Circular RNA CDR1as Exerts Oncogenic Properties Partially through Regulating MicroRNA 641 in Cholangiocarcinoma

It has been found that the circular RNA (circRNA) CDR1as is upregulated in cholangiocarcinoma (CCA) tissues. In this study, we tried to explore the roles of CDR1as in CCA. CDR1as was overexpressed or knocked down in human CCA cells to assess the effects of CDR1as on cell behaviors and tumor xenograft growth. In vitro, the CDR1as level was significantly increased in CCA cell lines. The results showed that CDR1as promoted the cell proliferation, migration, invasion, and activation of the AKT3/mTOR pathway in CCA cells.

ABSTRACT

It has been found that the circular RNA (circRNA) CDR1as is upregulated in cholangiocarcinoma (CCA) tissues. In this study, we tried to explore the roles of CDR1as in CCA. CDR1as was overexpressed or knocked down in human CCA cells to assess the effects of CDR1as on cell behaviors and tumor xenograft growth. In vitro, the CDR1as level was significantly increased in CCA cell lines. The results showed that CDR1as promoted the cell proliferation, migration, invasion, and activation of the AKT3/mTOR pathway in CCA cells. Moreover, miR-641, a predicted target microRNA (miRNA) of CDR1as, could partially reverse the effects of CDR1as on cell behaviors in CCA cells. Furthermore, CDR1as improved tumor xenograft growth, and it could be attenuated by miR-641 in vivo. Additionally, CDR1as expression was inversely correlated with miR-641 in CCA cells, and miR-641 could directly bind with CDR1as and its target genes, the AKT3 and mTOR genes. Mechanistically, CDR1as could bind with miR-641 and accelerate miR-641 degradation, which possibly leads to the upregulation of the relative mRNA levels of AKT3 and mTOR in RBE cells. In conclusion, our findings indicated that CDR1as might exert oncogenic properties, at least partially, by regulating miR-641 in CCA. CDR1as and miR-641 could be considered therapeutic targets for CCA.

INTRODUCTION

Cholangiocarcinoma (CCA) is one of the most common hepatic malignancies after hepatocellular carcinoma and accounts for 3% of all gastrointestinal cancers (1, 2). CCA is difficult to diagnose at an early stage, and surgical resection is still the only potential cure. However, because of early metastasis, only limited CCA tumors are resectable, and the 5-year survival rates are only about 20% to 40% (3–5). Therefore, it is urgent to understand the molecular pathogenesis of CCA and identify the potential molecular regulators for novel therapeutic approaches.

Circular RNAs (circRNAs) are a special type of noncoding RNA molecule, and circRNAs have location-specific expression levels in tissues (6, 7). circRNAs act as posttranscriptional regulators and function as an efficient microRNA (miRNA) sponge and a competing endogenous RNA (ceRNA) to regulate miRNA (8–10). circRNAs play important regulatory roles in multiple biological processes such as cell proliferation, migration, invasion, and apoptosis through regulating the levels of miRNA and their targeted genes (11–13). Furthermore, circRNAs may be involved in the progression of many diseases, including cancer development, and serve as new clinical diagnostic markers (6, 7, 14).

A previous study found that the circRNA CDR1as (ciRS-7) can function as an inhibitor/sponge for the miRNA miR-7 in embryonic zebrafish (15). CDR1as interacts with miR-7 and then downregulates the levels of miR-7, which leads to the upregulation of the expressions of miR-7 target genes, affecting multiple biological processes (16). For example, growing evidence suggests that CDR1as could promote proliferation, migration, and invasion in hepatocellular carcinoma, colorectal cancer, non-small-cell lung cancer, and pancreatic cancer (17–21). Moreover, it has been found that the expression of CDR1as is upregulated in CCA tumor tissues compared to adjacent normal tissues (22). However, the role of CDR1as in CCA remains to be elucidated.

miR-641 was predicted to be one of the potential target miRNAs of CDR1as on Starbase. miR-641 functions as a tumor suppressor by inhibiting cell proliferation, migration, and invasion and inducing cell apoptosis of cervical cancer cells (23). miR-641 can also suppress cell proliferation and induce apoptosis in gastric cancer and lung cancer cells (24, 25). However, the functions of miR-641 in CCA have not been investigated. It has been reported that miR-641 could inhibit the proliferation of gastric cancer cells and glioblastoma development by suppressing the phosphorylation of AKT (24, 26). It has been reported that the AKT/mTOR pathway contributes to cell proliferation, migration, and tumorigenesis in CCA (27, 28). The AKT3 gene was predicted to be a potential target gene of miR-641 on Starbase. Therefore, we wanted to know whether miR-641 could regulate the expression of AKT3. In this study, we aimed to reveal the expression of CDR1as in CCA cells and the effects of CDR1as on cell proliferation, migration, and invasion of CCA cells in vitro and in a xenograft mouse model. Furthermore, whether miR-641 is an important miRNA contributing to the functions of CDR1as in CCA was investigated. Our findings might provide a better understanding of tumor oncogenesis and a therapeutic target for CCA.

RESULTS

CDR1as is upregulated and promotes cell proliferation of CCA cells.

The expression levels of CDR1as in the cholangiocyte cell line HIBEpiC and the human CCA cell lines HCCC-9810 and RBE were determined by real-time quantitative PCR (RT-qPCR). As shown in Fig. 1A, the expression of CDR1as was upregulated in CCA cell lines. Next, CDR1as was overexpressed and knocked down in HCCC-9810 and RBE cells, respectively. Briefly, the vectors with CDR1as (CDR1as-OV) or without CDR1as (negative control 1 [NC1]) and the vectors containing short hairpin RNAs (shRNAs) for CDR1as (CDR1as-sh1/sh2) or negative control 2 (NC2) were transfected into HCCC-9810 and RBE cells，respectively (Fig. 1B and C). The results from a Cell Counting Kit-8 (CCK-8) assay showed that cell proliferation was enhanced by CDR1as overexpression (Fig. 1D and G) and impaired by CDR1as knockdown (Fig. 1E and H). Additionally, results from cell colony formation assays showed that the upregulation of CDR1as significantly promoted cell growth and that the downregulation of CDR1as inhibited the cell growth of HCCC-9810 and RBE cells (Fig. 1F and I).

