The Ceramide-Dependent EV Secretome Differentially Affects Prostate Cancer Cell Migration

Abstract

Tumor-derived extracellular vesicles (EVs) play an important role in cancer progression. Neutral sphingomyelinases (nSMases) are lipid-modifying enzymes that modulate the secretion of EVs from cells. How nSMase activity and therefore ceramide generation affect the composition and functionality of secreted EVs is not fully understood. Here, we aimed to investigate the expression of nSMases 1 and 2 in prostate cancer (PCa) tissue and their role in EV composition and secretion for prostate cancer cell migration. Reduced nSMase 1 and 2 expression was found in prostate cancer and correlated with the age of the patient. When nSMase 2 was inhibited by GW4869 in PCa cells (PC3 and DU145), the EV secretome was significantly altered, while the number of EVs and the total protein content of released EVs were not significantly changed. Using proteomic analysis, we found that extracellular matrix proteins, such as SDC4 (Syndecan-4) and SRPX-2, were differentially secreted on EVs from GW4869-treated PC3 cells. In scratch wound migration assays, GW4869 significantly increased migration compared to control PC3 cells but not DU145 cells, while SDC4 knockdown significantly reduced the migration of PC3 cells. These and other nSMase-2-dependent secreted proteins are interesting candidates for understanding the role of stress-induced EVs in the progression of prostate cancer.

Keywords: SMPD2; SMPD3; cell migration; extracellular matrix; neutral sphingomyelinases.

Figure 1

SMPD2 and SMPD3 expression in prostate cancer. (A) Paraffin-embedded tissue arrays (52 duplicate cores) were stained for SMPD2 and SMPD3 (mouse antibodies) in healthy prostate tissue (n = 7) and (B) prostatic adenocarcinoma tissue (n = 45), scale bars = 10 µm; (C) androgen receptor, (D) SMPD2, and (E,F) SMPD3 stratified by age (<70 and ≥70 years). Median SMPD3, SMPD2, and androgen receptor levels significantly vary between age groups above and below 70 years; Mann–Whitney, statistical significance * p < 0.05, ** p < 0.01, *** p < 0.001 as indicated. (G) The corresponding means, 95% confidence intervals, and effect sizes from (C–F).

Figure 2

SMPD2 and SMPD3 expression in prostate cancer tissue and cell lines. (A) Data on SMPD2 and (B) SMPD3 expression in patient samples of prostate tissue (n = 245) and prostate cancer (n = 494) in fragments per kilobase per million mapped fragments from proteinatlas.org. Unpaired t-test with Welch’s correction, p < 0.0001. (C) Data on SMPD2 and SMPD3 expression in prostate cancer cell lines from proteinatlas.org. (D) Protein levels of SMDP2 and (E) SMPD3 in PC3 and DU145 cells; GAPDH was used as a loading control. (F–I) Scratch-wound-healing assays in (F) PC3 cells or (H) DU145 cells treated with DMSO or 5 µM GW4869. The relative wound density was measured in percent. (G) Single data points from independent scratch wounds in PC3 or (I) DU145 cells after 24 h. Paired t-test, significance level: ns not significant, ** p< 0.01, *** p < 0.001, **** p < 0.0001.

Figure 3

NSMase-2-dependent EV secretion from PC3 cells. (A) Experimental setup of EV purification (B) A comparison of the size distribution of secreted particles in P14 and (C) P100 after the treatment of PC3 cells with DMSO or GW4869, with particle diameter in nm and particle concentration in ×10⁶/mL. (D) Concentration differences in P14 with three size bins, 130–350 nm, and (E) P100 with three size bins, 50–210 nm, and their multiples compared to the control group. Statistical analysis was performed using a 2-way ANOVA with multiple testing. (F,G) Western blots of lysate and p100 of PC3 cells with Calnexin, CD63, CD81, Alix, and Syntenin. (H) Quantification of (F,G). Student’s t-test, significance level: ns not significant, ** p< 0.05, *** p < 0.001, **** p < 0.0001.

Figure 4

Differential EV proteomes from PC3 and DU145 cells were investigated by label-free mass spectrometry-based proteomics. (A) Experimental procedure with subsequent quality control and candidate selection. (B) Overlap of identified P14 and P100 proteins with TOP EV proteins from PC3 cells and (C) from DU145 cells. (D) Overlap of proteins enriched upon GW4869 treatment in P14 and P100 fractions in PC3 and (E) DU145 cells. (F) Of the 33 overlapping PC3 EV proteins, 13 overlapped with 26 DU145 EV proteins. (G) String database analysis of these 33 proteins from PC3 and (H) 26 proteins identified on EV from DU145 cells.

Figure 5

NSMase-2-dependent enrichment of cellular components and pathways. GO-term cellular components and Reactome pathway enrichment of proteins upregulated in GW4869-treated cells. (A) GO-term cellular components and (B) Reactome pathways in P100-EVs from PC3 and DU145 cells. (C) GO-term cellular components and (D) Reactome pathways in P14-EVs from PC3 and DU145 cells.

Figure 6

SDC4 knockdown affects PCa cell migration. (A) Heatmap of nSMase-2-dependent differentially regulated proteins in P14 and P100 from PC3 and DU145 cells. (B) Scratch-wound-healing assays of PC3 cells against SRPX2 or (C) SDC4 were treated with siRNA for 24 h before scratching Relative knockdown efficiency was determined by quantitative PCR. (D) Relative wound density was measured in percent after 24 h in PC3. (E) Single data points from independent scratch wounds in PC3 after 24 h. Similarly, knockdown efficiency was determined in DU145 cells (F) against SRPX2 or (G) SDC4, and then (I) relative wound density was measured in percent after 24 h. (J) Single data points from independent scratch wounds show a significant decrease in wound density for SDC4 after 24 h. Significance level: ns not significant, * p< 0.05, ** p< 0.01, *** p < 0.001, **** p < 0.0001.