The lion’s share of various LNPs gets distributed to the liver and to a lesser extent to the spleen, and the reason is the presence of a sinusoidal discontinuous endothelium with pores of a ∼100–200 nm diameter in these organs (Kumar et al., 2023). After intravenous administration and entry into the bloodstream, a protein corona forms on the surface of LNPs, and this topic is investigated in detail in Dilliard et al. (2021). The protein corona is capable of influencing the biodistribution of nanoparticles among tissues in various organs owing to the adsorption of the proteins on the tissue surface, leading to internalization via specific receptors. The best-studied mechanism of targeting of LNPs involves their uptake by hepatocytes via binding to low-density lipoprotein (LDL) receptors on their surface (Paunovska et al., 2022). The ApoE protein present in blood plasma is a ligand of LDL receptors and binds to cholesterol on the nanoparticles’ surface and ensures their predominant accumulation in liver cells (Paunovska et al., 2022; Da Silva Sanchez et al., 2022). This mechanism is confirmed by experimental data where LNPs incubated with ApoE have been found to be targeted to hepatocytes (Johnson et al., 2022); on the other hand, particles with a protein corona composed of albumin are distributed preferentially into Kupffer cells instead of hepatocytes (Johnson et al., 2022). Overall, the delivery of LNPs into cells by an ApoE-independent mechanism leads to their accumulation in cells other than hepatocytes (Wadhera et al., 2016). A decrease in the proportion of cholesterol also impairs the delivery of LNPs to hepatocytes (Kawaguchi et al., 2023). Because of the fairly widespread expression of LDL receptor in human tissues (https://www.proteinatlas.org/ENSG00000130164-LDLR/tissue), delivery of nanoparticles containing cholesterol and ApoE (either contained in the particle or immobilized from blood plasma) to all the above tissue types may be promising, e.g., to tissues of the lungs, stomach, liver, and kidneys. The second pathway of LNP uptake in the liver involves Kupffer cell scavenger receptors. For macrophages, there is in vitro evidence that they employ clathrin-mediated endocytosis to internalize LNPs (Hou et al., 2020a).

The size of LNPs also influences targeting efficiency of delivery. When administered intramuscularly, the nanoparticles end up in the liver and spleen, especially if the LNP size is less than 100 nm (Di et al., 2022). Larger nanoparticles, 320 nm, remain at the injection site because they do not seem to be able to penetrate capillaries effectively (Di et al., 2022). Furthermore, the biodistribution of LNPs may differ among animal species (Lam et al., 2023). It is believed that the optimal particle size for nonhuman primates (50–60 nm) is smaller than that for rodents (70–80 nm), possibly indicating a need for further optimization of LNP size for use in humans (Lam et al., 2023; Hassett et al., 2021).

Recent research (Labonia et al., 2023; Hou et al., 2023; Yu et al., 2023) shows that for some tissues/organs, such as the heart, kidneys, and cartilage, a desired protein’s expression can be achieved via injection of an mRNA–LNP complex directly into a target organ (Figure 1). To deliver an mRNA–LNP complex to the retinal pigment epithelium, the subretinal and intravitreal routes of administration have been used, which ensure high expression of a reporter protein (Patel et al., 2019). Expression of the luciferase reporter gene in the retinal pigment epithelium has been observed with the standard four-component composition of LNPs: ionizable lipid:DSPC:cholesterol:DMG-PEG2000 at 50:10:38.5:1.5. The lipid compositions guaranteeing reporter gene expression in the pigment epithelium have contained an ionizable lipid with a tertiary amine (Dlin-MC3-DMA [MC3], Dlin-KC2-DMA [KC2], or 1,2-dioleyloxy-3-dimethylaminopropane [DODMA]). By contrast, nanoparticles based on cationic lipids, N-(4-carboxybenzyl)-N,N-dimethyl-2,3-bis(oleoyloxy)propan-1-aminium (DOBAQ), 1,2-di-O-octadecenyl-3-trimethylammonium propane (DOTMA), 2-dioleoyl-sn-glycero-3-ethylphosphocholine (18:1 EPC), or 1,2-dioleoyl-3-trimethylammonium-propane (DOTAP), have not led to the expression of a reporter protein in the retinal pigment epithelium (Patel et al., 2019). MC3 and KC2 molecules with pK_a = 6.4 and two double bonds in each hydrocarbon tail have shown higher transfection activity as compared to DODMA with pK_a of 7.0 and one double bond in each lipid tail. Researchers attribute this outcome to the fact that the high unsaturation of hydrocarbon radicals in combination with pK_a within the pH range of endosomes (5.5–6.5) promotes membrane destabilization and facilitates the exit of nucleic acids from endosomes (Patel et al., 2019). Most of the reporter protein expression in that study was found in the retinal pigment epithelium and only weak expression in Müller glial cells. Because various types of blindness are associated with the retinal pigment epithelium dystrophy caused by monogenic diseases (Chawla and Vohra, 2023), genetic modification of the retinal pigment epithelium may hold promise for the fight against such diseases. In another study (Ryals et al., 2020), investigators developed nanoparticles based on ionizable MC3 that were able to transfect Müller glial cells, the optic nerve head, and trabecular meshwork with less traumatic (intravitreal) injection. For intravitreal administration, a higher concentration of mRNA is required relative to subretinal administration: 1,500 ng of RNA per dose of the therapeutic agent versus 200 ng. The best efficiency was manifested by particles with 0.5% of PEG and a size of ∼150 nm, which continued to express the fluorescent reporter in retinal cells even on the seventh day (Patel et al., 2019). Particles with a size of 150 nm were too large for phagocytosis, but with a size of <500 nm, they can still be taken up via endocytosis. Of note, reducing particle size to 50–60 nm for retinal delivery impaired reporter protein expression as compared to the larger particles. Similar deterioration of reporter protein expression was also observed when the molar percentage of PEG in the LNPs was increased to 5% (Patel et al., 2019). Five-component particles based on another ionizable lipid, LP01 (Gautam et al., 2023), have been successfully used for gene editing in the retina.

