Knockdown of lncRNA BDNF-AS inhibited the progression of multiple myeloma by targeting the miR-125a/b-5p-BCL2 axis

A total of 30 MM patients (14 males and 16 females) and 30 healthy donors (15 males and 15 females) were recruited at the Shidong Hospital Affiliated to University of Shanghai for Science and Technology between November 2017 and August 2019. Written informed consent was obtained from all participants. This study was approved by the Human Ethics Committee of the aforementioned hospital and conducted in accordance with the Declaration of Helsinki (2000). The serum samples were collected from MM patients and healthy donors, then the supernatant serum was collected through centrifugation at 3000 g for 15 min, followed by transferring into an RNase-free Eppendorf tube and stored in − 80 °C for future use. All patients did not receive any treatment before sample collection. The expression of BDNF-AS and clinicopathological characteristics of MM patients were shown in Supplementary Table 1.

Four MM cell lines (ARP-1, MM.1S, U266, and NCI-H929), normal plasma cell line (nPC), and HEK-293 T cells were purchased from the American Type Culture Collection (ATCC, Manassas, VA, USA). All cells were cultured in RPMI-1640 medium (GIBCO, Life Technologies, Carlsbad, CA, USA) supplemented with 10% fetal bovine serum (FBS; GIBCO) and 100 U/ml penicillin/streptomycin (GIBCO) at 37 °C in a humidified incubator with 5% CO₂.

The lentiviral vector carrying short hairpin RNA targeting BDNF-AS (sh-BDNF-AS), short hairpin RNA targeting Bcl-2 (sh-Bcl-2), and negative control (sh-NC) were designed and synthesized by GenePharma Co., Ltd. (Shanghai, China). The miR-125a-5p mimics, miR-125b-5p mimics, miR-125a-5p inhibitor, miR-125b-5p inhibitor, and their corresponding negative controls (miR-NC and inhibitor NC) were purchased from Ribobio (Guangzhou, China). Approximately 1.5 × 10⁷ viral particles and 100 pM miRNA mimics or inhibitor were transfected into HEK-293 T, MM.1S, and U266 cells using Lipofectamine 2000 (Thermo Fisher Scientific). After 48 h of transfection, puromycin (Sigma, USA) was added to screen HEK-293 T, MM.1S and U266 cells stably transfected with lentiviruses carrying sh-BDNF-AS, sh-Bcl-2, or sh-NC. The transfection efficiency was confirmed by qRT-PCR. The sequences of oligos used in this study were as follows: 5′-UCCCUGAGACCCUAACUUGUGA-3′ for miR-125b-5p mimic, 5′-TCACAAGTTAGGGTCTCAGGGA-3′ for miR-125b-5p inhibitor, 5′-UCCCUGAGACCCUUUAACCUGUGA-3′ for miR-125a-5p mimic, 5′-UCACAGGUUAAAGGGUCUCAGGGA-3′ for miR-125a-5p inhibitor, 5′-UUCUCCGAACGUGUCACGUTT-3′ for miR-NC, 5′-CAGUACUUUUGUGUAGUACAA-3′ for inhibitor-NC, 5′-GGCTCACCAGTTGTTTGTT-3′ for sh-BDNF-AS, and 5′-UAAGGCUAUGAAGAGA UAC-3′ for sh-NC.

Total RNAs were extracted using TRIzol reagent (Thermo Fisher, USA) and reversely transcribed into complementary DNA (cDNA) using a PrimeScript Reverse Transcription Reagent kit (Takara, Beijing, China). Then qRT-PCR reactions were prepared using a SYBR Green PCR kit (Applied Biosystems) and performed on an ABI Prism 7500 system (Applied Biosystems). The relative expression levels of genes were calculated using the 2^-ΔΔCt method with GAPDH and U6 as the internal reference for lncRNA, Bcl-2 and miRNA, respectively. The sequences of primers used in this study were as follows: miR-125a-5p forward 5′-ACACTCCAGCTGGGC AGCAGCACACTGTGG-3′ and reverse: 5′-TGGTGTCGTGGAGTCG-3′, miR-125b-5p forward 5′-ACACTCCAGCTGGGCAGCAGCACACTGTGG-3′ and reverse 5′-TG GTGTCGTGGAGTCG-3′, U6 forward 5′-TGCGGGTGCTCGCTTCGGCAGC-3′ and reverse 5′-CCAGTGCAGGGTCCGAGGT-3′, BDNF-AS forward 5′-TTGATGGCTT GAACATTTGG-3′ and reverse 5′-TCGTGATT TTCGGTGTCTGT-3′, and GAPDH forward 5′-GAGTCAACGATTTGGTCGT-3′ and reverse 5′-GACAAGCTTCCCGTT CTCAG-3′.

