miR‐152‐5p suppresses glioma progression and tumorigenesis and potentiates temozolomide sensitivity by targeting FBXL7

Abstract A generally used chemotherapeutic drug for glioma, a frequently diagnosed brain tumour, is temozolomide (TMZ). Our study investigated the activity of FBXL7 and miR‐152‐5p in glioma. Levels of microRNA‐152‐5p (miR‐152‐5p) and the transcript and protein of FBXL7 were assessed by real‐time PCR and Western blotting, respectively. The migratory and invasive properties of cells were measured by Transwell migration and invasion assay and their viability were examined using CCK‐8 assay. Further, the putative interaction between FBXL7 and miR‐152‐5p were analysed bioinformatically and by luciferase assay. The activities of FBXL7, TMZ and miR‐152‐5p were analysed in vivo singly or in combination, on mouse xenografts, in glioma tumorigenesis. The expression of FBXL7 in glioma tissue is significantly up‐regulated, which is related to the poor prognosis and the grade of glioma. TMZ‐induced cytotoxicity, proliferation, migration and invasion in glioma cells were impeded by the knock‐down of FBXL7 or overexpressed miR‐152‐5p. Furthermore, the expression of miR‐152‐5p reduced remarkably in glioma cells and it exerted its activity through targeted FBXL7. Overexpression of miR‐152‐5p and knock‐down of FBXL7 in glioma xenograft models enhanced TMZ‐mediated anti‐tumour effect and impeded tumour growth. Thus, the miR‐152‐5p suppressed the progression of glioma and associated tumorigenesis, targeted FBXL7 and increased the effect of TMZ‐induced cytotoxicity in glioma cells, further enhancing our knowledge of FBXL7 activity in glioma.

F-box proteins characteristically possess a 40-amino acid F-box domain with NH2-terminal that links to Cul1-Rbxl by binding to Skp1. 9 These F-box proteins are further sub-categorized as subfamilies L (containing a motif of leucine-rich repeats), W (containing a motif of WD repeats) and O (F-box only) based on the carboxyl-terminal domain. 10,11 Their WD/leucine-rich repeat motif enables these proteins to recognize an array of substrates. 12 The association between F-box protein and its substrates has facilitated better information on several SCF E3 ligases mediated biological functions, such as inflammation, mitotic cell cycle progression and gene expression. 13 FBXL7 has been shown to target Aurora kinase A for proteasomal degradation, 14 and polyubiquitylation, which has a crucial function in formation of mitotic spindles and segregation of chromosome. 15 Additionally, whereas FBXL7 exhibits pro-apoptotic activity, the role of FBXL7 on glioma remains to be extensively unravelled. 15,16 The small (nearly 22 nucleotides long), noncoding, endogenous transcripts, called the microRNAs (miRNAs), participate in translational repression and gene silencing via binding with target mRNAs. 17,18 Further, miRNAs have been closely associated with progression, initiation, diagnosis, and a prognosis of cancer and cytotoxicity induced by drugs in several cancerous states, such as glioma. 19,20 As miR-152 was first discovered in mouse colon in 2002, 21 more and more studies have shown that miR-152 is a tumour suppressor that is involved in cell proliferation, invasion and migration of various cancers, including ovarian, gastric and liver carcinomas. 22 In humans, the miR-152 gene is located at 17q21.32. 23 After transcription and nuclear cleavage, the precursor miR-152 (pre-miR-152) is transported to the cytoplasm and is cleaved by Dicer into miR-152 duplex. 24 Finally, two mature of lengths and sequences were generated from opposite arms of the miR-152 duplex, called miR-152-3p and miR-152-5p, respectively. 25 However, the role of miR-152-5p in glioma is not fully understood. Therefore, in this study, we examined the function of miR-152-5p in regulating the growth and progress of tumours by targeting FBXL7 as well as the function of miR-152-5p/FBXL7 axis on cytotoxicity induced by TMZ in glioma. We observed that miR-152-5p lies upstream of FBXL7 and regulates FBXL7 level and overexpressed miR-152-5p repressed glioma development and progression and targets FBXL7 to augment cytotoxicity induced by TMZ in glioma cells.

| Cell lines and samples of tissues
The tissues samples of glioma and NATs (normal adjacent brain tissues) were obtained from glioma patients at The Second Hospital of Hebei Medical University. The Ethics Committee of The Second Hospital of Hebei Medical University authorized this study. Before our study, each patient gave informed written consent.
NHAs (normal human astrocytes) were procured from Lonza (Switzerland), and H4, A172, LN229, U87 and U251 cell lines were got from ATCC (USA) and cultured using Astrocyte Growth Medium Bullet Kit from Lonza with media for astrocyte growth as well as supplements. Cell lines U87 and U251 were grown in Dulbecco's Modified Eagle's Medium containing foetal bovine serum (FBS, 10%) (Thermo Fisher Scientific).

