Hsa_circ_0063329 inhibits prostate cancer growth and metastasis by modulating the miR-605-5p/tgif2 axis

ABSTRACT

Circular RNAs play crucial regulatory roles in the progression of various cancers. Nevertheless, the underlying mechanisms of circRNAs in prostate cancer (PCa) proliferation and metastasis remain largely uncertain. Here, we performed circRNA microarray analyses to identify differentially expressed circRNAs in a normal prostate epithelial cell line and PCa cell lines. We found that hsa_circ_0063329 was significantly downregulated in PCa. A series of in vitro and in vivo functional assays showed that overexpression of hsa_circ_0063329 inhibits PCa cell progression, while silencing of hsa_circ_0063329 achieved the opposite effects. Mechanistically, bioinformatics analysis, RNA pulldown, RNA-seq and dual-luciferase reporter assays demonstrated that hsa_circ_0063329 exerts its effect by sponging miR-605-5p to derepress TG-interacting factor 2 (TGIF2) and inactivate the TGF-β pathway. In conclusion, hsa_circ_0063329 inhibits the proliferation and metastasis of PCa via modulation of the miR-605-5p/TGIF2 axis, and targeting hsa_circ_0063329 may provide a promising treatment strategy for aggressive PCa.

Introduction

Prostate cancer (PCa) is the second leading cause of cancer death among men in the United States [1]. In China, the incidence of prostate cancer has gradually increased in recent years [2]. Despite advances in comprehensive treatment, the prognosis of PCa sufferers remains poor because of recurrence and metastasis, and most cases inevitably develop castration-resistant prostate cancer (CRPC), the fatal stage of PCa [3,4]. Thus, finding biomarkers and potential therapeutic targets for PCa patients is urgently needed.

As a unique set of noncoding transcripts, circular RNAs (circRNAs) are linked by the back-splicing of introns or exons of pre-mRNAs [5,6]. Recently, studies have demonstrated the crucial roles of circRNAs in processes, such as protein translation, transcription, and splicing modification, as well as miRNA and RNA-binding protein sponging [7,8]. CircRNAs have been reported to outperform their linear transcripts as diagnostic and therapeutic targets due to their species conservation, structural stability, and tissue specificity [9–11]. An increasing number of studies have reported that circRNAs have a crucial function in tumor progression [7,12], including PCa progression [13,14]. For instance, circSCAF8 has been reported to promote PCa proliferation and metastasis by sponging miR-140-3p and miR-335 and then activating the LIF-STAT3 pathway [15]. CircPDE5A was found to be downregulated in human PCa tissue, and it can disrupt the translation of EIF3C by forming the circPDE5A-WTAP complex, which then inactivates the MAPK pathway, thereby inhibiting PCa progression [16]. Our previous study showed that circNOLC1 is significantly overexpressed in PCa; it acts as a sponge for miR-647 and then activates the AKT signaling pathway, thereby promoting PCa progression [17]. All these results suggest a crucial role for circRNAs in PCa development. However, there are still a large number of circRNAs that are involved in PCa progression. Hence, the aim of this study was to identify novel circRNAs in PCa.

In the current study, circRNA microarrays were applied to investigate the profiles of circRNAs in PCa cell lines and immortal prostate epithelial cell lines. We uncovered a new candidate circRNA termed hsa_circ_0063329, which was markedly downregulated in PCa cell lines and tissues. Then, we revealed the antitumour role of hsa_circ_0063329 in the progression of PCa. Further research demonstrated that hsa_circ_0063329 sponged miR-605-5p to increase the translation of TG-interacting factor 2 (TGIF2), which is a negative regulatory protein in the progression of PCa. In summary, hsa_circ_0063329 suppressed growth and metastasis in PCa via the hsa_circ_0063329/miR-605-5p/TGIF2 signaling pathway; thus, it is a potential therapeutic candidate for PCa.

Materials and methods

Patient tissue samples

A total of 90 PCa tissues were obtained from patients experiencing radical prostatectomy at the First Affiliated Hospital of Guangzhou Medical University (China), and 20 tissues from patients diagnosed with BPH (benign prostatic hyperplasia) were used as controls. The detailed clinical information is listed in Table 1. The Human Ethics Committee of the Third Affiliated Hospital of Guangzhou Medical University and the First Affiliated Hospital of Guangzhou Medical University agreed with this research protocol. In addition, all PCa patients approved and provided written informed consent.

