A lipid nanoparticle (LNP) is a small particle assembled primarily from lipids and used to carry biologically active substances. In pharmaceutical science, the term encompasses several kinds of lipid-based nanoparticles, but it is especially associated with formulations that encapsulate RNA within a mixture of ionizable lipids, structural lipids, cholesterol, and surface-stabilizing lipids. Many therapeutic LNPs have diameters of roughly 50–150 nanometres, although size depends on composition, preparation method, and intended application.
LNPs address a central problem in drug delivery: a molecule with useful biological activity may be unstable outside cells, unable to cross a cell membrane, or distributed to the wrong tissues. RNA is particularly difficult to deliver because it is large, negatively charged, and susceptible to degradation. An LNP can protect RNA during administration and help a fraction of it reach the cytoplasm, where it can alter gene expression or direct the production of a protein.
The clinical importance of the technology was demonstrated by the approval of the small interfering RNA medicine patisiran in 2018 and by the deployment of lipid-encapsulated messenger RNA vaccines against COVID-19 beginning in 2020. These applications share basic delivery principles but differ substantially in dose, administration route, target cells, and desired immune response. LNP development consequently combines lipid chemistry, molecular biology, materials characterization, pharmaceutical manufacturing, and the study of interactions between synthetic particles and living tissues.
Historical development and terminology
Modern lipid delivery systems developed from research on the behavior of phospholipids in water. In work published in 1964, Alec Bangham and R. W. Horne described membrane-like structures formed by hydrated phospholipids using electron microscopy. These observations helped establish the study of the liposome, a vesicle bounded by one or more lipid bilayers. Liposomes provided experimental models of biological membranes and offered a way to enclose water-soluble substances within an aqueous compartment while incorporating hydrophobic compounds into a membrane.
Early pharmaceutical work concentrated on delivering conventional drugs. Important obstacles included rapid removal from blood, leakage of the encapsulated material, variability in particle size, and difficulties producing reproducible batches. Surface modification with polyethylene glycol (PEG) helped extend the circulation of some formulations. In 1995, the United States Food and Drug Administration approved Doxil, a PEGylated liposomal formulation of doxorubicin. It demonstrated that a lipid carrier could become a regulated medicine with distinctive distribution and pharmacokinetic properties, rather than simply serve as an inert container.
The delivery of genetic material introduced different requirements. In 1987, Philip Felgner and colleagues reported lipid-mediated transfection using the cationic lipid DOTMA. Positively charged lipids could associate with negatively charged nucleic acids and promote their interaction with cells. However, formulations that remained strongly cationic under physiological conditions often had unfavorable interactions with proteins and membranes. Subsequent research developed lipids whose charge changes with acidity, allowing them to bind nucleic acids during preparation while remaining less charged in the bloodstream.
During the 1990s and 2000s, work by groups including those of Pieter Cullis and Thomas Madden advanced the encapsulation of nucleic acids in small, reproducible lipid particles. The discovery of RNA interference created a particularly strong need for intracellular delivery. Patisiran, approved in 2018, became the first approved therapeutic using small interfering RNA and an important clinical validation of ionizable LNP technology. Messenger RNA vaccines subsequently brought related delivery systems into large-scale public use.
Terminology remains context-dependent. “Lipid nanoparticle” sometimes functions as an umbrella term that includes liposomes, solid lipid nanoparticles, and nanostructured lipid carriers. In contemporary RNA-delivery literature, it usually refers more narrowly to an ionizable-lipid formulation containing an internally organized lipid–nucleic acid phase rather than a simple hollow bilayer vesicle. Liposomes and RNA LNPs therefore have overlapping histories but are not interchangeable structural categories. Nor does the label LNP identify a single composition, manufacturing process, or biological behavior.
Particle architecture and lipid components
Most established RNA LNP formulations contain four functional classes of lipid: an ionizable lipid, a helper phospholipid, cholesterol, and a PEG-conjugated lipid. Their proportions affect particle formation, internal organization, stability, and biological activity. A frequently cited research formulation uses an approximate molar ratio of 50:10:38.5:1.5 for these four components, respectively. This is an illustrative composition, not a definition or a universal recipe; successful products and experimental formulations use different ratios and different lipid structures.
The ionizable lipid is usually the principal component responsible for binding RNA and supporting intracellular release. Its molecular architecture commonly includes an amine-containing headgroup, hydrophobic tails, and connecting groups that influence packing and degradation. Acidic conditions protonate the headgroup, producing a positive charge that favors association with negatively charged RNA. At physiological pH, many such lipids become predominantly uncharged. Their effective acid–base behavior within a particle is described by an apparent , which need not equal that of an isolated lipid molecule in solution.
