The SV40-immortalized HCEC line was kindly provided by Dr. Kaoru Araki-Sasaki (14) (Osaka University, Osaka, Japan). HCECs were cultured in Dulbecco’s modified Eagle’s/F12 medium (DMEM/F12, HyClone; GE Healthcare Life Sciences, Logan, UT, USA) supplemented with 10% fetal bovine serum (FBS; Gibco; Thermo Fisher Scientific, Inc., Waltham, MA, USA) and 1% penicillin/streptomycin and were plated at a density of 1 × 10⁵ cells/cm². The medium was changed every 2 days, and the cells were passaged until full differentiation was reached. Passage five cells were used in the experiment.

Peripheral blood mononuclear cells (PBMCs) were isolated from peripheral venous blood obtained from normal healthy volunteers, according to the instructions provided in the Human Lymphocyte Separation Medium manuscript (Biolegend, San Diego, CA, USA). Additionally, NK cells and CD8+ T cells were labeled by fluorescent antibody (NK cells: FITC-CD3⁻, APC-CD16⁺, RPE-CD56⁺; CD8+ T cells: FITC-CD8⁺) and were isolated from the PBMCs by flow cytometry sorting. They were then activated by IL-2 (Chiron, NC, USA; 50 U/ml for NK cells and 100 U/ml for CD8+ T cells).

In apoptosis experiments, NK cells and CD8+ T cells were cocultured with HCECs in DMEM/F12 complete medium. Before the coculture procedure, HCECs were seeded onto 24-well U-bottom plates for 12 h at 37°C. Effector (E) to target (T) ratios (E:T) were 6:1 for NK cells and 10:1 for CD8+ T cells. For blocking studies, NK and CD8 + T cells were pre-incubated for 30 min with 100 µg/ml of NKG2D-mAbs (Biolegend).

To evaluate the concentration dependence, HCECs were treated with various doses of IFN-γ (100, 500, and 1,000 ng/ml) for 24 h, and the subsequent time-dependent experiments were performed using 500 ng/ml of IFN-γ, incubating for 1, 12, and 24 h. In NFκB-related experiments, HCECs were treated with 100 ng/ml pyrrolidine dithiocarbamate (PDTC) for 2, 4, 6, 8, and 10 h to determinate the time dependence. HCECs were then treated with different doses of PDTC (0, 10, 25, 50, 100, and 200 ng/ml) for 2 h.

A Cell Counting Kit 8 (Dojindo, Kumamoto, Japan) was used to detect cell proliferation according to the user guidelines. Briefly, 1 × 10³ HCECs were added to 96-well-plates in 100 µl cell culture medium per well. The cells were then treated with 500 ng/ml IFN-γ for 24 h or not treated at all. Next, CCK8 solution was injected into each well for 1 h at 37°C until the color changed. The absorbance of each sample was determined at 450 nm using an Epoch microplate spectrophotometer (Bio Tek, Biotek Winooski, VT, USA) and a Synergy hybrid reader (Bio Tek). The assay was repeated in triplicate.

Total RNA was isolated from HCECs using Trizol (Invitrogen, Carlsbad, CA, USA) according to the manufacturer’s instructions. cDNA for miRNA and mRNA real-time PCR was synthesized using a One Step PrimeScript^® miRNA cDNA Synthesis kit (cat. no. D350A; Takara Biotechnology Co., Ltd., Dalian, China) and a PrimeScript™ RT reagent kit with gDNA Eraser (cat. no. RR047A; Takara Biotechnology Co., Ltd.), respectively. The SYBR^® Premix Ex Taq™ II Real-Time PCR kit (cat. no. DRR081A; Takara Biotechnology Co., Ltd.) was used with the ABI ViiA7 Real-Time PCR System (Thermo Fisher Scientific, Inc.) for the qPCR analysis. U6 RNA was used as a control for the miRNA, and GAPDH was used as a control for the mRNA. The primers used are listed in Table 1.

The amount of MICA present in the supernatant of the cell culture medium was detected using a human MICA ELISA kit (ProSpec-Tany; TechnoGene, Ness-Ziona, Israel) according to the manufacturer’s protocol.

IFN-γ was added to the culture medium at a final concentration of 500 ng/ml, incubating for 24 h. The total RNA amount was obtained according to the Trizol method and then sent to the MicroRNA Microarray Service (Gminix Biologic Science Co., Ltd, Shanghai, China), which used an AffymetrixGeneChip miRNA array. Differential miRNA expression was defined using a cut-off value of a twofold change and analyzed by Gene Ontology (GO; www.geneontology.org) using KEGG analysis and Ingenuity Pathway Analysis (IPA). mirBase, miRWalk, and targetScan bioinformatics software were used to predict the potential target genes of the miRNA.