A separate difficult task is the delivery of mRNA therapeutics to the brain or fetal tissue because they are protected by the blood–brain barrier and blood–placental barrier. The passage of LNPs—without additional modifications—through the blood–brain barrier is still a matter of debate (Pardridge, 2023; Pateev et al., 2023). These data suggest that the delivery of LNPs to the brain remains an unsolved problem, and only a few authors have detected their trace amounts (Pateev et al., 2023). After traumatic brain injury, the permeability of the blood–brain barrier is disturbed, and delivery of LNPs up to 100 nm in size to the brain becomes possible (Waggoner et al., 2023). MC3-based four-component nanoparticles carrying luciferase mRNA were administered intravenously in the work just cited, with major distribution to the liver and spleen. When the content of DSPE-PEG was diminished from 1% to 0.1% (and accordingly the particle size was increased), the difference between the damaged and undamaged brain in distribution increased by more than an order of magnitude in the period 4–24 h. After administration of a Cy5-labeled mRNA–LNP complex into the brain via convection-enhanced delivery, the particles diffused only 1 mm away from the injection site and were found in neurons, microglia, and astrocytes (Waggoner et al., 2023). A relatively wide distribution of LNPs in the brain has been achieved only with the help of stereotactic injections of a reporter mRNA–LNP complex (Tuma et al., 2023). Administration of Cre mRNA as an mRNA–LNP complex into the striatum and hippocampus of Ai14 mice in that report led to the expression of the tdTomato reporter within a month in these brain regions (Tuma et al., 2023). Expression of the reporter protein was observed within a radius of up to 3 mm from the injection site, with the greatest intensity in the range 750–1,250 μm in the striatum, and as far away as 2,500 μm (with the greatest intensity in the range 1,250–2,250 μm) in the hippocampus; in this context, the system of single guide RNA and Cas9 induced tdTomato expression in fewer cells as compared to the Cre mRNA system. Furthermore, the expression of the reporter protein was observed there in different types of cells: neurons, astrocytes, and microglia.

LNPs also do not cross the blood–placental barrier (Pateev et al., 2023). To deliver mRNA within LNPs into embryonic tissues, direct injection into the liver of embryos is performed in pregnant mice (Gao et al., 2023). This route of administration leads to wide biodistribution of LNP-bound mRNA throughout organs of the embryo (Cy5 fluorescence of mRNA is observed in the heart, lungs, liver, kidneys, gastrointestinal tract, and brain) and is also detectable in the liver of a pregnant mouse but not other organs. After administration of Cre mRNA in utero to Ai14 mice in that study, mTomato fluorescence was observed in the same organs, with the exception of the brain, indicating the entry of the LNPs into the brain without the expression of the reporter protein (Gao et al., 2023). Notably, in that paper, of all embryonic organs, the highest percentage of cells expressing the reporter protein was observed in the heart: up to 7% (Gao et al., 2023). An analysis of embryonic cell populations after the administration of LNPs there revealed that CD90-positive stem cells get transfected most often in all organs (Gao et al., 2023).

