MSC-AS1 knockdown inhibits cell growth and temozolomide resistance by regulating miR-373-3p/CPEB4 axis in glioma through PI3K/Akt pathway

50 Glioma tissues and 25 peritumoral brain edema tissues (normal tissues) were collected at Chinese PLA General Hospital. The sample tissues were collected with glioma patient consent and all participants signed informed consent forms. The detailed information of patients is provided in Table 1. Glioma patients were divided into resistant group (n = 25) and sensitive group (n = 25) based on their sensitivity to TMZ treatment. Tissues were frozen in liquid nitrogen after collection. The study was approved by the Ethics Committee of Chinese PLA General Hospital.

Human glioma cell lines LN229 and SHG-44 and renal epithelial cell line 293T were obtained from Procell (Wuhan, China), cultured in Roswell Park Memorial Institute 1640 (RPMI 1640, HyClone, South Logan, UT, USA) supplemented with 10% fetal bovine serum (FBS, Gibco, USA) at 37 °C with 5% CO₂. TMZ-resistant cell lines were established as described by Tian et al. [23]. Briefly, LN229 and SHG-44 cells were treated with an initial TMZ concentration of 1 μM for 2 weeks. Then, the TMZ dose was doubled, and each dose was hatched for 2 weeks, and the final concentration was added to 400 μM. The established TMZ-resistant cells were named as LN229/TR and SHG-44/TR.

Small interfering RNA against MSC-AS1 1, 2, and 3 (si-MSC-AS1-1, si-MSC-AS1-2, si-MSC-AS1-3) or CPEB4 (si-CPEB4) and control (si-NC), miR-373-3p mimic, inhibitor, and their matched controls (mimic NC, inhibitor NC) were synthesized by RiboBio (Guangzhou, China). Lentivirus packaged sh-MSC-AS1 plasmid (RiboBio) was used to suppress the expression of MSC-AS1, and sh-NC acted as the negative control. These structures were transfected into cells by Lipofectamine 3000 (Invitrogen, Carlsbad, CA, USA).

Glioma tissues and cells were collected, and their total RNAs were extracted by TRIzol (Invitrogen). The reverse transcription was performed using a Reverse Transcription Kit (Takara, Wuhan, China). The relative expression was detected by qRT-PCR using SYBR Green PCR Master Mix (Takara) on the ABI 7500 Fast Real-Time PCR system (Applied Biosystems, Foster City, CA, USA). U6 and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) were used as endogenous controls for miR-373-3p and MSC-AS1 or CPEB4. The primers explored were listed as below: MSC-AS1, Forward (F): 5′-GCCAGTCAGAAAATGAGGAAC-3′ and Reverse (R): 5′-CCAGTTGGGTGAACAGGAC-3′; miR-373-3p, F: 5′-GAAGTGCTTCGATTTTGCC-3′ and R: 5′-GAATACCTCGGACCCTGC-3′; CPEB4, F: 5′-TGGGGATCAGCCTCTTCATA-3′ and R: 5′-CAATCCGCCTACAAACACCT-3′; GAPDH, F: 5′-GCACCGTCAAGGCTGAGAAC-3′ and R: 5′-ATGGTGGTGAAGACGC CAGT-3′; and U6, F: 5′-CTCGCTTCGGCAGCACA-3′ and R: 5′-AACGCTTCACGAATTTGCGT-3′.

The proteins from glioma tissues and cells were lysed in RIPA buffer (Thermo Fisher Scientific, Waltham, MA, USA), separated using sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE), and then transferred to polyvinylidene fluoride membranes (PVDF, Invitrogen). After blocked with 5% skim milk powder for 2 h, the membranes were probed with primary antibodies overnight at 4 °C and maintained in horseradish peroxidase-linked secondary antibodies (1:4000, Abcam, Cambridge, MA, USA) for 1 h. Finally, the proteins were visualized by chemiluminescence and the relative protein expression was analyzed by the ImageJ software and normalized to GAPDH. The primary antibodies were myeloid cell leukemia-1 (MCL-1, 1:1000, Abcam), multidrug resistance-associated protein-1 (MRP-1, 1:500, Abcam), cyclinD1 (1:2000, Abcam), B-cell lymphoma-2-associated x (Bax, 1:2000, Abcam), caspase-3 (1:2000, Abcam), CPEB4 (1:1000, Thermo Fisher Scientific), PI3K (1:1000, Abcam), P-PI3K (1:1000, Abcam), AKT (1:1000, Abcam), P-AKT (phospho Ser473) (1:1000, Abcam), and GAPDH (1:2000, Abcam).