Total proteins were extracted using RIPA lysis buffer. Protein concentration was determined using a BCA Protein Assay Kit (Tiangen, Beijing, China). Equal amount of protein samples were separated by 10% SDS-PAGE and transferred onto PVDF membranes (Millipore, USA). After blocking in 3% BSA, the membranes were incubated with specific primary antibodies against cleaved caspase 3, total caspase 3, cleaved PARP, total PARP, Bcl-2, and GAPDH (all from Abcam, MA, USA) at 4 °C overnight. After washing TBST for 3 times, the membranes were incubated with HRP-conjugated secondary antibody at room temperature for 2 h. Protein bands were visualized using an enhanced chemiluminescence (ECL) kit (Millipore, Billerica, MA, USA), and the band intensity was analyzed using ImageJ software (National Institutes of Health, Bethesda, MD).

Cell viability was detected using Cell Counting Kit-8 Reagent (CCK-8, Dojindo, Japan). In brief, the transfected MM.1S and U266 cells were plated into a 96-well plate and cultured for 24, 48, 72, and 96 h. Then 10 μl CCK8 solution was added into each well and incubated for another 4 h. Finally, the absorbance at 450 nm was measured using a microplate reader.

The binding sites between miR-125a/b-5p, or miR-125a/b-5p and Bcl-2 were predicted by Starbase 3.0 software (http://​starbase.​sysu.​edu.​cn/​). To determine their relationship, the entire Bcl2 3′-UTR (NM_000657.3) containing the wild-type (WT) or mutated (MUT) putative miR-125a/b-5p binding site were amplified and cloned into the XbaI-FseI sites of pGL3-Promoter Luciferase Expression Vector (Promega) and named Bcl-2-WT or Bcl-2-MUT, respectively. The primers were as follows: Bcl-2: forward 5′-GATCTCTAGAACATGCCTGCCCCAAACAAATATG-3′, reverse 5′-GATCTCTAGAACAGACAAGGAAAGTTTAATGGCAATGTG-3′, and mutation of the miR-125b-5p (position 2419–2426, Fig. 6A) and miR-125a-5p (position 2419–2426, Fig. 6A) binding sites was carried out using the QuikChange II Site-directed Mutagenesis kit (Agilent Technologies UK Ltd., Wokingham, UK). These luciferase reporter plasmids were co-transfected with miR-125a/b-5p mimics or miR-NC into HEK-293 T cells using Lipofectamine 2000. After transfection for 48 h, cells were lysed, and the relative luciferase activities were detected using the Dual Luciferase Reporter Assay system (Promega).

DDPCR was performed to quantify the absolute RNA copy numbers of BDNF-AS and miR-125a/b-5p in MM1S and U266 cells. In brief, 1 μg total RNA samples were reversely transcribed using a First-Strand cDNA Synthesis Kit. The droplets were generated using the QX200™ AutoDG™ Droplet Digital™ PCR System. The PCR reactions using the droplets were prepared using EvaGreen Supermix (1,864,033, Bio-Rad) containing 1 μl of cDNA. The absolute RNA copy numbers were assessed using QX200 Droplet Digital PCR System and calculated as previously described [37]. The copy numbers per cell were further estimated using a reference mRNA of known abundance as previously described [38].

RIP was performed using the EZ-Magna RIP Kit (Millipore, Billerica, MA, USA) following the manufacturer’s instructions. Briefly, HEK-293 T cells were lysed using RIP lysis buffer, and cell lysates were incubated with magnetic beads conjugated with anti-Ago2 antibody or negative control IgG (Millipore, USA). The immune-precipitated RNAs were then extracted, and the enriched binding targets were detected by qRT-PCR.

Cell proliferation was evaluated using EdU (5-ethynyl-2′-deoxyuridine) staining assay as previously described [39]. In brief, the transfected MM.1S and U266 cells were seeded into 96-well plates. At 24 h after seeding, 100 μl fresh medium containing 50 μM EdU (Ribobio) was added into each well. Cells were then incubated for another 2 h, fixed with 4% paraformaldehyde for 20 min, and then counterstained with DAPI solution. The images were observed under a fluorescence microscope (Nikon) (× 200 magnification) and merged by Adobe Photoshop 6.0 software. The proliferation rate of cells was counted as the percentage of EdU positive cells in 5 random fields.