| Determination of mRNA levels by reverse transcription quantitative PCR assay
Real-time PCR was performed as previously reported. 26

| Western blotting
Western blotting was performed as previously studied. 28,29 Using RIPA Lysis and Extraction Buffer, glioma cells were lysed in a system con- Blocking of the membrane was performed in skimmed milk (5%) and probed at 4°C overnight with primary antibodies (Abcam) against FBXL7 or β-actin (Abcam). Next, secondary antibody (goat-anti-rabbit) and HRP horseradish peroxidase-conjugated) (Abcam) was added.

| Reporter assays
FBXL7 3′-UTR partial wild-type (WT) sequence harbouring the potential binding site for miR-152-5p and its mutant sequences were cloned singly in luciferase vector psiCHECK-2 and FBXL7 3′-UTR-WT and FBXL7 3′-UTR-MUT to generate luciferase reporters, respectively. Then, the transfection of either reporter was performed into U87 and U251 cells along with the mimic of miR-152-5p or miR-con (negative control), and the luciferase activity was measured.

| Assay for cell invasion and migration
Seeding of cells transfected in a medium free of serum was performed in the upper chamber of Transwell (Corning) Matrigelprecoated (to assess invasion) or non-coated (to assess migration).
Then, 20% FBS containing medium was plated into chamber sublayers. After incubating for one full day, cotton swabs were used to erase cells on membrane upper surface, and fixing, staining and imaging of cells penetrating the membranes were performed followed by counting them in 15 random fields.

| Estimation of cell cytotoxicity and proliferative ability
Assessment of cell cytotoxicity and proliferative ability was examined using the CCK-8 assay kit from Sigma. In brief, seeding of cells was performed with 96-well plates. After infection or treatment, 10 μL of CCK-8 solution was added at specific time-points per well, kept for 120 min and the optical density was estimated at 450 nm. (lenti-miR-152-5p) in lentivirus and its control (lenti-miR-con) were constructed. Mice were arbitrarily categorized into groups indicating sh con, TMZ, sh con + TMZ, TMZ + sh FBXL7, sh FBXL7, lenti-miRcon, lenti-miR-con + TMZ, TMZ + lenti-miR-152-5p, lenti-miR-152-5p groups and the control group (saline), with six mice per group. U87 cells (Uninfected) or cells (5 × 10 5 ) infected with sh FBXL7, sh con, lenti-miR-152-5p lentiviruses or lenti-miR-con (5 × 10 5 /each) were given administered subcutaneously into the six-week-old male nude mice (Balb/c athymic) in its right hind limb. Tumour volume was recorded every 2 days, and the volume of tumours was estimated using a caliper and the formula: volume = ½ × length × width 2 .

| Mouse model
Seven days later, intraperitoneal administration of TMZ (20 mg/kg) was performed into mice each day for 10 days post-cell injection.
Nineteen days after injection, excision and weighing of tumours were performed.

| Statistical analysis
Each experiment was repeated thrice and expressed as means ± SD Analyses of data were performed using GraphPad Prism (USA).
Analysis of statistical difference was performed using one-way ANOVA along with Tukey's post hoc test or Student's t test, keeping P < .05 as statistically significant.

| The expression of FBXL7 was significantly enhanced in glioma cells and correlated with the grade of glioma and patient survival
The results of real-time PCR assay in different grades of glioma tissues (Grades I, II, III and IV according to WHO) and adjacent normal brain tissues (NAT) to determine level of FBXL7 expression reveal significantly up-regulated FBXL7 expression in glioma tissues all four grades compared with NAT specimens ( Figure 1A), which improved progressively as the glioma grades increased ( Figure 1A). To further assess this association, the glioma specimens were categorized into high and low FBXL7 groups with the cut-off point being the mean of FBXL7 expression. Poor survival in cases with higher FBXL7 expression than those with lower levels in the glioma patients was observed on Kaplan-Meier survival analysis ( Figure 1B; P < .001).
Furthermore, on analysing Grade IV of glioma samples, the cases with higher FBXL7 expression exhibited a shorter survival compared with those in low expression groups ( Figure 1C; P < .001). In addition, glioma cells exhibited a remarkably up-regulated FBXL7 level than that in NHAs (normal human astrocytes) ( Figure 1D,E). The above data indicated that FBXL7 level was increased in glioma tissues and cells, and relative to with the grade and poor prognosis of glioma.

| Knock-down of FBXL7 in glioma cells impeded proliferation, migration and invasion and enhanced TMZ-cytotoxicity
Next, the mRNA and protein levels of FBXL7 were estimated through real-time PCR and Western blotting experiments to be reduced significantly in sh FBXL7 lentiviruse-infected U87 and U251 cells than that in sh con lentiviruse-infected cells (Figure 2A
Additionally, with enhancement in tumour grade of glioma tissues, gradual down-regulation of miR-152-5p was observed ( Figure 3D).

