE3 [177]. Additionally, Siegwart et al. studied PEGylated BODIPY dyes (PBDs) as surface-stabilizing agents for LNP formulations with pH-responsive mRNA release. PBD LNPs greatly enhanced cytoplasmic protein production and better mRNA delivery in the liver compared to PEG-DMG LNPs [178]. Furthermore, a novel pH-sensitive LNP called BAMPA-O16B has been developed to enhance siRNA delivery and endosomal escape in GBM cells. This LNP effectively delivered siRNA against CD47 and PD-L1, improving T cell-dependent antitumor immunity in mice [176].

Fig. 8

Illustration of endosomal escape and stimuli-responsive RNA release mechanisms; pH-, enzyme-, and redox-responsive RNA release can be achieved following LNP endocytosis

Higher redox potential in the cytoplasm compared to the extracellular environment can trigger RNA release. The higher GSH concentration in the cytoplasm facilitates the disruption of disulfide bonds within LNP structures incorporating the CR8C peptide, enabling improved gene therapy [165]. An ionizable lipid with a degradable linker (4A3-SCC-PH) and branched tails significantly enhanced mRNA transfection, showing a 15.5-fold improvement in FLuc mRNA delivery compared to MC3 LNP. This was due to the asymmetric alkyl chains attached to the thioether and the GSH-responsive disulfide bond, creating a conical shape that facilitated efficient mRNA delivery in malignant cells [179]. Kamath et al. reported a redox-responsive LNP loaded with TP53-coding mRNA for inducing enhanced apoptosis in TP53-deficient Hep3B HCC and H1299 NSCLC cells in vitro and in vivo [180]. Also, ROS is significantly higher in cancer cells than in normal cells, making ROS a specific stimulus for triggering mRNA release [181]. Similarly, boric acid ester [182,183,184] and thioketal linkages [185] can create ROS-sensitive LNPs for controlled gene release. Cai et al. synthesized an ionizable lipid, BAmP-TK-12, containing ROS-degradable thioketal moieties for mRNA delivery. High ROS concentrations in tumor cells induce the oxidation and degradation of TK-12 moieties in BAmP-TK-12-based LNPs, facilitating mRNA release and subsequent translation. This approach successfully delivered mRNA encoding RAS protease DUF5 for RAS mutant depletion, suppressing tumor growth in A549 tumor-bearing mice and demonstrating the potential of ROS-responsive LNPs for tumor-specific mRNA delivery [52, 186].

Enzyme-responsive nanocarriers can trigger gene release, carrier dissociation, and improved cellular uptake in tumor environments. For example, esterase-responsive cationic polymers can reverse their charge to release therapeutic RNA in the presence of intracellular esterase [187]. Similarly, matrix metalloproteinase-sensitive nanocarriers can release their payload specifically in tumor tissues, enhancing therapeutic efficacy [188].

ATP-responsive LNPs for RNA delivery leverage the high ATP levels in tumor cells to achieve targeted and efficient gene release. The phenylboronic acid group in these nanocarriers can form and break ester bonds with ATP, leading to nanoparticle dissociation and RNA release [189]. Other examples include phenylboronic acid-modified polymers and PEI crosslinked with alginate, which enhances transfection efficiency and gene silencing in tumor cells [190, 191].

Multi-stimulus responsive LNPs for RNA delivery are designed to respond to multiple stimuli, enhancing accuracy and specificity. For example, Gao et al. developed a pH/redox dual-responsive polyplex for delivering MDR1 siRNA and doxorubicin, which releases its payload in acidic and high GSH environments, demonstrating potent antitumor efficacy [192]. Similarly, Zhang et al. designed a dual-locking nanoparticle responsive to pH and H₂O₂ levels in tumor microenvironments to control CRISPR/Cas13a gene editing selectively [193].

External-stimuli responsive systems are also being explored for RNA delivery, although less attention has been given to LNP formulations. Magnetic field-responsive LNPs can accumulate at tumor sites when guided by an external magnetic field. Studies with magnetic mesoporous silica Nps and magnetic Fe₃O₄ NPs coated with polydopamine have shown increased tumor targeting and therapeutic efficacy [194, 195]. UV light-responsive systems like coumarin-anchored PAMAM dendrimers, release their payload upon UV exposure [196]. Mo et al. developed a LASER strategy using porphyrin-LNPs, which generate ROS when exposed to near-infrared light, enhancing membrane disruption and endosomal release of siRNA. This method improved siRNA escape and target knockdown in prostate cancer cells. In vivo, porphyrin-LNPs with gold NPs confirmed the LASER effect, enhancing RNA therapeutic delivery [197]. Far-red light-mediated nanocarriers, such as those developed by Wang et al., utilize photosensitizers to produce ROS, promoting endosomal escape and gene release [198]. Another example is photolabile spherical nucleic acids that disassemble and release siRNA upon NIR light-triggered ROS production [199]. Ultrasound-targeted gene delivery employs ultrasound to direct genes to target sites, enhancing cellular uptake by increasing membrane permeability. Genes are loaded onto ultrasonic NPs using cationic lipids like DPPC, DSPC, and DOTAP. For example, mannose-modified bubble lipoplexes showed efficient DNA vaccination under ultrasound, achieving potent antitumor effects [200]. Biosynthetic nanobubbles from Halobacterium NRC-1 also improved gene transfection efficiency [201]. Ultrasound-enhanced ROS nanocarriers decorated with sonosensitizers like IR780 facilitate gene release through ROS generation [202].

Biodegradability plays a vital role in reducing the toxicity of nanovectors, and LNPs composed of ionizable lipids that break down into non-toxic metabolites are favored. Introducing ester bonds increases lipid biodegradability, as shown for lipids like L319 [44], Lipid 5 [160], SM-102 [41], and ALC-0315 [203]. Notably, SM-102 and ALC-0315 are used in COVID-19 vaccines [204]. Combining ester and disulfide motifs further accelerates degradation through disulfide bond cleavage, driving intraparticle nucleophilic attack on ester linkers [158, 205, 206]. Whitehead et al. synthesized a large library of degradable lipidoids created via the conjugate addition of alkyl-amines to alkyl-acrylate tails. Lipidoids have been formulated with cholesterol, DSPC, PEG2000-DMG, and siRNA for LNP preparation. The researchers discovered that particular structural and pKa parameters can predict over 95% of protein silencing in vivo, indicating that these criteria may serve as alternatives to expensive cell culture assays and animal testing [207]. Other advances include pH-sensitive PEG detachment [208] and employing materials like black phosphorus [209] that degrade under specific conditions, releasing their payload effectively. The use of biodegradable components like hyaluronic acid [210], polydisulfides [211], thioketal-crosslinked PEI [198], PEI-modified nanobubbles [201], and responsive materials like boric acid esters [182] enhances the efficacy and safety of RNA delivery. Therefore, properly designed LNPs can ensure precise control of targeted gene delivery while maintaining biocompatibility and safety [212].

Selection of the administration route

The route of administration significantly impacts the distribution, kinetics, and efficacy of RNA-LNP [6]. Intravenous gene delivery without a carrier leads to rapid degradation in the blood and minimal gene expression in liver non-parenchymal cells [213]. In contrast, local injections, particularly into muscle tissue, result in traceable gene expression [214]. After intravenous injections, LNPs predominantly localize within various liver cell types, including hepatocytes, LSECs, and Kupffer cells [215]. LNPs primarily target the liver because of their physicochemical properties and anatomical and physiological features, such as high blood flow and discontinuous hepatic capillaries [216]. To broaden the therapeutic potential of LNPs, achieving biodistribution to organs beyond the liver is essential. Intra-arterial and intraportal injections can alter the first-pass effect, whereas intravenous injection primarily targets the lungs as the first-pass organ. Intraperitoneal delivery, commonly used in animal research rather than human studies, is associated with lower systemic toxicity than intravenous methods. Melamed et al. observed that this administration route, when applied to LNPs with cationic phospholipids, resulted in significant protein expression in the pancreas. This finding suggests potential new treatments for pancreatic diseases, including cancer and diabetes [217].

