Amphiphile Blending Controls Micelle Morphology to Enhance In Vitro Performance and Selective In Vivo mRNA Delivery to Lung and Spleen

Abstract

Polymeric micelles assembled from amphiphilic block copolymers represent a versatile platform for mRNA delivery due to their modular chemical design and tunable self-assembly. Here, we report a library of ternary mixed micelles composed of amphiphiles displaying a cationic oligoamine (A7), anionic carboxyethyl acrylate (CEA), and neutral poly­(ethylene glycol) (PEG), engineered to enhance in vitro transfection and enable passive, ligand-free in vivo organ targeting. Systematic blending of these components allowed modulation of micelle surface chemistry and morphology, which are key parameters that govern mRNA complexation, stability, and biodistribution. Increasing incorporation of anionic amphiphiles resulted in a distinct evolution of morphology from spherical micelles to elongated wormlike structures, while maintaining uniformly high zeta potential and efficient mRNA complexation. Top-performing ternary formulations in vitro achieved over 95% transfection efficiency and cell viability across multiple cell types, representing a 2-fold performance enhancement relative to homomicelle controls. These formulations also exhibited robust mRNA internalization and efficient endosomal escape, resulting in a 7-fold improvement in GFP mean fluorescence intensity. Furthermore, by controlling micelle aggregation, we demonstrate that serum stability can be significantly improved, highlighting the tunability of this mixed micelle system. Finally, in vivo, increasing incorporation of anionic amphiphiles consistently shifted mRNA expression from lungs to spleen. We attribute this shift in organ tropism to morphology, where spherical micelles sequester in the lungs and wormlike ternary mixed micelles accumulate in the spleen. Transmission electron microscopy confirmed the elongated wormlike architecture that correlated with spleen localization, supporting morphology as the dominant factor driving organ selectivity. This morphology-driven organ selectivity marks a departure from traditional charge-centric targeting paradigms and establishes a tunable, composition-based framework for directing polymeric micelles to specific organs. Together, these findings introduce a simple yet powerful strategy for enhancing in vitro mRNA delivery and achieving organ-specific mRNA delivery using a blended micelle platform.

Introduction

The emergence of nucleic acid therapeutics has revolutionized modern medicine, offering highly precise interventions for the treatment of both genetic and acquired disorders, including thalassemia, sickle cell anemia, cystic fibrosis, and cancer. − With 33 current FDA-approved genetic therapies and thousands more in clinical trials, the potential for curative treatments continues to expand. Among these, messenger RNA (mRNA)-based therapies have attracted significant attention due to their ability to drive transient protein expression without necessitating genomic integration, thus mitigating the risks of insertional mutagenesis and other genetic perturbations. , Despite its widespread clinical adoption, including the development of vaccines, mRNA-based therapeutics face significant barriers, primarily stemming from their susceptibility to enzymatic degradation, instability in physiological environments, and limited cellular uptake. , Consequently, overcoming these barriers demands the development of advanced delivery systems capable of effectively protecting mRNA payloads while facilitating targeted cellular entry and efficient intracellular release. Furthermore, addressing the challenges associated with organ-specific localization remains crucial, as precise organ-targeted delivery not only enhances therapeutic efficacy but also minimizes off-target effects and systemic toxicity.

To address these challenges, nanoparticle-based delivery platforms have been extensively explored, with viral vectors and lipid nanoparticles (LNPs) leading the field. , However, while effective, these systems can present notable drawbacks, particularly challenges related to room temperature storage stability and effective manufacturing scale-up. − Consequently, polymeric nanocarriers have gained traction as versatile alternatives, offering broad chemical tunability, scalable synthesis, and the ability to precisely modulate their physicochemical properties for enhanced therapeutic performance. , Additionally, polymeric delivery systems provide access to less encumbered intellectual property space, as many LNP formulations are patented, which can hinder broad adoption and rapid translation. Among these, cationic polymeric micelles have emerged as particularly promising candidates, formed through the self-assembly of amphiphilic block copolymers to generate stable, nanoscale carriers with well-defined architectures. − These micelles exhibit remarkable potential in gene delivery by electrostatically condensing nucleic acids into micellar nanoparticles (MNPs) or micelleplexes, ensuring cargo protection, controlled release, and allowing access to a diverse range of structural motifs. − Furthermore, polymeric micelle platforms offer a compelling route toward effective extrahepatic targeting, whereas LNPs exhibit a strong hepatic bias, along with accumulation in other organs, driven by their cholesterol- and phospholipid-rich compositions.

Beyond their role in nucleic acid complexation, the physicochemical properties of cationic micellar nanoparticles, particularly surface charge and corona composition, critically influence their biodistribution, cellular interactions, and therapeutic efficacy. ,− The positive surface charge of these nanoparticles is well-known to trigger protein opsonization, leading to serum fouling, immune activation, and diminished delivery efficiency. , In contrast, neutral or mildly anionic formulations tend to exhibit extended circulation times and lower systemic toxicity. − Given these charge-dependent effects on biodistribution and cellular uptake, a particularly promising approach to achieve optimal charge balance and structural control is the engineering of mixed micelle formulations. By coassembling amphiphiles with distinct physicochemical properties, mixed micelles provide a dynamic platform for optimizing corona charge balance and modulating micelle size and morphology. Notably, incorporating neutral or anionic amphiphiles alongside cationic components has been shown to modulate nanoparticle-cell interactions, thereby affecting stability, cellular uptake, and release. ,,, For instance, the incorporation of PEGylated and cationic polymer blocks has been shown to facilitate siRNA release while simultaneously mitigating premature aggregation. Likewise, integration of anionic functionalities within mixed micelle formulations has led to enhanced pDNA delivery efficiency and improved nanoparticle serum stability in vitro, underscoring the critical role of charge modulation in improving therapeutic performance.

Importantly, modulation of nanoparticle physicochemical properties has demonstrated considerable promise in achieving passive targeting for organ-selective delivery. − Several findings have demonstrated charge-dependent organ-selective LNP systems, where cationic components preferentially promoted lung tropism, while anionic components facilitated spleen targeting. ,− Recently, Li et al. demonstrated that altering the charge ratio in polymer–nucleic acid complexes modulates not only the surface zeta potential but also particle morphology, resulting in distinct patterns of organ localization. Specifically, nanoparticles formulated with higher cationic charge densities maintained strongly positive surface potentials at physiological pH, which enhanced serum protein adsorption and promoted pulmonary accumulation. In contrast, formulations with near-neutral charge profiles favored splenic uptake. While these studies have emphasized surface charge as a determinant of organ tropism, other findings suggest that nanoparticle morphology and size may play an equally critical role in directing biodistribution. Studies from DeSimone and co-workers − and Mitragotri and co-workers − have shown that tuning particle shape, curvature, and mechanical deformability can redirect organ accumulation by altering how nanoparticles navigate vascular, interstitial, and cellular barriers. Taken together with insights from other related studies, ,, efficient and selective lung delivery appears to depend primarily on nanoparticle composition, where variations in surface chemistry, and the resulting protein corona, govern pulmonary accumulation more strongly than particle size. In contrast, spleen targeting is influenced predominantly by both particle morphology and size, as rigid nanoparticles exceeding 200 nm are physically restricted from passing through the splenic reticuloendothelial system, leading to preferential retention within the spleen. ,− Together, these reports highlight that both surface characteristics and morphology synergistically determine nanoparticle biodistribution, challenging the notion that surface charge alone governs organ selectivity.

A previous work from our group demonstrated that cationic micelles formulated from an amphiphilic diblock polymer incorporating oligoamine (A7 amphiphile) as the cationic functionality exhibited robust in vitro transfection performance (∼70% delivery efficiency) and precise lung-specific mRNA delivery in vivo. In the present study, we advance the A7 homopolymer micelle platform by systematically generating a diverse library of binary and ternary mixed micelles through controlled incorporation of neutral and anionic amphiphilic counterparts in variable proportions (Figure ). Specifically, polyethylene glycol (PEG)-based amphiphilic diblock copolymers are introduced as neutral components to mitigate cytotoxicity, and improve overall biocompatibility, whereas anionic carboxyethyl acrylate (CEA)-based amphiphiles are employed to modulate micelle charge, improve in vitro transfection efficacy, and induce tunable organ-selective biodistribution. For these ternary mixed micelles, we observe a unique morphological evolution wherein increasing the fraction of anionic amphiphile results in a progressive shift in micelle morphology from spherical to worm-like structures in a dose-dependent manner. Comprehensive in vitro transfection assays reveal that the ternary micelles comprising A7, CEA, and PEG exhibit remarkably high transfection efficiencies (∼95%), representing a significant advancement over A7 homomicelles. Additionally, the incorporation of PEG and CEA results in an impressive 7-fold enhancement in GFP mean fluorescence intensity (MFI), coupled with a substantial reduction in cytotoxicity, thus effectively doubling the overall effective efficiency (defined as transfection efficiency * cell viability). These improvements are particularly pronounced in serum-containing conditions, where we introduce a controlled aggregation strategy to achieve a remarkable 10-fold increase in GFP expression relative to native A7 micelles. Importantly, our in vivo investigations demonstrate, for the first time, that the compositional mixing–driven tuning of micelle morphology through noncationic amphiphile incorporation ultimately dictates precise tuning of organ tropism. Specifically, systematically varying the anionic component ratio, induces predictable changes in micelle morphology and surface presentation, which in turn enable targeted biodistribution from lung-specific to spleen-specific delivery, all without relying on ligand-based targeting strategies. Collectively, this study constitutes the first reported instance of polymeric micelle-based nucleic acid delivery systems exhibiting inherently morphology dependent tunable organ selectivity, underscoring the transformative potential of rationally engineered mixed micelle architectures for precision gene therapy applications.

