Silencing of circTASP1 inhibits proliferation and induces apoptosis of acute myeloid leukaemia cells through modulating miR‐515‐5p/HMGA2 axis

Abstract Acute myeloid leukaemia (AML) is a common hematopoietic disease that is harmful to the lives of children and adults. CircRNAs are aberrantly expressed in the haematologic malignancy cells. However, the expression of circTASP1 and its function in AML remain unclear. In this study, we showed that circTASP1 was significantly up‐regulated in AML peripheral blood samples and cells. Knockdown of circTASP1 inhibited proliferation and promoted apoptosis of HL60 and THP‐1 cells in vitro. Bioinformatics prediction and luciferase reporter assay proved that circTASP1 sponged miR‐515‐5p and negatively regulated miR‐515‐5p expression in HL60 and THP‐1 cells. High mobility group A2 (HMGA2) was proved to be a downstream target of miR‐515‐5p. The rescue experiments confirmed that knockdown of circTASP1 inhibited proliferation and induced apoptosis by modulating miR‐515‐5p/HMGA2 pathway. Moreover, the in vivo experiment indicated that knockdown of circTASP1 suppressed tumour growth. In conclusion, circTASP1 acts as a sponge for miR‐515‐5p to regulate HMGA2, thereby promoting proliferation and inhibiting apoptosis during AML progression. Thus, circTASP1 has the potential to be explored as a therapeutic target for AML treatment.


| INTRODUC TI ON
Acute myeloid leukaemia (AML) is a haematopoietic disease caused by aberrant self-renewal, proliferative and potential differentiation in the precursor of bone marrow cells. 1 AML in children is rare and heterogeneous, with an incidence of 7 cases per million children under the age of 15 years. In developed countries, intensive treatment combined with effective supportive care can improve the survival rate to 70%. 2 Stem cell transplantation is a common treatment strategy for AML. 3  accumulating studies have identified more prognostic markers, including a variety of molecular signals: DNA mutation, abnormal expression of miRNA and non-coding RNA (ncRNA). 5 These potential biomarkers will contribute to the diagnosis, monitoring or prognosis of AML.
Non-coding RNA (ncRNA) is a class of non-protein translation RNA, which includes long lncRNA, miRNA and circRNA. CircRNAs are characterized by covalent closed loops and are formed by backsplicing without 3'-end and 5'-end. 6 Nowadays, circRNAs are common in eukaryotes and are abundant and conserved in cytoplasm and blood. 7 Circ-ANAPC7 is reported to be up-regulated in AML. 8 In addition, circRNA-DLEU2 expression is enhanced in AML tissues and cells, which promotes proliferation and inhibits cell apoptosis of AML cells. It accelerates human AML in vivo by reducing miR-496 and enhancing PRKACB expression. 9 Knockdown of circ_0009910 represses AML cell growth in vitro and in vivo via increasing miR-20a-5p. 10 Considering the expression change of circRNAs in AML, the expression change of circRNAs might be involved in leukaemogenesis. However, there are still many circRNAs that need to be investigated, which may participate in the development or response to the treatment of AML.
In this study, circRNA microarray GSE11 6617 and circRNA sequencing are used to identify the expression of circRNAs in peripheral blood samples from 8 patients with AML and 4 normal individuals. Furthermore, circ_406083 (Gene symbol: TASP1) was identified to be significantly up-regulated in patients with AML compared with normal, which was consistent with a previous study. 11 In addition, high levels of circTASP1 predicted the poor prognosis of patients with AML. Further in vitro and in vivo experiments confirmed that down-regulation of circTASP1 inhibited the proliferation of AML cells and promoted apoptosis by increasing the expression of miR-515-5p in AML cells, thereby providing a novel target for AML treatment.

| MATERIAL S AND ME THODS
In this study, the animal experiments were conducted in accordance with the Guide for the Care and Use of Laboratory Animals, and the acquisition of human samples was based on the principle of Declaration of Helsinki and was authorized by the Medical Ethics Committee of The First Affiliated Hospital of Nanchang University.

