Patients and tissue specimens

This study was approved by the ethics committee of The First Hospital of Jilin University. Informed consent was provided from all 30 patients. Tumor tissues and neighboring normal tissues were harvested from GBC patients after radical tumor resection.

MiRNAs, plasmid construction and transfection

MiR-520b mimic (miR-520b mimic), negative control of miR-520b mimic (NC mimics), miR-520b inhibitor (miR-520b inhibitor) and negative control of miR-520b inhibitor (NC inhibitor) were bought from GenePharma (Shanghai, China). RAB22A cDNA was inserted into a pCMV expression vector (LQbiotech, Shanghai, ChinaEra Biotech) to construct the RAB22A overexpression vector. MiRNA or plasmid transfection was performed using Lipofectamine 2000 (Invitrogen, Beijing, China).

Cell culture

The NOZ cell line was purchased from Shanghai Cell Institute Country Cell Bank (Shanghai, China) and maintained in high-glucose Dulbecco’s Modified Eagle’s Medium (DMEM, Life Technologies, Beijing, China) supplemented with 10% fetal bovine serum (Life Technologies, Beijing, China) at 37°C.

MTT assay

Ten thousand transfected NOZ cells were seeded and incubated in 12-well plates for 48 h. Then, 0.5 mg/ml MTT was added to each well. After 4 h, formazan crystals were dissolved in DMSO and OD 490 nm was measured.

Colony formation assay

Two hundred transfected NOZ cells were plated and incubated in 6-well plates for 48 h. After fixation, the colonies were dyed using crystal violet. Images were taken and colonies were counted.

Wound healing assay

One hundred thousand transfected NOZ cells were plated in 24-well plates. When GBC cells reached around 80% confluency, the cell layer was scraped using a 200 μl pipette tip. Images of the wounded areas were taken 24 h after treatment via an inverted light microscope.

Transwell invasion assay

One hundred thousand transfected NOZ cells were suspended in 500 μl of DMEM medium without FBS and transferred into Matrigel-coated upper chambers. 500 μl of complete medium was supplemented into the lower chambers. 48 h after seeding, residual cells were removed and invasive cells were fixed, dyed and counted under a microscope.

Quantitative real-time PCR

Total RNA was extracted from frozen tissue samples or NOZ cells, and then transcribed into cDNA via a reverse transcription kit (TaKaRa, Beijing, China). RNA level was examined through quantitative real-time PCR (qPCR) using SYBR Green (Takara, Beijing, China). Primers used in this study were as follows:

miR-520b-fwd: 5’-AACACGCAAAGTGCTTCCTTTTAGA-3’

miR-520b-rev: 5’-CAGTGCAGGGTCCGAGGT-3’

RAB22A-fwd: 5’-GTGTGTCTGCTCGGGGATAC-3’

RAB22A-rev: 5’-GCCCCTATTGTTGGGTTGATGT-3’

β-Actin-fwd: 5’-GGAGATTACTGCCCTGGCTCCTA-3’

β-Actin-rev: 5’-GACTCATCGTACTCCTGCTTGCTG-3’

Immunohistochemical (IHC) staining

GBC tumor tissues and neighboring normal tissues were fixed, embedded, dewaxed and hydrated. After blocking, the sections were immersed with the primary antibody against RAB22A (Abcam, Beijing, China). After washing, sections were probed with the secondary antibodies for 1 h at 22°C. Sub-sequently, the sections were stained via DAB for 8 min and counterstained via hematoxylin for 2 min. Images were taken using a fluorescence microscope.

Western blot analysis

Protein lysates of frozen tissue or cells were prepared. After quantification, proteins were separated by SDS-PAGE and then transferred onto PVDF membranes. After blocking, PVDF membranes were probed via primary antibody against RAB22A (1 : 500; Abcam, Beijing, China); anti-GAPDH antibody (1 : 3000; Abcam, Beijing, China) was taken as the internal reference. Then, after incubation with secondary antibodies, the protein bands were quantified by an ECL chemiluminescence kit.

