# Lipid Nanoparticle Endosomal Escape: The Numbers Never Measured

> A plain-English guide to lipid nanoparticle endosomal escape, the four steps of mRNA delivery, and which numbers were never measured before rollout.

Canonical: https://amyloid-test.com/resources/lipid-nanoparticle-endosomal-escape/
Last updated: 2026-11-19
Author: Tim Ottowitz
Evidence level: preclinical

**Short answer:** Endosomal escape is the step where mRNA inside a lipid nanoparticle breaks out of the bubble a cell swallowed it in and reaches the cell's protein factory. Published work shows this step is rare and poorly quantified, so most particles are internalised without ever producing protein, and nobody can say exactly how many do.

## Key takeaways

- Arriving in a tissue, entering a cell, escaping the endosome, and making protein are four different events, and only the first two are routinely measured.
- Ionizable lipids differ: ALC-0315 and SM-102 carry different pharmacokinetics, so the two mRNA vaccines are not biologically interchangeable.
- Liver uptake is driven by ApoE binding and LDL receptors plus the fenestrated sinusoid lining, which is why so much injected material ends up in hepatocytes.
- After intramuscular injection in mice, 10.98% of splenic macrophages, 17.46% of dendritic cells and 4.08% of B cells expressed a reporter protein, so activity is not limited to the injection site.
- A 2019 study found that lipid nanoparticle mRNA can be repackaged into extracellular vesicles and carried to other cells, which initial biodistribution studies cannot capture.

A French thread by a physicians' collective went around in November 2026 asking a simple question about mRNA shots: how much of the injected material actually does anything, and where did the rest go? That question has a real scientific answer, and it starts with a step most people have never heard of. This guide walks through the four stages a lipid nanoparticle has to clear, explains what endosomal escape is, and shows which of those stages have hard numbers behind them and which do not.

If you want the background first, our guides to [lipid nanoparticle biodistribution](https://amyloid-test.com/resources/lipid-nanoparticle-biodistribution/) and [human blood distribution of mRNA and SM-102](https://amyloid-test.com/resources/lipid-nanoparticle-blood-distribution/) cover where the particles travel. This article is about what happens after they arrive.

## What is a lipid nanoparticle, in plain terms?

Messenger RNA (mRNA) is a set of instructions for building one protein. On its own it falls apart in minutes, because the body is full of enzymes that chop up loose RNA.

So the instructions get packed into a tiny ball of fats called a lipid nanoparticle (LNP). Picture a greasy bubble about a thousand times thinner than a hair, with the mRNA curled up inside. The bubble protects the message and helps it slip into a cell.

Four lipid types make up that bubble. The important one is the **ionizable lipid**, which changes its electrical charge depending on how acidic its surroundings are. Pfizer's shot uses ALC-0315 and Moderna's uses SM-102. A 2026 *Biophysical Journal* paper showed that the surrounding membrane itself shifts the point at which these lipids flip their charge, so their behaviour inside a cell isn't a fixed property of the molecule.

![Four narrowing stages of mRNA delivery: biodistribution, internalisation, endosomal escape and protein production](https://amyloid-test.com/resource-images/lnp-delivery-funnel-stages.jpg)

## The four steps, and why mixing them up matters

Here's the part that trips up most reporting on this topic. Getting mRNA to work in a cell isn't one event. It's four, and each one loses most of what the previous one delivered.

1. **Biodistribution.** The particle arrives somewhere in the body.
2. **Internalisation.** A cell swallows it.
3. **Endosomal escape.** The mRNA breaks out of the bubble the cell wrapped it in.
4. **Functional transfection.** A ribosome reads the mRNA and builds the protein.

Think of a parcel. Reaching the right city isn't the same as reaching the right house. Getting through the front door isn't the same as getting out of the locked porch. And none of that is the same as somebody opening the box and following the instructions inside.

| Step | What it means | How it's usually measured | What's still thin |
| :--- | :--- | :--- | :--- |
| Biodistribution | The particle reaches a tissue | Radiolabels, whole-body imaging, PCR of lipids or mRNA in organs | Human data over weeks rather than days, covered in [LNP blood distribution](https://amyloid-test.com/resources/lipid-nanoparticle-blood-distribution/) |
| Internalisation | A cell takes the particle in | Fluorescent lipid tags, microscopy, flow cytometry | How many particles per cell, and how many were still carrying cargo |
| Endosomal escape | mRNA reaches the cytosol | Specialised reporter assays in cell culture | Almost no organ-level numbers in living animals, let alone humans |
| Functional transfection | Protein actually gets made | Reporter proteins, antibodies against the protein | Which cell types, for how long, and after repeat doses |

The fourth column is the whole argument. Steps one and two are measured routinely. Step three, the one that decides whether anything happens at all, is the hardest to quantify and the least often reported.

