Structural Change of ApoE–Lipid Nanoparticles Alters Receptor-Mediated Endocytosis and Induces Endosomal Disruption

Abstract

Apolipoprotein E (ApoE) is a key regulator of lipid metabolism that binds to lipid nanoparticle (LNP) surfaces to mediate cellular interactions. However, the ApoE-LNP behavior is highly dependent on the LNP composition, and the underlying mechanisms remain unclear. Here, we show that subtle alterations in LNP surface lipids profoundly reshape the ApoE-LNP structure and intracellular trafficking. Using cryogenic electron microscopy and live-cell imaging, we demonstrate that replacing 10 mol % 1,2-distearoyl-sn-glycero-3-phosphocholine (DSPC) with a porphyrin–lipid conjugate induces highly faceted, irregular LNP membrane morphologies upon ApoE binding. These structural transitions alter receptor-mediated uptake, enhance endosomal disruption, promote cytosolic siRNA release, and ultimately trigger cell apoptosis. Through fluorescence lifetime imaging, we also demonstrate that ApoE promotes LNP intracellular disassembly. Overall, our findings identify lipid-driven structural remodeling of ApoE-LNPs as a critical determinant of their intracellular fate, offering mechanistic insights for the rational design of LNPs and new perspectives on ApoE’s role in disease pathogenesis.

Introduction

Apolipoprotein E (ApoE) is a multifunctional glycoprotein critical for lipid transport and cholesterol homeostasis, primarily mediating lipoprotein clearance through interaction with low-density lipoprotein receptors (LDLRs). − Structurally, ApoE exhibits great conformational plasticity, which allows it to modulate receptor binding depending on its lipidation state. Subtle structural variations in ApoE can, however, profoundly affect its biological function. For instance, the isoforms ApoE3 and ApoE4 differ by only a single amino acid, yet ApoE4 exhibits altered lipid binding and receptor interactions, making it a major genetic risk factor for late-onset Alzheimer’s disease (AD), while ApoE3 remains neutral. − These pathological consequences highlight the need to investigate how structural alterations in ApoE influence cellular interactions.

Recently, lipid nanoparticles (LNPs) have emerged as powerful delivery vehicles for RNA therapeutics, revolutionizing gene therapies and vaccines. − Upon systemic administration, LNPs rapidly adsorb plasma proteins to form a “protein corona”, predominantly enriched with apolipoproteins. , Among them, ApoE plays a pivotal role in mediating LNP biodistribution and cellular uptake, serving as a potent targeting ligand for LDLRs on hepatocytes. This mechanism underlies the clinical success of Onpattro, the first FDA-approved LNP-siRNA therapeutic for silencing abnormal transthyretin (TTR) protein production in the liver hepatocytes.

To fully understand ApoE’s role in guiding LNP–cell interactions, it is essential to elucidate the reciprocal impact between ApoE and LNPs. A recent study has demonstrated that ApoE binding can induce lipid composition redistribution between the core and shell of LNPs. Conversely, the lipid composition of LNPs can also affect their affinity for ApoE and subsequently modulate ApoE-mediated receptor interactions and cellular uptake. For instance, LNPs incorporating DOPE exhibit enhanced ApoE binding compared to those containing DSPC, despite having identical remaining lipid compositions. Furthermore, although ApoE–receptor interactions generally improve LNP intracellular uptake, the degree of enhancement significantly varies depending on specific ionizable lipids incorporated into LNP formulations. , In vivo studies have shown that LNPs can exhibit ApoE-independent biodistribution depending on ionizable lipid type , or the inclusion of charged lipids and their structures. , These findings suggest a complex interplay among LNP composition, ApoE binding affinity, and ApoE functionality, potentially driven by lipid-induced structural changes in ApoE. However, the structural alterations of ApoE upon association with distinct lipid environments remain largely unexplored. Furthermore, while ApoE’s extracellular roles in receptor engagement and cellular uptake are well documented, its intracellular functions after LNP internalization, particularly in key RNA delivery events like endosomal escape, are poorly characterized.

In this work, we systematically examine how lipid composition, specifically the incorporation of a model helper lipid, porphyrin-lipid, induces structural alteration of ApoE-coated LNPs and regulates their subsequent intracellular fate. Using cryogenic electron microscopy (cryo-EM), live-cell imaging, and biochemical analyses, we reveal that porphyrin-lipid incorporation uniquely reorganizes the ApoE on the LNP surface, producing faceted, irregular membrane morphologies. This structural change reduces LNP dependence on classic ApoE receptor-mediated uptake pathways while dramatically enhancing endosomal disruption, leading to cytosolic release of RNA payloads and triggering cell apoptosis. Collectively, our findings provide critical mechanistic insights into lipid-dependent modulation of ApoE-LNP structure and function, providing a foundation for the rational design of advanced LNP-based therapeutics and deeper understanding of ApoE’s biological roles.

Results

Distinct Intracellular Delivery Profiles of ApoE-Onpattro and ApoE-Porphyrin-LNP

In our study, we selected the clinically approved Onpattro formulation as a benchmark for comparison. Porphyrin-lipid is a fluorescent phospholipid which is well-compatible with LNP , and has been integrated as an imaging tool in the current work. siRNA-loaded Onpattro-LNPs and porphyrin-LNPs were both synthesized via a standard microfluidic mixing method. The only structural difference between Onpattro and porphyrin-LNP is the replacement of 10 mol % DSPC with porphyrin-lipid as the helper lipid (Figure a). Following synthesis, LNPs were incubated overnight with ApoE to generate ApoE-coated formulations. Live-cell imaging was then performed to investigate intracellular delivery across the following three cell lines: PC3-Luc6 (prostate cancer), Hep3B (liver cancer representing hepatic cells), and HDFn (human fibroblasts representing normal cells). As shown in Figure b,c, after 6 h of incubation, both ApoE-coated formulations exhibited significantly stronger intracellular fluorescence of FAM-labeled siRNA than their noncoated counterparts. However, notable differences were observed in the intracellular distribution of siRNA between the two ApoE-coated formulations. In ApoE-Onpattro, intense FAM-siRNA signals localized to the cell membranes and punctate endocytic organelles, indicative of strong ApoE–receptor interactions and consistent with a typical receptor-mediated endocytosis pattern. In contrast, ApoE-porphyrin-LNP showed minimal cell membrane association and reduced punctate endosomal accumulation. Instead, siRNA was predominantly distributed throughout the cytosol, most significantly in PC3-Luc6 cells and, to a moderate extent, in Hep3B and HDFn cells. Quantitative analysis further confirmed that ApoE-porphyrin-LNP enabled significantly higher average intracellular siRNA intensity in PC3-Luc6 cells compared to ApoE-Onpattro (Figure c), consistent with enhanced cytosolic siRNA release. Together, these data suggest that ApoE-porphyrin-LNPs follow a distinct cellular entry pathway compared to ApoE-Onpattro and facilitate the more efficient cytosolic delivery of siRNA.

1.

Distinct intracellular uptake profiles of ApoE-coated Onpattro and porphyrin-LNPs. (a) Schematic of the LNP surface composed of either DSPC or porphyrin-lipid and their corresponding chemical structure. (b) Live-cell imaging of FAM-siRNA (green) delivery in cells following a 6 h incubation of LNP at a 30 μM total lipid concentration. Markedly different uptake patterns are observed for Onpattro and porphyrin-LNP formulations with or without ApoE. (c) Average intracellular FAM-siRNA fluorescence intensity across cell lines after LNP uptake. The solid line is the median, and the dashed lines are the 25th and 75th percentiles. N = cells, indicated above the violins. (d) Experimental flow for quantifying LNP intracellular uptake via Gd-lipid labeling and ICP-MS. Gd³⁺ content in the digested cell samples was normalized to total cell counts. (e) Quantitative analysis of cellular uptake when LNPs were dosed at 12 μM total lipid, determined by Gd³⁺ measurements using ICP-MS; data are expressed as mean ± standard deviation (n = 3).