FIG 1.

Effects of CDR1as on the proliferation of CCA cell lines. (A) Expression levels of CDR1as in the cholangiocyte cell line HIBEpiC and two human CCA (cholangiocarcinoma) cell lines, HCCC-9810 and RBE, were determined by RT-qPCR. (B and C) Expression levels of CDR1as in CDR1as-OV/NC1- or CDR1as-sh1/sh2/NC2-transfected HCCC-9810 or RBE cells were determined by RT-qPCR. (D to F) Cell proliferation of CDR1as-overexpressing and CDR1as-silenced HCCC-9810 cells was demonstrated by using a CCK-8 assay and a colony formation assay. (G to I) Cell proliferation of CDR1as-overexpressing and CDR1as-silenced RBE cells was demonstrated by using a CCK-8 assay and a colony formation assay. Data are presented as the means ± SD. *, P < 0.05; **, P < 0.01; ***, P < 0.001 (versus the corresponding negative control [NC]). OV, overexpression; sh, short hairpin RNA; OD, optical density.

CDR1as promotes migration and invasion of CCA cells.

In order to determine the effects of CDR1as on the migration and invasion of CCA cells, we performed wound-healing migration and transwell invasion assays after CDR1as overexpression and CDR1as knockdown in HCCC-9810 and RBE cells. As shown in Fig. 2A and B, the migration rates of HCCC-9810 and RBE cells were remarkedly increased in CDR1as-overexpressing cells and decreased in CDR1as-silenced cells. The results shown in Fig. 2C and D reveal that the upregulation of CDR1as significantly induced the cell-invasive capacity and that the downregulation of CDR1as suppressed cell invasion in both cell lines.

FIG 2.

Effects of CDR1as on migration and invasion of CCA cell lines. (A and B) HCCC-9810 cells and RBE cells were transfected with CDR1as-OV/NC1 or CDR1as-sh1/sh2/NC2. Cell migration was measured by a wound-healing assay. Bars, 200 μm. (C and D) Invasion of CDR1as-overexpressing and CDR1as-silenced HCCC-9810 cells and RBE cells was measured by a transwell assay. Bars, 100 μm. Data are presented as the means ± SD. *, P < 0.05; **, P < 0.01; ***, P < 0.001 (versus the corresponding NC).

CDR1as activates the AKT3/mTOR signaling pathway in CCA cells.

It is well known that the AKT/mTOR pathway is involved in cell proliferation and migration in many types of cancer cells. In this study, we examined the protein levels of AKT3, mTOR and its downstream targets p70S6K and 4EBP1, and their phosphorylated forms in CDR1as-overexpressing and -knocked-down HCCC-9810 and RBE cells. As shown in Fig. 3A to D, the mRNA levels of AKT3 and mTOR were increased in CDR1as-overexpressing cells and decreased in CDR1as-silenced cells. Importantly, upregulated CDR1as significantly increased the protein levels of AKT3 and mTOR and the activation of the AKT3/mTOR/p70S6K/4EBP1 pathway (Fig. 3E and F). In contrast, the silencing of CDR1as caused decreases in the protein levels of AKT3, mTOR, and phosphorylated forms of AKT3, mTOR, p70S6K, and 4EBP1 in both HCCC-9810 and RBE cells (Fig. 3G and H).

FIG 3.

CDR1as regulates the AKT3/mTOR pathway in CCA cells. (A and B) HCCC-9810 cells and RBE cells were transfected with CDR1as-OV/NC1 or CDR1as-sh1/sh2/NC2. The expression levels of AKT3 in HCCC-9810 or RBE cells were determined by RT-qPCR. (C and D) Expression levels of mTOR in transfected HCCC-9810 or RBE cells were determined by RT-qPCR. (E and F) Levels of AKT3, mTOR, p70S6K, 4EBP1, and their phosphorylated forms in CDR1as-overexpressing HCCC-9810 cells and RBE cells were determined by Western blotting. (G and H) Levels of AKT3, mTOR, p70S6K, 4EBP1, and their phosphorylated forms in CDR1as knockdown HCCC-9810 cells and RBE cells were determined by Western blotting. Data are presented the means ± SD. **, P < 0.01; ***, P < 0.001 (versus the corresponding NC).

miR-641 is a target miRNA of CDR1as in CCA cells.

It is well known that circRNA can bind target functional miRNA and then regulate the miRNA-targeted mRNA. Bioinformatics analysis predicted that miR-641 might be one of the potential target miRNAs of CDR1as. There are more than 10 sites of binding between miR-641 and CDR1as. One of the sites of binding CDR1as and miR-641 is shown in Fig. 4A. Luciferase reporters containing wild-type CDR1as (CDR1as WT) fragments or mutant-type CDR1as (CDR1as MUT) were cotransfected with miR-641 mimics or negative-control (NC) mimics in CCA cells. The results showed that miR-641 significantly decreased the luciferase activity in HCCC-9810 and RBE cells transfected with CDR1as WT fragments (Fig. 4B). In addition, the expression of miR-641 was downregulated in HCCC-9810 and RBE cells compared to that in HIBEpiC cells, which was negatively correlated with CDR1as (Fig. 4C). The results in Fig. 4D and E show that the overexpression of CDR1as inhibited the level of miR-641 in CCA cell lines, whereas the knockdown of CDR1as upregulated the miR-641 level.

