Written informed consent was obtained from all participants, and our study was approved by Medical Ethics Committee of First Affiliated Hospital of Guangxi Medical University. A total of 15 HCC patients and 15 healthy donors, who were recruited at the Hepatobiliary Surgery Department of First Affiliated Hospital of Guangxi Medical University, were included in our study. Peripheral blood mononuclear cells (PBMCs) were isolated from HCC patients and healthy controls by density gradient centrifugation (Human lymphoprep, TBD, Tianjin, China). Primary NK cells were negatively selected from PBMCs using magnetic activated cell sorting human NK cell isolation kit (Sangon Biotech, Shanghai, China).

HCC cell line SMMC7721, HepG2, human NK cell line NK-92, and 293T cells were purchased from American Type Culture Collection (ATCC, Manassas, VA, USA). SMMC7721, HepG2, and 293T cells were maintained in DMEM with 10% heat-inactivated fetal bovine serum (FBS; Sigma-Aldrich, St. Louis, MO, USA) and 1% penicillin/streptomycin. NK-92 cells were maintained in α-minimum essential medium (Invitrogen, Carlsbad, CA, USA) supplemented with 12.5% FBS (Sigma-Aldrich), 12.5% horse serum (Gibco, Grand Island, NY, USA), 100 U/mL recombinant human IL (rhIL)-2 (PeproTech, Rocky Hill, NJ, USA), and 0.1 mM β-mercaptoethanol. The isolated primary NK cells were cultured in complete RPMI medium (Gibco) supplemented with 10% heat-inactivated FBS (Sigma-Aldrich), 2 mM L-glutamine, and 100 U/mL rhIL-2 (PeproTech). All cells were incubated at 37℃ in an incubator of 5% CO₂.

miR-506 mimic (miR-506), miRNA negative control (miR-NC), miR-506 inhibitor (anti-miR-506), inhibitor control (anti-miR-NC), pcDNA-STAT3, pcDNA empty control (pcDNA), siRNA specially against STAT3 (si-STAT3), and siRNA control (si-NC) were purchased from GenePharma (Shanghai, China). NK-92 and primary NK cells were seeded into 96-well plates and transfected with the abovementioned nucleotides or plasmids, using Lipofectamine 2000 (Invitrogen). Co-culturing or RNA extraction was performed 48 hours after transfection.

Total RNA, including miRNA, was isolated from the isolated primary NK cells or NK-92 cells using TRIzol reagent (Invitrogen). Complementary DNA was synthesized from 1 µg of total RNA using PrimeScript RT reagent kit (TaKaRa, Dalian, China). miR-506 detection was performed using TaqMan miRNA assay system (Applied Biosystems, Foster City, CA, USA), while STAT3 mRNA expression was examined using SYBR Green Real-time PCR Master Mix (TaKaRa). Quantitative real-time (qRT) PCR reaction was conducted on GeneAmp PCR 9700 Thermo cycler (Applied Biosystems). Expressions of miR-506 and STAT3 mRNA were evaluated using 2^−ΔΔCt method.

NK-92 cell-mediated cytotoxicity against HCC cells was assessed via flow cytometry using CFSE/7AAD cytotoxicity assay. NK-92 cells transfected with or without miR-506, miR-NC, anti-miR-506, anti-miR-NC, miR-506 + pcDNA, miR-506 + STAT3, anti-miR-506 + si-NC, or anti-miR-506 + si-STAT3 were labeled with 5 µM CFSE (Beyotime, Nanjing, China) for 10 min at 37℃. Stained cells were washed twice with phosphate-buffered saline (PBS) and mixed with target SMMC7721 cells at various effector-to-target ratios (E:T): 10:1, 5:1, and 2.5:1, followed by 6 h of incubation at 37℃. After washed twice with PBS, cell mixtures were and incubated with 7-AAD (KeyGEN BioTECH, Nanjing, China) for 15 min and detected using FACSCalibur system (BD Biosciences, San Jose, CA, USA).