The enhanced intracellular siRNA fluorescence observed in ApoE-LNPs compared to their non-ApoE counterparts generally suggests higher intracellular uptake of the formulations. However, during live-cell imaging, probe fluorescence intensity does not always reflect its abundance, as it can also be affected by the probe’s local environment and aggregation state. To address this limitation, we employed inductively coupled plasma mass spectrometry (ICP-MS) to quantitatively measure the absolute intracellular uptake of the formulations (Figure d). All formulations were labeled with 2% Gd-chelating lipids to enable precise lipid quantification by ICP-MS following cellular uptake and complete cell digestion. The measured Gd content per cell directly reflects the total lipid and thus LNP that is internalized. As shown in Figures e and S1, ApoE coating significantly enhanced Onpattro uptake across all tested cell lines, yielding approximately 3-, 8-, and 2-fold increases in PC3-Luc6, Hep3B, and HDFn cells, respectively.

To validate whether the lipid and siRNA components exhibited consistent cellular uptake profiles, the siRNA sequences was labeled with Alexa-647 and incorporated into Onpattro formulations. Following treatment, cells were subjected to a digestion protocol using DMSO to ensure complete disruption of both the cells and internalized LNPs. The process released the siRNA and restored any fluorescence quenched within intact LNPs, allowing for accurate fluorescence quantification. The resulting uptake profile closely matched the lipid uptake determined by ICP-MS analysis (Figure S2), confirming coordinated cellular internalization of both siRNA and lipid components. Consistently, both assays support that ApoE coating markedly enhances the overall cellular uptake of Onpattro. However, ApoE coating did not substantially enhance the cellular uptake of porphyrin-LNP. As shown in Figures e and S1, ICP-MS quantification revealed that cellular uptake of ApoE-porphyrin-LNPs was nearly unchanged compared to noncoated porphyrin-LNPs across all three cell lines, yielding only 1.1-, 1.4-, and 0.9-fold change in PC3-Luc6, Hep3B, and HDFn cells, respectively. This finding was particularly intriguing given that cells treated with ApoE-porphyrin-LNP exhibited markedly stronger intracellular siRNA fluorescence compared to porphyrin-LNP alone (Figure b,c). Since the ICP-MS quantification reflects absolute LNP internalization, the enhanced siRNA intracellular fluorescence cannot be attributed to increased LNP uptake. Rather, it likely arises from altered intracellular distribution and aggregation states of the FAM-siRNA. Specifically, ApoE-porphyrin-LNP promoted cytosolic dispersion of FAM-siRNA, thereby generating bright, diffuse, and unquenched fluorescence. In contrast, siRNA delivered by porphyrin-LNP largely remained confined within LNPs or endosomal compartments, producing dim, punctate signals that were likely quenched.

To further validate this interpretation and elucidate the intracellular delivery process, time-lapse live-cell confocal imaging was performed with Cy3-siRNA-loaded LNPs. PC3-Luc6 cells were incubated with ApoE-Onpattro or ApoE-porphyrin-LNP for 6 h, followed by continuous imaging for 24 h. As shown in Figure a and Video S1, ApoE-Onpattro followed a classic endocytic trafficking route: siRNA initially localized on the cell membrane and gradually internalized into punctate, vesicle-like structures, indicative of sustained endocytosis. Additionally, bright-field imaging confirmed that cells remained healthy and viable throughout the imaging period. In contrast, ApoE-porphyrin-LNP exhibited clearly distinct intracellular dynamics (Figure b,c, Video S2). During the first 0–2 h, Cy3-siRNA demonstrated punctate distribution within endocytic organelles. Between 2 and 8 h, however, the siRNA signal became strongly unquenched and dispersed throughout the cytosol, suggesting efficient endosomal escape. Remarkably, from 8 h onward, cells began to exhibit morphological hallmarks of cytotoxicity, including nuclear condensation, swelling, and extensive membrane blebbing, ultimately leading to cell death by ∼18–24 h. Bright-field imaging confirmed extensive membrane blebbing and complete loss of viability at 24 h.

2.

Time-lapse imaging of ApoE-LNPs during intracellular uptake. PC3-Luc6 cells were incubated with ApoE-Onpattro (a) or ApoE-porphyrin-LNPs (b) that were loaded with Cy3-siRNA for 6 h at 30 μM total lipid concentration, followed by 24 h of time-lapse imaging. Inset in panel a shows the Cy3-siRNA fluorescence of the cell outlined by the red dashed box. (c) Representative zoomed-in view of a single cell from panel b (red dashed box), showing pronounced endosomal disruption, cytosolic siRNA release, and associated cell morphological changes induced by ApoE-porphyrin-LNP.

These time-lapse imaging results provide crucial insight into the unique intracellular dynamics of ApoE-porphyrin-LNP compared to ApoE-Onpattro. While ApoE-Onpattro undergoes typical receptor-mediated endocytosis and remains largely confined within endosomal compartments, ApoE-porphyrin-LNP induces pronounced endosomal disruption, enabling robust cytosolic release of siRNA (as observed in Figures b and c). This burst-like release into the cytosol led to an immediate increase in intracellular siRNA fluorescence, explaining why ApoE-porphyrin-LNP exhibits stronger cellular fluorescence despite no measurable increase in overall uptake. Notably, the endosomal disruption induced by ApoE-porphyrin-LNP may also liberate endosomal substrates that activate cell death-associated pathways, contributing to the observed cell death. Control experiments with porphyrin-LNP lacking ApoE (Video S3) showed neither cytotoxicity nor altered siRNA distribution under the same imaging conditions, ruling out phototoxicity from porphyrin excitation. Rather, these effects arise from the structural interplay between ApoE and specific surface porphyrin lipids in the LNP.

ApoE Incorporation Promotes Intracellular Disassembly of LNPs

To further elucidate the intracellular disassembly behavior of the ApoE-coated LNPs, fluorescence lifetime imaging microscopy (FLIM) was employed to monitor structural changes following cellular uptake. Using FAM-siRNA-loaded LNPs as an example, the FLIM workflow is presented in Figure a: Following intracellular uptake, the FAM-siRNA fluorescence intensity and lifetime images were captured simultaneously. The fluorescence decay profile at each pixel was transformed into coordinates on a phasor plot, where pixels with similar lifetimes cluster together and clusters with higher phase values correspond to longer lifetimes. Each cluster was then back-mapped to the corresponding intensity image to reveal spatial localization within distinct subcellular compartments. Because fluorescence lifetime reflects the local molecular environment of FAM-siRNA, specifically, whether it remains encapsulated within LNPs (short τ) or is released into the intracellular milieu (long τ), this approach enables spatially resolved mapping of LNP structural transitions inside cells.

3.

Fluorescence lifetime imaging of LNP cellular uptake. (a) Phasor-FLIM analysis workflow: Pixels from the lifetime image are projected onto a phasor plot to identify clusters with distinct fluorescence lifetimes. Back-mapping of these clusters onto the corresponding intensity image reveals their spatial distribution, enabling the correlation of specific lifetime signatures with the molecular state of the probe (e.g., LNP-encapsulated versus free FAM-siRNA). (b, c) Solution-based FLIM analysis of free FAM-siRNA (b) or FAM-siRNA encapsulated in Onpattro or ApoE-Onpattro (c) at 1000 nM under different pH conditions. LNP disruption was induced using 2% Triton X-100 in PBS (pH 7.4). FLIM acquisition parameters: λ_ex = 488 nm, λ_em = 507–580 nm. (d) Cell-based FLIM analysis: PC3-Luc6 cells were incubated with LNPs encapsulating FAM-siRNA (total lipid concentration: 30 μM) for 6 h before imaging. Representative images show the FAM-siRNA fluorescence intensity, corresponding lifetime maps, phasor plots, and phasor-selected intensity mappings. Distinct phasor clusters with different lifetimes were identified and mapped back onto the grayscale intensity images to visualize their spatial distribution within cells.