Helper phospholipids provide structural and interfacial functions. Distearoylphosphatidylcholine, or DSPC, is used in several established formulations and has saturated hydrocarbon chains that support relatively ordered membrane structures. Dioleoylphosphatidylethanolamine, or DOPE, is used in other systems and has molecular packing properties that favor non-bilayer structures under appropriate conditions. Cholesterol occupies spaces between other lipids and changes packing, permeability, and mechanical properties. Its effects depend on the surrounding mixture rather than corresponding to a simple universal increase in either rigidity or fluidity.
PEG-lipids present a hydrated polymer chain toward the surrounding liquid. This surface layer helps control particle growth during preparation and reduces aggregation. It can also reduce some interactions with proteins and cells. The length and chemical structure of the lipid anchor influence how rapidly the PEG-lipid exchanges out of the particle. This creates a design trade-off: surface shielding that improves colloidal stability may also inhibit cell association or membrane interactions needed for delivery.
The resulting particle is not necessarily a miniature sphere with a uniform membrane around a water-filled cavity. Structural studies show that RNA LNPs can contain disordered or organized internal lipid-rich regions, aqueous spaces, and RNA associated with ionizable lipids. Some preparations display internal lamellar features or other nanoscale arrangements. Their structures depend on payload, acidity, lipid composition, and preparation history. Consequently, schematic diagrams showing a neatly layered particle are useful teaching devices but cannot represent every LNP formulation or every stage of its formation.
Formation, manufacturing, and scale-up
RNA LNPs are commonly prepared by rapidly mixing an aqueous nucleic-acid solution with lipids dissolved in a water-miscible organic solvent, usually ethanol. The aqueous solution is often acidic so that the ionizable lipid is positively charged during assembly. When the streams mix, the solvent environment changes, hydrophobic lipids associate, and RNA becomes incorporated into the developing particles. The process is therefore governed by coupled self-assembly, electrostatic association, solvent exchange, and mass transfer, rather than by mechanically filling previously manufactured containers.
Mixing speed matters because assembly can occur on very short timescales. If local lipid concentration and solvent composition differ widely across a mixer, particles can grow through different pathways and produce a broad size distribution. Microfluidic mixers, impinging-jet systems, and other controlled mixing devices create reproducible contact between streams. Important operating variables include total flow rate, the ratio of aqueous to organic flow, lipid concentration, nucleic-acid concentration, buffer composition, and temperature. The same nominal lipid recipe can give different particles when these variables change.
The ratio of protonatable nitrogen groups in the lipid to phosphate groups in the nucleic acid is often expressed as an N/P ratio. It provides a useful formulation descriptor, but it does not fully specify particle charge or encapsulation behavior. Different lipids have different numbers of ionizable groups, different apparent values, and different packing characteristics. RNA length and secondary structure also affect assembly. An N/P ratio that works for a short duplex cannot automatically be transferred to a much longer messenger RNA without evaluating the resulting product.
After initial formation, the dispersion typically undergoes solvent removal and buffer exchange. Tangential-flow filtration can retain the nanoparticles while allowing ethanol, salts, and other small solutes to pass through a membrane. Concentration, pH adjustment, addition of stabilizing excipients, sterile filtration where compatible, and aseptic filling follow as required by the process. Sterilization methods suitable for small-molecule products, such as strong heat treatment, may damage RNA or disrupt lipid structures. Control of microorganisms and contaminants therefore begins well before the final container is filled.
Industrial scale-up is a problem in chemical engineering as well as pharmaceutical formulation. Increasing vessel size or flow rate can change mixing behavior and residence-time distribution. Manufacturers may preserve local mixing conditions by using larger validated mixers or multiple parallel channels. Raw-material purity, equipment contact surfaces, filtration losses, and filling operations can also affect the final product. Reproducibility depends on demonstrating that a controlled process consistently produces particles with the required chemical composition, physical properties, and biological activity, not merely on reproducing an average diameter.
Cellular uptake and endosomal escape
An LNP must overcome several distinct barriers before its RNA can act. Reaching a tissue does not establish that the particle has entered the relevant cells, and entry into a cell does not establish that RNA has reached the cytoplasm. Many particles enter cells by endocytosis, in which the plasma membrane encloses extracellular material within a membrane-bound compartment. Uptake pathways vary with particle properties and cell type. Surface-bound proteins and cellular receptors can be as important as direct interactions between the synthetic lipids and the membrane.
