Introduction

Prostate cancer is a common health issue in males contributing to the major cause of cancer-related deaths [1]. In the past decades, great progress has been made in research on the pathogenesis of PCa [2]. With the development of diagnosis and treatment of PCa, the outcomes of patients have gained much improvement [3, 4]. However, the challenge remains on the limited effective strategies. Therefore, improving the understanding of the underlying pathogenesis of PCa is crucial for investigating mechanism-based therapeutic strategies.

Circular RNAs (circRNAs) are a unique class of endogenous RNA transcripts, in principle generated using a back-splicing mechanism, with higher stability and tissue-dependent expression [5]. The biology of circRNAs has rapidly advanced and demonstrated their diverse roles, including regulation of gene expression and possible coding potential, and competitive interactions with proteins and microRNAs (miRNAs) under various pathological conditions [6]. Advances in RNA identification techniques have provided a large number of circRNAs exhibiting upregulation or downregulation in PCa tissues compared to normal adjacent tissues [7]. For instance, circ_0006156 expressed at low levels in PCa tissues, and circ_0006156 lowered PCa cell malignancy by binding to S100A9 [8]. Circ-PFKP facilitated cell proliferative capacity via activation of IMPDH2 in PCa [9]. Hsa_circ_0081234 (circTRRAP), derived from the TRRAP gene, was uncovered as a facilitator in PCa [10]. Currently, little is known on the underlying mechanism of circTRRAP in PCa progression.

MiRNAs participate in disease progression via regulating oncogenes or tumor-suppressive genes [11, 12]. Dysregulation of miRNA expression is associated with a variety of cancers, including PCa [13]. For instance, miR-29b was lowly expressed in PCa, and miR-29b upregulation lessened PCa cell invasion and migration [14]. Another example was that miR-128 targeted BMI-1 to restrain PCa progression [15]. Moreover, several circRNAs have been demonstrated to be involved in PCa progression by interacting with miRNAs [16,17,18]. For instance, circ-NOLC1 prompted PCa development via interaction with miR-647 [19]. Circ-0086722 drove PCa progression by mediating STAT5A expression by sponging miR-339-5p [20]. Existing evidence suggests a vital action of the circRNA/miRNA/mRNA competitive endogenous RNA (ceRNA) network pathway in PCa [21, 22]. Based on the important role of the ceRNA mechanism, we found that miR-515-5p might interact with circTRRAP through the circinteractome online prediction tool (https://circinteractome.irp.nia.nih.gov/). miR-515 has been demonstrated as a tumor repressor in PCa [23, 24], but the association between miR-515 and circTRRAP is indistinct in PCa. In addition, homeobox A1 (HOXA1), which is an oncogene in multiple tumors [25, 26], was predicted as a possible miR-515-5p molecular target using the Targetscan online prediction tool (https://www.targetscan.org/vert_80/). Currently, whether miR-515-5p is associated with HOXA1 in PCa needs to be verified.

Hence, the research was to characterize the action of circTRRAP in PCa and the molecular mechanism of the circTRRAP/miR-515-5p/HOXA1 axis. All findings come together to confirm that circTRRAP might function as a miR-515-5p sponge to increase HOXA1 expression and consequently promote PCa progression. In conclusion, targeting circTRRAP to explore methods for PCa therapy is very promising.

Materials and methods

Patient and tissue collection

PCa patients (n = 50) who have signed the informed consent were recruited from Seventh Medical Center, PLA General Hospital. This study was performed under the approval of the Ethics Committee of Seventh Medical Center, PLA General Hospital.

Cell culture

Human prostatic epithelial cell line RWPE-1 (#165641, Mingzhoubio., Ningbo, China) and PCa cell lines DU145 (MZ-0058, Mingzhoubio.) and PC3 (MZ-0145, Mingzhoubio.) were cultured at 37 °C with 5% CO2. RWPE-1, DU145, and PC3 cell lines were cultured with the Keratinocyte Serum-Free Medium (K-SFM) Kit (#17005-042, Thermo, Waltham, MA, USA), MEM (Thermo) and Ham’s F-12K (Thermo), respectively.

