PEG-lipid shedding and biodistribution are shaped by nanocarrier morphology and lipid chemistry

SUMMARY

Targeting moieties are often attached to nanocarriers (NCs) using poly(ethylene glycol)-lipids (PEG-lipids) to direct uptake to target cells and organs. However, nanoparticles (NPs) can interact with proteins during systemic circulation, causing PEG-lipids to dissociate or “shed,” which limits targeting efficiency. Here, we investigate how lipid chemistry and poly(ethylene glycol)-b-poly(propylene sulfide) (PEG-PPS) NC morphology affect PEG-lipid shedding. We conjugated one fluorophore to the NC and another to the PEG-lipid to track shedding over time. By coupling Förster resonance energy transfer (FRET) and total internal reflection fluorescence microscopy (TIRF) with biological systems, we found that PEG-lipids with stable anchors shed slowly and are consistent across NC morphologies. In contrast, weakly anchored PEG-lipids exhibit morphology-dependent shedding and biodistribution in vivo. These findings underscore how PEG-lipids can have different biodistribution from their NCs, highlighting the importance of this phenomenon in the design of new targeted nanomedicines.

Graphical Abstract

In brief

Nanomedicines often rely on surface display of PEG through incorporation of PEG-lipids to tune circulation and targeting, yet PEG-lipid retention at the nanoparticle surface remains incompletely understood. Here, we investigate how nanocarrier morphology and PEG-lipid chemistry influence PEG-lipid shedding behavior. By decoupling nanocarrier stability from PEG-lipid retention, this work establishes molecular-level design principles for targeted drug delivery and offers insights into the chemical-biological interface governing nanotechnology-driven therapeutics.

INTRODUCTION

Functionalizing nanocarriers (NCs) with targeting moieties such as antibodies and peptides is a well-established strategy to limit non-specific distribution and direct uptake to specific targets.^1,2 For example, functionalizing lipid nanoparticles (LNPs) with anti-CD3 antibodies or antibody fragments has shown promise in generating in vivo chimeric antigen receptor (CAR)-T cells against B cell malignancies.^3-5 These targeting moieties are covalently bound to poly(ethylene glycol)-lipids (PEG-lipids), which consist of a hydrophobic lipid tail that embeds within the LNP surface and a hydrophilic PEG chain that extends outward to form a steric “stealth” layer. In doing so, PEG-lipids can minimize protein adsorption, reduce clearance by the mononuclear phagocyte system, and enhance LNP stability, circulation time, and targeting specificity.⁶

During systemic circulation, PEG-lipids can undergo “shedding” whereby they detach from the LNP surface, compromising particle stability, reducing targeting and delivery efficiency, and increasing off-target exposure and toxicity.⁷ Single-particle analysis of liposomes isolated from mice 4 h post-injection revealed loss of targeting antibodies, highlighting the need to further study this phenomenon.⁸ It is hypothesized that the adsorption of bio-molecules onto the LNP surface destabilizes the PEG-lipid molecule and exposes the underlying ionizable lipids that modulate the biomolecular corona by selecting for specific proteins.^9,10 Shedding has been demonstrated by exposure to serum where free PEG-lipids diffuse faster than those tethered to an LNP.^7,11 In vivo, PEG-lipids have been shown to remain in circulation adsorbed to lipoproteins while LNPs distribute to tissues in mice.¹² This mechanism has been leveraged for selective targeting. For example, in the clinically approved formulation Onpattro, PEG-DMG, a 14-carbon, double-tailed lipid, is competitively dis-placed by apolipoprotein E, enabling selective delivery to hepatocytes.¹³

Substituting PEG-DMG with PEG-DSPE, an 18-carbon, double-tailed lipid that exhibits minimal shedding,^14,15 can reduce mRNA delivery and expression due to stronger interactions with the LNP.¹⁰ However, this strong PEG-DSPE anchoring and enhanced half-life can be advantageous in applications where prolonged circulation is required, such as immune modulation for in vivo CAR-T cells and transport of LNPs across the blood brain barrier.^3,15,16 Beyond acyl chain length, modification of the PEG-lipid terminal functional group also influences delivery efficacy. Reactive PEG end groups are commonly used for ligand conjugation to promote tissue- and cell-specific uptake, but unmodified reactive termini can induce anti-PEG antibody production and complement activation, leading to accelerated blood clearance after intravenous administration.^17,18 In contrast, in immune-privileged applications, PEG-lipids terminated with carboxylic acid or N-hydroxysuccinimide groups have been shown to enhance endocytosis and transfection of retinal cells, likely by strengthening LNP-cell receptor interactions.¹⁹ However, systematic studies defining design heuristics for other PEG-lipid chemical properties, including lipid tail saturation and linker chemistry, remain limited.

PEG-lipids have also been employed as anchors for targeting moieties in polymeric NCs, notably in morphologically diverse PEG-PPS polymersomes (PSs) and micelles (MCs).^20-22 While shedding is actively being studied in LNPs, there is limited investigation of this phenomenon in polymeric NCs, especially with respect to morphology. Here, we utilize PEG-PPS PSs and MCs grafted with a small library of PEG-lipids to systematically evaluate the effects of both lipid chemistry and NC morphology on shedding. PEG-PPS provides a unique platform for these studies for two reasons. First, as an amphiphilic block copolymer, PEG-PPS self-assembles into a variety of morphologies depending on PEG weight fraction, enabling assessment of morphological effects independently of surface chemistry. Second, the enhanced structural integrity of these NCs allows decoupling of particle stability from PEG-lipid shedding,^23,24 unlike LNPs, which require PEG-lipids for colloidal stability.^25,26

In this study, we employed a variety of optical techniques, including Förster resonance energy transfer (FRET) and total internal reflection fluorescence microscopy (TIRF), to quantify lipid association and dissociation in controlled environments devoid of complex biological interactions. This approach allowed us to directly assess the effects of both lipid chemistry and NC morphology on PEG-lipid stability. To complement these studies, we designed in vitro uptake and in vivo biodistribution experiments in which both core NCs and PEG-lipids were fluorescently tagged, enabling independent tracking. This integrated strategy provides a framework to understand how PEG-lipid chemistry and NP morphology jointly influence shedding, informing the rational design of targeted nanomedicines with improved stability and efficacy.

RESULTS

PEG-lipid PEG-PPS nanocarriers are structurally stable

We first synthesized PEG-PPS polymers with varying end groups, including inert benzyl (Bz) and reactive thiol (SH) and azide (N₃) functionalities (Figures S1-10). PEG₁₇-PPS₃₀-Bz and PEG₄₅-PPS₂₁-Bz served as the bulk polymer for PS and MC formulations, respectively. To enable fluorescent tracking of the NCs, 2.5 mol % of PEG₁₇-PPS₃₀-SH or PEG₄₅-PPS₂₁-SH was added for conjugation to a maleimide-functionalized fluorescent core dye. For the covalent control formulations, 2.5 mol % N₃-PEG₂₄-PPS₄₀ or N₃-PEG₄₅-PPS₂₁ was added to provide a reactive group for a DBCO-functionalized outer dye.