1.

(A) Chemical structure of the three differently charged amphiphiles used for mixed micelle formulations. (B) Mixed micelle formulation strategy; the amphiphiles are dissolved in a good solvent at the required ratios, dried via N₂ stream to form a polymer film, and rehydrated with deionized water to form micelles exhibiting morphologies ranging from spherical to wormlike. (C) The formulated micelles are then complexed with mRNA (GFP mRNA for in vitro studies and Fluc mRNA for in vivo studies), followed by biophysical characterization and in vitro and in vivo screening to determine the delivery performance. The ternary mixed micelles enhanced GFP expression across multiple cell types, while in vivo studies revealed morphology-dependent organ targeting with spherical micelles preferentially accumulating in the lungs and wormlike micelles targeting the spleen.

Results and Discussion

Amphiphile Synthesis and Mixed Micelle Formulation

The mixed micelle formulations were prepared through the synthesis and assembly of three distinctly charged diblock copolymers: the cationic oligoamine amphiphiles (hereafter referred to as A7 amphiphiles), anionic carboxyethyl acrylate (CEA)-derived amphiphiles, and neutral poly­(ethylene glycol) (PEG)-based amphiphiles (Figure A-B). Leveraging our previously established postpolymerization modification methodology, the cationic A7 amphiphiles were synthesized from modular scaffolds comprising poly­(pentafluorophenyl acrylate)-b-poly­(n-butyl acrylate) (PPFPA-b-PnBA). These scaffolds were obtained through RAFT polymerization, subsequently functionalized via amidation with the A7 amine derivative. Selectivity in modification was enhanced through Boc protection of amines in A7, and quantitative conversion was confirmed by the complete disappearance of poly­(pentafluorophenyl acrylate) characteristic peaks, monitored via ¹⁹F NMR spectroscopy, as reported previously (Figures S1 and S4–S6). Furthermore, to promote stable micelle self-assembly, the hydrophobic PnBA block and cationic block lengths were fixed at 118 and 230 repeat units respectively, affording a hydrophilic fraction (f _hydrophilic) of 0.66, which exceeds the ∼0.5 criterion associated with spherical micelle assembly. The anionic CEA amphiphiles were similarly synthesized through RAFT-mediated chain extension of carboxyethyl acrylate monomer on a poly­(n-butyl acrylate) backbone used for synthesis of the A7 amphiphile, ensuring hydrophobic block consistency across amphiphiles (Figures S3A and S8). Finally, for the neutral amphiphile, a PEG (10K) macro chain transfer agent was first synthesized via acid-alcohol coupling of polyethylene glycol with DDMAT chain transfer agent (CTA). This PEG-functionalized macro-CTA was subsequently used for the RAFT polymerization of n-butyl acrylate, yielding the neutral PnBA-b-PEG amphiphile with matched hydrophobic chain length relative to the cationic and anionic amphiphiles (Figure S2). Each noncationic amphiphile possess hydrophilic blocks exhibiting consistent lengths of approximately 240–260 repeat units, ensuring uniformity in hydrophilic-to-hydrophobic ratios across the polymer library. Following synthesis, all amphiphiles underwent purification via dialysis, and purity was assessed using nuclear magnetic resonance (NMR) spectroscopy (Figures S6–S11). Furthermore, potentiometric titration of these amphiphiles revealed distinct ionization behaviors; notably, the A7 amphiphile exhibited dual pK _a values at 5.1 and 8.6, while the CEA amphiphile showed an acidic character with a single pK _a around 5.2 (Figure S12).

The assembly of the amphiphiles to generate “mixed” micelles (micelles containing 2 or 3 different block copolymers) was accomplished via an ultrasound-assisted thin-film hydration procedure (Figure B). To ensure the uniform composition and incorporation of the amphiphiles, each copolymer was individually dissolved in methanol and mixed at specified molar ratios to generate homogeneous mixtures. Subsequently, the solvent was evaporated under a nitrogen stream to yield a polymer film, that was hydrated with deionized water, followed by ultrasound sonication to yield the micelle formulations. To methodically explore ternary amphiphile compositions, we adopted the strategy of anchoring the internal molar ratio between the cationic A7 and neutral PEG amphiphiles at defined proportions (where each number represents the mole percentage): A7(100):PEG(0), A7(95):PEG(5), A7(90):PEG(10), and A7(85):PEG(15) (Table ). Each of these binary mixtures was incrementally doped with 10%, 20%, and 30% molar ratios of the CEA amphiphile. All components for a given formulation were combined in a single mixing step according to these final molar percentages yielding a comprehensive library of 16 distinct micellar formulations encompassing, 1 purely cationic homomicelle, 3 PEG containing binary mixtures, 3 CEA containing binary mixtures and 9 ternary mixtures (Table ). Importantly, practical limitations on amphiphile mixture solubility constrained maximum CEA doping to 30 mol %, as higher incorporation led to precipitation upon hydration due to potential charge neutralization. Similarly, increased PEG incorporation beyond 15 mol % was limited by aggregation-induced precipitation, thus establishing an upper bound for the A7:PEG ratio at 85:15. The nomenclature adopted for mixed micelle formulations is detailed in Table along with the respective molar composition ratios for the amphiphiles.

1. Mixed Micelle Formulations and Molar Ratios .

^a% indicates molar percentages of the respective component. Naming convention: A, P, and C denote Amine amphiphile, PEG amphiphile, and CEA amphiphiles, respectively. APC x/y indicates x mol % PEG in the A7:PEG binary mixture, followed by y mol % CEA doping. Thus, APC 10/20 corresponds to a 90:10 A7:PEG formulation containing 20 mol % CEA.

Micelle and Micelleplex Biophysical Characterization

Following the preparation of mixed micelles, dynamic light scattering (DLS) was employed to assess their hydrodynamic size distribution (Table S1). As shown in Figure A, a clear trend emerged, wherein hydrodynamic radii increased progressively with the incremental incorporation of PEG and/or CEA amphiphiles. PEG-containing binary formulations (APC 5/0, 10/0, and 15/0) exhibited a progressive increase in particle size, with hydrodynamic diameters increasing from ∼100 nm for APC 0/0 homomicelles to ∼180 nm for APC 15/0, corresponding to an approximately 2-fold increase. Similarly, CEA-containing binary formulations (APC 0/10, 0/20, and 0/30) displayed a more pronounced size expansion, reaching ∼250 nm for APC 0/30 micelles, representing an approximately 2.5-fold increase relative to APC 0/0. Ternary mixed formulations showed the largest size increases, with diameters reaching up to ∼400 nm for APC 15/30, corresponding to an approximately 4-fold increase compared to APC 0/0 homomicelles. The increase in size upon PEG incorporation likely results from neutral PEG chains displacing cationic amphiphiles within the micelle, thereby reducing electrostatic repulsion among neighboring chains in the corona. Literature reports have shown that strong charge–charge repulsion within highly cationic assemblies limits packing efficiency and typically yields smaller particles. , Thus, the reduction in interchain repulsion upon PEG incorporation permits denser amphiphile packing and greater aggregation, leading to larger micelles (∼150–200 nm). Additionally, the strong hydrophilicity of PEG enhances corona hydration and induces swelling, further expanding the hydrodynamic radius. In contrast, the size increase observed with CEA incorporation appears counterintuitive, as electrostatic complexation between anionic and cationic species should result in structural compaction. However, prior studies have shown that partial charge neutralization between anionic and cationic polymers can lead to the formation of a charge-neutral intrapolyelectrolyte complex (IPEC) layer surrounding the hydrophobic core rather than collapse of the assembly (Figure D). In our system, the molar ratio of cationic to anionic amphiphiles was maintained above 1 for all formulations (Table ), resulting in an excess of cationic amphiphiles. Under these conditions, we propose that the ternary micelles adopt a core–shell–corona architecture in which electrostatic pairing between cationic and anionic amphiphiles forms an intermediate IPEC shell. Importantly, this charge-neutral IPEC layer is highly hydrated and can swell significantly in aqueous media, providing a plausible explanation for the increased hydrodynamic radii observed for CEA-containing formulations. In addition, DLS measurements revealed negligible changes in hydrodynamic size and polydispersity index after two years of storage (Table S2), demonstrating excellent long-term stability of the mixed micelles. Notably, incorporation of oppositely charged amphiphiles did not result in long-term aggregation, underscoring the structural robustness and formulation stability of these assemblies for nucleic acid delivery applications.

2.