To establish reference lifetimes, we first measured the fluorescence lifetimes of free and LNP-encapsulated FAM-siRNA in solution (Figure b,c and Table S1). Free FAM-siRNA exhibited clear pH sensitivity, with the lifetime decreasing from 3.99 ns at pH 7.4 to 3.37 ns at pH 5.0. Both Onpattro and ApoE-Onpattro formulations showed similar pH-dependent decreases. Notably, at matched pH values, LNP-encapsulated FAM-siRNA displayed consistently shorter lifetimes than the free form (e.g., at pH 6.4, 3.87 ns for free siRNA, 3.32 ns for Onpattro, and 3.47 ns for ApoE-Onpattro). This reduction reflects aggregation-caused quenching (ACQ) and homo-FRET arising from the high local siRNA density within LNPs; upon complete LNP dissociation, the lifetime recovered to ∼4 ns (Figure c). However, in the porphyrin-LNP formulations, FAM-siRNA fluorescence was extensively quenched (96.2% compared with 62.7% in Onpattro, Figure S3) due to FRET to the porphyrin moieties, yielding insufficient photon counts for reliable fluorescence lifetime determination when the particles remained intact.

Next, we performed live-cell FLIM analysis after 6 h of LNP incubation with cells (Figure d). FAM-siRNA delivered by ApoE-porphyrin-LNPs exhibited three distinct phasor populations with lifetimes characteristic of intracellular organelles (AP_organelle, τ = 2.98 ns), the plasma membrane (AP_membrane, 0.89 ns), and the cytosol (AP_cytosol, 3.72 ns). In contrast, ApoE-Onpattro showed only two major lifetime populations associated with intracellular organelles (AO_organelle, 3.11 ns) and the plasma membrane (AO_membrane, 2.29 ns). LNPs lacking ApoE exhibited two populations corresponding to organelle localization and cellular autofluorescence. Notably, porphyrin-LNP displayed a shorter intraorganelle lifetime (P_organelle, 1.77 ns) compared to Onpattro (O_organelle, 2.80 ns).

These results reveal formulation-dependent LNP disassembly kinetics across different uptake stages. For example, by comparison of Onpattro and ApoE-Onpattro, a progressive lifetime increase was observed: AO_membrane (2.29 ns) < O_organelle (2.80 ns) < AO_organelle (3.11 ns). Importantly, because the solution studies demonstrated that acidification alone shortens the FAM-siRNA lifetime, the observed lifetime increase in more acidic organellar compartments relative to the less acidic plasma membrane can only be explained by stronger LNP dissociation that counteracts pH-driven quenching. Specifically, the AO_membrane population represents LNPs that have just engaged in receptor-mediated endocytosis and remain associated with the plasma membrane, where particles are highly concentrated and largely intact, resulting in strong ACQ and homo-FRET. Following internalization into endosomal compartments, partial LNP disassembly and siRNA release reduced quenching, leading to a moderate lifetime increase (O_organelle, 2.80 ns) despite the more acidic environment. The further increase in AO_organelle (3.11 ns) suggests enhanced particle destabilization facilitated by ApoE. A similar increasing trend was found in porphyrin-lipid containing formulations: AP_membrane (0.89 ns) < P_organelle (1.77 ns) < AP_organelle (2.98 ns), confirming that ApoE incorporation enhances LNP intracellular dissociation across different formulations. The overall shorter lifetimes of porphyrin-LNPs compared to non-porphyrin counterparts reflect FRET between FAM-siRNA and porphyrin-lipids, with FRET efficiency decreasing as particles disassemble. Importantly, the smaller difference between AO_organelle (3.11 ns) and AP_organelle (2.98 ns), compared to O_organelle (2.80 ns) and P_organelle (1.77 ns) further indicates that ApoE facilitates more complete intracellular LNP dissociation, leading to comparable siRNA dequenching in both Onpattro and porphyrin-LNP systems despite the latter’s additional FRET pathway.

Moreover, ApoE-porphyrin-LNP exhibited a distinct cytosolic population AP_cytosol (3.72 ns) with markedly reduced plasma-membrane association compared to ApoE-Onpattro, which remained largely membrane-bound. The cytosolic lifetime of 3.72 ns approaches that of free FAM-siRNA at PBS 7.4 (∼3.99 ns), indicating substantial siRNA release into the cytoplasm. This observation aligns with data presented in Figures and , suggesting stronger endosomal membrane disruption and altered ApoE-receptor interactions for ApoE-porphyrin-LNP. Finer mapping of the organelle-level lifetime distribution of ApoE-porphyrin-LNP uptake (2.14–3.30 ns, Figure S4) revealed that cells exhibiting stronger cytosolic siRNA signals also displayed longer organelle-associated lifetimes than neighboring cells with minimal cytosolic siRNA. This data may suggest that the level of endosomal disruption is related to the extent of ApoE-porphyrin-LNP intracellular disassembly. Collectively, these results demonstrate that ApoE facilitates intracellular disassembly of both Onpattro and porphyrin-LNPs, while porphyrin-lipid incorporation further amplifies endosomal membrane disruption and promotes cytosolic siRNA release.

Structural Basis Underlying Distinct Intracellular Delivery Profiles of ApoE-Onpattro and ApoE-Porphyrin-LNP

To understand the structural basis underlying the distinct intracellular delivery behaviors of ApoE-Onpattro and ApoE-porphyrin-LNP, we characterized their morphology using cryo-EM. As shown in Figure a, in the absence of ApoE, both the Onpattro and porphyrin-LNP displayed typical spherical morphologies characteristic of siRNA-loaded LNPs. Upon ApoE association, however, notable structural differences emerged. While ApoE-Onpattro still maintained spherical membrane structure, ApoE-porphyrin-LNPs exhibited markedly distorted and irregular morphologies compared to their ApoE-free counterparts. To better quantify these structural differences, 2D classification of LNPs from cryo-EM micrographs was performed using a previously developed analysis pipeline. As shown in Figure b, ApoE-porphyrin-LNP displayed a significant change in membrane morphology, characterized by faceted outlines composed of segments with locally reduced curvature, producing an overall polyhedral appearance. In contrast, ApoE-Onpattro maintained a smooth, spherical membrane with consistently positive curvature. Quantitative curvature analysis further supports these observations (Figure c–e). Specifically, the root-mean-square (RMS) curvature, which reflects both the average magnitude of membrane bending and the overall extent of curvature variation, was calculated. ApoE-porphyrin-LNP exhibited the highest RMS curvature at 55.8 μm^–1, with a broad distribution spanning both high and low curvature values. Notably, 20% of the ApoE-porphyrin-LNP surface exhibited locally concave (negative) curvature, whereas all other formulations remained fully convex (positive) (Figure d). To further assess spatial irregularity, we quantified the frequency of curvature sign switches per micrometer (Figure e). This metric captures how frequently the membrane alternates between convex and concave regions. Nearly 45% of ApoE-porphyrin-LNP membranes exhibited more than 100 convex-to-concave curvature transitions per micrometer, whereas other formulations showed exclusively convex profiles. Collectively, these results confirm that ApoE-porphyrin-LNP possesses extensive membrane domains with significantly reduced and even negative curvature distributed across the whole particle surface. Membranes with negative curvature are known to lower the energy barrier for membrane fusion events, , which may underlie the enhanced endolysosomal disruption observed in PC3-Luc6 cells treated by ApoE-porphyrin-LNPs compared to ApoE-Onpattro.

4.

ApoE-porphyrin-LNPs exhibit distinct membrane morphology without impacting particle stability. (a) Cryo-EM micrographs depicting structural changes in ApoE-coated versus noncoated Onpattro and porphyrin-LNP. Scale bar in zoomed-in view = 25 nm. (b) 2D-classification results reveal distinct surface morphology and curvature induced by ApoE coating between Onpattro and porphyrin-LNP. Scale bar = 10 nm. (c–e) Curvature analysis of Onpattro (O), ApoE-Onpattro (AO), Porphyrin-LNP (P), and ApoE-porphyrin-LNP (AP): (c) root-mean-square (RMS) curvature; (d) percentage of membrane area with positive or negative curvature; (e) curvature sign switching frequency. (f) Size and (g) zeta potential of LNPs stored at 4 °C for given days. Data are expressed as mean ± standard deviation (n = 3). (h) Pharmacokinetic profile of ApoE-LNPs. LNPs were dosed at siRNA of 0.75 mg/kg (n = 6). (i, j) Biodistribution of LNPs in PC3-Luc6 subcutaneous tumor bearing mice. LNPs were intravenously administered at a siRNA dose of 1 mg/kg. Mice were dissected at 24 h postinjection (n = 5).