LN229/TR and SHG-44/TR cells after stable transfection were tiled into the 96-well plates and treated with different concentrations of TMZ (0–3 μM) for 48 h. The concentration selection was based on the study of Tan et al. [24]. Then, 10 μL cell counting kit-8 (CCK-8, Dojindo, Shanghai, China) was added to the cells and incubated for another 3 h. The absorbance at 450 nm was measured by a Microplate reader. The half maximal inhibitory concentration (IC₅₀) value of TMZ was employed to assess the sensitivity of TMZ in glioma cells.

LN229/TR and SHG-44/TR cells were inoculated at the density of 800 cells/well in 6-well plates and transfected. Cells were cultured in RPMI 1640 with 10% FBS and 300 μmol/L TMZ. 2 weeks later, the supernatant of medium was discarded and cells were fixed and stained with 0.1% crystal violet (Beyotime, Beijing, China) for 20 min. The stained and visible colonies were counted and photographed.

Cell apoptosis rate was determined by using an Annexin V-fluorescein isothiocyanate (FITC)/propidium iodide (PI) apoptosis detection kit (Keygen, Beijing, China). TMZ-resistant glioma cells were first transfected. After 48 h later, cells were harvested and washed. Then, cells were stained with 5 μL Annexin V-FITC and 5 μL PI for 15 min in the dark. Cell apoptosis was analyzed by flow cytometry (BD Biosciences, Franklin Lake, NJ, USA).

Bioinformatics analysis (starBase v3.0) was used to predict targets of miR-373-3p. To verify the potential interaction between miR-373-3p and MSC-AS1 or CPEB4, we subcloned MSC-AS1 or CPEB4 with site-directed mutation or not in the miR-373-3p binding site into the pmirGLO vector (Promega, Fitchburg, WI, USA). Therefore, luciferase activity represents the expression and activity of miR-373-5p. The products were named as MSC-AS1 wt, MSC-AS1 mut, CPEB4 3′UTR wt, and CPEB4 3′UTR mut, and they were co-transfected into HEK-293T cells with miR-373-3p mimic or mimic NC, respectively. 24 h later, the cells were lysed and the luciferase activity was examined by a dual-luciferase reporter assay kit (Promega).

EZMagna RIP kit (Millipore, Billerica, MA, USA) was used in RIP assay. LN229/TR and SHG-44/TR cells were collected and cleaved by cell lysis, following incubated with magnetic beads conjugated to Argonaute-2 (Ago2) antibody or immunoglobulin G (IgG) antibody. Subsequently, the RNAs from the beads were purified and extracted by TRIzol. The levels of MSC-AS1 and miR-373-3p were measured by qRT-PCR.

LN229/TR and SHG-44/TR cells transfected with sh-MSC-AS1 or sh-NC were subcutaneously injected into 4-week-old male nude mice (n = 6 each group). At 10 days after the injection, tumor volume was measured and then detected every other 4 days. TMZ (25 mg/kg) was intraperitoneally injected into the mice every day when the tumors reached 100 mm³. The volume of the tumor was calculated by following formula: volume (mm³) = (longest tumor diameter × shortest tumor diameter²)/2. 30 days later, all mice were euthanized and tumor weight was measured. In addition, tumor tissues were stripped to extract the RNAs and proteins. The animal experiments were approved by the Animal Care and Use Committee of Chinese PLA General Hospital.

Data were acquired with at least three replicates and shown as the mean ± standard deviation (SD). Comparisons among multiple groups were evaluated by one-way analysis of variance (ANOVA) and the differences between two groups were analyzed by Student’s t-test through GraphPad Prism 7. Statistical significance was defined as the P value less than 0.05.