DLS hydrodynamic radii of (A) micelles in water, (B) micelleplexes in water, and (C) micelleplexes in OptiMEM. (D) Schematic representation of the core–shell–corona micelle architecture arising from blending oppositely charged amphiphiles, which leads to formation of an intrapolyelectrolyte complex (IPEC) intermediate shell. (E) TEM images showing morphology of (i) APC 0/0 homomicelle and (ii) APC 10/0 binary PEG containing micelle (spherical micelles), (iii) APC 10/10 and (iv) APC 10/20 ternary mixed micelles (core–shell-corona spherical micelles), and (v) APC 10/30 micelles (wormlike micelles). (E­(vi)–(x)) TEM images showing morphology of APC 0/0, 10/0, 10/10, 10/20, and 10/30 micelleplexes postcomplexation with mRNA.

To elucidate the underlying morphological transformations responsible for the observed size increase, transmission electron microscopy (TEM) was utilized to visualize the micellar structures directly. TEM imaging of a representative subset of formulations (APC 0/0, 10/0, 10/10, 10/20, and 10/30), selected to span the compositional space of the full library, revealed a progressive morphological evolution with increasing CEA incorporation (Figure E). Initially, the native cationic homomicelles (APC 0/0) and binary A7-PEG micelles (APC 10/0) predominantly exhibited spherical morphologies, consistent with previous reports , (Figure E (i)-(ii)). The APC 10/10 micelles containing all three different amphiphiles, retained a uniformly spherical shape; however, they displayed almost two times the size compared to the APC 0/0 homomicelles (Figure E (iii)). Furthermore, they displayed a distinct dark border around the core, indicating the likely formation of the core–shell-corona architecture (Figure E (i), (iii)). Interestingly, APC 10/20 micelles exhibited subtle micelle–micelle associations, manifested as chaining of spherical particles, consistent with the onset of higher-order assembly rather than discrete particle behavior (Figure E (iv)). We attribute this behavior to diminished corona charge–charge repulsion resulting from increased anionic incorporation and consequently reduced cationic density at the micelle surface, which increases the probability of micelle–micelle associations. In the APC 10/30 formulation, higher-order micelle architecture became strikingly pronounced, driving a distinct morphological transition from discrete spherical structures to elongated cylindrical/wormlike assemblies (Figure E (v)). We posit that this transformation is driven by increased incorporation of anionic amphiphiles, which enhances charge neutralization and expands the volume of the charge-neutral intermediate shell. Swelling of this shell increases the effective packing parameter and introduces internal strain within the micellar assembly, thereby destabilizing the spherical morphology. Simultaneously, the diminishing cationic surface charge density weakens electrostatic repulsions, further encouraging micelle–micelle fusion into extended wormlike structures. , Similar curvature transitions have been described in complex coacervate micelles (C3M), where modulation of electrostatic interactions alters packing efficiency and increases the effective packing parameter, resulting in a shift from spherical to wormlike geometries. Although our micelles differ in architecture, the underlying principle applies: increasing anionic incorporation reduces cationic surface density, expands the intermediate shell, and decreases the effective headgroup area, all of which push the assembly toward lower interfacial curvature. In our system, this results in evolution of compact spherical micelles into extended wormlike structures as the balance of charge neutralization, corona packing, and intermicellar association shifts with composition.

In addition, because the micelles were prepared by thin film rehydration, the resulting assemblies may represent kinetically trapped or metastable structures rather than fully equilibrated curvature-driven morphologies. Prior studies of block copolymer micelles have shown that preparation pathway and slow chain exchange kinetics can produce nonergodic assemblies whose structure depends on formation history rather than thermodynamic equilibrium. Under such conditions, polymer chain exchange between micelles may be restricted, leading to long-lived metastable morphologies that reflect precursor film organization or early aggregation events. Accordingly, the morphological evolution described here may arise from a combination of equilibrium self-assembly and kinetic trapping during the rehydration process. Furthermore, the thin film rehydration method may produce partially aggregated or interconnected structures, which could contribute to the large hydrodynamic radii observed for the ternary mixed micelles in DLS. Collectively, the transition from spherical to wormlike architecture appears to arise from composition-dependent modulation of electrostatic interactions, corona packing, and intermicellar association, while potential kinetic trapping during thin film rehydration may also influence the resulting structures. These findings highlight the importance of both compositional design and assembly pathway in determining the final architecture of polymeric micelles.

The proposed core–shell-corona architecture suggests that, due to the charge balancing effects and the presence of residual cationic and neutral PEG moieties, the micelle corona is likely to bear a net positive charge. To validate this hypothesis, zeta potential measurements were conducted across all formulations. Despite varying levels of anionic amphiphile incorporation, all micelles exhibited consistently high positive surface charges (∼+45–55 mV) (Figure S13), suggesting that residual cationic and PEG dominate the micelle surface, while anionic CEA is primarily confined to the neutralized intermediate shell. Thus, the zeta potential data complements DLS and TEM findings and underscores the need to interpret surface charge measurements within the broader context of micelle architecture and physicochemical environment.

Finally, we investigated the impact of mixed micelle composition on micelle complexation with mRNA cargo. Complexes were formulated at a fixed nitrogen-to-phosphate (N/P) molar ratio of 10, representing the micelle nitrogen (N) content relative to mRNA phosphate groups (P). Upon complexation in aqueous conditions, DLS measurements revealed a consistent decrease in the hydrodynamic diameter of the micelle-nucleic acid particles (micelleplexes or MNPs) when compared to respective native micelles (Figure B, Table S2). This reduction in size can be attributed to efficient cargo compaction within the micellar corona, thereby resulting in more compact structures. TEM analysis of the micelleplexes revealed morphologies consistent with the native micelles (Figure E (vi)-(x)). APC 0/0 and APC 10/0 micelles retained the spherical architecture upon complexation with mRNA while APC 10/30 MNPs displayed wormlike morphologies. However, dilution of the complexes into the transfection medium OptiMEM led to a pronounced increase in particle size for the ternary mixed micelles, suggestive of MNP aggregation mediated by ionic interactions with charged species present in the medium (Figure C). This aggregation phenomenon was particularly exacerbated in formulations containing higher levels of anionic amphiphile incorporation (APC XX/20 series and APC XX/30 series), as evidenced by the formation of aggregates extending into the micrometer scale. This behavior likely arises from increased charge neutralization, which reduces electrostatic repulsion and enhances susceptibility to salt-induced aggregation.

Crucially, successful mRNA delivery is immensely influenced by the micelle-mRNA binding affinity and the corresponding dynamics of cargo release. To investigate these parameters, two complementary dye exclusion assays employing RiboGreen dye were implemented (Figure S14). The RiboGreen binding assay aids quantification of the relative nucleic acid complexation by measuring the reduction in fluorescence that occurs when RiboGreen dye is excluded from RNA upon binding with the micelle. The initial RiboGreen assay conducted in pure water and at an N/P ratio of 10, indicated consistently high mRNA complexation efficiencies (∼98%) across all micelle formulations, independent of amphiphile composition (Figure S14A). Similarly, upon dilution into OptiMEM, an increase in fluorescence intensity, indicative of partial mRNA release, was observed (Figure S14B). However, this release did not exhibit clear compositional dependence, maintaining an average binding efficiency of approximately 80–85% across the library. To further assess the relative strength of micelle–mRNA interactions, we introduced heparan sulfate, an anionic glycosaminoglycan abundant on cell surfaces that competitively displaces nucleic acid cargo, into the RiboGreen assay. Its competitive binding with mRNA offers an indirect measure of MNP stability. However, despite the introduction of this competitive agent, no discernible trends emerged regarding amphiphile composition and binding strength (Figure S14C). These results corroborate earlier zeta potential observations, which demonstrated similar positive surface charges (∼+ 45 mV) across all micellar formulations, suggesting the surface charge plays a critical role in determining mRNA binding affinity. Thereby, the incorporation of PEG and anionic amphiphiles within the micelle architecture does not significantly compromise mRNA binding stability or complexation dynamics due to consistent surface charge, highlighting the robust and consistent biophysical behavior of these mixed micelle formulations.

Anionic Amphiphile Incorporation Enhances In Vitro mRNA Delivery

To elucidate the impact of mixed micelle composition on in vitro mRNA delivery efficiency, we employed a reporter gene assay with green fluorescent protein (GFP)-encoding mRNA. Flow cytometry was used to quantify two key delivery outcomes - transfection efficiency, defined by the percentage of GFP-positive cells, and geometric mean fluorescence intensity (MFI), indicative of functional mRNA expression per transfected cell. While transfection efficiency provides a global perspective on the overall ability of micelles to mediate cellular entry and transfection, GFP MFI offers deeper insights into intracellular mRNA release efficiency and subsequent translation efficacy, directly correlating to the amount of functional payload delivered per cell. Experiments were conducted across a range of therapeutically relevant cell lines, including HEK293T (kidney), A549 (lungs), and HuH7 (liver), at a fixed N/P ratio of 10, as previously we have shown A7 homomicelle to perform the best at this ratio. Since serum proteins commonly present in complete growth media can significantly influence MNP stability through aggregation or premature decomplexation, transfection was initially performed in serum-free OptiMEM for 4h. Postincubation, cells were supplemented with complete growth medium to maintain cellular viability. This experimental design enabled a clear assessment of micelle-specific delivery performance without interference from serum proteins.

