Characteristics of CAFs isolated from PCa patients and isolation of exosomes
We isolated and cultured three sets of primary CAFs from PCa samples. Next, we characterized the phenotypes of primary CAFs and hTERT PF179T CAF cell line. We performed immunofluorescence to detect the expression of CAF markers such as a-SMA, FAP, and Vimentin (Fig. 1a and Additional file 1: Figure S1a). we used western blotting assay to detect the androgen receptor (AR) expression in CAFs and found that AR was expressed in all CAFs but at a lower level than LNCaP cells (Additional file 1: Figure S1b).
CAF-secreted exosomes can contribute to the progression and metastasis by transferring different types of substances in various tumours [26]. Previous study showed that inhibiting AR signalling in CAFs might promote PCa cell migration by enhanced secretion of CCL2 and CXCL8 [11], we would like to determine whether exosomes derived from CAF exposed to androgen deprivation might promote migration or invasion of PCa cells. Exosomes were isolated from the conditioned medium of CAFs. Firstly, the exosomes were resuspended with PBS and characterized by transmission electron microscopy (TEM). As shown in Fig. 1b, particles isolated from the conditioned medium of CAFs ranged from 30 to 150 nm. Secondly, A western blotting analysis of purified exosomes revealed the expression of the exosomal positive markers CD81 and TSG101 (Fig. 1c). Lastly, the effect of DHT treatment on exosome release in CAFs were assessed. We used Nanoparticle Tracking Analysis to confirm the size and concentration of released exosomes. Upon DHT treatment, CAFs did not display increased levels of exosome secretion (Fig. 1d). Based on the results above, we verified that the particles isolated from the supernatants of CAFs were exosomes and DHT did not influence the number of exosomes released from CAFs.
Exosomes from CAFs after ADT increase migration and invasion of PCa cells via activating EMT both in vitro and in vivo
We next evaluated the effect of CAF-derived exosomes on PCa cell migration and invasion while inhibiting AR signalling in CAFs. Exosomes were collected as described and added to recipient cells in freshly 10% CSFBS DMEM medium for 48 h. Compared with DHT-treated CAFs-derived exosomes (DHT-treated CAF Exo), ETOH-treated CAFs-derived exosomes (ETOH-treated CAF Exo) significantly enhanced the migration and invasion ability of LNCaP and DU145 (Figs. 2a-d). In addition, we detected the expression of EMT markers E-cadherin and Vimentin in PCa cells (Fig. 2e, f). Due to the low expression of Vimentin in LNcaP cells, we only detected the protein expression of Vimentin in DU145. The results showed that PCa cells in the ETOH-treated CAF Exo group expressed lower level of E-cadherin than in the DHT-treated CAF Exo group (Fig. 2e, f). And we also found DHT-treated CAFs-derived exosomes decreased the expression of Vimentin in DU145 than the control group; however, exosomes isolated from ETOH-treated CAFs increased the expression of Vimentin in DU145 than in the DHT-treated CAF Exo group (Fig. 2f).
To confirm the effects of exosomes derived from CAFs exposed to ADT on PCa metastasis in vivo, androgen ablation was induced by surgical castration in mice. Prior to tail vein injection, DU145 luciferase cells were pre-incubated with exosomes derived from DHT/ETOH-treated CAFs or PBS twice a day for 4 days. Then we injected the indicated cell lines (106 cells in 100 ul PBS per mouse) through the tail vein. After injection of tumour cells, mice were treated with CAF-derived exosomes or PBS via tail vein injections every other day for 2 weeks. Then, bioluminescence imaging was employed to monitor tumour metastasis. Consistent with the experiments in vitro, luciferase signals and the number of lung metastases were higher in the ETOH-treated CAF Exo group than DHT-treated CAF Exo group and control group (Fig. 3a, b). Significantly, immunohistochemical analysis demonstrated that the expression level of E-cadherin was lower, while the abundance of Vimentin was higher in ETOH-treated CAFs Exo group (Fig. 3c, d).
In addition, we assessed the effect of CAF-derived exosomes on PCa cell proliferation under castration condition while inhibiting AR signalling in CAFs. However, we found that ETOH/DHT-treated CAFs-derived exosomes did not affect the proliferation of PCa cells under castration condition (Additional file 1: Figure S2a-d).
In conclusion, after androgen deprivation therapy, exosomes from CAFs might promote the migration and invasion of PCa cells via activating EMT both in vitro and in vivo.
The level of miR-146a-5p significantly decreased in the exosomes isolated from CAFs after ADT
MiRNAs are principal and important cargos in exosomes. To figure out the molecular mechanism of ETOH-treated CAFs-derived exosomes promoting migration and invasion of PCa cells, we performed miRNA sequencing to analyse the miRNA profiles of exosomes derived from DHT-treated CAFs and matched ETOH-treated CAFs. As shown in Figs. 4a-c, 6 miRNAs were significantly upregulated, and 5 miRNAs were significantly downregulated in the ETOH-treated CAFs-derived exosomes. Significantly, we observed that the expression of miR-146a-5p had a marked decrease in ETOH-treated CAFs-derived exosomes. Therefore, we performed real-time PCR to confirm the sequence result. The result showed that the level of miR-146a-5p was significantly lower in exosomes derived from ETOH-treated CAFs than that of exosomes derived from DHT-treated CAFs (Fig. 4d). Furthermore, the level of miR-146a-5p was also downregulated in ETOH-treated CAFs compared with matched DHT-treated CAFs (Fig. 4e). Therefore, among these miRNAs, we chose miR-146a-5p for further study.