Table 1.

The correlation between hsa_circ_0063329 expression and clinicopathological characteristics was analyzed in PCa by FISH (n = 90).

Variables	Total N	circDDX17		χ2	p valueb
		High expression (n, %)	Low expression (n, %)
Type
BPH	20	11 (55.0)	9 (45.0)	11.283	0.001
PCa	90	15 (16.7)	75 (83.3)
Age (years)
≤70a	47	5 (10.6)	42 (89.4)	2.574	0.109
>70	43	10 (23.3)	33 (76.7)
ISUP Grade
1–2	28	3 (10.7)	25 (89.3)	0.508	0.476
3–5	72	12 (19.4)	50 (80.6)
T stage
T1-T2	59	14 (23.7)	45 (76.3)	6.151	0.013
T3-T4	31	1 (3.2)	30 (96.8)

a: Mean age. b: p value is from χ²-test.

Cell lines

PCa cell lines (PC-3, 22Rv1, DU145, LNCaP and C4–2) were obtained from the Stem Cell Bank, Chinese Academy of Sciences. The human prostatic epithelial cell line (RWPE-1) was obtained from ATCC (the American Type Culture Collection). PCa cells were maintained in RPMI-1640 medium containing 10% FBS (fetal bovine serum) (GIBCO, BRL). EGF (epidermal growth factor, 5 ng/mL) in keratinocyte serum-free medium (KSFM) (Gibco, No. 10450–013) was used to culture RWPE-1 cells. ALL cell lines were kept at 37°C with 5% CO₂.

CircRNA microarrays

The circRNAs from PCa cell lines and control RWPE-1 cells were obtained with TRIzol (Invitrogen, USA), and microarray analysis was performed as described in our previous reports [17]. Briefly, Arraystar Human circRNA Array v2 (Kangcheng Biotech, Shanghai, China) was applied to analysis circRNA microarray. Total RNA from each sample was quantified with the NanoDrop ND-1000. Sample preparation and microarray hybridization were performed as outlined in the standard protocols stipulated by Arraystar. Normalized Intensity of each group (averaged normalized intensities of replicate samples, log2 transformed) were analyzed by paired t-test (P: 0.05). Quantile normalization and subsequent data processing was performed through the R 4.0.2 software limma package. Hierarchical Clustering was used to perform the distinguishable circRNAs expression pattern among the samples. The differentially expressed circRNAs with p < 0.05 and absolute fold changes>1.5 were considered significant.

Reverse transcription-quantitative polymerase chain reaction (Rt‒qPCR) and RNase R treatment

TRIzol reagent was used to isolate the total RNA of cells and tissues. Complementary DNA was generated utilizing the Prime Script RT Reagent Kit (TaKaRa). The MiRNA First Strand cDNA Synthesis Poly A Tailing Kit (iGene, China) was used for reverse transcription of miRNAs. TB Green PCR Master Mix (TaKaRa) was used to perform quantitative real-time PCR. Data were quantified in comparison to GAPDH as a control. RNase R (Epicenter, Madison, WI) at 3 U/μg was mixed with total RNA for 15 min at 37°C, and the stability of hsa_circ_0063329 and DDX17 mRNA was assessed using RT‒qPCR. The primers used in this study were purchased from Tsingke Biotechnology Co., Ltd. and are shown in Supplementary Table S1.

Fluorescence in situ hybridization (FISH)

18S probes and Cy3-labeled hsa_circ_0063329 probes were obtained from RiboBio Co., Ltd. (Guangzhou, China). Alexa 488-labeled miR-605-5p probes were produced by Focofish Co., Ltd. (Guangzhou, China). The experiment was performed with a FISH Kit obtained from RiboBio following the manufacturer’s instructions. Then, images were visualized by laser scanning confocal microscopy (LSM 880, Germany).

Nuclear-cytoplasmic fractionation

Cytoplasmic and nuclear RNA was separated utilizing the PARIS™ Kit (Invitrogen, USA) according to the manufacturer’s instructions. In short, the cells were centrifuged at 4°C to separate the cytoplasmic fraction from the pellet containing the nuclear fraction. Subsequently, the pellet was rewashed with cell fraction buffer and lysed with disruption buffer. Next, the cytoplasmic and nuclear fractions were mixed with 2× Lysis/Binding Solution, and then 100% ethanol was added. Nuclear and cytoplasmic RNAs were collected with the eluate after centrifugation and washing.