Cell transfection

The overexpression plasmids of circTRRAP (circTRRAP) and HOXA1 (pcDNA-HOXA1) were generated with the pCD5-ciR (Geneseed, Guangzhou, China) or pcDNA3.1 (YouBio, Changsha, China) vectors with their respective empty vectors as negative controls (circ-NC and pcDNA-NC). Three small interfering RNAs (siRNAs) for circTRRAP circTRRAP [si-circTRRAP#1 (5ʹ-UGACGUGACAUUUGUUCGAGU-3ʹ), si-circTRRAP#2 (5ʹ-ACGUGACAUUUGUUCGAGUCU-3ʹ), and si-circTRRAP#3 (5ʹ-UGACAUUUGUUCGAGUCUCAC-3ʹ)] and non-target siRNA (si-NC, 5ʹ-AUAUCAUAGGUUGAUAAUGGU-3ʹ), miR-515-5p inhibitor (5ʹ-CAGAAAGUGCUUUCUUUUGGAGAA-3ʹ) and its control (inhibitor-NC, 5ʹ-CAGUACUUUUGUGUAGUACAA-3ʹ), as well as miR-515-5p mimic (miR-515-5p, 5ʹ-UUCUCCAAAAGAAAGCACUUUCUG-3ʹ) and its control (miR-NC, 5ʹ-UUCUCCGAACGUGUCACGUTT-3ʹ) were generated by Fulengen (Guangzhou, China). The lipofecter liposomal transfection reagent (Beyotime, Shanghai, China) was used for transfection of DU145 and PC3 cells.

Quantitative real-time polymerase chain reaction (qRT-PCR)

Trizol reagent (Solarbio, Beijing, China) was used for isolation of total RNA from PCa samples and cells. The PARIS Kit (Thermo) was utilized for isolation of nuclear and cytoplasmic fractions. RNase R (Epicentre Technologies) was utilized for digestion of total RNA (2 μg). RNA was reversely transcribed to complementary DNA (cDNA) using the reverse transcription PrimeScript RT Master Mix (Takara, Dalian, China) or RiboBio reverse transcription kit (Guangzhou, China). Quantification was done with SYBR (Takara) and specific primers (Table 1). The 2−ΔΔCt method was utilized for the calculation of the fold change [27].

Table 1 Primer sequences used for qRT-PCR
Full size table

Estimation of cell viability

PCa cells (5 × 103 cells) were plated into 96-well plates. After the incubation for 0, 24, 48 and 72 h, 10 μL of the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl-tetrazolium bromide (MTT) solution (Beyotime) was added into each well. The cells were cultured for another 4 h and then each well was filled with 100 μL of dimethyl sulfoxide (Solarbio) for dissolving the crystal. The measurement of the optical density value was conducted with a microplate reader (Bio-Rad, Hercules, CA, USA).

Assessment of cell proliferation

PCa cells (5 × 103 cells) were seeded into 96-well plates and their proliferative capacity was measured using the Cell-Light™ 5-ethynyl-2ʹ-deoxyuridine (EdU) Apollo 488 Kit (RiboBio) following the manufacturer’s protocol. Briefly, cells were labeled with 50 μM EdU for 2 h, followed by fixation in 4% paraformaldehyde for 30 min. After staining with Apollo 488 and Hoechst, the number of positive cells exhibiting green fluorescence was recorded under a fluorescence microscope (Olympus, Tokyo, Japan).

Transwell assay

Estimation of cell migratory and invaded capacities was conducted with transwell inserts (Corning, Corning, NY, USA). For the invasion assay, the inserts were pre-coated with Matrigel. Meanwhile, the inserts without Matrigel were used for the migration assay. PCa cells (1 × 104) resuspended in serum-free medium were plated into the upper chambers. The lower chambers were filled with fresh medium containing 10% serum. After the incubation for 24 h, cells that crossed the membrane were stained with 0.5% crystal violet (Solarbio). Observation of migrating or invasive cells was done with a microscope (Olympus) at 100× magnification.

