PRR34-AS1 promotes exosome secretion of VEGF and TGF-β via recruiting DDX3X to stabilize Rab27a mRNA in hepatocellular carcinoma

HCC cells including HLF, Huh-7, SNU-449, HepG2 and LM3 were obtained from COBIOER (Nanjing, China). HLF, Huh-7 and LM3 cells were grown in DMEM. SNU-449 cells were grown in RPMI-1640 medium. HepG2 cells were grown in MEM. All mediums were obtained from Gibco (Grand Island, USA). Human liver epithelial cell (THLE-3) was obtained from ATCC (Manassas, VA, USA) and kept in BEGM (Lonza/Clonetics Corporation, Walkersville). Cells were left to grow at 37 °C under a humid environment with 5% CO₂.

Total RNAs were extracted with the application of Trizol reagent (Invitrogen, USA). For the evaluation of gene expression, cDNA was synthesized with the application of PrimeScript RT master mix (Takara, Japan). Next, SYBR Green PCR Master Mix (Applied biosystems) was utilized to conduct qPCR with 2^−△△CT calculation. GAPDH served as control.

Specific shRNAs targeting PRR34-AS1 (sh/PRR34-AS1), as well as DDX3X (sh/DDX3X), Rab27a (sh/Rab27a) and negative control (sh/NC) were synthesized by Genechem (Shanghai, China). Besides, NC or PRR34-AS1 or Rab27a were obtained from RiboBio (Guangzhou, China). Cells were collected for further experiments after 48 h of transfection.

Cell proliferation was assessed with the application of EdU kit (RiboBio, Guangzhou, China). Cells in 96-well plates were incubated to 90% confluence, and then cultured with 100 μL of 50 μM EdU diluent for 2 h. After fixation with 4% paraformaldehyde and 0.5% Triton X-100 treatment, cells were incubated in the dark with 100 μL of 1× Apollo® staining reaction. DAPI was applied for nuclear redyeing. Images were acquired by a laser confocal microscope.

Cells were planted into 6-well plates (600 cells/well) for 2 weeks, followed by the treatment of 4% paraformaldehyde and 0.5% crystal violet. Finally, the number of colonies (> 50 cells) was counted.

Cells were planted into 6-well plates until completely adhered to the wall, and then cells were scratched for cell culture. Images of wound healing were obtained at 0 h and 24 h after scratching, and the relative wound width was calculated.

Transwell chambers (8-μm pores, Corning, USA) pre-coated with Matrigel (BD, USA) were used for invasion assay. Cells were planted into the upper chamber, and the bottom chamber was filled with medium containing 10% FBS. After 24 h of incubation, cells were removed from the upper surface, and the invaded cells were subjected to the fixation with 4% paraformaldehyde for 10 min at room temperature, and stained in 0.5% crystal violet. The average number of invaded cells per field was assessed using a microscope.

Cells were cultured in RIPA buffer (Beyotime) for the extraction of proteins. BCA protein assay kit II (Beyotime) was performed for determining protein concentrations. The extracted proteins were subjected to the separation on 10% SDS-PAGE and then transferred onto PVDF membrane (Invitrogen), followed by being blocked with 3% BSA. Blots were incubated overnight at 4 °C with antibodies against E-cadherin (1/1000), N-cadherin (1/1000), Vimentin (1/1000), Slug (1/1000), Twist (1/1000), GAPDH (1/1000), CD9 (1/2000), CD63 (1/2000), HSP70 (1/1000), TSG101 (1/1000), DDX3X (1/1000), VEGF (1/1000), TGF-β (1/1000) and Rab27a (1/1000) and all these antibodies were procured from Abcam. Next, blots were incubated with horseradish peroxidase (HRP)-labeled secondary antibodies. The blot signals were measured through enhanced chemiluminescence reagents.

Cells were treated with 4% paraformaldehyde and 0.5% Triton X-100 for fixation and permeabilization, and then blocked in 3% bovine serum albumin. Subsequently, cells were cultured overnight at 4 °C with primary antibodies and cultured for 1 h at 37 °C with specific secondary antibodies. After washing, DAPI was utilized for nuclear redyeing. The images of Immunofluorescence staining were obtained under a confocal microscope.

Exosomes were separated from cells using Ribo exosome isolation reagent (RiboBio, Guangzhou, China) based on the user manual. Cells were centrifuged at 2000×g for 30 min, and mixed with Ribo exosome solution reagent. The mixtures were refrigerated at 4 °C overnight, followed by centrifugation at 1500×g for 30 min. Exosomes were obtained after removing the supernatants.

PKH67 (1 μM, Sigma-Aldrich) was commercially obtained to label the exosomes according to the supplier’s protocol. Cell nuclear was double-stained by DAPI. Images were observed under a laser scanning microscope.