We then formulated NCs via thin film hydration with PEG-PPS polymers and N₃-PEG-lipids or N₃-PEG-PPS, followed by labeling with fluorescent core and outer lipid dyes (Figures 1A and 1B). All PS formulations exhibited narrow size distributions with polydispersity indices (PDIs) around 0.1, consistent with monodisperse populations (Figures 1C and 1E and Data S1).²⁷ The Z-average diameter increased with the addition of N₃-PEG-lipid or N₃-PEG-PPS (Figure 1E and Data S1), and one-way ANOVA confirmed significant differences in size across formulations depending on PEG-lipid composition (Table S1). These variations are consistent with known differences in lipid packing and compression arising from distinct chemical structures.^25,28 Quantification of surface ligands by nano flow cytometry revealed comparable loading across PS formulations, except for PEG-stearate (Figures 1D and S11 and Data S1). The low loading of PEG-stearate, a single-tailed PEG-lipid, suggests limited anchoring in PS, highlighting the importance of lipid-tail structure.

Figure 1. Characterization of PEG-b-PPS nanocarriers.

(A) Schematic of nanocarrier formulation.

(B) Chemical characteristics of N₃-PEG-lipids used in this work.

(C) Size distributions of PS formulations were determined by DLS.

(D) Loading of N₃-PEG-lipids in PS was determined by quantifying molecules of equivalent soluble fluorophores (MESF) per particle using nano flow cytometry. Data are presented as mean ± SD. Statistical significance was determined by a one-way ANOVA with Tukey’s post hoc test. LOD, limit of detection.

(E) Summary of PS properties including size, polydispersity index (PDI), and zeta potential.

(F) Size distributions of MC formulations were determined by DLS.

(G) Loading of N₃-PEG-lipids in MC was determined by quantifying fluorescence intensity using a spectrophotometer. Data are presented as mean ± SD. Statistical significance was determined by a one-way ANOVA with Tukey’s post hoc test.

(H) Summary of MC properties including size, PDI, and zeta potential.

All MC formulations exhibited PDIs below 0.3, indicative of relatively monodisperse populations (Figures 1F and 1H; Data S1).²⁹ Similar to PS, the Z-average diameter varied with ligand composition, albeit to a lesser extent (Table S2). PEG-PPS MCs were significantly larger than all other formulations (Figure 1H). Quantification of ligands by nano flow cytometry was not feasible for MCs due to their small size being below the instrument’s limit of detection. Therefore, loading was assessed by spectrophotometry. In contrast to PS formulations, PEG-stearate exhibited appreciable loading, whereas PEG-DMG, a C14 lipid, showed the lowest incorporation into MC (Figure 1G and Data S1). The zeta potential for both PS and MC formulations ranged between −10 mV and +10 mV, consistent with a neutral surface charge and previous work (Figures 1E and 1H; Data S1).^29-32

FRET reveals rapid shedding of PEG-lipids in mouse serum

To assess the colocalization of PEG-lipids with their NCs over time, we formulated dual-tagged AF594/AF647 NCs and measured FRET, which was inspired by previous work.^31,33,34 In this study, excitation of the AF594-core dye donor molecule transfers energy to excite the AF647-lipid dye acceptor molecule, resulting in AF647 emission when the two fluorophores are in close proximity (around 1–10 nm). Thus, stronger AF647 emission indicates greater colocalization, and a reduction in AF647 or increase in AF594 emission represents PEG-lipid shedding (Figure 2A; Data S2). After incubating PSs containing no-lipid or PEG-stea-rate in serum, minimal FRET was observed, consistent with low loading efficiencies (Figures 2B and 2C). For PSs with PEG-DMG, PEG-DSPE, and PEG-DOPE, AF647 emission rapidly decreased (Figures 2D-F). Among these, PEG-DSPE exhibited the most gradual shedding over 30 min (Figure 2E). In contrast, PEG-PPS PSs showed sustained FRET with no loss of AF647 emission indicative of stable colocalization (Figure 2G).

Figure 2. FRET shows rapid shedding of PEG-lipids from nanocarriers tagged with AF594/AF647 FRET pair.

FRET shedding assay was performed by using λEx at 530 nm and recording emission spectra from 560 to 750 nm for 30 min. (A) Schematic of experimental setup and serum-induced PEG-lipid shedding monitored using FRET. Emission spectra are shown for PS with (B) no lipid, (C) stearate, (D) DMG, (E) DSPE, (F) DOPE, and (G) PPS. Emission spectra are also shown for MC with (H) no lipid, (I) stearate, (J) DMG, (K) DSPE, (L) DOPE, and (M) PPS. All samples were completed in replicates (n = 3).

We then repeated this FRET assay with MC formulations. Similar to PSs, minimal FRET was observed in the no-lipid control and PEG-DMG, reflecting their low PEG-lipid loading (Figures 2H and 2J). PEG-DSPE and PEG-DOPE MCs exhibited a gradual loss of AF647 emission accompanied by an increase in AF594 emission, indicating progressive shedding over time (Figures 2K and 2L). Interestingly, both PEG-stearate and PEG-PPS MCs maintained stable FRET with no reduction in AF647 emissions over 30 min (Figures 2I and 2M).

Long-term experiments were also performed by incubating NCs with serum for up to 24 h; however, AF647 emission remained unchanged or was lost after 1 h, indicating that formulations were no longer shedding (Figure S12). The loss of FRET was also confirmed by disrupting the NCs with Triton X-100 or hydrogen peroxide for PSs and MCs, respectively. Disruption increased the diffusion of dyed PEG-lipid molecules away from the core dye, which prevented efficient energy transfer between fluorophores (Figure S13). An equimolar mixture of free AF594 and AF647 dyes also confirmed minimal FRET between the molecules (Figure S14).

To model FRET decay over time, non-linear regression analysis was performed to elucidate any differences in lipid chemistry or morphology. After normalizing FRET ratios to their disassembled controls at each time point, we modeled the FRET decay as a one-phase exponential decay. Formulations that shed too quickly (i.e., PEG-stearate PSs and PEG-DMG MCs) and those that did not shed (i.e., PEG-PPS PS, PEG-stearate MC, and PEG-PPS MC) were excluded from this analysis, as a decay constant could not be computed. We found that the fitted curves were significantly different when comparing morphologies (Figures S15A and S15B) and lipids (Figures S15C and S15D). However, when constructing 95% confidence intervals (CIs) for the FRET decay constant, we found that CIs for a given PEG-lipid overlapped with each other (Figure S15E). This suggests that shedding may not be morphologically dependent for PEG-lipids that stably anchor with their NCs.

When testing for PEG-lipid chemistry, we found that PEG-DSPE anchored most stably with both NC morphologies, as evidenced by the more gradual decline in FRET ratio over time compared to PEG-DOPE (Figures S15C and S15D). This finding is further supported by the 95% CIs that do not overlap with each other (Figure S15E), suggesting that PEG-DOPE may interact less strongly with the NCs. Collectively, these results suggest that strongly anchored PEG-lipids exhibit similar shedding profiles across NC morphology and highlight how lipid chemistry is a determinant of shedding for PEG-lipids that weakly anchor.

