Japan.co.jp

For an RNA medicine, choosing the right genetic target is only half the problem. RNA is fragile, water-soluble and highly charged. Left unprotected, it can be degraded before reaching its destination, and it does not readily cross the lipid membrane of a cell.

That is why delivery has become one of the defining engineering problems of RNA therapeutics. Lipid nanoparticles, or LNPs, can protect nucleic acids, carry them through biological fluids, help cells take them up and, if everything works, release the RNA where it can act.

Yet two LNP preparations can have nearly the same particle size, contain similar amounts of siRNA and still perform differently. Researchers at Chiba University and Tohoku University asked whether the missing information lies inside the particle.

The team, led in Chiba by Keisuke Ueda with Yui Sakagawa, Tomoki Saito, Fumie Sakuma, Kenjirou Higashi and Kunikazu Moribe, used solution-state proton NMR together with small-angle X-ray scattering and other nanoscale techniques to examine how the method used to mix siRNA into an LNP changes its molecular organization. The paper was published online August 2, 2024 in the Journal of Controlled Release and appeared in volume 373 that September. Chiba University released a research summary on October 1, 2024.

Particles that looked similar from outside were different inside

The researchers used DLin-MC3-DMA, or MC3, as the ionizable lipid and prepared siRNA-loaded LNPs three ways.

In the pre-mixing method, an acidic siRNA solution was mixed with an ethanolic lipid solution through a microfluidic mixer, followed by dialysis into phosphate-buffered saline at pH 7.4.

For post-mixing, the team first made empty LNPs and then introduced siRNA under acidic conditions. One post-mix condition included ethanol, producing post-mix LNP A; the other omitted ethanol, producing post-mix LNP B.

All three preparations formed particles with an average diameter of about 50 nanometers. The ratio of encapsulated siRNA to lipid content was also broadly constant across preparation methods.

If size and encapsulation were the only important quality attributes, the formulations might have appeared essentially equivalent. They were not.

Pre-mixing produced a more homogeneous core

Small-angle X-ray scattering showed that the ordered siRNA–MC3 stacked bilayer structure differed among the preparations.

Solution-state ¹H NMR went further. The data indicated that siRNA was incorporated comparatively uniformly through the core of pre-mixed LNPs.

Post-mixed particles were more heterogeneous. Their cores contained regions depleted of siRNA alongside areas with locally enriched siRNA. That heterogeneity was most pronounced in post-mix B, where no ethanol had been added. Post-mix A was intermediate, suggesting that ethanol helped lipids and siRNA reorganize and mix more uniformly.

Gene silencing fell in the same order

The biological result tracked the structural result.

Gene-silencing activity was strongest for the pre-mixed LNP, lower for post-mix A and lowest for post-mix B—the same order in which internal siRNA distribution became less homogeneous.

The study therefore supports the idea that molecular-scale heterogeneity inside an LNP can reduce the functional performance of its siRNA cargo.

That does not mean the NMR experiment directly proved every downstream mechanism. Cellular uptake, endosomal escape, intracellular release and RISC loading are separate biological steps. The paper establishes a structural-function relationship, not a complete causal map of every intracellular event.

How NMR can look inside a nanoparticle

Nuclear magnetic resonance measures how atomic nuclei respond in a magnetic field. MRI uses related physical principles at the scale of tissues; pharmaceutical NMR uses them to examine molecular environments and motion.

For LNPs, that matters because molecular mobility can reveal whether a lipid or RNA component is tightly associated, relatively exposed, constrained or located in a different local environment.

Ueda described this problem in a 2026 Chiba University researcher profile as an effort to open what had remained a “black box” inside LNP formulations. Empirical formulation rules can tell researchers that one recipe works better than another. Molecular characterization can begin to explain why.

A 2023 study had already shown that lipid composition changes molecular motion

The mixing study sits in a broader Chiba research program.

In 2023, Ueda and colleagues used ¹H NMR relaxometry to examine how the neutral lipids DSPC and cholesterol alter siRNA-loaded LNPs. DSPC reduced the overall molecular mobility of ionizable lipids, while cholesterol preferentially reduced mobility in their hydrophobic tails, consistent with cholesterol filling space among those tails.

The study connected lipid composition to maintenance of the stacked siRNA–ionizable-lipid structure, encapsulation efficiency and PEG-chain flexibility. It also showed that too much DSPC or cholesterol could deform particles or cause cholesterol demixing.

Together, the 2023 and 2024 papers ask two different formulation questions: what should the particle contain, and how should its ingredients be assembled?

What siRNA actually does

Small interfering RNA is a short double-stranded RNA that harnesses the cell’s RNA-interference machinery to reduce production of a specific protein.

Rather than editing DNA, siRNA targets messenger RNA—the temporary molecular message copied from DNA before protein is made. Guided by sequence complementarity, the RNA-induced silencing complex can cleave the targeted mRNA, reducing expression of that gene.

The modern field traces back to Andrew Fire and Craig Mello’s 1998 experiments in Caenorhabditis elegans, which established that double-stranded RNA can trigger powerful, sequence-specific gene silencing. They received the 2006 Nobel Prize in Physiology or Medicine for the discovery of RNA interference.