The concentration and distribution of exosome size was measured using ZetaView PMX 10 (Particle Metrix, Germany). Exosomes were re-suspended in granular PBS and added to the instrument. Each sample was recorded in 5 videos to confirm the numbers and size distribution of exosomes.

Cells were centrifuged and treated with 2.5% glutaraldehyde and then treated by osmium tetroxide. Next, cells were embedded with epoxy resin and cut into sections at a thickness of 100 nm, followed by the treatment of uranyl acetate and lead citrate. The images of microscopic samples were captured using a transmission electron microscope.

Cy3-labeled PRR34-AS1 probe was synthesized (RiboBio, Guangzhou, China). A FISH Kit (RiboBio) was applied for assessing the subcellular localization of PRR34-AS1 base on the manufacturer’s instructions.

Subcellular isolation of RNAs was performed with the application of Cytoplasmic and Nuclear RNA Purification Kit which was procured from Norgen (Thorold, ON, Canada). Nuclear and cytoplasmic fractions were measured via RT-qPCR.

For Rab27a promoter analyses, the sequence of Rab27a promoter was firstly sub-cloned into pGL3-basic reporter vectors (Promega, Madison, WI, USA). After the co-transfection of downregulated PRR34-AS1 and plasmids, the activity of Rab27a promoter was evaluated using a luciferase assay kit (Promega).

Magna RNA-binding protein immunoprecipitation kit (Millipore, USA) was utilized to assess the binding between RNAs. Cell lysates were cultured with RIP buffer and antibodies (anti-Ago2, anti-SNRNP70 and anti-DDX3X) conjugated with magnetic beads for 1 h. The enrichment of RNAs was evaluated by RT-qPCR.

Pierce Magnetic RNA–Protein Pull-down kit (Thermo fisher scientific) was used in this assay in the light of supplier’s instructions. PRR34-AS1 sense and PRR34-AS1 antisense were in vitro transcripted and labeled. Biotin-labeled PRR34-AS1 (50 pmol) was cultured with streptavidin magnetic beads (50 μL) with rotation, and then hatched in cell lysates. The eluted proteins were measured by western blot after mass spectrometry.

After reaching 80% confluence, cells were treated with actinomycin D (4 µg/mL, Sigma-Aldrich, USA) to inhibit transcription. The relative mRNA levels were analyzed by RT-qPCR at the indicated time points.

Data were processed by GraphPad Prism 6.0 Software and displayed as means ± standard deviation of three replicates. The Student’s t-test and two-way ANOVA were used to analyze differences between two groups or more than two groups. Data were defined as statistical significance if P < 0.05.

First, GEPIA 2 (http://​gepia2.​cancer-pku.​cn) database was applied to assess PRR34-AS1 expression in liver hepatocellular carcinoma (LIHC) tissues. PRR34-AS1 exhibited a markedly high expression level in LIHC tissues (Additional file 1: Fig. S1A). Consistently, RT-qPCR analysis indicated that PRR34-AS1 was highly expressed in HCC cells including HLF, Huh-7, SNU-449, HepG2 and LM3 in comparison to THLE-3 cells (Fig. 1A). Since HepG2 and LM3 cells presented the highest PRR34-AS1 expression among the selected cells, both HepG2 and LM3 cells were selected for subsequent assays. We stably silenced PRR34-AS1 expression via transfecting two shRNAs targeting PRR34-AS1 into HepG2 and LM3 cells (Additional file 1: Fig. S1B). As shown in EdU and colony formation assays, PRR34-AS1 depletion reduced the percent of EdU positive cells and the number of colonies (Additional file 1: Fig. S1C, D). Meanwhile, the relative wound width was obviously increased when PRR34-AS1 was knocked down, suggesting that PRR34-AS1 silencing inhibited cell migratory ability in HCC (Additional file 1: Fig. S1E). Additionally, the number of invaded cells was also decreased in PRR34-AS1-silenced HepG2 and LM3 cells (Additional file 1: Fig. S1F). Moreover, we also used western blot to evaluate the role of PRR34-AS1 depletion on EMT in HCC cells. It turned out that PRR34-AS1 absence caused an upregulation in the level of epithelium marker (E-cadherin) whereas diminished the levels of mesenchymal markers (N-cadherin and Vimentin) and EMT transcription factors (Slug and Twist), implying that PRR34-AS1 silencing suppressed the EMT process in HCC cells (Additional file 1: Fig. S1G). Collectively, PRR34-AS1 enhances cell proliferative, migratory, invasive and EMT phenotypes in HCC.