Western blot

Extraction of total protein was done with the total protein extraction kit (#BC3710, Solarbio). The protein samples (20 μg) were subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis. The membranes were blocked with the western blocking buffer (Solarbio) after transferring the membrane to a nitrocellulose membrane (Bio-Rad). Following incubation with specific primary antibodies, the membranes were incubated with a secondary antibody, the bands were visualized by SuperSignal West Pico Stable Peroxide Solution (Thermo). Primary antibodies were against PCNA (ab92552, 1:5000, Abcam, Cambridge, UK), MMP2 (ab181286, 1:1000, Abcam), MMP9 (ab76003, 1:5000, Abcam), HOXA1 (ab230513, 1:500, Abcam), and β-actin (ab115777, 1:200, Abcam).

Glucose and lactate level

PCa cells (1 × 104) were cultured for 24 h and then the culture medium was collected. Measurement of glucose uptake and lactate production was done with the Glucose Colorimetric Fluorometric Assay Kit (#K606-100, BioVision, USA) and the Lactate Colorimetric Assay Kit (#K627-100, BioVision) following manufacturer’s instructions, respectively.

Dual-luciferase reporter assay

PCa cells were transfected with a constructed luciferase reporter vector (circTRRAP-WT, circTRRAP-MUT, HOXA1 3ʹUTR-WT, or HOXA1 3ʹUTR-MUT) along with miR-515-5p or miR-NC. Cells were collected for measurement of the luciferase activity with a dual-luciferase assay kit (Promega, Madison, WI, USA) after 48 h of transfection. The luciferase plasmids used above were generated by inserting the wild-type and mutant sequences of circTRRAP and HOXA1 into the psiCHECK™-2 vector (Promega).

Xenograft models

The animal experiment permission was granted by the Animal Care Committee of the Seventh Medical Center, PLA General Hospital. Male BALB/c nude mice (4 weeks old) (Vital River, Beijing, China) were reared in conditions deprived of specific pathogens. PC3 cells (2 × 106) with Lenti-sh-NC or Lenti-sh-circTRRAP were injected into the flank of nude mice (n = 6). Tumor volume was monitored and calculated every 5 days (volume = (width2 × length)/2). After 30 days of injection, mice were anesthetized and then sacrificed for subsequent analysis.

IHC staining

Paraffin-embedded xenograft tumor sections were deparaffinized with xylene, rehydrated with graded ethanol, antigen retrieved with citrate buffer (10 mM, pH 6.0), and blocked for endogenous peroxidase with 3% H2O2 in TBS. Samples were incubated with primary antibodies against ki-67 (ab243878, 1:100, Abcam), MMP2 (ab97779, 1:500, Abcam), and MMP9 (ab76003, 1:1000, Abcam). After removal of excess antibody, tissue samples were reacted with DAB chromogen and substrate mixture (Thermo), followed by counterstaining with hematoxylin.

Statistical analysis

Bar plots were generated by GraphPad Prism 7 software (GraphPad, La Jolla, CA, USA). Each data represented the mean ± standard deviation (SD) of three biological replicates, with each biological replicate in triplicate. The difference was assessed by Student’s t-test or analysis of variance. Results were considered statistically significant when the P-value was less than 0.05.

Results

Validation of circTRRAP as an upregulated circRNA in PCa samples and cell lines

Hsa_circ_0081234 (circTRRAP) is derived from the TRRAP gene (exons 4 to 8) (Fig. 1A) and has been reported as a promoting regulator in PCa [10]. To explain the action of circTRRAP in PCa, the alteration in circTRRAP expression was validated. Results of qRT-PCR showed that circTRRAP expression was significantly elevated in PCa samples (n = 50) relative to matching normal samples (n = 50) (Fig. 1B). Similar circTRRAP trends were also seen in PCa cell lines when compared with the RWPE-1 cell line (Fig. 1C). Subcellular fractionation combined with qRT-PCR analysis exhibited that circTRRAP had a higher proportion in the cytoplasmic fraction of PCa cell lines (Fig. 1D). The combination of RNase R digestion and qRT-PCR analysis confirmed that circTRRAP has a more stable structure than the linear mRNA of the parental gene TRRAP (Fig. 1E). Collectively, high circTRRAP expression might be associated with PCa progression.