While FRET is commonly used as a measure of colocalization, it is possible that the loss of energy transfer could occur as a result of fluorophore reorientation, PEG rearrangement, or partial structure reorganization.³⁵ Therefore, to confirm that loss of FRET is due to PEG-lipid shedding, we performed size exclusion chromatography on PEG-DSPE and PEG-PPS PS formulations after incubation in serum for 2 h. The chromatogram for PEG-DSPE PSs shows that a significant portion of the tagged PEG-DSPE remains associated with the NC, as evidenced by signal in fractions collected at ~8.5 mL. As fractionating continues, PEG-DSPE appears at three distinct peaks between ~13 and 20 mL, consistent with PEG-DSPE shedding, and associates with serum proteins of varying sizes (Figure S16A). In contrast, the chromatogram for PEG-PPS PSs shows greater stability, as the tagged PEG-PPS largely elutes with the NC at ~8.5 mL. Only minimal tagged PEG-PPS appears at later elution volumes, indicating reduced shedding compared to PEG-DSPE and supporting FRET data (Figure S16B).

TIRF microscopy yields single-particle insight into shedding

TIRF was performed to gain single-particle insight into PEG-lipid shedding behavior over time. Chambered cover glass was functionalized with silane-PEG₂₄-DBCO to enable covalent attachment of N₃-functionalized NCs. Serum was then introduced to induce PEG-lipid shedding and NC desorption from the surface. Representative images from 0 to 2 h illustrate shedding kinetics from a single-particle perspective (Figures S17 and S18; Data S3). Surface-area coverage was quantified over time, and the desorption rates were determined using linear regression (Figure 3).

Figure 3. TIRF demonstrates shedding of tethered nanocarriers over time.

N₃-grafted nanocarriers were tethered to the surface of DBCO-functionalized cover glass and observed over time after addition of PBS or mouse serum. (A) Schematic of experimental setup and serum-induced PEG-lipid shedding monitored using TIRF. Images were acquired for PS with (B) no lipid, (C) stearate, (D) DMG, (E) DSPE, (F) DOPE, and (G) PPS. The mean surface area coverage is shown. (H) 95% confidence intervals of desorption rates of PS using linear regression were constructed. Images were also acquired for MC with (I) no lipid, (J) stearate, (K) DMG, (L) DSPE, (M) DOPE, and (N) PPS. The mean surface area coverage is shown. (O) 95% confidence intervals of desorption rates of MC using one-phase exponential decay were constructed.

Consistent with FRET studies, our analysis shows that PEG-PPS PSs exhibited the least change in surface-area coverage over time with the lowest desorption rate constant (Figures 3G and 3H). PEG-DSPE PSs maintained consistent surface coverage over time in phosphate buffered saline (PBS) but showed decreased retention in 50% mouse serum (Figure 3E). Quantitatively, the desorption rate coefficient for PEG-DSPE in serum is lower than that of PEG-DOPE, PEG-stearate, and no-lipid formulations, indicating stronger tethering of the PS to the DBCO-functionalized cover glass (Figure 3H). Although PEG-DMG exhibited greater shedding in PBS compared to PEG-DSPE, both showed comparable desorption rates in serum, as evidenced by overlapping CIs (Figures 3D, 3E, and 3H). The no-lipid control and PEG-stearate displayed the highest degree of desorption over time, consistent with weak surface attachment (Figures 3B and 3C). Overall, these results reveal lipid-dependent differences in the shedding behavior of tethered PSs, underscoring the role of lipid chemistry in surface stability.

We next repeated this experiment with MCs and observed rapid loss of surface coverage for all formulations immediately following the addition of serum, indicating a faster rate of shedding for this morphology (Figures 3I-3N). Therefore, desorption rates were estimated using a one-phase exponential decay model, and 95% CIs were constructed (Figure 3O). PEG-PPS MCs followed the same trend as PEG-PPS PSs, with the lowest desorption rate and the least shedding over time (Figures 3N and 3O). PEG-stearate and PEG-DSPE MCs exhibited comparable shedding behavior (Figures 3J, 3L, and 3M), consistent with trends observed in FRET studies showing minimal changes in AF647 emission over time (Figures S12H and S12J). PEG-DMG and PEG-DOPE MCs displayed the greatest desorption, with highly overlapping CIs indicating similar shedding kinetics (Figures 3K, 3M, and 3O). The no-lipid MCs desorbed so rapidly that the initial reduction was not captured in the one-phase exponential decay model, resulting in an artificially low calculated desorption coefficient (Figures 3I and 3O). This formulation also had the least surface-area coverage at the end of the 2 h (Figure 3I). Overall, these results reinforce the idea that shedding behavior is dictated by PEG-lipid chemistry and further highlight morphology-dependent differences.

PEG-lipids insert into the membrane of cells prior to NC uptake

We next evaluated PEG-lipid shedding in the context of cellular uptake. To determine subcellular trafficking over time, RAW 264.7 macrophage-like cells were incubated with dual-tagged AF647/AF594 NCs for 2 and 24 h (Data S4). After 2 h, flow cytometry revealed a high percentage of cells positive for the AF647 lipid dye, with minimal internalization of the AF594 core dye (Figure 4A). While PEG-PPS PSs exhibited limited AF647 uptake, the MC morphology showed a substantially higher percentage of AF647⁺ cells (Figure 4A). After 24 h, uptake of the AF594 core dye increased, and the proportion of AF647⁺AF594⁺ cells closely matched that of AF594⁺ cells, indicating colocalization of the lipid and core dyes (Figure 4B). Interestingly, PEG-DSPE and PEG-DOPE formulations displayed lower overall uptake in both morphologies compared to other PEG-lipid compositions, for which nearly 100% of cells internalized both dyes (Figure 4B). Based on these results, we hypothesized a two-step mechanism in which PEG-lipids initially shed from the NC and intercalate into the cell membrane at early time points, followed by uptake of the NC core at later times.

Figure 4. Treatment of RAW 264.7 macrophage-like cells shows differential uptake between PEG-lipid and nanocarriers.

RAW 264.7 cells were treated with 0.5 mg/mL polymer of AF594/AF647 dual-tagged nanocarriers and treated for 2 or 24 h. Uptake was then evaluated using flow cytometry to measure % positive cells after (A) 2 h and (B) 24 h. Widefield fluorescence microscopy was used to evaluate subcellular location after (C) 2 h and (D) 24 h. A 100× objective with 1.5× magnification was used with the following dyes: blue = Hoechst 33342 nuclear stain, red = AF594-core dye, and yellow = AF647-lipid dye. Scale bars: 10 μm.