Cell transfection

The siRNAs (small interfering RNAs) of hsa_circ_0063329 (The sequence is shown in Supplementary Table S1), miR-605-5p inhibitor and mimics were produced by RiboBio. The plasmids of hsa_circ_0063329 and the corresponding empty vector were obtained from GeneSeed Co., Ltd. (Guangzhou, China). Lipofectamine® 3000 Reagent (Invitrogen, USA) was used to perform the transfection assay. Stable cell lines expressing OE-hsa_circ_0063329 or sh-hsa_circ_0063329 were constructed using lentiviral vectors (GeneChem Bio-Medical Biotechnology, China) based on the manufacturer’s protocol.

RNA sequencing

RNA-seq (RNA sequencing) of LNCaP-vector and OE-hsa_circ_0063329 cells were conducted by LC-Bio Co., Ltd. (Hangzhou, China). The heatmap and gene set enrichment analyses were performed by the BGI Omicstudio system.

Wound-healing assay

PCa cells (2×10⁵) were seeded in a six-well plate, and after growing to 90% confluence, each well was scraped using a 10 μL plastic pipette tip. Then, cell debris was removed with PBS, the cells were further cultured in 6-well plates in FBS-free medium for another 48 hours. The width of each wound was measured using ImageJ software (NIH, USA).

Cell proliferation

A Cell Counting Kit-8 (Dojindo, Japan) was used to detect cell proliferation. Briefly, 3 × 10³ cells were resuspended in 96-well plates with 3 replicates. At Days 0, 1, 2, 3, 4, and 5, 10 μL CCK-8 and 90 μL fresh medium were added to each well. Then, the cells were cultured for another 2 hours, and the OD 450 nm of each well was measured.

Transwell assay

Migration assays were performed using 24-well transwell migration vessels (Corning-3422, USA). Briefly, cells (5 × 10⁴ PC-3, 1 × 10⁵ C4–2, or 3 × 10⁴ DU145) were resuspended in 200 μL of serum-free 1640 medium and seeded into the inner chamber. Then, 600 µL of medium containing 10% FBS was added to the bottom insert. After an appropriate incubation time, 4% paraformaldehyde was used to fix the cells that migrated through the wells, and they were stained with 0.1% crystal violet. The results were then analyzed through an optical microscope.

Dual-luciferase reporter gene assay

A Dual Luciferase Reporter Assay Kit (Promega, USA) was utilized to carry out this assay. In short, 3 × 10⁵ cells were seeded in a 24-well plate, and hsa_circ_0003998-WT/MUT and miR-605-5p mimic/NC were cotransfected into the cells using Lip3000. After 48 h, firefly and Renilla luciferase fluorescence was detected as described in the manuscript. Three replicates were performed for each assay.

RNA pull-down

A total of 2 × 10⁷ cells were prepared with cold PBS and then lysed in lysis buffer. Biotin-coupled probes were bound on magnetic beads and then incubated with lysates overnight at 4°C. Biotin-labeled hsa_circ_0063329 and oligonucleotide probes were purchased from RiboBio. After purification, the pull-down miRNAs were quantified by qRT‒PCR.

Tumor xenograft in vivo

Nude male BALB/c mice (4 weeks old) were purchased from Guangdong Animal Center and raised on a standard pelleted diet and water under pathogen-free conditions. Subcutaneous injection of PCa cells was performed to establish tumor xenograft models. Approximately 2 × 10⁶ PCa cells (PC-3-shcircRNA#3 cells or PC-3-NC cells and C4-2OE cells or C4–2-vector cells) were injected subcutaneously into mice. The formula used to calculate the tumor volume was V = (length × width²)/2. After 5 weeks, these mice were anaesthetized and intraperitoneally injected with D-luciferin (#D-Luciferin, Apexbio), and then IVIS Lumina II (Calliper Life Science, USA) was applied to visualize the fluorescence value of luciferase. Finally, the evidence of tumor cells in the xenograft tissues was verified with hematoxylin-eosin (H&E) staining, and Ki-67 (Abcam, #ab16667) antibody was used to confirm the proliferation index of cancer cells via immunohistochemistry (IHC) staining. The animal experiments were in agreement with laboratory animal welfare and ethics guidelines and were authorized by the Animal Care Commission of the First Affiliated Hospital of Guangzhou Medical University (Guangzhou, China).