FIG 4.

miR-641 is a target miRNA of CDR1as in CCA cells. (A) Predicted binding site between CDR1as and miR-641 and the mutant binding site on CDR1as. (B) Luciferase activities of CDR1as WT (wild type) and CDR1as MUT (mutant type) in HCCC-9810 and RBE cells were evaluated after transfection with miR-641 mimics or NC mimics. (C) Expression levels of miR-641 in HIBEpiC, HCCC-9810, and RBE cells were determined by RT-qPCR. (D and E) Expression levels of miR-641 in HCCC-9810 and RBE cells transfected with CDR1as-OV/NC1 or CDR1as-sh1/sh2/NC2 were determined by RT-qPCR. (F and G) Fold enrichment levels of CDR1as and miR-641 after an RNA immunoprecipitation (RIP) assay with AGO2 antibody or immunoglobulin (IgG) antibody in HCCC-9810 and RBE cells. (H) Relative levels of CDR1as in CCA cells transfected with miR-641 mimics or NC mimics. (I) Fold enrichment levels of CDR1as pulled down by biotin-coupled miR-641 and the NC in HCCC-9810 and RBE cells. (J and K) Fold enrichment levels of CDR1as pulled down by biotin-coupled CDR1as probes or oligonucleotide probes in RBE cells transfected with CDR1as-OV or NC1. (L) Fold enrichment levels of miR-641 pulled down by biotin-coupled CDR1as probes or oligonucleotide probes in HCCC-9810 and RBE cells. Data are presented as the means ± SD. **, P < 0.01; ***, P < 0.001 (versus the corresponding NC or IgG).

Next, direct binding between CDR1as and miR-641 was verified. RNA binding protein immunoprecipitation (RIP) results showed that the levels of CDR1as and miR-641 were significantly increased in the AGO2-containing samples compared with immunoglobulin G (IgG) (Fig. 4F and G). This indicated that CDR1as and miR-641 could be enriched by AGO2 in both cell lines. Additionally, the upregulation of miR-641 could decrease the level of CDR1as (Fig. 4H). Moreover, a pulldown assay showed that biotin-labeled miR-641 captured more CDR1as than biotin-labeled NC miRNA (Fig. 4I), and the biotin-coupled specific CDR1as probe captured and enriched more miR-641 than the oligonucleotide probe in both HCCC-9810 and RBE cells (Fig. 4J to L). These results indicated that CDR1as can bind to miR-641.

CDR1as regulates cell behaviors partially through miR-641 in CCA cells.

We also studied whether the oncogenic properties of CDR1as were related to miR-641 in CCA. The vectors containing CDR1as-OV, miR-641 mimics, CDR1as-OV/miR-641 mimics, or their NCs were transfected into HCCC-9810 cells. The results showed that the upregulation of miR-641 inhibited the cell migration and invasion of HCCC-9810 cells and reduced the increased cell migration and invasion that was caused by the overexpression of CDR1as (Fig. 5A and B). The results in Fig. 5C show that the upregulation of miR-641 significantly suppressed cell proliferation and inhibited the increased cell proliferation that was induced by the overexpression of CDR1as. Additionally, miR-641 mimics significantly abated the increased protein levels of AKT3 and mTOR and the activation of the AKT3/mTOR/p70S6K/4EBP1 pathway that was induced by CDR1as overexpression in RBE cells (Fig. 5D and E).

FIG 5.

Upregulation of miR-641 abates increased migration, invasion, and proliferation in CDR1as-overexpressing CCA cells. (A) HCCC-9810 cells were transfected with CDR1as-OV/NC1 or miR-641/NC mimics or cotransfected with CDR1as-OV and miR-641/NC mimics. Migration was measured by a wound-healing assay. Bar, 200 μm. (B) Invasion was measured by a transwell assay. Bar, 100 μm. (C) Cell proliferation was measured by a CCK-8 assay. (D and E) Levels of AKT3, mTOR, p70S6K, 4EBP1, and their phosphorylated forms were determined by Western blotting. Data are presented as the means ± SD. *, P < 0.05; **, P < 0.01; ***, P < 0.001 (versus the corresponding NC).

In order to confirm our findings, RBE cells were transfected with CDR1as-sh2, miR-641 inhibitor, CDR1as-sh2/miR-641 inhibitor, or NCs. As expected, the miR-641 inhibitor promoted cell migration and invasion. Moreover, the miR-641 inhibitor attenuated the decreased cell migration and invasion caused by CDR1as knockdown (Fig. 6A and B). Moreover, the proliferation of RBE cells, protein levels of AKT3 and mTOR, and phosphorylation levels of AKT3, mTOR, p70S6K, and 4EBP1 were increased in miR-641 inhibitor-transfected cells (Fig. 6C to E). Our results suggested that the downregulation of miR-641 at least partially reversed the effects of CDR1as-sh2 on cell proliferation and the activation of the AKT3/mTOR/p70S6K/4EBP1 pathway.

FIG 6.

Downregulation of miR-641 attenuates decreased migration, invasion, and proliferation in CDR1as-silenced CCA cells. (A) RBE cells were transfected with CDR1as-sh2/NC2 or the miR-641/NC inhibitor or cotransfected with CDR1as-sh2 and the miR-641/NC inhibitor. Migration was measured by a wound-healing assay. Bar, 200 μm. (B) Invasion was measured by a transwell assay. Bar, 100 μm. (C) Cell proliferation was measured by a CCK-8 assay. (D and E) Levels of AKT3, mTOR, p70S6K, 4EBP1, and their phosphorylated forms were determined by Western blotting. Data are presented as the means ± SD. *, P < 0.05; **, P < 0.01; ***, P < 0.001 (versus the corresponding NC).

