Engineering RNA Payload Distribution In LNPs To Unlock Higher In Vivo Gene Editing Efficiency

By Sixuan Li, Ph.D., Hai-Quan Mao, Ph.D., Denny Truong, Ph.D., and Jeff Wang, Ph.D., John Hopkins University

Lipid nanoparticles (LNPs) have become one of the most advanced systems for modern nucleic acid delivery. Their clinical validation through siRNA therapeutics and mRNA vaccines has positioned them as one of the leading nonviral vectors for next-generation modalities, including CRISPR-based gene editing. But as the field pushes into more complex therapeutic territory, a critical and often overlooked variable is beginning to surface: how RNA payloads are actually distributed within and across LNPs.

For simpler systems, such as single-cargo mRNA delivery, successful LNP formulation can often be approximated by bulk metrics like encapsulation efficiency, particle size, and overall RNA concentration. These measurements have served the field well, offering a practical framework for formulation development and quality control.

However, gene editing introduces a fundamentally different challenge. It requires the co-delivery of at least two distinct RNA species, typically messenger RNA (mRNA) encoding a nuclease such as Cas9 and guide RNA (gRNA) that directs the editing machinery to a specific genomic locus. Therapeutic success depends not only on delivering these components efficiently but on delivering them together, into the same cell, in the correct proportions, and within a functional time window.

This additional layer of complexity exposes a gap in how LNP systems are evaluated. Bulk measurements can confirm that RNA is present, but they cannot reveal whether that RNA is meaningfully organized at the level required for function.

The Hidden Complexity Of Multi-RNA Delivery

At a conceptual level, co-encapsulation of multiple RNA species may seem straightforward: mix the components, formulate the particles, and deliver the payload. In practice, however, this process produces a highly heterogeneous population of nanoparticles.

Rather than a uniform distribution, LNP formulations typically contain a mixture of particle types. Some successfully encapsulate both mRNA and gRNA. Others contain only one of the two. Still others contain no RNA at all.

This heterogeneity introduces a fundamental inefficiency into the system. Only co-encapsulated particles are capable of driving productive gene editing. Particles carrying only one RNA species may not form the necessary ribonucleoprotein complex. Empty particles contribute nothing to efficacy while still potentially influencing biodistribution and immune response.

The result is a disconnect between what is theoretically deliverable and what is functionally achieved in vivo.

Despite its importance, this payload distribution has historically been difficult to measure. Traditional analytical tools provide averaged values that mask underlying variability. As a result, payload heterogeneity has remained largely invisible — an unmeasured variable with potentially significant consequences.

A Single-Particle View With CICS

To address this blind spot, researchers applied cylindrical illumination confocal spectroscopy (CICS), a technique that enables single-particle analysis of LNPs. Unlike bulk methods, leveraging the single-molecule detection sensitivity, CICS can resolve the RNA content of individual nanoparticles, providing a direct view into payload distribution.

Using this approach, four distinct LNP subpopulations were identified:

• co-encapsulated LNPs containing both mRNA and gRNA (50.7%–60.4%)
• gRNA-only LNPs (30.0%–36.5%)
• mRNA-only LNPs (2.0%–3.4%)
• empty LNPs (4.2%–13.8%).

These findings immediately challenge the assumption of uniformity. Even in optimized formulations, a substantial fraction of particles is functionally incomplete.

More importantly, the analysis revealed significant variability within each subpopulation. Co-encapsulated particles did not all carry the same number of RNA molecules. Instead, there was a broad distribution of mRNA and gRNA copy numbers from particle to particle.

This intra-class variability adds another layer of complexity. Even among “functional” particles, differences in RNA stoichiometry can influence the efficiency of ribonucleoprotein formation and, ultimately, gene-editing activity.

Taken together, these results shift the conversation from whether RNA is encapsulated to how it is organized, both across and within particles.

Formulation Drives Functional Outcomes

One of the most important insights from this work is that payload distribution is not merely a byproduct of formulation. It is a direct consequence of it.

By comparing LNPs formulated with different ionizable lipids, including ALC-0315 and DLin-MC3-DMA, and varying mixing strategies, the researchers demonstrated that formulation choices significantly influence payload heterogeneity.

Changes in lipid composition and processing conditions altered the proportion of co-encapsulated particles, the prevalence of empty particles, and the RNA loading per particle. These effects were not subtle; they represented meaningful shifts in the functional landscape of the formulation.

A particularly notable observation was the inverse relationship between empty particle fraction and RNA loading. Formulations that minimized the number of empty LNPs tended to exhibit higher RNA content in the particles that did carry cargo.

This relationship suggests that improving payload distribution is not simply about increasing total RNA input. It requires controlling how that RNA is partitioned during particle formation.

In practical terms, formulation becomes an exercise in probability management: increasing the likelihood that any given particle will contain the right combination and quantity of RNA to be functionally effective.

Connecting Structure To In Vivo Editing Efficiency

To determine whether these structural differences translate into biological outcomes, the researchers evaluated gene editing performance in vivo using mouse models.

The results provide a clear link between payload distribution and therapeutic efficacy.

LNPs with higher RNA loading per co-encapsulated particle achieved significantly greater editing activity:

• ~9.8 mRNA copies and ~25.4 gRNA copies per particle → 55.4% editing (indels)
• ~8.0 mRNA copies and ~20.3 gRNA copies per particle → 36.3% editing

This represents approximately a 1.5-fold increase in editing efficiency, a substantial gain in a system where incremental improvements are often hard-won.