Western blot

Western blot analysis was conducted as described previously [18]. Briefly, PCa cells were lysed in RIPA buffer (#KGP250, KeyGEN) following the manufacturer’s instructions. Then, 25 µg of protein was separated by SDS‒PAGE and transferred to PVDF membranes (Millipore). Subsequently, membranes were blocked with 5% nonfat milk. After washing, the primary antibodies were immunoblotted. The following primary antibodies were used in this research: DDX17 (ab180190) was purchased from Abcam, and TGIF2 (sc -81,989) and GAPDH (sc -365,062) were purchased from Santa Cruz. After incubation overnight, HRP-conjugated anti-mouse (sc-2005) and anti-rabbit IgG (sc-2004) secondary antibodies (Santa Cruz) were added. Finally, the signals were visualized by Bio-Rad ChemiDoc XRS+.

Statistical analysis

Statistical analyses were conducted using Prism version 7.0 (GraphPad Software, USA) and SPSS 22.0 software (SPSS, Inc., USA). All data are presented as the mean ± standard deviation (SD). Student’s t test (unpaired, two-tailed) was conducted to compare two groups for continuous variables. One-way analysis of variance (ANOVA) was executed for continuous variables in more than two groups. The χ² test was used to analyze the correlation between the expression of hsa_circ_000663329 and clinicopathological profiles. P < 0.05 was considered statistically significant.

Results

Characteristics of hsa_circ_0063329

First, the expression of circular RNAs was analyzed by microarray using PCa cell lines (DU145, PC-3, 22Rv1 and C4–2) and the RWPE-1 cell line (normal prostate epithelial cell line). Heatmap results revealed that 23 circRNAs were significantly downregulated in the PCa cell lines (Figure 1(a)). Furthermore, we compared our results with two other datasets [19,20] (Figure 1(b)), and we discovered that hsa_circ_0063329 was downregulated in all of these datasets. Therefore, hsa_circ_0063329 was chosen for the next study. Subsequently, we demonstrated that hsa_circ_0063329 was downregulated in PCa cell lines (Figure 1(c)). In addition, RT‒qPCR and FISH assays revealed that circ_0063329 was markedly downregulated in PCa tissues compared to adjacent normal tissues. (Figure 1(d,e) & Table 1). Moreover, downregulation of hsa_circ_0063329 was notably correlated with aggressive pT stage in patients with PCa (Table 1). Hsa_circ_0063329, containing 200 nucleotides, is produced by back-splicing of DDX17 (ADP ribosylation factor guanine nucleotide exchange factor 2) located at chr22:38894089–38894578. Sanger sequencing further confirmed the back-splicing in the RT‒qPCR output of hsa_circ_0063329 (Figure 1(f)). In addition, the stability of hsa_circ_0063329 was evaluated, and RNase R failed to digest hsa_circ_0063329; in contrast, RNase R was able to digest DDX17 mRNA (Figure 1(g)). Meanwhile, the actinomycin D assay revealed that the stability of hsa_circ_006332 was better than that of its linear DDX17 mRNA (Figure 1(i,j)). Moreover, RT‒qPCR and FISH were utilized to verify the localization of circular hsa_circ_0063329, and the results showed that hsa_circ_0063329 was largely expressed in the cytoplasm rather than in the nucleus (Figure 1(h,k-l)). Taken together, these results showed that the expression of hsa_circ_0063329 was markedly reduced in PCa cells and tissues and that it was largely expressed in the cytoplasm of PCa cells.

Figure 1.

Hsa_circ_0063329 was downregulated in PCa cells and tissues. (a) Heatmap shows the top 23 circRnas downregulated in PCa cell lines. (b) Venn diagram shows hsa_circ_0063329 was downregulated in PCa cell lines and tissues. (c) RT-qPCR was applied to confirm the expression of hsa_circ_0063329 in PCa cells and normal prostate cell (RWPE1). (d) the expression of hsa_circ_0063329 was examined in PCa tissues and normal samples by RT-qPCR. (e) FISH was further used to analysis the expression of hsa_circ_0063329 in BPH and PCa tissues. (f) Sanger sequencing to confirm the back splicing site of hsa_circ_0063329. (g) DDX17 mRNA and hsa_circ_0063329 were validated, RNA samples were treated with RNase R or mock (without the enzyme). (h) FISH was further used to analysis the expression of hsa_circ_0063329 in BPH and PCa tissues. (i and j) the stability of DDX17 mRNA and hsa_circ_0063329 was detected with actinomycin D in 22rv1 and LNCaP cells. (k and l) RNA FISH and Nuclear and Cytoplasmic Extraction experiments were used to detected the cellular localization of hsa_circ_0063329. ns, no significance, *P<0.05, **P<0.01, ***P<0.001.