The delivery problem almost stopped the therapeutic promise

siRNA can be designed against many genes in principle. In practice, naked RNA is a difficult drug.

It is negatively charged, strongly hydrophilic and unable to cross cell membranes efficiently. It can also be degraded by nucleases in biological fluids.

LNPs solve part of that problem by combining ionizable lipids with helper phospholipids, cholesterol and PEG-lipids. Under acidic manufacturing conditions, ionizable lipids carry positive charge and bind negatively charged RNA. Near physiological pH, they become less charged, reducing some of the toxicity associated with permanently cationic lipids.

In 2018, an LNP helped turn RNA interference into an approved medicine

A major milestone came with patisiran, marketed as Onpattro and approved by the U.S. FDA in 2018 for adults with the polyneuropathy of hereditary transthyretin-mediated amyloidosis.

Patisiran contains an siRNA directed against transthyretin mRNA and is formulated in an ionizable lipid nanoparticle for delivery to hepatocytes. FDA materials describe it as an LNP containing four lipids and administered intravenously.

That approval demonstrated that RNAi was no longer only a molecular-biology tool: with the right delivery system, it could become a medicine.

COVID-19 made lipid nanoparticles familiar to the world

LNPs later became widely recognized through mRNA vaccines. mRNA, like siRNA, is fragile and difficult to move across cell membranes without help.

But the payloads do opposite jobs. Vaccine mRNA instructs cells to produce a protein. siRNA is used to reduce production of a chosen protein by destroying its messenger RNA.

That distinction matters for formulation. Payload length, structure, amount and required release profile differ, meaning an LNP optimized for one RNA may not be optimal for another.

A 50-nanometer size measurement cannot tell the whole quality story

Particle diameter, surface properties and encapsulation efficiency are standard quality attributes in nanomedicine.

The Chiba study demonstrates why those macroscopic measurements may be insufficient. All three formulations were roughly 50 nm and carried similar siRNA-to-lipid ratios, yet they did not organize siRNA the same way and did not silence genes equally well.

That raises an important manufacturing possibility: molecular-level NMR signatures may eventually become useful in formulation development or quality control, revealing internal differences that a size analyzer cannot see.

NMR works best as part of a measurement toolkit

The study did not rely on NMR alone. Its toolkit included SAXS, cryogenic transmission electron microscopy and nano-flow cytometry alongside NMR.

SAXS can report nanoscale ordering. Cryo-TEM can visualize particle morphology in a frozen hydrated state. Nano-flow cytometry can measure individual nanoscale particles. NMR contributes information on molecular environments and dynamics.

Together they turn an LNP from a featureless sphere into a physical structure that can be interrogated layer by layer and interaction by interaction.

Manufacturing is part of the drug

Nanomedicine forces a shift in how pharmaceutical manufacturing is understood.

For a conventional small molecule, the active chemical identity is central. For a self-assembled nanoparticle, mixing rate, solvent composition, pH, concentration, fluid path and order of addition can all affect the structure that forms.

That means the manufacturing process can determine a critical quality attribute rather than merely reproduce one.

The Chiba pre-mix/post-mix comparison is a clear example: nearly the same ingredient list produced different internal arrangements because the assembly pathway differed.

By 2026, the field is moving beyond liver delivery

A 2026 review in Nature Reviews Bioengineering describes RNA-LNP systems being developed beyond prophylactic vaccines for cancer immunotherapy, protein replacement and gene therapy.

Researchers now tune lipid chemistry and formulation to change tissue targeting, circulation, cellular uptake, endosomal escape and clearance.

At the same time, toxicity remains a central problem. Recent 2026 work has shown that ionizable lipids and the membrane-disrupting processes required for endosomal escape can also drive inflammatory responses. Efficient delivery and tolerability must therefore be engineered together.

This study did not demonstrate a new treatment in animals or patients

The Chiba work is formulation science. It examined preparation conditions, internal structure, siRNA distribution and gene-silencing performance.

It did not establish safety or efficacy for a specific disease in patients, nor did it demonstrate that NMR optimization alone will improve every LNP in vivo.

This is also a 2024 paper, not a newly published 2026 result

The underlying Journal of Controlled Release paper was published online August 2, 2024; Chiba University’s news release followed on October 1, 2024.

Its relevance in the October 2026 Chiba edition is the continuity of Ueda’s formulation-science program. He published a 2025 review on NMR-based molecular characterization of LNP formulations, and Chiba University highlighted his broader “molecular pharmaceutics” work again in 2026.

The accurate framing is therefore a deeper science feature on a foundational Chiba result, not a breaking 2026 discovery.

The next generation of RNA drugs will design the sequence and the container together

Early RNA-therapy development focused heavily on sequence: which gene to target and which RNA sequence would silence it.

Delivery science showed that the carrier matters just as much. The Chiba work goes one step further: even choosing the right carrier ingredients may not be enough. How those ingredients are mixed can determine where RNA ends up inside the nanoparticle.

Drug performance therefore depends on information at several scales at once: the nucleotide sequence, the lipid chemistry, the particle architecture and the manufacturing pathway that creates it.

If NMR can make that hidden architecture measurable, RNA medicines can move a little further from empirical recipe-making toward engineering—build the particle, inspect its molecular organization, and redesign it before the biology fails.