Switching to the intratracheal method of LNP delivery (Massaro et al., 2023) has made it possible to eliminate the expression of the reporter protein in the spleen and liver and to ensure the delivery of LNPs that are based on ionizable lipid MC3 and on phospholipid DPPC (1,2-dipalmitoyl-sn-glycero-3-phosphocholine) exclusively to the lungs. An analysis of lung cell populations there indicated that Cy5-labeled mRNA is detectable in approximately half of all type 1 and type 2 alveolar epithelial cells as well as in half of fibroblasts. Of note, targeting to the lungs after intratracheal administration was not dependent on particle size. In Massaro et al. (2023), nanoparticle size was 34 nm; in another study (Zhang et al., 2020), where the particles were also effectively targeted, their size was ∼150 nm. Thus, several different compositions of four-component particles based on the MC3 lipid and helper lipid DOPE or DPPC have turned out to be targetable to lungs when administered intratracheally (Zhang et al., 2020). It should be pointed out that in both studies on delivery to lungs (Massaro et al., 2023; Zhang et al., 2020), the proportion of a helper lipid (15–20 mol%) was higher as compared to standard (10 mol%) systems of LNPs, possibly indicating the necessity of increasing the helper lipid content for delivery to lungs.

The route of administration also has a significant impact on kinetics of nanoparticle distribution. In Pardi et al. (2015), effects of various administration routes were studied toward the pharmacokinetics of four-component LNPs based on new ionizable lipid L319. After subcutaneous or intradermal administration of mRNA with LNPs, reporter protein expression was observed at the injection site, whereas after intramuscular, intraperitoneal, or intravenous (retro-orbital) administration, at the injection site and in the liver. Upon intratracheal administration, the expression was noted in the lungs and liver. In that paper, intradermal administration caused the most prolonged reporter protein expression: 9 days. In all likelihood, this outcome is due to longer retention of LNPs in the dermis. There are reports of other patterns for nanoparticle distribution. For instance, in Huysmans et al. (2019), it was demonstrated that when administered subcutaneously, LNPs remain at the injection site, with partial migration to a draining lymph node after 24 h. Moreover, additional use of electroporation immediately after the subcutaneous administration of LNPs enhanced the expression of the reporter protein by two orders of magnitude (Huysmans et al., 2019).

Therefore, the route of administration of RNA–LNP therapeutics has a significant impact on their biodistribution. Existing research articles show that even without special manipulation of lipid composition, intracerebral, subretinal/intravitreal, or intratracheal administration of LNPs causes specific delivery to the brain, retina, or lungs, respectively.

The classic composition of LNPs for RNA delivery contains a zwitterionic (neutral) helper phospholipid. In the last few years, many studies were focused on targeted delivery of LNPs to organs (mainly to the spleen and lungs) via changes in the charge of the nanoparticles, which is determined by the helper lipid. This strategy is called Selective ORgan Targeting (SORT), and the lipid responsible for the selective delivery of nanoparticles to organs is called a SORT-lipid (Cheng et al., 2020). Besides, the SORT strategy can be applied in two ways: complete replacement of a helper lipid or partial replacement (introduction of a fifth component).

In LoPresti et al. (2022), it was found that complete replacement of a zwitterionic helper lipid with an anionic one induces greater accumulation of LNPs in the spleen as compared to the liver, whereas replacement with a cationic helper lipid leads to LNP accumulation in lungs. It must be noted that the nature of the helper lipid in that report did not have as much influence on nanoparticle biodistribution as did the charge of the lipid’s polar head.

In Cheng et al. (2020); LoPresti et al. (2022); Radmand et al. (2023), it has been convincingly proven that the use of a cationic helper lipid drives the accumulation of LNPs in lungs. Besides, in Radmand et al. (2023); Kauffman et al. (2018), it has been demonstrated that the greatest accumulation of LNPs occurs in endothelial cells, which are most accessible to the bloodstream. The effect of cationic lipid DOTAP’s proportion in LNPs on the biodistribution of nanoparticles is intriguing, as described in Cheng et al. (2020). For example, LNPs devoid of DOTAP accumulated in the liver; LNPs having a DOTAP content of 10%–15% predominantly ended up in the spleen, whereas with an increase in the DOTAP content above 50%, the LNPs were detectable almost exclusively in lungs. A similar change of the organ showing the highest accumulation (liver, spleen, or lungs) when the proportion of a cationic lipid was increased was documented for other cationic helper lipids: DDAB (dimethyl-dioctadecyl-ammonium bromide) and EPC.

The use of an anionic helper lipid leads to the accumulation of LNPs in the spleen after intravenous administration (Alvarez-Benedicto et al., 2022; Luozhong et al., 2022; Gomi et al., 2023). In Luozhong et al. (2022), it was found that when 1,2-dioleoyl-sn-glycero-3-phospho-L-serine (DOPS) served as the anionic lipid, aside from the accumulation of LNPs in the spleen, their specific accumulation was registered in superficial cervical lymph nodes, in contrast to LNPs containing phosphatidic acid. Authors of Gomi et al. (2023). demonstrated that phosphatidylserine-containing LNPs are mainly taken up by hepatic phagocytes. In this context, the presence of an ionizable lipid in the LNPs enhanced the uptake of the nanoparticles by cells. It is worth mentioning that specific delivery of phosphatidylserine-containing LNPs to lymphoid organs was achieved there only with intravenous administration.