Primary NK cell cytotoxicity was determined using a lactate dehydrogenase (LDH) assay according to the manufacturer's instructions. Briefly, SMMC7721 or HepG2 cells were seeded into 96-well plates at a density of 5×10³ cells per well before coculturing. Primary NK cells transfected with or without miR-506, miR-NC, anti-miR-506, anti-miR-NC, miR-506 + pcDNA, miR-506 + STAT3, anti-miR-506 + si-NC, or anti-miR-506 + si-STAT3 were added to each well and co-cultured with SMMC 7721 or HepG2 cells at an E:T ratio of 5:1. After 4 h of incubation, supernatant was collected and cytolytic activity of NK cells was examined using LDH Activity Assay kit (Jiancheng Bioengineering, Nanjing, China).

Wild-type or mutated STAT3 3′UTR fragments containing the potential miR-506 binding sites were synthesized and cloned into pMIRNA-Reporter vector (Ambion, Austin, TX, USA) to produce pMIRNA-STAT3-WT (STAT3-WT) and pMIRNA-STAT3-MUT (STAT3-MUT). 293T cells were seeded into 24-well plates and cotransfected with miR-506 or miR-NC and constructed luciferase plasmids using Lipofectamine 2000 (Invitrogen). After 48 h of transfection, firefly and Renilla luciferase activity was performed using Dual Luciferase Reporter Assay Kit (Promega, Madison, WI, USA).

Cell protein was extracted from primary NK cells transfected with miR-506, anti-miR-506, or matched controls using RIPA lysis buffer (Sigma-Aldrich) for 10 min at 4℃ and quantified using Bio-Rad protein assay system (Bio-Rad, Hercules, CA, USA). Equal amounts of cellular proteins were loaded on 10% sodium dodecyl sulfate-polyacrylamide (SDS-PAGE) gel electrophoresis and transferred to polyvinylidene difluoride membranes (PVDF, Millipore, Billerica, MA, USA). Membranes were blocked with 5% nonfat milk at room temperature for 2 h, and then incubated with primary antibody against STAT3 (1:1000; Cell Signaling Technology, Beverly, MA, USA) and β-actin (1:1000; Cell Signaling Technology) at 4℃ overnight, followed by incubation with horseradish peroxidase-conjugated secondary antibodies (Santa Cruz Biotechnology, Santa Cruz, CA, USA) for 1 h. Protein bands were visualized using SuperSignal enhanced chemiluminescence kit (Pierce, Rockford, IL, USA).

RNA immunoprecipitation (RIP) assay was performed with anti-Ago1 (Abcam, Cambridge, MA, USA) or IgG (Sigma-Aldrich). SMMC7721 cells at 48 h after transfection with miR-506 or miR-NC were lysed in complete RIP lysis buffer, and then 100 µL whole-cell lysate of SMMC7721 cells was incubated with RIP buffer containing magnetic beads conjugated with human anti-Ago1 or negative control IgG. Samples were incubated with Proteinase K with shaking to digest the protein, and RNA coimmunoprecipitated with anti-Ago1 or IgG antibodies was isolated using TRIzol (Invitrogen) and then subjected to qRT-PCR analysis.

All experimental results are shown as mean±SD, and all statistical analyses were performed using SPSS 19.0 (IBM Corp., Armonk, NY, USA). Comparison between two or more groups was performed using Student's t-test and one-way ANOVA. A two-tailed p<0.05 was considered statistically significant.

To characterize the function of miR-506 in NK cells, we determined the expression of miR-506 in primary NK cells isolated from HCC patients and healthy controls. As shown in Fig. 1A, miR-506 expression was significantly downregulated in primary NK cells from 15 HCC patients as compared with that from 15 healthy controls. In addition, primary NK cells from HCC patients showed remarkably reduced cytotoxicity against SMMC7721 or HepG2 cells, with respect to primary NK cells from healthy controls (Fig. 1B and 1C). Interestingly, we found that miR-506 expression was positively correlated with NK cell cytotoxicity against SMMC7721 or HepG2 cells (Fig. 1D and E).

