Three amino acids boost mRNA delivery 20-fold and push gene editing to 90%

Biohub (Chan Zuckerberg Initiative)

Lipid nanoparticles carried COVID-19 mRNA vaccines into billions of arms. But when researchers try to use the same delivery vehicles for cancer therapies, anti-inflammatory treatments, or CRISPR gene editing, they hit a wall: the particles work far better in lab dishes than inside living bodies.

A team at Biohub has figured out why, and the fix is almost absurdly simple.

The cell, not the particle, was the problem

For years, the field assumed the efficiency gap was a nanoparticle engineering problem. Labs screened hundreds of novel lipid formulations. AI-driven searches explored billions of possible combinations. Yet clinical delivery efficiency remained disappointing.

Daniel Zongjie Wang and Shana O. Kelley at Biohub flipped the question. Instead of redesigning the vehicle, they asked whether the body's own cells might be the bottleneck. The answer, published March 11, 2026, in Science Translational Medicine, is yes.

Standard lab culture media, formulated decades ago to maximize cell growth, contain nutrient concentrations far higher than what cells encounter in the bloodstream. When the team grew cells in a medium that mimics actual human plasma, lipid nanoparticle (LNP) uptake dropped by 50 to 80 percent. Metabolic and genetic analyses pinpointed the cause: several amino acid pathways were significantly suppressed in the body-like conditions. Cells running on a leaner metabolic diet simply could not internalize nanoparticles as efficiently.

Methionine, arginine, and serine

Through systematic screening, the team identified an optimized supplement containing three common amino acids: methionine, arginine, and serine. Co-injecting this cocktail with LNPs produced striking results: a 5- to 20-fold increase in target protein expression across diverse cell types, in both lab dishes and living animals.

The boost was consistent across three major administration routes: intramuscular, intratracheal (into the lungs), and intravenous. It worked regardless of which specific lipid formulation or mRNA cargo was used. Mechanistic studies showed the amino acids enhance a specific cellular uptake pathway, essentially widening the door through which nanoparticles enter cells.

Every mouse survived

The team tested the supplement in two demanding scenarios. In a mouse model of acetaminophen-induced acute liver failure, the leading cause of drug-induced liver failure in human patients, mice treated with growth hormone mRNA in LNPs alone had a 33% survival rate. When the same treatment was paired with the amino acid supplement, survival jumped to 100%. Serum levels of the therapeutic protein rose nearly ninefold, and markers of liver damage dropped to near-healthy levels.

In a second set of experiments, CRISPR-Cas9 components were delivered to mouse lungs via LNPs. Without the supplement, a single dose achieved editing efficiencies of 20 to 30 percent, consistent with published results. With the amino acids, efficiency soared to 85 to 90 percent from one administration. For diseases like cystic fibrosis, which require efficient gene correction in lung tissue, that kind of single-dose efficiency could be transformative.

Pharmaceutical-grade and already manufactured at scale

What makes the approach attractive for clinical translation is its simplicity. The three amino acids are pharmaceutical-grade compounds already manufactured at industrial scale and widely regarded as safe. Unlike strategies requiring genetic manipulation of target cells or redesign of the nanoparticle itself, the cocktail could simply be mixed into the injection buffer alongside existing LNP formulations.

Kelley noted that any LNP formulation being developed today could potentially benefit from the approach.

Mouse results, human questions

All efficacy data come from mice. Whether the same metabolic bottleneck exists in human cells in vivo, and whether the same amino acid supplement resolves it, remains to be tested in clinical trials. Mouse and human cellular metabolism differ in ways that could affect the magnitude of the benefit.

The long-term safety of repeated amino acid co-administration alongside LNP therapies has not been evaluated. While individual amino acids are safe at normal dietary levels, pharmacological doses delivered alongside potent gene-editing tools are a different matter.

The 85-90% editing efficiency in mouse lungs is impressive, but off-target editing rates at those high efficiencies need careful characterization. More efficient delivery means more CRISPR activity in more cells, which could amplify both intended and unintended edits.

Still, if the approach translates to humans, it would address one of the central obstacles in the mRNA and gene therapy field without requiring any changes to the nanoparticles themselves.