Biomimetic Lipopolysaccharide-Free Bacterial Outer Membrane-Functionalized Nanoparticles for Brain-Targeted Drug Delivery

Abstract

The blood-brain barrier (BBB) severely blocks the intracranial accumulation of most systemic drugs. Inspired by the contribution of the bacterial outer membrane to Escherichia coli K1 (EC-K1) binding to and invasion of BBB endothelial cells in bacterial meningitis, utilization of the BBB invasion ability of the EC-K1 outer membrane for brain-targeted drug delivery and construction of a biomimetic self-assembled nanoparticle with a surface featuring a lipopolysaccharide-free EC-K1 outer membrane are proposed. BBB penetration of biomimetic nanoparticles is demonstrated to occur through the transcellular vesicle transport pathway, which is at least partially dependent on internalization, endosomal escape, and transcytosis mediated by the interactions between outer membrane protein A and gp96 on BBB endothelial cells. This biomimetic nanoengineering strategy endows the loaded drugs with prolonged circulation, intracranial interstitial distribution, and extremely high biocompatibility. Based on the critical roles of gp96 in cancer biology, this strategy reveals enormous potential for delivering therapeutics to treat gp96-overexpressing intracranial malignancies.

Keywords: bacterial outer membrane; biomimetic; blood-brain barrier; drug delivery; nanoparticles.

Figure 1

Design of EC‐K1 LPS‐free outer membrane‐coated dOMV@NPs for systemic drug delivery to the brain. a) Schematic representation of the brain entry of EC‐K1 to induce the occurrence of bacterial neuroinflammation and the use of a detoxified outer membrane from EC‐K1 to construct the brain‐targeted dOMV@NPs. b) In the blood vessels of the brain, dOMV@NPs are proposed to penetrate the BBB through OmpA‐gp96 interaction‐mediated transcytosis and are further distributed within the intracranial interstitial space.

Figure 2

Characterization of dOMV and dOMV@NPs. a) Representative TEM images of OMV and dOMV. Scale bar, 50 nm. b,c) SDS–PAGE (b) and western blot (c) of OMV and dOMV. d) Endotoxin activity of OMV and dOMV was analyzed by LAL assay. e) Representative TEM images of OMV@NPs and dOMV@NPs. Scale bar, 50 nm. f,g) SDS–PAGE (f) and western blot (g) of OMV@NPs and dOMV@NPs. h) Endotoxin activity of OMV@NPs and dOMV@NPs was analyzed by LAL assay. i) Distribution of the hydrodynamic diameter of dOMV@NPs by dynamic light scattering. j) The mean hydrodynamic diameter of the indicated NPs. k) The zeta potential of the indicated NPs by dynamic light scattering. Error bars represent the SD (n = 3). ns (not significant), **p < 0.01, one‐tailed unpaired t‐test.

Figure 3

dOMV@NPs penetrate the BBB through the OmpA‐gp96 interaction‐mediated transcellular vesicle transport pathway. a) Levels of claudin‐5 on bEND.3 BBB endothelial cells after stimulation with the indicated bacterial components (20 µg mL⁻¹ protein or 2.24 ng mL⁻¹ LPS for 24 h) were analyzed by western blot. b) Paracellular penetration of 40 kDa FITC‐dextran (1 mg mL⁻¹, 6 h incubation) across the in vitro BBB (which was prestimulated with the indicated bacterial components at 20 µg mL⁻¹ protein or 2.24 ng mL⁻¹ LPS for 12 h). c) Claudin‐5 expression on bEND.3 cells after stimulation with the indicated NPs (20 µg ml⁻¹ protein for 24 h) was analyzed by western blot. d) Cellular uptake of the indicated DiR‐labeled NPs (1.5 µg DiR mL⁻¹ for 6 h) by bEND.3 cells was measured by flow cytometry. e) gp96 on RAW264.7 macrophages and bEND.3 cells stimulated with either OmpA or dOMV was analyzed by western blot. f–h) Cellular uptake of doxorubicin‐labeled dOMV@NPs (5 µg doxorubicin mL⁻¹ for 1.5 h) by bEND.3 cells in the presence of anti‐OmpA antibody (f), anti‐gp96 antibody (g), or the indicated endocytosis inhibitors (h) was measured by flow cytometry. i) Intracellular localization of dOMV@NPs (12 µg doxorubicin mL⁻¹ for 1 h) after further incubation with fresh medium for different times, imaged using a confocal laser scanning microscope. Scale bar, 5 µm. j) Penetration of the indicated doxorubicin‐labeled NPs (5 µg doxorubicin mL⁻¹ for 6 h) across the in vitro BBB. Error bars represent the SD (n = 3). ns (not significant), *p < 0.05, **p < 0.01, ***p < 0.001, one‐tailed unpaired t‐test.

Figure 4

Pharmacokinetics and brain accumulation in normal mice. a) Cellular uptake of the indicated DiR‐labeled NPs (1.5 µg DiR mL⁻¹ for 3 h) by RAW264.7 macrophages was measured by flow cytometry. b,c) Blood concentration‐time profiles (b) and pharmacokinetic parameters (c) in normal mice (n = 5). d–f) Brain accumulation of IR780 in normal mice at 8 and 24 h after injection of the indicated IR780‐labeled NPs was qualitatively imaged (d) and semiquantitatively analyzed (e,f) by measuring the fluorescence of IR780. g) Quantitative measurement of intracranial Fe concentration by ICP–MS at 8 h after intravenous injection of NPs loaded with SPIO (n = 5). h) The brains were further excised, and IR780 was extracted to quantitatively investigate the exact % ID/g brain (n = 5). Error bars represent the SD (n = 3, unless otherwise specified). ns (not significant), **p < 0.01, ***p < 0.001, one‐tailed unpaired t‐test.

Figure 5

Intracranial microdistribution and immunohistochemical localization. a) Microdistribution of doxorubicin (red) in the indicated intracranial regions at 12 h after the second intravenous injection of the indicated doxorubicin‐labeled NPs into normal mice. Nuclei (blue) were stained with DAPI. Yellow scale bars and white scale bars (in insets) indicate 100 and 50 µm, respectively. b) Immunohistochemical analysis of cerebral sections for the indicated intracranial regions was performed to check the colocalization of dOMV@NPs with various types of intracranial cells. Cerebral sections obtained at 12 h after the second intravenous injection of doxorubicin‐labeled dOMV@NPs were stained with anti‐CD31, anti‐GFAP, anti‐Iba1, and anti‐NeuN antibodies (green) to label brain endothelial cells, astrocytes, microglia, and neurons, respectively. Nuclei (blue) were stained with DAPI. Yellow scale bars and white scale bars (in insets) indicate 20 and 10 µm, respectively.

Figure 6

Biosafety assessment of dOMV@NPs. a) mRNA levels of TNF‐α, IL‐6, and IL‐1β in brains were quantitatively measured 24 h after intravenous injection of the indicated NPs into normal mice. Mice treated with saline were used as the control (n = 5). b,c) Expression levels of TNF‐α, IL‐6, and IL‐1β in the brain (b) and serum (c) were measured by ELISA 24 h after intravenous injection of the indicated NPs into normal mice. d,e) Blood markers of liver function (d) and kidney function (e) were quantitatively measured 24 h after intravenous injection of the indicated NPs into normal mice (n = 5). Error bars represent the SD (n = 3, unless otherwise specified). ns (not significant), *p < 0.05, **p < 0.01, ***p < 0.001, one‐tailed unpaired t‐test.