Next, we evaluated whether these pronounced curvature changes compromise particle stability and cargo retention. ApoE-LNP formulations were incubated in cell culture medium containing 10% FBS at 37 °C. As shown in Figure S5, both ApoE-Onpattro and ApoE-porphyrin-LNPs remained highly stable, preserving over 90% siRNA encapsulation over 24 h, comparable to their non-ApoE counterparts. Moreover, ApoE-porphyrin-LNPs exhibited comparable colloidal stability to porphyrin-LNPs over 24 h in PBS and 50% fetal bovine serum at 37 °C, as assessed by porphyrin fluorescence quenching efficiency, a surrogate marker of LNP integrity (Figure S6). The hydrodynamic diameter, polydispersity index (PDI), and zeta potential of all the LNP formulations were monitored over 7 days of storage at 4 °C and remained largely unchanged (Figure f,g). On day 7, the average particle size was ∼75 nm with PDI < 0.2 across all formulations. The average ζ-potentials also showed no significant change between D7 and D0, measuring +5.6 mV for both Onpattro and porphyrin-LNPs, −5.7 mV for ApoE-Onpattro, and −10.3 mV for ApoE-porphyrin-LNPs. This overall shift toward a more negative surface potential in the ApoE-LNPs is consistent with the net negative charge of ApoE3 at neutral pH (measured ζ = −2.4 mV) and its isoelectric point of 5.55.

We next investigated whether these ApoE-induced alterations in membrane properties affect the in vivo behavior of LNPs (Figure h–j). Both ApoE-porphyrin-LNP and ApoE-Onpattro exhibited rapid blood clearance (t _1/2 = 1.4 and 9.1 min, respectively), consistent with previous findings that ApoE adsorption promotes accelerated hepatic uptake. Interestingly, despite similar initial clearance kinetics, ApoE-porphyrin-LNP showed a longer terminal elimination half-life (2.91 h vs 1.28 h), suggesting altered LDLR recycling dynamics that may reduce repeated ApoE–LDLR-mediated clearance. In PC3-Luc6 tumor-bearing mice, both ApoE-LNP formulations showed biodistribution patterns similar to their non-ApoE counterparts, characterized by predominant accumulation in the liver and spleen and minimal distribution to other tissues (<2%, ID%/g). ApoE-porphyrin-LNP and ApoE-Onpattro showed nearly identical profiles, except for the latter’s higher spleen uptake (44.2 vs 15.7, ID%/g).

Overall, these results indicated that despite its distinct faceted membrane morphology, ApoE-porphyrin-LNP maintains good physicochemical stability in vitro and in vivo, exhibiting biodistribution patterns comparable to those of ApoE-Onpattro.

Conformational Change of ApoE and Impact on Receptor-Mediated Cell Uptake

When ApoE binds to spherical lipid membranes, its C-terminal domain acts as the primary lipid anchor, associating with the membrane via shallow interfacial insertion. The amphipathic α-helices within this region lie parallel to the membrane plane, an orientation that minimizes membrane deformation and favors stable binding on positively curved surfaces. , This structural model aligns with our cryo-EM observations of ApoE-Onpattro, which maintains a uniformly spherical morphology upon ApoE binding. In contrast, porphyrin-LNPs, which differ from Onpattro only in the substitution of 10% DSPC with porphyrin-lipid, exhibit dramatic morphological remodeling upon ApoE incorporation. ApoE binding induces membrane faceting and irregularity, suggesting that ApoE may adopt a distinct conformation in the context of porphyrin-lipid-enriched surfaces, which can affect the LNP’s surface properties.

For example, ζ-potential measurements (Figure g) showed that ApoE-porphyrin-LNPs (−10.3 mV) were more negatively correlated than ApoE-Onpattro (−5.7 mV), whereas their non-ApoE counterparts exhibited nearly identical surface potentials. The greater negative ζ-potential of ApoE-porphyrin-LNPs implies increased exposure of acidic residues and reduced presentation of cationic domains of ApoE. Importantly, the receptor-binding domain (amino acid 136–150) of ApoE is enriched in Lys and Arg residues that form a cationic patch typically solvent-exposed for LDLR interaction. This region may therefore be partially shielded in ApoE-porphyrin-LNP while remaining exposed in ApoE-Onpattro. To further evaluate the surface properties of ApoE-LNP, we employed Quartz Crystal Microbalance with Dissipation Monitoring (QCM-D) using a negatively charged supported lipid bilayer (DOPC/DOPS = 90:10 mol %). , In QCM-D, decreases in resonance frequency (Δf) correspond to mass adsorption, while increases in dissipation (D) reflect the viscoelasticity of the adsorbed layer. As shown in Figure a, Onpattro alone caused negligible changes in Δf ₇ or D ₇, whereas ApoE-Onpattro induced a pronounced, progressive decrease in Δf ₇ to approximately −25 Hz accompanied by a corresponding strong increase in D ₇ that persisted after rinsing. The result indicates strong adsorption of ApoE-Onpattro onto the negatively charged bilayer, likely mediated by electrostatic interactions involving the exposed cationic residues of the ApoE receptor-binding domain. In contrast, ApoE-porphyrin-LNPs exhibited markedly weaker adsorption (Δf ₇ ≈ −10 Hz), aligning with their more negative ζ-potential and supporting an altered ApoE orientation that reduces exposure of positively charged LDLR-binding residues. To account for residual unpurified protein, free ApoE was tested as a benchmark control (Δf ₇ ≈ −18 Hz, Figure S7). Notably, the response of ApoE-Onpattro exceeded this baseline, while the signal from ApoE-porphyrin-LNP fell below it. This divergence indicates that the observed QCM-D responses predominantly reflect the behavior of the ApoE-LNP complexes rather than the free ApoE in solution. Specifically, the reduced signal associated with the porphyrin-LNP formulation suggests that ApoE is effectively sequestered by the nanoparticles into an altered configuration that suppresses its direct interaction with model membranes, thereby diminishing the level of adsorption.

5.

Structural change of ApoE-porphyrin-LNP changes its interaction with LDLR. (a) QCM-D analysis of LNP interaction with DOPC/DOPS (90/10 mol %) supported planar bilayers. (b, c) Effects of LDLR knockdown by siLDLR (b) and PCSK9 inhibition (c) on the intracellular distribution of FAM-labeled siRNA delivered by ApoE-Onpattro and ApoE-porphyrin-LNPs. (d, e) Quantification of LDLR expression at the protein (d) and mRNA (e) levels across multiple cell lines, highlighting differential LDLR abundance that may influence ApoE-LNP performance. (f) Immunofluorescence staining of LDLR (yellow) in PC3-Luc6 after cells were treated with LNPs for 6 h and then fixed immediately or after 18 h.

With altered surface properties and potential changes in ApoE receptor-binding domain presentation, ApoE-porphyrin-LNP may engage the LDLR differently from ApoE-Onpattro. To assess this, we examined whether LDLR remained involved in the cellular uptake of the LNPs by performing LDLR knockdown via siRNA or functional inhibition by PCSK9. As shown in Figure b, PC3-Luc6 cells pretreated with siLDLR exhibited reduced intracellular punctate FAM-siRNA signals and increased membrane-associated signals upon ApoE-Onpattro treatment, compared to siCtrl or PBS, confirming LDLR involvement in typical ApoE-mediated endocytosis. In contrast, siLDLR pretreatment in the case of ApoE-porphyrin-LNP markedly reduced cytosolic FAM-siRNA fluorescence, suggesting inhibited endosomal disruption. This indicates that LDLR remains engaged with ApoE in the porphyrin-LNP context and contributes to endosomal membrane disruption during ApoE-porphyrin-LNP uptake. Consistent results were observed with PCSK9 inhibition, which modestly decreased punctate signals for ApoE-Onpattro but significantly weakened cytosolic siRNA release for ApoE-porphyrin-LNP (Figure c), further supporting LDLR involvement in ApoE-porphyrin-mediated endosomal disruption. These findings also help explain the differential effects observed across cell types. In earlier live-cell imaging (Figure b), ApoE-porphyrin-LNP induced the strongest endosomal disruption and cytosolic siRNA release in PC3-Luc6 cells, with a weaker effect in Hep3B and HDFn cells. To investigate this discrepancy, we quantified LDLR expression levels across the three cell lines (Figure d,e), which revealed that PC3-Luc6 cells express >6-fold higher LDLR protein and 2-fold higher LDLR mRNA compared to Hep3B and HDFn. The elevated LDLR expression likely accounts for the pronounced endosomal disruption in PC3-Luc6 cells following ApoE-porphyrin-LNP treatment.