Cell migration ability was examined as previously described [40]. In brief, cells were incubated with a normal cell growth medium in 6-well plates. The cell layers at approximately 90% adherence were scratched with a 10 μl sterile pipette tip. The shredded cells were washed with sterile PBS. After 24 h, images of cells in plates were acquired using a microscope and analyzed using Image J software. Cell migration rate was calculated as the percentage of the total cell-free area.

A total of 8 BALB/c nude mice (male, 5–6 weeks old, approximately 18–20 g) were purchase from Beijing Vital River Laboratory Animal Technology Co., Ltd. (Beijing, China). All animal procedures were performed following the guidelines of the Institutional Animal Care and Use Committee (Permission No: 655). This study was approved by the Ethics Committee of Shidong Hospital Affiliated to University of Shanghai for Science and Technology. A xenograft tumor model with MM was established as previously described [41]. In brief, mice were randomly assigned into 2 groups with 4 mice in each group and injected subcutaneously with 1 × 10⁶ MM.1S cells stably transfected sh-BDNF-AS or sh-NC. The tumor volume was evaluated every week for 4 weeks uisng the formula: V = (L× W²)/2, where L is the tumor length and W is the tumor width. After 4 weeks, mice were sacrificed by cervical dislocation and the xenograft tumors were excised and weighed. The xenograft tumors were stored at − 80 °C for subsequent experiments.

The xenograft tumors were embedded in paraffin and cut into 4-μm thick sections. The sections were then dewaxed with xylene, hydrated with gradient ethanol, and incubated with primary anti-Ki67 antibody (Bioss Antibodies, Inc., 1:200) at 4 °C overnight. On the next day, the sections were incubated with secondary antibody at room temperature for 1 h. The expression of Ki-67 was detected using a DAB immunohistochemistry color development kit (Sangon Biotech, China) following the manufacturer’s instructions. Cell images were captured under a fluorescence microscope (IX-51; Olympus) at 200x magnification.

All data were presented as the mean ± standard deviation (SD). Each experiment was replicated for 3 times. Statistical analyses were performed using SPSS v. 19.0. The difference between two groups was determined using student’s t-test. Differences among multiple groups were determined using one-way analysis of variance (ANOVA). Spearman’s correlation analysis was used to analyze correlation in a data set. Kaplan-Meier curve was used to evaluate the overall survival of MM patients with different expression levels of BDNF-AS using the median expression level as the cutoff value (without considering the stage and prognosis of MM patients). P < 0.05 was considered as the significant threshold.

To explore the role of BDNF-AS and miR-125a/b-5p in MM, we firstly detected their expression in the serum from MM patients (MM serum). The results showed that BDNF-AS was significantly upregulated in MM serum (n = 30) compared with that from healthy subjects (n = 30) (p < 0.01, Fig. 1A). However, the expression levels of BDNF-AS were significantly lower in the serum of MM patients at Stage III than that in Stage I and II (p < 0.01, Supplementary Table 1). Meanwhile, the expression of miR-125a-5p and miR-125b-5p were markedly downregulated in MM serum (n = 30) compared with that from healthy subjects (n = 30) (p < 0.01, Fig. 1B and C). All patients were divided into high and low BDNF-AS expression groups with the median expression level of BDNF-AS as the cutoff value. The survival analysis showed that high expression levels of BDNF-AS predicted a poorer overall survival of MM patients than that with low expression levels of BDNF-AS (p < 0.01, Fig. 1D). Spearman’s correlation analysis showed a significantly negative correlation between the expression of BDNF-AS and miR-125a/b-5p in MM tissues (n = 30, R² = 0.772, Fig. 1E; R² = 0.664, Fig. 1F). In addition, their expression levels in MM cell lines were also detected. The results showed that BDNF-AS was significantly upregulated (p < 0.05), while the expression of miR-125a-5p and miR-125b-5p were markedly downregulated in the 4 MM cell lines (p < 0.05, Fig. 1G-I). Due to the highest expression level of BDNF-AS and the lowest expression level of miR-125a/b-5p in MM.1S and U266 cells, these two cell lines were selected for the subsequent experiments.