Local administration of LNPs, including intradermal, subcutaneous, and intramuscular injections, can elicit systemic immune responses by delivering mRNA to lymph nodes, where APCs and T cells are present [6]. Compared with intravenous injection, these routes often lead to prolonged protein expression at the injection site [41], potentially bypassing systemic circulation and offering sustained therapeutic effects. However, mRNA and/or lipid components may also disseminate to other tissues such as the lungs, liver, non-draining lymph nodes, and spleen [41]. LNPs employed in mRNA vaccines, including COVID-19 vaccines, predominantly generate antigen expression in the deltoid muscle and draining lymph nodes [218]. The intramuscular route allows for larger injection volumes compared with the intradermal and subcutaneous routes [38], whereas the intranodal route holds promise for naked mRNA delivery [219].

Intraocular delivery is an effective strategy for addressing ocular diseases, as conventional methods fail to reach the posterior region of the eye [198]. Techniques such as intravitreal and subretinal injections have proven successful in targeting the retina and retinal pigment epithelium with LNPs [220]. Research by Patel et al. demonstrated that ICLs with specific chemical features, such as low pKa and unsaturated tails, enhance gene expression in the retinal pigment epithelium following subretinal injection [221]. Similarly, LNPs designed with minimal PEGylation have shown increased expression in retinal cells after subretinal and intravitreal administration. Targeting the neural retina and photoreceptors, which are typically difficult to access, has been achieved through receptor-mediated uptake using LNPs modified with retina-specific peptides, as demonstrated by Herrera et al. This method holds significant potential for advancing treatments for hereditary retinal diseases [222].

Pulmonary routes such as intratracheal, intranasal, and aerosol inhalation offer direct lung-targeted delivery of therapeutics, reducing systemic adverse effects. This method circumvents the sequestration of mRNA-LNPs by Kupffer cells in the bloodstream [223]. Intratracheal administration of LNPs, whether unmodified, cationic, or ligand-tethered, is effective for siRNA delivery, with mannose-modified LNPs being particularly efficacious against pulmonary fibrosis [224]. Lokugamage et al. refined LNP formulations for pulmonary delivery through a nebulizer, termed NLD1, enriched with PEG-lipid and helper cationic lipids, achieving superior mRNA delivery to the lungs compared to LNPs designed for systemic circulation [69].

Oral and transdermal routes present significant challenges for LNP delivery due to physiological barriers. Microneedles offer a promising solution for delivering NAs across the skin [225]; however, oral delivery of LNP-NA formulations faces various hurdles in the gastrointestinal tract, including acidity, bile salts, digestive enzymes, mucus, and membrane permeability [226]. Various formulation strategies have been explored to overcome these challenges, with modifications aimed at enhancing stability, retention, and efficacy [227]. For instance, pH-responsive polymer coatings can prevent pepsin digestion and bile salt interaction, whereas lyophilization enables regular oral administration via solid dosage forms [227]. Orally available siRNAs have shown promise in preclinical and clinical studies, attributed to conjugation technologies like GalNAc conjugation and formulations containing intestinal permeation enhancers [228].

Overall, the physiological barriers to LNP delivery should be carefully considered to develop non-invasive administration methods, refine LNP-NA interactions, and enhance therapeutic effects.

LNP for tissue-specific gene delivery

LNPs exhibit promise as carriers for RNA therapies in living organisms, notably highlighted by the emergency authorization of mRNA COVID-19 vaccines. However, challenges persist, including achieving extrahepatic delivery due to liver accumulation via ApoE-LDL interaction and LNP clearance by the hepatic or renal systems or MPS [229]. Effective strategies for targeting specific organs and tissues are vital for expanding RNA-LNP applications. Recent research has concentrated on routes of administration and engineering functionalized LNPs to tackle these challenges [230]. Targeted drug delivery systems aim to precisely transport therapeutic agents to specific cells or tissues, improving treatment efficacy and minimizing mRNA-related side effects, particularly in genome editing. Controlling the distribution of LNPs is crucial for addressing conditions localized to specific tissues or organs. Active and passive targeting are two strategies used to improve the efficacy and specificity of therapeutic agents. Pathophysiological conditions can create unique environments for tissue-specific drug delivery. Passive targeting relies on the natural physiological and pathological differences between healthy and diseased tissues. Tumors often have leaky vasculature and poor lymphatic drainage, so the EPR effect allows NPs to accumulate more in tumor tissue than in normal tissue. In ischemic conditions, areas with reduced blood flow can be targeted as NPs can accumulate in these regions due to the disrupted blood flow. Active targeting involves the use of specific ligands that bind to receptors on diseased cells. Tumors often have abnormal vasculature, overexpressed receptors, and an acidic environment. These features can be targeted by designing functionalized and stimulus-responsive LNPs. Inflammatory conditions like rheumatoid arthritis or inflammatory bowel disease have overexpressed inflammatory markers that can be targeted with specific ligands conjugated with LNPs [231]. Other pathophysiological changes, including glutamate excitotoxicity, oxidative stress, mitochondrial dysfunction, cell death, driver gene mutation, angiogenesis, and aberrations in the cell cycle and signaling pathways elucidate potential targets for nanoparticle-based drug delivery systems [232, 233].

The following section elaborates on the latest advancements in LNP passive and active targeting approaches.

Passive targeting

Passive targeting in LNPs is mainly influenced by the adsorption of the protein corona, which dictates their biodistribution. Biodistribution is also determined by the LNPs’ physicochemical characteristics, including size, charge, morphology, composition, and ratio of components. LNPs interact with ApoE, after PEG detachment, facilitating passive delivery. Modulating particle size is one approach to achieving targeted organ delivery [234, 235]. LNPs under 10 nm in size are swiftly eliminated by renal clearance, predominantly gathering in the liver, and traversing into the lymphatic system. In contrast, larger LNPs tend to accumulate in the spleen and lungs or may persist at the administration site [234]. LNPs of intermediate sizes, around 78 nm, exhibit optimal delivery efficiency targeting the liver. Adjusting particle size via formulation parameters is essential for targeted delivery. LNPs incorporating pSar instead of PEG lipid showed elevated mRNA expression levels in the spleen, with a tendency for increased spleen accumulation as their size diminished [236]. The leaky nature of tumor vasculature allows for the passive targeting of NPs to tumors via the EPR effect [237]. However, the efficacy of the EPR effect is generally limited in human tumors [123]. Consequently, smaller NPs (specifically those under 50 nm) are considered more effective for human cancer treatment. Additionally, the LNP surface charge significantly impacts biodistribution and mRNA delivery efficiency. Positively charged LNPs target the lung, neutrally charged LNPs preferentially enter the liver, and negatively charged LNPs mostly accumulate in the spleen [238, 239].

LNP components, especially ICLs, which interact with specific serum proteins, are crucial for designing new delivery materials targeting organs. For instance, LNP multi-tailed iPhos have been developed to enhance endosomal escape and mRNA/sgRNA delivery for in vivo genome editing. These lipids can function together with variously charged helper lipids to allow organ-specific delivery [240]. Modifying the ICL linker structure impacts protein adsorption, consequently influencing mRNA delivery from the liver to the lungs [136]. LNPs having alkyne lipid tails can significantly increase membrane fusion to enhance mRNA release in the liver [53]. Lee et al. chemically synthesized 4A3-Cit using unsaturated thiols, resulting in an 18-fold improvement in transfection efficiency in vitro and significant liver targeting [85]. Lokugamage et al. synthesized a constrained adamantane cationic lipid, forming an armchair structure, capable of delivering siRNA to T cells in vivo [241]. The ionizable lipid OF-Deg-Lin with ester linkages generates a non-toxic fatty acid facilitated targeting of the spleen and transfection of B lymphocytes [239]. Qiu et al. utilized N-series LNPs containing an amide bond in the tail, selectively delivering mRNA to the mouse lung [136]. Ding et al. showed enhanced in vivo mRNA delivery by editing lipid amine-head groups, with spermine-derived head groups exhibiting the strongest cellular uptake, endosomal escape, and transfection potency of IL-12 siRNA in acute liver failure [242]. A study by Mitchell et al. demonstrated that suitable adjustments to the chemical structure of ionizable lipids by incorporating siloxane composites into the ionizable lipids can direct LNPs to specific tissues such as the liver, lungs, and spleen [243]. Zhang et al. demonstrated the effectiveness of functionalized N1,N3,N5-tris(2-aminoethyl)benzene-1,3,5-tricarboxamide derivatives (FTT) in promoting in vivo mRNA delivery. They found that FTT-lipids with branched ester chains, such as FTT2 to FTT6, facilitated efficient mRNA delivery to the liver and spleen compared with linear ester chain analogs like FTT7 to FTT10 [244]. In another study, Kowalski et al. demonstrated that amino polyesters synthesized from tertiary amino alcohol and lactone could more effectively express mRNA in the lung endothelium and liver hepatocytes, resulting in nearly tenfold higher systemic mRNA delivery efficacy [245].