To test this hypothesis and validate our results from flow cytometry, we used widefield fluorescence microscopy to visualize subcellular trafficking. Representative images after 2 and 24 h demonstrate distinct uptake behaviors between PEG-lipids and NCs (Figures 4C, 4D, and S19). After 2 h, the PEG-lipid fluorescence appeared as a distinct corona outlining the cell membrane with minimal internalization of the NC core, supporting our earlier hypothesis (Figure 4C). In contrast, PEG-PPS formulations were not confined to the periphery, as indicated by their more diffuse intracellular distribution. After 24 h, both lipid and core dye fluorescence were dispersed throughout the cyto-plasm, consistent with internalization (Figure 4D). AF594 core dye signal intensity was lower overall, aligning with flow cytometry results (Figures 4B and 4D).

To test whether these conclusions could be generalized to other cell types, we repeated these experiments in the DC2.4 dendritic cell line. Our results revealed very similar behaviors to those of RAW 264.7 cells. Initially, a large percentage of cells were positive for AF647 lipid dye, with minimal uptake of AF594 core dye. Similar to RAW264.7 cells, PEG-PPS PSs exhibited minimal AF647 lipid dye as opposed to the MC counter-part (Figure S20A). After 24 h, both the uptake of AF594 core dye and the proportion of AF647⁺AF594⁺ cells increased for all grafted PS formulations and remained relatively low for PEG-DSPE, PEG-DOPE, and PEG-PPS MC formulations (Figure S20B). Notably, the proportion of AF647⁺AF594⁺ DC2.4 cells was greater for PEG-DSPE and PEG-DOPE PSs than in RAW 264.7 cells after 24 h, highlighting how cell type may be a determinant in uptake behavior (Figures 4B and S20B). Microscopy data supported the flow cytometry data and showed similar trends to the RAW 264.7 cells (Figure S21). Based on these results, we can conclude that shedding at early time points seems to remain consistent between cell types. At later time points, morphology and cell type play a larger role in uptake kinetics.

To assess applicability to primary cells and determine whether PEG-lipids exhibit preferential affinity for certain immune cell types, we performed flow cytometry following incubation with splenocytes from healthy female C57BL/6J mice for 30 min. Similar to NC treatment in immortalized cells, a high percentage of cells were positive for the AF647 lipid dye, with minimal internalization of the AF594 core dye for PEG-lipid formulations (Figures 4 and S20-S22). Across both morphologies, splenocytes treated with PEG-DSPE NCs exhibited a lower proportion of AF647⁺ cells for all immune cell types compared with PEG-DOPE NCs. Due to the differential stability of the PEG-lipids, uptake diverges for less-stable PEG-stearate and PEG-DMG formulations (Figure S22I). Overall, PEG-lipids have a higher affinity toward macrophages and dendritic cells than they do toward B cells and T cells. When comparing molecular characteristics, PEG-PPS demonstrated the strongest anchoring, followed by PEG-DSPE, PEG-DOPE, PEG-DMG, and PEG-stearate.

Stably anchored PEG-lipids localize with their NC following subcutaneous injection

To assess PEG-lipid shedding in a physiologically relevant context, dual-labeled NCs were administered subcutaneously (s.c.) to healthy mice. To minimize background tissue autofluorescence, we employed red-shifted fluorophores and labeled the PEG-lipid with AF647 and the NC core with AF750. We validated that these fluorophores also form a FRET pair, although some re-sidual AF750 fluorescence was observed in the absence of AF647, limiting the strength of conclusions that can be drawn from FRET signal alone (Figure S24). To capture the initial kinetics of PEG-lipid shedding in vivo, we evaluated biodistribution at early time points (2, 4, and 8 h) (Figures 5 and S25-S34; Data S5-S25).

Figure 5. In vivo imaging system highlights differences in PEG-lipid and nanocarrier biodistribution.

Mice were injected with NC s.c., and organs were harvested following perfusion 4 h post-injection.

(A) Representative in vivo imaging system (IVIS) images of AF647 lipid dye for PS formulations, and quantification of signal in the liver, kidney, and draining lymph node (dLN).

(B) Representative IVIS images of AF750 core dye for PS formulations, and quantification of signal in the liver, kidney, and dLN.

(C) Representative IVIS images of AF647 lipid dye for MC formulations, and quantification of signal in the liver, kidney, and dLN.

(D) Representative IVIS images of AF750 core dye for MC formulations, and quantification of signal in the liver, kidney, and dLN.

Data points represent individual mice; n = 5 mice/group. Statistical significance was determined by a one-way ANOVA with Dunnet’s post hoc test compared to saline.

At 4 h post-injection, PS formulations containing PEG-stearate and PEG-DMG exhibited predominant AF647 signal in the kidneys, suggesting lipid shedding and renal clearance, whereas AF750 signal from the core PS dye was primarily observed in the liver (Figures 5A, 5B, and S25; Data S5). In contrast, formulations containing more stably anchored PEG-DSPE, PEG-DOPE, and PEG-PPS displayed localized AF647 and AF750 signals in the liver and draining lymph nodes (dLNs), indicative of retained lipid-core association and lymphatic transport. A strong FRET signal in these organs also suggests lipid-core colocalization (Figure S26; Data S9 and S11). MC formulations showed similar AF647 lipid biodistribution, but AF750 core fluorescence was detected in both the liver and the kidneys, owing to their smaller hydrodynamic size (Figures 5C, 5D, S27, and S28; Data S10).

Analysis of the time-dependent biodistribution over 2, 4, and 8 h provided further insight into PEG-lipid shedding kinetics in vivo (Figures S29-S32; Data S6 and S12-S15). For no-lipid PSs, the AF750 signal from the core dye increased linearly over time in the liver but remained low in the kidneys, consistent with hepatic accumulation and limited renal clearance of these larger particles (Figures 6A and 6B). Among PEG-lipid-modified PSs, PEG-stearate and PEG-DMG exhibited strong AF647 signal in the kidneys at 2 h that decreased over time, suggesting rapid lipid shedding followed by renal clearance (Figures 6B and 6D). In contrast, more stably anchored PEG-DSPE and PEG-DOPE formulations showed lower early kidney accumulation that gradually increased with time, indicative of slower, progressive shedding (Figures 6B and 6D). PEG-PPS PS formulations, which possess the most stable anchoring chemistry, displayed negligible AF647 signal in the kidneys across all time points, consistent with minimal dissociation.

Figure 6. Longitudinal IVIS analysis reveals in vivo PEG-lipid shedding kinetics.

Mice were injected with NC s.c., and organs were harvested following perfusion 2, 4, or 8 h post-injection.

(A and B) Quantification of AF647 lipid dye and AF750 core dye fluorescence in the (A) liver and (B) kidney for PS formulations over time.

(C and D) Quantification of AF647 lipid dye and AF750 core dye fluorescence in the (C) liver and (D) kidney for MC formulations over time.

Data points represent individual mice, and lines represent mean values; n = 5 mice/group.

For no-lipid MCs, AF750 signal increased in both the liver and the kidney between 2 and 4 h, followed by a slight decrease by 8 h (Figures 6C and 6D; Data S6). PEG-DSPE and PEG-DOPE MCs exhibited parallel trends in AF647 and AF750 signal intensity, suggesting coordinated lipid-core transport and minimal early dissociation. Across both morphologies, the kinetics of accumulation in the spleen and dLN mirrored those in the liver, consistent with gradual lymphatic drainage and systemic distribution following s.c. administration (Figure S33 and Data S16).

