Long non-coding RNA AFAP1-AS1 accelerates the progression of melanoma by targeting miR-653-5p/RAI14 axis

Human epidermal melanocytes HEMa-LP (C0245C) were provided by Thermo Fisher Scientific (Waltham, MA, USA). Four human melanoma cell lines A375 (CRL-1619), M21 (BAA-1539), B16F10 (CRL-6475) and SK-MEL-2 (HTB-68), were obtained from American Type Culture Collection (ATCC; Manassas, VA, USA). All cell lines were cultured at 37 °C with 5% CO₂. Then, cells were all maintained continuously in DMEM medium (Thermo Fisher Scientific) supplied with 10% fetal bovine serum (FBS; Thermo Fisher Scientific) and 1% antibiotics (Invitrogen, Carlsbad, CA, USA).

Using the TRIzol Reagent (Invitrogen), total RNAs were extracted from cells, which were reverse transcribed to generate first-stand cDNA. Next, qRT-PCR was carried out using SYBR® Premix Ex Taq™ II kit (Takara, Dalian, China) on 7500 Real-Time PCR System (Applied Biosystems, Foster City, CA, USA) to determine the quantification of AFAP1-AS1 expression. Results were calculated by the 2^-ΔΔCt method, followed by normalization to GAPDH/U6.

Under the standard conditions, A375 or M21 cells were incubated and seeded into 6-well plates. The lentivirus vector bearing short hairpin RNAs (shRNAs) sequences targeting AFAP1-AS1 (sh-AFAP1-AS1#1/2), RAI14 (sh-RAI14#1/2), their corresponding negative control (sh-NC), miR-653-5p mimics/inhibitors, NC mimics/inhibitors, the RAI14 overexpressing plasmids and empty pcDNA3.1 vector were constructed by GenePharma (Shanghai, China). Using Lipofectamine 2000 (Invitrogen), transfection of above plasmids was implemented in A375 or M21 cells. 48 h lately, transfected cells were collected.

Transfected A375 or M21 cells were seeded in 96-well plates and then incubated. Subsequently, the same amount of CCK-8 solution was added to each well. Cells were then incubated for another 4 h. The absorbance values were tested by Thermo-max microplate reader (Thermo Fisher Scientific) at 450 nm.

After being placed into 6-well plates, transfected A375 or M21 cells were incubated for 14 days. Following that, cells were fixed by paraformaldehyde (PFA; Sigma-Aldrich, St. Louis, MO, USA) and stained by crystal violet solution (Sigma-Aldrich). Colonies in each well were visualized, followed by quantitated via Image J software (Brainlab, Munich, Germany).

In order to determine cell migration, a 24-well chamber containing an aperture of 8 μm was employed. Transfected A375 or M21 cells were inoculated into the top chamber filled with 300 μl serum-free medium. DMEM (Thermo Fisher Scientific) with 10% FBS was added into the lower chamber. After incubating 1 day, remained cells from the upper surface were treated with a cotton swab in the top chamber. Cells that migrated to the bottom of the membrane were washed and fixed with 4% paraformaldehyde, then dyed. Finally, migrated cells were counted via a 200 × microscope (Olympus, Tokyo, Japan).

For cell invasion, the condition of culture is no difference. However, the membrane of the chamber was coated with Matrigel solution (BD Diagnostics, Franklin Lakes, NJ). 24 h later, cells that invaded to the lower chambers were pictured and counted under a microscope.

Total proteins were isolated from A375 or M21 cells and the concentration was detected via BCA Protein Assay Kit (Takara, Tokyo, Japan). Protein samples in equal quantity were separated via sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) (Millipore, Bedford, MA, USA), and then transferred into the PVDF (Millipore) membranes. Membranes blocked with non-fat milk were incubated with primary antibodies from Abcam (Cambridge, USA) of anti-E-cadherin (ab194982), anti-N-cadherin (ab202030), anti-RAI14 (ab137118), anti-Ki67 (ab16667) and anti-GAPDH (ab8245), followed by with secondary antibody. GAPDH was regarded to be an endogenous control. Western bands were observed using ECL detection system.

Thirty-six male BALB/c nude mice (age: 5 weeks ±1 week; weight: 23 g ± 2 g) obtained from the National Laboratory Animal Center (Beijing, China) were fed under the condition of specific pathogen-free. Then, the suspension of A375 cells transfected with sh-NC or sh-AFAP1-AS1#1 (with a concentration of 2 × 10⁷ cells/ml) was subcutaneously injected into lower right flank of each mouse that was randomly divided into two groups (with 6 mice each group). Another group (two subgroups) of 12 mice was injected with cells transfected with NC mimics or miR-653-5p mimics (RiboBio Co., Ltd., Guangzhou, China) in the same way. The left 12 mice were injected with cells transfected with sh-NC or sh-RAI14#1 similarly. Every 4 days, the volumes and weights of tumors were measured and calculated via the formula of length × width² × 0.5. After injection of 4 weeks, mice were euthanized via dislocation of cervical vertebra method and then tumor was excised, followed by extraction for further Ki67 staining. All procedures during the in vivo experiments were approved by the Institutional Committee for Animal Research and kept to the national guidelines for the care and use of laboratory animals (GB14925–2010).