The LDLR knockdown results indicate that ApoE, although likely adopting an altered conformation on porphyrin-LNPs, retains its ability to interact with LDLR. This observation, however, presents an apparent paradox, as ApoE-porphyrin-LNPs did not exhibit enhanced cellular uptake compared with porphyrin-LNPs alone (Figure e). A possible explanation is that while ApoE retains its ability to bind LDLR, its altered orientation on the porphyrin-LNP surface may lead to atypical receptor interactions that perturb normal LDLR recycling dynamics, thereby limiting sustained LNP uptake. To test this hypothesis, we performed immunofluorescence staining to examine the intracellular distribution of LDLR in cells after ApoE-LNP treatment (Figure f). PC3-Luc6 cells were incubated with LNPs for 6 h and fixed either immediately or following an additional 18-h chase. At the immediate post-treatment time point, intracellular LDLR levels were comparable between cells treated with ApoE-Onpattro and ApoE-porphyrin-LNP, with both groups showing slightly elevated intracellular LDLR levels relative to untreated controls, consistent with LDLR internalization during LNP uptake. In contrast, after 18 h post-treatment, cells treated with ApoE-porphyrin-LNP exhibited markedly greater intracellular LDLR accumulation compared to ApoE-Onpattro-treated cells. This observation indicates that LDLR becomes retained within intracellular compartments following ApoE-porphyrin-LNP internalization. Such impaired LDLR recycling may account for the lack of ApoE-mediated enhancement in the overall intracellular uptake of porphyrin-LNPs.

In summary, the distinct membrane morphology and surface properties of ApoE-Onpattro and ApoE-porphyrin-LNP suggest that ApoE can have altered conformation depending on the LNP surface lipid environment, which in turn modulates its interaction with LDLR. In porphyrin-LNP, this altered ApoE conformation impairs efficient LDLR recycling. Consequently, the resulting entrapment of unrecycled LDLR, ApoE, and other LNP components within endosomal organelles collectively promotes endosomal membrane disruption.

Impact of Main Components on ApoE-Porphyrin-LNP Induced Endosomal Disruption

Our findings highlight several key observations. First, the surface lipid type largely affects LNP membrane morphology upon ApoE binding, suggesting that ApoE adopts distinct conformations depending on the lipid environment. Second, ApoE-porphyrin-LNP demonstrates two unique biological outcomes, distinct from ApoE-Onpattro: (1) altered receptor-mediated endocytosis with minimal enhancement of intracellular uptake and (2) notably enhanced endosomal membrane disruption, leading to cytosolic siRNA release and subsequent cell death. Given that LNP endosomal escape can be mediated by specific components, particularly ionizable lipids, we next dissected the contributions of individual components toward the enhanced endosomal disruption observed in ApoE-porphyrin-LNP treatment.

First, the role of ApoE was evaluated by varying the lipid-to-ApoE ratio from 50:1 (high ApoE) to 1000:1 (low ApoE). Endosomal membrane disruption was assessed by monitoring the secondary outcomes, including cytosolic siRNA release and a cell viability decrease. As shown in Figure a, across all the tested lipid-to-ApoE ratios, ApoE-Onpattro consistently demonstrated classical receptor-mediated endocytosis, characterized by pronounced membrane-bound and punctate intracellular FAM-siRNA signals. Conversely, ApoE-porphyrin-LNP exhibited negligible membrane binding and prominent diffuse cytosolic siRNA fluorescence. Cell viability assays further supported these observations (Figure b): ApoE-Onpattro induced minimal cytotoxicity, whereas ApoE-porphyrin-LNP consistently led to >50% cell death across all tested ratios. Importantly, empty porphyrin-LNP lacking siRNA exhibited comparable cytotoxicity (Figure c), indicating that siRNA itself was not a critical factor driving endosomal disruption. Notably, both imaging and viability data revealed that even at the lowest ApoE concentration (1000:1 lipid-to-ApoE), ApoE was sufficient to induce endosomal disruption in porphyrin-LNP. Increasing the ApoE amount did not further enhance membrane disruption. To directly track ApoE dynamics during intracellular uptake, ApoE was labeled with Cy3. Imaging results (Figure d) demonstrated the strong intracellular uptake and colocalization of ApoE-Cy3 with siRNA puncta in ApoE-Onpattro. In contrast, ApoE-porphyrin-LNP exhibited significantly reduced intracellular ApoE signal. Together, these findings suggest that ApoE-porphyrin-LNP achieves limited uptake enhancement through receptor-mediated endocytosis compared with ApoE-Onpattro, and its pronounced endosomal disruption is driven primarily by ApoE conformational changes rather than ApoE abundance within endocytic organelles.

6.

Contribution of ApoE and lipid compositions to endosomal disruption. (a) FAM-siRNA intracellular distribution at varying ApoE ratios after 6 h intracellular uptake of different ApoE-LNP formulations. (b, c) Viability of cells treated by ApoE-LNPs with (b) and without (c) siRNA loading. (d) Intracellular distribution of cy3-labeled ApoE. (e, f) Impact of porphyrin-lipid/DSPC (e) and porphyrin-lipid/MC3-lipid (f) molar ratios in ApoE-porphyrin-LNP on the FAM-siRNA cellular delivery pattern.

Next, we evaluated the roles of two key lipid components: the porphyrin-lipid (surface lipid) and the MC3-lipid (ionizable lipid). Increasing the porphyrin-lipid ratio from 0% to 20% induced a clear transition from typical receptor-mediated endocytosis to reduced membrane binding with increased cytosolic siRNA distribution (Figure e), underscoring the critical role of porphyrin-lipid in facilitating endosomal disruption, likely through ApoE conformational changes. To assess the contribution of MC3-lipid, its fraction was progressively reduced from 50% to 10% by substitution with porphyrin-lipid (Figure f). Cytosolic siRNA distribution remained at MC3-lipid down to 25% but was lost at 10% at 6 h postuptake. At 24 h post-treatment (Figure S8), ApoE-porphyrin-LNP containing 40% MC3-lipid and 20% porphyrin-lipid showed the highest cytotoxicity, as evidenced by extensive cellular disintegration and debris under bright-field view. This increase in cytotoxicity observed when the porphyrin-lipid content rose from 10% to 20% (with a corresponding reduction in MC3 from 50% to 40%) highlights the critical contribution of porphyrin-lipid to endosomal disruption. However, further increasing porphyrin-lipid beyond this level while depleting MC3-lipid reduced cytotoxicity, suggesting that MC3-lipid remains essential for effective endosomal membrane disruption.

Collectively, these results demonstrate that ApoE-porphyrin-LNP-induced endosomal disruption is driven by the interplay of multiple factors including intracellular LDLR, ApoE, porphyrin-lipid, and MC3-lipid. While MC3-lipid contributes to endosomal escape, it is insufficient on its own to induce a strong membrane disruption. Instead, the surface lipid is the dominant determinant, since it governs ApoE conformation and its interaction with LDLR. Importantly, the siRNA cargo itself does not influence endosomal disruption, underscoring that the observed endosomal disruption primarily results from the combined effects of lipid composition and ApoE conformation changes rather than the RNA cargo.

Mechanism of Cell Death Induced by ApoE-Porphyrin-LNP

Next, we sought to elucidate the mechanism underlying ApoE-porphyrin-LNP-induced cell death. Using Annexin V and propidium iodide (PI) staining, we evaluated cell apoptotic profiles by flow cytometry analysis. (Figure a) ApoE-porphyrin-LNP treatment led to a 3.6-fold increase in Annexin V and PI double-positive cells (28.5%) compared to ApoE-Onpattro (6.2%), indicative of late-stage apoptosis or necrosis accompanied by increased cell membrane permeability. Confocal microscopy imaging further supported these findings (Figure b). Importantly, ApoE-Onpattro, Onpattro, and porphyrin-LNP alone did not significantly alter apoptotic profiles, clearly suggesting that neither ApoE nor porphyrin-lipid alone is sufficient to trigger notable cell death. Instead, the combination of ApoE and porphyrin-lipid is required to induce apoptosis.

