Engineered Exosomes as Vehicles for Biologically Active Proteins

Abstract

Exosomes represent an attractive vehicle for the delivery of biomolecules. However, mechanisms for loading functional molecules into exosomes are relatively unexplored. Here we report the use of the evolutionarily conserved late-domain (L-domain) pathway as a mechanism for loading exogenous proteins into exosomes. We demonstrate that labeling of a target protein, Cre recombinase, with a WW tag leads to recognition by the L-domain-containing protein Ndfip1, resulting in ubiquitination and loading into exosomes. Our results show that Ndfip1 expression acts as a molecular switch for exosomal packaging of WW-Cre that can be suppressed using the exosome inhibitor GW4869. When taken up by floxed reporter cells, exosomes containing WW-Cre were capable of inducing DNA recombination, indicating functional delivery of the protein to recipient cells. Engineered exosomes were administered to the brain of transgenic reporter mice using the nasal route to test for intracellular protein delivery in vivo. This resulted in the transport of engineered exosomes predominantly to recipient neurons in a number of brain regions, including the olfactory bulb, cortex, striatum, hippocampus, and cerebellum. The ability to engineer exosomes to deliver biologically active proteins across the blood-brain barrier represents an important step for the development of therapeutics to treat brain diseases.

Keywords: ESCRT; L-domain; blood-brain barrier; drug delivery; extracellular vesicles; intraluminal vesicles; nasal delivery; therapy; ubiquitin.

Graphical abstract

Figure 1

The Packaging of Targeted Fusion Proteins into Exosomes (A) Flow cytometric analysis of transfected mT/mG MEFs showed that the WW-tagged Cre protein (WW-Cre) maintained Cre-recombinase activity compared to wild-type Cre protein. Unpaired t test. (B) Cre protein is not exported in exosomes with or without Ndfip1 expression. WW-Cre protein is efficiently exported into exosomes in the presence of Ndfip1 in LN18 cells. The asterisk indicates the WW-Cre band above the Tsg101 band after reprobing the western blot. (C) NanoSight tracking analysis was used to measure the particle size of the purified exosomes. (D) GW4869 inhibited the release of WW-Cre protein into exosomes. Data points are the average ± SEM of three independent experiments.

Figure 2

Exosomes Containing Target WW-Cre Proteins Are Functionally Active in Recipient Reporter Cells (A–D) Exosomes purified from LN18 cells transfected with Cre only (A), Cre and Ndfip1 (B), WW-Cre only (C), or WW-Cre and Ndfip1 (D) were delivered to mT/mG MEFs. (E) Quantification of the number of positive reporter mT/mG MEFs after exosome delivery. Scale bar, 50 μm. Data points are the average ± SEM of three independent experiments. One-way ANOVA tests with Bonferroni post-tests, ****p < 0.0001.

Figure 3

Ndfip1 Expression Is Required for Functional WW-Cre Exosomes and Mediates the Ubiquitination of WW-Cre (A) Immunoprecipitation assay of overexpressed WW-Cre and Ndfip1 in HEK293T cells showed an interaction between the two proteins. (B) Ndfip1^+/+ and Ndfip1^−/− MEFs used in the co-culture assay in (C) expressed similar levels of WW-Cre, as shown by western blotting. (C) Ndfip1^+/+ and Ndfip1^−/− MEFs infected with lentiviral construct of WW-Cre were co-cultured with mT/mG reporter MEFs. Only Ndfip1^+/+ MEFs resulted in a positive recombination of mT/mG MEFs. (D) Denaturing ubiquitin assays show that Cre protein is not ubiquitinated with or without Ndfip1. In contrast, WW-Cre protein is predominantly monoubiquitinated in the presence of Ndfip1. Scale bar, 50 μm.

Figure 4

The Stability and Function of Stored Exosomes (A) Proteinase treatment of exosomes did not result in the degradation of WW-Cre unless Triton X-100 was added to permeabilize the exosome membrane. (B) Exosomes were stored for 3 days under different conditions before western blotting was performed to determine the stability of WW-Cre protein compared to a control fresh preparation of WW-Cre exosomes. (C) Stored exosomes were analyzed for the ability to deliver active WW-Cre to mT/mG reporter MEFs. Data points are the average ± SEM of three independent experiments. One-way ANOVA tests with Bonferroni post-tests, *p < 0.05.

Figure 5

In Vivo Activity of Exogenous Exosomes (A) Quantification of positive reporter cells after nasal delivery of exosomes derived from LN18 cells transfected with WW-Cre and Ndfip1, Cre and Ndfip1, or untransfected to Ai14 mice. Mouse brains were collected 14 days after last nasal delivery, and sections from multiple brain regions were analyzed for red reporter cells indicating Cre-recombination (n = 4 mice per condition). (B) Overview of the coronal brain section analyzed for Cre recombination, with the white box indicating the area imaged in (C). (C) Higher magnification of brain regions showing Cre recombination after delivery of WW-Cre containing exosomes. (D) Immunocytochemistry with anti-NeuN antibody shows colocalization in cells with Cre-recombinase activity (tdTomato) in Ai14 mice after WW-Cre exosome delivery. (E) Immunocytochemistry with anti-Iba-1 antibody shows colocalization with Cre-recombinase activity (tdTomato) in Ai14 mice after WW-Cre exosome delivery. One-way ANOVA tests with Bonferroni post-tests, *p < 0.05. Scale bars, (B) 2.5 mm, (C) 250 μm, and (D and E) 50 μm.