Analysis of transfection outcomes across all three cell types demonstrated a clear composition dependent impact on micelle performance. For clarity, subsequent discussion focuses on HEK293T cells, as comparable trends were observed across all cell lines (Figures and S16–S17). Notably, incremental incorporation of the anionic CEA amphiphile in binary mixed formulations (APC 0/10, 0/20, and 0/30) displayed a strong positive correlation with GFP expression, with APC 0/30 formulation achieving 95% transfection efficiency (Figure A). In contrast, increasing incorporation of the PEG-based amphiphile in binary formulations (APC 5/0, 10/0, and 15/0) led to reduced transfection, with GFP-positive cells decreasing to ∼50% compared to ∼70% for APC 0/0 homomicelles (Figure A). Remarkably, ternary mixed formulations containing 30 mol % CEA (APC 5/30, 10/30, and 15/30) displayed significantly elevated transfection efficiency, achieving GFP-positive populations exceeding 90%, representing a substantial enhancement (∼30% improvement) compared to the APC 0/0 (∼70% GFP+ve) micelle control. Overall, across both binary and ternary formulation subsets, a clear stepwise increase in GFP expression was observed as CEA incorporation increased from 10 to 30 mol %.

3.

(A) Percentage of live cells positive for GFP expression upon full library transfection in HEK293T cells. (B) Mean fluorescence intensity (MFI) of GFP positive cells for full library transfection in HEK 293T cells. (C) Live cell images representing the change in levels of GFP expression upon incremental anionic amphiphile incorporation. U = Untreated, N = Naked mRNA, J = JetPEI (positive control). Error bars represent standard deviation (n = 3).

Interestingly, GFP MFI analysis revealed even more pronounced performance enhancements attributable to micelle composition (Figure B). Incremental incorporation of the anionic amphiphile produced a consistent increase in GFP MFI across both binary and ternary formulations, with APC 5/30 exhibiting the highest values (∼3500), compared to ∼500 for the APC 0/0 homomicelle control, corresponding to approximately 7-fold enhancements in MFI. Conversely, increasing PEG amphiphile content negatively impacted GFP MFI, with values decreasing from ∼500 for APC 0/0 to ∼250 for binary mixed formulations (APC 5/0, 10/0, and 15/0), reinforcing the inverse relationship between PEG incorporation and mRNA transfection efficiency. Consequently, ternary compositions mirrored these trends, demonstrating reduced GFP MFI values with elevated PEG amphiphile content, yet maintaining a positive correlation with CEA incorporation. Importantly, formulations APC 0/30 and APC 5/30 not only exhibited superior performance relative to other micellar formulations but also markedly outperformed the commercial standard, JetPEI, in terms of both transfection efficiency and GFP MFI. Consistent trends were observed in A549 and HuH7 cells (Figures S16–S17), where mixed formulations containing 30 mol % anionic CEA again exhibited the highest GFP expression, underscoring the robustness and broad applicability of these ternary mixed MNPs for efficient mRNA delivery across multiple cell lines.

Since all the evidence thus far points to the importance of anionic incorporation, we set out to investigate the impact of anionic amphiphile identity on delivery performance. This was done by replacing carboxyethyl acrylate (CEA) with other anionic amphiphiles, each featuring distinct chemical functionalities (Figure A). We synthesized three new amphiphiles from the following monomers: succinic ethyl acrylate (SEA), benzoic acid ethyl acrylate (BEA), and 2-(phosphonooxy)­ethyl acrylate (PEA) via RAFT polymerization as described before (Figures S3 and S9–S11). Given the superior delivery performance previously demonstrated by the APC 5/XX micelle formulation incorporating CEA, analogous ternary micelle formulations were prepared with SEA, BEA, and PEA at an identical A7:PEG molar ratio of 95:5 with different amounts of anion doping (10–30%). Transfection assays in HEK293T cells revealed that SEA-containing ternary formulations exhibited transfection efficiency and GFP MFI trends closely resembling those observed with CEA, while formulations incorporating BEA or PEA demonstrated no significant performance enhancement over the APC 0/0 homomicelle control (Figure A-B). Potentiometric titration assays revealed that the distinct ionization properties of the anionic amphiphiles were critical to the performance observed. SEA (pK _a ∼ 5.9) exhibited a similar acidic profile to CEA (pK _a ∼ 5.2), whereas BEA (pK _a ∼ 7.2) and PEA (pK _a ∼ 7.7) possessed substantially higher, near neutral pK _a values (Figure S12). These findings strongly suggest that amphiphiles with acidic pK _a values may play a key role in modulating pK _a dependent cellular mechanisms, which helps in improving delivery performance. Together, these results highlight the critical role of micelle composition in modulating transfection outcomes, with acidic amphiphile doping emerging as a key driver of enhanced gene expression. Overall, this establishes a tunable formulation strategy that can be broadly leveraged to optimize polymer-based delivery systems for high-efficiency, cell-type agnostic mRNA transfection.

4.

(A) Chemical structure of the four different anionic amphiphiles used to study the impact of anion pK _a on delivery performance. (B) Transfection efficiency of the ternary mixed micelles containing different anions upon transfection in HEK293T. (C) GFP MFI of the ternary mixed micelles containing different anions upon transfection in HEK293T. U = Untreated, N = Naked mRNA, J = JetPEI (positive control). Error bars represent standard deviation (n = 3).

Mechanistic Understanding of the Role of Anionic Amphiphile in Improving Performance

To probe the underlying mechanistic drivers governing the delivery performance trends, we conducted cellular internalization assays using Cy5-labeled GFP mRNA (Figure A). Uptake studies at 4 h postincubation demonstrated consistently high Cy5 positive populations (>90%) across all micelleplex formulations, which was sustained even at 24 h (Figure S15A,C). Despite comparable uptake percentages, Cy5 MFI measurements at 4 h revealed pronounced composition-dependent differences in internalization efficiency (Figures B and S15B). To facilitate direct comparison of mRNA uptake and intracellular consumption across formulations, Cy5 MFI values in Figure B are presented relative to the APC 0/0 homo micelle control at each time point, while absolute MFI values are reported in the Supporting Information (Figure S15B,D). Increasing anionic amphiphile incorporation led to systematic increases in Cy5 MFI across both binary and ternary formulations. Specifically, APC 0/30 exhibited an approximately 3-fold increase in Cy5 MFI (∼3000 units) relative to the APC 0/0 control (∼1000 units), while ternary formulations APC 5/30, 10/30, and 15/30 showed approximately 2-fold higher Cy5 MFI. In contrast, PEG-only binary formulations (APC 5/0, 10/0, and 15/0) did not enhance internalization, exhibiting Cy5 MFI values (∼800–1000 units) comparable to APC 0/0, thus correlating directly with their reduced transfection performance previously observed. The enhanced uptake observed for anion-containing mixed formulations is attributed to increased aggregation in OptiMEM, as supported by DLS measurements discussed above, which likely promoted accelerated MNP sedimentation onto the cell surface and facilitated internalization. At 24 h postincubation, absolute Cy5 MFI increased for all formulations (with Cy5MFI for APC 0/0 MNPs increasing 4× to ∼4000 units), reflecting continued cellular internalization of mRNA. Notably, CEA-containing binary formulations (APC 0/10, 0/20, and 0/30) maintained Cy5 MFI levels comparable to APC 0/0 (∼4000 units). In contrast, ternary formulations exhibited substantially lower Cy5 MFI (∼2000 units), corresponding to a 30–50% reduction relative to APC 0/0. This pronounced loss of Cy5 signal at later time points is indicative of more efficient intracellular release and subsequent translation-coupled catabolism of the delivered mRNA. Together, these results suggest that enhanced early internalization, followed by rapid intracellular payload utilization, underpins the superior GFP expression levels and elevated GFP MFI observed for ternary mixed micelle formulations.

5.

(A) Internalization studies with MNPs containing Cy5 tagged mRNA shows that high performing MNPs display higher mRNA consumption over time. (B) Relative Cy5 MFI (MFI normalized to APC 0/0 values) as measured at 4 h and 24 h post-transfection in HEK293T cells. (C) Scheme for inhibition of proton-sponge mechanism endosomal escape via incubation in Bafilomycin-A1. (D) Plot showing the drastic decrease in GFP expression upon BAF-A1 incubation demonstrating the reliance on high endosomal escape efficiency for high performing MNPs. U = Untreated, N = Naked mRNA, J = JetPEI (positive control). Error bars represent standard deviation (n = 3).