7.

Mechanistic insights into ApoE-porphyrin-LNP-induced cell death. (a, b) Annexin V/PI staining of PC3-Luc6 cells following LNP treatment, analyzed by flow cytometry (a) and confocal microscopy (b). (c) Caspase-3/7 activation visualized by confocal microscopy post-LNP treatment. (d) Quantification of extracellular ATP release in PC3 cells. O, Onpattro; P, porphyrin-LNP; AO, ApoE-Onpattro; AP, ApoE-porphyrin-LNP. Data are expressed as mean ± standard deviation (n = 3). (e) Quantification of HMGB1 release in PC3-Luc6 cells.

To further confirm the apoptotic mechanism, we assessed activation of Caspase-3/7, key mediators of apoptosis. , ApoE-porphyrin-LNP significantly increased Caspase-3/7 activity compared to the controls (Figure c), strongly supporting apoptosis as the primary pathway. Recent study has suggested Caspase-3 activation as an upstream event in Gasdermin E (GSDME)-mediated pyroptosis. Given the pronounced cell swelling and membrane blebbing observed (Figure b,c), which are morphological characteristics commonly associated with pyroptosis, , we next investigated whether pyroptosis also contributed to cell death. We examined extracellular release of ATP and HMGB1, known markers associated with inflammatory cell death such as immunogenic cell death (ICD) and pyroptosis. ApoE-porphyrin-LNP induced ∼2.5-fold higher ATP release compared to ApoE-Onpattro and ∼5-fold higher ATP release compared to non-ApoE LNPs (Figure d). However, HMGB1 release was not significantly elevated relative to ApoE-Onpattro, and the increase compared with non-ApoE LNPs was modest (less than 2-fold, Figure e). To directly evaluate pyroptosis involvement, we analyzed GSDME cleavage via immunoblotting. Neither GSDME cleavage nor changes in full-length GSDME protein expression were detected after treatment with ApoE-porphyrin-LNP or other control LNP formulations (Figure S9), excluding pyroptosis as the dominant mechanism. Taken together, these results demonstrate that ApoE-porphyrin-LNP induces cell death predominantly through apoptosis, driven by endosomal membrane disruption. The modest increases in extracellular ATP and HMGB1 suggest a minor contribution from inflammatory mechanisms, although these appear secondary to apoptosis.

Discussion

A central insight from this study is the profound influence of lipid composition on the ApoE structure and function within LNPs. This sensitivity arises from ApoE’s modular domain architecture and conformational plasticity. ApoE3 comprises two major structural domains: a C-terminal lipid-binding domain and an N-terminal receptor-binding domain. In its lipid-free state, the receptor-binding domain is sterically masked by the lipid-binding domain, making ApoE3 functionally inactive in receptor interaction. Lipidation, such as association with LDL particles, initiates conformational rearrangements that expose the receptor-binding domain, enabling interaction with LDL receptor family members. Notably, this conformational mechanism exhibits exquisite sensitivity: even single-residue substitutions, as seen in ApoE2 or ApoE4, , can dramatically alter LDLR binding through changes in domain–domain interactions and structural flexibility.

Our findings extend this model to synthetic lipid nanoparticles, revealing that surface lipid composition, exemplified by porphyrin-lipid incorporation, actively reconfigures ApoE conformation on the LNP surface. Cryo-EM imaging revealed pronounced curvature changes in the ApoE-porphyrin-LNP membrane, consistent with structural remodeling of ApoE’s lipid-binding domain. Given that receptor engagement depends on domain exposure, these lipid-induced structural changes likely modify ApoE’s docking orientation on LNPs and, subsequently, its receptor-binding interface presentation. This mechanistic insight explains the attenuated ApoE-mediated uptake of porphyrin-LNPs and aligns with previous reports demonstrating that lipid composition, including both surface and ionizable lipids, substantially impacts ApoE receptor binding in vitro and in vivo. −

Moreover, we uncovered a previously unrecognized intracellular role of ApoE3 in promoting endosomal membrane disruption when it was presented in a porphyrin-lipid context. This disruption promotes cytosolic RNA release and cell apoptosis, a property not commonly associated with ApoE3 in physiological settings. Interestingly, this phenomenon mirrors ApoE4-associated pathology in Alzheimer’s disease, where ApoE4 has been implicated in defective receptor recycling, − endolysosomal leakage, , and neuron apoptosis. , In particular, ApoE4 conformational changes are thought to underlie its retention in endosomes and its role in lysosomal instability, , which resonates with our observations that ApoE3-porphyrin-LNPs impair LDLR recycling and promote endosomal rupture. These parallels suggest that whether in engineered nanoparticles or endogenous disease states, lipid-induced ApoE conformational changes may converge on shared mechanistic pathways. We acknowledge that the ApoE-porphyrin-LNP formulation characterized in this work is not suitable for standard gene therapy applications due to the potent endosomal-disruption phenotype that leads to rapid cell death. However, from an application perspective, endosomal-disruption-mediated cytotoxicity can be intentionally exploited as a therapeutic mechanism (for example, to induce cancer cell death) independent of the RNA cargo. Conversely, this phenomenon highlights a design consideration in LNP engineering, since specific surface chemistries and their interactions with ApoE may inadvertently trigger cytotoxic effect. Importantly, the phenotype is tunable (e.g., by adjusting the porphyrin-lipid mol %), enabling the decoupling of delivery efficiency from cytotoxicity when desired.

Collectively, our study demonstrates lipid composition as a structural switch that governs the ApoE conformation, receptor interaction, and intracellular fate. These findings have dual implications. In LNP-based RNA therapeutics, they offer a novel strategy to modulate ApoE-mediated cellular trafficking by tuning lipid components, allowing LNPs to harness or bypass ApoE pathways depending on the therapeutic goals. Simultaneously, they also highlight the potential of synthetic lipid systems as tools to model, modulate, or even reverse pathological ApoE conformations, providing a new perspective for neurodegeneration research.

Conclusion

Our work establishes lipid composition as a critical determinant of ApoE-LNP structure and function. By substitution of porphyrin-lipid for DSPC, ApoE-LNPs acquired multifaceted, irregular membrane morphologies that fundamentally reshaped their cellular interactions: altering receptor-mediated endocytosis while enhancing endosomal disruption. These structural changes also promoted cytosolic RNA delivery and triggered downstream cellular apoptosis. Together, our findings provide mechanistic insights into LNP–protein interactions and their structure–function relationships. More broadly, this work underscores lipid membrane engineering as a promising strategy to modulate ApoE functionality in both therapeutic applications and pathologies associated with ApoE dysfunction.

Materials and Methods

Materials

1,2-Distearoyl-sn-glycero-3-phosphocholine (DSPC), cholesterol, 1,2-dimyristoyl-rac-glycero-3-methoxy­(poly­(ethylene glycol))-2000 (DMG-PEG₂₀₀₀), and diethylene-triamine-penta-acetic acid-bis­(stearylamide) (gadolinium salt) (DTPA-BSA­(Gd)) were purchased from Avanti Polar Lipids (Alabaster, AL, USA). DLin-MC3-DMA was purchased from Nanosoft Polymers (NC, USA). Porphyrin-lipid was synthesized by the previously reported methods. Recombinant human ApoE3 was purchased from PeproTech (NJ, USA). Human PCSK9 (D374Y) was purchased from Acrobiosystem (DE, USA).

Small Interfering RNA

All siRNAs were custom synthesized and purchased from Horizon Discovery (USA). siLuc: sense, 5′-GAU UAU GUC CGG UUA UGU AdTsdT-3′; antisense, 5′-UAC AUA ACC GGA CAU AAU CdTsdT-3′. For dye-labeled siRNA (FAM, Cy3, and Alexa-647 conjugated), the fluorophore was attached to the 5′ end of the sense strand.