Cholesterol modification is another option for passive targeting. Paunovska et al. formulated LNPs with esterified cholesterol, demonstrating enhanced efficiency in gene delivery to liver microenvironment cells compared to regular or oxidized cholesterol [64]. LNPs containing cholesteryl oleate efficiently delivered CRISPR sgRNAs to liver endothelial cells in vivo [246]. Jung et al. revealed LNPs with histidinamide-modified cholesterol, resulting in improved intracellular delivery of SARS-CoV-2 spike mRNA [247]. CHEMS, an anionic phospholipid, was combined with other lipids to target lymph nodes [234].

Passive targeting of LNPs can also be achieved through phospholipid and PEG-lipid modification. Gan et al. developed LNPs containing constrained phospholipids with adamantyl groups, enhancing targeting to liver endothelial cells and Kupffer cells compared to unconstrained phospholipids [248]. Replacement of phospholipids with 18:1 BMP (S, R), imparting a negative charge, resulted in spleen targeting [249]. Sago et al. altered the lipid tail of PEG-lipid, thereby modifying the tropism of LNPs to target BMECs [250]. Furthermore, lipid-polymer hybrid NPs have shown effectiveness in delivering RNA through the mucus-covered epithelium in the human airway, indicating their potential in inhaled siRNA therapy for CF [251].

The SORT approach entails incorporating an additional SORT molecule into standard LNP components, guiding them to precise anatomical sites based on their biophysical characteristics. The SORT molecule can manipulate the internal and/or external charge of formulated LNPs, enabling targeted gene delivery to the lung, spleen, or liver. For instance, adding higher percentages of permanently positively charged DOTAP alters tissue tropism from the liver to the lungs. Conversely, incorporating a negatively charged 18PA SORT molecule at 10–40% in an LNP formulation leads to spleen-specific Luc expression [128]. Figure 9 shows the factors affecting passive targeting.

Fig. 9

Factors affecting lipid nanoparticle (LNP) passive targeting: A Fluorescence and bioluminescence biodistribution patterns of 1,1′-dioctadecyl-3,3,3′,3′-tetramethylindotricarbocyanine iodide (DiR)-tagged Fluc-mRNA LNPs varying in size and the corresponding luciferase expression after intramuscular administration [235]; B biodistribution of Luc-LNP at various charge ratios in BALB/c mice 24 h after intravenous administration [238]; C incorporating an additional element, known as a selective organ targeting (SORT) molecule, into conventional LNPs modifies their distribution within the body and facilitates targeted delivery to specific tissues [128]

Active targeting

Active targeting utilizes specific ligands or molecules that bind to receptors or surface structures on target cells, enhancing the uptake of the delivery system [252]. Common ligands, including proteins, antibodies, peptides, carbohydrates, aptamers, and vitamins, can be incorporated into LNPs through non-covalent interactions or covalent bonds, primarily to PEG-lipids [253]. Generally, several conjugation strategies can be employed to attach targeting ligands to the surface of LNPs or lipid compounds. These strategies include direct covalent conjugation via carbodiimide, maleimide, or click chemistry, and non-covalent conjugation through electrostatic, hydrophobic, or hydrogen bonding interactions. Additionally, lipid anchor incorporation of lipopeptides or lipids modified with targeting moieties can be incorporated into the lipid bilayer during fabrication or post-insertion into preformed LNPs. Hybrid strategies combine multiple conjugation approaches to design ligand-targeted mRNA-LNPs with improved therapeutic properties. These ligands facilitate interactions with target cell surface molecules, such as ligand-receptor or Ab-antigen pairs, promoting retention at the target site and NP uptake by target cells [254].

Sugar modification has emerged as a valuable tool for targeting specific cell types. Galactose and GalNAc target liver parenchyma cells, whereas mannose modification targets non-parenchyma cells and APCs [255]. Ligands like GalNAc and mannose have been coupled to PEG lipids for efficient liver and hepatic sinusoidal endothelial cell targeting [256]. Combinations like mannosylation of bubble lipoplexes with ultrasound irradiation have shown effective delivery of NF-κB decoy to TAMs, enhancing survival in animal models [257]. LNPs functionalized with HA offer promise for active targeting of CD44-positive cells in various cancers and certain types of inflammatory and immune cells [258].

Conjugation of antibodies, aptamers, or peptides to LNPs enhances targeted delivery [259, 260] (Table 2). For example, Rurik et al. showed that modifying LNP surfaces with T cell-targeting antibodies enables mRNA delivery for specific targeting of cardiac fibroblasts, a novel approach for heart failure treatment [261]. LNPs functionalized with CD38 mAb specifically target lymphoma cells [262], whereas LNPs modified with CD5 antibodies deliver mRNA to CAR-T cells for heart disease treatment [261]. Coupling LNPs with CD31 antibodies achieves targeted therapy of pulmonary endothelial cells [263]. Lipoprotein incorporation into siRNA-loaded LNPs creates ASSET, facilitating interaction with targeting antibodies. Arginine-rich cell-penetrating peptides, combined with targeting ligands, enhance gene delivery [264]. Kwong et al. engineered APNs that target CD8⁺ T cells, facilitating the in vivo delivery of mRNA to antigen-specific T cells [265]. Ligands like RGD peptide and HER-2-targeting peptide show promise for active tumor targeting [266, 267]. LNPs conjugated with ligands specific to T cell markers (CD3, CD4, CD5, and CD8) enhance the targeting efficiency. In addition, functionalized polymers or polymerized peptides with GALA repeats, which provide fusogenicity to APCs, have been explored as non-liver RNA-delivery therapeutics [268]. Furthermore, the CH6 aptamer enables targeted gene delivery to osteoblasts through LNPs [269].

Table 2 Active targeting strategies for LNP-RNA therapeutics

Exploring small-molecule ligands has diversified targeted delivery with LNPs. Folate-functionalized LNP shows promise for brain and breast cancer therapy [296]. LNPs modified with phenylboronic acid target cancer cells by interacting with surface sialic acid residues, enabling selective binding and uptake [297]. LNPs modified with imidazole target splenic T lymphocytes for mRNA delivery, facilitating T cell immunotherapy [298]. LNPs equipped with piperazine-derived lipid target immune cells in the liver and spleen [299]. LNPs modified with heterocyclic rings in their lipid head groups enhance mRNA delivery and IFN-γ secretion, subsequently activating the cGAS-STING type I IFN signaling pathway [300]. LNPs functionalized with alendronate offer targeted delivery to bone cells via calcium chelation [297]. Vitamin-conjugated LNPs deliver antimicrobial peptides and mRNA-encoding cathepsin B to macrophages [301]. BP lipid-like compounds demonstrated exceptional targeting of the bone microenvironment and efficiency in mRNA delivery [280]. An LNP targeting secondary lymphoid tissues has been developed with novel phosphatidylserine derivatives, enhancing mRNA delivery to the spleen. The phosphatidylserine-LNP showed increased uptake by macrophages in the spleen’s red pulp and marginal zone [302]. In another study, anisamide-tethered lipidoid (AA-T3A-C12) was employed for the fabrication of siRNA-LNP to activate HSCs for silencing HSP47 in liver fibrosis [295]. Figure 10 depicts LNP modifications with targeting moieties for active targeting.