Given that cationic nanoparticles predominantly utilize endosome-mediated pathways for cellular internalization, − successful nucleic acid delivery critically hinges upon overcoming endosomal entrapment. , A key strategy to facilitate efficient endosomal escape involves enhancing the buffering capacity of delivery vehicles, thereby triggering the proton sponge effect. In this process, protonatable amine groups within nanoparticles buffer against endosomal acidification, resulting in counterion influx (primarily Cl^–) and osmotic swelling, ultimately compromising endosomal integrity and enabling cargo release into the cytosol. For the mixed micelles, the combination of acidic CEA amphiphiles with A7 amphiphiles (already display broad buffering capacity) , likely enhances endosomal escape via the proton sponge effect. To experimentally validate this hypothesis, we employed bafilomycin-A1 (Baf-A1), an inhibitor of V-ATPase proton pumps, to block endosomal acidification (Figure C). Treatment with Baf-A1 substantially reduced GFP MFI across all tested formulations, confirming the crucial role of endosomal acidification in mediating efficient cytosolic delivery (Figure D). While all micelles containing anionic amphiphiles exhibited dramatic reductions (>90%) in GFP MFI following Baf-A1 treatment, detailed analysis further emphasized this effect, revealing that the top-performing ternary micelle formulations (e.g., APC 0/30 and APC 5/30) experienced approximately 7-fold greater reductions in GFP MFI relative to APC 0/0 homomicelle control. On the other hand, binary formulations containing only PEG as the neutral amphiphile demonstrated considerably lower sensitivity to Baf-A1 treatment, experiencing only about a 50% reduction in GFP MFI. These results suggest that incorporating anionic amphiphiles significantly enhances the buffering capacity of mixed micelles, synergizing with the intrinsic buffering capabilities of polyamine-rich A7 amphiphile to amplify proton sponge-mediated endosomal escape. This enhanced buffering directly correlates with the substantial increase in functional mRNA delivery, as evidenced by the pronounced elevation in GFP expression (MFI) observed for anionic-containing formulations.

Overall, the exceptional performance of anion-doped mixed micelles can be attributed to a synergistic mechanism involving enhanced cellular uptake and endosomal escape. We propose that the anionic amphiphile drives micelle aggregation and sedimentation in culture media, promoting rapid and efficient cellular uptake. Once internalized, its low pK _a (∼5.2) value complements the broad buffering capacity of A7 (pK _a = 5.1, 8.6), synergistically enhancing the proton sponge effect to facilitate endosomal escape. This is evidenced by the pronounced Baf-A1 sensitivity and sharp drop in Cy5 signal at later time points, indicating efficient and rapid mRNA translation. Altogether, these findings establish that anion-driven modulation of micelle assembly, uptake, and endosomal trafficking is key to achieving potent and broadly effective mRNA delivery with ternary mixed micelle systems.

PEG Amphiphile Incorporation Enhances In Vitro Viability

To evaluate how amphiphile composition influences cellular viability and overall delivery performance, we performed CCK-8 viability assays concurrently with the transfection experiments. The CCK-8 assay is based on the enzymatic reduction of a tetrazolium-based salt by NADH-dependent oxidoreductases present in viable cells. This generates a formazan product, which is detectable by its strong absorbance at 450 nm, allowing the determination of relative viability by comparing MNP treated cell populations against untreated controls. Analysis of the CCK-8 results (Figure A) indicated that incorporation of anionic amphiphiles into binary mixed micelles (APC 0/10, 0/20, and 0/30) negatively impacted cellular viability relative to APC 0/0 homomicelles, indicating potential cytotoxic effects associated with increased anionic incorporation. Since higher anionic content necessitates the addition of more total micelle to maintain the N/P ratio at 10, this inadvertently increases the overall polymer dose resulting in increased toxicity. Conversely, binary mixed micelles containing PEG amphiphiles (APC 5/0, 10/0, and 15/0) demonstrate enhanced viability compared to both APC 0/0 homomicelles and anion containing binary mixtures (APC 0/10, 0/20, and 0/30). This observation aligns with existing literature, suggesting that PEG incorporation mitigates cytotoxicity through its ability to effectively shield excessive positive surface charges, thereby reducing detrimental cellular interactions. Finally, all ternary mixed micelles containing both PEG and anionic amphiphiles exhibited viability trends aligning closely with PEG-containing binary micelles, suggesting that PEG’s protective effects persist within these ternary mixtures, mitigating the detrimental effects of increased polymer load.

6.

(A) Relative viabilities of HEK293T cells upon transfection with the full ternary micelle library as determined via CCK-8 assay. (B) Effective efficiencies (transfection efficiency × viability) of the different ternary mixed MNPs. U = Untreated, N = Naked mRNA, J = JetPEI (positive control). Error bars represent standard deviation (n = 3).

To integrate the competing impacts of transfection efficiency and viability on overall micelle performance, we introduced the concept of effective efficiency (E_eff), which is defined as the product of transfection efficiency and cell viability (Figure B). Examining E_eff across different formulation subsets, three distinct trends emerged: First, the incorporation of anionic amphiphiles alone (APC 0/10, 0/20, and 0/30) enhanced transfection efficiency but concurrently decreased viability, resulting in effective efficiencies comparable to APC 0/0 homomicelles. Second, PEG incorporation alone (APC 5/0, 10/0, and 15/0) improved cell viability but correspondingly reduced transfection efficiency, again yielding effective efficiencies similar to APC 0/0 homomicelles. Finally, ternary formulations (APC XX/YY series) uniquely demonstrated a synergistic benefit from incorporating both anionic and PEG amphiphiles. Anionic amphiphile presence improved transfection efficiency, while PEG incorporation preserved or improved viability, cumulatively driving a marked increase in effective efficiency. Among these ternary formulations, the APC 15/10, 15/20, and 15/30 series displayed the highest E_eff, achieving nearly a 2-fold enhancement (∼0.75 E_eff) over native APC 0/0 homomicelles (∼0.4 E_eff). This enhancement in E_eff far outperforms the commercial transfection standard JetPEI (∼0.6 E_eff). Collectively, these findings underscore the necessity of balancing amphiphile composition to achieve optimal transfection performance, highlighting that strategic integration of PEG and acidic anionic amphiphiles synergistically enhances the overall efficacy and biocompatibility of polymeric micelle-based delivery systems.

Ternary Mixed Micelles Show Tunable Organ Tropism In Vivo

To assess the in vivo performance and biodistribution of the mixed micelle formulations, we employed luciferase-encoding mRNA as a bioluminescent reporter. MNPs were formulated at an N/P ratio of 10 and administered intravenously to immune-competent BALB/c mice at a dose of 0.5 mg/kg via tail-vein injection. Bioluminescence imaging was performed to quantify luciferase expression, serving as a direct readout of mRNA delivery and translation efficiency across anatomical sites. Whole-body luminescence analysis revealed a composition-dependent decrease in total signal upon incremental incorporation of PEG or anionic amphiphiles (Figure A,B). However, despite this global attenuation in signal in mixed formulations (Figure B), striking shifts in organ-specific biodistribution emerged, underscoring the impact of amphiphile composition on nanoparticle tropism (Figure A). Homomicelles composed solely of the cationic amphiphile (APC 0/0) exhibited strong pulmonary accumulation, with strong luminescence localized to the thoracic region. Binary A7-PEG micelles (like APC 5/0) maintained this lung targeting ability but showed a stepwise reduction in intensity with increased PEG content. In contrast, ternary micelles incorporating anionic amphiphiles demonstrated a pronounced redistribution of organ tropism. Intermediate anionic content (like APC 5/10 and 10/20) produced dual-site localization in both thoracic and splenic regions, while the highest anionic formulation (APC 10/30, 15/30) exhibited predominant splenic accumulation (Figure A, C). To quantify this shift, formulations showing elevated splenic signal (e.g., APC 0/30, 5/30, 10/30, and 15/30) were selected for follow-up organ-level biodistribution studies (Figure C,D). APC 0/0 homomicelles demonstrated strong pulmonary targeting, with approximately 2 orders of magnitude higher luminescence in the lungs (2 × 10⁵ p/s/cm²/sr) compared to the spleen (1 × 10³ p/s/cm²/sr) (Figure C). In contrast, formulations with increasing anionic content progressively inverted this distribution. Notably, APC 0/30 yielded equivalent luminescence in both lungs and spleen (∼4 × 10⁴ p/s/cm²/sr), whereas APC 15/30 exhibited a pronounced splenic preference, with almost 1.5 orders of magnitude higher signal in the spleen (∼6 × 10⁴ p/s/cm²/sr) relative to the lungs (∼2 × 10³ p/s/cm²/sr) (Figure C). The spleen-targeting capability of these ternary mixed micelles is illustrated in Figure D, which shows pronounced spleen-localized luminescence in contrast to the lung-specific luminescence observed for the APC 0/0 homomicelle control. These results establish that oppositely charged amphiphile blending enables tunable reprogramming of organ specificity from lung to spleen.

7.

(A) Representative images for in vivo average luciferase radiance obtained via an IVIS imager for Fluc mRNA delivery with ternary mixed MNPs via IV tail vein injection. (B) Fluc radiance distribution obtained from whole animal imaging. (C) Organ level (lungs and spleen) Fluc radiance distribution for select formulations showed high levels of splenic radiance during whole animal imaging (n = 3). (D) organ-level biodistribution image showcasing lung selective mRNA delivery by APC 0/0 MNP formulation and spleen selective mRNA delivery by APC 0/30, 5/30, 10/30, and 15/30 MNP formulation. T = Thymus, S = Spleen, H = Heart, K = Kidney, Li = Liver, Lu = Lungs. U = Untreated, N = Naked mRNA, J = JetPEI in vivo (positive control). Error bars represent standard deviation (n = 3).