Using Ribo™ Fluorescent in Situ Hybridization Kit (RiboBio, Guangzhou, China), FISH assay was performed. Nucleus was stained via DAPI, which were designed and synthesized by Ribobio. Cy3 fluorescent dye was selected to label AFAP1-AS1. Fluorescence detection was conducted via the confocal laser-scanning microscope (Leica Microsystems, Wetzlar, Germany).

In short, biotinylated AFAP1-AS1 probe (AFAP1-AS1 probe biotin) or a negative control probe (AFAP1-AS1 probe-no biotin) were separately synthesized from Thermo Fisher Scientific. Next, the biotinylated lncRNA (lncRNA biotin) was transfected into A375 or M21 cells. After 2 days, cells were subjected to RNA pull-down assay. The results were analyzed by qRT-PCR.

The wild-type or mutant binding sequence of miR-653-5p in AFAP1-AS1 or RAI14 3′-UTR was synthesized and sub-cloned into pmirGLO dual-luciferase vector (Promega, Madison, WI, USA). AFAP1-AS1-Wt/Mut vector was co-transfected with miR-653-5p mimics or NC mimics into A375 or M21 cells. Besides, A375 or M21 cells were co-transfected with RAI14-Wt/Mut vector or NC mimics/miR-653-5p mimics/miR-653-5p mimics + pcDNA3.1/AFAP1-AS1. 48 h later, Dual Luciferase Report Assay System (Promega) was employed to monitor luciferase activity.

After transfection, A375 or M21 cells were cross-linked with formaldehyde (Sigma-Aldrich), and then lysed in RIP lysis buffer (Thermo Fisher Scientific). Subsequently, cells were incubated overnight with magnetic beads (Invitrogen) conjugated to antibodies of anti-Ago2 (Abcam) and anti-IgG (Abcam) at 4 °C. The levels of AFAP1-AS1, miR-653-5p and RAI14 were determined via qRT-PCR.

Firstly, we studied the potential role of AFAP1-AS1 in melanoma. qRT-PCR results showed that AFAP1-AS1 exhibited conspicuous high expression in melanoma cell lines (A375, M21, B16F10 and SK-MEL-2) in comparison to normal human epidermal melanocyte HEMa-LP cells (Fig. 1a). The expression of AFAP1-AS1 in A375 and M21 cells remarkably declined by the transfection of sh-AFAP1-AS1#1/2 (Fig. 1b). Through CCK-8 and colony formation assay, we discovered that the cell proliferation was inhibited by suppressing AFAP1-AS1 (Fig. 1c-d). Transwell assay was conducted to detect the migrating and invasive abilities of melanoma cells. We found that the migrating and invasive abilities of A375 and M21 cells were impaired by AFAP1-AS1 knockdown (Fig. 1e-f). Western blot assay showed that depletion of AFAP1-AS1 notably enhanced E-cadherin expression and reduced N-cadherin expression (Fig. 1g). To further explore the role of AFAP1-AS1 in melanoma, in vivo assays were performed using healthy nude mice (weight: 23 g ± 2 g; no infection and any treatment previously). The tumor size, volume and weight (analyzed using 6 animals in each group) were measured and found down-regulated in mice inoculated with sh-AFAP1-AS1#1 (Fig. S1A-B). AFAP1-AS1 was significantly lowly expressed in mice inoculated with sh-AFAP1-AS1#1 (Fig. S1C). In addition, E-cadherin expression was elevated, while N-cadherin and Ki67 expression was decreased in response to AFAP1-AS1 repression (Fig. S1D). These results demonstrated that AFAP1-AS1 is up-regulated in melanoma cell lines and facilitates melanoma cells growth, migration, invasion and EMT process.