To identify glioma resistance-related lncRNAs, we analyzed the expression profile of lncRNAs in the GSE113510↗ microarray dataset downloaded from the GEO database. As shown in Fig. 1a, 40 lncRNAs were up-regulated in TMZ-resistant glioma cells (229R). Excluding the miRNAs which its roles in cancers have not previously been reported, MSC-AS1 was the top up-regulated lncRNA. Research has shown that MSC-AS1 acts as a tumor promoter in hepatocellular carcinoma [25], nasopharyngeal carcinoma [26], and kidney renal clear cell carcinoma [12]. However, the effect of MSC-AS1 in glioma cancer is still limited. Thus, MSC-AS1 was selected for further research. We used qRT-PCR to validate MSC-AS1 expression in glioma tissues, and MSC-AS1 was significantly higher in glioma tissues than that normal tissues, and higher level of MSC-AS1 was observed in TMZ-resistant glioma tissues compared to TMZ-sensitive glioma tissues (Fig. 1b). Consistently, qRT-PCR results showed that MSC-AS1 expression was markedly increased in TMZ-resistant (TR) glioma cells (LN229/TR and SHG-44/TR) compared with their parental cells (LN229 and SHG-44) (Fig. 1c). Besides, glioma patients with high MSC-AS1 expression showed a shorter 5-year overall survival than that in glioma patients with low MSC-AS1 expression (Fig. 1d). And the receiver operating characteristic (ROC) analysis of the sensitivity and specificity showed that the area under the RPC curve (AUC) was 0.8052 for MSC-AS1 in detecting glioma patients from the healthy people (Fig. 1e).

MCL-1, an anti-apoptotic member of the BCL-2 family proteins, is frequently overexpressed in a variety of cancers [27]. A number of studies have confirmed the pivotal roles of MCL-1 in cancer cell survival and apoptosis resistance, as in vitro evidence documenting the pro-apoptotic and chemosensitization effects of Mcl-1 knockdown [28–30]. In TMZ-resistant glioma tissues, the expression levels of resistant marker proteins MCL-1 and MRP-1 were increased (Fig. 1f). To confirm that we have successfully generated the TMZ-resistant glioma cell lines, we also measured the expression of resistant marker proteins. As shown in Fig. S1a, the protein levels of MCL-1 and MRP-1 were distinctly facilitated in LN229/TR and SHG-44/TR cells in contrast to their parental cells. In addition, the viability of LN229/TR and SHG-44/TR cells was drastically enhanced in a TMZ (0–400 μM) dose-dependent manner, and the half maximal inhibitory concentration (IC₅₀) value of TMZ in LN229/TR and SHG-44/TR cells was increased (Fig. S1b), suggesting TMZ-resistant glioma cell lines were successfully established. The results demonstrated that MSC-AS1 might play an important role in glioma resistance to TMZ.

Given that MSC-AS1 was highly expressed in TMZ-resistant glioma cell lines, we knocked down MSC-AS1 in LN229/TR and SHG-44/TR cells by transfection of si-MSC-AS1-1, si-MSC-AS1-2, and si-MSC-AS1-3. The efficiency of MSC-AS1 knockdown was determined by qRT-PCR. The data showed that MSC-AS1 expression was significantly decreased in LN229/TR and SHG-44/TR cells transfected with si-MSC-AS1-1, si-MSC-AS1-2, and si-MSC-AS1-3 compared to control group and the si-MSC-AS1-1 had the most obvious inhibitory effect (Fig. 2a, b). When MSC-AS1 was silenced in LN229/TR and SHG-44/TR cells, TMZ could decrease cell viability in a dose-dependent manner and the IC₅₀ value of TMZ was reduced (Fig. 2c). Furthermore, colony formation assay showed that repression of MSC-AS1 degraded the rate of colony formation in LN229/TR and SHG-44/TR cells (Fig. 2d). On the contrary, flow cytometry assay demonstrated that interference with MSC-AS1 fortified the apoptosis rate of LN229/TR and SHG-44/TR cells (Fig. 2e). Meanwhile, western blot results indicated that the levels of MCL-1, MRP-1, and cell cycle-related protein cyclinD1 were down-regulated in LN229/TR and SHG-44/TR cells transfected with si-MSC-AS1-1 with respect to that cells transfected with si-NC, and the levels of pro-apoptotic proteins Bax and caspase-3 were up-regulated. In addition, the PI3K/Akt pathway-related proteins P-PI3K and P-AKT were markedly decreased (Fig. 2f). Moreover, we also utilized the si-MSC-AS1-3 to knockdown the expression of MSC-AS1. The data also showed that the cell colony was significantly decreased after silencing MSC-AS1 (Fig. 2a). These results suggested that silencing MSC-AS1 could suppress cell growth and TMZ resistance in TMZ-resistant glioma cells through PI3K/Akt pathway.