To determine whether PEG-lipid trafficking is driven by association with NCs or by intrinsic biodistribution, labeled PEG₄₅-lipids were administered s.c. at a molar-equivalent dose to the previous experiment, and biodistribution was evaluated 4 h post-injection (Figure S34A and Data S17). Free PEG-lipids exhibited biodistribution profiles similar to those grafted onto MCs, with strong AF647 signals observed in the kidneys for PEG-DSPE and PEG-DOPE (Figures 5C and S34B). In contrast, free PEG-stearate showed low kidney signal after 4 h, whereas PEG-stearate MCs had strong signals in the kidneys at all time points. We attribute this difference to rapid clearance of free PEG-stearate, with grafting onto MCs extending its apparent half-life. Based on these observations, we hypothesize a two-step process in which PEG-lipids first shed from the NC and sub-sequently traffic according to their intrinsic biodistribution.

NCs primarily colocalize with phagocytes in dLNs

Flow cytometry was performed on single-cell suspensions of dLNs to identify the immune cell populations responsible for NC uptake (Figures 7 and S35-S37; Data S18-S19). Across both morphologies, the highest levels of uptake were observed in macrophages (CD11b⁺F4/80⁺), followed by dendritic cells (CD11c⁺MHCII⁺) (Figures 7A, 7B, 7I, and 7J). PEG-DSPE and PEG-DOPE formulations showed greater uptake of the PEG-lipid than the NC core across both morphologies, indicative of PEG-lipid shedding. Lymphocytes exhibited a comparatively low fluorescent signal, consistent with limited phagocytic capacity (Figures 7C, 7D, 7K, and 7L). Among PS formulations, PEG-PPS PSs displayed the highest association with both B cells (CD19⁺) and T cells (CD3⁺), accompanied by a strong double-positive AF647⁺AF750⁺ signal, consistent with intact NCs. In contrast, PEG-PPS MCs exhibited minimal association with lymphocytes. Instead, PEG-DSPE and PEG-DOPE resulted in the highest percentage of AF647⁺ lymphocytes, with negligible uptake of AF750 across all formulations. Consistent with in vitro data, PEG-PPS MCs showed greater uptake of the lipid dye than the core dye, suggesting reduced structural stability of this morphology compared to PSs.

Figure 7. Phagocytes in dLNs are primarily responsible for NC uptake.

Mice were injected with NC s.c., and dLNs were harvested and processed for flow cytometry after 2, 4, or 8 h.

(A–D) Percentage of cells positive for AF647 lipid dye, AF750 core dye, or both dyes for PS formulations in (A) macrophages, (B) dendritic cells, (C) B cells, and (D) T cells.

(E–H) Percentage of total immune cells positive for AF647 lipid dye or AF750 core dye for (E) DSPE, (F) DOPE, (G) PPS, and (H) no-lipid PS.

(I–L) Percentage of cells positive for AF647 lipid dye, AF750 core dye, or both dyes for MC formulations in (I) macrophages, (J) dendritic cells, (K) B cells, and (L) T cells.

(M–P) Percentage of total immune cells positive for AF647 lipid dye or AF750 core dye for (M) DSPE, (N) DOPE, (O) PPS, and (P) no-lipid MC.

Data points represent individual mice, and lines represent mean values; n = 5 mice/group.

Consistent with the dLN in vivo imaging system (IVIS) data, a higher proportion of immune cells (CD45⁺) were positive for the AF647 lipid dye than for the AF750 core dye over time for both PEG-DSPE and PEG-DOPE PS formulations (Figures 7E and 7F). For PEG-PPS PSs, the percentages of AF647⁺ and AF750⁺ cells were comparable and increased modestly over time, suggesting progressive uptake of intact NCs by immune cells in the dLN (Figure 7G).

The kinetics of MC uptake were distinct, with PEG-DSPE and PEG-DOPE formulations showing peaks in AF647⁺ cells at 4 h (Figures 7M and 7N). These formulations exhibited substantially higher AF647 uptake, with over 30% of immune cells staining positive in some cases, while AF750 uptake remained minimal, consistent with PEG-lipid shedding after lymphatic transport. In contrast, PEG-PPS MCs showed minimal uptake of both dyes at all time points, similar to no-lipid, PEG-stearate, and PEG-DMG for both morphologies (Figures 7H, 7O, 7P, S35, and S36). Together, these data suggest that PS formulations maintain greater structural integrity after reaching the dLN, whereas MC formulations undergo rapid PEG-lipid loss and limited delivery of intact NCs to immune cells.

DISCUSSION

The lack of PEG-lipid shedding heuristics and design principles limits the clinical translation of targeted NPs due to safety and efficacy concerns. In this study, we begin to address this gap by developing quantitative methods to measure PEG-lipid shedding. We engineered synthetic NCs grafted with PEG-lipids that enable systematic interrogation of PEG-lipid behavior. Using FRET and TIRF along with in vitro and in vivo studies, we quantified shedding as a function of lipid chemistry and NC morphology.

FRET experiments revealed that shedding is dependent on both lipid chemistry and NC morphology. PEG-DSPE and PEG-DOPE formulations had similar rates of FRET loss over time across PSs and MCs but differed from each other (Figure S16), suggesting that lipid chemistry plays a larger role than morphology when PEG-lipids are strongly anchored. This contrasts previous reports highlighting morphology as a determinant of therapeutic efficacy dependent on structural stability.³¹ Differences in FRET loss between PEG-DSPE and PEG-DOPE point to the importance of lipid saturation, consistent with previous work showing that unsaturation in PEG-DOPE increases lateral mobility and promotes shedding.^36,37 Although we could not directly quantify the effects of acyl chain number and length, both appear to influence shedding, as indicated by low energy transfer for PEG-stearate and PEG-DMG in serum (Figure 2). These results align with prior knowledge that stipulates that longer acyl chains enhance anchoring, reducing PEG-lipid shedding and potentially limiting cargo release.^10,14,38,39

We leveraged the high-throughput capabilities of TIRF to elucidate chemical determinants of shedding on a single-particle basis. PEG-stearate PSs exhibited a shedding profile similar to PSs without lipid, indicating that a single acyl chain provides limited surface stabilization (Figures 3B and 3C). Differences between PEG-DSPE and PEG-DOPE PSs further highlight the idea that unsaturation enhances shedding. While PEG-DMG shed rapidly from PSs in FRET studies, its desorption coefficient is between that of PEG-DSPE and PEG-DOPE for TIRF. Interestingly, in both FRET and TIRF, PEG-DMG exhibited shedding in PBS alone, suggesting that this behavior is largely independent of protein-mediated effects. In contrast to PSs, MCs exhibited a shedding profile best described by exponential decay, highlighting morphology-dependent differences between NCs. Trends for MCs were largely consistent between TIRF and FRET, with PEG-PPS MCs showing the strongest surface tethering.