| RNA-FISH assay
The location of circTASP1 in cells was determined by RNA-FISH assay. Cy3-labelled circTASP1 probes were purchased from RiboBio, and the RNA-FISH assay was performed by fluorescent in situ hybridization kit (RiboBio). After incubated with prehybridization solution for 3 hours at 37℃, the slices were blocked with Cy3-labelled circTASP1 probes at 55℃ overnight. After washed with PBS for 3 times, the slices were counterstained with DAPI for 90 seconds.
Finally, the images were captured with a fluorescence microscope (Olympus).

| Cell Counting Kit-8 (CCK-8) assay
The cell viability of HL60 and THP-1 cells was detected by CCK-8 assay using Cell Counting Kit-8 (CCK-8; Beyotime). After treatment, a total of 5 × 10 3 cells were seeded in each 96-well plate and incubated for 24 hours, 48 hours and 72 hours, respectively. Each well was added with 10 μL CCK8 reagent for 1-hour incubation. Finally, the OD 450 nm value of each sample was examined by a microplate reader.

| 5-Ethynyl-20-deoxyuridine (EdU) incorporation assay
The EdU detection was performed by Cell-Light EdU DNA Cell proliferation kit (RiboBio). After incubated with 50 mmol/L EdU for 2 hours, the HL60 and THP-1 cells were fixed in 4% paraformaldehyde and then stained with Apollo Dye Solution (cell nucleus was stained by DAPI). The EdU-positive cells were photographed and counted using a fluorescence microscope.

| Apoptosis and cell cycle analysis
After washed with PBS, the HL60 and THP-1 cell apoptosis and cell cycle were detected by Annexin V-FITC Apoptosis Detection Kit (Beyotime) and EZCellTM Cell Cycle Analysis Kit (BioVision), respectively. The cell apoptosis and cell cycle were examined using flow cytometer (Beckman).

| RNA immunoprecipitation (RIP) assay
EZ-magna RIP RNA-binding protein immunoprecipitation kit (Millipore) was used for RIP detection to investigate the interaction between circTASP1 and mir-515-5p. After lysis of cells, the cells were incubated with Millipore antibody against human Ago2 coated on magnetic beads in RIP buffer solution. The input and IgG were considered as controls. The RNA was isolated and reverse transcribed into cDNA, and the circTASP1 and miR-515-5p levels were analysed by qRT-PCR.

| Quantitative real-time PCR (qRT-PCR) assay
Total RNA from peripheral blood and cells was isolated by TRIzol reagent (Invitrogen). The reverse transcription kit (Takara) was used to synthesize cDNA. The gene expression was examined by qRT-PCR using SYBR-Green kit (TaKaRa) and detected using StepOnePlus™ real-time PCR system (Applied Biosystems). The GAPDH or U6 is considered as an internal control. Each sample is repeated at least 3 times.

| Tumour xenografts
A total of 10 BALB/c-Nude mice (Charles River, China) were randomly divided into two groups: sh-NC group and sh-circTASP1 group (5 mice for each group). A total of 5 × 10 6 HL60 cells were stably transfected with lentivirus wrapped sh-circTASP1 or sh-NC plasmids and then subcutaneously injected into the right flank of each mouse.
After 15 days, the tumours were measured every 5 days and then removed on 40th day.

| Statistical analysis
SPSS software 20.0 was used for statistical analyses. The t test or one-way analysis of variance was utilized to analyse the differences between groups. The correlations between circTASP1 and miR-515-5p expression, circTASP1 and HMGA2 expression, and miR-515-5p and HMGA2 expression in patients with AML were evaluated by Pearson correlation analysis. Kaplan-Meier method with log-rank test was used for survival analysis. The data (mean ± SD) were calculated, and P < .05 was considered significant.