Similar trends in biodistribution upon anionic doping have been reported for lipid nanoparticle systems, suggesting that changes in surface charge may mediate organ targeting. , However, zeta potential measurements in our system consistently demonstrated a uniformly high positive surface charge (∼+45–55 mV) across all formulations, indicating that net surface charge of micelles is not the dominant driver of tropism in this context. Furthermore, all MNPs were formulated at the same N/P ratio, which is defined exclusively by the stoichiometric ratio of the cationic amphiphile to anionic mRNA and is therefore independent of the amounts of neutral or anionic amphiphiles. This maintains a constant overall charge balance across the formulation library. Taking all these factors together, we hypothesize that nanoparticle morphology, together with concurrent changes in particle size and deformability, may play a central role in dictating the observed organ selectivity. Several reports have demonstrated that particle shape, size and deformability strongly impact in vivo splenic distribution: rigid or irregular particles are often trapped within the spleen, whereas softer, more deformable structures can evade the reticuloendothelial system (RES) and circulate more broadly. ,, Our own TEM and DLS analyses revealed a transition from small spherical micelles to larger, elongated wormlike structures with increasing anionic amphiphile incorporation, accompanied by an increase in particle size (up to ∼400 nm by DLS). Furthermore, TEM analysis of the mRNA-bound micelleplexes confirmed the persistence of these morphological differences, where lung-targeting APC 0/0 homomicelles formed spherical aggregates (Figure E­(vi)), while APC 10/30 micelleplexes displayed wormlike or tubular structures (Figure E­(x)), indicating that micelle morphology is retained upon complexation and represents the form encountered by organs during biodistribution. We propose that this morphological and size heterogeneity dictates the observed shifts in biodistribution: smaller spherical micelles (APC 0/0) preferentially localize to pulmonary capillaries, whereas larger, anisotropic worm-like structures (like APC 10/30 and APC 15/30) are more readily sequestered by the splenic reticuloendothelial system. This model also explains the dual-site localization observed in APC 10/10 and APC 10/20 formulations, which likely consist of mixed morphologies, where smaller spherical micelleplexes readily slip through the RES and accumulate in the lungs and larger, elongated aggregates are preferentially retained and cleared by the splenic RES (Figure E­(viii-ix)). Collectively, these results underline the complex interplay between amphiphile composition, nanoparticle morphology and size, and resultant biodistribution patterns. They further emphasize the importance of morphological control as a key factor in optimizing targeted organ-specific delivery. The ability to precisely modulate nanoparticle morphology through strategic amphiphile selection and incorporation represents a potent approach for enhancing the precision and efficacy of systemic nucleic acid delivery platforms, paving the way for tailored therapeutic strategies.

Optimizing in Vitro Delivery Performance in Serum Presence

While the observed organ tropism can be rationalized by morphology-dependent biodistribution, the ternary mixed micelles also exhibited a global attenuation in overall signal in vivo, prompting further investigation. One likely contributor is interaction with serum proteins, which are known to reduce the performance of cationic polymer–nucleic acid complexes by inducing premature disassembly, reducing payload protection, and limiting effective cellular uptake. , To assess how mixed micelle composition influences serum tolerance and functional mRNA delivery under physiologically relevant conditions, we performed in vitro transfection studies in HEK293T cells using OptiMEM supplemented with 10% fetal bovine serum (FBS) as a controlled proxy for serum exposure in vivo (Figure A­(i)). While these experiments evaluate functional delivery outcomes in the presence of serum, they do not directly measure particle integrity or RNA protection, and conclusions regarding physicochemical stability are therefore limited. For delivery under serum conditions ((−) OptiMEM maturation), MNP formulations were directly diluted 3× in OptiMEM containing 10% FBS rather than OptiMEM alone, while the remaining transfection protocol described in the Materials and Methods section was unchanged. Flow cytometry analysis revealed significant differences in performance under serum conditions compared to serum-free conditions (Figure B–C, (−) OptiMEM maturation). Incremental incorporation of PEG amphiphile into binary micelles (APC 5/0, 10/0, and 15/0) did not yield additional performance improvements relative to APC 0/0 homomicelles. However, binary and ternary micelles containing anionic amphiphiles exhibited improved transfection efficiency (∼70–80% GFP+ve cells) over APC 0/0 homomicelles (∼50% GFP+ve cells). Notably, this improvement was maximal at intermediate levels of anionic doping (10–20 mol % CEA) and diminished at higher CEA content (30 mol %), where transfection efficiency declined back to ∼50% GFP+ve. Analysis of GFP MFI revealed a pronounced serum-dependent attenuation across all formulations (Figure C). Relative to serum-free conditions, where GFP MFI spanned approximately 500–4000 across the formulation library, MFI values decreased by ∼80% under serum conditions to ∼100–350. This reduction was particularly severe for ternary mixed formulations, which exhibited a decrease from 3000 to ∼300 MFI, compared to a more modest reduction for APC 0/0 homomicelles (∼500 to ∼250 MFI). These results suggest reduced functional intracellular payload delivery in the presence of serum. Collectively, these results suggest that excessive incorporation of PEG and anionic amphiphiles may lower effective micelle charge density, which could promote premature decomplexation in the presence of serum proteins. While micelle morphology governs organ-selective biodistribution in vivo, the magnitude of the overall expression signal appears to be dictated by micelle complexation and decomplexation dynamics under serum-rich physiological conditions.

8.

(A) Scheme for MNP transfection under (i) normal serum conditions and (ii) serum conditions but with OptiMEM maturation. (B) Percentage of HEK293T cells positive for GFP expression under two different serum-based transfection conditions. (C) GFP MFI of GFP positive cells under two different serum-based transfection conditions. U = Untreated, N = Naked mRNA, J = JetPEI (positive control). Error bars represent standard deviation (n = 3).

To address this serum-associated reduction in functional mRNA delivery in vitro, we employed an innovative strategy leveraging controlled micelleplex aggregation. Previously, we observed rapid aggregation of mixed micelles upon dilution into serum-free OptiMEM. Leveraging this behavior, we developed a novel strategy where micelles were first allowed to aggregate in serum-free OptiMEM, after which serum-supplemented medium was introduced (Figure A­(ii)). For this OptiMEM maturation strategy, MNP formulations were first diluted 2× in OptiMEM and incubated for 15 min to allow preaggregation, followed by further dilution with OptiMEM containing 20% FBS to achieve a final 3× dilution and 10% total FBS concentration, while the remaining transfection protocol was unchanged. By initially inducing micelle aggregation in serum-free conditions, followed by subsequent serum introduction, we hypothesized that serum proteins would preferentially opsonize micellar aggregates rather than individual micelles, thereby improving functional mRNA delivery. Implementing this strategy, transfection outcomes demonstrated performance trends closely resembling those of serum-free experiments, with significant improvements in transfection efficiency and GFP MFI observed for formulations incorporating higher levels of anionic amphiphiles (Figure B–C, (+) OptiMEM maturation). Particularly, the APC 10/30 formulation exhibited superior performance, surpassing both other tested formulations and the commercial JetPEI control. Viability remained high (>70%) across all formulations, though slight reductions were noted for formulations with higher anionic incorporation that can be attributed to MNP aggregate sedimentation (Figure S18). DLS characterization of serum-induced aggregate proved challenging, as interference from serum protein size distributions produced unreliable autocorrelation measurements. Nevertheless, robust transfection outcomes strongly support the effectiveness of our aggregation-based strategy (referred to as OptiMEM maturation), underscoring its potential to mitigate serum-associated reductions in transfection performance. Collectively, these findings emphasize the critical interplay between micelle composition, serum interactions, and controlled aggregation. This strategic aggregation of micelles prior to serum exposure presents a compelling approach for optimizing polymeric micelle-based nucleic acid delivery in physiologically relevant in vitro environments.

Conclusion

In summary, this work establishes amphiphile composition as a central design parameter governing the structure, stability, and biological performance of micellar polymeric nanoparticles for mRNA delivery. By systematically varying cationic oligoamine amphiphile, acidic anionic amphiphile (CEA, SEA, BEA, PEA), and neutral PEG amphiphile mixing ratios, we delineated clear structure–function relationships linking micelle composition to physicochemical properties, cellular interactions, and in vivo biodistribution. Progressive incorporation of anionic amphiphiles drove a dose-dependent morphological transition from spherical to elongated, wormlike micelles, revealing morphology as an emergent and tunable feature of ternary mixed systems. In vitro, incorporation of acidic anionic amphiphiles markedly enhanced transfection efficiency and GFP expression across multiple cell lines (HEK293T, A549, HuH7), driven by increased uptake, efficient intracellular release, and improved endosomal escape. Notably, amphiphiles with lower pK _a values (CEA and SEA) consistently outperformed near-neutral analogues (BEA and PEA), underscoring the importance of amphiphile acidity in enhancing mRNA delivery and expression. Concurrent PEG incorporation mitigated cytotoxicity through charge shielding, yielding optimized ternary formulations (APC 10/30 and APC 15/30) that achieved the highest effective efficiencies by balancing transfection performance with cell viability. Furthermore, during in vitro transfection in serum presence, introducing controlled preaggregation significantly improved serum resistance, preserving cargo stability, and ultimately displaying superior transfection efficiency and GFP expression levels comparable to serum-free conditions, highlighting a practical strategy for improving serum robustness. Finally, in vivo biodistribution revealed a clear morphology-dependent shift in organ tropism from lung to spleen as noncationic amphiphile content increased. While spherical, cationic micelles preferentially displayed thoracic luminescence, mixed micelles with elongated wormlike morphology showed splenic luminescence. Notably, this transition occurred despite comparable positive zeta potentials across formulations, indicating that micelle morphology, rather than surface charge, governs organ selectivity. Transmission electron microscopy confirmed the correlation between wormlike morphology and splenic accumulation, establishing morphology as the key determinant of biodistribution and highlighting a tunable, composition-driven strategy for directing polymeric micelles to specific organs. Collectively, these findings demonstrate that deliberate control over amphiphile composition and careful manipulation of micelle morphology enables simultaneous optimization of in vitro performance and in vivo organ selectivity. This blended micelle platform provides a versatile framework for designing polymeric mRNA delivery systems with enhanced efficacy and programmable biodistribution, advancing the rational development of next-generation nucleic acid therapeutics.