Starbase 3.0 was used to determine the binding relationship between BDNF-AS and miR-125a/b-5p. The prediction revealed a putative binding site between BDNF-AS and miR-125a-5p, as well as miR-125b-5p (Fig. 2A). To further confirm their relationship, miR-125a/b-5p mimics and inhibitors were transfected into HEK-293 T cells. qRT-PCR assay showed that the expression levels of miR-125a-5p and miR-125b-5p were significantly increased by mimics, and their expression levels were reduced by inhibitors (p < 0.001, Fig. 2B). The results of RIP assay showed that BDNF-AS was significantly enriched in cells transfected with miR-125a/b-5p mimics compared with that in cells transfected with miR-NC (p < 0.01, Fig. 2C), suggesting that BDNF-AS and miR-125a/b-5p were binding to Ago2. Then luciferase reporter assay was performed in HEK-293 T cells, and the results showed that overexpression of miR-125a-5p and miR-125b-5p both obviously reduced the relative luciferase activity of BDNF-AS-WT (p < 0.01, Fig. 2D), suggesting that BDNF-AS acted as the sponge of miR-125a/b-5p. In addition, overexpression or knockdown of miR-125a/b-5p did not affect the expression of BDNF-AS (Fig. 2E). And overexpression or knockdown of BDNF-AS had no effect on the expression of miR-125a/b-5p (Fig. 2F). Moreover, as shown in Supplementary Table 2, the copy numbers of BDNF-AS and miR-125a/b-5p per cell were comparable with their stoichiometric interaction. These results demonstrated that BDNF-AS served as a sponge for miR-125a/b-5p.

To explore the role of BDNF-AS in MM, MM.1S and U266 cells were transfected with the lentiviral vector carrying sh-BDNF-AS and sh-NC. The transfection efficiency was evaluated by qRT-PCR, and the results showed that sh-BDNF-AS significantly reduced the expression levels of BDNF-AS compared with sh-NC in MM.1S (p < 0.001) and U266 cells (p < 0.001, Fig. 3A). By performing CCK-8 assay, we found that knockdown of BDNF-AS significantly decreased the viability of MM.1S (p < 0.01) and U266 cells compared with sh-NC (p < 0.01, Fig. 3B). The wound healing assay results also showed that knockdown of BDNF-AS significantly reduced the migration capacity of both MM.1S and U266 cells compared with sh-NC (p < 0.01, Fig. 3C). The EdU staining assay showed that knockdown of BDNF-AS markedly reduced the percentage of EdU positive cells in both MM.1S (p < 0.01) and U266 cells compared with sh-NC (p < 0.01, Fig. 3D). Furthermore, the expression levels of apoptosis-related proteins were evaluated using Western blot analysis, and the results showed that knockdown of BDNF-AS significantly increased the ratio of cleaved caspase 3/total caspase 3 and cleaved PARP/total PARP in both MM.1S and U266 cells (all p < 0.01) (Fig. 3E). These results demonstrated that knockdown of BDNF-AS significantly reduced proliferation and promoted apoptosis of MM cells.

To further determine BDNF-AS/miR-125a-5p-mediated effects on MM cells, MM.1S and U266 cells were transfected with miR-125a-5p inhibitor or inhibitor NC, and the transfection efficiency was evaluated by qRT-PCR. The results showed that miR-125a-5p inhibitor significantly reduced the expression levels of miR-125a-5p compared with inhibitor NC in MM.1S and U266 cells (p < 0.01, Fig. 4A). Then MM.1S and U266 cells were transfected with sh-BDNF-AS, miR-125a-5p inhibitor, or co-transfected with sh-BDNF-AS and miR-125a-5p inhibitor. CCK-8 assay showed that knockdown of BDNF-AS reduced the viability of MM.1S and U266 cells (p < 0.05), miR-125a-5p inhibitor obviously enhanced their viability (p < 0.05), and co-transfection of sh-BDNF-AS and miR-125a-5p inhibitor significantly reversed the inhibitory effects of sh-BDNF-AS on the viability of MM.1S and U266 cells (p < 0.05, Fig. 4B). In addition, knockdown of BDNF-AS significantly reduced the migration rate of MM.1S and U266 cells (p < 0.05), miR-125a-5p inhibitor significantly increased their migration rate (p < 0.05), and co-transfection of sh-BDNF-AS and miR-125a-5p inhibitor significantly reversed the inhibitory effect of sh-BDNF-AS on the migration rate of MM.1S and U266 cells (p < 0.05, Fig. 4C). Moreover, EdU staining assay showed that the number of EdU positive MM.1S and U266 cells was significantly reduced by knockdown of BDNF-AS (p < 0.05) and increased by miR-125a-5p inhibitor (p < 0.05). And co-transfection of sh-BDNF-AS and miR-125a-5p inhibitor obviously attenuated the inhibitory effect of sh-BDNF-AS on the number of EdU positive cells (p < 0.05, Fig. 4D). Furthermore, the expression levels of apoptosis-related proteins were evaluated by Western blot, and the results showed that miR-125a-5p inhibitor significantly reduced the ratio of cleaved caspase 3/total caspase 3 and cleaved PARP/total PARP (all p < 0.05), and co-transfection of sh-BDNF-AS and miR-125a-5p inhibitor obviously reversed the effects of sh-BDNF-AS on the expression of apoptosis-related proteins (all p < 0.05) but not that of total caspase 3 and total PARP in MM.1S and U266 cells (Supplementary Fig. 1A). These results revealed that downregulation of miR-125a-5p reversed the effects of knockdown of BDNF-AS on the proliferation and apoptosis of MM cells in vitro.