As stated above, targeted delivery of LNPs to organs is determined by the charge, not by the nature of the polar group in a helper lipid. For instance, for nanoparticles containing anionic phosphatidylserine, phosphatidylglycerol, or phosphatidic acid, researchers have observed more specific delivery to the spleen (60%, 50%, and 75%, respectively) than to the liver (20%, 45%, and 5%, respectively) (LoPresti et al., 2022). On the other hand, another group of authors (Pattipeiluhu et al., 2022) has reported that for LNPs containing phosphatidylglycerol, accumulation in the liver occurs predominantly in sinusoidal endothelial cells and Kupffer cells, in contrast to phosphatidylcholine-containing LNPs, whose accumulation takes place equally in hepatocytes, liver sinusoidal endothelial cells, and Kupffer cells.

The introduction of an additional charged helper lipid as a fifth component also enables investigators to alter targeting efficiency of LNPs. Not so long ago (Dilliard et al., 2021), scientists achieved almost 100% reporter mRNA expression in the liver, spleen, or lungs by means of an additional neutral, negatively, or positively charged lipid, respectively. Targeting to an organ depended there on apparent pK_a of nanoparticles: at pK_a < 5, the particles accumulated mainly in the spleen; at pK_a of 6.2–6.4, mainly in the liver; whereas at pK_a > 11, mainly in lungs. It is noteworthy that at pK_a of 6.5–11, there was selective distribution of LNPs into the lungs and spleen at different levels, indicating a nonobvious dependence of the LNP biodistribution on the charge. It is likely that pK_a can serve as a good predictor of bioactivity. For classic four-component LNPs, apparent pK_a was 6.3–6.6 in the above study, depending on the ionizable cationic lipid used. The addition of an anionic lipid reduced pK_a, whereas the addition of a cationic lipid increased pK_a, which affected the interaction with plasma proteins and the formation of the protein corona (Dilliard et al., 2021). Results of their proteomic analysis of the protein corona suggested that targeting to the liver takes place with predominance of the ApoE protein in the protein corona; to lungs, with predominance of the β2-GPI protein; and to the spleen, with predominance of the VTN protein. Incubation of these proteins with the LNPs in vitro also enabled specific LNP targeting (Dilliard et al., 2021).

Those authors proposed a three-stage mechanism underlying the targeting of a five-component LNP (Dilliard et al., 2021). The first stage is the dissociation of the PEGylated lipid from the surface of LNPs; the second is the binding of SORT-lipid to blood proteins; and the third stage is the binding of proteins (adhering to the particle) to tissue-specific receptors. In the case of PEG with a longer aliphatic anchor, which shows worse desorption, the expression of the reporter protein in all organs was noticeably lower, thus confirming the above mechanism. In another study (Meyer et al., 2022), LNPs containing a positively charged group of the additional lipid and a PLGA polymer also got distributed into lungs and the spleen, and the distribution was influenced by the ratio of a cationic lipid (DOTAP) to an ionizable lipid (MC3). Of note, switching the method of mRNA introduction from encapsulation to adsorption made it possible to achieve up to 80% expression of a target protein in the lungs when composition with a DOTAP/MC3 ratio of 1:1 was employed, and up to 50% expression in the spleen with composition featuring a DOTAP/MC3 ratio of 3:1. Nanoparticles carrying adsorbed mRNA had a size of ∼200 nm and did not accumulate in kidneys during excretion, likely indicating the suitability of such particles for subsequent pharmaceutical research and development. In addition, these nanoparticles did not contain cholesterol, and Cy5-labeled mRNA was detectable but not expressed in the liver; this finding may be predictive of the behavior of other cholesterol-free particles. These data are consistent with another study (Gueguen et al., 2024), where LNPs of five-component composition containing cationic lipid IM21.7c also got distributed into lungs and the spleen (and very little into the liver), and >95% of the reporter protein was expressed in the lungs when the LNPs were administered intravenously. The zeta-potential of the LNPs containing cationic lipid IM21.7c was positive, in contrast to the previously obtained nanoparticles (Dilliard et al., 2021), which also got distributed into lungs and the spleen but had negative zeta-potential. We believe that zeta-potential has insufficient predictive value to determine a target organ.

Thus, the addition of a charged lipid into LNPs affects pK_a of the particles, which in turn determines the formation of a protein corona on the surface of the LNPs and accordingly their subsequent distribution throughout organs, which is modulated by proteins of the corona. This principle allows to redirect particles between the liver, spleen, and lungs through the addition of neutral, negatively charged, or positively charged lipids, respectively. It is possible that a combination of an administration route with charged lipids can ensure targeting to new types of tissues and cells.

