Lipid Nanoparticles for mRNA Delivery: How They Work

Lipid nanoparticles, or LNPs, work by wrapping fragile mRNA molecules inside a tiny shell of specially designed fats, sneaking them past cell membranes, and releasing them into the cell’s interior where the mRNA can be read and translated into protein. The shell is built from four lipid components, each with a specific job, and the whole system relies on a pH-sensitive trick that keeps the particle stable in the bloodstream but reactive once it reaches the acidic interior of a cell’s recycling compartment. This delivery method powered the Pfizer-BioNTech and Moderna COVID-19 vaccines, and it is now being developed for cancer immunotherapies, rare genetic diseases, and even antibody production directly inside the body.

The Four Ingredients

Every LNP formulation is built from the same four categories of lipid: an ionizable cationic lipid, a helper phospholipid, cholesterol, and a PEG-lipid (a lipid with a polyethylene glycol chain attached).¹ Each component pulls its weight differently. The ionizable lipid is the star performer, responsible for binding the mRNA and later helping it escape from inside the cell. The helper phospholipid supports the structural integrity of the particle and influences how proteins in the blood interact with it. Cholesterol fills gaps between lipid molecules, stiffening the particle and keeping the mRNA locked inside during storage and circulation. The PEG-lipid coats the outer surface like a molecular raincoat, preventing particles from clumping together or being gobbled up by immune cells before they reach their target.²

The ratios matter. Too little cholesterol and the particle falls apart in storage; one study found that LNPs with only 10 mol% cholesterol dropped to about 67% encapsulation efficiency after just one week, while formulations with more cholesterol stayed above 75%.³ Too much PEG-lipid and the particle circulates well but struggles to fuse with cell membranes. The formulation is a balancing act, and even small changes to any of the four components can redirect where the particle ends up in the body or how much protein it produces.

The pH Switch

The ionizable lipid is engineered to change its electrical charge depending on the acidity of its surroundings. At the near-neutral pH of your bloodstream (around 7.4), the lipid carries no net charge, keeping the LNP relatively inert and invisible to the immune system. But when the pH drops below about 6.5, the lipid picks up a positive charge. This shift is the core mechanism that makes LNP delivery work, because it governs two critical steps: loading the mRNA during manufacturing and releasing it once the particle is inside a cell.

During manufacturing, the mRNA and lipids are mixed at a low pH where the ionizable lipid is positively charged. The negatively charged mRNA is electrostatically attracted to the lipid, and the two bind tightly together as the particle assembles. Then the pH is raised toward neutral, the lipid loses its charge, and the particle closes up around the mRNA cargo. Molecular simulations show that at low pH, mRNA sits close to the lipid headgroups, but at neutral and intermediate pH, the interaction becomes repulsive and the mRNA redistributes deeper into the aqueous layers of the particle.⁴

Not all ionizable lipids are equal in this regard. Synchrotron X-ray studies have revealed that different ionizable lipids undergo distinct structural transitions as pH drops, forming different ordered phases. The specific transition from one phase to another at around pH 6 appears to be a hallmark of how well a given lipid performs in LNPs, with measurable differences in spacing and hydration between lipids like DLin-MC3-DMA (the lipid used in Onpattro, the first FDA-approved LNP drug) and newer candidates.⁵ Researchers have also shown that the internal repeat distances within LNPs shift with pH in a pattern that mirrors the lipid’s protonation curve, directly linking the particle’s nanoscale structure to the pH responsiveness of its ionizable lipid.⁶

How LNPs Assemble

LNPs are not built piece by piece under a microscope. They form spontaneously through self-assembly when the lipid mixture (dissolved in ethanol) is rapidly mixed with the mRNA (dissolved in an acidic water-based buffer). Simulations of this process suggest it unfolds in three stages: individual lipids first cluster into small disc-like patches, those patches merge into larger membrane sheets, and the sheets then close into spherical vesicles that trap the mRNA inside.⁷ If you simply stirred the components together gently, diffusion alone would take days to produce usable particles. Instead, manufacturers use specialized microfluidic mixers that slam the two streams together in milliseconds, producing uniform particles typically in the range of 60 to 100 nanometers across.

The speed and geometry of mixing affect particle size, uniformity, and how efficiently the mRNA gets encapsulated. This is one reason why scaling up LNP production from a lab bench to a factory floor is a genuine engineering challenge, not just a matter of making bigger batches.

