miR-1227 Targets SEC23A to Regulate the Shedding of Large Extracellular Vesicles

Abstract

Cancer cells shed a heterogenous mixture of extracellular vesicles (EVs), differing in both size and composition, which likely influence physiological processes in different manners. However, how cells differentially control the shedding of these EV populations is poorly understood. Here, we show that miR-1227, which is enriched in prostate cancer EVs, compared to the cell of origin, but not in EVs derived from prostate benign epithelial cells, induces the shedding of large EVs (such as large oncosomes), while inhibiting the shedding of small EVs (such as exosomes). RNA sequencing from cells stably expressing miR-1227, a modified RISCTRAP assay that stabilizes and purifies mRNA-miR-1227 complexes for RNA sequencing, and in silico target prediction tools were used to identify miR-1227 targets that may mediate this alteration in EV shedding. The COPII vesicle protein SEC23A emerged and was validated by qPCR, WBlot, and luciferase assays as a direct target of miR-1227. The inhibition of SEC23A was sufficient to induce the shedding of large EVs. These results identify a novel mechanism of EV shedding, by which the inhibition of SEC23A by miR-1227 induces a shift in EV shedding, favoring the shedding of large EV over small EV.

Keywords: extracellular vesicle biogenesis; extracellular vesicles; large oncosomes.

Figure 1

MiR-1227 induces shedding of L-EV and inhibits the shedding of S-EV in PC cells. (A) MiR-1227 (teal) and miR-21 (grey) expression in different biological fluids from Weber et al. 2010 [23]. (B) Taqman qPCR quantification of miR-1227 in EV isolated from pooled normal or PC patient plasma (n = 1). (C) Transwell migration of RWPE-2 PC and U87 glioma cells stably expressing miR-1227 (teal) or empty vector control (grey) (n = 3) normalized to corresponding vector control; (D) qPCR for miR-1227 from S-EV (blue) and L-EV (red) isolated from PC3 cells stably expressing miR-1227(n = 1). (E,F) Digital droplet qPCR quantification of miR-1227 per S-EV (E) and L-EV (F) (n = 3). (G,H) Atomic force microscopy quantifications of S-EV (G) and L-EV (H) shed from PC3 cells stably expressing miR-1227 (n = 3). (I–L) Quantification of S-EV (I) and L-EV (J) shed from PC3 cells stably expressing miR-1227 by TRPS with size distribution plots in (K,L), respectively (n = 3); * p < 0.05.

Figure 2

Identification of miR-1227 targets. (A) Summary of the bioinformatics strategy used to identify miR-1227 targets. (B) Waterfall plot of putative miR-1227 targets, identified by stable cell RNA sequencing in RWPE-2 cells stably expressing miR-1227. Prostate cancer-related genes downregulated by miR-1227 are highlighted. (n = 1) (C) Schematic of the modified RISCTRAP assay. In standard RNA sequencing, some mRNA targets are rapidly degraded by the miRNA and are, thus, not detectable by RNA sequencing. The RISCTRAP assay stabilizes mRNA-miRNA complexes with a FLAG-tagged dominant negative GW182 protein, trapping the mRNA-miRNA complexes in an inactive state. The cells are transfected with a miRNA of interest, and all mRNA-miRNA complexes are pulled down using FLAG for RNA sequencing. We modified this assay by transfecting the cells with a biotinylated miR-1227, which is used to specifically isolate the mRNA-miR-1227 complexes. (D) Waterfall plot of putative miR-1227 targets, identified by the modified RISCTRAP RNA sequencing. Select high confidence genes with an R_RC over 1.67 that are altered in prostate cancer are highlighted. (n = 1) (E) Overlap of the genes identified in the stable cell and RISCTRAP RNA sequencing experiments. Shared genes identified in both miR-1227 stable cell and RISCTRAP sequencing from cancer cells are in blue. (F) Gene ontology of the shared miR-1227 target genes.

Figure 3

SEC23A is a direct target of miR-1227; (A) qPCR for high confidence miR-1227 target genes involved in vesicle formation from PC3 cells stably expressing miR-1227 (n = 3). (B) Western blot validation of SEC23A reduced expression in PC3, RWPE-2, and U87 cells stably expressing miR-1227. (C) Prostate cancer transcriptome atlas analysis of SEC23A in prostate cancer, showing reduced expression levels of SEC23A in cancer versus benign prostate tissue and in metastatic castration resistant versus primary prostate cancer. (D–F) Luciferase assays from RWPE-2 cells transfected with miR-1227 or control miRNA, plus the SEC23A 3′UTR luciferase construct (D) or the SEC23A 3′UTR luciferase construct with the deletion of 2 different predicted miR-1227 binding sites (E,F) (n = 3); * p < 0.05; ** p < 0.005.

Figure 4

Inhibition of SEC23A induces L-EV shedding. (A) Quantification of L-EV shedding by TRPS in PC3 cells transiently transfected with SEC23A shRNA (n = 3). (B) Quantification of L-EV shedding by TRPS after transient transfection with increasing doses of SEC23A shRNA (n = 1). (C) Expression of SEC23A in PC3 cells stably expressing SEC23A shRNA. (D–I) Quantification of L-EV (D,E) and S-EV (G,H) shedding by TRPS in PC3 (D,G) and RWPE-2 (E,H) cells stably expressing SEC23A shRNA (n = 2). (H,I) Size distribution of L-EV (F) and S-EV (I) from PC3 cells stably expressing SEC23A shRNA. (J) Western blot for EV markers from EV isolated from PC3 cells stably expressing SEC23A shRNA; * p < 0.05.