The LNPs utilized in most studies contain a substantial amount of cholesterol (up to 40 mol%). In Kawaguchi et al. (2023), researchers investigated the effect of the cholesterol content on physicochemical properties of LNPs and on the magnitude of expression and distribution of a reporter protein with various routes of administration. In that study, LNPs were obtained with the classic cholesterol content of 40%, as were LNPs with a lowered cholesterol content of 20% and 10%. From the standpoint of physicochemical properties, the nanoparticles of all types turned out to be almost identical. Their sizes ranged from 75 to 140 nm, with a relatively narrow size distribution (polydispersity index [PDI] <0.25). It is worth noting that LNPs of the classic composition were smaller compared to the others and had a narrower size distribution. At the same time, the stability of the obtained LNPs decreased with the diminishing proportion of cholesterol: within a week, particles with 10% of cholesterol increased in size while losing encapsulated mRNA. Cholesterol only weakly affects the degree of mRNA encapsulation into nanoparticles. Usually, the correlation between the cholesterol amount and mRNA loading is within systematic error, and therefore it cannot be said definitively that cholesterol increases the loading. A weak correlation between RNA encapsulation and the level of cholesterol in LNPs has been observed only in two studies (Kawaguchi et al., 2023; Zhang et al., 2020). In Kawaguchi et al. (2023), the effect of cholesterol was demonstrated the most clearly because concentrations of the other components, except for phospholipid, were fixed. Taking into account their reported experimental errors, encapsulation efficiency remained constant at the ratios used. The only parameter that cholesterol affected is the stability of the resulting nanoparticles. In the same work (Kawaguchi et al., 2023), it was noted that with a decrease in the cholesterol amount, the nanoparticles tend to increase in size over time, and they also have a higher PDI and their encapsulation efficiency declines over time.

For subcutaneous and intramuscular administration of the LNPs, a decrease in the cholesterol content reduced the biodistribution of the nanoparticles to the liver and the expression of the reporter protein there, whereas at the injection site, the level of reporter protein expression did not change. This evidence suggests that the creation of LNPs that do not end up in the liver requires reducing the amount of (or replacing) cholesterol, in agreement with the results from Radloff et al. (2023), which indicates that the use of cholesterol-free three-component LNPs when administered intravenously leads to reporter protein expression almost exclusively in lungs, with small levels in the spleen. Reducing the cholesterol content also improved the targeting of LNPs to the spleen when the LNPs were based on known lipids and were administered intravenously (Radloff et al., 2023). In the optimized composition of LNPs in Shimosakai et al. (2022), which showed >95% efficiency of LNP delivery to the spleen, the cholesterol content was reduced to 10 mol%, whereas the concentration of a helper lipid was increased to 60 mol%. Among splenocytes, the LNPs got successfully distributed into antigen-presenting cells but not T lymphocytes (Shimosakai et al., 2022).

Modified cholesterol can also be employed for LNP targeting. In Paunovska et al. (2019), scientists assessed the effect of cholesterol modifications on the delivery of Cre mRNA within LNPs into different types of cells of Ai14 mice in seven organs: the liver, spleen, heart, kidneys, lungs, pancreas, and bone marrow. Those authors used oxidized forms of cholesterol, where rings or the tail of cholesterol were modified with a hydroxyl or ketone group. The distribution of these LNPs among cell and tissue types did not depend on the cholesterol being tested; tdTomato-positive cells were found in the spleen and liver; in other tissues, the percentage of such cells was negligible. The proportion of liver cells expressing tdTomato was 100%, 80%, 60%, and 20% among Kupffer cells, endothelial stem cells, immune cells, and hepatocytes, respectively, whereas in the spleen, 17%, 9%, and 3% among T cells, B cells, and macrophages, respectively (Paunovska et al., 2019). The administration of LNPs containing modified 20α-hydroxycholesterol enhanced reporter protein expression in endothelial cells, hepatocytes, Kupffer cells, and liver immune cells by an order of magnitude as compared to LNPs containing unmodified cholesterol (Paunovska et al., 2019). Authors of Hatit et al. (2023) have obtained additional data on the use of 20α-hydroxycholesterol in LNPs thus showing that the observed effect is not due to a change in physicochemical properties of the LNPs but is explained by biological effects. Their LNPs containing a mixture of stereoisomers, when entering a cell, activated genes associated with phagocytosis and inflammation to a greater extent. Consequently, replacing unmodified cholesterol with stereopure 20α-hydroxycholesterol may be promising for increasing the expression of a target protein.