Mounting evidence has revealed that lncRNAs act as molecular sponges and effectively inhibit miRNA function [8, 9]. Based on online lncRNA-miRNA binding databases (starBase v2.0), more than 50 miRNAs have the potential binding sites for MSC-AS1. We performed a systematic Literature search in pumped and screened 6 miRNAs (miR-29b-3p, miR-142-5p, miR-373-3p, miR-302a-3p, miR-524-3p, and miR-330-3p), which was down-regulated in glioma specimens. qRT-PCR assay was performed to detect the expression of these miRNAs in glioma specimens and TMZ-resistant glioma tissues. As shown in Fig. S3a, compared with normal tissues, miRNAs expression was noticeably decreased in glioma tissues. While in TMZ-resistant glioma tissues, the expression of miRNAs was further reduced compared to TMZ-resistant glioma tissues (Fig. S3b). Besides, the expression of miRNAs was also reduced in TMZ-resistant glioma cells LN229/TR (Fig. S3c) and SNG-44/TR (Fig. S3d). Among the top de-regulated lncRNAs was miR-373-3p. Considering miR-373 was decreased in glioma [15], and it could elevate the chemosensitivity of pancreatic carcinoma cells [18], we further investigated whether miR-373-3p was involved in glioma resistance to TMZ. We first measured miR-373-3p expression in glioma tissues by qRT-PCR. As presented in Fig. 3a, miR-373-3p expression was markedly reduced in glioma tissues and TMZ-resistant glioma tissues compared to normal tissues and TMZ-sensitive glioma tissues. Meanwhile, we observed that miR-373-3p expression was down-regulated in LN229/TR and SHG-44/TR cells versus their parental cells (LN229 and SHG-44) (Fig. 3b). Interestingly, miR-373-3p was predicted as a potential target gene for MSC-AS1 using starBase v3.0 online database (Fig. 3c). To verify whether miR-373-3p was a direct target gene of MSC-AS1, dual-luciferase reporter assay was performed in 293T cells. As shown in Fig. 3d, miR-373-3p mimic degraded the luciferase activity of MSC-AS1 wt, but not that of the MSC-AS1 mut, indicating a direct binding between them. RIP assay demonstrated that the enrichments of MSC-AS1 and miR-373-3p were enhanced in LN229/TR and SHG-44/TR cells combined with Ago2 antibody compared with the control group (Fig. 3e). These findings supported that MSC-AS1 could interact with miR-373-3p in TMZ-resistant glioma cells.

We then evaluated whether MSC-AS1 regulated the proliferation, apoptosis, and TMZ resistance by targeting miR-373-3p in TMZ-resistant glioma cells. LN229/TR and SHG-44/TR cells were co-transfected with si-MSC-AS1-1 and miR-373-3p inhibitor. The results showed that miR-373-3p inhibitor could reverse the facilitated impact of si-MSC-AS1-1 on miR-373-3p expression (Fig. 4a). The cell viability and TMZ IC₅₀ value (Fig. 4b) as well as the rate of colony formation (Fig. 4c) of LN229/TR and SHG-44/TR cells were obviously repressed by silencing MSC-AS1, whereas facilitated by interfering with miR-373-3p. The promoting effect of MSC-AS1 knockdown on TMZ-resistant glioma cells apoptosis could be counteracted by suppression of miR-373-3p (Fig. 4d). Meanwhile, MSC-AS1 knockdown decreased the levels of MCL-1, MRP-1, cyclinD1, P-PI3K, and P-AKT, and raised Bax and caspase-3 levels, whereas counteracted by inhibition of miR-373-3p (Fig. 4e). These results supported that MSC-AS1 knockdown inhibited cell proliferation and induced TMZ sensitivity and cell apoptosis in TMZ-resistant glioma cells by targeting miR-373-3p through regulating PI3K/Akt pathway.

In view of the key role of CPEB4 in glioma [21], we wondered whether there was a connection between CPEB4 and miR-373-3p. The expression of CPEB4 was first detected in glioma tissues. As shown in Fig. 5a, b, CPEB4 was up-regulated in glioma tissues and TMZ-resistant glioma tissues at mRNA and protein levels. Consistent with the results in tissues, the mRNA and protein levels of CPEB4 in LN229/TR and SHG-44/TR cells were notably enhanced than their parental cells (Fig. 5c, d). Since CPEB4 and miR-373-3p had opposite expression patterns, we then predicted the target genes of miR-373-3p by starBase v3.0 and found that there were complementary sites that bind to miR-373-3p in the 3′UTR of CPEB4 (Fig. 5e). Dual-luciferase assay revealed that the luciferase activity of CPEB4 3′UTR wt was degraded by miR-373-3p mimic, but there was no significant change in the luciferase activity of CPEB4 3′UTR mut in 293T cells (Fig. 5f), indicating CPEB4 was the target for miR-373-3p.