Crucially, these differences occurred despite nearly identical bulk characteristics, including particle size, PDI, and overall encapsulation efficiency. From a traditional analytical perspective, the formulations would appear comparable.

This disconnect underscores the limitations of conventional metrics. Two LNP systems can look the same on paper while performing very differently in vivo due to differences in payload organization.

Moving Beyond Encapsulation Efficiency

Encapsulation efficiency has long been a cornerstone of LNP development, serving as a key indicator of formulation success. But as this work demonstrates, it is no longer sufficient for evaluating multicomponent systems.

A high encapsulation percentage does not guarantee functional delivery. It does not reveal how many particles contain both required components, nor does it capture the variability in RNA loading among those that do.

For gene editing applications, more meaningful metrics include:

• the fraction of co-encapsulated particles
• the distribution of RNA copy numbers within those particles
• the prevalence of single-cargo and empty particles.

These parameters provide a more direct link to biological function and therapeutic potential.

In this context, encapsulation efficiency becomes a necessary but incomplete measure — one piece of a larger, more nuanced picture.

Implications For LNP Design And Development

The findings from this study have broad implications for how RNA therapeutics are designed, developed, and evaluated.

Payload Distribution As A Design Parameter

Payload distribution should be treated as an explicit design objective. Formulation strategies should aim not only to encapsulate RNA but to optimize how it is distributed across particles and within each particle.

Process Development As A Lever For Potency

Mixing methods, lipid selection, and manufacturing conditions all influence payload heterogeneity. This positions process development as a critical lever for improving therapeutic performance.

Single-Particle Analytics As A Standard Tool

Techniques like CICS provide insights that are inaccessible through bulk analysis. Incorporating single-particle characterization into development workflows can enable more informed decision-making and faster optimization cycles.

Redefining Quality Attributes

As the field evolves, quality frameworks may need to expand to include distribution-based metrics. Co-encapsulation rate, empty LNP fraction, and per-particle RNA loading could become key indicators of product consistency and efficacy.

A Broader Shift Toward Precision Delivery

While this work focuses on CRISPR-based gene editing, its implications extend across the broader RNA therapeutics landscape.

As therapies become more sophisticated — incorporating multiple RNA species, targeting specific cell populations, or requiring precise temporal control — the importance of payload organization will only grow.

The field is moving beyond a binary view of delivery (delivered vs. not delivered) toward a more nuanced understanding of precision delivery, where the spatial and quantitative arrangement of therapeutic components determines outcome.

This shift mirrors trends in other areas of medicine, where increasing complexity demands more refined control and characterization.

From Hidden Variable To Engineering Target

Payload heterogeneity has long existed in LNP systems, but it has largely gone unmeasured and unoptimized. This work brings it into focus and demonstrates that it is not merely a source of variability but a controllable determinant of therapeutic performance.

Establishing a clear link between nanoparticle structure and in vivo function provides a framework for more rational design of multicomponent delivery systems.

For developers, the takeaway is both simple and consequential: if you are not measuring payload distribution, you are likely overlooking one of the most important levers in your system.

As RNA therapeutics continue to evolve, success will increasingly depend on the ability to engineer not just what is delivered, but how it is delivered, down to the level of individual particles.

In that context, payload distribution is no longer a hidden variable. It is an engineering target.

About The Authors:

Sixuan Li, Ph.D., is an assistant research scientist in mechanical engineering at Johns Hopkins University. His research focuses on developing advanced analytical methods for characterizing RNA therapeutics and lipid nanoparticles. He is the lead developer of cylindrical illumination confocal spectroscopy (CICS), and cofounder of CICS Analytics Corp. Li also serves on the USP Expert Panel on mRNA Vaccines and Therapeutics and collaborates with academic, industry, and government partners to establish new analytical standards for RNA-based medicines.

Tza-Huei “Jeff” Wang, Ph.D., is the Louis M. Sardella Professor of Mechanical Engineering at Johns Hopkins University. His expertise spans micro- and nano-biotechnologies, molecular diagnostics, single-particle analysis, and quantitative measurement platforms for biomedical applications. His research is guided by a vision to advance global health equity through technologies that combine high sensitivity, specificity, affordability, and accessibility. His team has advanced CICS-based methods for lipid nanoparticle characterization, enabling single-particle measurements of nucleic acid payload, lipid content, and formulation heterogeneity. Wang has authored more than 200 peer-reviewed journal articles and holds more than 30 patents. He is a fellow of AAAS, AIMBE, ASME, IEEE, RSC, and IAMBE.

Hai-Quan Mao, Ph.D., is director of the Institute of NanoBioTechnology and professor of materials science and engineering and biomedical engineering at Johns Hopkins University, with a joint appointment in the Center for Translational Immunoengineering and Translational Therapeutic and Regenerative Engineering Center. His research integrates materials science, immuno-engineering, and nanotechnology to develop biomaterials and nanotherapeutic platforms for tissue repair, stem cell delivery, immune modulation, and nucleic acid and protein delivery. He is a fellow of the National Academy of Inventors and the American Institute for Medical and Biological Engineering and serves on multiple editorial boards.

Linh (Denny) B. Truong, Ph.D., is an associate scientist working in the Research & Development team at Arbor Biotechnologies. Arbor Biotechnologies is a clinical stage, next-generation gene editing company based in Cambridge, MA, and is advancing a portfolio of first-in-class genomic medicines addressing serious diseases for which there are no existing functional cures.

Additional contributions to this article were made by Lyn Zhu.