Impact of the Endosomal Escape Activity of Cell-Penetrating Peptides on the Endocytic Pathway

Abstract

Cell-penetrating peptides (CPPs) are routinely used for the delivery of macromolecules into live human cells. To enter the cytosolic space of cells, CPPs typically permeabilize the membrane of endosomes. In turn, several approaches have been developed to increase the endosomal membrane permeation activity of CPPs so as to improve delivery efficiencies. The endocytic pathway is, however, important in maintaining cellular homeostasis, and understanding how endosomal permeation impacts cells is now critical to define the general utility of CPPs. Herein, we investigate how CPP-based delivery protocols affect the endocytic network. We detect that, in some cases, cell penetration induces the activation of Chmp1b, Galectin-3, and TFEB, which are components of endosomal repair, organelle clearance, and biogenesis pathways, respectively. We also detect that cellular delivery modulates endocytosis and endocytic proteolysis. Remarkably, a multimeric analogue of the prototypical CPP TAT permeabilizes endosomes efficiently without inducing membrane damage responses. These results challenge the notion that reagents that make endosomes leaky are generally toxic. Instead, our data indicates that it is possible to enter cells with minimal deleterious effects.

Figure 1.

CPP-based cytosolic delivery approaches and endosome-disrupting reagents used in this study.

Figure 2.

Reagent-mediated leakage of fluorescent probes trapped in endosomes. SNAP-Surface 488, Fl-k5 and AF488-H1 were incubated with delivery reagents for 1h and their cytosolic egress was quantified by counting the number of cells displaying nuclear or nucleolar staining by the probes. For all experiments, the data provided are the mean of triplicate experiments (approximately 500 cells quantified per condition for each experiment).

Figure 3.

Triggering of endosomal membrane damage responses by CPPs. A) Scheme highlighting the activity of probes Chmp1b, Gal3, and TFEB as reporters of membrane repair, autophagy, and lysosomal biogenesis, respectively. B) Cells were incubated with delivery reagents, washed and imaged. Representative fluorescence of images treated and untreated cells are shown (scale bars: x100: 10 μm). Images are pseudocolored green for Chmp1b, Gal3 and TFEB, and magenta for Hoechst. To measure the recruitment of Chmp1b and Gal3 to damaged endosomes, the number of fluorescent puncta per cell, as well as the sum of pixels from fluorescent puncta were quantified. Colocalization with the late endosome/lysosome marker Lamp1 was used to establish that the puncta correspond to endocytic organelles. To assess whether lysosomal biogenesis is induced, the percentage of cells displaying TFEB nuclear accumulation was determined. The data shown represent means determined from biological duplicates. (C) Colocalization analysis of Gal3 and Chmp1b upon treatment with LLOME or dfTAT/UNC7938. Pearson’s colocalization coefficient (R) and Manders’ overlap coefficient (MOC) were determined for each condition. Fluorescence images are pseudocolored red for Gal3 and green for Chmp1b.

Figure 4.

Impact of CPPs on endocytic uptake and endosomal degradation post incubation. A) Cellular uptake of endocytic probes after treatment with delivery agents. Cells were treated with CPPs for their respective incubation times, washed, and incubated with AF488-labelled probes for endocytic uptake, namely transferrin (T), cholera toxin subunit B (CT), and Histone H1 (H1). The uptake of each probe was quantified by flow cytometry. Four representative flow cytometry traces of the uptake of AF488-H1 are shown. The AF488 intensities that corresponded to a 50% threshold (dotted lines) was determined for each condition. The fold change in uptake for each condition as compared to the vehicle/probe control is shown. The values reported represent the mean of biological duplicates. The standard deviations (not shown) are 0.2 or less. (B) Impact of dfTAT on endosomal proteolysis. The degradation of the probe AF488-H1 endocytosed by cells was monitored over time by quantifying the formation of fluorescent protein fragments by SDS-PAGE. The degradation rate of AF488-H1 endocytosed was compared for cells preincubated for 1h with either vehicle or dfTAT. Each plotted value represents the mean fluorescence intensity of intact or degraded AF488-H1 from biological duplicates. Fluorescence microscopy was used to compare the localization of H1 incubated with cells during or after incubation with dfTAT. Images are pseudocolored green for AF488-H1 and magenta for Hoechst (scale 10 μm). (C) Transmission electron microscopy imaging of cells treated with dfTAT or dfTAT/UNC7938. The number of multivesicular bodies/late endosomes, endolysosomes, and lysosomes were counted from the cross sections of 5 cells. (D) Representative TEM images of cells treated with vehicle (left) and dfTAT/UNC7938-treated (right). Organelles are designated as multivesicular bodies/late endosomes (MV), endolysosomes (EL), lysosomes (L) based on their morphologies.

Figure 5.

Recovery kinetics of membrane damage probes. (A) and (B) The number Gal3 or Chmp1b positive puncta per cell was quantified over time after LLOME or dfTAT/UNC7938 treatment. The data are represented as box and whiskers plots. (C) Percentage of cells displaying a nuclear localization of TFEB post LLOME or dfTAT/UNC7938 treatment. (D) Cellular recovery of AF488-H1 uptake after treatment with dfTAT for 1h. (E) Recovery of LysoTracker Green staining in cells were treated with LLOME or dfTAT. For all experiments, the data are obtained from biological triplicates.

Figure 6.

Schematic representation of the membrane responses induced by CPPs in this study. (A) Endosomal leakage and toxicity associated with each condition are shown to highlight the lack or correlation between cell penetration by endosomal escape and endosomal membrane damage. Extent of membrane damage is indicated by arrows showing the progression of cellular responses in order of expected severity. B) LLOME and UNC7938 generate membrane disruptions that trigger recruitment of Gal3 and components of the ESCRT-III complex, but that fail to promote the endosomal escape of the SNAP-S488, DEAC-k5, and AF488-H1 probes. In contrast, dfTAT releases these probes into the cytosol of cells effectively without inducing Gal3/ESCRT-III membrane repair pathways.