We further investigated the combined effect of miR-373-3p and CPEB4 on the proliferation, apoptosis, and TMZ resistance of TMZ-resistant glioma cells. As displayed in Fig. 6a, b, the mRNA level and protein level of CPEB4 were obviously down-regulated in glioma cells transfected with si-CPEB4, while this promoting effect was reversed in glioma cells co-transfected with si-CPEB4 and miR-373-3p. CCK-8 showed that the cell viability and IC₅₀ value of TMZ were restrained by CPEB4 knockdown, whereas facilitated by miR-373-3p suppression in LN229/TR and SHG-44/TR cells (Fig. 6c). The inhibition of si-CPEB4 on the number of cell colony (Fig. 6d) and the promotion on cell apoptosis (Fig. 6e) could be reversed by miR-373-3p deficiency. Moreover, knockdown of CPEB4 degraded the levels of MCL-1, MRP-1, P-PI3K, P-AKT, and cyclinD1, and augmented the levels of Bax and caspase-3, whereas these effects were weakened by silencing miR-373-3p in LN229/TR and SHG-44/TR cells (Fig. 6f, g). These data suggested that CPEB4 could be targeted by miR-373-3p to regulate the proliferation, apoptosis, and TMZ resistance in TMZ-resistant glioma cells via activating PI3K/Akt pathway.

Since MSC-AS1 could act as miR-373-3p sponge, and CPEB4 was the target for miR-373-3p, we further explored whether MSC-AS1 could regulate CPEB4 expression by targeting miR-373-3p. As shown in Fig. 7a, b, transfection with si-MSC-AS1-1 distinctly decreased the mRNA and protein levels of CPEB4. We also detected the CPEB4 expression was significantly down-regulated in si-MSC-AS1-3 group shown in Fig. S2b. However, the inhibitory effect of si-MSC-AS1-1 on CPEB4 expression was counteracted by miR-373-3p inhibitor in LN229/TR and SHG-44/TR cells, suggesting that MSC-AS1 could regulate CPEB4 expression by binding to miR-373-3p in TMZ-resistant glioma cells.

To further explore the impact of MSC-AS1 knockdown on the growth of implanted LN229/TR tumors in vivo, LN229/TR cells stably transfected with sh-NC or sh-MSC-AS1 were inoculated into the flanks of nude mice. The results showed that MSC-AS1 knockdown inhibited tumor growth by reducing tumor volume and weight (Fig. 8a, b). qRT-PCR results indicated that lower levels of MSC-AS1 and CPEB4 were observed in tumors of LN229/TR-sh-MSC-AS1 group than that in the tumors of LN229/TR-sh-NC group, and higher level of miR-373-3p was shown in tumors of LN229/TR-sh-MSC-AS1 group (Fig. 8c), and the protein expression of CPEB4 was also decreased in LN229/TR-sh-MSC-AS1 group (Fig. 8d). Consistently, SHG-44/TR stably transfected with sh-NC or sh-MSC-AS1 was inoculated into the flanks of nude mice. Tumor volume and weight were reduced by MSC-AS1 inhibition (Fig. 8e, f). In addition, MSC-AS1 and CPEB4 levels were down-regulated and miR-373-3p was up-regulated in SHG-44/TR-sh-MSC-AS1 group (Fig. 8g); the protein expression of CPEB4 was also degraded in SHG-44/TR-sh-MSC-AS1 group (Fig. 8h). Additionally, compared to control group, the protein levels of MCL-1, MRP-1, P-PI3K, P-AKT, and cyclinD1 were dwindled in LN229/TR and SHG-44/TR-sh-MSC-AS1 groups, and the levels of Bax and caspase-3 were promoted (Fig. 8i). Therefore, MSC-AS1 knockdown impaired the growth of TMZ-resistant glioma cells in vivo by sponging miR-373-3p and hampering CPEB4 expression via PI3K/Akt pathway.

Below is the link to the electronic supplementary material.

All the authors declare that they have no conflict of interest or financial ties to disclose.