Exosomes Transfer Mir-146a-5p From Cafs To Pca Cells
To determine if miR-146a-5p could be transferred via exosomes from CAFs to PCa cells, CAFs were transiently transfected with Cy3-labeled miR-146a-5p, then co-cultured with DU145 for 24 h. We used fluorescence microscopy to detect the red signals (Cy3-labeled miR-146a-5p) and green signals (Vimentin) in DU145. Cy3-tagged miR-146a-5p was observed in DU145 and the red signals were abolished when treating CAFs with the exosome inhibitor, GW4869 (Fig. 5a).
We then isolated exosomes from CAFs transfected with Cy3-labeled miR-146a-5p. The purified exosomes were added to DU145 and incubated for 24 h. After that, we detected the red fluorescent signals in DU145 with fluorescence microscopy (Fig. 5b). In conclusion, these results suggested that miR-146a-5p could be transferred from CAFs to PCa cells via exosomes.
Significantly, RT-PCR analysis showed that miR-146a-5p expression was decreased in PCa cells after treating with exosomes from ETOH-treated CAFs than that of cells treated with DHT-treated CAFs-derived exosomes (Fig. 5c). In addition, we analysed the expression of miR-146a-5p in lung metastasis of the nude mice among three groups. It showed that the expression of miR-146a-5p in ETOH-treated CAF Exo group was much lower than that in DHT-treated CAF Exo group (Fig. 5d). Furthermore, RT-PCR analysis confirmed that the expression of miR-146a-5p was much lower in the tumours from patients who received ADT than that of tumours from patients who underwent operations without ADT therapy (Fig. 5e). Taken together, CAFs may decrease the expression of miR-146a-5p in PCa cells through downregulating the transfer of exosomal miR-146a-5p after ADT.
MiR-146a-5p impairs cell migration and invasion by inhibiting EMT in PCa cells
To determine whether miR-146a-5p influenced the migration and invasion of PCa cells, we transfected cells with miR-146a-5p mimics and conducted a Transwell assay. Interestingly, we found that miR-146a-5p overexpression decreased the migration and invasion of both LNcaP and DU145 cells (Figs. 6a-d). In addition, western blotting showed that relative expression of E-cadherin was increased in both LNcaP and DU145 cells after transfection of miR-146a-5p mimics (Fig. 6e, f). And we found that miR-146a-5p suppressed the expression of Vimentin in DU145 (Fig. 6f). In brief, miR-146a-5p impairs the migration and invasion by inhibiting EMT in PCa cells.
EGFR is a direct target of exosomal miR-146a-5p in PCa cells
To delineate the potential targets of miR-146a-5p in prostate cancer, we used publicly available bioinformatics tools (miRWalk, miRTarBase) to predict 89 target genes. Among these genes, EGFR was reported to promote prostate cancer bone metastasis [27]. Using miRTarBase, there are three predicted binding sites of miR-146a-5p in EGFR 3’UTR region and we designed a luciferase vector containing a wild-type or mutant EGFR 3’UTR region (Fig. 7a). To verify whether miR-146a-5p could target the 3’UTR of EGFR directly, a dual luciferase reporter assay was conducted. The result showed that the luciferase activities of EGFR WT 3’UTR could be significantly reduced after cotransfection with miR-146a-5p mimics compared with cotransfection with miR-NC mimics. In contrast, no significant direct interaction was observed between miR-146a-5p and the vector containing EGFR MUT 3’UTR (Fig. 7b).
In both LNcaP and DU145, overexpression of miR-146a-5p markedly inhibited the EGFR mRNA level (Fig. 7c). Additionally, western blot analysis showed that transfection of miR-146a-5p repressed the protein expression of EGFR in both LNcaP and DU145 cells (Fig. 7d). As expected, after incubation with CAF-derived exosomes, the protein level of EGFR in both cell lines were increased in ETOH-treated CAF Exo group than that in the DHT-treated CAF Exo group (Fig. 7e). Moreover, IHC analysis confirmed that EGFR staining was significantly stronger in lung metastasis of mice treated with ETOH-treated CAFs-derived exosomes than that of mice treated with DHT-treated CAFs-derived exosomes (Fig. 7f). Together, EGFR is a direct target of exosomal miR-146a-5p in PCa cells.
Exosomal miR-146a-5p exhibits its functions by inhibiting EGFR/ERK pathway in PCa cells
To further study whether miR-146a-5p exerted its effects by suppressing EGFR expression in PCa cells, we restored EGFR in PCa cells co-transfected with lenti-miR-146a-5p. Transwell assays showed that EGFR overexpression reversed the effects of miR-146a-5p in both LNcaP and DU145 (Figs. 8a-d).
Interestingly, miR-146a-5p also downregulated the protein level of p-ERK1/2 but no effect on total ERK1/2 protein. Meanwhile, Upregulating EGFR expression ectopically counteracted the effects of miR-146a-5p (Fig. 8e). Furthermore, immunohistochemical staining showed that the expression level of p-ERK was elevated in lung metastasis of ETOH-treated CAF Exo group than that of DHT-treated CAF Exo group (Fig. 8f). Significantly, we found that the expression of E-cadherin was decreased and the protein level of Vimentin, EGFR and p-ERK were enhanced in androgen deprived prostate tissues (Fig. 8g).
Collectively, these results indicated that loss of exosomal miR-146a-5p promotes EMT and cell migration and invasion in PCa cells through activating EGFR/ERK pathway (Fig. 9).