Dual luciferase reporter assay

Wide-type RAB22A 3′UTR (containing putative miR-520b binding sites) or mutant RAB22A 3′UTR (harboring the mutant binding sequences of miR-520b) was generated and inserted into a luciferase reporter plasmid (LQbiotech, Shanghai, China) to generate wild-type (RAB22A 3′UTR-WT) or mutant report plasmid (RAB22A 3′UTR-MUT) respectively.

miR-520b mimic, as well as RAB22A 3′UTR-WT or RAB22A 3′UTR-MUT, was transfected into NOZ cells. 48 h after transfection, luciferase reporter assay was conducted using a luciferase assay kit (Promega, Shanghai, China).

Statistical analysis

Data were analyzed using SPSS 18.0 software and presented as mean ± SD. Comparisons between two groups were assessed via one-way ANOVA. The correlation between miR-520b and RAB22A expression levels was examined via Pearson’s χ² test. A difference was considered statistically significant when p < 0.05 in this study.

MiR-520b inhibited the proliferation of NOZ cells

qPCR confirmed there was an increase of miR-520b mRNA in miR-520b mimic groups compared to that in the NC mimic group (Figure 2 A, p < 0.001). In contrast, miR-520b inhibitor significantly decreased miR-520b mRNA level in NOZ cells in comparison to the NC inhibitor group (Figure 2 A, p < 0.01). MTT assay showed that miR-520b mimic inhibited the proliferation in NOZ cells time-dependently (Figure 2 B, p < 0.01). However, Figure 2 B demonstrates that down-regulation of miR-520b facilitated the proliferation of NOZ cells in a time-dependent manner (p < 0.01). In addition, colony formation assay indicated miR-520b mimic overexpression NOZ cells had a much lower number of colonies in comparison with those in the NC mimic group (Figure 2 C, p < 0.01), whereas miR-520b inhibitor stimulated NOZ colony formation compared to NC inhibitor (Figure 2 C, p < 0.05).

Figure 2

miR-520b inhibited NOZ cell proliferation. NOZ cells were transfected with miR-520b mimic or miR-520b inhibitor. A – MiR-520b mRNA level in transfected NOZ cells was determined by qPCR. B, C – Cell viability of transfected NOZ cells was determined via MTT assay (B) and colony formation assay (C) (mean ± SD)

*p < 0.05, **p < 0.01, ***p < 0.001.

MiR-520b restricted the migration and invasion of NOZ cells

Figure 3 A shows that miR-520b mimic inhibited the migration of NOZ cells (68.09%, p < 0.01) compared to NC mimic, whereas miR-520b inhibitor exhibited no significant influence on the migratory capability of NOZ cells. Additionally, miR-520b mimic limited the invasion ability of NOZ cells (Figure 3 B, 52.94%, p < 0.05), whereas miR-520b inhibitor fostered their invasion capacity (Figure 3 B, 90.91%, p < 0.05).

Figure 3

miR-520b repressed the migration and invasion of GBC cells. The migration and invasion of NOZ cells transfected with miR-520b mimic or miR-520b inhibitor were investigated via wound healing (A) and transwell invasion assay (B) respectively (mean ± SD)

*p < 0.05, **p < 0.01.

RAB22A was a direct downstream target of miR-520b

Bioinformatics analysis using TargetScan revealed that there was a potential miR-520b binding site in the RAB22A 3′UTR (Figure 4 A). Dual luciferase reporter assay further revealed that miR-520b mimic decreased the luciferase activity of wild-type RAB22A 3′UTR in comparison to NC mimic (Figure 4 A, 55.56%, p < 0.01). However, miR-520b mimic exhibited little effect on the luciferase activity of mutant RAB22A 3′UTR (Figure 4 A). In addition, Figure 4 B shows that NOZ cells transfected with miR-520b mimic had a significantly lower RAB22A protein level in comparison to cells transfected with NC mimic (p < 0.001). However, miR-520b inhibitor stimulated RAB22A expression compared to the NC inhibitor group (Figure 4 B, p < 0.05).