Hsa_circ_0063329 inhibited the proliferation and migration of PCa cells

To clarify the role of hsa_circ_0063329 in PCa cells, hsa_circ_0063329 overexpression and knockdown systems in PCa cell lines were constructed. RT‒qPCR and Western blotting confirmed the overexpression and knockdown effects of this system, but it did not change the expression of DDX17 at either the transcript or protein level (Fig S1A-C) . Then, the viability of PCa cells was assessed by CCK-8 assay. The results demonstrated that downregulating the expression of hsa_circ_0063329 increased PCa cell viability; however, upregulating hsa_circ_0063329 significantly inhibited PCa cell viability (Figure 2(a)). The colony formation assay revealed that knockdown of hsa_circ_0063329 remarkably promoted the proliferation of PCa cells; in contrast, overexpression of hsa_circ_0063329 had the opposite result (Figure 2(b,c)). In addition, cell cycle assays demonstrated that after hsa_circ_0063329 knockdown, the percentage of PCa cells in G1 phase decreased and the percentage of PCa cells in S phase increased significantly. Silencing hsa_circ_0063329 had the opposite result.

Figure 2.

Hsa_circ_0063329 suppressed the proliferation and migration capability of PCa cells in vitro. The effect of hsa_circ_0063329 on PCa cell proliferation was detected by CCK-8 (a) and colony formation (b and c). (d and e) Cell cycle was determined in hsa_circ_0063329-silenced and hsa_circ_0063329 overexpressed PCa cells. The effect of hsa_circ_0063329 on PCa cell migration was confirmed by transwell migration assays (f,g,j) and wound healing assay (h,i,k). ns, no significance, *P<0.05, **P<0.01, ***P<0.001.

To further explore the function of hsa_circ_0063329 in PCa cell migration, transwell and wound scratch healing assays were carried out. The results confirmed that migrating cells increased significantly after hsa_circ_0063329 overexpression and decreased after knockdown of hsa_circ_0063329 (Figure 2(f,g,j)). Moreover, as shown in Figure 2(h,i,k), similar results were acquired by performing wound healing assays, which showed that knockdown or overexpression of hsa_circ_0063329 significantly changed PCa cell migration capability.

Collectively, these results proved that hsa_circ_0063329 impeded the proliferation and migration abilities of PCa cells.

Hsa_circ_0063329 functions as a sponge for binding miR-605-5p in PCa cells

It has been reported that circRNAs mostly function as miRNA sponges in the cytoplasm or as nuclear transcriptional regulators in the nucleus to regulate gene expression [21,22]. We found that hsa_circ_0063329 was mainly localized in the cytoplasm (Figure 1(l-k)). Then, we estimated the possible miRNA targets of hsa_circ_0063329 via bioinformatics analysis, and we obtained four overlapping potential miRNAs from the miRanda, CircBank and CircInteractome databases. Next, using the hsa_circ_0063329 probe for the circRNA pulldown assay in PCa cells, the pulldown efficiency of hsa_circ_0063329 was significantly enhanced in cells overexpressing hsa_circ_0063329 (Figure 3(c)). Subsequently, the levels of candidate miRNAs in particles pulled down with the hsa_circ_0063329 probe were detected by RT‒qPCR. We revealed that miR-605-5p was notably enriched in the pellets of PC3 and DU145 cells (Figure 3(d,e)). In addition, the luciferase activity was significantly reduced after cotransfection with these miRNA mimics and hsa_circ_0063329 (Figure 3(f)). According to these results, we selected miR-605-5p for further study. Additionally, RNA FISH assays were adopted to evaluate the localization of hsa_circ_0063329 and miR-605-5p in PCa cells, and we found that they were largely colocalized in the cytoplasm (Figure 3(g)). Moreover, a luciferase reporter gene assay by transfection of wild-type or mutant hsa_circ_0063329 sequences showed that luciferase activity was significantly lower in cells cotransfected with miR-605-5p mimics and wild-type sequences, but luciferase activity was not affected by site-mutated hsa_circ_0063329 (Figure 3(h,i)). These experiments suggested that hsa_circ_0063329 functions as a sponge of miR-605-5p.