Fig. 1
figure 1

CircTRRAP was overexpressed in PCa. A Schematic illustration of the structure of circTRRAP. B CircTRRAP expression was measured in PCa tissues (n = 50) and corresponding normal tissues by qRT-PCR. C CircTRRAP abundance was detected in PC cell lines by qRT-PCR. D CircTRRAP abundance in the nuclear and cytoplasm fractions was assessed by qRT-PCR. E RNase R was administrated to digest the extracted RNA for analysis of the stability of circTRRAP. *P < 0.05

Full size image

Knockdown of circTRRAP restrained PCa cell malignant phenotypes and lowered PCa cell glycolysis

To assess the effect of circTRRAP silencing on PCa cell malignant phenotypes, we transfected PCa cells with circTRRAP-specific siRNAs or a non-target control si-NC. PCa cells transfected with circTRRAP-specific siRNAs displayed substantially lower circTRRAP expression validated by qRT-PCR, and si-circTRRAP#1, which caused the greatest change in circTRRAP expression, was used for subsequent functional analysis (Fig. 2A). Functionally, circTRRAP-deficient PCa cell lines had low cell viability and proliferative capacity, as confirmed by MTT assay and EdU assay (Fig. 2B and C). Transwell assays showed that depletion of circTRRAP resulted in reduced migratory and invasive abilities of PCa cell lines (Fig. 2D and E). Also, circTRRAP knockdown decreased PCNA, MMP2, and MMP9 protein levels (Fig. 2F). Insufficiency of circTRRAP resulted in a significant reduction in both glucose uptake and lactate production in PCa cell lines (Fig. 2G and H). Together, circTRRAP prompted malignant phenotypes and glycolysis of PCa cells.

Fig. 2
figure 2

Knockdown of circTRRAP reduced PCa cell malignancy and glycolysis. A Transfection efficiencies of three siRNAs targeting circTRRAP were determined by qRT-PCR. B–E Cell viability, proliferation, migration, and invasion in si-circTRRAP#1 or si-NC-transfected PCa cells were estimated by MTT, EdU, and transwell assays. F and G PCNA, MMP2, and MMP9 protein levels in si-circTRRAP#1 or si-NC-transfected PCa cells were detected by western blot. G and H The glucose uptake and lactate production of si-circTRRAP#1 or si-NC-transfected PCa cells were measured. *P < 0.05

Full size image

MiR-515-5p was sponged by circTRRAP

Because cytoplasmic circRNAs can affect miRNAs and corresponding downstream genes through a ceRNA mechanism, we explored the function of circTRRAP as a miRNA molecular sponge. An online prediction tool Circular RNA Interactome showed that circTRRAP possessed bases that may adsorb miR-515-5p (Fig. 3A). For validation of this association, the luciferase reporter vectors circTRRAP-WT and circTRRAP-MUT were constructed. The data of dual-luciferase reporter assay showed that the luciferase reporter circTRRAP-WT had a lower luciferase activity in the miR-515-5p-overexpression group, but the luciferase reporter circTRRAP-MUT did not change significantly (Fig. 3B and C). Moreover, circTRRAP insufficiency led to an elevation in miR-515-5p expression, but circTRRAP overexpression had the opposite function (Fig. 3D and E). Furthermore, the abundance of miR-515-5p was significantly decreased in PCa samples and cells (Fig. 3F and G). Meanwhile, there was a negative correlation between miR-515-5p and circTRRAP expression levels in PCa samples (r = − 0.455, P = 0.0009) (Fig. 3H). These findings suggested circTRRAP as a miR-515-5p molecular sponge.

