RNA helicase DDX5-induced circPHF14 promotes gastric cancer cell progression

Cell culture

The AGS and 7901 gastric cancer cells were cultured in RPMI-1640 medium containing 10% fetal bovine serum (FBS). Cell lines were obtained from American Type Culture Collection (https://www.ATCC.org, ATCC) for this study. A mycoplasma test was negative for the cell line.

Plasmid construction, and transfection with short interfering RNA (siRNA)

DNA fragments encoding Flag-DDX5 was created by PCR with cloning into the p3×Flag-CMV-10 empty vector. All siRNAs in this study were obtained from Ribobio (Guangzhou, China). All siRNAs and plasmids in this study were transfected into cells by transfection reagent Lipofectamine 3000 (Invitrogen). CircPHF14 siRNA targeting sequences (5’- TCATCTTCCAAGAGAACCA -3’ or negative control sequences (5’-TTCTCCGAACGTGTCACGT-3’). CircPHF14 knockdown was identified by qPCR.

Real-time quantitative PCR (qPCR)

2 μg of total RNA was reverse transcribed into cDNA using Reverse transcription kit (Vazyme, Nanjing, China). The qPCR kit was obtained from Takara. All primers in the present research were synthesized by Sangon (Shanghai, China). Human GAPDH or U6 GAPDH or U6 were used as negative controls. The expression of each genetic profile was quantified by measuring the cycle threshold, and the relative change of gene expression was calculated by mathematical method. Primer sequences are shown as follows: GAPDH, F-5’-GGGAGCCAAAAGGGTCAT-3’, R-5’- GAGTCCTTCCACGATACCAA-3’. CircPHF14, F-5’-GTGATTCTTCATCTTCCAAGAGAAC-3’, R-5’- TGCAACACAGACTGTCTTTGTCTTC-3’.

Western blot

The proteins were extracted using RIPA lysate containing protease inhibitors, and the extracted proteins were quantified by BCA. 30 μg of protein was used for SDS-PAGE in each treatment group, and the proteins were transferred to PVDF (Millipore) and incubated sequentially with primary and secondary antibodies for the target proteins.

Bioinformatics analysis

The analysis of Oncomine database (http://www.oncomine.org) was made as described before [27]. Shortly, we used the following threshold values to evaluate the expression of DDX5 in gastric cancer tissues and corresponding normal gastric tissues: fold-change of 2, P value of 0.05. The UALCAN database (http://ualcan.path.uab.edu/index.html) is used to analyze public TCGA samples. The Human Protein Atlas database (https://www.proteinatlas.org/) was used to obtain the immunohistochemical detection data of DDX5 in gastric cancer specimens and corresponding non-cancerous samples.

RNase R and actinomycin D treatment

Actinomycin D (Sigma, USA) treated cells blocked RNA transcription to detect the stability of circPHF14 and linear PHF14 mRNA, and dimethyl sulfoxide (DMSO) was used as negative control. Take the final concentration of 1 μg/mL DMSO or actinomycin D treated cells for 0, 8, 16 and 24 h, RNA was extracted and detected by reverse transcription q-PCR. For the RNase R digestion, TRIzol was used to extract total RNAs using RNase R processing in buffer added with Ribonuclease Inhibitor at 37° C for 1 h. Reverse transcription Kit (Vazyme, Nanjing, China) reverse transcribed RNA with RNase R-treated to synthesize cDNA and detected by q-PCR.

circRNA sequencing

Following the manufacturer’s instructions, total RNA was isolated using the RNeasy Mini kit (Qiagen). RNA quality was ensured by adding DNase I in column digestion. Two biologically replicated samples were conducted for each treatment. For circRNA sequencing, before the RNA library was constructed, total RNA was digested with RNase R at 37° C for 1 h. NEBNext® Ultra™ Directional RNA Library Prep Kit for Illumina (E7420L) was used to construct an RNA library for circular RNA sequencing. A paired-end sequencing experiment was performed at RiboBio Co., Ltd. or Amogene Biotech Co., Ltd. with Illumina HiSeq 3000. Under accession number GSE207636, the raw expression files and details are available in NCBI GEO.

Cell proliferation assay

To analyze cell proliferation by Cell Counting Kit-8 (CCK-8; Biotool, China). Briefly, 96 well plates were used in each group, and 1000 cells were cultured in each well. From 1 day to 4 days after incubation, 10 ml of CCK-8 reagent was added to each of the well and placed at 37° C for 2 hours. Measuring the absorbance value of 450nm.