## What is endosomal escape?

When a cell takes something in from outside, it doesn't just let it float around. It pinches off a piece of its own outer membrane and wraps the visitor in a bubble called an **endosome**. From the cell's point of view, that's a holding cell. The usual next step is to pump acid into it and fuse it with a lysosome, the cell's recycling bin, where the contents get taken apart.

For an mRNA shot to work, the message has to get out of that bubble before the bin opens. That's endosomal escape. The ionizable lipid is what makes it possible. In the blood it's close to neutral, so it doesn't stick to everything. Inside the acidifying endosome it picks up positive charge, grabs the bubble's negatively charged inner wall, and destabilises it enough for some mRNA to spill into the cytosol.

Some. Not all, and not most. Once the mRNA is in the cytosol, ribosomes can read it, and the cell starts producing the encoded protein. Everything still inside the bubble gets degraded.

A 2026 *ACS Nano* review of how these particles cross biological barriers makes a related point worth keeping in mind. Barriers differ. The liver's blood vessels are leaky by design, so particles reach liver cells easily. The blood-brain barrier needs active transport mechanisms such as transcytosis, so leakiness in one place tells you nothing about crossing in another. If you see a claim that liver uptake proves brain uptake, that step doesn't follow.

{{CTA:mid_article}}

## Why does so much end up in the liver?

Two reasons, and both are structural rather than sinister.

First, the liver's smallest vessels, the sinusoids, have fenestrations: openings in the vessel wall that let things pass straight through to the liver cells behind them. Most capillaries elsewhere don't have those.

Second, a circulating nanoparticle doesn't stay bare. Blood proteins stick to it within seconds and form a coat, and one of those proteins is apolipoprotein E (ApoE), which the body uses to route fat around. LDL receptors on liver cells recognise ApoE and pull the whole package in. A 2025 review in *Molecular Therapy Methods and Clinical Development* walks through both routes in detail.

So liver tropism isn't an accident of dosing. It's built into how the particle meets the bloodstream.

## Do the two vaccine formulations behave the same way?

No, and a 2025 study in the *Journal of Controlled Release* measured the difference. Comparing ionizable lipid types after intravenous administration, the researchers found different pharmacokinetic and biodistribution profiles depending on which lipid was used, with protein expression concentrated in the liver.

That matters for how the two products are discussed. ALC-0315 and SM-102 are often treated as minor formulation details, as though the mRNA sequence were the only variable that counts. The lipid choice changes how much of the body is exposed, for how long, and where the protein gets made. Our review of [lipid nanoparticle membrane dysfunction](https://amyloid-test.com/resources/lipid-nanoparticle-membrane-dysfunction/) covers what the lipids do to cell membranes once they arrive.

![Lipid nanoparticles passing through fenestrated liver sinusoid pores on one side, and glowing macrophage, dendritic and B cells in lymphoid tissue on the other](https://amyloid-test.com/resource-images/lnp-liver-spleen-uptake.jpg)

## Which cells actually read the instructions?

This is where a 2024 *ACS Nano* paper on the inflammatory response to intramuscular lipid nanoparticles is useful. The team injected Cre-mRNA packaged in SM-102 particles into muscle and then looked at the spleen, using a reporter system that lights up only in cells that produced the protein.

They found reporter-positive cells well away from the injection site:

- Macrophages: 10.98%
- Dendritic cells: 17.46%
- B cells: 4.08%

Two things follow. First, functional protein production isn't confined to the shoulder. Second, note what those percentages are and aren't. They count cells that made the reporter protein. They do not tell you how many particles arrived, how many still carried intact mRNA, how many were internalised, or how many mRNA molecules escaped the endosome in each cell. The step that the whole process hinges on remains unquantified even in a study that measured the outcome carefully.

## Is the carrier biologically inert?

No, and researchers don't claim it is. A 2025 *FASEB Journal* review describes the lipid nanoparticle's own adjuvant properties: the lipids help activate innate immunity, which is part of why the platform produces strong responses.