Figure 3.

Hsa_circ_0063329 suppressed PCa cell proliferation and migration by acting as a sponge of miR-605-5p. (a) Overlap region showed the predicted miRnas from CircBank, CircInteractome and miRANDA. (b) Schematic illustration showed the potential binding sited between miRnas and hsa_circ_0063329 with Targetscan. (c-e) RNA-Pulldown was performed in PCa cells using hsa_circ_0063329 and oligo probes. (f) Luciferase reporter assay was used to determine the effects of four predicted miRnas on the luciferase activity of hsa_circ_0063329. (g) Colocalization between hsa_circ_0063329 and miR-605-5p was analyzed using RNA-FISH in 293T cells. DAPI stained for nuclei. Scale bar = 40 μm. (h) the potential binding site of hsa_circ_0063329 and miR-605-5p. (i) Dual-luciferase reporter assay was utilized to confirm the binding site of hsa_circ_0063329 and miR-605-5p. (j, k, m and n) Transwell and CCK-8 assays (l) were performed to detect the proliferation and migration of PCa cells after transfection with miR-605-5p (mimics, inhibitor and NC) and hsa_circ_0063329. *P<0.05, **P<0.01, ***P<0.001.

MiR-605-5p has been found to facilitate non-small cell lung cancer (NSCLC) progression by targeting TNFAIP3 [23]. To investigate the function of miR-605-5p in PCa cells, transwell assays and CCK-8 assays were carried out after transfection with miR-605-5p mimics or inhibitors. RT‒qPCR was used to confirm the efficiency of the miR-605-5p inhibitor and mimics (Fig. S1D). The results of transwell and CCK-8 assays showed that miR-605-5p promoted PCa cell migration and proliferation (Figure 3(j-l)). However, the oncogenic effects of miR-605-5p were reversed by hsa_circ_0063329 overexpression (Figure 3(l-m)). Therefore, we speculated that hsa_circ_0063329 suppressed the proliferation and migration abilities of PCa cells, possibly by acting as a sponge of miR-605-5p.

TGIF2 is a target gene of the hsa_circ_0063329/mir-605-5p axis in PCa

To better understand the mechanism of hsa_circ_0063329 in PCa progression, we performed mRNA-seq in LNCaP cells (transfected with hsa_circ_0063329 and empty vector), and compared with the control group, 69 genes were downregulated and 77 genes were upregulated in hsa_circ_0063329-overexpressing cells (Figure 4(a)). The results of Gene Ontology and KEGG analyses of the differentially expressed genes are shown in Figure 4(b,c). MicroRNAs in cancer and the TGF-beta signaling pathway were highly enriched. Next, to further determine the possible target genes of hsa_circ_0063329 and miR-605-5p, we applied three databases (miRDB, miRTarBase, and TargetScan) to estimate the possible target genes of miR-605-5p. Furthermore, TGIF2 was obtained from overlapping genes of these databases and was upregulated in LNCaP-hsa_circ_0063329 cells (Figure 4(d)). Then, we analyzed the mRNA and protein expression levels of TGIF2 in PCa cells after overexpression of hsa_circ_0063329. We found that the expression of TGIF2 was significantly enhanced in hsa_circ_0063329-overexpressing cells (Figure 4(e)). As expected, the expression level of TGIF2 was markedly increased after transfection with the miR-605-5p inhibitor (Figure 4(f)). Moreover, the increased effect of TGIF2 expression levels was rescued in PCa cells cotransfected with hsa_circ_0063329 and miR-605-5p mimics (Figure 4(f)). In summary, these results provide evidence supporting the hypothesis that TGIF2 is a target of hsa_circ_0063329 and miR-605-5p in PCa cells.

Figure 4.