Synthesis of ApoE-LNPs

Base LNPs were synthesized by using a microfluidic rapid mixing method as previously described. Lipids were dissolved in ethanol at a molar ratio of 50:10:38.5:1.5 (DLin-MC3-DMA/DSPC/Cholesterol/DMG-PEG2000) for the Onpattro formulation, replacing 10% DSPC with porphyrin-lipid for porphyrin-LNP. siRNA was prepared in 25 mM sodium acetate buffer (pH 4.0). The aqueous siRNA solution and lipid–ethanol solution were rapidly mixed using herringbone microfluidic chips (Microfluidic ChipShop, Germany) at a volumetric flow rate ratio of 3:1 (aqueous/ethanol) and a total flow rate of 10 mL/min. The resultant LNP suspension was dialyzed overnight against PBS (pH 7.4), concentrated using Amicon centrifugal filters (Sigma-Aldrich), and sterilized using 0.22 μm filters. For ApoE-LNPs, LNPs were further incubated overnight with recombinant human ApoE at specified total lipid-to-protein ratios at 4 °C before use.

In Vitro Characterization of LNPs

The hydrodynamic size, polydispersity, and zeta-potential of LNPs was characterized with a Zetasizer Nano ZS (Malvern Instruments). siRNA concentration and encapsulation efficiency were measured by Quant-it RiboGreen RNA Assay based on the manufacturer’s protocol (Thermofisher). The concentration of each lipid component in the LNP after synthesis was quantified using an ACQUITY UPLC H-Class system (Waters, Canada) with a BEH C18 column (1.7 μm, 2.1 mm × 50 mm) as reported previously.

Cell Culture

PC3-Luc6 (prostate cancer), Hep3B (liver cancer), and HDFn (normal fibroblast) cells were maintained at 37 °C with 5% CO₂ in their respective recommended media supplemented with 10% fetal bovine serum (FBS). PC3-Luc6 and Hep3B cells were cultured in Eagle’s Minimum Essential Medium (EMEM), while HDFn cells were cultured in Dulbecco’s Modified Eagle Medium (DMEM).

LNP Cellular Uptake Imaging

For cellular uptake imaging, cells were seeded into eight-well coverglass-bottom chambers (Nunc LabTek, Sigma-Aldrich) at a density of 2 × 10⁴ cells per well and incubated for 48 h. Subsequently, LNPs loaded with FAM-siRNA were added at a concentration of 30 μM (based on total lipid) and incubated for 6 h. Cells were washed twice with culture medium prior to imaging. Fluorescence images were captured using a Leica TCS SP8 confocal microscope or Leica STELLARIS confocal microscope equipped with a 63× oil immersion objective. Cells were excited at 488 nm, laser power 30%. Two HyD detectors were used to collect signal for FAM-siRNA (Em: 507–580 nm) or porphyrin-lipid (Em: 670–765 nm).

LNP Intracellular Time-Lapse Imaging

PC3-Luc6 cells were seeded into eight-well coverglass-bottom chambers at a density of 2 × 10⁴ cells per well and cultured for 48 h. Cells were treated with ApoE-coated LNPs loaded with Cy3-siRNA at a concentration of 30 μM (based on total lipid) for 6 h, washed twice, and then continuously monitored over 24 h using a Leica TCS SP8 confocal microscope (Ex: 561 nm, 1% power; Em: 565–654 nm). Images were captured every 10 min to observe real-time LNP intracellular trafficking, endosomal escape, and cell morphological changes.

Fluorescence Lifetime Imaging Microscopy (FLIM)

PC3-Luc6 cells were treated with FAM-siRNA-loaded LNPs as described above. Fluorescence lifetime imaging was performed on a Leica TCS SP8 confocal microscope equipped with a FALCON FLIM module. FAM was excited using a 40 MHz pulsed white-light laser tuned to 488 nm, and emission was collected between 507 and 580 nm with a HyD detector. Photon arrival times were recorded by using time-correlated single-photon counting (TCSPC). For each image, pixelwise fluorescence decays were transformed into the frequency domain and represented on a phasor plot. Phasor analysis and extraction of the average fluorescence lifetime were carried out using LAS X FLIM software (Leica Microsystems, Germany). A wavelet filter at a threshold of 50 was applied to all of the phasor plots to remove background noise. For solution FLIM analysis, 30 μL of solution was applied to a slide and covered with a coverslip. Imaging was performed using a Leica TCS SP8 confocal microscope equipped with a FALCON FLIM module using consistent settings. The entire image was fitted to a three-component exponential decay model, selected based on minimal reduced chi-square (χ²) values. The intensity-weighted average lifetime is reported as the mean lifetime.

Quantitative LNP Cellular Uptake Assay

PC3-Luc6 cells were seeded at 2.5 × 10⁴ cells per well in 24-well plates and cultured for 48 h. Cells were incubated with LNPs at concentrations of 2 μM (based on porphyrin) or equivalent to 12 μM (total lipid) for indicated times. LNPs were all formulated with 2% DTPA-BSA­(Gd) lipid. After incubation, cells were trypsinized, counted using a DeNovix cell counter, collected in capped Eppendorf tubes, and dried using SpeedVac. For ICP-MS measurement, all samples were digested in a 3:1 mixture of 70% (w/v) nitric acid and concentrated 37% (w/v) hydrochloric acid for 24 h. Samples were diluted with ultrapure water and analyzed by ICP-MS (NexION 350×, PerkinElmer) for Gd³⁺ quantification. Gd³⁺ levels were normalized against cell counts based on a calibration curve from 1 × 10^–6 to 1 × 10^–2 mg·L^–1. For fluorescence-based measurements, cells were lysed with DMSO to disrupt LNPs, and the fluorescence of Alexa647-siRNA was measured using a CLARIOstar plate reader (Ex/Em: 650 nm/665–690 nm).

Cryo-EM Sample Preparation and Image Acquisition

Cryo-EM grids were prepared by applying 4 μL of LNP samples (∼5–15 mg/mL lipid) onto glow-discharged holey gold grids, vitrified in liquid ethane using a ThermoFisher Mark IV Vitrobot. Blotting conditions of 2 s of blotting, +3 blot force, 5 °C ambient temperature, and 90% relative humidity were used. Cryogrids were loaded on a Glacios 200 kV microscope equipped with a Falcon 3EC camera (ThermoFisher). Data sets were collected semiautomatically with EPU (ThermoFisher) at a nominal magnification of 92,000× (1.566 Å per pixel) at a target defocus range of 1–3 μm. Image processing was performed in cryoSPARC (Structura Biotechnology). 2D classification was performed as reported previously.

Curvature Analysis

After obtaining the 2d-classification results of each formulation, images of the individual classes were analyzed using the Kappa curvature analysis plugin in Fiji to extract the point curvature κ distribution along the selected membrane contour. The resulting data were subsequently processed to calculate the root-mean-square (RMS) curvature, the fraction of the membrane exhibiting positive or negative curvature, and curvature sign-switching frequency according to the following definitions:

The RMS curvature was calculated as the length-weighted root-mean-square of curvature along the membrane:

$$\mathrm{RMS}(\kappa )=\sqrt{\frac{1}{L}{\int }_0^L\kappa {(s)}^2\mathrm{d}s}\approx \sqrt{\frac{\sum _{i=1}^N{{\kappa }_i}^2\Delta s_i}{\sum _{i=1}^N\Delta s_i}}$$

where κ­(s) is the point curvature at arc-length position s, κ _i is the curvature at the ith point, Δs _i is the arc-length increment between adjacent points, and L = ∑ _i Δs _i is the total membrane length.

The fraction of the membrane length exhibiting positive or negative curvature was calculated as

$$\%\kappa >0=100\times \frac{\sum _{i:{\kappa }_i>0}\Delta s_i}{\sum _{i=1}^N\Delta s_i},\%\kappa <0=100\times \frac{\sum _{i:{\kappa }_i<0}\Delta s_i}{\sum _{i=1}^N\Delta s_i}$$

The frequency of curvature sign changes along the membrane was determined by counting the number of sign switches in the curvature profile and normalizing the parameters by membrane length:

$$N_{\mathrm{switch}}=\sum _{i=1}^{N-1}\mathbf{1}[\mathrm{sgn}({\kappa }_{i+1})\ne \mathrm{sgn}({\kappa }_i)],f_{\pm }=\frac{N_{\mathrm{switch}}}{L}$$

Here, N _switch is the number of curvature sign changes, 1[·] is the indicator function, and f _± is the sign-switching frequency expressed as switches per μm of membrane length.