Fig. 10

A Schematic representation of common targeting ligands used for specific LNP-RNA delivery; B examples of recent LNP surface modification for active targeting: 1 bisphosphonate (BP) lipid-like component for mRNA delivery to the bone microenvironment: a illustration depicting the construction of BP-LNPs for mRNA delivery to the bone environment, leveraging the interaction of BP-LNPs with Ca²⁺; b Cryo-TEM image showing BP-LNP morphology (100 nm scale bar) and ex-vivo imaging of bones post-LNP delivery. LNPs encapsulate luciferase (Luc) mRNA and are labeled with 1′-dioctadecyl-3,3,3′,3′-tetramethylindotricarbocyanine iodide (DiR) dye [280]; 2 phosphatidylserine-loaded LNP for delivering mRNA to secondary lymphoid tissues: a illustration outlining the preparation of phosphatidylserine LNPs and strategies for mRNA delivery to macrophages in secondary lymphoid organs; b images of Luc activity in tissues from mice treated with subcutaneous and intravenous injections of phosphatidylserine-LNPs encapsulating Luc mRNA [302]; 3 ligand-tethered lipid nanoparticles for RNA delivery to treat liver fibrosis: a fabrication of AA-T3A-C12/siHSP47 LNP by microfluidic mixing and specific delivery to activated hepatic stellate cells (HSCs) for silencing heat shock protein 47 (HSP47) in liver fibrosis; b immunofluorescence staining of HSP47 in LNP-treated activated 3T3 fibroblasts [295]

While active targeting offers advantages, challenges such as variability in ligand-receptor specificity among patients, biomolecular corona interference, and complex formulation should be addressed for efficient delivery. Continued research into effective ligand-receptor combinations is essential for optimizing this strategy. In addition, combining multiple targeting ligands on LNP surfaces can enhance specificity and binding affinity.

Current clinical applications of RNA-LNPs

Initially considered impractical because of instability and immunogenicity, mRNA has become a focal point in therapeutics because of advancements in stabilizing modifications and delivery systems. This has led to its potential to replace plasmid DNA or recombinant proteins [303]. Moreover, the widespread success of mRNA-LNP vaccines for COVID-19 has catalyzed further exploration of mRNA therapeutics, including protein replacement therapy, regenerative medicine, vaccines for infectious diseases, anticancer therapies, and organ-specific disease targeting (e.g., CF, chronic obstructive pulmonary disease, or asthma), as well as anti-inflammatory and gene editing applications (Table 3). Beyond therapeutic use, RNA-LNPs have found utility in diverse bioengineering realms such as medical imaging, bioprinting, and basic scientific research.

Table 3 Application of LNP-RNA therapeutics based on the target disease

Protein replacement therapy

Protein replacement therapy, aimed at generating missing or defective proteins, has traditionally relied on recombinant proteins. However, mRNA encoding specific proteins or antibodies has emerged as a promising alternative, offering advantages such as rapid production, flexible design, and lower risk of contamination [324]. This approach has garnered interest across various fields, including hematology, metabolism, cardiology, pulmonology, neurology, and oncology. For instance, mRNA-loaded LNPs have been explored for treating hemophilia, a genetic bleeding disorder, by encoding missing clotting factors [325]. LNPs carrying mRNA for factor VIII or factor IX variants have shown promising results in mouse models, rapidly inducing and sustaining therapeutic protein levels [326]. Similarly, metabolic disorders like hepatorenal tyrosinemia and alpha 1-antitrypsin deficiency have been targeted using LNPs encapsulating mRNA for deficient enzymes, leading to improved organ function and increased protein levels [327]. Moreover, mRNA-LNPs have been investigated for lysosomal storage diseases such as Fabry disease, where high levels of therapeutic proteins were achieved systemically, demonstrating the potential for disease management [328]. Additionally, LNPs for delivering mRNA-encoding enzymes have shown efficacy in acute intermittent porphyria, methylmalonic acidemia, and glycogen storage disease, significantly reducing disease symptoms and improving survival rates [311].

Beyond liver-related diseases, mRNA-LNPs have been explored as therapeutic proteins. For instance, LNPs carrying mRNA encoding vascular endothelial growth factor have been studied for cardiac repair after myocardial infarction and for improving placental function in pre-eclampsia and fetal growth restriction [329]. Furthermore, LNPs for delivering mRNA encoding human frataxin have shown potential for treating Friedreich’s ataxia, a neurological disorder [330]. Moreover, mRNA-LNPs have been investigated for lung diseases through nasal administration, showing promise in restoring lung function in conditions like CF [331]. Davies et al. demonstrated that subcutaneous administration of mRNA within LNPs can effectively deliver therapeutic proteins, but it causes inflammatory responses. The integration of anti-inflammatory steroid prodrugs within LNPs has led to a reduction in inflammation and an extension of protein expression duration, thereby rendering this approach appropriate for chronic treatment regimens and self-administration [332]. DeRosa et al. demonstrated that delivering mRNA encoding human α-galactosidase protein via LNPs significantly increased serum GLA protein levels in mice and non-human primates. This method shows promise as a new treatment approach for Fabry disease, potentially surpassing current therapies [328]. Overall, the versatility and efficacy of mRNA-LNPs for protein replacement therapy offer promising avenues for treating several genetic and acquired diseases, representing a significant advancement in medical therapeutics.

Genome editing

Gene therapy, which encompasses genome editing, seeks to deliver mRNA to specific targets for gene modulation, including silencing or expression. CRISPR/Cas9 has gained prominence for gene editing due to its precision and versatility [333]. However, challenges in delivering pDNA or the Cas9 protein/sgRNA complex have prompted the exploration of mRNA-LNP platforms for efficient and safe genome editing [334]. CRISPR/Cas9-mediated gene editing using LNPs has shown feasibility and efficacy, particularly in liver-related disorders [335]. For instance, hypercholesterolemia, characterized by elevated LDL levels, has been targeted by reducing PCSK9 expression. Studies have demonstrated significant reductions in serum PCSK9 and cholesterol levels following systemic delivery of Cas9 mRNA and sgRNA using LNPs [333]. LNPs have also been employed for base editing, achieving efficient PCSK9 knockdown and LDL reduction in animal models [336]. Furthermore, LNPs have been used to target other genes implicated in various diseases. For example, Angptl3 and transthyretin have been targeted to reduce serum protein levels associated with cardiovascular and hTTR amyloidosis diseases, respectively [337]. CRISPR/Cas9 LNPs have also shown promise in addressing hemophilia and metabolic disorders like PKU and tyrosinemia type 1 [338]. Beyond liver-related conditions, LNP-mediated CRISPR/Cas9 delivery has been explored in extrahepatic organs. Intramuscular administration of LNPs has been investigated for Duchenne muscular dystrophy treatment, whereas LNPs targeting PLK1 and EGFR have shown efficacy against glioblastoma and ovarian cancer cells, respectively [339]. Im et al. developed ionizable LNPs to effectively deliver CRISPR/Cas9 ribonucleoproteins (RNPs), achieving high-efficiency gene editing in vitro and in vivo, potentially addressing diseases like cancer and genetic disorders [340]. Herrera-Barrera formulated an array of enhanced LNP (eLNPs) to deliver prime editors (PEs) RNA efficiently for gene editing, achieving a 54% prime editing rate. eLNPs led to improved endosomal escape, eventually causing onset of editing within 9 h and reaching maximum efficiency after 24 h. This technology showcased potential for new therapies targeting various disease-causing mutations [341]. Overall, LNP-mediated CRISPR/Cas9 gene editing holds great potential for treating various diseases, offering a promising avenue for precise and efficient therapeutic interventions.