In the plate colony formation experiment, 1000 treated gastric cancer cells were cultured in 6-wellplates and cultured for two weeks in RPMI-1640 medium supplemented with 10% fetal bovine serum for two weeks. Then, the cells were cleaned with PBS, fixed with 4% paraformaldehyde, and most stained with Viola crystals. Microscope and Image J software were used to estimate the percentage and intensity of the area covered by stained cell colonies.

EdU incorporation assay was performed after 48 hours of culture of the treated cells based on the instruction manual. Then, the cells were washed and stained with buffer containing DAPI, and imaged were collected using a fluorescence microscopy.

Cell migration assays

Transwell containing polycarbonate filters with 8-μm pores was used for cell migration testing. 5×10⁵ cells were suspended in serum-free medium and added to the upper chamber; Add RPMI-1640 medium (10% fetal bovine serum) to the small chamber underneath. The migrated cells were cleaned with PBS, fixed in 4% paraformaldehyde and finally stained with Viola crystal.

Immunofluorescence

Cells were fixed using formalin, then the cell membrane was permeabilized with Triton X-100, incubated with DDX5 antibody and circPHF14 probe followed by fluorescent secondary antibody, and finally photographed with fluorescent microscope.

Animal experiment

All animal euthanasia and care protocols were approved by the Institutional Animal Care and Use Committee of Changzhi Medical College (Changzhi, China). In the research of cancer cell xenograft, about 3-4 weeks old nude mice were injected subcutaneously 5×10⁶ AGS cells in the right abdomen of the mice transfected with DDX5/ DDX5+si-circPHF14/negative control (nc) in 100 μl serum-free medium that was mixed with matrigel (1:1). The mice were dissected for about 4 weeks, and the tumor tissues were taken out for fixation, dehydration and photography.

Statistical analysis

Statistical analysis was conducted utilizing Graphpad prism 8 and SPSS 19.0 software. Data were typically expressed as the mean ± SD, and the differences between groups were assessed with the unpaired Student t-test. P < 0.05 was regarded as be significant.

Availability of data and materials

The datasets, cell lines, plasmids and other reagents described in this manuscript are available upon a reasonable request.

DDX5 expression is correlated with poor patient survival in gastric cancer

To further characterise the connection of DDX5 and gastric cancer, we analysed its expression in the gastric tissues of tumour patients. Oncomine database is used to obtain the public gene expression data of DDX5 [28]. We analyzed three datasets from DErrico et al., Deng et al., Chen et al. [29–31]. And a significantly higher DDX5 expression in tumour tissues compared to non-cancerous normal tissues. (Figure 1A). Through The Cancer Genome Atlas database [32], we validated that DDX5 expression was significantly higher in gastric cancer samples than in non-tumour controls. (Figure 1B). Using the gene expression dataset of gastric cancer patients from the KM Mapper database, an association study between DDX5 and overall survival was conducted. Patients with higher DDX5 expression were found to have a lower overall survival when compared to those with low DDX5 expression (Figure 1C). Next, the expression of DDX5 in normal tissues and gastric cancer patients was analysed in The Human Protein Atlas database. In agreement with the results from public databases, DDX5 expression was higher in gastric cancer specimens than in normal specimens. (Figure 1D). We also further examined the protein expression levels of DDX5 in multiple cell lines by western blot assay and found that the protein expression levels of DDX5 were generally higher in gastric cancer cells than in normal cells GES1 (Supplementary Figure 1).

Figure 1. DDX5 expression is correlated with poor patient survival in gastric cancer. (A) The expression of DDX5 was analyzed in patients with gastric. DDX5 expression in gastric cancer samples and non-cancer controls; the publicly accessible gene expression data of DDX5 was obtained from Oncomine database. (B) DDX5 expression in gastric cancer samples and non-cancer controls; The publicly accessible gene expression data of DDX5 was obtained from The Cancer Genome Atlas (TCGA) database. (C) Kaplan-Meier overall survival curves according to DDX5 expression in patient cohorts in KM-plotter database. The percentage of survival patients in high DDX5 and low DDX5 groups at different time points are presented. (D) DDX5 expression in gastric cancer patient with or non-cancer controls was measured by immunohistochemical staining. The publicly accessible protein expression data of DDX5 was obtained from The Human Protein Atlas database. Data were analyzed with Student’s t-test, p values were shown. Scale bar 200μm.