To further validate the role of miR-506 in NK cells, gain-of-function and loss-of-function approaches were performed in NK-92 and primary NK cells by transfecting with miR-506, anti-miR-506, or matched controls. qRT-PCR analyses demonstrated that miR-506 expression was dramatically elevated in miR-506-transfected NK-92 and primary NK cells, but significantly declined in anti-miR-506-treated NK-92 and primary NK cells (Fig. 2A and B). Moreover, CFSE/7AAD cytotoxicity assay showed that cytotoxicity of miR-506-mimicked NK-92 cells to SMMC7721 cells was evidently higher than that of miR-NC-transfected cells (Fig. 2C). On the contrary, anti-miR-506-introduced NK-92 cells exhibited decreased cytotoxicity to HepG2 cells at different E:T ratios relative to anti-miR-NC group (Fig. 2D). Consistently with the results from CFSE/7AAD cytotoxicity assay, LDH assay demonstrated that enhanced expression of miR-506 by miR-506 mimic drastically increased primary NK cell cytotoxicity at E:T of 5:1, while knockdown of miR-506 by anti-miR-506 strikingly inhibited the cytotoxicity of primary NK cells from HCC patients against SMMC7721 or HepG2 cells (Fig. 2E and F). Collectively, these results demonstrated that miR-506 overexpression enhanced NK cell cytotoxicity against HCC cells.

To further investigate the underlying mechanism of miR-506 in regulating NK cell cytotoxicity against HCC cells, bioinformatics analysis was performed to predict the potential targets of miR-506. The results demonstrated that STAT3 was predicted as a target of miR-506 based on putative target sequence at position 1303–1309 of STAT3 3′UTR, as shown in Fig. 3A. To validate whether STAT3 was a direct target of miR-506, luciferase reporter plasmids containing wild-type or mutated STAT3 3′UTR were constructed and luciferase reporter assay was performed. As displayed in Fig. 3B, overexpression of miR-506 led to a remarkable reduction of luciferase activity of STAT3-WT in 293T cells, but did not affect luciferase activity of STAT3-MUT. To further explore whether miR-506 and STAT3 mRNA interact and form miRNA ribonucleoprotein (miRNP) complexes, RIP assay was conducted and the results implicated that STAT3 could be specifically recruited to the miRNP complexes isolated using Ago1 antibody following forced expression of miR-506 (Fig. 3C). qRT-PCR analysis demonstrated that STAT3 mRNA expression was aberrantly higher in primary NK cells from 15 HCC patients compared to that from 15 healthy controls (Fig. 3D), and negatively correlated with miR-506 expression in HCC patients (Fig. 3E). Moreover, we found that miR-506 mimic-transfected primary NK cells exhibited a significant decrease of STAT3 level, while anti-miR-506-treated primary NK cells displayed an evident increase of STAT3 level (Fig. 3F). Therefore, we concluded that miR-506 could suppress STAT3 expression by directly targeting 3′UTR of STAT3.

We found that STAT3 mRNA expression was negatively correlated with NK cell cytotoxicity against SMMC7721 cells (Fig. 4A). To demonstrate the effects of miR-506-STAT3 axis on NK cell cytotoxicity, NK-92 and primary NK cells were transfected with miR-NC, miR-506, miR-506 + pcDNA, miR-506 + STAT3, anti-miR-NC, anti-miR-506, anti-miR-506 + si-NC, or anti-miR-506 + si-STAT3. Unsurprisingly, STAT3 mRNA level was inhibited in NK-92 or primary NK cells transfected with miR-506, whereas introduction of STAT3 overexpression supported the abundance (Fig. 4B–E). Subsequently, CFSE/7AAD cytotoxicity assay showed that ectopic expression of miR-506 strikingly promoted cytotoxicity of NK-92 cells against SMMC7721 cells, while cotransfection with miR-506 and STAT3 dismissed this effect (Fig. 4F). In contrast, STAT3 knockdown significantly abolished anti-miR-506-mediated suppression of NK cell cytotoxicity against HepG2 cells at different E:T ratios (Fig. 4G). LDH assay also revealed that increased expression of miR-506 distinctly accelerated the cytotoxic activity of primary NK cells from HCC patients against SMMC7721 cells, which was obviously counteracted after overexpressing STAT3 (Fig. 4H). On the contrary, anti-miR-506-mediated inhibition of the cytotoxic activity of primary NK cells from HCC patients against HepG2 cells was dramatically abolished following STAT3 knockdown (Fig. 4I). Taken together, these results demonstrated that overexpressing STAT3 greatly reversed miR-506-mediated promotion of NK cell cytotoxicity against HCC cells.