TGIF2 is a direct target gene of the hsa_circ_0063329/mir-605-5p axis in PCa. (a) the heatmap showed the differentially expressed mRnas in LNCaP cells after overexpression hsa_circ_0063329. (b and c) GO&Kegg enriched pathway of DEGs in LNCaP cells. (d) Venn diagram of potential target mRnas of miR-605-5p identified in three databases (miRDB, miRtarbase, and TargetScan) and upregulated mRnas in LNCaP cells. (e) Relative TGIF2 mRNA levels in PCa cells after overexpression hsa_circ_0063329. (f) Western blot showed the TGIF2 protein expression level in PCa cells transfected with miR-605-5p (inhibitor, mimic and NC) and hsa_circ_0063329. **P<0.01, ***P<0.001.

Hsa_circ_0063329 was involved in tumor progression in vivo

Finally, to further assess the pathophysiological role of hsa_circ_0063329 in vivo, 2 × 10⁶ PCa cells with stable knockdown or overexpression of hsa_circ_0063329 were injected subcutaneously into nude mice. After 4 weeks, the mice were sacrificed, and the subcutaneous tumors were isolated and weighed. The results showed that knockdown of hsa_circ_0063329 significantly facilitated tumor growth in vivo (Figure 5(a-c)). H&E staining confirmed the histopathological characteristics of the tumor tissues (Figure 5(d)). IHC staining revealed that the number of Ki-67-positive cells was remarkably elevated in the hsa_circ_0063329-knockdown group compared with the control group (Figure 5(e,f)). In contrast, the tumor volumes and weights were dramatically reduced in nude mice injected with hsa_circ_0063329-overexpressing PCa cells compared with the control group (Figure 5(g-i)). In addition, the expression of Ki-67 was lower in the hsa_circ_0063329 overexpression group (Figure 5(j,k)). These data demonstrated that silencing hsa_circ_0063329 promotes tumor proliferation, while overexpression of hsa_circ_0063329 delays tumor development.

Figure 5.

Effect of hsa_circ_0063329 on tumor progression in vivo. (a) Tumor was detected by in vivo images transfected with sh-hsa_circ_0063329 and sh-Control. (b and c) the effect of hsa_circ_0063329 silencing on tumor volume and tumor weight were analyzed. (d-f) H&E staining for tissue morphology and IHC staining analysis of proliferation (Ki-67). (g and h) the pictures showed tumors of nude mice from different groups (transfected with vector or hsa_circ_0063329). (I) the weight of tumor from each group was analyzed. (j and k) the expression of Ki-67 was determined by IHC. *P<0.05, **P<0.01, ***P<0.001.

Discussion

With technological advances in high-throughput RNA sequencing, an openly accessible dataset containing comprehensive cancer-related circRNA expression profiles has been discovered. Diverse circRNAs have been reported to be abundantly and differentially expressed in multiple cancer types [24,25], such as breast cancer [26], hepatocellular carcinoma [27], and ovarian cancer [28]. Undoubtedly, circRNAs play vital roles in modulating tumor proliferation and metastasis and influence cancer hallmarks, including PCa [29,30]. Interestingly, our previous report identified circRNAs from PCa cells and proved that circNOLC1 serves as an oncogene in PCa [17]. Here, we focused on a significantly downregulated circRNA, hsa_circ_0063329, which was dramatically downregulated in PCa clinical samples and negatively correlated with clinical T stage. Then, functional experiments of hsa_circ_0063329 in PCa progression were conducted. Gain- and loss-of-function experiments proved that upregulating hsa_circ_0063329 remarkably suppressed the proliferation and migration capabilities of PCa cells in vitro and inhibited PCa oncogenesis in vivo. In contrast, silencing hsa_circ_0063329 exerted the opposite effects. These results implied that hsa_circ_0063329 plays a critical role in the pathogenesis of PCa.