Innate immune cells also appear to carry changes beyond the acute phase. A 2022 *JCI Insight* study reported short-term epigenetic memory in innate immune cells after consecutive BNT162b2 doses, and a 2025 *Molecular Systems Biology* paper described persistent epigenetic memory in monocyte-derived macrophages after mRNA vaccination. Neither study establishes clinical harm. Both make it hard to treat the carrier as a wrapper that disappears without trace.

A 2026 *Nature Communications* paper went further on one specific question, reporting that mitochondrial vulnerability underlies myocarditis after COVID-19 mRNA vaccination. That's a mechanism for a recognised and rare adverse event, which is exactly the kind of work that resolves an open question rather than leaving it open.

## Can the material move on to other cells?

Possibly, and this is the finding that undercuts the idea that initial biodistribution settles the matter.

A 2019 *Nature Communications* study found a direct link between endosomal escape and loading into extracellular vesicles. Extracellular vesicles are small packets cells release to pass material to their neighbours. The researchers showed that lipid nanoparticle mRNA which escaped the endosome could end up in these vesicles and be transported to other cells, where it was still functional.

So a map of where particles land in the first 48 hours doesn't necessarily describe where the cargo ends up. The same logic shows up elsewhere in this field: our guide to [how SARS-CoV-2 may move between cells](https://amyloid-test.com/resources/sars-cov-2-cell-to-cell-spread/) covers a parallel route for viral material.

## What would close the gaps?

Reading the open questions honestly, the studies that would answer them are specific and doable:

1. Quantitative biodistribution over time, not just at one early timepoint.
2. The fate of the lipids themselves, separately from the mRNA, organ by organ.
3. Direct measurement of endosomal escape rates in living tissue.
4. Which cell populations are transfected outside the injection site, and for how long.
5. Persistence and clearance of each component.
6. Immune effects, including after repeat doses.
7. Longer follow-up for any effects that take time to appear.

None of these is an exotic request. They're the standard characterisation questions for a delivery platform, and the bar reasonably rises with the number of healthy people receiving it.

## What this does and doesn't prove

Unanswered isn't the same as dangerous. Three categories deserve to stay separate:

- **Demonstrated experimentally.** Liver tropism via ApoE. Splenic immune cell transfection in mice. Extracellular vesicle transfer in cell models. Adjuvant activity of the lipids.
- **Biologically plausible but under-characterised.** Long-term redistribution of lipid components. The clinical weight of the epigenetic findings. Escape rates in human tissue.
- **Open questions needing work.** Cumulative effects of repeat dosing. Component persistence past the first weeks.

Most of the evidence above is preclinical, in mice or in cell culture. Mouse circulation, dose per body weight and immune responses all differ from ours. Absence of data is a reason to ask for studies, not a result in itself. Collapsing those three categories into one is how both overclaiming and dismissal happen.

## Where microclot testing fits

None of this is measured by a microclot test, and it's worth being direct about that.

What the research above does establish is that spike protein gets produced in places beyond the injection site, including inside blood vessel linings. Separately, published work shows spike protein can drive fibrin into an amyloid form that resists normal breakdown, which is the subject of our guide to [spike protein amyloid fibrils and fibrinolysis](https://amyloid-test.com/resources/spike-protein-amyloid-fibrils-fibrinolysis/).

Our [at-home amyloid test](https://amyloid-test.com/) looks for those fibrin amyloid microclots in a dried blood sample from a finger prick. The sample goes to a partner laboratory in Mexico City, where it's stained with Thioflavin T and recorded under blue light with video microscopy, then reviewed with AI-assisted detection and trained human oversight. You get a Videographic Microscopy Report in a secure portal to share with your own doctor.

To be clear about its limits: it does not measure lipid nanoparticles, does not measure circulating spike protein, is not an Alzheimer's test, and is not a diagnostic test. It tells you whether amyloid-positive microclots are present in your sample, and nothing about where they came from.

## What to do next

If you're trying to make sense of symptoms after an infection or an injection, the useful move is to separate mechanism from measurement. The mechanism questions in this article are for researchers to settle. What you can act on is what's measurable in you now.

- Start with [how to test for microclots](https://amyloid-test.com/resources/how-to-test-for-microclots/) for the practical options and what each one shows.
- Read [post-vaccination microclots and fibrinolysis](https://amyloid-test.com/resources/post-vaccination-microclots-fibrinolysis/) for what's been published on clotting after vaccination.
- Bring papers, not screenshots, to your doctor. Every study cited here is listed below with a link.