U251 cells, and significantly enhanced in cells depleted with miR-
152-5p ( Figure 3F), and this relationship was observed in glioma tissues of Grade IV (WHO) ( Figure 3G). Thus, our findings indicated that FBXL7 is clearly a target of miR-152-5p in glioma.

| MiR-152-5p overexpression enhanced TMZ sensitivity in glioma cells by targeting FBXL7
Subsequently, the overexpression of FBXL7 in U87 and U251 cells could abolish miR-152-5p-mediated repression of FBXL7, as confirmed by Western blotting and real-time PCR (Figure 5A,B).

| miR-152-5p overexpression or FBXL7 knockdown in U87 cell-derived models of glioma xenograft improved anti-tumour effect mediated by TMZ and inhibited tumour growth
We further observed that the silencing of FBXL7, TMZ stimula-

| D ISCUSS I ON
Gliomas are characterized as anaplastic astrocytomas, well-differentiated low-grade astrocytomas, and glioblastoma multiforme (GBM), a highly aggressive and deadly brain tumour in adults. 30,31 The treatment of GBM using post-surgical chemotherapeutic agents and radiotherapy provides limited relief and the median survival of less than 1 year in GBM patients. 6,32,33 FBXL7 is evolutionally highly conserved protein with a sequence identity of 98% between mouse and humans, and >95% in other species. 34 However, in terms of the biological and regulatory role in cells, the characterization of FBXL7 remains relatively poor. 16,35 Extensive investigation is required to investigate whether FbxL7 is a part of a complex that initiates mitotic arrest causing cell death, or it acts as an oncoprotein. In the current study, we observed markedly up-regulated expression of FBXL7 glioma cells, in accordance with earlier reports. 16,36 Further, FBXL7 expression correlated with the grade of glioma and poor prognosis. 15 In vitro functional analysis showed that knock-down of FBXL7 repressed invasion, prolifera- In contrast, induced expression of FBXL7, indicating the effect of endogenous miR-152-5p on FBXL7 expression.
In several cancers, much focus has been paid on the examining molecular biomarkers carrying vital information in glioma prevention, diagnosis and prognosis, as well as treatment. 37  that was TMZ-resistant; however, no significant TMZ sensitivity restoration was seen in associated clinical trials. Therefore, there is a need for developing new approaches to make patients more sensitive to efficient TMZ chemotherapy. 45,46 We could show here after miR-152-5p overexpression, the suppression of glioma cell migration, proliferation and invasion, as well prognosis. However, the downstream or upstream regulatory molecules or pathways in gliomas must be explored to assess the molecular basis of the biomarker capacity of FBXL7. Also, the combined effect of TMZ, miR-152-5p, and FBXL7 on xenograft growth, glioma cell migration, proliferation, invasion and TMZ resistance must be investigated.

ACK N OWLED G EM ENTS
This research was supported by the National Natural Science

CO N FLI C T O F I NTE R E S T
There is no conflict of interest.

AUTH O R CO NTR I B UTI O N S
SK, YF, BW, YC, RH and ZZ performed the experiments. SK and ZZ designed the study. SK and YF analysed the data. SK and ZZ wrote the draft. All authors reviewed and approved the manuscript.
F I G U R E 6 FBXL7 knock-down or miR-152-5p overexpression inhibited tumour growth and enhanced TMZ-mediated antitumour effect in U87 cell-derived glioma xenograft models. (A-F) Uninfected U87 cells or U87 cells infected with sh con, sh FBXL7, lenti-miR-con or lenti-miR-152-5p lentiviruses were injected subcutaneously into the right hind limb of mice. One week later, TMZ was administered intraperitoneally into mice daily at a dose of 20 mg/kg bodyweight for a total of 10 days. (A, C and E) Tumour volume was measured every 2 days using a caliper.
(B, D and F) At the end of experiments, tumours were excised and weighed. **P < .01

DATA AVA I L A B I L I T Y S TAT E M E N T
The data used to support the findings of this study are available from the corresponding author upon request.