Cancer

Recent investigations into cancer have highlighted the therapeutic potential of mRNA-based LNPs for various cancer types, including melanoma, lymphoma, hepatocellular carcinoma, colon cancer, breast cancer, and prostate cancer [342]. In non-Hodgkin’s lymphoma, targeting lymphocytes with mRNA delivery presents a rational approach, with the spleen and lymphoid glands serving as potential delivery sites [343]. mRNA-based LNPs have shown promise in melanoma, leveraging specific tumor antigens like human MART1 and LAMP1 to induce robust antitumor immune responses and enhance survival rates in murine models [344]. In addition, cationic liposomes carrying Bax-mRNA demonstrated superior antitumor effects compared with Bax-plasmid cationic liposomes, both in vitro and in vivo [345]. Modified OVA mRNA-lipid formulations stimulate CD8⁺ T-cell responses against melanoma models, with specific lipid characteristics (e.g., addition of sodium lauryl sulfate) influencing clonal expansion and function of CD8⁺ T-cells [346]. In Ai14D reporter mice, delivery of Cre-recombinase mRNA via optimized LNPs resulted in distinct protein expression rates in APCs. In addition, comparing unmodified and modified OVA mRNA revealed enhanced CD8⁺ T cell production, particularly through IFN induction. Moreover, OVA mRNA-modified liposomes containing galactosyl ceramide as an immune adjuvant significantly reduced tumor growth in the B16-OVA melanoma model, albeit with a moderate survival rate [282]. For hepatocellular carcinoma, LNP-mediated mRNA delivery has shown efficacy in reducing tumor size and increasing survival in a transgenic mouse model with an overexpressed MYC oncogene, with low toxicity in normal cells [347]. Modifications of LNPs have been explored in colon cancer, where protamine complexation enhanced mRNA condensation, lowered enzymatic degradation, and improved delivery efficiency [348]. In breast cancer, mRNA delivery systems have shown superior pharmacokinetic profiles compared with synthetic antibodies that results in increased serum levels of therapeutic agents and improved tumor suppression with minimal adverse effects [349]. For instance, compared with synthetic trastuzumab (Herceptin), in vivo PK evaluation of the mRNA-based system in C57BL/6 mice showed significantly higher serum trastuzumab levels, despite lower dosages. In addition, HER2⁺ mice exhibited higher morbidity-free survival, reduced average tumor volume, and no significant toxicity with trastuzumab mRNA compared to Herceptin [321]. In prostate cancer, delivery of modified PTEN mRNA via PEG-coated polymer-lipid hybrid NPs has demonstrated promising results in increasing tumor suppression potential, inducing apoptosis, inhibiting cell viability, and suppressing the PI3K pathway [350]. Silva et al. compared three different mRNA vaccine modalities to target HPV-16-related tumors in mice. These vaccines, encapsulated in LNPs, showed strong activation of E7-specific CD8⁺ T cells, prevented tumor relapses, and eradicated subcutaneous tumors. All three mRNA vaccines were superior to DNA and protein vaccines, indicating their promising potential for further clinical trials [351]. Chen developed an endogenous lymph node-targeting LNP named 113-O12B for mRNA cancer vaccines. The LNP increased CD8⁺ T cell response and showed excellent tumor inhibition, with long-term immune memory in treated mice [54]. These findings underscore the potential of mRNA-based therapies across a spectrum of cancer types, offering new avenues for effective and targeted cancer treatment.

Infectious diseases

mRNA vaccines encapsulated in LNPs have emerged as a groundbreaking technology for combating infectious diseases. Their appeal lies in several advantages, including ease of manufacturing, acceptable immunogenicity, and excellent safety profile [352]. This innovation represents a significant leap forward in vaccine development, particularly for diseases characterized by high genetic instability and infectivity. The development of mRNA-based vaccines traces back to 1989 when they were first explored as a therapeutic approach [353]. mRNA encapsulation within LNPs controls biodistribution and targeting specific cells. Notably, in response to the COVID-19 pandemic, mRNA vaccines developed with LNPs entered clinical trials, with mRNA-1273 (Moderna vaccine) and BNT162b2 (Pfizer/BioNTech vaccine) receiving emergency use authorization from the FDA in 2020 [354]. Beyond COVID-19, mRNA-LNP vaccines hold promise for various infectious diseases. Candidates targeting viruses such as HIV, seasonal influenza, Zika virus, RSV, and EBV are undergoing preclinical and clinical evaluation [355]. Studies have demonstrated the effectiveness of mRNA-based influenza vaccines in eliciting protective immune responses against diverse influenza virus strains [356].

In addition to vaccination, mRNA therapeutics can address infectious diseases. For example, mRNA encoding anti-HIV antibodies delivered via LNPs has shown promise for passive immunotherapy against HIV-1 [357]. The efficacy of mRNA delivery systems is influenced by factors such as the type of delivery system and administration route, which affect biodistribution. Studies have shown that mRNA-LNP favors spleen accumulation after intravenous administration [358]. In addition, biodistribution studies in mice have revealed distinct expression patterns following various administration routes, highlighting the importance of route selection in optimizing efficacy [41]. Li et al. developed a multiply adjuvanted mRNA vaccine using LNPs for optimized mRNA delivery and enhanced immune responses. In mice, this vaccine increased antibody titers against SARS-CoV-2 tenfold compared to unadjuvanted vaccines, indicating a potential for improved efficacy and safety in mRNA-based immunization [359]. SAM vaccine delivered by LNPs demonstrated a robust immune response and favorable safety profile in preclinical models, neutralizing multiple SARS-CoV-2 variants and reducing viral load. It has progressed to phase 1 clinical trials [360]. Clinical studies investigating mRNA-LNP-based vaccines for both cancer and infectious diseases are underway, further underscoring the potential of this innovative approach in advancing preventive and therapeutic interventions.

Inflammation

LNPs can trigger an immune response by interacting with receptors of the innate immune system, such as PRRs or TLRs, potentially diminishing therapeutic efficacy [361]. Incorporating anti-inflammatory agents like dexamethasone into LNPs can mitigate inflammation while maintaining therapeutic effectiveness. Zhang et al. demonstrated that the co-delivery of dexamethasone with mRNA-LNPs suppressed local inflammation while enhancing hepatic mRNA expression [361]. LNPs containing dexamethasone at a molar ratio of 9:1 with cholesterol (referred to as C9D1 LNP) showed reduced serum TNF levels compared with dexamethasone-free LNPs, without compromising transfection efficiency or inducing cytotoxicity. Davies et al. incorporated hydrophobic aliphatic ester prodrugs, including anti-inflammatory compounds like rofleponide and budesonide, within LNPs to mitigate innate immune responses [332]. Rofleponide-C14 prodrug, predominantly distributed in the outer shell of LNPs, significantly reduced local and systemic inflammation when included in mRNA-LNPs encoding a model-secreted protein (hFGF21). Similarly, budesonide conjugated with longer chain length esters (C16) improved local tolerability compared with shorter chain lengths. These findings indicate the potential of LNPs to suppress inflammatory responses. In addition to the direct incorporation of corticosteroids and chemically modified substances, modifications to ionizable lipids or optimized mRNA sequences have been explored to reduce anti-inflammatory reactions [362]. Kawase et al. showed that LNPs loaded with Irf5 siRNA effectively downregulated inflammatory factors (TNF and IL-6) in macrophages using biodegradable ionized lipid L120 [363]. Verma et al. engineered gemini LNP (GLNPs) to deliver siRNA effectively to colon cells and mitigate gut inflammation by reducing TNF-α expression. These GLNPs demonstrated stability, reduced toxicity, and inhibited key immune cell infiltration, highlighting their potential for next-generation nucleic acid delivery in treating inflammatory bowel disease [364].