| CircTASP1 was up-regulated in AML patients and cells
Based on the circRNA microarray GSE11 6617, 19 remarkably differentially expressed circRNAs were screened based on the fold change value between 8 patients with AML and 4 normal healthy individuals via bioinformatics analyses. The data were visualized in a clustered heat map, among which hsa_circRNA_406083 (hsa_circ_0007340; chromosome 20, 13509092 to 13561628; Gene symbol: TASP1) was highly expressed in patients with AML compared with normal individuals ( Figure 1A). The Sanger sequencing of the head-to-tail splicing in the RT-PCR product of circTASP1 was performed in Figure S2. In order to validate the bioinformatics analyses, the expression of circTASP1 was determined by qRT-PCR assay in 8 patients with AML and 4 normal individuals, and the results showed that circTASP1 was remarkably up-regulated in patients with AML compared with normal individuals ( Figure 1B). In addition, after the RNase R digestion, the expression of circTASP1 remained unchanged, indicating that circTASP1 was a circle RNA and could not be digested by RNase R ( Figure 1C). Subsequently, we evaluated circTASP1 expression in peripheral blood from 60 patients with AML and 20 normal individuals. CircTASP1 expression was obviously higher in patients with AML ( Figure 1D). In addition, the CD34 + cells were isolated from peripheral blood of 60 patients with AML and 20 normal individuals. As we expected, the cirTASP1 was up-regulated in CD34 + cells of patients with AML compared with healthy individuals ( Figure S4). According to the cut-off value, we classified the whole-cohort of patients with AML into two groups (high and low) in order to further analyse the clinical significance of circTASP1 expression in AML. High circTASP1 expression was found to be associated with CEBPα (P = .037) and haemoglobin (HGB) (P = .018, Table 1). Moreover, a multivariate analysis showed significant differences between circTASP1 and HGB (P = .002; Table 2). In addition, the Kaplan-Meier analysis indicated that high circTASP1 levels led to poor survival of patients with AML ( Figure 1E). Furthermore, circ-TASP1 was also significantly up-regulated in AML cell lines (HL60, THP-1, U937 and KG-1) compared with normal cell lines (HS-5) ( Figure 1F).
To confirm the location of circTASP1 in cells, the RNA-FISH assay was performed, and the results indicated that circTASP1 was located in the cytoplasm of HL60 and THP-1 cells ( Figure 1G). Given that circTASP1 was up-regulated in patients with AML and cells, we decided to focus on the function of circTASP1 in further experiments.

| Knockdown of circTASP1 inhibited proliferation and induced apoptosis in AML cell lines
To verify the role of circTASP1 in AML, sh-circTASP1#1 or sh-circTASP1#2 was transfected into HL60 and THP-1 cells. The results showed that sh-circTASP1#1 or sh-circTASP1#2 could significantly reduce the expression of circTASP1 but could not change the linear TASP1 mRNA levels in both HL60 and THP-1 cells (Figure 2A and Figure S1). Through CCK-8 assay, we found that knockdown of circTASP1 impaired the cell viability of HL60 and THP-1 cells ( Figure 2B). Additionally, Edu staining confirmed that knockdown of circTASP1 decreased the Edu-positive cells, indicating that the cell proliferation was inhibited by knockdown of circTASP1 in HL60 and THP-1 cells ( Figure 2C). Flow cytometry analysis results suggested that knockdown of circTASP1 increased cell apoptosis rate in HL60 and THP-1 cells ( Figure 2D). Moreover, Western blot assay indicated that knockdown of circTASP1 induced Bax and cleaved caspase 3 expression but inhibited Bcl-2 expression in HL60 and THP-1 cells ( Figure 2E). Aberrant cell cycle results in decreased proliferation and increased apoptosis. Cell cycle analysis suggested that knockdown of circTASP1 led to more cells arrested in G1 phase while less cells in S and G2 phase in HL60 and THP-1 cells ( Figure 2F). These findings indicated that knockdown of circTASP1 inhibited proliferation and induced apoptosis in HL60 and THP-1 cells.