Fig. 3
figure 3

CircTRRAP functioned as a miR-515-5p decoy. A The complementary binding sites of circTRRAP with miR-515-5p. B and C Luciferase activity was measured in PCa cells transfected with circTRRAP-WT or circTRRAP-MUT and miR-NC or miR-515-5p. D and E miR-515-5p abundance was measured in PCa cells transfected with si-NC, si-circTRRAP#1, circ-NC or circTRRAP by qRT-PCR. F and G miR-515-5p abundance was detected in PCa tissues and cell lines by qRT-PCR. H Pearson’s correlation analysis determined the correlation between the levels of circTRRAP and miR-515-5p in PCa tissues. *P < 0.05

Full size image

Knockdown of miR-515-5p alleviated circTRRAP insufficiency-mediated impacts on PCa cell malignant phenotypes and glycolysis

Whether circTRRAP-mediated regulation of PCa progression is linked to miR-515-5p was further elucidated. As shown in Fig. 4A, the increased expression of miR-515-5p urged by circTRRAP knockdown was weakened by the introduction of miR-515-5p inhibitor. Furthermore, circTRRAP knockdown-mediated repression of PCa cell viability, proliferation, migration, and invasion was whittled by downregulation of miR-515-5p (Fig. 4B–F). Additionally, interference of circTRRAP-induced downregulation of MMP2 and MMP9 was mitigated by miR-515-5p deficiency (Fig. 4G). Moreover, the decreased levels of glucose uptake and lactate production in PCa cells mediated by circTRRAP inhibition were impaired after miR-515-5p silencing (Fig. 4H and I). These results uncovered that circTRRAP interacted with miR-515-5p to mediate PCa progression.

Fig. 4
figure 4

Deficiency of miR-515-5p attenuated circTRRAP knockdown-mediated effects on PCa cell malignancy and glycolysis. A miR-515-5p abundance was measured in PCa cells transfected with si-NC, si-circTRRAP#1, si-circTRRAP#1 + inhibitor-NC, or si-circTRRAP#1 + miR-515-5p inhibitor by qRT-PCR. B–F Cell viability, proliferation, migration and invasion were detected in the above cells by MTT, EdU and transwell assays. G PCNA, MMP2, and MMP9 protein levels were detected in the above cells by western blot. H and I The glucose uptake and lactate production of the above cells were measured. *P < 0.05

Full size image

HOXA1 was a miR-515-5p target

TargetScan predicted that HOXA1 contained the complementary binding sites with miR-515-5p (Fig. 5A). To ascertain the impact of miR-515-5p on HOXA1, a fragment containing the 3ʹUTR of HOXA1 was spliced into the psiCHECK™-2 vector. The reduced luciferase activity was observed in cells co-transfected with the HOXA1 3ʹUTR-WT vector with miR-515-5p mimic (Fig. 5B and C). Moreover, HOXA1 mRNA and protein levels were remarkably repressed by miR-515-5p overexpression in PCa cells (Fig. 5D and E). In addition, the upregulation of HOXA1 mRNA and protein levels was got in PCa tissues (n = 50) compared with that in normal tissues (Fig. 5F and G). Similarly, the abundances of HOXA1 mRNA and protein were also significantly higher in PCa cells than that in RWPE-1 cells (Fig. 5H and I). Besides, the HOXA1 mRNA level in PCa tissues had a negative correlation with miR-515-5p (r = − 0.437, P = 0.0015) (Fig. 5J). Also, circTRRAP silencing led to an overt reduction in the mRNA and protein levels of HOXA1, but miR-515-5p inhibition partly reversed this reduction (Fig. 5K and L). Expectedly, the HOXA1 mRNA level in PCa tissues had a positive correlation with circTRRAP (Fig. 5M). The above results illustrated the targeting relationship between miR-515-5p and HOXA1 and that circTRRAP interacted with miR-515-5p to mediate HOXA1 expression.