Regenerative medicine

LNPs have versatile applications beyond therapeutics, spanning bioengineering fields such as medical imaging [73], bioreactors [365], and bioprinting with 3D reconstructed scaffolds [366]. LNPs play a pivotal role in life science studies [367] and intracellular trafficking [368]. For instance, bio-printed nanofibers facilitate the design and validation of optimal RNA-LNP vaccines, thereby enhancing gene editing efficiency [369]. LNPs are also employed in phototherapy and chemiluminescence imaging, particularly in conjunction with quantum dots [370]. In tissue engineering and regenerative medicine, LNPs enable spatially controlled payload release in both 2D and 3D tissue models, addressing challenges in programable cell-to-cell signal transport [366]. Yang et al. integrated LNP-mediated gene editing complexes into a Caco-2/HT29-MTX 3D tissue-engineered system to validate and screen gene-editing functions [371]. Similarly, Chen et al. used LNPs loaded with siRNA in a cartilage scaffold composed of GelMA hydrogel, chondrocytes, and adipose-derived stem cells to suppress microvascular infiltration and promote new cartilage formation in mice [372]. In another study, LNP was utilized in the determination of NIH-3T3 fibroblast growth on a polylactic acid-based 3D-printed human auricular model [373]. LNPs containing cyclosporine A and coenzyme Q10 were also used to evaluate whether cell proliferation on scaffolds could be induced. Rizvi et al. developed an efficient, safe method to transiently express growth factors in hepatocytes using mRNA-LNP delivery, promoting liver regeneration. In mice, this approach reversed liver damage and accelerated restoration, offering a promising therapeutic intervention for conditions like non-alcoholic fatty liver disease [109]. Therefore, LNPs are a valuable toolbox for bioengineering and fundamental life science research.

Current challenges and strategies for in vivo RNA-LNP delivery

Safety concerns

RNA-based therapeutics are generally well-tolerated but can cause adverse effects, including inflammation, liver and kidney toxicity, and thrombocytopenia, as observed with free siRNAs [374]. The safety of mRNA as a therapeutic or vaccine modality has been the subject of extensive research and scrutiny. Naked mRNA can induce innate immune responses upon delivery into cells. The innate immune system recognizes foreign RNA molecules, such as viral RNA or exogenous mRNA, through PRRs. This recognition may trigger inflammatory cascades and the production of cytokines and chemokines. Unmodified mRNA molecules may have off-target effects, leading to unintended gene expression or interference with endogenous cellular processes. These off-target effects can contribute to toxicity or alter the normal function of cells. Modified mRNA, which incorporates various chemical modifications, enhances stability and translational efficiency and reduces immunogenicity.

Non-viral delivery systems are recognized as safer alternatives to viral vectors, which carry risks such as producing neutralizing antibodies and the potential for hepatitis and leukemia. However, non-viral systems still raise safety concerns [375]. mRNA delivery theoretically avoids oncogenesis by viral vectors and plasmid DNA transfer by electroporation [376]. In the context of non-viral vectors, cytotoxic effects are often observed in cell culture studies, possibly due to the polycationic characteristics of the vectors. For instance, cationic liposomes have been shown to induce apoptosis [377]. Additionally, issues such as immunogenicity, hemagglutination, and inflammatory conditions like hepatitis pose challenges in live organisms [378]. Cationic vectors can induce capillary embolization by aggregating erythrocytes [379]. The likelihood of aggregation is influenced by the ratio of pseudo blood to cationic vectors, with a 1:1 ratio typically resulting in agglutination, reflecting the post-dosage period, while ratios exceeding 10:1 are less prone to aggregation [380]. PEGylation inhibits agglutination even at a 1:1 ratio [116]. Additionally, cationic non-viral vectors can elicit innate immune responses upon administration, leading to localized inflammation and tissue damage at the administration site [381].

ICLs, crucial for the efficacy of LNPs in delivering RNAs, can induce immune responses and cause toxicity by activating inflammatory pathways [382]. Their metabolites can induce inflammation by activating PPARs and cause liver damage [383]. PEGylated lipids can provoke immune responses, leading to the production of anti-PEG antibodies [383, 384]. Except for lipids, lysosomal cysteine proteases are released into the cytosol after LNP endocytosis, causing inflammation due to the activation of the NLRP3 inflammasome and release of pro-inflammatory cytokines [385]. Cathepsins also lead to cellular toxicity by promoting cell death and necroptosis by activating apoptosis pathways and rupturing the plasma membrane [386]. To mitigate toxicity, it is possible to substitute ICLs with biodegradable materials such as trehalose glycolipids or to alter the lipid head by incorporating 6,6′-trehalose dimycolate. This modification allows the LNPs to break down into non-toxic byproducts following the successful delivery of their cargo [387]. Another solution is to replace PEGylated lipids with pSar, which improves safety and maintains effectiveness [236].

Vaccine reactogenicity, a physical manifestation of the inflammatory response, includes local side effects like swelling and pain, and systemic side effects like fatigue, fever, and anaphylaxis, especially prevalent after mRNA vaccines [388, 389]. PEGylated lipids and NK cells contribute to these responses, which can sometimes lead to severe adverse events [389, 390]. Solutions to reduce reactogenicity include using a needle-free, pyro-drive liquid jet injector (PYRO) to deliver naked mRNA directly to APCs ensuring localized mRNA distribution and reducing systemic inflammation [391]. Another approach is using noncationic thiourea LNPs, which avoid high cationic charge density, demonstrating higher transfection efficiency and fewer inflammatory effects [392].

LNPs can activate the immune system through TLRs, NLRP3 inflammasome, IL-6 receptor, and MyD88, leading to beneficial immune responses but also potential adverse effects like severe allergic reactions and autoimmune phenomena [393, 394]; so, comprehensive long-term studies, careful preclinical evaluation, and clinical monitoring of immune responses are essential to ensure safety and efficacy. Strategies to enhance immunogenicity for vaccines include using adjuvant lipidoids and modifying lipid formulations with intrinsic adjuvant properties, such as incorporating TLR agonists [359, 395, 396]. Conversely, to reduce immunogenicity for repeated administrations, researchers suggest adjusting LNP compositions, using cleavable PEG variants or biodegradable polymers, and optimizing nanoparticle size and surface charge [397, 398]. For instance, smaller, neutrally charged particles exhibit reduced immunogenicity and superior lymph node targeting [398]. Additionally, incorporating immunomodulatory agents like dexamethasone and modifying vector surfaces can help minimize interactions with host immune cells, promoting immune tolerance [399]. Alternative administration routes such as IV and IN can refine immunogenic profiles [400]. More details about the immunogenicity of LNP-RNA therapeutics are mentioned in the following section.

Immunogenicity

Immunogenicity poses a significant challenge for LNPs in RNA therapy, affecting their therapeutic efficacy. Repeated administration may activate T and B cells, prompting the release of antibodies and cytokines, potentially reducing therapeutic efficacy. T cell activation can induce inflammation through proinflammatory cytokines. B cell activation may result in antibodies that recognize and neutralize LNPs, potentially compromising their function and eliciting adverse reactions. Chronic diseases requiring prolonged treatment could intensify these issues, potentially leading to CARPA [401]. The following sections address the immune recognition of RNA therapeutics and LNPs, highlighting the importance of meticulous monitoring and regulation of LNP immunogenicity through their structural, physicochemical, and biological attributes to ensure clinical effectiveness and safety.

Immune recognition of RNA therapeutics

Synthetic mRNA is sensed by innate immune cells through three major categories: (1) uridine-based RNA identification [402], (2) dsRNA recognition [403], and (3) sensing of mRNA’s uncapped 5′ terminus [404]. Vaccines using nucleoside-modified mRNA, notably those with Ψ or m1Ψ, have achieved greater clinical use and success, attributed to enhanced protein synthesis and reduced detection by RNA-specific intracellular sensors [405]. As shown in Fig. 11, innate immune cells detect RNA through endosomal PRRs (like TLRs and NLRs) and cytoplasmic RLRs. TLR7, which attaches to guanosine and UUU, and TLR8, which binds uridine and UG, are key sensors for unmodified ssRNA [402, 406]. TLR3 is responsive to dsRNA, a byproduct of mRNA. RLRs, including RIG-I, MDA-5, and LGP2, identify dsRNA based on structure and length. NLRP3 and NOD2 are specific for dsRNA and ssRNA, respectively. Introducing Ψ modifications can mitigate the response of dsRNA sensors like TLR3, RIG-I, MDA5, PKR, and OAS, thereby reducing the release of pro-inflammatory cytokines such as IL-1, IL-6, TNF, and type I IFN α/β [407]. Additionally, methods to remove dsRNA contaminants from unmodified IVT mRNA, such as HPLC purification [408], cellulose adsorption [409], and RNase III digestion [409] or adjusting reaction conditions, are crucial for enhancing protein production and minimizing immune responses. Modified nucleosides like Ψ, m¹Ψ, and m⁵C can decrease the formation of dsRNA byproducts in mRNA synthesized using T7 RNA polymerase [407]. Lowering the Mg²⁺ concentration and using a thermostable T7 RNA polymerase during IVT can reduce dsRNA formation [410]. The choice of capping strategy also impacts mRNA performance, with cap1 structures preferred to avoid recognition by innate immune sensors like RIG-I [411]. These strategies minimize innate immune activation and enhance protein expression via mRNA therapeutics.