After uptake, the particle commonly enters an endosome, part of the cell’s sorting and degradation system. Endosomes become more acidic as they mature. This changes the protonation state of an ionizable lipid, potentially increasing its positive charge. In an idealized description for a basic headgroup, the protonated fraction is
This relationship illustrates why the same lipid can behave differently during acidic formulation, in near-neutral extracellular fluid, and within an acidifying intracellular compartment. Real particles are more complex because neighboring lipids, local electrostatics, and internal organization influence the apparent .
A leading mechanistic explanation for escape involves interactions between protonated ionizable lipids and negatively charged endosomal membrane lipids. The resulting lipid mixtures can favor packing arrangements incompatible with a stable bilayer, promoting local membrane defects and release of some cargo. Experimental observations also support localized membrane damage and heterogeneous release events. The mechanism is formulation-dependent, and no single sequence of molecular events has been established for all LNPs. Acidification alone is therefore not sufficient to predict effective delivery.
Endosomal escape is generally inefficient. Some studies have estimated cytoplasmic release of only a few percent of internalized small interfering RNA under particular experimental conditions. That figure is not a universal property of all LNPs, but it illustrates why cellular uptake can greatly exceed productive delivery. Cargo that remains trapped may be recycled out of the cell or transported toward lysosomal degradation. Fluorescence microscopy showing particles inside cells consequently does not, by itself, demonstrate successful functional delivery.
Once released, the payload follows its own biological pathway. Messenger RNA can be translated by cytoplasmic ribosomes into a protein. Small interfering RNA is incorporated into the RNA-induced silencing complex, which uses sequence complementarity to reduce expression of a target RNA. Delivery of DNA generally introduces an additional barrier because transcription normally requires access to the nucleus. LNP design must therefore be matched not only to the destination tissue but also to the intracellular destination and molecular requirements of the cargo.
Biodistribution and tissue targeting
Biodistribution describes where a formulation and its components travel after administration. It is strongly influenced by route: intravenous injection immediately exposes particles to blood, whereas intramuscular injection initially places them within muscle tissue and its extracellular environment. Local uptake, lymphatic drainage, vascular transport, and clearance then compete. The distribution of an intact LNP, a lipid label, and functional RNA expression may differ, so measurements of one cannot automatically substitute for measurements of the others.
After entering a biological fluid, nanoparticles acquire a changing layer of adsorbed proteins and other molecules called a protein corona. This layer can alter their effective surface properties and recognition by cells. Apolipoprotein E has an important role in the liver uptake of several ionizable LNP formulations. Its association with particles can support interactions with receptors involved in lipoprotein uptake. This contributes to the effectiveness of certain intravenously administered LNPs in delivering RNA to hepatocytes, although the process depends on formulation and biological context.
The liver is also favored by its anatomy and clearance functions. Specialized capillaries facilitate access to hepatic cells, while liver-resident macrophages can capture circulating particles. Other phagocytic cells, including those in the spleen, also participate in particle removal. These processes help explain why the liver and spleen are common sites of accumulation after intravenous administration. Accumulation nevertheless does not prove delivery to a particular cell population: signal in liver tissue may originate from hepatocytes, endothelial cells, macrophages, extracellular material, or degraded components.
Targeting strategies seek to redirect this natural distribution. Lipid composition can be modified to alter interactions with plasma proteins and cell membranes. Surface ligands, such as sugars, peptides, or antibodies, can be attached to promote receptor recognition. Some experimental approaches introduce additional charged lipids to shift delivery toward organs such as the lung or spleen. These effects are sensitive to dose, route, species, and assay conditions. A formulation that preferentially transfects one organ in a mouse does not necessarily reproduce that distribution in humans.
Several tissues remain especially difficult to reach. Vascular barriers, dense extracellular matrices, rapid clearance, and limited access to the appropriate cell surface can all restrict delivery. In the brain, the blood–brain barrier limits the entry of many circulating materials. In solid tumors, abnormal vessels and interstitial pressure create spatially uneven exposure. For any targeting claim, a useful distinction is whether a carrier merely accumulates in an organ, enters the intended cell type, or produces the required molecular effect there. These are separate experimental outcomes with different implications.
Therapeutic payloads and established applications
LNP technology supports several classes of nucleic-acid medicine, each with different delivery requirements. Small interfering RNA is usually a short double-stranded molecule designed to reduce expression of a selected gene. Messenger RNA is considerably longer and encodes a protein. Gene-editing systems can require delivery of an editor-encoding messenger RNA together with one or more guide RNAs. Other proposed cargos include self-amplifying RNA, circular RNA, and combinations of nucleic acids. Differences in length, charge density, structure, and required intracellular dose affect formulation behavior.