DDX5 promotes cancer cell growth and migration

Next, to assess the regulatory function of DDX5 in the migratory capacity of gastric tumour cell lines, DDX5 was transfected into AGS cells and 7901 cells using a DDX5 expression vector. DDX5 overexpression notably accelerated the proliferation and growth of gastric cancer cells as shown by tumor cell clonogenesis assay, EdU staining and CCK8 cell proliferation assay (Figure 2A–2C). For the purpose of testing the role of DDX5 in cell migration, we carried out transwell experiments in gastric cancer cells. DDX5 expression enhanced the migration capability in AGS and 7901 cell (Figure 2D). Together, it is suggested that DDX5 stimulates the growth and migration on gastric cancer cells in vitro.

Figure 2. DDX5 promotes cancer cell growth and migration. (A) Colony formation of AGS and 7901 cells two weeks after transfection with the indicated vector. (B) CCK-8 assay of proliferation in AGS and 7901 cells following a time course of transfection with NC or DDX5. (C) EdU assay was conducted to test the effect of DDX5 expression on the proliferation of AGS and 7901 cells. Scale bar 100μm. (D) Transwell assay was conducted to test the effect of DDX5 expression on the migration of AGS and 7901 cells. Number of cells were counted and shown in the column graph on the right of the corresponding pictures. Data are mean ± SD of three independent experiments. * P < 0.05, ** P < 0.01. Scale bar 500μm.

circRNAs are induced by DDX5

We know that circular RNA is obtained by covalently linking a downstream splicing donor to an upstream splicing acceptor, a backsplicing process catalyzed by the spliceosome [33]. DDX5, as a well-known splicing factor, it has been reported to regulate a large number of splicing events [34, 35]. Thus, we speculated whether DDX5 could promote the progression of gastric cancer cells via regulating the splicing formation of tumor-related circular RNAs. To identify whether the expression of DDX5 was linked to circular RNAs splicing processes, we transfected AGS cell with FLAG-DDX5 or FLAG-tagged empty vector (negative control, NC), and analyzed the differentially expressed circRNAs by Ribo zero rRNA removal kit and circRNA-seq (RNase R-treated RNA sequencing) (Figure 3A and Supplementary Figure 2). We analyzed the circRNA-seq data by CIRCexplorer2 and CIRI2software and 7587 circular RNAs were found in the control group and DDX5 overexpressed cells (Supplementary Table 1, Supporting Information). Next, we asked whether DDX5 could change the program of circRNAs in AGS cells via comparing confirmed circRNAs in DDX5-overexpressed cells to that of control cells. We found that DDX5 overexpression promoted or inhibited the expression of 72 and 80 circular RNAs, respectively (Supplementary Table 2, Supporting Information and Figure 3B). To reveal the potential biological roles of these DDX5-induced circRNAs, we carried out Gene Ontology (GO) and found that a lot of DDX5-induced circRNAs are components of mRNA processing (Figure 3C). Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis also revealed that a lot of DDX5-regulated circRNAs are mainly enriched in tumor-related signaling pathway, including VEGF and insulin signaling pathway (Figure 3D). These results suggested that DDX5 may regulate circRNA biogenesis.

Figure 3. A large number of circRNAs are induced by DDX5. (A) Workflow for sequencing and analysis of DDX5-treated specimens. (B) Heatmap of differentially expressed circRNAs between NC and DDX5-overexpression AGS cells identified by circRNA-seq. (C) GO enrichment analysis was used to analyze the biological functions of the DDX5-regulated differentially expressed circRNAs. (D) KEGG pathway was used to analyze the pathways related to the DDX5-regulated differentially expressed circRNAs.