The lentiviral vector, pLenti-CMV-miRNA4448, negative control (NC), pLenti-CMV-NC, and lentiviral packaging plasmids (Biofavor Biotechnology, Wuhan, China) were co-transfected into 293FT packaging cells. The sequence of the miRNA4448 was as follows: gcaattattctttattccaatattataataatcctcgctctataatcataacctaggaaaaaccgggccatacagaggcaggagctgaggggacatagtgaggagtgaccaaaagacaagagtgcgagccttctattatgcccagacagggccaccagagggctccttggtctaggggtaatgccagcgtctgggaagatgcccgttgccaagcagactgtggtctagcggtagcatgtcaaggaaaaacacctgctacttagtagtccctgggggagtttagaga.

The virus-containing supernatant was harvested and filtered 72 h post-transduction. The virus was then used to infect the HCEC line with Polybrene^® (Sigma-Aldrich, St. Louis, MO, USA) at a final concentration of 5 µg/ml. Cell culture medium was replaced with fresh complete medium after 24 h of transfection, and the HCEC line was incubated for an additional 24–72 h prior to observing the expression of green fluorescent protein (GFP). Following flow-cytometer (Epics Altra; Beckman Coulter, Inc., Brea, CA, USA) selection for GFP (+) cells, the stable overexpression of miRNA4448 clones was obtained. The analysis software used was Expo32 V1.2 Analysis Multicycle for Windows (Beckman Coulter, Brea, CA, USA).

Western blotting was performed using standard protocols. In brief, nuclear proteins were isolated using a nuclear extraction kit (Active Motif, Carlsbad, CA, USA). In addition, cell lysates were separated by sodium dodecyl sulfate polyacrylamide gel electrophoresis, transferred to a polyvinylidene fluoride membrane, and blocked in 5% skimmed milk. Membranes were incubated overnight at 4°C with primary antibodies: anti-rabbit P-P65 (1:1,000; Santa Cruz Biotechnology, Inc., Dallas, TX, USA) or anti-rabbit Lam B1 (1:1,000; Miao Tong Biological Science & Technology Co., Ltd., Shanghai, China). An anti-rabbit secondary antibody (1:2,000; Miao Tong Biological Science & Technology Co., Ltd.) was added to membranes at room temperature (RT) for 1 h. The specific protein bands were visualized using enhanced chemiluminescence plus Western blotting detection reagents (GE Healthcare Life Sciences, Pittsburgh, PA, USA) and an LAS4000 luminescent image analyzer (Kodak, Rochester, NY, USA).

Human corneal epithelial cells (HCEs) were plated in 24-well plates 24 h before the experiment. Cells were fixed in 4% paraformaldehyde for 30 min at RT and then washed twice with cold PBS. In the membrane rupture period, a cold 1% BSA/PBS solution containing 0.2% Triton X-100 was added to cells for 5 min at 4°C. After three rinses with PBS for 3 min each time, the cells were blocked with 1% BSA/PBS for 1.5 h at RT and then incubated with a P-P65 antibody (1:100; Santa Cruz Biotechnology, Inc.) for 2 h. The cells were washed three times with PBS and then incubated with a FITC-conjugated secondary antibody (1:1,000; Sigma) for 1 h. After rinsing with PBS, the cells were counterstained with DAPI (10 ng/ml) and then mounted. Each sample was analyzed using a fluorescence microscope.

To investigate the influence of NFκB on MICA, the binding site between NFκB and MICA was first predicted. The MICA promoter sequence contains a putative binding site (178–192: atcggaatcacctag) with NFκB. The wild type (WT) or mutant MICA promoter were cloned into the pGL4.10 plasmid (Obio, Shanghai, China). The WT construct contained the potential binding sequence, and the mutant construct contained the sequence with a –ggaa– and –cc– deletion. The NFκB 3Flag sequence was constructed and cloned into the pcDNA3.1 vector (Obio, Shanghai, China). 293 T cells grown in 96-well plates were harvested for assays 48 h after transfection with pcDNA3.1-NFKB-3Flag and either the pGL4.10-MICA promoter-WT or pGL4.10-MICA promoter-mutant via Lipofectamine 2000 (Invitrogen). The firefly and Renilla luciferase activities were measured using a dual-luciferase reporter assay system (Promega Corporation) with a microplate luminometer, and each sample’s luciferase activity was normalized to that of Renilla.

