The Hidden Physics Of RNA-LNPs: Tuning RNA Encapsulation For Potency And Safety

By Turash Haque Pial, postdoctoral scholar, Department of Materials Science and Engineering, Whiting School of Engineering, Johns Hopkins University

From Bench To Bedside: Why LNP Design Determines RNA Therapeutic Success

RNA therapeutics have rapidly transitioned from experimental platforms to clinically validated medicines, enabled largely by lipid nanoparticles (LNPs). These delivery systems now underpin vaccines, gene-silencing therapies, and emerging applications in protein replacement and gene editing. As clinical pipelines expand toward chronic and repeat-dose indications, the focus is shifting from proof-of-concept efficacy to a more demanding challenge: achieving consistent, safe, and predictable delivery.

LNP formulations are commonly characterized by using simplified metrics such as particle size and encapsulation efficiency. However, growing evidence suggests that these measurements do not fully capture functional performance. Formulations with nearly identical bulk properties can produce markedly different biological outcomes, raising fundamental questions about how nanoparticle structure, composition, and formation history influence therapeutic effectiveness.

A central challenge in RNA delivery is therefore not only the discovery of new lipid chemistries but also understanding how formulation processes govern the distribution of RNA cargo among individual nanoparticles and ultimately influence biological performance.

Hidden Heterogeneity In RNA Delivery Systems And In Vitro Consequences

LNPs form through rapid mixing of lipids dissolved in organic solvent with RNA dissolved in an aqueous buffer. This process occurs on millisecond timescales and involves simultaneous diffusion, electrostatic complexation, and hierarchical self-assembly. Although the resulting LNPs appear uniform, individual particles may contain vastly different amounts of RNAs.

Recent analytical advances have revealed that LNP populations frequently include a substantial fraction of particles containing little or no RNA payload, while others carry disproportionately large amounts. These empty carriers contribute to total lipid exposure without delivering therapeutic benefit, potentially limiting dose efficiency and increasing safety concerns during repeated administration.

For acute applications such as vaccination, this variability may be tolerable. However, as RNA therapeutics move toward chronic treatment paradigms, including metabolic disease, genetic disorders, and long-term protein replacement therapies, the implications become more significant. Repeated dosing amplifies the need to minimize inactive material that may trigger unwanted biological responses while maintaining therapeutic potency.

To better understand the functional consequences of nanoparticle variability, we examined gene-silencing performance across LNP formulations produced under different mixing conditions. Although all formulations exhibited comparable encapsulation efficiencies, they differed substantially in uniformity. Using cylindrical illumination confocal spectroscopy (CICS), we observed significant heterogeneity in RNA distribution among individual LNPs. We systematically varied both the fraction of empty LNPs and the width of the RNA payload distribution to evaluate the effects of heterogeneity on biological performance.

At high RNA concentrations, therapeutic outcomes were similar across formulations, consistent with conventional expectations. However, at lower doses, formulations containing fewer empty LNPs and a more uniform RNA distribution demonstrated significantly enhanced in vitro activity. Because the total number of RNA molecules was held constant, the improvement could not be attributed to increased total RNA loading. Instead, the results suggest that more uniform RNA partitioning across nanoparticles enables more consistent cellular uptake and delivery.

These results highlight an important principle for chronic therapies: maintaining efficacy through improved nanoparticle uniformity may allow lower administered doses, reducing toxicity risk, improving tolerability, and easing manufacturing demands. Nanoparticle uniformity, therefore, emerges as a key determinant of therapeutic performance, beyond traditional metrics such as total payload or encapsulation efficiency.

The Challenge Of Understanding Formulation Physics

Despite its importance, the physics of payload heterogeneity has remained poorly understood. Traditional formulation development has largely relied on empirical optimization because the underlying physics of nanoparticle formation is difficult to observe directly. Explaining how such heterogeneity arises requires confronting the intrinsic complexity of LNP assembly, a process that spans multiple physical regimes simultaneously: molecular-scale lipid–lipid and lipid–RNA interactions occurring on nanosecond timescales; mesoscale particle growth driven by diffusion and aggregation; and macroscopic fluid mixing governed by hydrodynamics.

No single experimental or computational method can directly capture this entire hierarchy. Experimental measurements provide endpoint characterization but offer limited visibility into early assembly events. Conversely, atomistic simulations capture molecular interactions but cannot realistically follow particle evolution over experimentally relevant timescales. As a result, formulation development has historically relied on iterative experimentation rather than predictive modeling.

To address this gap, we developed a multiscale framework integrating coarse-grained molecular simulations, kinetic Monte Carlo modeling, and machine learning analysis. Experimental measurements of single-particle RNA content were used to validate predictions and connect simulation outcomes to biological performance. By combining methods operating across complementary time and length scales, this framework identifies the dominant physical mechanisms governing nanoparticle formation while enabling predictive control of the formulation process. Rather than attempting to reproduce every molecular detail, the framework focuses on identifying the dominant physical mechanisms that govern nanoparticle formation. By linking process conditions to emergent structural outcomes, the approach provides a pathway toward mechanistic understanding and engineering control of LNP assembly.