DDX5 alters circPHF14 expression

In order to identify novel circRNAs that plays a significant biological role in gastric cancer, particularly its growth and migration. In these DDX5-induced circular RNAs, we found that hsa_circ:chr7:11101418-11151094 has a very top expression change after DDX5 overexpression in AGS cells. Moreover, the expression of hsa_circ:chr7:11101418-11151094 in gastric cancer cell lines was significantly higher than that in normal immortalized GES1 cells (Figure 4A). RT-PCR and Sanger sequencing showed that hsa_circ:chr7:11101418-11151094 was formed by back-splicing of exons 14-16 of PHF14 gene (reference sequence: NR_033436.2). The total length of hsa_circ:chr7:11101418-11151094 is 291 nt, and its splicing sequence is consistent with that of hsa_circ_0079440 in circBase; Therefore, it is named circPHF14 (Figure 4B). We further detected the effect of DDX5 on the expression of circPHF14 in gastric cancer cells by qRT–PCR. The results showed that the expression of circPHF14 changed continuously with the increase of DDX5 level in AGS and 7901 cells (Figure 4C). We treated gastric cancer cells with actinomycin D or RNase R to inhibit intracellular RNA transcription. The qPCR experiment showed that the circPHF14 was more stable than the linear mRNA of the PHF14 gene (Figure 4D, 4E).

Figure 4. DDX5 alters circPHF14 expression. (A) Endogenous circPHF14 expression levels in multiple gastric cancer cells by RT-PCR. (B) We confirmed circPHF14 was formed by circularization of exons 14, 15, 16 in the PHF14 mRNA by Sanger sequencing, and also determined its genomic size and sequence. (C) RT-PCR was performed in AGS or 7901 cells transfected with DDX5, and then relative expression of circPHF14 was detected by qPCR. (D, E) The relative expression of circPHF14 and PHF14 mRNA in AGS or 7901 cells was detected by RT-PCR after RNase R treatment for 30 min (D) or actinomycin D treatment for 0 h, 8 h, 16 h, and 24 h (E). Data are mean ± SD of three independent experiments. * P < 0.05, ** P < 0.01, ns stands for no significance.

We also found a significant co-localization between DDX5 and circPNF14 using immunofluorescence experiments (Supplementary Figure 3).

Silencing circPHF14 reverses the DDX5-induced tumor-promoting effects in gastric cancer cells

To prove that DDX5 promotes the development of gastric cancer cells, at least in part by regulating the expression of circPHF14, it is important to confirm whether inhibition of circPHF14 can reverse the effect of DDX5 on gastric cancer cells. Thus, rescue assays were performed via co-transfecting DDX5 expression vector and si-circPHF14 or si-NC (Supplementary Figure 4). Knockdown of both circPHF14 could partly rescue the excessive proliferation abilities of overexpressed DDX5 gastric cancer cell lines compared with the negative control group by plate colony formation, CCK-8 and EdU staining assays (Figure 5A–5C). Additionally transwell migration assay shown that the promotion of cell motility ability was reversed via the inhibition of circPHF14 expression (Figure 5D).

Figure 5. DDX5 promotes gastric cancer cells growth and migration by regulating circPHF14. (A) AGS or 7901 cells transfected with control siRNA or siRNA specifically targeting circPHF14 (si-circPHF14) in the presence DDX5 were subjected to colony formation assay. (B) CCK-8 assay of proliferation in AGS and 7901 cells following a time course of transfection with NC or DDX5 or DDX5 and si-circPHF14. (C) EdU assay was conducted to test the effect of DDX5 and circPHF14 on the proliferation of AGS and 7901 cells. Scale bar 100μm. (D) A migration rescue experiment verified the effect of DDX5 and circPHF14 on the migration ability of AGS and 7901 cells. Data are mean ± SD of three independent experiments. * P < 0.05, ** P < 0.01. Scale bar 500μm.

DDX5 promote gastric cancer cells progression in vivo via circPHF14

To determine whether DDX5 and circPHF14 an extremely significant function in the development of gastric cancer cells in vivo, we established a xenograft nude tumor model. Stable transfection of AGS gastric cancer cells with NC or co-transfected with DDX5 and sh-NC or co-transfected with DDX5 and sh-circPHF14, and injected separately into the nude mice. As shown in Figure 6A, 6B, the DDX5 overexpressed group had higher proliferation rate than the negative control group, whereas cells silenced for circPHF14 partially reversed the inhibition in proliferation. In addition, the volume of AGS-derived tumors was inhibited by reducing the expression of circPHF14 in vivo (Figure 6C, 6D). These findings suggest that DDX5 may play a key role in accelerating the growth of gastric cancer cells through circPHF14 in vivo.

Figure 6. DDX5 promote gastric cancer cells progression in vivo via circPHF14. (A) Tumor growth curves of AGS xenografts in nude mice. Group 1, 2, 3 mice were injected with DDX5, NC, DDX5 and si-circPHF14 cells, respectively. (B–D) Xenograft tumors collected on day 30 post subcutaneous implantation.