Surface chemistry-mediated modulation of adsorbed albumin folding state specifies nanocarrier clearance by distinct macrophage subsets

Abstract

Controlling nanocarrier interactions with the immune system requires a thorough understanding of the surface properties that modulate protein adsorption in biological fluids, since the resulting protein corona redefines cellular interactions with nanocarrier surfaces. Albumin is initially one of the dominant proteins to adsorb to nanocarrier surfaces, a process that is considered benign or beneficial by minimizing opsonization or inflammation. Here, we demonstrate the surface chemistry of a model nanocarrier can be engineered to stabilize or denature the three-dimensional conformation of adsorbed albumin, which respectively promotes evasion or non-specific clearance in vivo. Interestingly, certain common chemistries that have long been considered to convey stealth properties denature albumin to promote nanocarrier recognition by macrophage class A1 scavenger receptors, providing a means for their eventual removal from systemic circulation. We establish that the surface chemistry of nanocarriers can be specified to modulate adsorbed albumin structure and thereby tune clearance by macrophage scavenger receptors.

Fig. 1. PEG-b-PPS polymersome morphology and hydrophilic loading characterization.

a Illustration depicting the poly(ethylene glycol)-b-poly(propylene sulfide) (PEG-b-PPS) polymersomes (PS) of three different surface chemistries. b Cryo-TEM of Phos PS, OH PS, and MeO PS. TEM micrographs were acquired at ×10,000 magnification (scale = 100 nm). c Illustration of PS characterization by small angle x-ray scattering (SAXS). A schematic of the spherical vesicle model is shown (core radius, r_c; total radius, r_t; scattering length density of the solvent and shell, ρ_solv and ρ_shell, respectively). d–f SAXS scattering profiles and fit models for blank d Phos PS, e OH PS, and f MeO PS nanocarriers. SAXS was performed using synchrotron radiation. In each case, the scattering profile (gray dots) was fit using a spherical vesicle model. The model fit is represented as a solid line. The intensity (arbitrary units; a.u.) is plotted with the scattering vector (q). The chi square (X²) for each model fit is displayed. A X² value of <1.0 indicates a reasonable model fit. g Zeta potential (mV) of Phos PS, OH PS, and MeO PS. The mean ± s.d. (n = 3) is displayed. h, i, Loading of the high molecular weight hydrophilic cargo. h The concentration of encapsulated 70 kDa Dextran-TMR (Dex-TMR) after purifying PS with size exclusion chromatography. i Dex-TMR encapsulation efficiency. For h, i, error bars represent s.e.m. from three parallel experiments (n = 3). In all cases, significance was determined by ANOVA with post hoc Tukey’s multiple comparisons test (5% significance level). *p < 0.05, ***p < 0.001, ****p < 0.0001.

Fig. 2. Diverse murine macrophage populations are sensitive to polymersome surface chemistry in vivo.

a Overview of nanocarrier biodistribution study in four treatment groups of C57BL/6 J mice (n = 5) receiving intravenously administered PBS or near infrared dye, DiR-loaded PS (DiR PS). After 4 h, organ and cellular biodistribution was assessed by IVIS and multicolor flow cytometry, respectively. b Percent change in DiR PS fluorescence (λ_Ex = 750 nm; λ_Em = 780 nm) between serum (4 h) and a formulation-specific t₀ proxy. The t₀ proxy is the input DiR PS formulation diluted in untreated mouse plasma in a 1:15 ratio matching the ratio of DiR PS (100 μL) administered to the ~1.5 mL blood volume per mouse. c–h Organ biodistribution. c Representative IVIS images of organs. The radiant efficiency is displayed. d Distribution of total adjusted radiant efficiency within each treatment group, representing the percentage of the total adjusted radiant efficiency accounted for by the spleen, liver, kidneys, and lungs. e–h Adjusted radiant efficiency comparison between treatment groups in the e spleen, f liver, g kidneys, and h lungs. i–m Cellular biodistribution. The percentage of DiR PS+ cells of the specified type is displayed for the i spleen, j liver, k lymph nodes (LNs), l lungs, and m kidneys. Statistical significance was determined by ANOVA with post hoc Tukey’s multiple comparisons test (5% significance level; *p < 0.05, **p < 0.01, ***p < 0.0005, ****p < 0.0001). Error bars represent s.e.m. n Percentage of the 31 statistically significant comparisons (annotated in i–m) accounted for by each cell type.