Thus, cholesterol is necessary for membrane stability but at the same time directs LNPs to the liver. Replacing cholesterol with analogs does not affect the biodistribution but may influence the efficiency of cell transfection. Accordingly, to redirect particles from the liver to other organs, one possible solution is to devise cholesterol-free formulations or utilize additional features that will ensure targeting to a desired organ while promoting stability of the nanoparticles’ membrane.

The introduction of PEG into lipid composition allows for self-assembly of LNPs and reduces nanoparticle aggregation and heterogeneity owing to steric stabilization. Furthermore, PEG improves the stability and durability of LNPs because of the formation of a shell that stabilizes the membrane (Hald et al., 2022). The downside of PEG addition is that an overly stable membrane prevents LNPs from fusing with an endosome, and hence the mRNA is unable to enter the cell. Some articles (Gueguen et al., 2024; Ryals et al., 2020; Lam et al., 2023) suggest that reducing the percentage content of a PEGylated lipid improves the expression of a target protein in animals. By varying the molar percentage and molar mass of PEG, an investigator can control the size and size distribution of the resultant nanoparticles. In one study (Sarode et al., 2022), a series of PEGylated lipids was investigated that differed in length (550–2000 Da) and architecture (linear or branched) of the polymer chain and in the nature of the lipid moiety (phosphoglyceride, diglyceride, or ceramide). Those authors revealed that LNPs containing a PEGylated lipid based on phosphoglycerides are characterized by a pronounced dependence of particle size on the molar proportion and molecular weight of PEG in contrast to lipids based on diglycerides and ceramides. Among particles with a high content of a long-chain PEGylated lipid, the size distribution was broad, which those authors attributed to the formation of a subpopulation of PEGylated lipid micelles. There was no obvious impact of the lipid moiety on the size of the LNPs.

Of note, the use of LNP therapeutics containing a PEGylated lipid can cause unwanted immune responses in the body (Tenchov et al., 2021; Shiraishi and Yokoyama, 2019). In particular, the emergence of accelerated blood clearance has been documented. In the blood, anti-PEG antibodies appear that quickly remove the administered PEGylated nanoparticles. As a result, the effectiveness of the mRNA therapeutic agent diminishes, and there is a need for repeated administration of the therapeutic. Thus, the incorporation of a PEGylated lipid into LNPs can be an obstacle to repeated use of mRNA–LNP therapeutics. As an alternative to PEG, a polysarcosine (pSar) has been investigated, which is a polypeptide derived from sarcosine and possesses all the properties of PEG (Birke et al., 2018). Moreover, a polysarcosine does not induce an immune response even after repeated administration, in contrast to PEG.

Polysarcosines can serve as analogs of PEG (Nogueira et al., 2020). Just as for PEG, an increase in the molar proportion of a polysarcosine within LNPs reduces particle size: this pattern is observed when the molar proportion of a polysarcosine within LNPs changes from 1.5% to 10%, and the shorter the polysarcosine, the more pronounced is the effect. In the article just cited, the use of the shortest polysarcosine called pSar11, containing 11 units, within LNPs diminished particle size from 300 to 125 nm, whereas the addition of the longest polysarcosine (pSar65, containing 65 units) into the LNPs led to a decrease in particle size from 150 to 75 nm. With increasing length of the polysarcosine component of the LNPs, the accumulation of the particles shifted from the spleen to the liver. Those authors ascribe this redistribution of LNPs to their size: larger particles with a shorter polysarcosine accumulate more efficiently in the spleen. A similar redistribution pattern was registered with an increase in the proportion of a polysarcosine of the same chain length within the LNPs: an increase in the proportion of the polysarcosine reduced particle size and hence shifted their accumulation from the spleen to the liver. Additionally, LNPs based on a polysarcosine, as compared to conventional PEG, have lower reactogenicity, as shown in experiments on mice (Nogueira et al., 2020). A recent study (Gautam et al., 2023) addressed various modifications of PEG within LNPs. For subretinal delivery, LNPs were tested containing additional PEG modified with a carboxyl group, an amine group, or a carboxyester group. Depending on the charge, the LNPs showed different tropism for eye photoreceptor cells. Control four-component LNPs induced tdTomato expression only in the retinal pigment epithelium, whereas LNPs with all types of modified PEG were able to transfect nuclei of cone and rod photoreceptor cells (Gautam et al., 2023). In these experiments, LNPs containing PEG with the carboxyester group caused expression of the reporter protein in the outer segment of a photoreceptor, in the nucleus, and in the retinal pigment epithelium. In addition to these modifications, PEG conjugated with the BODYPI dye is used too, but LNPs based on this lipid have shown weaker reporter protein luminescence as compared to particles based on DMG-PEG2000 (1,2-dimyristoyl-rac-glycero-3-methoxypolyethylene glycol-2000), possibly because of lower pK_a of the BODYPI-PEG nanoparticles (∼5.0) in comparison with pK_a of the LNPs based on DMG-PEG2000: ∼6.5 (Xiong et al., 2020).