We next up-regulated PRR34-AS1 expression in THLE-3 cells (Fig. 1B) and found that the EdU positive stained cells were sharply increased in THLE-3 cells after PRR34-AS1 overexpression (Fig. 1C). Likewise, PRR34-AS1 overexpression led to an augment on the number of colonies (Fig. 1D), suggesting that PRR34-AS1 up-regulation facilitated the proliferation ability of THLE-3 cells. Besides, we found that the relative wound width was shorter in PRR34-AS1 groups compared to NC group, implying that up-regulated PRR34-AS1 promoted the migratory capacity of THLE-3 cells (Fig. 1E). Simultaneously, transwell assays disclosed that PRR34-AS1 overexpression enhanced the invasive capacity of THLE-3 cells (Fig. 1F). Additionally, we also assessed the influence of PRR34-AS1 up-regulation on EMT in THLE-3 cells. E-cadherin level was obviously reduced whereas the levels of N-cadherin, Vimentin, Slug and Twist were enhanced in PRR34-AS1-up-regulated THLE-3 cells (Fig. 1G). Similarly, the intensity of E-cadherin was weakened whereas N-cadherin intensity was strengthened in THLE-3 cells when PRR34-AS1 was overexpressed, implying that PRR34-AS1 overexpression accelerated EMT in THLE-3 cells (Fig. 1H). Taken together, PRR34-AS1 up-regulation facilitated the malignant capacities of THLE-3 cells.

Then, we wondered whether HCC cells could secrete exosomal PRR34-AS1 and transmit them to THLE-3 cells, leading to the promotion of cell proliferation, migration, invasion and EMT in THLE-3 cells. After the co-culture of HepG2 or LM3 cells with THLE-3 cells, we performed functional assays in co-cultured cells. Cell proliferation ability was enhanced in co-cultured cells relative to THLE-3 cells (Fig. 2A, B). Consistently, the potential of cell migration and invasion was markedly promoted in co-cultured cells (Fig. 2C, D). Moreover, we found that E-cadherin level was reduced while the level of N-cadherin, Vimentin, Slug and Twist were elevated in co-cultured cells (Fig. 2E). Meanwhile, the data from immunofluorescence staining also uncovered that the co-cultured cells had a stronger EMT capacity than THLE-3 cells (Fig. 2F). However, PRR34-AS1 expression had no marked change in co-cultured cells (Fig. 2G), which excluded the hypothesis of exosomal PRR34-AS1 in HCC cells. Collectively, the increased cell malignant phenotypes were observed in co-cultured HCC and THLE-3 cells.

Then, we speculated the existences of other exosome transmitted from HCC cells to THLE-3 cells. Exosomes from HCC cells were isolated and labeled with PKH67 dye (green). The result showed that PKH67-labeled exosomes dispersed in HCC cell cytoplasm (Fig. 3A). Besides, we found that exosomes ranged from 50 to 150 nm in diameter (Fig. 3B). Moreover, we found that the exosomes under TEM had the typically round or cup-shaped morphology (Fig. 3C). Additionally, western blot analysis further confirmed that HCC cells had abundant exosomal proteins (CD9, CD63, HSP70 and TSG101) and extracellular vesicle proteins (CD81, ERBB2 and HLA-A) (Fig. 3D). All above results validated the existence of exosomes in HCC cells. Subsequently, we treated HCC cell-secreted exosomes into THLE-3 cells to observe the changes of their cellular phenotypes. As shown in Fig. 3E, F, the proliferation ability was obviously increased in THLE-3 cells (Fig. 3E, F). Meanwhile, the capacities of migration and invasion were apparently enhanced in THLE-3 cells treated with HCC cell-secreted exosomes (Fig. 3G, H). Moreover, we found that the EMT process was also accelerated in THLE-3 cells treated with HCC cell-secreted exosomes (Fig. 3I). Expectedly, the expression of PRR34-AS1 had no significance change in THLE-3 cells treated with HCC cell-secreted exosomes (Fig. 3J). Collectively, HCC cell-secreted exosomes enhanced cell proliferation, migration, invasion and EMT.

We extracted the exosomes secreted from HCC cells and found that the level of total exosomal protein, exosome markers (CD9, CD63, HSP70 and TSG101) and extracellular vesicle proteins (CD81, ERBB2 and HLA-A) was obviously decreased when PRR34-AS1 was downregulated (Fig. 4A–D). Therefore, we speculated that the level of HCC cell-secreted exosomal proteins might be regulated by PRR34-AS1. We then analyzed the level of exosomal proteins in HCC cells treated with downregulated PRR34-AS1. The level of Rab27a was substantially inhibited in HCC cells after PRR34-AS1 silencing and other exosomal proteins had no marked changes (Fig. 4E). Taken together, PRR34-AS1 regulated exosomal protein Rab27a in HCC cells.