Similarly, we also explored whether the effect of BDNF-AS was mediated by miR-125b-5p. MiR-125b-5p inhibitor or inhibitor NC was transfected into MM.1S and U266 cells, and qRT-PCR assay showed that miR-125b-5p inhibitor significantly reduced the expression levels of miR-125b-5p compared with inhibitor NC in MM.1S and U266 cells (p < 0.01, Fig. 5A). Then MM.1S and U266 cells were transfected with sh-BDNF-AS, miR-125b-5p inhibitor, or co-transfected with sh-BDNF-AS and miR-125b-5p inhibitor. CCK-8 assay showed that the viability of MM.1S and U266 cells were significantly reduced by knockdown of BDNF-AS (p < 0.05) and enhanced by miR-125b-5p inhibitor (p < 0.05). Co-transfection of sh-BDNF-AS and miR-125b-5p inhibitor significantly reversed the inhibitory effect of sh-BDNF-AS on cell viability (p < 0.05, Fig. 5B). Wound healing assay showed that knockdown of BDNF-AS significantly reduced cell migration rate (p < 0.05), miR-125b-5p inhibitor significantly increased cell migration rate compared with that in the control group (p < 0.05), and co-transfection of sh-BDNF-AS and miR-125b-5p inhibitor reversed the inhibitory effect of sh-BDNF-AS on cell migration (p < 0.05, Fig. 5C). Meanwhile, the number of EdU positive cells was significantly reduced by knockdown of BDNF-AS (p < 0.05) and increased by miR-125b-5p inhibitor (p < 0.05). Co-transfection of sh-BDNF-AS and miR-125b-5p inhibitor obviously attenuated the inhibitory effect of sh-BDNF-AS on the number of EdU positive cells (p < 0.05, Fig. 5D). In addition, the expression levels of apoptosis-related proteins were evaluated by Western blot analysis. The results showed that miR-125b-5p inhibitor significantly reduced the ratio of cleaved caspase 3/total caspase 3 and cleaved PARP/total PARP (all p < 0.01), and co-transfection of sh-BDNF-AS and miR-125a-5p inhibitor obviously reversed the effects of sh-BDNF-AS on the expression of apoptosis-related proteins (all p < 0.01) but exhibited no effects on total caspase 3 and total PARP in MM.1S and U266 cells (Supplementary Fig. 1B). These results demonstrated that downregulation of miR-125b-5p reversed the effects of knockdown of BDNF-AS on proliferation and apoptosis of MM cells in vitro.