Thus, PEG, just as cholesterol, is necessary to stabilize a nanoparticle and to increase circulation time, but an excessive PEG content negatively affects properties of LNPs. Introduction of a polysarcosine (which has lower immunogenicity) instead of PEG may be promising.

RNA–LNP-based therapeutics have become popular after successful use of two mRNA vaccines (RNA-1273 and BNT162b2) against SARS-CoV-2. To date, the number of approved mRNA vaccines has increased, but not all of them take advantage of LNPs for delivery (Hu et al., 2024). Aside from being used as vaccines, RNA–LNP constructs are employed to treat cancers (as encoded adjuvants, neoantigen vaccines, in vivo chimeric antigen receptor T-cell therapy [CAR-T], or epigenetic regulators) and hereditary diseases (as genetic editors) (Hu et al., 2024). On the ClinicalTrials.gov website, we searched for all clinical trials that have involved LNPs. In most studies, information about the delivery system is not disclosed, only 16 papers mention LNPs, and detailed composition is described in only five papers (Table 2). Lipid composition is often not disclosed by the companies that have developed the drug; hence, it is not possible to fully study the composition of the formulations in question. All known compositions from the above-mentioned clinical trials contain four lipids: an ionizable lipid, a phospholipid, cholesterol, and a PEGylated lipid, and the compositions themselves are already known (trials No. NCT03323398, CT05945485, and NCT05755620). Moderna utilizes a new ¹³C-labeled lipid, which is an analog of SM-102 and will allow for more accurate assessment of pharmacokinetics of mRNA–LNP therapeutics (Schellekens et al., 2011). Another new ionizable lipid, XH-07, is used by Duke University (NCT06468605). Side effects of the new lipids have not yet been investigated.

Cancer vaccines have involved mRNAs encoding cytokines IL-23, IL-36γ, IL-12, or OX40L (NCT03739931, NCT03323398, and NCT06468605) as well as hepatocyte nuclear factor 4α (NCT06572189), which participates in hepatocyte proliferation. Linear mRNA-based vaccines are also being tested against infectious diseases such as influenza (NCT05829356 and NCT05755620), malaria (NCT06069544 and NCT05581641), and HIV (a booster vaccine; NCT05903339). Separately, drugs are being developed against hereditary disorders, including orphan diseases such as cystic fibrosis (NCT05712538), ornithine transcarbamylase deficiency (NCT04442347, NCT05526066, NCT06488313, NCT04416126, and NCT03767270), glucose-6-phosphatase deficiency (NCT05095727), and hereditary transthyretin amyloidosis with polyneuropathy (NCT04601051).

When new cationic/ionizable lipids are designed, addition of new functional groups or replacement of existing ones may not affect the biodistribution of nanoparticles within the body. Typical biodistribution with subsequent expression in the liver and spleen is characteristic of LNPs containing i) cationic/ionizable lipids linked through disulfide bonds and carrying an unsaturated tail and amine head (Shen et al., 2023; Lee et al., 2020), ii) lipids with tertiary-amine headgroups modified with a terminal hydroxyl group, ester bond, and branched saturated tails (Long et al., 2023), iii) modified aminoglycosides that are synthesized from natural existing aminoglycosides coupled with alkyl epoxides and acrylates (Yu et al., 2020), or iv) a cationic lipid carrying an amine head with two unsaturated tails (Hashiba et al., 2020).

During the construction of new delivery systems based on four-component LNPs, researchers try to synthesize new ionizable lipids having unique properties. As mentioned above, targeting of LNPs to desired organs can be achieved by means of ionizable lipids and/or via introduction of an additional helper lipid. For example, after intravenous injection of LNPs, positively charged DOTAP as a helper lipid ensures delivery to lungs (Dilliard et al., 2021).

For new LNP compositions, we evaluated the influence of chemical modifications of a cationic lipid on biodistribution and transfection efficiency (Tables 3–5), then analyzed in detail the distribution of an mRNA or protein product among organs and cell populations as well as methods and doses of administration, chemical and physical properties of the LNPs based on new lipids, and strategies for the design of new lipids (Table 6).

Separately, we would like to discuss such specific use cases of LNPs as delivery to the placenta and to a malignant tumor.