The AKT3 and mTOR genes are target genes of miR-641 in CCA cells.

Our results showed that CDR1as and miR-641 can regulate AKT3 and mTOR expression in CCA cells. This made us wonder whether the miR-641/AKT3/mTOR axis is related to the role of CDR1as in cell behaviors in CCA cells. According to the bioinformatics analysis, the AKT3 and mTOR genes were potential target genes of miR-641. The predicted binding sites of the wild-type AKT3 and mTOR 3′ untranslated regions (UTRs) for miR-641 are displayed in Fig. 7A and C. Next, a dual-luciferase reporter assay demonstrated that miR-641 decreased the luciferase activity of AKT3- and mTOR 3′-UTR WT-transfected cells but not of 3′-UTR MUT-transfected cells (Fig. 7B and D). In our experiments, we chose AKT3 for further study. As shown in Fig. 8A, a pulldown assay showed that biotin-labeled miR-641 captured more AKT3 than that in the control group. The results demonstrated that miR-641 can directly bind to AKT3. In a rescue assay, two small interfering RNAs (siRNAs) against AKT3 were synthesized, and the efficiencies of siRNAs were assessed by the measurement of AKT3 mRNA levels in RBE cells (Fig. 8B). Next, AKT3 siRNA-1 was used for the rescue assay. RBE cells were cotransfected with the miR-641 inhibitor and AKT3 siRNA-1. Figure 8C shows that the downregulation of miR-641 enhanced the protein level of AKT3, and it could be abated by AKT3 knockdown. Moreover, the silencing of AKT3 significantly attenuated the inhibited effects of miR-641 on cell proliferation, migration, and invasion of RBE cells (Fig. 8D to F).

FIG 7.

The AKT3 and mTOR genes are target genes of miR-641 in CCA cells. (A) Predicted site of binding between miR-641 and the 3′ UTR of AKT3 and the mutant binding site on the 3′ UTR of AKT3. (B) Luciferase activities of the 3′ UTRs of AKT3 WT and MUT in HCCC-9810 and RBE cells were evaluated after transfection with miR-641 mimics or NC mimics. (C) Predicted site of binding between miR-641 and the 3′ UTR of mTOR and the mutant binding site on the 3′ UTR of mTOR. (D) Luciferase activities of the 3′ UTRs of mTOR WT and MUT in HCCC-9810 and RBE cells were evaluated after transfection with miR-641 mimics or NC mimics. Data are presented as the means ± SD. *, P < 0.05; **, P < 0.01 (versus the corresponding NC).

FIG 8.

Knockdown of AKT3 abates the effects of miR-641 inhibitors on CCA cells. (A) Fold enrichment levels of AKT3 pulled down by biotin-coupled miR-641 and NC in HCCC-9810 and RBE cells. (B) RBE cells were transfected with AKT3 siRNA-1 or siRNA-2 or NC siRNA for 24 h, and the relative mRNA levels of AKT3 were measured by RT-qPCR. (C) Protein levels of AKT3 were determined by Western blotting. (D) Cell viability was measured by a CCK-8 assay. (E) Migration was measured by a wound-healing assay. Bar, 200 μm. (F) Invasion was measured by a transwell assay. Bar, 100 μm. Data are presented as the means ± SD. *, P < 0.05; **, P < 0.01 (versus the corresponding NC).

In order to fully understand the potential underlying regulatory mechanisms among CDR1as, miR-641, and AKT3/mTOR, a transcription inhibitor, actinomycin D (Act D), was added to RBE cells transfected with CDR1as-OV, CDR1as-sh2, miR-641 mimics, or NC controls to block new RNA transcription. Therefore, the remaining levels of CDR1as, miR-641, and mRNA of AKT3/mTOR were related to RNA degradation. First, the results in Fig. 9A to C show that CDR1as accelerated miR-641 degradation and increased the stability of the AKT3 and mTOR transcripts in CDR1as-overexpressing RBE cells in 24 h. Expectedly, in RBE cells transfected with CDR1as-sh2, the remaining level of miR-641 was increased, and the levels of mRNA of AKT3 and mTOR were decreased (Fig. 9D to F). Moreover, the remaining level of CDR1as in cells transfected with miR-641 mimics was significantly lower than that in cells transfected with the NC mimics in 24 h, and miR-641 significantly decreased the stability of mRNAs of AKT3 and mTOR (Fig. 9G to I). These results suggested that CDR1as could accelerate miR-641 degradation. In contrast, miR-641 could degrade CDR1as and mRNAs of AKT3 and mTOR in RBE cells.

FIG 9.

Regulatory mechanisms among CDR1as, miR-641, and AKT3/mTOR. RBE cells were transfected with CDR1as-OV, CDR1as-sh2, miR-641 mimics, or their NCs in the presence of 20 μg/ml actinomycin D for the times indicated. (A to C) Relative levels of miR-641 (A) and mRNAs of AKT3 (B) and mTOR (C) in RBE cells transfected with CDR1as-OV or NC1 were measured by RT-qPCR. (D to F) Relative levels of miR-641 (D) and mRNAs of AKT3 (E) and mTOR (F) in RBE cells transfected with CDR1as-sh2 or NC2 were measured by RT-qPCR. (G to I) Remaining levels of CDR1as (G) and mRNAs of AKT3 (H) and mTOR (I) in RBE cells transfected with miR-641 mimics or NC mimics were measured by RT-qPCR. Data are presented as the means ± SD. *, P < 0.05; **, P < 0.01 (versus the corresponding NC group).