Engineering Insight And Design Rules

The combined modeling and experimental analysis revealed that RNA loading heterogeneity originates primarily during the earliest stages of mixing. When lipid and RNA streams first meet, mixing is not instantaneous. Instead, transient concentration gradients emerge, creating localized regions in which lipids aggregate before encountering RNA molecules. Nanoparticles nucleated within these lipid-rich regions become effectively “empty,” and subsequent growth preserves this initial compositional state.

In this framework, empty LNPs are not defects introduced during later processing steps but kinetic products of incomplete mixing during early assembly. The analysis further showed that RNA incorporation is governed by a competition between two characteristic timescales: lipid aggregation kinetics and RNA diffusion into forming particles. Rapid mixing shortens diffusion distances, enabling RNA to participate more uniformly in nucleation events. In contrast, slower mixing increases the likelihood that lipid aggregates form independently, leading to heterogeneous nanoparticle populations.

Beyond mixing kinetics, we also investigated additional formulation parameters regularly used in experiments and industry, including PEGylation levels and solution salt concentration. These factors were found to influence RNA payload distribution through mechanisms distinct from mixing-driven effects.

Several engineering principles emerged from this analysis. First, mixing dynamics primarily control RNA distribution but, within turbulent mixing regimes, exert minimal influence on LNP size. This decoupling provides a pathway to independently tune payload uniformity without substantially altering particle dimensions. Second, thermodynamic factors such as PEGylation and ionic strength regulate nanoparticle merging behavior. Increased merging promotes the formation of larger LNPs while reducing the fraction of empty particles. Unlike mixing kinetics, however, these thermodynamic parameters do not independently control payload distribution; instead, RNA loading scales volumetrically with particle size.

Together, these findings establish complementary kinetic and thermodynamic control knobs for LNP design. By integrating mixing-driven control of RNA distribution with thermodynamic regulation of particle growth, it becomes possible to engineer nanoparticles with targeted combinations of size and payload uniformity. Such precision is critical because LNP size strongly influences cellular uptake and biodistribution, while homogeneous RNA loading maximizes therapeutic potency and supports improved safety profiles.

Implications For Manufacturing And Chronic Therapy Development

The ability to decouple particle size from payload uniformity has practical implications for both safer therapy development and manufacturing.

For chronic indications, improved RNA distribution may enable therapeutic efficacy at lower administered doses, reducing cumulative lipid exposure and improving safety margins. Lower dosing requirements could also ease manufacturing scale demands and improve supply sustainability for long-term treatments.

From a process development perspective, the findings suggest that mixing conditions should be treated as critical design variables alongside lipid composition. Small changes in fluid dynamics may have outsized effects on therapeutic performance, even when standard quality metrics appear unchanged.

More broadly, integrating predictive modeling with experimental validation offers a path toward rational formulation development, reducing reliance on trial-and-error optimization.

Looking Forward

This work highlights that nanoparticle performance is largely determined during formation itself. Early assembly dynamics shape RNA distribution across particles, and this structural uniformity directly influences biological function by governing the consistency of cellular delivery and therapeutic response. Clarifying these links between nanoscale organization and biological outcome represents an important step toward understanding LNPs not simply as carriers but as structured delivery systems whose internal heterogeneity can dictate efficacy and safety.

More broadly, advancing the field will require integrating mechanistic insight with data-driven design strategies capable of connecting formulation parameters to biological performance. Multiscale modeling, single-particle characterization, and machine learning approaches together offer a pathway to map complex formulation spaces and identify predictive relationships between process conditions, nanoparticle structure, and therapeutic outcomes.

As RNA therapeutics expand beyond acute applications toward chronic and repeat-dose treatments, demands on delivery reliability, safety, and reproducibility will continue to grow. Moving toward precision RNA medicine will therefore depend on predictive data-informed design rules that link formulation physics to biological function. Such approaches will enable rational optimization tailored to specific clinical contexts, supporting scalable manufacturing, consistent patient responses, and a more precise and controllable generation of RNA-based therapies.

This article summarizes key findings from a recent research paper published in Advanced Functional Materials, “Controlling Payload Heterogeneity in Lipid Nanoparticles for RNA-Based Therapeutics,” in which Turash Haque Pial, and colleagues from John Hopkins University, examine how the physical processes governing lipid nanoparticle (LNP) formation shape RNA delivery performance.

About The Expert

Turash Haque Pial, MD, is a postdoctoral researcher at Johns Hopkins University, specializing in computational materials science. He uses multiscale modeling and machine learning to study soft and bio-inspired material assemblies. His research connects formulation processes with the structure of drug delivery materials and their biological responses. He is also invested in creating open-access tools to translate mechanistic insight into deployable RNA technologies that can broaden access to chronic and precision healthcare.