A limitation of TIRF is the nonspecific adsorption of the no-lipid controls, which persisted despite passivation of the glass surface. This effect may be reduced in serum due to protein competition for the NC surface and is supported by the dramatic loss in surface coverage observed for no-lipid PSs and MCs, which may not be fully captured by desorption rate constants alone (Figures 3H and 3I). Future studies should focus on developing improved coatings to further reduce nonspecific adsorption for single-particle analysis. Overall, TIRF results matched trends observed by FRET, with stable PEG-PPS association, slower shedding of PEG-DSPE, and rapid shedding of PEG-DMG and PEG-DOPE. While PEG-DSPE appeared to be the most stable PEG-lipid, its phospholipid linkage may contribute to lateral mobility within the membrane.³⁷ Future studies using PEG-DSG, an 18-carbon PEG-lipid without a phospholipid, could disentangle the effects of acyl length and linkage chemistry.

Subcellular localization and uptake studies revealed distinct internalization pathways for PEG-lipids and PEG-PPS NCs. PEG-lipids rapidly shed from their carriers and associate with the cell periphery, followed by slow internalization of the NC (Figure 4). For PEG-PPS ligands, uptake differed between PS and MC morphologies. The labeled PEG-PPS ligand was internalized before the core for MCs, while it remained stably associated with the core for PSs (Figures 4A and 4B). These observations are consistent with the lower stability and faster cargo release reported for MCs compared to PSs.^31,40

Our dual-labeled polymeric NCs enabled independent tracking of PEG-lipids and core NCs to capture the kinetics of PEG-lipid dissociation following s.c. administration through simultaneous analysis of biodistribution, lymphatic transport, and cellular uptake. Together, these results reveal that the stability of the PEG-lipid and the NC architecture govern not only the extent of lipid retention but also the delivery of intact particles to immune cells in the dLN. Across both PSs and MCs, the behavior of PEG-lipids in vivo reflected their anchoring stability observed in vitro. PEG-stearate and PEG-DMG, which possess shorter or saturated acyl chains, exhibited strong early kidney accumulation, consistent with rapid desorption from the NC surface and renal clearance. In contrast, PEG-DSPE and PEG-DOPE displayed slower, progressive increases in kidney signal, indicating more gradual shedding (Figures 5 and 6). These findings underscore the strong influence of anchor hydrophobicity and linkage chemistry on the residence time of PEG-lipids at the NC surface in vivo.

NC morphology also played a central role in lipid retention and biodistribution. PS formulations exhibited more coordinated AF647 and AF750 signals, indicating that lipid and core components trafficked together, particularly for DSPE-, DOPE-, and PPS-anchored formulations. In contrast, MC formulations showed dissociation, with AF647 and AF750 signals diverging over time (Figures 6 and S25-S33). The smaller hydrodynamic size of MCs likely facilitated systemic distribution and enhanced clearance, which may contribute to the observed divergence in biodistribution independent of PEG-lipid shedding. At the same time, the increased curvature and dynamic assembly of MCs may reduce the energetic stability of surface-anchored PEG-lipids, rendering them more susceptible to desorption during transport through interstitial and lymphatic environments.

Within the dLNs, immune cell uptake patterns further reflected PEG-lipid stability trends (Figure 7). Flow cytometry revealed that macrophages and dendritic cells were the predominant cell types associated with NCs, consistent with their phagocytic roles. Among PS formulations, PEG-DSPE and PEG-DOPE showed modest but consistent uptake of both lipid and core dyes, suggesting the delivery of partially intact particles. In contrast, PEG-PPS PSs exhibited strong double-positive AF647⁺AF750⁺ signals in both phagocytes and lymphocytes, consistent with preserved lipid-core association (Figure 7). We note that double-positive signals alone do not definitively indicate intact NC uptake, as co-uptake of shed PEG-lipids and NC cores could also generate this pattern. To limit this concern, we also evaluated FRET measurements by IVIS, which support PEG-PPS PS lipid-core colocalization. In contrast to PSs, PEG-PPS MCs displayed minimal cellular uptake, while PEG-DSPE and PEG-DOPE MCs exhibited pronounced lipid dye staining with negligible core signal, particularly after 4 h (Figure 7). This pattern, together with reduced FRET, is consistent with PEG-lipid shedding and subsequent membrane association following arrival in the dLN.

The kinetics of immune cell association mirrored these mechanistic differences. For PS formulations, the proportion of lipid- and core-dye positive CD45⁺ cells increased gradually over time, suggesting steady trafficking and uptake of intact particles. In contrast, PEG-DSPE and PEG-DOPE MCs exhibited a sharp increase in lipid-positive cells at 4 h followed by a decline, consistent with transient lipid association and rapid clearance or redistribution (Figure 7). Together, these data support a model in which PS formulations maintain structural integrity during lymphatic transport, enabling delivery of intact NCs to phagocytic cells, whereas MC formulations undergo early PEG-lipid shedding, limiting core delivery but potentially promoting surface exchange with cell membranes.

Our labeling strategies to fluorescently tag the PEG-lipid and PEG-PPS polymer utilize strain-promoted azide-alkyne cyclo-addition (SPAAC) and thiol-maleimide click chemistry, respectively. While the bonds formed from SPAAC are known to be stable under physiological conditions, the thioether bond formed from thiol-maleimide click chemistry is known to be prone to hydrolysis.⁴¹ Therefore, cleavage of the bond between PEG-PPS-SH and AF750 could lead to increased signal in the kidney. Both AF647 and AF750 also contain amide bonds between the clickable region and the fluorophore. While amides are generally stable in neutral buffers, enzymatic degradation of these bonds can occur in cells.⁴² Exocytosis of these cleaved fluorophores and subsequent excretion may lead to false signals in the kidneys as well. To decouple fluorophore cleavage and metabolic degradation from stability, future work should incorporate secondary, biological readouts, such as an application in a disease model, to evaluate how shedding affects pathological outcomes.

Our study focuses exclusively on s.c. administration where transport barriers and lymphatic drainage play key roles in PEG-lipid shedding and targeted delivery to immune cells. Interstitial fluid differs from blood in protein composition and concentration due to size and charge barriers,⁴³ which may result in shedding behaviors that diverge from those of intravenously (i.v.) administered NCs, due to differences in biomolecular corona formation.^44,45 Additional factors such as shear stress and clearance mechanisms may further contribute to differences between i.v. and s.c. administration. Studies comparing i.v. and s.c. administration of LNPs have demonstrated differences in biodistribution and anti-PEG antibody titers, highlighting delivery-route-dependent pharmacokinetics and clearance mechanisms.^46-48 Therefore, our proposed shedding mechanism is specific to s.c. administration and may not directly translate to other delivery routes.