Hitching a Ride to the Liver

Once injected into the bloodstream, LNPs quickly become coated with proteins from the blood. The most important of these is apolipoprotein E, or ApoE, a protein your body normally uses to shuttle cholesterol around. ApoE binds to the surface of the LNP and, in effect, disguises it as a natural cholesterol-carrying particle. Liver cells are studded with receptors that recognize ApoE, so the coated LNPs are efficiently taken up by the liver through receptor-mediated endocytosis.⁸

The choice of helper lipid influences this process. Unsaturated helper lipids like DOPE increase membrane fluidity, which appears to enhance ApoE adsorption compared to the saturated helper lipid DSPC.⁹ Helper lipid properties like headgroup chemistry, chain length, and degree of saturation all modulate how the particle’s membrane behaves and how stable it remains.¹⁰ One liver-targeting helper lipid called DEPE has been shown to modulate both LNP structure and ApoE binding to enhance liver-specific delivery.¹¹

Recent work with cryo-electron microscopy has shown that the shape of the LNP surface changes when ApoE binds. Replacing a fraction of the standard DSPC helper lipid with a porphyrin-lipid conjugate, for instance, produced highly faceted, irregular membrane shapes upon ApoE binding, which altered how cells took up the particles and actually enhanced the release of cargo inside the cell.¹² The surface geometry of the particle, in other words, is not just cosmetic. It feeds directly into how efficiently the mRNA gets delivered.

Breaking Out of the Endosome

Getting inside a cell is only half the battle. When LNPs are taken up by receptor-mediated endocytosis, they end up trapped inside membrane-bound compartments called endosomes. If nothing else happens, the endosome matures, fuses with a lysosome, and the mRNA is destroyed. The whole point of the ionizable lipid is to prevent this by exploiting the endosome’s gradually dropping pH.

As the endosome acidifies, the ionizable lipid becomes positively charged. This charge interacts with the negatively charged phospholipids in the endosomal membrane, destabilizing it and allowing the mRNA to escape into the cell’s cytoplasm, where ribosomes can translate it into protein. A recent re-evaluation of the evidence argues that the dominant escape mechanism is a process called vesicle budding-and-collapse, where the LNP causes small buds to form on the endosomal membrane that then rupture.¹³ That same analysis identified a second bottleneck: after escape, the lipid and mRNA form an insoluble aggregate in the cytoplasm that must slowly dissolve before the mRNA becomes available for translation.

This bottleneck has drawn increasing attention. Research measuring intracellular mRNA dissociation directly has found that the separation of mRNA from the ionizable lipid inside the cell, not just getting out of the endosome, is a key rate-limiting step. Adjusting the amount and type of ionizable lipid can substantially improve how much mRNA becomes functionally available.¹⁴

Different ionizable lipids show strikingly different escape behavior. Using a fluorescence-based imaging platform, researchers found that LNPs made with the lipid C12-200 showed rapid signal diffusion consistent with efficient endosomal release, while LNPs made with MC3 accumulated in endosomes with limited escape.¹⁵ This helps explain why lipid chemistry matters so much: two LNPs that look identical in size and encapsulation efficiency can perform very differently inside a cell depending on which ionizable lipid they contain.

Making the mRNA Invisible to the Immune System

Unmodified mRNA triggers alarm bells inside cells. Your immune system has evolved sensors, particularly toll-like receptors in the endosome, that detect foreign RNA and mount an inflammatory response. This would both destroy the mRNA before it could be translated and cause unwanted side effects. The solution was a breakthrough that earned Katalin Karikó and Drew Weissman the 2023 Nobel Prize in Physiology or Medicine: swapping one of the natural building blocks of RNA for a chemically modified version.

The original modification, pseudouridine, replaced the natural uridine nucleoside. mRNA containing pseudouridine showed higher protein production and dramatically lower immune activation when tested in cells and in mice. Even at higher doses, the modified mRNA did not trigger the flood of interferon-alpha that unmodified mRNA provoked.¹⁶ A further refinement, N1-methylpseudouridine, outperformed pseudouridine by producing up to roughly 13-fold higher protein expression in mice and further reduced innate immune signaling, partly by evading detection by endosomal toll-like receptor 3.¹⁷ Both COVID-19 mRNA vaccines use N1-methylpseudouridine in their mRNA.

Immune Activation From the Particle Itself

Even with modified mRNA, the LNP shell is not immunologically silent. In human whole-blood assays, mRNA-LNPs triggered a robust release of pro-inflammatory cytokines including TNF-α, IL-1β, IL-6, and others.¹⁸ This is partly why mRNA vaccines produce the sore arms, fevers, and fatigue that many people experience: the LNP itself is acting as an adjuvant, priming the immune system alongside the antigen the mRNA encodes.

Interestingly, the lipid nanoparticle and the mRNA each activate different parts of the immune system. Studies comparing empty LNPs (no mRNA inside) with mRNA-loaded LNPs found that the lipid component primarily activated monocytes and upregulated antigen-presentation molecules, while the mRNA component was responsible for activating B cells and a specific subset of dendritic cells.¹⁹ For vaccines, this dual activation is a feature. For therapeutic applications where you just want the protein produced without inflammation, it is a problem that researchers are actively trying to solve.