Patisiran, marketed as Onpattro, provides an established example of systemic siRNA delivery. Approved in the United States and European Union in 2018, it treats the polyneuropathy associated with hereditary transthyretin-mediated amyloidosis in adults. Its LNP delivers siRNA to the liver, where it reduces production of transthyretin. The formulation includes the ionizable lipid DLin-MC3-DMA, DSPC, cholesterol, and a PEG-lipid. Its activity illustrates a particularly favorable match between the natural hepatic distribution of an intravenously administered LNP and a disease mechanism involving a liver-produced protein.
COVID-19 vaccines demonstrated a different use: transient production of an antigen to stimulate the immune system. The Pfizer–BioNTech and Moderna vaccines introduced in 2020 used LNP-encapsulated messenger RNA encoding the SARS-CoV-2 spike protein. Their original formulations used different ionizable lipids, ALC-0315 and SM-102, respectively. Following intramuscular administration, antigen production and subsequent immune processing support adaptive immune responses. The lipid formulation is integral to RNA protection and delivery, rather than a separate optional accessory to an otherwise readily deliverable molecule.
Therapeutic protein expression is another major application. Instead of administering a purified protein, an mRNA formulation can temporarily supply cells with the instructions to produce it. Investigational programs have examined this approach for secreted proteins, intracellular enzymes, and antigens used in cancer immunotherapy. Personalized cancer-vaccine approaches can encode selected tumor-associated mutations. Such applications differ from infectious-disease vaccines in patient population, antigen selection, treatment setting, and accompanying therapies; the shared use of mRNA does not make their development requirements identical.
LNPs also support experimental in vivo gene editing, including delivery of components of CRISPR–Cas systems. A formulation may carry messenger RNA encoding an editing protein together with guide RNA. Transient delivery limits the duration of editor production, but the resulting genomic change can be permanent. This separates the pharmacokinetics of the carrier from the duration of the biological effect. Clinical progress with one payload or indication cannot establish the safety or effectiveness of another: the nucleic-acid sequence, expressed product, target tissue, and formulation all contribute to the final medicine.
Physical characterization and biological testing
LNP characterization requires multiple methods because no single measurement describes a particle’s composition, architecture, payload integrity, and biological activity. Particle diameter and size distribution are commonly assessed using dynamic light scattering. This method infers a hydrodynamic size from fluctuations in scattered light caused by particle motion. It is rapid and useful for process monitoring, but larger particles contribute disproportionately to the signal. A small number of aggregates can therefore influence the reported distribution, and an average diameter can conceal heterogeneous populations.
Imaging supplies complementary information. Cryogenic electron microscopy, performed with an electron microscope on rapidly vitrified specimens, can reveal particle shape, internal features, and some forms of structural heterogeneity. It avoids several drying artifacts associated with conventional specimen preparation, although freezing, sample selection, and image interpretation still require care. Small-angle X-ray scattering and other techniques based on spectroscopy or scattering can probe internal organization across larger sample populations. Results from imaging and ensemble methods are most informative when interpreted together.
Chemical analysis establishes what the particles contain. Chromatographic methods can quantify individual lipids and detect degradation products. RNA assays assess identity, concentration, length distribution, and integrity. Encapsulation is often measured using dyes that detect accessible nucleic acid before and after particle disruption. A typical calculation expresses encapsulation efficiency as
where represents RNA detected under conditions intended to exclude intact particles. The result depends on assay design and does not necessarily establish that every encapsulated molecule is intact or functional.
Biological potency testing measures an effect relevant to the intended mechanism. For an mRNA formulation, this may involve production of the encoded protein in a defined cell system. For siRNA, it may involve reduction of a target transcript or protein. Such assays integrate several steps—uptake, intracellular trafficking, escape, and payload activity—but may not reproduce the barriers encountered in a whole organism. Conversely, a simple uptake assay can be highly reproducible while correlating poorly with functional delivery.
Interpreting distribution studies also requires attention to labels and controls. A fluorescent lipid can exchange into cellular membranes, and a label attached to RNA may persist after degradation. Measurements of labeled material should therefore be distinguished from measurements of intact cargo and biological activity. Preclinical evaluation commonly combines tissue analysis, pharmacokinetics, histology, immune measurements, and functional assays. Human efficacy and safety ultimately require clinical evidence, including appropriately designed randomized controlled trials when suited to the development program. Neither particle uniformity nor high encapsulation efficiency independently establishes clinical performance.
Safety, immune interactions, and component clearance
The biological effects of an LNP arise from the complete formulation rather than from a single ingredient considered in isolation. Relevant variables include lipid chemistry, RNA sequence and purity, dose, administration route, tissue distribution, and frequency of exposure. A formulation intended for vaccination deliberately interacts with immune processes, whereas a formulation intended for repeated protein replacement may seek to minimize inflammatory signaling. Safety assessment must therefore be specific to the product and use; the broad category “lipid nanoparticle” does not define a uniform toxicological profile.