To further investigate the specific mechanism of BDNF-AS in MM, Starbase 3.0 was used to predict the potential targets of miR-125a/b-5p. It showed that Bcl-2 had a putative binding site with miR-125a-5p and miR-125b-5p (Fig. 6A). In addition, Ago2 RIP assay was performed, and the results showed that BCL2 was significantly enriched in the cells transfected with miR-125a/b-5p mimics compared with that in cells transfected with miR-NC (p < 0.01, Fig. 6B). Luciferase reporter assay was also performed in HEK-293 T cells, and the results showed that overexpression of miR-125a-5p or miR-125b-5p decreased the relative luciferase activity of Bcl-2-WT compared with miR-NC (p < 0,01) but had no effect on Bcl-2-MUT (Fig. 6C). Western blot analysis showed that both miR-125a-5p and miR-125b-5p mimics significantly decreased the expression levels of Bcl-2 compared with miR-NC in MM.1S and U266 cells (p < 0.001), and knockdown of BDNF-AS also decreased the expression levels of Bcl-2 compared with sh-NC in these two cells (p < 0.05, Fig. 6D). Next, Bcl-2 was silenced in MM.1S and U266 cells by the transfection of lentiviral vector carrying sh-Bcl-2 and sh-NC, and qRT-PCR assay showed that sh-Bcl-2 significantly reduced the expression levels of Bcl-2 in MM.1S and U266 cells compared with sh-NC (p < 0.01, Fig. 6E). To determine whether the effect of miR-125a/b-5p in MM was mediated by Bcl-2, MM.1S and U266 cells were transfected with miR-125a-5p inhibitor, miR-125b-5p inhibitor, or co-transfected with miR-125a-5p inhibitor and sh-Bcl-2, or co-transfected with miR-125b-5p inhibitor and sh-Bcl-2. CCK-8 assay showed that miR-125a-5p inhibitor and miR-125b-5p inhibitor significantly enhanced cell viability compared with that in the control (p < 0.05), and this effect was obviously blunted by co-transfection of miR-125b-5p inhibitor and sh-Bcl-2, or miR-125b-5p inhibitor and sh-Bcl-2 (p < 0.05, Fig. 6F). These results demonstrated that Bcl-2 was the target of miR-125a/b-5p. To confirm the relationship among BDNF-AS, miR-125a/b-5p, and BCL-2, further experiments were performed. NM cells were transfected with the wild-type (WT) or mutant (Mut) BDNF-AS overexpression vector. Ago2 RIP assay showed that overexpression of BDNF-AS-WT increased the enrichment of BCL-2 binding to Ago2, while overexpression of BDNF-AS-Mut had no effect on the enrichment of BCL-2 (Fig. 7A). In addition, NM cells were transfected with luciferase reporters containing BCL2 3′-UTRs and the wild-type (WT) or mutant (Mut) BDNF-AS overexpression vector. The results showed that overexpression of BDNF-AS-WT increased the luciferase activity in cells co-transfected with BCL2 3′-UTRs, while overexpression of BDNF-AS-Mut did not affect the luciferase activity of cells co-transfected with BCL2 3′-UTRs (Fig. 7B). Moreover, overexpression of BDNF-AS-WT increased the expression levels of BCL2 and enhanced cancer cell viability, while overexpression of BDNF-AS-WT had no effect on the expression of BCL2 and cell viability (Fig. 7C-D). Taken together, these data suggested that BDNF-AS functions in NM cells by regulating the miR-125a/b-5p/BCL2 axis.

To further determine the protective role of knockdown of BDNF-AS in MM in vivo, MM.1S cells stably transfected with sh-BDNF-AS or sh-NC were subcutaneously injected into nude mice to establish the xenograft tumor model. We found that knockdown of BDNF-AS effectively reduced tumor volume (p < 0.01) and tumor weight compared with that in sh-NC group (p < 0.05, Fig. 8A-C). Meanwhile, IHC staining in the tumors revealed that the number of Ki-67-positive cells was obviously reduced in sh-BDNF-AS group compared with that in sh-NC group (p < 0.05, Fig. 8D). The expression of BDNF-AS, miR-125a-5p, and miR-125b-5p in tumor tissues were detected by qRT-PCR, and the results showed that BDNF-AS was significantly downregulated (p < 0.001), while miR-125a-5p and miR-125b-5p in tumor tissues were markedly upregulated in sh-BDNF-AS injected mice compared with that in sh-NC injected mice (p < 0.01, Fig. 8E and F). Moreover, the expression of Bcl-2 protein in tumor tissues was significantly downregulated in sh-BDNF-AS injected mice compared with that in sh-NC injected mice (p < 0.01, Fig. 8G). In addition, the expression of apoptosis-related protein in tumor tissues were evaluated by Western blot analysis, and the results showed that knockdown of BDNF-AS significantly increased the ratio of cleaved caspase 3/total caspase 3 and cleaved PARP/total PARP (both p < 0.01) (Fig. 8H). These results demonstrated that knockdown of BDNF-AS suppressed tumor growth in vivo.