Another limitation is that animal studies were performed in young, healthy female mice, which may not capture sex- or age-dependent differences in lymphatic transport or immune cell uptake. Moreover, the biodistribution and shedding kinetics observed here likely differ in pathological contexts where inflammation, vascular permeability, and macrophage activation are altered. Our analysis was restricted to early time points and relied on fluorophore intensity as a surrogate for lipid and core localization, which may not fully resolve intact versus dissociated nanocarriers. Additionally, IVIS imaging and flow cytometry provide limited spatial resolution, and immune cell profiling focused on uptake rather than functional consequences. Despite these limitations, this study establishes a mechanistic framework for understanding how PEG-lipid anchoring chemistry and NC structure govern delivery in vivo.

Overall, these findings highlight how both anchoring chemistry and NC morphology determine the stability and fate of PEG-lipid-modified systems in vivo. Stable anchors such as PEG-DSPE and PEG-PPS preserve surface coating and integrity to a higher degree than weaker anchors do across morphologies, which may be advantageous for sustained delivery or antigen presentation. In contrast, formulations with faster-shedding lipids may facilitate transient PEG exposure or lipid transfer to biological membranes, potentially enhancing immune recognition or clearance. Understanding and tuning these molecular and structural parameters will be essential for designing NCs with predictable in vivo behavior and optimized delivery to immune targets.

METHODS

NC formulation and characterization

PEG-PPS polymers were first synthesized and characterized using ¹H-NMR. Further details can be found in the supplemental information. Core-dyed NCs were formed via thin film hydration. Briefly, PEG-PPS-Bz was mixed with 2.5 mol % PEG-PPS-SH, dissolved in organic solvent, and dried to form thin films for several hours. The films were then hydrated with 400 μL of 1× PBS and shaken vigorously overnight. To dye the NCs, 1.1 eq AZ-Dye594-maleimide (Vector FP-1102) or IR750-maleimide (Vector FP-1557) with respect to PEG-PPS-SH was added to the solutions, and shaking was continued overnight. Unencapsulated dyes were removed using 7K MWCO Zeba Desalting Columns (Thermo Fisher Scientific) followed by dialysis against 1× PBS in 7K MWCO cassettes (Thermo Fisher Scientific). After dialysis, NCs were sterile-filtered multiple times using a 0.2 μm filter.

To functionalize NCs with N₃-PEG-PPS-Bz, 2.5 mol % N₃-PEG₁₇-PPS₃₀-Bz and N₃-PEG₄₅-PPS₂₁-Bz were added into the organic solvent during thin film hydration for PSs and MCs, respectively. For N₃-PEG-lipid-grafted NCs, core-dyed NCs were then shaken with 2.5 mol % N₃-PEG-lipid overnight. For PSs, N₃-PEG₂₄-Stearate (BroadPharm-41904), N₃-PEG₂₄-DMG (BroadPharm-41903), N₃-PEG₂₄-DSPE (BroadPharm-26187), and N₃-PEG₂₄-DOPE (BroadPharm-28699) were used. For MCs, N₃-PEG₄₅-Stearate (BroadPharm-43810), N₃-PEG₄₅-DMG (BroadPharm-43716), N₃-PEG₄₅-DSPE (BroadPharm-26188), and N₃-PEG₄₅-DOPE (BroadPharm-28700) were used. Therefore, for each PS and MC, there were 6 NC formulations, including a group with no N₃ grafting. Ungrafted PEG-lipid was removed from formulations using multiple 7K MWCO Zeba Desalting Columns. N₃-grafted NCs were then incubated with 1.1 eq AZDye647-DBCO (Vector CCT-1302) with respect to the N₃ ligand. Unconjugated dye was removed using multiple 7K MWCO Zeba Desalting Columns. NCs were characterized using dynamic light scattering to verify monodispersity and zeta potential.

To quantify loading of the N₃ ligand, NCs were incubated with 1.1 eq AZDye488-DBCO (Vector CCT-1278) with respect to the N₃ ligand. Unconjugated dye was removed using multiple 7K MWCO Zeba Desalting Columns. For PSs, loading was measured using nano flow cytometry (NanoFCM) by quantifying the molecules of equivalent soluble fluorophores (MESF) against a standard curve determined by MESF beads provided by NanoFCM as previously described.⁴⁹ Briefly, the 488 nm was used with 488/10 nm and 525/40 nm bandpass filters used for SSC and AF488 dye, respectively. Quality control (NanoFCM) and S16M-Exo beads (NanoFCM) were used in 1:1000 dilutions to calibrate the instrument. PSs were diluted to a concentration of 0.3 μg/mL polymer in 0.2 μm-filtered 1× PBS, and a minimum of 2,000 events were recorded for each formulation. For MCs, loading was measured using a spectrophotometer (Tecan Infinite MPlex) using 494 nm excitation/517 nm emission. Each formulation had three independently prepared NCs for a n = 3 for each group. A one-way ANOVA was performed to determine any differences in loading of N₃ ligands.

Fluorescence resonance energy transfer shedding assay

AF594-core dye/AF647-lipid dye NCs were diluted to 1 mg/mL polymer. 50 μL of NC was added to a black 96-well plate along with 50 μL of normal mouse serum (Invitrogen 10410). The spectra from 560 to 750 nm was recorded every 5 min for 30 min using λ_Ex of 530 nm. A different well was used for each time point to ensure minimal bleaching, and this was repeated three times (n = 3). The fluorescence was normalized to the maximum value recorded to ensure standardization across samples. To determine shedding over time, the FRET ratio was computed as shown below in Equation 1:

$$FRETRatio=\frac{I_{647}}{I_{647}+I_{594}},$$ (Equation 1)

where $I_{647}$ is the emission intensity of the AF647 acceptor and $I_{594}$ is the emission intensity of the AF594 donor. The FRET ratios were then normalized to the destructed particles treated with Triton X or hydrogen peroxide. A one-phase decay was fitted, and an extra sum of squares F-test was performed to determine significance between lipid chemistry and NP morphology.

Treatment and uptake of NCs in RAW 264.7 macrophages

RAW 264.7 macrophages were plated overnight in a 96-well plate at a density of 80,000 cells per well in Dulbecco’s Modified Eagle Medium (DMEM) supplemented with 10% fetal bovine serum (FBS), 10 mM 4-(2-hydroxyethyl)-1-piperazineethan esulfonic acid (HEPES), 1 mM pyruvate, and 100 U/mL penicillin/streptomycin. The next day, the cells were washed twice with unsupplemented DMEM. Macrophages were treated with PSs or MCs at a concentration of 0.5 mg/mL polymer for both 2 and 24 h in unsupplemented media. Each group contained an n = 3. Afterward, macrophages were dissociated from the plate using 1 mM ethylenediaminetetraacetic acid (EDTA) and 1 mM ethylene glycol bis(2-aminoethyl ether)-N,N,N′,N′ tetra-acetic acid (EGTA) in PBS. Macrophages were then stained with Live/Dead fixable blue. Samples were recorded on a BD LSR-Fortessa and analyzed using FlowJo software. The entire sample (~80,000 cells) was recorded.