Fig. 11

Pattern recognition receptors (PRRs) identify both mRNA and lipid nanoparticles (LNPs) extracellularly and intracellularly via toll-like receptors (TLRs), RIG-I-like receptors (RLRs), and NOD-like receptors (NLRs)

Immune recognition of LNPs

Both the innate and adaptive immune systems play crucial roles in detecting LNPs through various components, including (1) the entire particle, (2) individual lipids such as ICLs and cholesterol [412,413,414], and (3) modified or metabolized lipid products such as oxidative impurities of ICLs or oxidized phospholipids [415, 416]. For instance, single-cell RNA-sequencing analysis of draining lymph nodes from BNT162b2-vaccinated mice revealed spiked mRNA predominantly in myeloid DCs, monocytes, and macrophage clusters [417]. The following immune cells play an important role in LNP recognition:

• (a) Neutrophils, vital in acute inflammation and tissue entry, respond differently to cationic and neutral lipid NPs. Cationic variants induce LDH release, a cytotoxicity marker, and degranulate neutrophils, triggering elastase and superoxide anion release. They also stimulate NETosis, an activation marker, worsening inflammation.

• (b) Macrophages, APCs originating from fetal myeloid progenitors or circulating monocytes, migrate to inflamed tissue. Resident classical macrophages are professional phagocytes expressing PRRs, such as PAMP and DAMP receptors. NPs are recognized as ligands for NAMP receptors, and their interaction with PRRs induces inflammatory mediators. The route of administration, as well as the size and charge of NPs, affect their clearance by macrophages, with larger NPs being phagocytosed by liver Kupffer cells and splenic red pulp macrophages. In addition, cationic LNPs can induce inflammatory cytokines from macrophages. Strategies to mitigate LNP immunotoxicity on macrophages include targeting NPs to specific subsets of macrophages to reduce inflammation by avoiding cationic lipids or ICLs.

• (c) DCs, professional APCs found on epithelial surfaces such as the skin (Langerhans cells) and mucosal surfaces of the gastrointestinal, respiratory, and urogenital tracts, present antigens to naive T cells in lymph nodes, eliciting T cell responses. DCs are among the first immune cells to encounter NPs administered through intradermal, subcutaneous, and intra-lymphatic routes. The immune response induced by mRNA-LNP depends on the PRR pathways and DCs involved. Targeting NPs without cationic lipids or ICLs to a subset of DCs could reduce inflammation caused by mRNA-LNP.

• (d) The complement system, comprising over 30 proteins synthesized by the liver, plays a vital role in innate immunity by cooperating with the adaptive immune system. The system encompasses three pathways: classical, alternative, and lectin, all culminating in the splitting of C3 into C3b (phagocytosis stimulant) and C3a (anaphylatoxins recruiting phagocytes and triggering CARPA). Therefore, intravenous NPs must avoid C3b interaction and alternative pathway activation. Sabnis et al. developed a novel amino lipid series for mRNA-LNPs, demonstrating their tolerability of up to 1 mg/kg without inducing complement activation in rats and primates [160].

• (e) The adaptive immune system recognizes antigens through specific T- or B-cell receptors. Antigens are presented to T cells via MHC molecules by APCs such as DCs, macrophages, and B cells [418]. Antigen identification by T and B lymphocytes initiates the clonal proliferation of active and memory cells, which may result in hypersensitivity responses. Figure 12 represents the recognition of mRNA-LNP by innate and adaptive immune cells and their activation mechanism leading to immunogenicity.

Fig. 12

Adopted from [419]

Activation of innate and adaptive immune cells upon mRNA-LNP administration: A intramuscular injection of mRNA-LNP vaccines leads to local inflammation, recruiting neutrophils, monocytes, and dendritic cells (DCs) from the bloodstream to the site of injection through the release of chemokines and cytokines, facilitating the migration of other immune cells; B mRNA-LNPs alone, or along with antigen-presenting cells (APCs), are localized in the nearby lymph nodes; C DCs, monocytes or macrophages present antigens and initiate the activation of T cells; D Tfh cells assist B cells during germinal center (GC) reactions, alongside follicular DCs, to refine antibody affinity. In mouse models, LNP-induced IL-6 is crucial for developing T follicular helper (Tfh) and GC B cell responses, while type I IFNs are known to enhance cytotoxic T lymphocyte (CTL) reactions

LNPs can induce immunogenicity by binding to biological proteins and forming corona proteins like ApoE and complement proteins. These proteins activate T and B lymphocytes through proinflammatory cytokines such as IL-1, TNF, and IL-12 [420]. Lipid components in LNPs can unexpectedly amplify IL-1 to enhance TLR activation, whereas cationic lipid compositions are recognized by extracellular TLR2 and TLR4, as well as intracellular NLRP3 and STING [421]. To address these challenges, ICLs have been developed for mRNA delivery to reduce positive charge and mitigate immune and inflammatory responses [422]. Unlike common lipid materials, phospholipids in LNPs do not elicit reactive T cells [423]. Helper lipids, such as DSPC, may induce systemic inflammation by stimulating liver Kupffer cells and releasing eicosanoids and allergens that trigger inflammatory cytokines [424].

The immunogenicity induced by LNPs can be harnessed to enhance therapeutic effects, including acting as adjuvants, activating complement pathways, and facilitating mRNA delivery to specific cell types for immune recognition and presentation [425]. LNPs incorporating mRNA vaccines can evoke robust immune responses, engendering the production of diverse cell types such as Tfh cells, CD4⁺ Th1 cells, CD8⁺ T cells, and GC B cells, along with the generation of neutralizing antibodies [37]. Moreover, LNPs provoke the secretion of IL-12, IL-6, TNF, IFN-γ, and GM-CSF, thereby contributing to immune activation and enhancing vaccine efficacy [414]. The induction of IL-12, a key mediator of innate immunity, occurs through TLRs, PAMPs, or DAMPs, indicating diverse pathways for mRNA-LNP adjuvanticity, possibly independent of inflammasome components like NLRP3 and ASC [417]. IL-6 is potentially significant in enhancing B-cell responses driven by CD8⁺ Tfh cells, including the maturation of germinal center B cells, memory B cells, and long-lived plasma cells [426].

Few researchers have examined the impact of LNP characteristics on immune responses. Hassett et al. found that LNP size significantly influenced the CMV mRNA-LNP vaccine immunogenicity. Larger LNPs (up to 100 nm) induced higher Ab titers in mice, whereas in non-human primates, immunogenicity is independent of size [398]. Another study found that liposome size influenced the Th1 and Th2 ratio, with vesicles ranging from 250 to 750 nm inducing stronger Th1 responses. However, multilamellar vesicles tended to favor Th2 responses. The surface charge of LNPs also plays a role in their immunostimulatory potential. Positive or negative vesicles generated higher Ab-neutralizing responses than neutral vesicles [427]. The functionalization of LNPs also influences their immunogenicity. For example, lipid-anchored gadolinium chelates on the LNP surface activate the complement system by IgM antibodies [428]. Mannosylated LNPs carrying SAM vaccines elicit stronger Ab production and antigen-specific CD8⁺ T cell responses than their non-glycosylated counterparts [281].