| CircTASP1 functioned as a molecular sponge of miR-515-5p
CircRNAs may function as sponges for miRNAs and interact with RISC complex. 13 To verify the mechanism of circTASP1, the potential target of circTASP1 was predicted by circinteractome (https:// circi ntera ctome.nia.nih.gov/). Fortunately, miR-515-5p was identified as the potential target of circTASP1. Subsequently, luciferase activity assay confirmed that miR-515-5p mimic inhibited the activity of circTASP1-WT reporter in HL60 and THP-1 cells ( Figure 3A). Further, miR-515-5p mimic significantly promoted miR-515-5p expression in HL60 and THP-1 cells ( Figure 3B). To examine the interaction between miR-515-5p and circTASP1, the 3'-terminal-biotinylated-miR-5 15-5p probe or 3'-terminal-biotinylated-miR-515-5p-MUT probe was designed. The probe was validated to pull down circTASP1 in HL60 and THP-1 cells ( Figure 3C). AGO2 is a unique protein of RNA-induced silencing complex (RISC) with endonuclease activity and recognizes specific targets through base-pairing of small interfering RNA or mi-croRNA (miRNA), resulting in degradation of mRNA. 14 To validate this interaction, we performed RIP of AGO2 in HL60 and THP-1 cells, and the results showed that both circTASP1 and miR-515-5p were specifically bound by AGO2 ( Figure 3D). As shown in Figure 3E, miR-515-5p expression was promoted by the knockdown of circTASP1 in HL60 and THP-1 cells. In addition, miR-515-5p was down-regulated in peripheral blood from 60 patients with AML compared with 20 normal individuals ( Figure 3F). More importantly, miR-515-5p expression was negatively correlated with circTASP1 in peripheral blood from 60 patients with AML ( Figure 3G). Collectively, these data indicated that circTASP1 functioned as a molecular sponge of miR-515-5p.
F I G U R E 1 CircTASP1 was up-regulated in AML patients and cells. A, Heat map of 19 expressed circRNAs AML patients and normal healthy individuals according to circRNA microarray data set (GSE11 6617). B, Expression of circTASP1 was up-regulated in 8 patients with AML compared with 4 normal individuals, which was detected by qRT-PCR assay. C, CircTASP1 expression could not be influenced by RNase R digestion. D, CircTASP1 expression in patients with AML was higher than that in normal individuals as detected by qRT-PCR. E, Kaplan-Meier curve of survival of patients with AML according to circTASP1 expression (mean value as cut-off). F, CircTASP1 was up-regulated in AML cell lines (HL60, THP-1, U937 and KG-1). G, RNA-FISH was utilized to examine the localization of circTASP1 in HL60 and THP-1 cells. *P < .05; **P < .01. Each experiment was repeated 3 times

| HMGA2 was a direct target of miR-515-5p
To clarify the molecular mechanism downstream of miR-515-5p during the development of AML, the TargetScan (www.targe tscan.org) was used to predict the direct target of miR-515-5p, and the results showed that 3'-UTR of HMGA2 gene bound with miR-515-5p ( Figure 5A).
Besides, the expression of HMGA2 was positively correlated with circ-TASP1 and negatively correlated with miR-515-5p in peripheral blood from 60 patients with AML ( Figure 5I). Therefore, HMGA2 was a direct target of miR-515-5p in AML.

| Knockdown of CircTASP1 suppressed AML progression in vivo
The role of circTASP1 in AML pathogenesis was determined by in vivo model. After transfected with sh-circTASP1 or sh-NC, the HL60 cells were subcutaneously injected into mice. As illustrated in Figure 7A-C, knockdown of circTASP1 significantly reduced tumour volume and tumour weight. In addition, immunohistochemistry staining indicated that sh-circTASP1 significantly reduced Ki67

| D ISCUSS I ON
Multidisciplinary research on circRNA is increasing, indicating that circRNA can be used as regulators and valuable diagnostic markers for disease. 6 In the present study, circRNA expression profiles were screened with circRNA microarray, and the results showed that circTASP1 was significantly up-regulated in patients with AML.
Circular TASP1 (circTASP1) is encoded by TASP1 gene. 15 Protease Taspase 1 (threonine aspartase 1; TASP1) is first identified as an endopeptidase, and its modulation is responsible for the cleavage of mixed lineage leukaemia 1 (MLL1) protein to appropriately regulate HOX expression. 16 Aberrant expression of TASP1 has been found TA B L E 2 Analysis of prognostic factors in acute myeloid leukaemia