CDR1as promotes CCA tumor growth partially through the miR-641/AKT3 axis.

In order to assess whether CDR1as regulates the growth of CCA tumors in vivo, we stably transfected CDR1as-OV, NC1, CDR1as-sh2, or NC2 into RBE cells and then injected the cells into nude mice. After tumor formation, the miR-641 agomir, the miR-641 antagomir, or their NCs were respectively injected into the mice every 3 days by tail vein injection, and the tumor sizes were measured every 3 days. We found that the upregulation of CDR1as promoted tumor growth, while the knockdown of CDR1as inhibited tumor growth, in a mouse xenograft model (Fig. 10A to D). In addition, the results of immunohistochemical assays and Western blotting showed that AKT3 and phosphorylated AKT3 (p-AKT3) levels were elevated in CDR1as-overexpressing tumor tissues, whereas they were decreased in CDR1as knockdown tumor tissues (Fig. 10E to J). We also found that the effects of CDR1as on tumor growth and AKT3 expression in tumor tissues were partially attenuated by treatment with the miR-641 agomir or miR-641 antagomir. The results indicated that CDR1as could promote CCA tumor growth and that the miR-641/AKT3 axis is possibly one of the axes related to CDR1as’ functions in vivo.

FIG 10.

CDR1as promotes CCA tumor growth partially through the miR-641/AKT3 axis. CDR1as-OV/NC1- or CDR1as-sh2/NC2-transfected RBE cells were injected into BALB/c nude mice to establish subcutaneous xenograft tumors. Next, the miR-641/NC agomir or miR-641/NC antagomir was injected into mice every 3 days after tumor formation by tail vein injection. The mice were sacrificed at 28 days. (A to D) Xenograft tumors were photographed, and the tumor volumes were also measured from day 7 to day 28. (E to J) Levels of AKT3 and p-AKT3 in xenograft tumor tissues were determined by immunohistochemical staining and Western blotting. Data are presented as the means ± SD. *, P < 0.05; **, P < 0.01; #, P < 0.05; ##, P < 0.01 (versus the corresponding NC group).

DISCUSSION

In this present study, we aimed to uncover the functions of CDR1as in CCA. We performed gain- and loss-of-function assays to assess the effects of CDR1as on the proliferation, migration, and invasion of CCA cells and tumor xenograft growth in a mouse model. Moreover, we investigated whether the miR-641/AKT3 axis is one of the regulatory axes regulated by CDR1as in CCA.

Previous studies found that the expression levels of CDR1as were significantly higher in cancer tissues than in paracancerous tissues such as hepatocellular carcinoma, colorectal cancer, non-small-cell lung cancer, esophageal squamous cell carcinoma, and pancreatic cancer (14, 21, 29–31). Mechanically, CDR1as could inhibit the expression of miR-7 to positively regulate miR-7- target genes, including those for epidermal growth factor receptor (EGFR), CCNE1, PIK3CD, and Krüppel-like factor 4 (KLF4), to promote cell proliferation and invasion and tumor growth (14, 17, 20, 30). CDR1as also acts as a sponge of miR-876-5p to accelerate esophageal squamous cell carcinoma progression (32). However, CDR1as plays a role as a tumor suppressor through inhibiting miR-135b-5p expression and increasing the expression of hypoxia-inducible factor 1-alpha inhibitor (HIF1AN) in ovarian cancer (33). In CCA, the expression of CDR1as in tumor tissues was upregulated compared to that in adjacent normal tissues (22). However, the functions of CDR1as in CCA remain unclear. Our results showed that the expression of CDR1as was increased in the human CCA cell lines HCCC-9810 and RBE. The overexpression of CDR1as could promote cell proliferation, migration, and invasion in both HCCC-9810 and RBE cells in vitro and tumor xenograft growth in vivo, whereas it showed the opposite trends when CDR1as was knocked down.

Our results showed that the expression of CDR1as was positively correlated with the protein levels of AKT3 and mTOR. It is known that CDR1as can mediate gene functions through regulating miRNA. Therefore, we tried to find an miRNA that is correlated with CDR1as, AKT3, and mTOR in CCA. miR-641, our target miRNA, has been extensively reported to be a tumor suppressor in many kinds of cancer through regulating its downstream target genes, such as those for murine double minute 2 (MDM2), neurofibromatosis 1 (NF1), and zinc finger E-box binding homeobox 1 (ZEB1) (23, 34, 35). Additionally, several long noncoding RNAs (lncRNAs), including lncRNA OIP5-AS1, lncRNA CRNDE, and lncRNA DLX6-AS1, promote cell proliferation and induce cell apoptosis through regulating miR-641 in gastric cancer, lung cancer, and osteosarcoma (24, 25, 36). However, it was found that lncRNA TUSC8 could inhibit invasion and migration through the miR-641/PTEN pathway in cervical cancer cells (37). In this study, the results showed that miR-641 could attenuate the changes in proliferation, migration, and invasion of CCA cells that are induced by CDR1as.