Pharmacokinetic Study

Healthy athymic male nude mice (Envigo, USA; 8 weeks, 20–25 g) were intravenously injected with ApoE-LNPs containing 2 mol % DTPA-BSA­(Gd) lipid at an siRNA dose of 0.75 mg/kg. Blood samples were collected from the saphenous vein at 2 min and 0.25, 0.5, 1, 2, 4, 8, 24, and 48 h (n = 6). Plasma was separated by centrifugation (10 min at 10000 rpm), and the collected supernatant plasma was dried using a speed vacuum. The samples were further analyzed by ICP-MS to quantify Gd³⁺ content following the same procedure described in the “Quantitative LNP Cellular Uptake Assay” section.

Biodistribution Study

Athymic male nude mice under general anesthesia were inoculated with 3 × 10⁶ PC3-Luc6 cells in 100 μL of PBS into the right flank. Once tumors reached 100 mm³, mice were intravenously injected with LNPs containing 2 mol % DTPA-BSA­(Gd) lipid at siRNA dose of 1 mg/kg. At 24 h postinjection, mice were sacrificed and tissues of interest (liver, spleen, lung, kidney, heart, tumor, intestine, skin, muscle) were excised, weighed, and dried using a speed vacuum. The samples were further analyzed by ICP-MS to quantify Gd³⁺ content following the same procedure described in the “Quantitative LNP Cellular Uptake Assay” section.

Quartz Crystal with Dissipation Monitoring (QCM-D)

The adsorption kinetics of LNPs on lipid bilayers were studied by using a QCM-D E4 instrument from Q-Sense (Gothenburg, Sweden). The silicon dioxide (SiO₂) coated quartz crystals (f ₀ = 5 MHz) were supplied by Q-Sense (Gothenburg, Sweden). Prior to their use, the crystals were soaked in a 10 mM sodium dodecyl sulfate (SDS) solution for 30 min. Then, they were rinsed and sonicated for 5 min with ultrapure water. Next, the crystals were dried under a nitrogen stream and treated in a UV–ozone chamber for 20 min. For the formation of a supported planar bilayer, DOPC/DOPS (90:10 mol %) vesicles were prepared in HEPES buffer (10 mM, NaCl 150 mM, pH 7.4) using thin-film hydration followed by extrusion 15 times through a 200 nm pore-sized polycarbonate membrane and then 15 times through a membrane with 100 nm pores. The DOPC/DOPS vesicle suspensions were then diluted to 0.5 mM in HEPES buffer (HEPES 10 mM, NaCl 150 mM, CaCl₂ 2 mM, pH 7.4) and flowed onto the SiO₂-coated quartz crystals for 7 min until the change in frequency (Δf) and dissipation (D) signals were stable. Then, the lipid bilayer was rinsed in HEPES buffer for 10 min. Finally, the suspensions of LNPs were diluted in HEPES buffer to 0.5 mM in total lipids prior to their injection and monitored for 1 h. The peristaltic pump flow rate was set to 100 μL/min, and the temperature was stabilized at 37 ± 0.1 °C.

LDLR Knockdown and Inhibition Assays

For the knockdown experiment, cells were transfected with siRNA against LDLR or control siRNA using Lipofectamine RNAiMax (50 nM, 48 h). Afterward, the medium was replaced, and cells were cultured another 24 h. Then, the cells were treated with LNPs for 6 h before imaging. For the inhibition experiment, cells were treated with PCSK9 (D374Y) at 10 μg/mL for 48 h. Then, cells were treated with LNPs for 6 h before imaging.

Immunofluorescence Staining

To examine the intracellular distribution of LDLR, PC3-Luc6 cells cultured on a coverglass were treated with ApoE-LNPs, fixed with 4% paraformaldehyde, and permeabilized using 0.05% Triton X-100. Cells were then blocked with 5% BSA in PBS for 1 h at room temperature, followed by overnight incubation at 4 °C with a mouse anti-LDLR antibody (Invitrogen, 1B10H10) diluted in 1% BSA (1:200). After washing, cells were incubated with Alexa Fluor 555-conjugated donkey anti-mouse IgG (1:1000) for 45 min at room temperature. Following thorough washing, samples were mounted with ProLong Glass Antifade Mountant (Thermo Fisher), cured, and imaged using a confocal microscope.

Western Blot Analysis

The expression of LDLR and cleavage of GSDME were detected by Western blot analysis. Cells with different treatments were washed twice with PBS and lysed by lysis buffer (Cell signaling, no. 9803) supplied with cocktail proteinase inhibitor (Sigma). The whole cell lysate was centrifuged at 12,500g for 20 min. The supernatant was transferred to a new tube to measure the protein concentration using the bicinchoninic acid method. To perform immunoblotting, the cell lysates (10 μg of protein per lane) were loaded on a 10% sodium dodecyl sulfate (SDS)-polyacrylamide gel at 120 V and transferred to poly­(vinyl difluoride) membranes using standard methods. Blots were then blocked with 5% skim milk in Tris-buffered saline plus 0.1% Tween-20 and probed with mouse anti-human LDLR (Thermofisher, #MA5-38556) at 1:1000 dilution. Afterward the blot was incubated with HRP-conjugated secondary antibody (1:2000) and visualized using the ChemiDoc imaging system (BioRad).

Reverse Transcription Quantitative Real-Time PCR

Total RNA of cells after treatments was extracted using TRIzol reagent (Invitrogen Life Technologies, Inc.) and an RNeasy Mini kit (Qiagen). cDNA was next synthesized using a High-Capacity cDNA Reverse Transcription Kit (ThermoFisher). For cDNA amplification, PowerUp SYBR Green Master Mix (ThermoFisher) and the following primers were used: GAPDH (forward: 5′-CAT GAG AAG TAT GAC AAC AGC CT-3′; reverse: 5′-AGT CCT TCC ACG ATA CCA AAG T-3′); LDLR (forward: 5′-CAG CTA CCC CTC GAG ACA GA-3′; reverse: 5′-CAC TGT CCG AAG CCT GTT CT-3′). The comparative cycle threshold (Ct) method was used to calculate the relative abundance of luciferase mRNAs (GAPDH mRNA as internal control). Each assay was conducted in triplicate.

Cell Viability and Cell Death Mechanisms

PC3-Luc6 cells were seeded at a density of 4,000 cells per well in 96-well plates and incubated for 48 h before LNP treatment. Following a 6 h incubation with LNPs, cells were washed, replenished with fresh culture medium, and incubated for an additional 18 h. Cell viability was assessed using the AlamarBlue assay by incubating cells for 2 h in medium containing 0.05 mg/mL AlamarBlue reagent (Invitrogen), followed by fluorescence measurement (excitation/emission: 540/590 nm) using a CLARIOstar microplate reader. To investigate cell death mechanisms post-LNP treatment, Annexin V Alexa Fluor 488 and Propidium Iodide (PI) staining (Invitrogen, no. V13245) was performed, and cells were analyzed via confocal microscopy and flow cytometry. Additionally, caspase-3/7 activation was assessed using CellEvent Caspase-3/7 Detection Reagents (Invitrogen); Extracellular ATP and HMGB1 release were quantified using RealTime-Glo Extracellular ATP Assay (Promega) and Lumit HMGB1 Immunoassay (Promega) respectively. All the assays were conducted according to the manufacturers’ protocols.

Statistical Analysis

Data were expressed as mean ± SD from at least three independent experiments. Statistical significance was determined using Student’s t test or one-way ANOVA, with p-values <0.05 considered significant.

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/jacs.5c16025.

• Figures of Additional experimental details, methods, LNP characterizations and intracellular imaging results, immunoblots (PDF)

• Time-lapse imaging of ApoE-Onpattro (Video S1) (AVI)

• Time-lapse imaging of ApoE-porphyrin-LNP (Video S2) (AVI)

• Time-lapse imaging of Porphyrin-LNP (Video S3) (AVI)

The authors declare no competing financial interest.