Technology

The hard part of RNA medicine is not the RNA

It is getting it where it needs to go.

A short interfering RNA is a precise instrument. Give a cell the right twenty-one letters and it will stop making one specific protein, and nothing else. The difficulty has never been the instruction. It is that a bare strand of RNA injected into the bloodstream is destroyed within minutes, and that the cells most worth reaching are the ones best equipped to refuse it.

Immune cells are the hardest of all. A macrophage carries an arsenal of sensors, including several toll-like receptors, whose entire job is to detect foreign genetic material and raise an alarm when they find it. Delivering RNA into a macrophage means getting past a system designed by evolution to stop exactly that.

Carrier

A carrier that assembles itself.

Our carrier is built from a short, branched polyethyleneimine of about 1.2 kilodaltons with lipids attached along its length. Mixed with siRNA in buffer, it self-assembles into a particle of roughly 180 nanometres carrying a positive surface charge of about 25 millivolts. The whole formulation takes a vortex and thirty minutes at room temperature. There is no microfluidic step and no specialised equipment.

  1. 1

    Holds on through the bloodstream

    Grips the cargo tightly enough to survive the journey, where bare genetic material is destroyed within minutes.

  2. 2

    Gets inside the hardest cells

    A positive surface pulls it through membranes that turn most carriers away. Immune cells exist to reject exactly this cargo.

  3. 3

    Escapes the compartment

    Once swallowed, most carriers stay trapped and are digested. This one is built to get out.

  4. 4

    Lets go on arrival

    Tuned to release its cargo rather than hold on, so more of each dose reaches the place it can work.

Four things the carrier has to do, described without the laboratory units. The measurements sit in the table below. (RSC Pharmaceutics 2026 and Adv Healthcare Mater 2025;14:e02019)

Limits

Three reasons a fatty carrier is the wrong vehicle for this drug.

Lipid nanoparticles are a genuine achievement and we do not dispute their record. They carried the first approved siRNA drug and two coronavirus vaccines. For this particular job, three of their properties work against us.
  1. One

    They go to the liver.

    A lipid nanoparticle entering the bloodstream acquires a coat of apolipoprotein E, which the low-density lipoprotein receptor on liver cells recognises and pulls in. That mechanism is the reason liver-directed RNA medicine works, and it is also a cage. All six approved siRNA drugs act on the liver. Immune cells do not carry the receptor that makes the mechanism work.

  2. Two

    They raise the very signal we are trying to lower.

    Work published in Nature Immunology in 2022 showed that the inflammatory response to RNA lipid particles depends on the same inflammasome we are targeting, and that blocking the resulting signal prevented the downstream release of other inflammatory messengers. For most applications this is a tolerable side effect. For a drug whose entire purpose is to reduce that signal, it is a contradiction at the centre of the design.

    (Nature Immunology 2022;23:532–542)

  3. Three

    Very little of the cargo gets out of the endosome.

    Even where a particle is taken up, most of its cargo never reaches the part of the cell where it can act. Quantitative imaging published in Nature Biotechnology in 2013 put endosomal escape at 1.34 per cent in cultured cells and 1.66 per cent in liver cells in a living animal. More recent single-vesicle work published in Nature Communications in 2025 found release to be incomplete and confined to a subset of damaged compartments. The dose a patient receives and the dose that works are separated by close to two orders of magnitude.

    (Nature Biotechnology 2013 · Nature Communications 2025;16:5354)

Conventional lipid nanoparticle

carries the drug inside
trips the same alarm on the way
pro-inflammatory
cytokines rise

A drug built to lower inflammatory signals cannot arrive in a particle that raises them.

Cytokine release driven by RNA lipid particles depends on the same alarm our therapy is designed to quiet.

Nature Immunology 2022;23:532–542

Delmune lipopolymer · the signals fall instead

The main inflammatory messengers stay essentially unchanged in human blood cells.

The contradiction: the argument this company rests on, in one line. (Nature Immunology 2022;23:532–542)

Comparison

Two carriers compared.

Conventional lipid nanoparticle compared with the Delmune lipopolymer across six properties.
PropertyConventional lipid nanoparticleDelmune lipopolymer
How it is madeMicrofluidic mixing on specialised equipment. Batch-to-batch variation is a recognised risk when scaling up.Mixed by hand in half an hour at room temperature. Reproducible at any bench.
Where it ends upMostly the liver. Blood proteins coat it and a liver receptor pulls it in.Concentrates where immune cells are dense. The lung receives four times the liver dose.
Getting into immune cellsPoor. These cells exist to detect and reject foreign genetic material.Built for them. Roughly six times the uptake of a leading commercial reagent.
Effect on inflammationCan itself trigger the release of inflammatory signals, through the very alarm the drug must quiet.Leaves the main inflammatory messengers essentially unchanged in human blood cells.
Releasing the cargoOnly a small fraction of what enters a cell ever reaches the place it can work.Tuned to let go on arrival rather than hold on.
TolerabilityReactogenicity limits dosing in several clinical settings.Liver, kidney and blood chemistry unchanged after repeat dosing in animals.
Written for a reader who is not a bench scientist. The underlying measurements are on the publications page. (RSC Pharmaceutics 2026 · Adv Healthcare Mater 2025;14:e02019 · ACS Pharmacol Transl Sci 2024;7:2840 · Nat Immunol 2022;23:532–542)

Breadth

siRNA today. Other payloads as the direction of travel.

Today · published

siRNA

Short interfering RNA delivered into myeloid and lymphoid cells, with silencing of the inflammasome reported in primary human cells.

Direction · planned

mRNA

Larger messages that instruct a cell to make a protein. A stated direction for the platform, not a result.

Direction · planned

Gene editing

Editing machinery delivered to immune cells. A stated direction for the platform, not a result.

Manufacturing

Vortex, wait thirty minutes, done.

Every step that needs specialised equipment is a step that can vary between batches, between sites and between operators. Our formulation needs a vortex and a timer. That matters for cost, for reproducibility at any bench, and for the conversation with a manufacturing partner once the program scales.

Could the carrier take your payload into cells you cannot reach?