Lipids can influence inflammation and innate immune activation. Permanently cationic formulations may interact strongly with negatively charged biological surfaces, contributing to membrane perturbation and toxicity. Ionizable lipids were developed partly to separate efficient nucleic-acid association under acidic conditions from persistent positive charge in circulation. This design reduces certain undesirable interactions but does not make all ionizable lipids biologically equivalent. Headgroup structure, hydrophobic tails, impurities, degradation products, and tissue exposure can all affect tolerability.
The payload also contributes to immune responses. Cells possess receptors that recognize foreign or unusually structured RNA. RNA sequence design, nucleoside modification, and purification can alter this recognition. Double-stranded RNA contaminants generated during manufacturing are particularly relevant to innate immune activation and can interfere with translation. Consequently, an inflammatory response after administration cannot automatically be assigned to the carrier, the RNA, or the encoded protein without experiments that distinguish their contributions. Product evaluation considers their combined effects.
Interactions with the complement system and pre-existing immune recognition of formulation components are additional considerations. Infusion-related reactions are recognized for some intravenously administered lipid medicines. Antibodies against PEG have been detected in human populations and can affect the behavior of some PEG-containing products. Their clinical significance varies with formulation, administration conditions, and exposure history. Rare serious allergic reactions have also been documented for particular medicines and vaccines, but identifying the responsible molecular mechanism can be difficult. Such observations do not establish that all PEG-containing particles cause the same reactions.
Clearance involves disassembly, exchange of components with biological membranes and lipoproteins, and chemical degradation. Lipids containing suitable ester linkages can undergo enzymatic hydrolysis; the resulting products then enter additional clearance or metabolic pathways. The rate depends on molecular structure and where the component accumulates. RNA is ultimately degraded by cellular nucleases, while proteins translated from delivered mRNA follow their own turnover pathways. A persistent biological effect, especially after gene editing or immune activation, should therefore not be interpreted as proof that the original nanoparticle remains intact for the same duration.
Stability, quality control, and platform limits
An LNP medicine must remain functional during storage and distribution, not merely immediately after manufacture. RNA can degrade through hydrolysis and other chemical reactions, while lipid components may undergo oxidation or hydrolysis. Particle aggregation, fusion, leakage, or redistribution of components can alter size and activity. Interactions between reactive lipid impurities or degradation products and RNA can also impair the payload. Stability is therefore a combination of chemical integrity, colloidal behavior, structural preservation, and retained biological potency.
Temperature control slows many degradation processes, but freezing introduces its own stresses. Ice formation concentrates salts and other solutes in the remaining liquid, changes local conditions, and can bring particles into closer contact. Cryoprotective excipients such as sucrose help reduce damage in suitable formulations. Repeated freezing and thawing can differ from continuous frozen storage in its effects. Storage specifications are established for individual products through testing; the requirements of one LNP formulation cannot be inferred reliably from those of another.
The cold chain became especially visible during the distribution of mRNA vaccines. Differences in RNA design, lipid chemistry, buffer systems, concentration, and container presentation contribute to different storage conditions and allowable handling periods. These specifications can change when additional stability data support updated regulatory labeling. Freeze-drying is being investigated as a way to improve distribution and shelf life, but removing and restoring water can damage particle structure or RNA. A dried product must demonstrate acceptable reconstitution, size distribution, chemical integrity, and potency rather than merely appearing physically intact.
Under good manufacturing practice, quality control links raw materials, process parameters, analytical methods, and release specifications. Relevant attributes can include lipid identity and purity, lipid ratios, RNA identity and integrity, particle size distribution, encapsulation, potency, residual solvent, endotoxin, sterility, and subvisible particles. Not every attribute is equally informative for every formulation. Changes in a lipid supplier, mixer, production site, or batch scale may require comparability studies because apparently small changes can affect assembly or biological activity.
The platform’s modularity has limits. Replacing one RNA sequence with another can preserve much of a manufacturing process, but changes in RNA length, structure, concentration, or biological function may alter particle formation and performance. Likewise, replacing an ionizable lipid can change clearance, immune responses, and tissue targeting even when particle diameter remains constant. Important unresolved technical problems include efficient delivery beyond the liver, reproducible targeting of defined cell populations, improved endosomal escape, and maintenance of potency through repeated administration. These problems are coupled: a modification that improves one step may worsen another, making optimization dependent on the complete formulation and its intended biological task.