Fig. 3. Adsorbed albumin increases macrophage uptake of all three PS surface chemistries but only significantly changes the zeta potential for Phos PS.

a PS–BSA complexes formed after polymersome (PS) incubation with bovine serum albumin (BSA; 0.1, 1.0, or 10.0 mg/mL). b Pfam domain annotations and Columbic electrostatic surface mapped to the BSA crystal structure (PDB: 4F5S). c Albumin adsorption concentration dependence (2.5 mg/mL PS; x-axis: input BSA concentration). Significance was determined by ANOVA with post hoc Tukey. d Zeta potential (mV) of PS and PS–BSA complexes. The mean ± s.d. is displayed from three parallel experiments (n = 3). e PS zeta potential as a function of adsorbed albumin. Non-linear fits (dashed lines) for one-phase association (Phos, MeO) and exponential decay (OH) models. f–h Flow cytometric analysis of PS–BSA uptake by RAW 264.7 macrophages in serum-free media (2 h, 37 °C). PS encapsulated Dex_70 kDa-TMR to quantify vesicle uptake. f Illustration. g %PS+ cells (1.1% false positive rate). h Median fluorescence intensity (MFI). Error bars represent s.e.m. from three biological replicates (n = 3) unless indicated otherwise. Significant differences (PS versus PS–BSA) were determined by ANOVA with post hoc Dunnett’s test. Statistical tests used a 5% significance level. *p < 0.05, **p < 0.005, ***p < 0.0005, ****p < 0.0001.

Fig. 4. Polymersome surface chemistry modulates the conformation of adsorbed albumin.

a Tryptophan-quenching assay. BSA (PDB: 4F5S) W134 and W213 residues (orange). b BSA fluorescence (λ_Ex = 280 nm) quenching after chaotropic guanidine hydrochloride (Gdn-HCl) denaturation. The AUC obtained by integrating emission spectra (inset) was normalized to folded BSA (third-order polynomial model (r² = 0.90), 95% confidence interval displayed). c BSA Trp quenching after adsorption to PS (Phos, OH, MeO); AUC normalized to folded BSA. G_5M = Gdn-HCl-denatured BSA control. The mean ± s.e.m. is displayed from three parallel experiments (n = 3). Significant differences (versus folded BSA) were determined by ANOVA with post hoc Dunnett’s test. d, e Limited proteolysis assessment of BSA cleavage susceptibility. d Trypsin proteolysis kinetics. e Full-length BSA cleavage with time monitored by MALDI-TOF MS (control shown). f Free BSA and PS-adsorbed BSA proteolysis kinetics (t₀-normalized). One phase decay models, y = (y₀ − y_p) × e^−kx, were fit (x = time (min); y₀ = AUC_t0; y_p = AUC plateau; k = proteolysis rate (min⁻¹)). Model r² values: 0.97 (BSA_control), 0.87 (Phos PS), 0.99 (OH PS), and 0.99 (MeO PS). g–I Circular dichroism (CD) spectroscopy of BSA, heat-denatured BSA control (85 °C, 10 min), and PS-adsorbed BSA. The g mean residue ellipticity (MRE) (10³deg cm²/dmol), h difference spectra w.r.t. folded BSA (PS–BSA_MRE − BSA_MRE), and i α-helicity percentage. The mean ± s.e.m. from three parallel experiments is displayed (n = 3). α-Helicity significant differences (versus folded BSA) were determined by ANOVA with Dunnett’s test. j Protein structure assay correlations (Pearson correlation coefficient (r), significance inset). Fit linear regression models are displayed with 95% confidence intervals. k R_g-normalized Kratke analysis of BSA SAXS profiles after adsorption. Error bars represent s.e.m. (n = 3). All statistical tests used a 5% significance level. **p < 0.001; ***p < 0.005; ****p < 0.0001.

Fig. 5. SR-A1 recognizes polymersomes with surface chemistry that denatures albumin and contributes to their non-exclusive uptake by macrophages in vitro.

a RAW 264.7 macrophages pre-treated with either Fucoidan (+F) SR-A1 competitor, or PBS, for 30 min at 37 °C. Afterwards, pristine PS (Phos, OH, MeO) were administered for 2 h at 37 °C prior to flow cytometry. b Fluorescence microscopy of macrophage SR-A1/CD204 expression. Brightfield, DAPI, FITC (anti-CD204-FITC) channels are shown. Scale bar = 50 μm. c Fucoidan (2.5 mg/mL) pre-treatment does not significantly alter surface SR-A1/CD204 expression after 4 h (MFI: median fluorescence intensity; anti-FITC-CD204 fluorescence). Significance was determined by t test (unpaired, two-tailed). d Fucoidan pre-treatment (2.5 mg/mL) increases pristine PS uptake in vitro. e, f Differential recognition of PS–BSA complexes by macrophage SR-A1 in vitro. e SR-A1 hypothesis illustration. f Macrophage uptake of PS–BSA after fucoidan pre-treatment. g The percent change in MFI (±fucoidan). Significance was determined by ANOVA using post hoc Tukey’s multiple comparisons test. h SR-A1 recognition mechanism summary. Serum-free conditions were used for all uptake studies. PS encapsulated hydrophilic Dex_70 kDa-TMR to quantify vesicle uptake. PS–BSA complexes were formed by incubating PS ([polymer] = 2.5 mg/mL) with 10 mg/mL BSA for 2 h at 37 °C. For c, d, f, g, the mean ± s.e.m. is displayed from three biological replicates (n = 3). In d, f, significant differences between ±fucoidan pre-treatment pairs was determined by ANOVA with Sidak’s test. All statistical tests used a 5% significance level. ns = not significant, **p < 0.01, ***p < 0.0005, ****p < 0.0001. Error bars represent s.e.m. (n = 3).