Furthermore, miR-641 was found to directly bind with the 3′ UTRs of AKT3 and mTOR. AKT has three isoforms in cells: AKT1, AKT2, and AKT3. These isoforms play distinct roles in different types of tumor progression. For instance, the silencing of AKT2 and AKT3, but not AKT1, can inhibit cell migration and invasion in human lung adenocarcinoma A549 cells (38), and the downregulation of AKT1 suppresses cell migration, invasion, and proliferation through the phosphatidylinositol 3-kinase (PI3K)/AKT/mTOR/p70S6K1 pathway in ovarian cancer cells (39). However, AKT1 and AKT2, but not AKT3, function as inhibitors of cell migration and invasion in prostate cancer cells (40). Mostly, studies have found that AKT3 promotes cancer growth and metastasis through the PI3K/AKT pathway, such as in thyroid cancer, cervical cancer, and uveal melanoma (41–43). Growing evidence proved that the activation of the AKT/mTOR/p70S6K/4EBP1 pathway promotes cell proliferation, migration, and invasion and inhibits apoptosis in leukemia PVTL-1 cells, renal cancer 786-O cells, and meningioma CH-157MN cells (44–46). Therefore, in this study, we assessed the effects of CDR1as on the expression of AKT3 and whether it was related to miR-641. Our results showed that miR-641 was negatively related to the expression of AKT3, and the knockdown of AKT3 could abate the inhibitory effects of miR-641 on CCA cells. Additionally, mTOR is an important activator of tumorigenesis and drug resistance in biliary tract cancers, and the PI3K/AKT/mTOR pathway positively regulates the proliferation of CCA cells (28, 47). Similarly, our results showed that mTOR could be positively regulated by CDR1as and negatively regulated by miR-641. Our results showed that miR-641 could directly bind CDR1as, AKT3, and mTOR (Fig. 4 and 7). Moreover, in the presence of Act D, CDR1as was negatively correlated with miR-641 degradation and mediated the stability of mRNAs of AKT3 and mTOR in RBE cells (Fig. 9A to F). Mechanistically, the miR-641 decay induced by CDR1as might be related to the occurrence of a target-directed miRNA degradation (TDMD) mechanism (48, 49). In the process of TDMD, specific RNAs can bind to target miRNAs and promote the degradation of miRNAs (50). Moreover, miR-641 also degraded CDR1as and downregulated the mRNA levels of AKT3 and mTOR in 24 h (Fig. 9G to I). Our findings demonstrated that CDR1as could bind to miR-641 and degrade miR-641, and it very likely will decrease the interaction between miR-641 and its target mRNAs, possibly leading to an improvement in the stability of AKT3 and mTOR in RBE cells. Therefore, our results indicated that CDR1as can promote the cell proliferation, migration, and invasion of CCA cells, and the miR-641/AKT3/mTOR axis might be one of regulatory axes regulated by CDR1as in CCA.

In conclusion, our results suggested that the CDR1as gene may play a role as an oncogene in CCA, at least partially, through miR-641, which indicates that CDR1as and miR-641 could be considered prognostic biomarkers of CCA, and the CDR1as/miR-641 axis might be a novel therapeutic target for CCA.

MATERIALS AND METHODS

Cell culture and transfection.

The cholangiocyte cell line HIBEpiC and two human CCA cell lines, HCCC-9810 and RBE, were obtained from Zhongqiaoxinzhou (Shanghai, China). HIBEpiC cells were cultured in endothelial cell medium (ECM; Zhongqiaoxinzhou), and HCCC-9810 and RBE cells were cultured in RPMI 1640 medium (Gibco, Grand Island, NY) supplemented with 10% fetal bovine serum (FBS; HyClone, South Logan, UT) at 37°C in a humidified atmosphere with 5% CO₂.

The vector containing CDR1as (CDR1as-OV) was purchased from Geneseed Biotech Co. Ltd. (Guangzhou, China), and the empty vector was negative control 1 (NC1). The vectors containing short hairpin RNA 1 (shRNA-1) for CDR1as (CDR1as-sh1) or short hairpin RNA-2 for CDR1as (CDR1as-sh2) (Table 1) were purchased from Wanleibio (Shenyang, China), and the empty vector was negative control 2 (NC2). The miR-641 mimic, the miR-641 inhibitor, siRNAs aimed at AKT3, and their negative controls (Table 1) were synthesized by JTS (Wuhan, China). Cell transfection was then performed using Lipofectamine 2000 (Invitrogen, Camarillo, CA) according to the manufacturer’s instructions. The transfected cells were incubated for 24 h, and the experiments and analyses were then performed. For stable transfection, the HCCC-9810 cells were selected using 300 μg/ml G418 (Invitrogen), and RBE cells were selected using 400 μg/ml G418.

TABLE 1.

RNA sequences of oligonucleotides

Oligonucleotide	Sequence(s) (5′–3′)a
CDR1as shRNA-1	gatccgGCCATCGGAAACCCTGGATttcaagagaATCCAGGGTTTCCGATGGCttttta, agcttaaaaaGCCATCGGAAACCCTGGATtctcttgaaATCCAGGGTTTCCGATGGCcg
CDR1as shRNA-2	gatccgGAAACCCTGGATATTGCAGttcaagagaGAAACCCTGGATATTGCAGttttta, agcttaaaaaGAAACCCTGGATATTGCAGtctcttgaaGAAACCCTGGATATTGCAGcg
NC shRNA (NC2)	gatccgCTTTGGCTGGATATTGCAGttcaagagaCTGCAATATCCAGCCAAAGttttta, agcttaaaaaCTTTGGCTGGATATTGCAGtctcttgaaCTGCAATATCCAGCCAAAGcg
AKT3 siRNA-1	GGAUGCCUCUACAACCCAUTT, AUGGGUUGUAGAGGCAUCCTT
AKT3 siRNA-2	GGAGGUUACCUUUCUACAATT, UUGUAGAAAGGUAACCUCCTT
NC siRNA	UUCUCCGAACGUGUCACGUTT, ACGUGACACGUUCGGAGAATT
miR-641 mimic	AAAGACAUAGGAUAGAGUCACCUC, GGUGACUCUAUCCUAUGUCUUUUU
NC mimic	UUCUCCGAACGUGUCACGUTT, ACGUGACACGUUCGGAGAATT
miR-641 inhibitor	GAGGUGACUCUAUCCUAUGUCUUU
NC inhibitor	UUGUACUACACAAAAGUACUG
miR-641 probe	AAAGACAUAGGAUAGAGUCACCUC-biotin
miR-NC probe	UUCUCCGAACGUGUCACGUTT-biotin
Oligonucleotide probe	TGTCTGCAATATCCAGGGTTTCCGATGGCACC-biotin
CDR1as probe	GGTGCCATCGGAAACCCTGGATATTGCAGACA-biotin

^aLowercase type indicates the restriction site sequence and sequence connecting the target sequence and the complementary sequence. For oligonucleotides with two sequences, the sense sequences is given first, followed by the antisense sequence.