The PEG Problem

The PEG coating that protects LNPs during circulation comes with a downside. Repeated dosing of PEGylated drugs can lead to the formation of anti-PEG antibodies, which accelerate clearance of subsequent doses from the blood, reduce how well the drug works, and can trigger allergic reactions.²⁰ Even more concerning, pre-existing anti-PEG antibodies have been detected in people who have never received a PEGylated drug, likely because PEG is found in many consumer products from cosmetics to laxatives. For a vaccine given once or twice, this is a manageable concern. For therapies that require repeated infusions over months or years, such as mRNA-encoded antibodies for cancer, it becomes a serious engineering challenge.

Redirecting LNPs Beyond the Liver

The liver tropism of standard LNPs is useful for diseases like hepatic enzyme deficiencies but useless for conditions affecting the lungs, spleen, or brain. One of the most promising approaches to redirecting LNPs is called selective organ targeting, or SORT. The idea is simple in principle: add a fifth lipid, the SORT molecule, whose chemical properties determine which organ the particle ends up in. By varying the SORT molecule’s charge and structure, researchers have engineered LNPs that selectively deliver mRNA to the lungs, spleen, or liver, and have demonstrated gene editing in specific cell types within those organs, including epithelial cells, endothelial cells, B cells, T cells, and hepatocytes.²¹

The mechanism behind SORT appears to involve changing which blood proteins adsorb onto the particle surface, which in turn changes which cell-surface receptors the particle engages. Adding a permanently cationic SORT molecule, for example, shifted protein adsorption away from ApoE and toward other proteins, redirecting the particles from liver to lung.²² This protein-corona-based targeting represents a fundamentally different strategy from traditional approaches that attach targeting ligands to the particle surface, and it has opened the door to applications like inhaled mRNA for lung disease and systemic mRNA delivery to immune cells.

Storage and Stability

One of the practical headaches with mRNA-LNPs is that they are fragile. The mRNA can degrade through hydrolysis or oxidation, and the lipid shell can aggregate or lose its structure over time. The Pfizer-BioNTech vaccine famously required ultra-cold storage at around minus 70°C, though formulation improvements later relaxed that requirement. A large-scale analysis of over 600 LNP formulations found that the critical threshold for long-term storage is around 90 days: beyond that duration, median biological activity retention dropped below 80%, and lyophilization (freeze-drying) became protective only after that point.²³ For short-term storage, liquid formulations held at refrigerator temperature can work, but the push toward room-temperature-stable LNPs remains a major focus of formulation science.

Cholesterol content plays a direct role in storage stability. As noted earlier, LNPs with too little cholesterol showed both a drop in mRNA encapsulation and an increase in particle size within a week, while higher-cholesterol formulations held steady.³ The interplay between lipid composition and shelf life is one of the less glamorous but most practically important areas of LNP research, especially for distributing mRNA therapies in regions without reliable cold chains.

Next-Generation Ionizable Lipids

The ionizable lipids used in approved products, like SM-102 (Moderna) and ALC-0315 (Pfizer-BioNTech), work well enough for vaccines but are not necessarily optimal for every application. One concern is that some ionizable lipids linger in the body longer than desired. Newer lipid designs incorporate built-in degradation points, such as ester bonds or disulfide bridges, that allow the body to break down the lipid more quickly after it has done its job.

Several research groups have developed modular chemistry platforms to rapidly generate large libraries of biodegradable ionizable lipids. One approach introduced a disulfide bond-bridged ester linker to create 96 degradable lipid candidates in a single campaign.²⁴ Another used the Passerini chemical reaction to quickly assemble chemically diverse biodegradable lipids from simple starting materials.²⁵ A third devised a plug-and-play system combining amines or thiols with dialkyl maleates to build structurally diverse biodegradable lipids.²⁶ The common thread is speed and variety: instead of optimizing a single lipid over years, researchers can now screen dozens or hundreds of candidates in parallel, looking for the sweet spot between potency, tolerability, and clearance.

From Mice to Humans

One persistent challenge in LNP development is that a formulation that works brilliantly in mice does not always translate to the same performance in humans. Body size, liver physiology, blood protein composition, and immune system differences all affect how LNPs behave. Researchers developing mRNA-encoded antibodies (where the LNP delivers mRNA that instructs the body’s own cells to manufacture therapeutic antibodies) have begun building allometric scaling frameworks to predict human pharmacokinetics from animal data. For two clinical-stage mRNA-LNP antibody programs, dose-normalized antibody exposure in mice predicted human levels within roughly four-fold, while nonhuman primate data predicted within about two-fold.²⁷ Those margins are encouraging but still wide enough that clinical trials remain essential. The field is getting better at predicting cross-species translation, but LNP pharmacology is complicated enough that surprises still happen.

The interplay between the helper lipid, the ionizable lipid, the mRNA cargo, and the patient’s own blood proteins creates a system with more moving parts than most drug-delivery platforms. Changing the helper lipid can alter not just LNP structure but also the immune cells that respond to the particle and the organs where the mRNA ends up being expressed.²⁸ That complexity is why no single LNP formulation dominates across all applications. The vaccines were the proof of concept. The next decade of work is about tuning each knob in the system for specific diseases, specific organs, and the practical demands of manufacturing and distribution at scale.