F I G U R E 2
Knockdown of circTASP1 inhibited proliferation and induced apoptosis in AML cell lines. The HL60 and THP-1 cells were transfected with sh-circTASP1#1 or sh-circTASP1#2. A, The expression of circTASP1 in HL60 and THP-1 cells was examined by qRT-PCR assay. B, The cell viability of HL60 and THP-1 cells was evaluated by CCK-8 assay. C, The HL60 and THP-1 cell proliferation was detected by Edu staining assay. D, Cell apoptosis in HL60 and THP-1 cells was determined using flow cytometry assay. E, The Bax, cleaved caspase 3 and Bcl-2 expression in HL60 and THP-1 cells were detected by Western blot. F, The cell cycle distributions in HL60 and THP-1 cells were determined using flow cytometry assay. **P < .01. Each experiment was repeated 3 times in leukaemia and alters cell cycle progression, cell proliferation and apoptosis. 17 AML is a malignant clonal disease with multiple prognostic outcome: 5-year overall survival is less than half, and approximately 20 per cent of elderly patients survive within 2 years after diagnosis. 18 CircTASP1 is one of the products of back-spliced RNAs of TASP1. 19 Considering the vital role of TASP1 in leukaemia, the potential function of circTASP1 in the prognosis of AML needs to be investigated. A recent study has demonstrated that circtasp1 is upregulated in T-cell acute lymphoblastic leukaemia (T-ALL). 11 Similarly, we found that circTASP1 expression was significantly up-regulated in patients with AML and high circTASP1 levels predicted poor survival of patients with AML, indicating that circTASP1 might be a promising diagnostic and prognostic biomarker for patients with AML. In the present study, knockdown of circTASP1 inhibited proliferation and induced apoptosis in AML cell lines, which was consistent with the function of TASP1. Our results for the first time suggested that circ-TASP1 positively regulates proliferation and apoptosis.
Previous studies have shown that the function of circRNAs is mainly to release miRNAs-mediated inhibition of corresponding genes by targeting miRNAs. 20 In AML, circ-0004136 promotes cell proliferation via sponging miR-142. 21 Circular RNA-100290 promotes cell proliferation and inhibits apoptosis through sponging miR-203 in AML. 22 Circ_0009910 suppresses cell proliferation and induces apoptosis through sponging miR-20a-5p in AML. 10 These studies prove that circRNAs can sponge miRNAs and regulate cell proliferation and apoptosis, and ultimately participate in the occurrence of AML. Mechanistically, knockdown of circTASP1 inhibited proliferation and induced apoptosis via sponging miR-515-5p in AML cells.
MiR-515-5p belongs to miR-515 family and acts as a suppressor in breast cancer and lung cancer. 23 In breast cancer and bladder cancer, miR-515-5p is involved in the ceRNA mechanism with LINC00673 and LncRNA SNHG3, respectively. 24,25 As the role of miR-515-5p in AML has never been illustrated, the data in this study proved that miR-515-5p was down-regulated in AML cells and sponged by circTASP1.
HMGA2 is a member of the high mobility group superfamily and is widely considered as a novel oncogene. 26 Accumulating studies have confirmed that HMGA2 is predominantly amplified and expressed in AML cells. 27 In addition, highly expressed HMGA2 independently predicted the adverse clinical outcome for AML. 28 Shuo Y et al have demonstrated that HMGA2 regulates AML progression via F I G U R E 3 CircTASP1 functioned as a molecular sponge of miR-515-5p. A, The binding site of miR-515-5p and circTASP1 was predicted by circinteractome, and the interaction of miR-515-5p and circTASP1 was examined by luciferase activity assay. B, QRT-PCR assay suggested that miR-515-5p mimic significantly increased miR-515-5p expression in HL60 and THP-1 cells. C, RNA pull-down assay was performed with miR-515-5p-biotin, miR-515-5p-biotin-MUT or miR-NC-biotin using HL60 and THP-1 cell extracts. RNA levels of circTASP1 in immunoprecipitates were detected by qRT-PCR. D, RIP assay indicated that circTASP1 and miR-515-5p were specifically bound by AGO2. E, QRT-PCR assay confirmed that knockdown of circTASP1 promoted miR-515-5p expression in HL60 and THP-1 cells. F, The qTR-PCR analysis of the relative expression of miR-515-5p in patients with AML and normal individuals. G, Pearson's correlation coefficients were used to evaluate the correlation between circTASP1 and miR-515-5p in patients with AML. **P < .01. Each experiment was repeated 3 times activating Wnt/β-catenin signalling. 29 In the current study, HMGA2 was identified as a direct target of miR-515-5p. Our study clarified as ceRNA for other miRNAs remains to be further studied.
In conclusion, our results showed that circTASP1 is up-regulated in AML tissues and cells, and acts as a sponge for miR-515-5p to regulate HMGA2, thereby promoting proliferation and inhibiting