Actinomycin D assay.

RBE cells were transiently transfected with CDR1as-OV, CDR1as-sh2, miR-641 mimics, or controls in the presence of 20 μg/ml actinomycin D (Act D) (catalog number HY-17559; MCE, Monmouth Junction, NJ) for the indicated times (0, 4, 8, and 24 h). Act D, an inhibitor of DNA-dependent RNA synthesis, was used to assess the stability and degradation of miRNA and mRNA (51–53). The remaining levels of miRNA and mRNA were related to RNA degradation for the indicated times.

Cell Counting Kit-8 assay.

Cell proliferation and cell viability were analyzed using the Cell Counting Kit-8 (catalog number KGA317; KeyGen Biotech, Nanjing, China) reagent. Briefly, after transfection, cells were seeded into 96-well plates at a density of 5 × 10³ cells/well. After being incubated for the indicated times, the cells were added to 10 μl of the CCK-8 solution. Two hours after incubation at 37°C, the absorbance of each well was measured at 450 nm using a microplate reader (ELX-800; BioTek, Winooski, VT).

Colony formation assay.

The transfected cells were seeded at a density of 1 × 10³ cells per well in a 35-mm culture plate and then cultured for about 2 weeks in complete medium. The cells were fixed with 4% paraformaldehyde (Sinopharm, Shanghai, China) for 20 min at room temperature and then stained with a Wright-Giemsa stain kit (catalog number D010; Nanjing Jiancheng Bioengineering Institute, Nanjing, China) for 5 min. The stained cell colonies were counted using a microscope (IX53; Olympus, Tokyo, Japan) and photographed using an imaging system (DP73; Olympus).

Wound-healing assay.

The transfected cells were cultured in 1 μg/ml mitomycin C serum-free medium (catalog number M0503; Sigma, St. Louis, MO) for 1 h and then scratched using a sterile 200-μl plastic tip. The images of wounds were measured and photographed by using a microscope 0 and 24 h after scratching.

Transwell assay.

A transwell invasion assay was performed using 24-well chemotaxis Matrigel chambers (8 μm, catalog number 3422; Corning, NY). Briefly, the transfected cells (1 × 10⁴) in 200 μl serum-free medium were added to the upper chambers, which were precoated with diluted Matrigel. The lower chambers were filled with 800 μl of medium containing 10% FBS to induce cells to invade through the membrane. After incubation for 24 h, the chambers were washed with phosphate-buffered saline (PBS) twice. The invasive cells that passed through the membrane into the lower chambers were fixed with 4% paraformaldehyde for 25 min and stained with 0.4% crystal violet (catalog number 0528; Amresco, Solon, OH) for 5 min. The cells were counted using a microscope in five random fields and photographed.

RNA extraction and real-time quantitative PCR.

The total RNA of cell lines and tumor tissues was isolated by using the RNAsimple total RNA kit (catalog number DP419; Tiangen, Beijing, China) with an RNase inhibitor (catalog number DP418; Tiangen) according to the manufacturer’s instructions. The RNA concentration was measured and used to synthesize cDNA using TIANSeq Moloney murine leukemia virus (M-MLV) (RNase H-minus) reverse transcriptase (catalog number NG212; Tiangen). The RT-qPCR experiments were conducted with SYBR (catalog number SY1020; Solarbio, Beijing, China) and 2× Taq PCR master mix (catalog number KT201; Tiangen). U6 small nuclear RNA (snRNA) was used as an internal control for miR-641 and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was used as an internal control for mRNA and CDR1as, respectively. The sequences of primers for RT-qPCR are presented in Table 2. Relative expression levels were analyzed by the $2^{-\Delta \Delta C_T}$ method.

TABLE 2.

Sequences of primers for RT-qPCR

Primer	Sequences (forward, reverse; 5′–3′)
miR-641	AAAGACATAGGATAGAGTCACCT, GCAGGGTCCGAGGTATTC
U6	GCTTCGGCAGCACATATACT, GGTGCAGGGTCCGAGGTAT
CDR1as	CCATCAACTGGCTCAATATCC, CACAGGTGCCATCGGAAA
AKT3	AAAACAGAACGACCAAAG, TCTGCTACAGCCTGGATA
mTOR	GCTGTCATCCCTTTATCG, TCTTCTTCTTCTCCCTGTAGTC
GAPDH	GACCTGACCTGCCGTCTAG, AGGAGTGGGTGTCGCTGT

Western blot analysis.