RAW 264.7 macrophages were plated overnight in an 8-well cover glass plate at a density of 10,000 cells per well in DMEM supplemented with 10% FBS, 10 mM HEPES, 1 mM pyruvate, and 100 U/mL penicillin/streptomycin. The next day, the cells were washed twice with unsupplemented DMEM. Macrophages were treated with PSs or MCs at a concentration of 0.5 mg/mL polymer for both 2 and 24 h in unsupplemented media. After treatment, cells were washed with PBS multiple times to remove treatment conditions and stained with Hoescht 33342 (Invitrogen) for 15 min at 37°C. Cells were then washed with PBS multiple times to remove excess dye. Cells were then imaged with widefield fluorescence using a Nikon Ti2 inverted microscope stand with a 100× objective and 1.5× magnification. The 385, 561, and 620 nm lasers were used for Hoescht 33342, AF594-core dye, and AF647-lipid dye, respectively.

Total internal reflection fluorescence microscopy

To functionalize the surface of cover glass with DBCO, 18-well chambered cover glass (Cellvis C18-1.5H) was plasma treated for 3 min to introduce hydroxyls. The cover glass was then incubated with 100 μL silane-PEG₁₀₀₀-DBCO (Nanosoft Polymers) dissolved in ethanol/water solution (95%/5% w/w) for several hours at 4°C. Unreacted silane-PEG₁₀₀₀-DBCO was removed by washing each well 5 times with ultrapure water. The surface was then blocked with 1% w/w bovine serum albumin (BSA) solution for 1 h at room temperature. Excess BSA was washed off 5 times with 1× PBS. 100 μL N₃-modified NCs loaded with AF594 dye were added to each well at a concentration of 0.25 μg/mL polymer for both PSs and MCs for 1 h at 4°C. Finally, excess NCs were washed from each well 5 times using 1× PBS. Each well was left in 100 μL PBS until TIRF imaging occurred immediately after.

For PSs, TIRF images were captured with a Nikon Ti2 (Nikon Instruments Inc) inverted microscope stand with Nikon H-TIRF and a Photometrics Prime95B 25 mm camera. A Nikon Apo TIRF 100× (NA 1.49) oil immersion objective and perfect focus system were used. For MCs, a Nikon TiE inverted microscope stand with Nikon TIRF-E and a Hamamatsu FLASH 4 V3 camera with a Nikon Apo TIRF 100× (NA 1.49) oil immersion objective and perfect focus system were used. TIRF on both systems was aligned by Center for Advanced Microscopy staff at the start of the experiment. The 561 nm laser line was used for both systems, and a TokaiHit incubation chamber was set to 37°C without CO₂. During acquisition, two baseline images were captured for 2 min, after which either an additional 100 μL PBS or normal mouse serum (Invitrogen) was added. Images were then captured every minute for 10 min followed by every 5 min for 110 min. Five different positions (n = 5) were captured for each well. Images were analyzed in FIJI by quantifying surface coverage over time.¹⁸ Briefly, bleach correction was applied using histogram matching followed by thresholding using Otsu. Surface coverage was measured in all wells and presented as percent coverage normalized to when serum or PBS was added. Curves were fitted, and 95% CIs were constructed for each replicate.

In vivo biodistribution experiments

All experiments were performed in accordance with the Institutional Animal Care and Use Committee at Northwestern University (Protocol IS00028761). Healthy female 10-week-old C57Bl6/J mice were s.c. injected with 100 μL dual-tagged AF750-core/AF647-lipid PS or MC at 20 mg/mL polymer. A group was also injected with 100 μL PBS as a negative control. Each group contained five mice. Mice were intraperitoneally injected with 250 μL ketamine/xylazine cocktail at 10 mg/mL ketamine and 1 mg/mL xylazine after injection with NCs at 2, 4, and 8 h. Mice were perfused with PBS, and the heart, liver, lung, kidney, spleen, dLNs (inguinal and axillary), and ndLNs (contralateral inguinal and axillary) were harvested. An in vivo imaging system (IVIS) was used to fluorescently image organs. AF647-lipid dye was detected using λ_Ex of 640 nm and λ_Em of 680 nm. AF750-core dye was detected using λ_Ex of 745 nm and λ_Em of 800 nm. Small binning and 0.5 ms exposure were used to acquire all images. IVIS data were analyzed using Living Image software (PerkinElmer/Revvity). Spleen and dLNs were then processed into single-cell suspensions. Single-cell suspensions of LNs were prepared by digestion in collagenase D and collagenase IV for 30 min at 37°C in DMEM supplemented with calcium chloride. Splenocytes and LN cells were filtered with 70 mm cell strainers. Splenocytes were then incubated with ACK lysis buffer to remove red blood cells. Both splenocytes and LN cells were stained with the following markers: BUV396 anti-CD3 (Biolegend), Live/Dead fixable violet (Invitrogen), BV605 anti-CD11b (Biolegend), BV711 anti-CD45 (Biolegend), FITC anti-CD19 (Biolegend), PerCP-Cy5.5 anti-IA/IE (Biolegend), PE anti-F4/80 (Biolegend), and PE-Cy7 anti-CD11c (Biolegend). Samples were recorded on a BD LSR-Fortessa and analyzed using FlowJo software. A minimum of 1,000,000 single events were recorded. A representative gating strategy is shown in Figure S37.

Software packages and statistical analyses

Statistical analysis was performed using GraphPad Prism 10 (GraphPad). Data were analyzed using one-way ANOVA with Dunn’s or Tukey’s post-hoc correction for multiple hypothesis testing unless otherwise stated. All flow cytometry data were analyzed using FlowJo_v10.7.2 software (FlowJo LLC, BD Biosciences).

Supplementary Material

Supplemental information can be found online at https://doi.org/10.1016/j.celbio.2026.100419.

Highlights.

• PEG-PPS, PEG-DSPE, and PEG-DOPE anchor with nanocarriers over time

• PEG-stearate and PEG-DMG anchor unstably in nanocarriers

• PEG-lipids show divergent uptake behavior from nanocarriers

THE BIGGER PICTURE.

PEG-lipids are widely incorporated into lipids and synthetic nanocarriers to enhance systemic circulation, improve particle stability, and provide sites for attaching targeting moieties. However, these molecules can desorb or “shed” from the nanocarrier surface, leading to off-target effects that compromise safety and hinder clinical translation. Despite its significance, PEG-lipid shedding remains poorly understood across most nanocarrier systems.

In this work, we systematically investigate how lipid chemistry and nanoparticle morphology influence PEG-lipid shedding from polymeric nanoparticles. These particles exhibit high colloidal stability in absence of the PEG-lipid, enabling independent tracking of the nanocarrier and the PEG-lipid. Using fluorescently labeled polymers and PEG-lipids, we integrate quantitative microscopy with biological models to elucidate how molecular structure and particle architecture govern shedding kinetics and biodistribution. Our results reveal that strongly anchored PEG-lipids maintain surface stability across nanoparticle morphologies, whereas weakly anchored variants display shedding behavior dependent on acyl chain chemistry. Collectively, these findings establish fundamental design principles for controlling PEG-lipid stability and, ultimately, modulating nanoparticle performance in biological environments.