Materials and Methods

Synthesis of Poly­(nBA)-b-poly­(A7 acrylamide) Amphiphilic Diblock Polymers

All reagents were purchased from Sigma-Aldrich or BocSciences and used without further purification. Pentafluorophenol acrylate was synthesized and used as previously reported. Poly­(nBA)-b-poly­(A7 acrylamide) was synthesized following previously reported procedures. Briefly, the poly­(nBA) macro-chain transfer agent (poly­(nBA)-DDMAT) was prepared by bulk RAFT polymerization of n-butyl acrylate (nBA) using 2-(dodecylthiocarbonothioylthio)-2-methylpropionic acid (DDMAT) as the chain transfer agent. After purification and characterization, poly­(nBA)-DDMAT macroCTA (1 equiv) was weighed into a 10 mL scintillation vial, and pentafluorophenyl acrylate (PFPA, 300 equiv) was added. The mixture was dissolved in 1,4-dioxane to achieve a final monomer concentration of 2 M, yielding a yellow homogeneous solution. Azobisisobutyronitrile (AIBN, 0.05 equiv) was then introduced, and the solution was sparged with nitrogen gas for 30 min to remove dissolved oxygen. The reaction mixture was stirred at 70 °C for 18 h under a nitrogen atmosphere. The resulting diblock copolymer, poly­(nBA) _n -b-poly­(PFPA) _m , was purified by precipitation in methanol. Block composition was confirmed by integration of diagnostic signals in the ¹H NMR spectrum (CDCl₃): 4.04 ppm (2H, poly­(nBA)), 3.08 ppm (1H, poly­(PFPA)), and 0.94 ppm (3H, poly­(nBA)).

To introduce the functional amine group, 200 mg of purified poly­(nBA)-b-poly­(PFPA) was dissolved in 5 mL of anhydrous DMF and stirred until fully homogeneous. Hydroxyethyl acrylate (20 equiv) was added as a capping agent, followed by addition of 2n equivalents of Boc-protected A7 amine, where n corresponds to the number of PFPA repeat units. The reaction vial was sealed, purged with nitrogen, and stirred at 40 °C for 18 h. Upon completion, the reaction was cooled to room temperature, and an aliquot was analyzed by ¹⁹F NMR to confirm the disappearance of PFPA signals, indicating successful substitution. For Boc deprotection, the crude polymer was dissolved in trifluoroacetic acid (TFA) at a ratio of 1 mL TFA per 100 mg of polymer and stirred open to air for 4.5 h. The reaction mixture was subsequently dialyzed against deionized water using a SnakeSkin dialysis membrane (MWCO 3.5 kDa) to remove excess acid and byproducts. The deprotected A7-functionalized amphiphile was recovered by lyophilization, and purity was verified via ¹H NMR spectroscopy.

Synthesis of Poly­(nBA)-b-poly­(anionic) Amphiphilic Diblock Polymers

The anionic amphiphiles were prepared via chain extension on poly­(nBA)-DDMAT macroCTA using RAFT polymerization. poly­(nBA)-DDMAT macroCTA was synthesized by the bulk polymerization of n-butyl acrylate (nBA) with 2-(Dodecylthiocarbonothioylthio)-2- methylpropionic acid (DDMAT) CTA as previously reported. The anionic monomer (Carboxyethyl acrylate (CEA), or 4-(3-(Acryloyloxy)­propoxy)­benzoic acid (BEA), or Mono­(2-acryloyloxyethyl) Succinate (SEA), or 2-(phosphonooxy)­ethyl acrylate­(PEA)) (300 Eq, purified on an alumina plug), and poly­(nBA)-DDMAT macroCTA (1 equiv) were first weighed into a 20 mL scintillation vial, dissolved in DMF (1M), and AIBN (0.05 Eq) was added to it. The reaction mixture was degassed by sparging with N ₂(g) for 30 min and subsequently immersed in an oil bath that was preheated at 70 °C. The reaction was stirred at 70 °C for 18h under N ₂(g) atmosphere and subsequently quenched by immersing the flask in liquid nitrogen and opening to air simultaneously. The polymer was purified via dialysis in methanol. After drying under vacuum, the polymer was recovered as an off-white powder.

Synthesis of Poly­(nBA)-b-poly­(ethylene glycol) Amphiphilic Diblock Polymers

The PEG-based chain transfer agent (PEG macroCTA) was synthesized following previously reported procedures with minor modifications. In a typical reaction, PEG–OH (M _n = 10 kDa, 1 equiv), DDMAT (2 equiv) and 4-dimethylaminopyridine (DMAP, 1 equiv) were dissolved in 10 mL of anhydrous dichloromethane (DCM) and cooled in an ice bath to 0 °C. Dicyclohexylcarbodiimide (DCC, 2 equiv), predissolved in DCM, was added dropwise under stirring. After 2 h at 0 °C, the reaction mixture was allowed to warm to room temperature and stirred for an additional 70 h. The crude product was precipitated using cold diethyl ether and collected by filtration. The resulting PEG­(10K)-DDMAT macroCTA was dried under high vacuum overnight to afford a yellow solid (yield: 95%).

RAFT polymerization of n-butyl acrylate (nBA) was then carried out using PEG­(10K)-DDMAT macroCTA. Specifically, PEG-CTA (1 equiv), n-butyl acrylate (140 equiv, prepurified by passage through a neutral alumina plug), and azobisisobutyronitrile (AIBN; 0.05 equiv) were weighed into a 20 mL scintillation vial. The reaction mixture was degassed by nitrogen sparging for 1 h and subsequently transferred to a preheated oil bath at 70 °C. Polymerization was allowed to proceed for 45 min under a nitrogen atmosphere before quenching by immersion in liquid nitrogen while simultaneously exposing the reaction to air. The resulting block copolymer was purified by dialysis in methanol, followed by vacuum drying to yield a yellow solid.

Mixed Micelle Formulation and Characterization

Micelles were prepared via the thin-film hydration method. First, individual stock solutions of each amphiphile were prepared by dissolving the dry powder in methanol to achieve 1 mg/mL concentration. Appropriate amounts of each amphiphile stock (A7, PEG, and Anionic (CEA, SEA, BEA, or PEA)) were mixed in a vial to achieve the desired mixing ratio. Subsequently, these mixtures were dried under N₂ stream to yield polymer films. Finally, the polymer films were hydrated with deionized RNase and DNase free water to obtained 1.5 mg/mL polymer dispersions. The dispersions were stirred at room temperature for 3 days and sonicated for 30 min at every 24-h interval using an ultrasonic bath. During the sonication process the temperature of the bath was monitored and never exceeded 30 °C. The size distribution of the mixed micelles was measured with a Malvern Zetasizer Nano ZS (λ = 633 nm) with a scattering angle of 173° at a temperature of 25 °C.

For TEM imaging, images of polymer micelles were obtained using a FEI Tecnai Spirit Bio-Twin TEM. Carbon coated mesh 300 copper grids (Electron Microscopy Sciences) were glow-discharged using a PELCO glow discharged. 10 μL of polymer micelles or micelleplexes (MNPs) at a concentration of 0.25 mg/mL were deposited on the grid. The grid was blotted after 30 s to remove excess sample. μL of 2% uranyl formate solution were deposited on the grid and blotted after 30 s. The grid was washed using Milli-Q water to remove excess UA. The grid was left to dry for 30 min. The grid was then mounted on a single tilt sample holder. The microscope was operated at 120 kV with a magnification of 10,000–120,000x to assess particle morphology and size distribution.

Micellar Nanoparticle Formulation and Characterization

MNP solutions were prepared by diluting micelle and mRNA components in DNA/RNase free water. All micelle stocks were diluted down to a final concentration of 3 nMol of nitrogen per μL. These stocks were then diluted to appropriate concentration necessary so that upon mixing (equal volumes) with diluted mRNA solution, the appropriate N/P is achieved. For example, 5 μL of the polymer stock (3 nMol N/ul) is necessary for every 0.5 μg of plasmid at an N/P 10. To make a polyplex sample containing 1.5 μg of plasmid at N/P 10 for a transfection, 15 μL of polymer stock was diluted in 60 μL of PBS to yield 75 μL of micelle stock. For mRNA at 1 mg/mL concentration, 1.5 μL of mRNA stock was added to 73.5 μL PBS. To form the MNPs, 75 μL of micelle stock was added to 75 μL mRNA stock, to total of 150 μL, mixed, and then allowed to incubate at room temperature for 45 min prior to use.