Transfected cells were collected and lysed using radioimmunoprecipitation assay (RIPA) lysis buffer (Solarbio) with 1 mM phenylmethylsulfonyl fluoride (PMSF) (Solarbio). After protein concentration measurement, an equal amount of protein was separated by SDS-PAGE and then transferred onto polyvinylidene difluoride (PDVF) membranes (Millipore, Billerica, MA). The membranes were blocked with 5% nonfat milk (catalog number A600669; Sangon Biotech, Shanghai, China) or a 5% bovine serum albumin (BSA) solution (catalog number BS043; Biosharp, Hefei, China) for 1 h at room temperature. The membranes were then incubated at 4°C overnight with the following antibodies: anti-p-AKT3 (catalog number ab192623; Abcam, Cambridge, UK), anti-AKT3 (catalog number 14982; CST, Danvers, MA), anti-p-mTOR (catalog number 5536; CST), anti-mTOR (catalog number 2972; CST), anti-p-p70S6K (catalog number 9234; CST), anti-p70S6K (catalog number 9202; CST), anti-p-4EBP1 (catalog number 2855; CST), anti-4EBP1 (catalog number 9644; CST), and anti-GAPDH (catalog number 60004-1-Ig; Proteintech, Wuhan, China). After washing, the membranes were then incubated with horseradish peroxidase-conjugated secondary antibodies (Solarbio) for 1 h at 37°C. The protein bands were visualized by using an electrochemiluminescence (ECL) solution (catalog number PE0010; Solarbio), and the relative protein levels were determined by normalization to GAPDH levels.

RNA binding protein immunoprecipitation.

Immunoprecipitation of CDR1as and miR-641 bound to AGO2 was performed using an EZ-Magna RNA binding protein immunoprecipitation (RIP) kit (catalog number 17-701; Merck Millipore, Germany). HCCC-9810 and RBE cells were collected and lysed using RIP lysis buffer. The lysates were centrifuged at 14,000 rpm for 10 min, and the supernatant was incubated with magnetic beads coated with 5 μg of anti-AGO2 (Proteintech) or normal control anti-IgG (Wanleibio) at 4°C overnight. The RNA was extracted, and the enrichment levels of CDR1as and miR-641 were determined by RT-qPCR.

Biotinylated RNA pulldown assay.

Biotin-coupled miR-641 and CDR1as pulldown assays were performed using a Pierce magnetic RNA-protein pulldown kit (catalog number 20164; Thermo Fisher Scientific, Waltham, MA). The biotin-labeled miRNA-641 mimics, miRNA NC mimics, CDR1as probe, or NC oligonucleotide probe (GenScript, Nanjing, China) (Table 1) was labeled by using a Thermo Scientific Pierce RNA 3′ desthiobiotinylation kit. HCCC-9810 and RBE cells were collected and lysed using Thermo Scientific Pierce immunoprecipitation lysis buffer. The biotin-labeled oligonucleotides were each incubated with streptavidin magnetic beads for 30 min at room temperature with agitation. After washing, the RNA-bound beads were mixed with the cell lysates in protein-RNA binding buffer and incubated for 60 min at 4°C with agitation. The beads were then washed, and the RNAs were extracted by using TRIpure (catalog number RP1001; BioTeke, Beijing, China). The enrichment levels of CDR1as and AKT3 were determined by RT-qPCR.

Dual-luciferase reporter assay.

The fragments of AKT3 3′ UTR, mTOR 3′ UTR, and CDR1as containing the specific miR-641 wild-type and mutant-type sequences were amplified by PCR and inserted into the reporter luciferase vector pmirGLO by GenScript. HCCC-9810 and RBE cells were seeded into 12-well plates and grown to 70% confluence. The cells were then cotransfected with either miR-641 mimics or NC mimics with a firefly luciferase vector containing target sequences using Lipofectamine 2000. After 48 h of cell transfection, luciferase activities were detected using a dual-luciferase reporter assay kit (catalog number KGAF040; KeyGen Biotech) according to the manufacturer’s instructions. The relative luciferase activity was normalized to the renilla luciferase activity.

Xenograft mouse model.

Nude mice (BALB/c nude; 6 weeks old and ∼18 to 20 g) were purchased from HFK Bioscience (Beijing, China) and kept under pathogen-free conditions. The mice were randomly divided into 8 groups with 6 mice per group. RBE cells (5 × 10⁶ cells/100 μl) stably transfected with CDR1as-OV, NC1, CDR1as-sh2, or NC2 were subcutaneously injected into the backs of the nude mice. After tumor formation, miR-641, the NC agomir, miR-641, or the NC antagomir was injected into the appropriate mice mentioned above every 3 days by tail vein injection. The tumor sizes were measured every 3 days. The mice were sacrificed, and the tumor nodes were separated to measure tumor size on day 28 after cell injection. The tumor tissues were fixed or kept in liquid nitrogen for analysis. All animal studies were approved by the Ethics Committee of the First Affiliated Hospital of Zhengzhou University, and animal experiments were performed according to guidelines for the care and use of laboratory animals (54).

Immunohistochemistry.

To determine the levels of AKT3 and p-AKT3 in CCA tumors, tumor tissues of mice were evaluated using immunohistochemistry (IHC). The tissues were fixed, paraffined, deparaffinized, rehydrated, and incubated with a 3% H₂O₂ solution (Sinopharm) for 15 min at room temperature. The sections were blocked with goat serum (Solarbio) for 15 min at room temperature and then incubated with AKT3 (catalog number 10176-2-AP; Proteintech) and p-AKT3 (catalog number Bs-15963R; Bioss, Woburn, MA) primary antibodies at 4°C overnight. After washing with PBS, the sections were incubated with the secondary antibody (catalog number 31460; Thermo Fisher Scientific) for 1 h at 37°C. After washing, the sections were added to a diaminobenzidine (DAB) solution (catalog number DA1010; Solarbio) and counterstained with hematoxylin (catalog number H8070; Solarbio). After dehydration and mounting, the stained sections were observed by using a microscope, and the images were photographed.

Statistical analysis.

All data were analyzed using GraphPad Prism 8.0 (GraphPad Software Inc., San Diego, CA) and are presented as the means ± standard deviations (SD). The data were analyzed using Student’s t test and analysis of variance (ANOVA). A P value of <0.05 was considered statistically significant.

Data availability.

The data that support the findings of this study are fully available from the corresponding author upon reasonable request.