To uncover the regulatory mechanism of PRR34-AS1 in HCC cells, we firstly assessed the cellular distribution of PRR34-AS1 in HCC cells. As shown in Fig. 5A and Additional file 2: Fig. S2A, PRR34-AS1 was primarily distributed in cell cytoplasm, indicating that PRR34-AS1 might post-transcriptionally regulate Rab27a in HCC cells. Meanwhile, luciferase reporter assays further confirmed this point, as indicated that PRR34-AS1 could not affect the luciferase activity of Rab27a promoter (Fig. 5B). It is well-acknowledged that lncRNA acts as a ceRNA to regulate cancer progression at post-transcriptional levels. We wondered whether PRR34-AS1 existed in the RNA-induced silencing complex (RISC) to bind to miRNAs and thereby regulated Rab27a expression. However, the outcome from RIP assays dispelled this hypothesis, as indicated that the enrichment of PRR34-AS1 had no significant change in Anti-Ago2 groups (Fig. 5C). Thus we further treated protein synthesis inhibitor (actinomycin D) into HCC cells to assess Rab27a mRNA level. The half-life of Rab27a mRNA was obviously declined after PRR34-AS1 silencing, indicating the regulatory effect of PRR34-AS1 on the stability of Rab27a mRNA (Fig. 5D). Then, the electrophoretic gel indicated that one specific band appeared at approximately 73 kDa and we finally identified it as DDX3X protein (Fig. 5E). Furthermore, the combination between PRR34-AS1 and DDX3X was confirmed by western blot and RIP analyses (Fig. 5F, G). Intriguingly, we also found the combination between Rab27a and DDX3X in HCC cells via RNA pull down assays (Fig. 5H). Moreover, we found that the combination of Rab27a and DDX3X was lessened in HCC cells after the downregulation of PRR34-AS1 (Fig. 5I). Thus we conjectured that PRR34-AS1 might interact with DDX3X to regulate Rab27a mRNA stability. We knocked down the levels of DDX3X and revealed that the levels of Rab27a were markedly decreased in HCC cells after DDX3X depletion (Additional file 2: Fig. S2B, C). Meanwhile, the stability of Rab27a mRNA was also reduced when DDX3X was silenced in HCC cells treated with actinomycin D (Fig. 5J). Collectively, PRR34-AS1 interacted with DDX3X to mediate Rab27a mRNA stability.

Next, we investigated the role of Rab27a in THLE-3 cells. We overexpressed Rab27a expression in THLE-3 cells and observed that the proliferation ability was obviously enhanced when Rab27a was up-regulated in THLE-3 cells (Fig. 6A–C). In parallel, the capacities of migration and invasion were promoted in THLE-3 cells after Rab27a up-regulation (Fig. 6D, E). Moreover, we found that Rab27a overexpression elevated the EMT process in THLE-3 cells (Fig. 6F, G). It has been reported that Rab27a promotes tumor growth via increasing VEGF and TGF-β secretion in vitro [22]. Herein, we also found that the protein concentration of VEGF and TGF-β in THLE-3 cells as well as their protein levels in HCC cell-secreted exosomes was increased after Rab27a up-regulation (Fig. 6H, I). All above results suggested that Rab27a increased VEGF and TGF-β secretion in HCC cells and transmitted them into THLE-3 cells to elevate cell proliferative, migratory and invasive phenotypes as well as EMT.

To explore the effect of PRR34-AS1 and Rab27a on the malignant phenotypes of THLE-3 cells, rescue experiments were arranged after the verification of the knockdown efficiency of Rab27a (Additional file 3: Fig. S3A). As indicated in Additional file 4: Fig. S4A, B, Rab27a silencing reversed the enhanced proliferation ability in THLE-3 cells after PRR34-AS1 overexpression. Besides, it was found that the enhanced migration and invasion capacities induced by PRR34-AS1 up-regulation were abolished by co-transfection of sh/Rab27a#a (Additional file 4: Fig. S4C, D). Moreover, the decreased protein level of E-cadherin and the enhanced levels of N-cadherin, Vimentin, Slug and Twist in THLE-3 cells transfected with PRR34-AS1 overexpression was overturned by co-transfection of sh/Rab27a#a (Additional file 4: Fig. S4E). Meanwhile, the data of immunofluorescence assays further confirmed that Rab27a depletion abrogated the increased EMT process in THLE-3 cells after PRR34-AS1 augment (Additional file 4: Fig. S4F). Taken together, PRR34-AS1 facilitated THLE-3 cells proliferation, migration, invasion and EMT via elevating Rab27a expression.