Fig. 5
figure 5

HOXA1 was a miR-515-5p target. A The binding sites of miR-515-5p with HOXA1. B and C Luciferase activity was detected in PCa cells transfected with HOXA1 3ʹUTR-WT or HOXA1 3ʹUTR-MUT and miR-515-5p or miR-NC. D and E HOXA1 mRNA and protein levels were detected in PCa cells transfected with miR-NC or miR-515-5p. F–I The abundances of HOXA1 mRNA and protein were examined in PCa tissues and cells. J Pearson’s correlation analysis assessed the correlation between the levels of miR-515-5p and HOXA1 mRNA in PCa tissues. K and L HOXA1 mRNA and protein levels were estimated in PCa cells transfected with si-NC, si-circTRRAP#1, si-circTRRAP#1 + inhibitor-NC, or si-circTRRAP#1 + miR-515-5p inhibitor. M Pearson’s correlation analysis determined the correlation between the levels of circTRRAP and HOXA1 mRNA in PCa tissues. *P < 0.05

Full size image

miR-515-5p targeted HOXA1 to lower PCa cell malignancy and glycolysis

To explore whether HOXA1 is associated with miR-515-5p-mediated effects on PCa cell malignancy and glycolysis, rescue experiments were conducted. The reduced mRNA and protein levels of HOXA1 promoted by miR-515-5p overexpression were whittled by introduction of the pcDNA-HOXA1 plasmid (Fig. 6A and B). Furthermore, miR-515-5p overexpression restrained PCa cell viability and reduced proliferative, migratory and invasive capacities of PCa cells, while HOXA1 overexpression weakened these impacts (Fig. 6C–F). Moreover, the decreased protein levels of PCNA, MMP2 and MMP9 caused by miR-515-5p upregulation were whittled by HOXA1 overexpression (Fig. 6G and H). In addition, miR-515-5p increase-induced reduction of glucose uptake and lactate production was impaired by the overexpression of HOXA1 (Fig. 6I and J). These data indicated that miR-515-5p targeted HOXA1 to inhibit PCa cell malignancy and glycolysis.

Fig. 6
figure 6

miR-515-5p lessened PCa cell malignancy and glycolysis by targeting HOXA1. A and B HOXA1 mRNA and protein levels were estimated in PCa cells transfected with miR-NC, miR-515-5p, miR-515-5p + pcDNA-NC, or miR-515-5p + pcDNA-HOXA1. C–F Cell viability, proliferation, migration and invasion were determined in the above cells. G and H Detection of PCNA, MMP2, and MMP9 protein levels was done in the above cells. I and J Measurement of glucose uptake and lactate production was conducted in the above cells. *P < 0.05

Full size image

Interference of circTRRAP reduced PCa growth in vivo

To validate the action of circTRRAP in vivo, PC3 cells stably expression of Lenti-sh-NC or Lenti-sh-circTRRAP were constructed and injected into mice. In vivo experiments demonstrated that circTRRAP silencing impaired tumor growth (volume and weight) relative to the control group (Fig. 7A and B). Moreover, tumor sample-derived from mice with circTRRAP silencing had lower levels of circTRRAP, HOXA1, and PCNA, as well as higher levels of miR-515-5p (Fig. 7C and D). IHC analysis also showed that the number of ki-67/MMP2/MMP9-positive cells was lowered in tumor sample-derived from mice with circTRRAP silencing (Fig. 7E). Together, circTRRAP prompted PCa growth in vivo.

Fig. 7
figure 7

Knockdown of circTRRAP decreased PCa growth in vivo. A Tumor volume was measured every 5 days. B The mean weight of tumors in both groups was at the end point. C The abundances of circTRRAP and miR-515-5p were detected in tumors in both groups. D Western blot analyzed HOXA1 and PCNA protein levels in tumors in both groups. E IHC analysis of ki-67, MMP2 and MMP9 protein levels in tumors in both groups. *P < 0.05

Full size image

Discussion

The unique molecular structure of circRNAs and their stage-specific and cell/tissue-specific expression characteristics make them more likely to be cancer therapeutic targets than linear transcripts [28]. Some scholars have confirmed the important role of some dysregulated circRNAs in PCa progression [29]. However, the molecular mechanism by which most circRNAs play an important role in PCa remains unclear.