CircRNAs have been reported to participate in many biological and pathological processes via protein binding, miRNA sponges, transcriptional regulation and even protein translation [6,31]. The miRNA sponge concept is that circRNAs regulate the function of miRNAs by combining with their complementary pairs, thereby controlling the expression of their target genes, and this is considered to be the most extensive and mature theory of how circRNAs function. For example, circKCNN2 sponges miR-520c-3p, thereby increasing the expression of MBD2 and inhibiting the recurrence of hepatocellular tumors [32]. Hsa_circ_0003258 has been reported to be elevated in PCa and may promote metastasis by acting as a sponge for miR-653-5p and forming a complex with IGF2BP3/HDAC4 [33]. However, the underlying mechanism of hsa_circ_0063329 in PCa progression remains uncertain. In the present study, we verified that hsa_circ_0063329 was mainly enriched in the cytoplasm of PCa cells. Subsequently, the potential targets of hsa_circ_0063329 were predicted using bioinformatics tools, and miR-605-5p was finally identified as one of the biological targets of hsa_circ_0063329. As previously reported, miRNAs are a class of noncoding RNAs containing 22–25 nucleotides [17]. They exert their effect by repressing gene expression by targeting the 3’-UTR of mRNA posttranscriptionally. MiR-605-5p has been reported to promote NSCLC invasion and proliferation by targeting TNFAIP3 [23]. In the current study, we proved that miR-605-5p promoted PCa progression by transfection miR-605-5p mimics into PCa cells. In the cotransfection system of the circRNA hsa_circ_0063329 and a miR-605-5p inhibitor, interfering with miR-605-5p reduced the effect of hsa_circ_0063329 silencing on cell proliferation and migration, suggesting that hsa_circ_0063329 could affect PCa progression by regulating miR-605-5p.

The TGF-β-induced factor homeobox 2 (TGIF2) protein is a Smad corepressor that inhibits the TGF-β/Smad pathway [34,35]. It is widely known that the TGF-β/Smad pathway participates in promoting the epithelial-mesenchymal transition of cancer, especially PCa [36,37]. Recently, it was reported that TGIF2 was downregulated via miR-181a in PCa, resulting in activation of the Smad pathway, thereby facilitating PCa metastasis [38]. In the current research, we discovered that TGIF2 was a downstream target of miR-605-5p. The expression of TGIF2 was modulated not only by hsa_circ_0063329 but also by miR-605-5p. In addition, we found that the TGF-beta signaling pathway was enriched in PCa cells after overexpression of hsa_circ_0063329 and that the expression of TGIF2 was rescued by cotransfection of miR-605-5p and hsa_circ_0063329. Therefore, we assumed that downregulated hsa_circ_0063329 in PCa activates the TGF-beta/Smad pathway via the miR-605-5p/TGIF2 axis. However, whether hsa_circ_0063329 has other mechanisms of regulating PCa progression requires further investigation.

Conclusion

Our study provided evidence that hsa_circ_0063329 plays a crucial role in suppressing growth and metastasis in PCa. Mechanistically, hsa_circ_0063329 serves as a sponge for miR-605-5p to modulate the expression of TIGF2 and then activate the TGF-β pathway (Figure 6). The current research provides a potential therapeutic target for PCa.

Figure 6.

Schematic illustrating the biological function and mechanism of hsa_circ_0063329 in PCa.

Funding Statement

This work was financed by grants from the National Natural Science Foundation of China [No.82203710], the Science and Technology Plan Project of Guangzhou [No. 202102010150], the Guangdong Basic and Applied Basic Research Foundation [No. 2020A15150100667], the Distinguished Young Talents in Higher Education Foundation of Guangdong Province [No.2019KQNCX115], the China Postdoctoral Science Foundation funded project [No.2019M662865], and the Achievement Cultivation and Clinical Transformation Application Cultivation projects of the First Affiliated Hospital of Guangzhou Medical University [No.ZH201908].

Author’s Contributions

Ping Liu had full access to all the data in the study and takes responsibility for the data and the accuracy of the data analysis; Daojun Lv, Shenren Cen and Shuxin Yang performed most of the experiments and wrote the manuscript. Zihao Zou, Jianfeng Zhong, Zhaojun Pan, Nan Deng, Yubin Li, Kaihui Wu, Jiamin Wang participated in study design, study implementation and manuscript revision. All authors approved the final version of the submitted manuscript.

Availability of data and materials

The datasets used and/or analyzed during the current study are available in the supplementary files. The sequencing data supporting the results of this study will be made available by the authors, without undue reservation.

Consent for publication

We have obtained consent to publish from the participants to report patient data.

Disclosure statement

No potential conflict of interest was reported by the authors.

Ethics approval and consent to participate

The study was reviewed and approved by the ethics committee of the Third Affiliated Hospital of Guangzhou Medical University (Guangzhou, China).

Supplementary material

Supplemental data for this article can be accessed online at https://doi.org/10.1080/15384101.2023.2174658.