Figure 4

RAB22A was a direct target of miR-520b. A – Schematic diagram of miR-520b binding sites in RAB22A 3′UTR. Luciferase reporter activities driven by wild-type or mutant RAB22A 3′UTR were examined in NOZ cells transfected with miR-520b mimic or NC mimic. B – RAB22A protein and mRNA level in NOC cells transfected with miR-520b mimic or miR-520b inhibitor were examined via Western blot and qPCR, respectively. β-actin was used as an internal control for qPCR assay (mean ± SD)

*p < 0.05, **p < 0.01, ***p < 0.001.

RAB22A overexpression blocked the anti-tumor effects of miR-520b on NOZ cells

RAB22A overexpression restored the inhibition of RAB22A mRNA and protein expression in NOZ cells caused by miR-520b mimic (Figure 5 A, p < 0.001). MTT assay showed that miR-520b mimic time-dependently decreased the viability of NOZ cells in comparison with NC mimic (Figure 5 B, p < 0.01), whereas RAB22A overexpression completely abolished this miR-520b mimic-induced reduction of viability (Figure 5 B, p < 0.01). Colony formation assay proved that miR-520b mimics significantly inhibited NOZ cell proliferation in comparison with NC mimic, whereas RAB22A overexpression inhibited the above phenomenon (Figure 5 C, p < 0.01). Additionally, wound healing assay demonstrated that miR-520b mimic inhibited GBC cell migration by ~50% compared to NC mimic, whereas the inhibitory effect of miR-520b on GBC migration could be restored upon RAB22A overexpression (Figure 5 D, p < 0.05). Transwell invasion assay also showed that RAB22A overexpression restricted the inhibitive effect of miR-520b on the invasion ability of NOZ cells (Figure 5 E, p < 0.01).

Figure 5

RAB22A overexpression blocked the anti-tumor effects of miR-520b on NOZ cells. NOZ cells were transfected with NC mimic, miR-520b mimic, or miR-520b mimic with RAB22A overexpression plasmid (miR-520b mimic + RAB22A). A – RAB22A mRNA and protein levels were detected via qPCR and Western blot, respectively. B – The viability of transfected NOZ cells was detected via MTT assay. C – The viability of transfected NOZ cells was detected via colony formation assay. D, E – The migration and invasion of transfected NOZ cells were investigated via wound healing (D) and transwell invasion assay (E) respectively (mean ± SD)

*p < 0.05, **p < 0.01, ***p < 0.001 compared to NC mimic group; ^#p < 0.05, ^##p < 0.01, ^###p < 0.001 compared to miR-520b mimic group.

RAB22A was up-regulated in GBC tumor tissues and had a negative relationship with miR-520b expression

Western blot analysis indicated that tumor samples of GBC patients had a higher RAB22A protein expression level (p < 0.001, p < 0.01 and p < 0.01 for T1, T2 and T3 compared to matched neighboring normal tissues respectively) than that in matched neighboring normal tissues (Figure 6 A). Also, IHC analysis showed that RAB22A expression was enhanced in GBC tumor samples compared to neighboring normal tissues (Figure 6 B). Moreover, Figure 6 C further confirmed that there was a negative regulatory relationship between miR-520b and RAB22A.

Figure 6

RAB22A was up-regulated and had negative relationship with miR-520b expression in GBC tumor tissues. A – Western blot analysis of RAB22A expression in tumor tissues (T) and neighboring normal tissues (N) from 3 GBC patients. GAPDH was used as the internal loading control. B – IHC staining of RAB22A in GBC tumor sample and neighboring normal tissue sample. C – Correlation of RAB22A and miR-520b expression in GBC tumor samples (mean ± SD)

**p < 0.01, ***p < 0.001.