F I G U R E 4
Overexpressed miR-515-5p inhibited cell proliferation and promoted cell apoptosis. The HL60 and THP-1 cells were transfected with miR-515-5p mimic or miR-NC mimic. A, The cell viability of HL60 and THP-1 cells was evaluated by CCK-8 assay. B, The HL60 and THP-1 cell proliferation was detected by Edu staining assay. C, Cell apoptosis in HL60 and THP-1 cells was determined using flow cytometry assay. D, The Bax, cleaved caspase 3 and Bcl-2 expression in HL60 and THP-1 cells were detected by Western blot. E, The cell cycle distributions in HL60 and THP-1 cells were determined using flow cytometry assay. **P < .01. Each experiment was repeated 3 times F I G U R E 5 HMGA2 was a direct target of miR-515-5p. A, The binding of miR-515-5p with HMGA2 gene 3'-UTR predicted by TargetScan. B, The direct binding between miR-515-5p and HMGA2 gene promoter was performed by dual-luciferase reporter assay. C, HMGA2 mRNA levels in HL60 and THP-1 cells transfected with miR-515-5p mimics. D, HMGA2 protein levels in HL60 and THP-1 cells transfected with miR-515-5p mimics. E, The anti-miR-515-5p decreased the miR-515-5p levels in HL60 and THP-1 cells. F, QRT-PCR assay confirmed that knockdown of circTASP1 reduced HMGA2 expression, which was reversed by anti-miR-515-5p. G, Western blot confirmed that knockdown of circTASP1 reduced HMGA2 expression, which was reversed by anti-miR-515-5p. H, QRT-PCR assay proved that HMGA2 expression in patients with AML was higher than that in normal individuals. I, Negative correlation between miR-515-5p and HMGA2 expression in patients with AML. Positive correlation between circTASP1 and HMGA2 expression in patients with AML. **P < .01, ##P < .01. Each experiment was repeated 3 times F I G U R E 6 Knockdown of circTASP1 inhibited proliferation and induced apoptosis by modulating miR-515-5p/HMGA2 pathway. A, The transfection efficiency of HMGA2 was validated by Western blot. The CCK-8 assay (B), Edu straining assay (C) and flow cytometry assay (D) indicated that knockdown circTASP1 inhibited cell viability and proliferation, while promoted apoptosis in HL60 and THP-1 cells, and when co-transfected with anti-miR-515-5p or HMGA2, these effects were reversed. E, The Bax, cleaved caspase 3 and Bcl-2 expression were determined by Western blot. F, The cell cycle distributions in HL60 and THP-1 cells were determined using flow cytometry assay. **P < .01, #P < .05, ##P < .01. Each experiment was repeated 3 times apoptosis during AML progression. More importantly, circTASP1 can serve as a potential target for AML treatment.

CO N FLI C T O F I NTE R E S T
The authors state that there are no conflicts of interest to disclose.

S TATE M E NT O F H U M A N A N D A N I M A L R I G HT S
All procedures in studies involving human participants were performed in accordance with the standards upheld by the Ethics

S TATE M E NT O F I N FO R M E D CO N S E NT
Written informed consent was obtained from a legally authorized representatives for anonymized patient information to be published in this article.

DATA AVA I L A B I L I T Y S TAT E M E N T
All data generated or analysed during this study are included in this F I G U R E 7 CircTASP1 knockdown suppressed AML progression in vivo. A, CircTASP1 knockdown significantly inhibited the tumour growth in vivo. Tumour volume (B) and tumour weight (C) of nude mice were analysed, n = 5. The expression of Ki67 (D) and cleaved caspase 3 (E) was determined by Immunohistochemistry. The expression of circTASP1 (F) and miR-515-5p (G) was determined using qRT-PCR assay. H, The expression of HMGA2 was evaluated by Western blot. **P < .01