LncRNAs have been reported to modulate gene expression in multifold mechanisms. Next, we further investigated the regulatory mechanism related to AFAP1-AS1. We firstly studied the subcellular distribution of AFAP1-AS1 in A375 and M21 cells. FISH assay revealed that AFAP1-AS1 was mainly located in the cytoplasm of A375 and M21 cells (Fig. 2a, Fig. S1E). We speculated that AFAP1-AS1 may realize its biological function by acting as a ceRNA. Through ENCORI (The Encyclopedia of RNA Interactomes) website, we found 29 miRNAs have binding sites for AFAP1-AS1. qRT-PCR assay was performed to detect the expression of miRNAs in sh-NC or sh-AFAP1-AS1#1 groups, and 5 miRNAs that up-regulated in sh-AFAP1-AS1#1 transfected cells were shown in Fig. 2b and Fig. S1F. RNA pull down assay was used to confirm the binding ability between AFAP1-AS1 and miRNAs. Results showed that only miR-653-5p was significantly enriched by biotinylated AFAP1-AS1 probe (Fig. 2c, Fig. S1G). In subsequence, we observed that miR-653-5p was down-regulated in melanoma cells (Fig 2d). Then, the overexpression and knockdown efficiency of miR-653-5p was detected by qRT-PCR (Fig. 2e, Fig. S1H). MiR-653-5p overexpression decreased AFAP1-AS1 expression, whereas AFAP1-AS1 repression increased miR-653-5p expression (Fig. 2f, Fig. S1I). We found the putative binding sites between AFAP1-AS1 and miR-653-5p, as demonstrated in Fig. 2g. Luciferase reporter assay has been utilized in the confirmation of the affinity between genes [19, 20]. If gene A could interact with gene B, the luciferase activity of the pmirGLO vector containing the wild type binding sites of gene A on gene B could be weakened by the overexpression of gene A. As depicted in Fig. 2h and Fig. S1J, the luciferase activity of AFAP1-AS1-Wt was impaired by miR-653-5p mimics, while that of AFAP1-AS1-Mut was unaffected by miR-653-5p mimics, verifying the combination between AFAP1-AS1 and miR-653-5p. Rescue assays revealed that the decrease of cell proliferation induced by AFAP1-AS1 deficiency could be reversed by miR-653-5p silencing (Fig. 2i-j). The migrating and invasive abilities were blocked by repressing AFAP1-AS1, but this effect was countervailed by miR-653-5p inhibition (Fig. 2k-l). MiR-653-5p knockdown could rescue the increase of E-cadherin expression and the decrease of N-cadherin expression caused by AFAP1-AS1 knockdown (Fig. 2m). As shown in Fig. S2A-C, the tumor size, volume and weight were repressed by miR-653-5p overexpression. Moreover, the expression of miR-653-5p was elevated in mice transfected with miR-653-5p mimics (Fig. S2D). In the end, western blot assay indicated that E-cadherin expression was increased when miR-653-5p was up-regulated, and N-cadherin as well as Ki67 levels were inhibited by miR-653-5p overexpression (Fig. S2E). Collectively, AFAP1-AS1 directly binds to miR-653-5p and miR-653-5p overexpression suppresses melanoma progression in vivo.

Through microT, miRanda and TargetScan bioinformatics tools, we screened out 3 probable genes that may interact with miR-653-5p, as displayed in the Venn diagram (Fig. 3a). Overexpression of miR-653-5p could lead to evident decline on the expression of RAI14 (Fig. 3b, Fig. S2F). Besides, RAI14 was revealed to be markedly high-expressed in melanoma cells (Fig. 3c). MiR-653-5p overexpression or AFAP1-AS1 repression reduced the mRNA and protein expressions of RAI14 (Fig. 3d, Fig. S2G). The corresponding sequences between miR-653-5p and RAI14 were exhibited in Fig. 3e. Result of luciferase reporter assay exhibited that miR-653-5p up-regulation-induced reduction on the luciferase activity of RAI14-Wt could be rescued by overexpressing AFAP1-AS1 (Fig. 3f, Fig. S2H). Finally, RIP assay was performed to study the molecular interplay among AFAP1-AS1, miR-653-5p and RAI14. Ago2 is an essential component of RNA-induced silencing complex (RISC) and therefore RISC could be pulled by antibody targeting Ago2. Besides, mRNAs and lncRNAs that could bind with miRNA all existed in RISC. Results demonstrated that AFAP1-AS1, miR-653-5p and RAI14 were enriched by antibody targeting Ago2 indicating their coexistence in the miR-653-5p induced silencing complex (Fig. 3g, Fig. S2I). In addition, the expressions of AFAP1-AS1 and miR-653-5p exhibited no change in response to RAI14 overexpression in A375 cells (Fig. S2J-K). Afterwards, to study the role of RAI14 in melanoma, RAI14 was effectively knocked down by sh-RAI14#1/2 for conducting loss-of-function assays (Fig. 3h). We noticed that RAI14 depletion remarkably restrained the tumor size, volume and weight of mice (Fig. 3i-k). Furthermore, the mRNA level of RAI14, the protein level of N-cadherin and Ki67 were detected to be down-regulated by RAI14 silencing in mice. However, the protein expression of E-cadherin was promoted via RAI14 deficiency (Fig. 3l-m, Fig. S3A). All in all, these findings unveiled that RAI14 binds with miR-653-5p and its depletion impairs tumor growth.

To confirm whether AFAP1-AS1/miR-653-5p/RAI14 axis was involved in melanoma progression, rescue assays were conducted. The mRNA and protein expressions of RAI14 were dramatically increased by RAI14 up-regulation (Fig. 4a-b). The cell proliferation was weakened by knocking down AFAP1-AS1, but this effect was counteracted by RAI14 overexpression (Fig. 4c-d). Cell migration and invasion phenotypes were weakened by AFAP1-AS1 depletion, while strengthened in response to RAI14 overexpression (Fig. 4e-f). Up-regulation of RAI14 could abolish the interfering impacts of AFAP1-AS1 depletion on EMT progress, as evidenced by the altered expressions of E-cadherin and N-cadherin (Fig. 4 g). Therefore, it was safe to reach the conclusion that AFAP1-AS1 elicited cellular function changes of melanoma cells via targeting the miR-653-5p/RAI14 axis.