Furthermore, the vaccination route influences the LNP immune response. Nebulized LNPs containing mRNA for an anti-hemagglutinin Ab protected mice against influenza [69]. Subcutaneous immunization with BNT162b2 resulted in reduced weight loss and potentially enhanced cytotoxic CD8⁺ T-cell activity, along with increased neutralizing antibodies compared to intramuscular administration [429, 430]. However, intradermal or intramuscular routes were more effective in eliciting polyfunctional CD8⁺ T-cell responses [430]. Therefore, the choice of vaccination route should align with the intended immunological outcome.

Repeated-dose pharmacokinetics and ABC phenomenon

Non-viral vectors typically result in transient gene expression dynamics, where gene activity in the liver, for instance, decreases exponentially by a factor of 10 each day [431]. In contrast, intramuscular administration exhibits a more sustained gene expression over time although the underlying mechanisms remain unclear [432]. Transcription factors such as NF-κB and AP-1 intricately regulate gene expression [433], with their transient activation correlated with the rapid decline in gene expression. To extend therapeutic efficacy, alternative strategies like administering multiple smaller doses of mRNA-LNP therapeutics have been explored. This approach can enhance treatment effectiveness and duration while mitigating adverse effects and immune reactions [6]. Notably, multiple dosing is crucial for the treatment of chronic diseases such as rheumatoid arthritis, cancer, multiple sclerosis, and genetic disorders, as it maintains steady drug levels to manage symptoms and halt disease advancement [434]. Pharmacokinetic challenges in repeat dosing stem from uncertainties about drug absorption, distribution, metabolism, and elimination [435]. Upon administration, LNPs are primarily distributed into the bloodstream and various tissues, with clearance involving the degradation of the mRNA payload and lipid components. Although clearance mechanisms may vary, LNPs are typically eliminated via renal and biliary excretion and tissue degradation. The half-life of LNPs in circulation varies based on factors such as particle size, surface modification, and administration route. In addition, patient physiology and immune responses can intensify these challenges [436].

PEG coating can induce anti-PEG antibodies upon repeated dosing [437], significantly impacting the efficacy of multiple doses [438]. PEG-specific antibodies produced in spleen B cells play a key role in the ABC phenomenon and complement activation [439]. Once injected, PEGylated NPs reach the spleen, where they bind and activate B cells, producing anti-PEG IgM. Subsequent doses may trigger complement activation if anti-PEG IgM is still circulating, resulting in CARPA [440]. For example, PEG has been associated with some cases of anaphylaxis to BioNTech/Pfizer Comirnaty and Moderna Spikevax COVID-19 vaccines in individuals with a known allergy to PEG [441]. In addition, the ABC phenomenon, driven by anti-PEG antibodies, causes opsonization and phagocytosis [442]. Opsonization by C3 fragments and accelerated absorption by Kupffer cells via complement receptor-mediated endocytosis contribute to the rapid clearance of LNPs from the body into the MPS of the lymphoid and other tissues [443].

Understanding the properties of PEGylated NPs is crucial for overcoming ABC challenges, as these characteristics significantly influence the stability, circulation duration, biocompatibility, biodistribution, and cellular uptake of LNPs. The following factors determine the efficacy of PEGylation [444]:

• (a) PEG density: molecular weight and chain length of PEG are critical for reducing immune reactions [445]. Longer PEG chains lead to thicker surface coating on LNPs, limiting immune interaction. Extended chains prolong circulation, while shorter chains enhance cellular internalization. Higher PEG density reduces immunogenicity by concealing antigenic features and decreasing immune detection. The proportion of PEG influences the LNP clearance rate, with higher content accelerating the elimination [446].

• (b) PEG structure: conformation of PEG affects immune system accessibility. NPs with branched PEGs reduce anti-PEG Ab production and complement activation [447].

• (c) PEG shedding: extended PEG chains provide “stealth” characteristics but may shed, reducing immunogenicity [149]. Rapid shedding minimizes immune interactions, while slow shedding maintains effectiveness. LNPs with short-chain PEG-lipids like C14 are preferred for hepatic gene silencing. Regulating the PEG shedding rate is crucial for repeated dosing success [448].

To circumvent the ABC phenomenon, hydrophilic polymers like PVP [449], polyglycerol [450], and pSar [451] may serve as alternatives to PEG, although their tolerability and efficacy require further investigation. In addition, the dosing interval significantly affects the ABC phenomenon, with longer intervals resulting in reduced clearance [437, 443]. Although intravenous administration is less potent in inducing ABC compared with slow infusion, subcutaneous injection enhances ABC upon repeated administration [437]. Further investigation is warranted for LNPs administered via alternative routes of administration.

Prospects

LNPs are promising delivery vectors that encapsulate and deliver RNA therapeutics to target cells and tissues while protecting them from degradation and immune reactions. Chemical modifications can reduce the immunogenicity of RNA therapeutics. In addition, RNA stability and biological function, such as gene expression, can be improved by chemically modified RNA sequences. The market for RNA-based treatments is expanding rapidly, but the sustainability of these opportunities hinges on developing effective delivery mechanisms, notably LNPs. The success of mRNA LNP COVID-19 vaccines has spurred increased interest and investment in mRNA technology, driving the commercial potential of LNP-based delivery systems (Table 4). LNPs enable gene therapy for various medical conditions, such as administering siRNA to reduce cholesterol levels, using the liver as a bioreactor to produce therapeutic proteins, and developing mRNA vaccines for cancer and infectious diseases. This includes innovative mRNA vaccines, personalized cancer immunotherapies, and CAR T cell therapy. Recent clinical trials are investigating their ability to target and destroy cancer cells by immune cells. These trials are showing promising results in terms of both safety and efficacy [452]. Beyond COVID-19 vaccines, LNPs are being utilized in the development of vaccines for other infectious diseases and genetic disorders. Innovations in this area are focused on improving the immune response and ensuring long-term protection [453].

Table 4 Recent LNP-RNA therapeutics in clinical trials

Microfluidic technologies facilitate the high-precision, reproducible, and scalable synthesis of LNPs, overcoming some technical challenges in the field, and a QbD approach can assist in the development of optimized formulations. However, LNPs face challenges in delivering genetic materials to target tissues, such as tissue specificity and immune reactions. LNPs accumulate in the liver, which is beneficial for treating liver disease but problematic for extrahepatic delivery. To passively or actively target other tissues, LNPs can be modified by varying their physicochemical characteristics or introducing targeting ligands. The protein corona formed on LNPs after their administration affects in-vivo fate by altering their stability, integrity, release, biodistribution, excretion, and accumulation in specific organs. The protein corona composition is also specific to each disease and individual, which calls for a personalized nanomedicine approach that can predict optimal treatment for each patient. Recent research has focused on improving the safety of LNPs for widespread long-term clinical uses [454]. Strategies such as modifying the lipid composition and optimizing the particle size are being explored to enhance their biocompatibility [455]. Repeated administration of RNA therapeutics via LNPs is hindered by the immune and complement systems’ recognition of these particles. Future studies aim to optimize LNP formulations to reduce the production of anti-PEG antibodies and minimize complement system activation triggered by LNPs. The design of LNPs involves careful consideration of their physicochemical properties, such as size, charge, and lipid composition. These properties influence the biodistribution, cellular uptake, and safety of the NPs [456]. The ongoing research and clinical trials are paving the way for the development of innovative treatments that have the potential to more safely and effectively tackle a diverse array of diseases.

Conclusion

The advancement of RNA therapeutics follows a cyclical process, starting with designing a delivery vector such as LNPs based on existing knowledge. These particles are manufactured using validated fabrication methods, followed by assessing formulation properties and in vivo behavior. Additionally, addressing safety issues, immunogenicity, repeated-dose effectiveness, and vector functionality for tissue-specific gene delivery are crucial steps. Ongoing research and development efforts are focused on optimizing LNP formulations to boost their stability, efficacy, and tissue-specific targeting. Advances in microfluidic technologies and formulation strategies are expected to improve the performance of RNA-LNP systems, making them more effective and versatile for clinical applications. Furthermore, continued research into the biodistribution and pharmacokinetics of RNA-LNPs will provide valuable insights into their in vivo behavior and tissue targeting capabilities. This knowledge will inform the design of more efficient and targeted delivery strategies that optimize therapeutic efficacy while minimizing off-target effects.