The hydrodynamic diameter (D _h) of the thus formed MNPs in water was determined via DLS using a DynaPro Plate Reader at a temperature of 25 °C. For determining their size distribution upon dilution in cell culture media, the MNPs were first formed in water, their size distributions recorded and then diluted (3X) with OptiMEM. The OptiMEM diluted MNPs were incubated further for 2 h prior to DLS measurements. The interval of time between the addition of media and sample measurement was kept constant for all samples.

RiboGreen Dye Exclusion Assay

Dye exclusion assays were performed using Quant-iT RiboGreen RNA reagent (ThermoFisher Scientific). mRNA solutions were prepared by diluting the primary mRNA stock (1 mg/mL) into RNase free water containing RiboGreen to obtain a final dye concentration of 0.5% v/v. Equal volumes of this RiboGreen-labeled mRNA solution and micelle dilutions were mixed to form micelle–mRNA polyplexes (MNPs), following aforementioned protocols. The N/P ratio was maintained at N/P 10 for all the MNPs in the library. Controls included mRNA with RiboGreen in the absence of polymer and RiboGreen in water alone. Aliquots (100 μL) of each formulation were dispensed in triplicate into black, flat-bottom 96-well microplates. Fluorescence was measured using a microplate reader (λ_ex= 485 ± 10 nm, λ_em = 528 ± 10 nm). Raw fluorescence values were background-subtracted using dye-free controls and normalized to the polymer-free sample (N/P = 0). Each well was subsequently diluted 3-fold with serum-free OptiMEM, and fluorescence end points were recorded 30 min postdilution. To assess mRNA release, 10 μL of heparan sulfate (1.875 mg/mL in water; target S/N = 10) was added to each well (containing 100 μL of MNPs), and fluorescence increase was recorded using the same parameters. The difference in normalized fluorescence intensity before and after heparin treatment was used to quantify mRNA release from the MNP complexes.

Delivery Reporter Assay

GFP-encoding mRNA was utilized as a reporter to assess in vitro delivery efficiency in HEK293T, HuH7, and A549 cell lines. Unless otherwise indicated, identical procedures were followed for all three cell types. Cells were seeded in 48-well plates at a density of 25,000 cells per well 24 h prior to transfection. Micelle-based nanoparticles (MNPs) were prepared as previously described and diluted 3-fold in serum-free OptiMEM before transfection. The culture media was aspirated, and 150 μL of the diluted MNP formulation was added to each well in triplicate, followed by incubation at 37 °C for 4 h. As a representative formulation, 75 μL of A7 micelle stock was mixed with 75 μL of mRNA solution (containing 1.5 μg of mRNA), incubated at room temperature for 45 min, and subsequently diluted with 300 μL of OptiMEM to reach a total volume of 450 μL. This solution was then distributed across three wells (150 μL per well), yielding a final mRNA dose of 0.5 μg per well. Following the 4 h incubation, 500 μL of cell-type-specific complete growth medium supplemented with 10% fetal bovine serum (FBS) was added to each well. A positive control using jetPEI (Polyplus) was included and prepared according to the manufacturer’s protocol at an N/P ratio of 5. All experimental formulations were tested at an N/P ratio of 10 with an mRNA dose of 0.5 μg per well. All biological experiments were performed in triplicate, and error bars in all figures represent the standard deviation.

Twenty-four hours post-transfection, cell viability was assessed using a CCK-8 assay (Dojindo). For this, CCK-8 reagent was diluted into FluoroBrite DMEM supplemented with 10% FBS to a final concentration of 10% (v/v). The media in each well was aspirated and replaced with 500 μL of the CCK-8 working solution, followed by 2 h incubation at 37 °C. Subsequently, 150 μL of media from each well was transferred to a clear-bottom 96-well plate, and absorbance was measured at 450 nm using a Synergy H1 Hybrid microplate reader (BioTek, Winooski, VT). Cell viability was calculated by normalizing absorbance values to untreated control wells.

Following viability assessment, cells were harvested for flow cytometry. Media was removed, and adherent cells were trypsinized and transferred to 2 mL deep-well plates. Cells were washed with PBS, centrifuged at 1000 rpm for 10 min at 4 °C, and resuspended in PBS containing 2% FBS and Calcein Violet AM to label live cells. GFP-positive cell populations were quantified using a ZE5 flow cytometer (Bio-Rad, Hercules, CA) equipped with 405 and 488 nm lasers. Data was analyzed using FlowJo software (BD Biosciences); representative gating strategies are provided in Figure S19.

Delivery Assays in Serum

For delivery in the presence of serum the methodology outlined above was followed with two different modifications. For (−) OptiMEM maturation strategy, the MNP formulations were 3X diluted with OptiMEM containing 10% FBS instead of normal OptiMEM. For (+) OptiMEM maturation strategy, the MNP formulations were 2X diluted with OptiMEM and allowed to incubate for 15 min, followed by addition of 1X OptiMEM containing 10% FBS, to a total of 3X dilution. The rest of the protocol was kept the same.

Cellular Uptake Assay

For cellular uptake studies, Cy5-labeled GFP-encoding mRNA (EZ Cap Cy5 EGFP mRNA [5-moUTP], ApexBio Technology) was employed as the fluorescent reporter. The transfection protocol was identical to that described above, with the exception that, during flow cytometry sample preparation, cells were washed three times with CellScrub buffer (Amsbio) to facilitate removal of cell surface-associated MNPs. Internalization was assessed at 4 and 24 h post-transfection by quantifying Cy5 signal via flow cytometry. Specifically, the percentage of live cells positive for Cy5 fluorescence, as well as the mean fluorescence intensity (MFI), was measured using a ZE5 flow cytometer (Bio-Rad, Hercules, CA) equipped with a 640 nm laser. Representative gating strategies used to define live and Cy5-positive populations are provided in Figure S19. All uptake experiments were conducted at an N/P ratio of 10 with an mRNA dose of 0.5 μg per well.

Bafilomycin-A1 Inhibition Assay

Transfections for the bafilomycin A1 (Baf-A1) treatment group followed the same workflow as previously described, with minor modifications. For the Baf-A1 (+) condition, culture medium was removed and replaced with 500 μL of DMEM containing 200 nM Baf-A1 10 min prior to transfection. After the brief preincubation period, the Baf-A1-containing media was aspirated, and transfection was performed as outlined above. Cells were harvested and analyzed via flow cytometry 24 h post-transfection. The positive control jetPEI was included at an N/P ratio of 5. All experimental conditions were carried out at an N/P ratio of 10 with a final mRNA dose of 0.5 μg per well.

In Vivo Studies

All animal studies were approved by the Institutional Animal Care and Use Committee (IACUC, protocol number -2207-40246A) at the University of Minnesota. Mice were housed within the University of Minnesota’s Research Animal Resource facilities, an AAALAC accredited program. Specific Pathogen Free female Balb/C mice were purchased from The Jackson Laboratory (Strain #:000651). Animals were dosed at 8 weeks old with either PBS, MNP formulations, or JetPEI carrying 10 μg luciferase mRNA in a total of 200 μL via tail vein injection. Six hours after injection, delivery was assessed by in vivo bioluminescence imaging on the IVIS Spectrum 2 (University of Minnesota Imaging centers, SCR_020997, NIH grant 1S10OD032261). Briefly, animals are intraperitoneally injected with D-luciferin substrate (100 μL total volume, 28 mg kg^–1). Animals are then anesthetized with inhaled isoflurane and placed in the imager where luminescent signal is measured 5 min after substrate injection. Image acquisition is set to 5 min exposure with fStop = 1 and Binning = 2.

For organ specific expression determination, mice were intraperitoneally injected with D-luciferin substrate, then euthanized by CO₂ inhalation for sample collection. The animals were perfused with PBS, and organs including the lungs, thymus, liver, spleen, heart, and kidneys were collected (6 h). Organs were imaged immediately after collection on the IVIS Spectrum 2. Bioluminescence data was analyzed using Living Image software (version 4.7.3).

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/jacsau.6c00068.

• Materials and instrumentation information; polymer synthesis and characterization information in the form of NMR spectra; potentiometric titration data; micelle and micelleplex DLS characterization data; detailed biophysical characterization data (dye exclusion assay, zeta potential data); additional in vitro data; flow gating schemes (PDF)

S.P. conceived the study, designed and conducted the chemical, biophysical, and biological experiments (in vitro and in vivo), performed data analysis, and wrote the manuscript. E.J.E. designed, conducted, and analyzed the in vivo studies and contributed to manuscript preparation. J.P.S., E.M.S., A.K.G., S.I., and J.T.B. assisted with in vivo experimentation. T.M.R. and B.S.M. supervised the research, contributed to experimental design, and assisted with manuscript preparation.

The authors declare no competing financial interest.