A previous study indicated the upregulation of circTRRAP in PCa samples with spinal metastases, and exosomal circTRRAP facilitated cell epithelial-mesenchymal transition and invasion by binding to miR-1 and subsequently increasing MAP3K1 expression in PCa [10]. Our data validated the overexpression of circTRRAP in PCa samples in comparison to matching normal samples. Functional experiments manifested that circTRRAP knockdown lowered PCa cell viability and repressed PCa cell proliferative, migratory, and invasive capacities, as well as weakened tumor growth in mouse xenograft models. These results highlighted the promoting effect of circTRRAP on PCa growth. The novelty of our study lied in the discovery that circTRRAP was involved in glycolysis in PCa cells and circTRRAP mediated PCa cell malignancy and glycolysis by the miR-515-5p/HOXA1 axis.

Cancer cells have the ability to produce energy in an oxygen-independent manner, that is, their metabolic phenotype is preferentially dependent on glycolysis [30]. Targeting glycolysis has great potential to treat tumors based on the dependence of cancer cells on glycolysis to influence tumorigenesis [31]. Herein, circTRRAP silencing reduced glucose uptake and lactate production in PCa cells, manifesting that circTRRAP took part in glycolysis in PCa.

Existing evidence points to the importance of circRNAs in cancer progression by regulating the miRNA/mRNA axis [32]. Currently, circTRRAP had been shown to regulate the miR-1/MAP3K1 axis to participate in PCa progression [10]. Here, we discovered the possible association of circTRRAP with miR-515-5p by bioinformatics analysis and subsequently confirmed their targeting relationship by dual-luciferase reporter assays. miR-515-5p as a suppressor has been exposed in diverse cancers, such as gastric cancer [33] and breast cancer [34]. In PCa, the downregulation of miR-515-5p mediated by circ-0057553 overexpression facilitated cancer cell aerobic glycolysis, invasion, and migration [24]. Furthermore, circ-PAPPA [35] and circ-FOXM1 [36] prompted PCa cell malignancy by sequestering miR-515-5p. Our results displayed the downregulation of miR-515-5p in PCa samples and cells. Downregulation of miR-515-5p abated the suppressive effects of circTRRAP silencing on PCa cell malignancy and glycolysis. All results supported that circTRRAP mediated PCa cell malignancy and glycolysis by interacting with miR-515-5p.

Next, this research probed into the targets of miR-515-5p. Here, Bioinformatics analysis and validation experiments confirmed that HOXA1 was a miR-515-5p target. HOXA1 is a conserved member of a family of homeobox transcription factors that coordinately regulate early developmental patterns, organogenesis, and cell fate [37]. There is increasing evidence that dysregulation of HOXA1 affects tumorigenesis, including gastric cancer [25], glioblastoma [38], and melanoma [39]. In PCa, HOXA1 silencing suppressed tumor growth and metastasis in xenograft models [40]. Moreover, circ_0074032 sequestered miR-198 and subsequently elevated HOXA1 expression, thus leading to prompting PCa progression [41]. Here, HOXA1 had high levels in PCa samples and cells, and HOXA1 upregulation alleviated miR-515-5p mimic-mediated repression of cell glycolysis and malignancy in PCa cells. Additionally, circTRRAP regulated HOXA1 expression by competitively sponging miR-515-5p, highlighting the network mechanism of the circTRRAP/miR-515-5p/HOXA1 axis in PCa cells.

In conclusion, circTRRAP in PCa samples and cells was highly expressed. Furthermore, knockdown of circTRRAP lessened PCa cell malignancy and glycolysis via decreasing HOXA1 expression via sponging and sequestering miR-515-5p. This research disclosed a novel mechanism for PCa progression.