Polarized Secretion of Extracellular Vesicles by Mammary Epithelia

Abstract

Extracellular vesicles (EVs) are secreted by many cell types and are increasingly investigated for their role in human diseases including cancer. Here we focus on the secretion and potential physiological function of non-pathological EVs secreted by polarized normal mammary epithelial cells. Using a transwell system to allow formation of epithelial polarity and EV collection from the apical versus basolateral compartments, we found that impaired secretion of EVs by knockdown of RAB27A or RAB27B suppressed the establishment of mammary epithelial polarity, and that addition of apical but not basolateral EVs suppressed epithelial polarity in a dose-dependent manner. This suggests that apical EV secretion contributes to epithelial polarity, and a possible mechanism is through removal of certain intracellular molecules. In contrast, basolateral but not apical EVs promoted migration of mammary epithelial cells in a motility assay. The protein contents of apical and basolateral EVs from MCF10A and primary human mammary epithelial cells were determined by mass spectrometry proteomic analysis, identifying apical-EV-enriched and basolateral-EV-enriched proteins that may contribute to different physiological functions. Most of these proteins differentially secreted by normal mammary epithelial cells through polarized EV release no longer showed polarized secretion in MCF10A-derived transformed epithelial cells. Our results suggest an essential role of EV secretion in normal mammary epithelial polarization and distinct protein contents and functions in apical versus basolateral EVs secreted by polarized mammary epithelia.

Keywords: Epithelial polarization; Extracellular vesicles; Mammary gland; Proteomics.

Fig. 1. Polarization of mammary epithelial cells

(a) Transwell system for the polarization of mammary epithelial cells and collection of apical and basolateral EVs. (b) Immunofluorescence assay of polarized MCF10A cells showing polarized expression of Podocalyxin (an apical marker) and Syntaxin-4 (a basolateral marker). DAPI staining indicates the nuclei. Z-stack images are reconstructed in 3D and the XZ-view is shown. Green and red arrows indicate apical staining of Podocalyxin and basolateral staining of Syntaxin-4, respectively. (c) Time course for diffusion of phenol red from the upper chamber to the lower chamber of MCF10A transwell culture. (d) Time course of trans-epithelial electrical resistance (TEER) from MCF10A cells grown on 10-cm transwell filters.

Fig. 2. Characterization of EVs

(a) EVs pelleted at 110,000 ×g were analyzed by nanoparticle tracking analysis. The black line indicates mean, whereas the red shaded area indicates SEM. (b) A schematic representation of the gradient separation procedure to further characterize the 110,000 ×g pellets. (c) Western blot and density measurement of fractions collected from gradient separation or unfractionated EV.

Fig. 3. EV secretion is required for the establishment of epithelial polarity

(a) Western blot showing indicated protein levels in MCF10A control and RAB27 KD cells. (b) Numbers of EVs secreted by equal number of indicated cells were determined by nanoparticle tracking analysis. *** P < 0.001 compared to the MCF10A control cells. (c) MCF10A (control or RAB27 KD), MCFDCIS, and HuMEC cells were seeded on 24-well transwell inserts and TEER was measured over the indicated time course. *** P < 0.001 compared to the MCF10A control cells. (d) Left: Phase contrast images of indicated cells grown on tissue culture dishes. Right: The XZ-view of reconstructed Z-stack images showing the immunofluorescence staining of Podocalyxin (green) and Syntaxin-4 (red) in cells grown on filters. DAPI staining (blue) indicates the nuclei. Green and red arrows indicate apical staining of Podocalyxin and basolateral staining of Syntaxin-4, respectively.

Fig. 4. Different effects of apical and basolateral EVs on cell polarity and migration

(a) MCF10A cells grown on filters were incubated with CFSE-labelled apical or basolateral MCF10A EVs (green) for 24 h and fixed before fluorescent and phase contrast images were captured. DAPI staining (blue) indicates the nuclei. (b) MCF10A (control or RAB27B KD) cells seeded on transwell filters were treated with apical or basolateral MCF10A EVs isolated from equal number (1:1) or three folds (3:1) of polarized MCF10A cells on day 1 and then on day 3. EVs were added to the upper chambers of transwell inserts. TEER was measured on day 3, 5, and 7. *** P < 0.001 compared to the PBS group. (c) MCF10A cells were mixed with apical or basolateral MCF10A EVs isolated from equal number (1:1) or three folds (3:1) of polarized MCF10A cells, and added to the upper chamber in transwell migration assays. After 24 h, cells that had migrated to the underside of transwell filters were counted. * P < 0.05, *** P < 0.001 compared to the PBS group.

Fig. 5. Identification of proteins in apical and basolateral EVs by mass spectrometry

(a) Venn diagrams showing the numbers of proteins with at least 4 peptide matches in the apical and basolateral EVs from the MCF10A, MCFDCIS (which shows impaired polarity), and HuMEC models. Numbers of proteins that are uniquely detected from one sample (top), or also including whose enriched by at least 2 folds (bottom) are shown. (b, c) EV proteins identified by mass spectrometry with at least 4 peptide matches and enriched by at least 2 folds in apical (b) or basolateral (c) EVs from both MCF10A and HuMEC models were analyzed by Ingenuity Pathway Analysis for the prediction of associated diseases and functions.