*This article is educational and does not provide medical advice, diagnosis, or treatment recommendations.*

## FAQ

### What is endosomal escape in simple terms?

When a cell swallows a lipid nanoparticle, it wraps it in a bubble called an endosome. That bubble is a holding cell, not the factory floor. For anything to happen, the mRNA has to break out of the bubble and reach the cytosol, where ribosomes can read it. That break-out is endosomal escape. Most particles never manage it, and they are degraded inside the bubble instead.

### Does internalisation mean the cell made the protein?

No, and this is the most common mix-up. A cell can take up a particle, keep it in an endosome, and never translate a single mRNA molecule. Studies that count particles in a tissue or label cells as positive measure arrival or uptake, not functional transfection. Protein production has to be shown separately, with a reporter protein or an antibody against the finished protein.

### Why do lipid nanoparticles concentrate in the liver?

Two features push them there. The liver's sinusoid vessels have fenestrations, small openings that let particles leave the bloodstream and reach hepatocytes directly. And once lipid nanoparticles circulate, blood proteins coat them, including apolipoprotein E, which LDL receptors on liver cells recognise and pull in. A 2025 review in Molecular Therapy Methods and Clinical Development describes both routes in detail.

### Are lipid nanoparticles immunologically inert?

No. The lipid components themselves activate innate immune signalling, which is part of why the formulations work as well as they do. A 2025 FASEB Journal review describes the carrier's adjuvant properties directly. Separate work reports changes in innate immune cells after mRNA injection that last beyond the acute response, so the carrier is biologically active rather than a passive wrapper.

### Does the Amyloid Test kit detect lipid nanoparticles?

No. The Amyloid Test kit is an at-home finger-prick test that looks for fibrin amyloid microclots in a dried blood sample using Thioflavin T fluorescence microscopy. It does not measure lipid nanoparticles, does not measure circulating spike protein, and is not a diagnostic test. It gives observational laboratory data you can bring to a clinician.

## Sources

- [Why do lipid nanoparticles target the liver? Understanding of biodistribution and liver-specific tropism](https://doi.org/10.1016/j.omtm.2025.101436) (Molecular Therapy Methods and Clinical Development, 2025)
- [Impact of ionizable lipid type on the pharmacokinetics and biodistribution of mRNA-lipid nanoparticles after intravenous administration](https://doi.org/10.1016/j.jconrel.2025.113945) (Journal of Controlled Release, 2025)
- [Distinct Inflammatory Programs Underlie the Intramuscular Lipid Nanoparticle Response](https://doi.org/10.1021/acsnano.4c08490) (ACS Nano, 2024)
- [Linkage between endosomal escape of LNP-mRNA and loading into EVs for transport to other cells](https://doi.org/10.1038/s41467-019-12275-6) (Nature Communications, 2019)
- [Lipid Nanoparticle: Beyond Delivery Vehicle, Unveiling Its Immunological Adjuvant Potential](https://doi.org/10.1096/fj.202500622R) (FASEB Journal, 2025)
- [Persistent epigenetic memory of SARS-CoV-2 mRNA vaccination in monocyte-derived macrophages](https://doi.org/10.1038/s44320-025-00093-6) (Molecular Systems Biology, 2025)
- [Consecutive BNT162b2 mRNA vaccination induces short-term epigenetic memory in innate immune cells](https://doi.org/10.1172/jci.insight.163347) (JCI Insight, 2022)
- [Mechanistic and Strategic Insights into How Lipid Nanoparticles Interact with Membrane-Based Biological Barriers](https://doi.org/10.1021/acsnano.5c21857) (ACS Nano, 2026)
- [Mitochondrial vulnerability underlies myocarditis from COVID-19 mRNA vaccine](https://doi.org/10.1038/s41467-026-71295-1) (Nature Communications, 2026)
- [Membrane environment sets the functional pKa of ionizable lipids](https://doi.org/10.1016/j.bpj.2026.06.030) (Biophysical Journal, 2026)

## Related

- [Spike Protein Test: Antibody, Antigen, and Downstream Testing](https://amyloid-test.com/resources/spike-protein-test/)
- [Spike Protein Amyloid Fibrils: How Spike685 Makes Clots Resist Lysis](https://amyloid-test.com/resources/spike-protein-amyloid-fibrils-fibrinolysis/)
- [Amyloid Fibril Test: Methods Used to Detect Amyloid Fibrils](https://amyloid-test.com/resources/amyloid-fibril-test/)