There are strategies for LNP design that differ from the standard four-component formulation. For example, lipid CAP-2 (cholesterol‐amino‐phosphate) has been created (Du et al., 2023), which combines an ionizable MC3 group and two molecules of cholesterol instead of alkyl tails, and this approach reduces the number of lipid components to three. When such an ionizable lipid was being designed, the layout of the amine core was altered, and alkyl substituents of different lengths and different linkers were utilized. During the construction of CAP molecules, the critical packing parameter, which characterizes the resultant phase during endosomal fusion, was calculated, and the molecules were designed in such a way that after the fusion, an inverted hexagonal (HII) phase formed, which is characterized by the greatest release of mRNA from the endosome. In animal trials, three-component LNPs containing CAP-2 have proven to be more effective at transfecting mouse seminiferous tubules as compared to four-component nanoparticles containing lipids MC3 and ALC-0315, which have not shown transfection activity after administration of LNPs based on them into testes. It is important to mention that other organs have not been studied, and therefore the question of specificity of such nanoparticles remains open (Du et al., 2023).

In another study (Yang et al., 2020), researchers assembled three-component nanoparticles in which they used lipid iBL0713, which combines functions of ionizable and helper lipids. For these LNPs, the biodistribution and expression of a reporter gene were evaluated by the IVIS (in vivo imaging system) method for the detection of luciferase and of Cy5-labeled mRNA encoding it, respectively. After detection of Cy5-labeled mRNA, its distribution was mainly seen in kidneys and to a lesser extent in the liver and spleen, while luciferase luminescence was detectable mainly in the liver and less in the spleen. Notably, Cy5-labeled mRNA was present in all organs, including the pancreas, brain, and thymus, where the reporter protein expression was not detectable (Yang et al., 2020).

LNPs made of two lipids—DOPE and a tri-substituted benzaldehyde derivative (Li et al., 2021)—have been found to cause the expression of a luminescent product exclusively in the spleen when administered intravenously but not subcutaneously or intramuscularly. To choose a final lipid, various alkyl saturated and cyclic substituents in para-substituted and tri-substituted benzaldehyde molecules were screened in that work, followed by selection of the most favorable ratio of the two lipids to mRNA. In tests on cultured cells, lipid-based particles with unsaturated tails manifested the greatest efficiency. The resulting nanoparticles were internalized mainly by macrophages and dendritic cells.

Three-component LNPs have been applied to antimicrobial treatments (Hou et al., 2020b). These LNPs contain DOPE, cholesterol, and vitamins modified by alkyl groups with tertiary amines; these compounds replace cationic lipids. To find the best lipid compositions, researchers have compared all the vitamin nanoparticles with each other. The greatest effectiveness was shown by composition with a lipid based on vitamin C (Hou et al., 2020b). Vitamin C–based nanoparticles containing antimicrobial peptide AMP-IB367 fused with cathepsin via a cathepsin-cleavable linker got delivered into macrophage lysosomes, thereby enabling the targeting of the payload to this cell compartment. The targeting was mediated not by the decoration of the particles but by the protein synthesized from mRNA directly in the cell. These LNPs were able to reduce the bacterial load and improve the survival rate of mice with sepsis.

Thus, proportions (%) of lipids within LNPs can vary. The minimal composition of effectively functioning LNPs includes an ionizable lipid and structural lipid, but three-component compositions of LNPs have been used successfully too. All these modifications of LNP composition may be interesting regarding the targeting of new tissues and cell types.

When nanoparticles are fabricated, a modified sequence of formulation is also utilized (Li Q. et al., 2023): mRNA is encapsulated into short peptides containing arginine, then this complex is formulated into cationic lipid DOTAP, resulting in the formation of particles with a separate membrane and core. In this procedure, different versions of arginine peptides contain stearyl, cholesterol, and various amino acids, but in cell assays, the best results have been shown by a peptide consisting of nine arginines and six histidines. After intraperitoneal administration of radioactive iodine and LNPs containing the mRNA of receptor NIS (which ensures the transfer of radioiodine into cancer cells), the localization of radioactive iodine exclusively in the thyroid gland was proven by computed tomography. The treatment reduced tumor size by a factor of five at 18 days after the induction as compared to control (Li Q. et al., 2023). Evidence from human skin explants administered intradermally (Blakney et al., 2019) indicates that the application of three-component LNPs containing a charged lipid and formulated together with DOPE and cholesterol can lead to high expression of a reporter protein. In that work, LNPs based on cationic lipids DDAB and DOTAP, ionizable lipid C12-200, or zwitterionic Cephalin manifested an ability to mediate the expression of small activating RNA in the human explants, for a maximum period of up to 21 days with DOTAP. Transfection of the largest number of cells was achieved by particles based on Cephalin. All compositions showed a similar distribution among cell types: ∼40% in the epithelium, ∼10% in fibroblasts, and ∼50% in immune cells (natural killers, leukocytes, T lymphocytes, dendritic cells, monocytes, and Langerhans cells).

To summarize, unconventional approaches to encapsulation of RNA into LNPs also ensure its efficient delivery. Rearranging the order of addition of lipids to mRNA and the use of peptides for mRNA encapsulation are new methods that may be suitable for the development of LNP compositions.