We used two types of MICA plasmids in our research. In order to overexpress MICA in the HCECs, in Annexin V-PI staining, we used pcDNA3.1-MICA and pcDNA3.1-empty plasmid (GeneCopoeia, Rockville, MD, USA), which were transfected into HCECs using Lipofectamine 3000 (Invitrogen). In the Caspase and Survivin mRNA expression study, a MICA ORF clone was transfected into the pcDNA3.1-GFP plasmid, and the pcDNA3.1-GFP-MICA or empty control vector was transfected into HCECs using Lipofectamine 2000 (Invitrogen). The cells were cultured for 24–72 h following transfection, and flow-cytometry was used to obtain GFP (+) cells.

After washing three times with PBS, HCECs were treated with 0.25% trypsin to prepare the cell suspension. FBS was used to neutralize the trypsin, and the cells were then rinsed three additional times followed by centrifugation for 5 min at 1,000 rpm. Next, 300 µl of binding buffer was added to suspend the cells, and the cells were incubated with Annexin V/PI (5 µl, 20 min for Annexin V and 5 µl, 5 min for PI). The apoptosis rate of each group was analyzed using flow cytometry.

Each experiment was performed in triplicate, and the data were expressed as mean ± SD. Differential miRNA expression, miRNA4448 expression after pLenti-CMV-GFP-miR4448 transfection, and the CCK8 proliferation assay were analyzed using a Student’s t test (SPSS 13.0, USA). Differences among more than two groups, such as the MICA and P65 expression, luciferase reporter assay and Caspase and Survivin mRNA expression, were evaluated using a one-way ANOVA. P < 0.05 was considered significant.

With the IFN-γ treatment for 24 h, the expression levels of MICA protein and MICA mRNA in HCECs were upregulated in a concentration-dependent manner (Figure 1). MICA mRNA and protein expression levels reached their peak at 12 and 24 h after IFN-γ (500 ng/ml) treatment, respectively (Figure 1). Therefore, treating HCECs with 500 ng/ml of IFN-γ for 24 h was identified as the optimal condition. Additionally, the CCK8 assay showed that 500 ng/ml of IFN-γ treating HCECs for 24 h had no toxic effect on cells (450 nm optical density value: con VS IFN-γ: 1.10 ± 0.10 vs 1.12 ± 0.07, p > 0.05).

There were two upregulated miRNAs and four downregulated miRNAs in the IFN-γ-treated HCECs according to the Affymetrix Gene Chip miRNA array. GO and IPA analysis showed that four of the six miRNAs—hsa-miR-17-5p, hsa-miR-4448, hsa-miR-4481, and hsa-miR-3136-5p—might regulate several critical pathways that influence MICA expression. The regulator including the NFκB, human protection of telomeres 1, cluster of differentiation 28 (CD28), and mitogen-activated protein kinase (MAPK) pathways. According to qRT-PCR (Figure 2; Table 2), the fold change values for the miRNA in the IFN-γ group were as follows: miRNA4448, 0.31 ± 0.14-fold (p < 0.05); miRNA3136-5p, 0.91 ± 0.13-fold (p > 0.05); miRNA4481, 1.98 ± 0.33-fold (p < 0.05); and miRNA17-5p, 1.64 ± 0.19-fold (p < 0.05). Among them, miRNA4448, which displayed the most significant change, was selected for further investigation.

Human corneal epithelial cells revealed normal cell morphology up to 72 h after transfection with pLenti-CMV-GFP-miR-4448 or pLenti-CMV-GFP-NC. GFP fluorescence showed a small, medium, and large amount of expression at 24, 48, and 72 h post-transfection, respectively, and the fluorescence intensity increased in a time-dependent manner (Figure 3). Furthermore, qRT-PCR indicated that miRNA4448 expression was 2.118 ± 0.29 times increased in pLenti-CMV-GFP-miR-4448 HCEC lines (p < 0.05; Figure 3).

In the pLenti-CMV-GFP-NC HCEC line, MICA mRNA levels increased 2.94 ± 0.24 times, and protein levels increased 1.32 ± 0.08 times after IFN-γ treatment (Figure 4). Overexpression of miRNA4448 significantly decreased the MICA mRNA and protein expression to 55.53 and 85.48%, respectively, compared to the pLenti-CMV-GFP-NC HCEC lines (Figure 4). Furthermore, IFN-γ treatment could partially reverse the MICA mRNA expression (Figure 4).

Based on immunofluorescent staining, IFN-γ treatment markedly promoted the phosphorylation and nuclear translocation of P65, while this effect was substantially inhibited by overexpression of miRNA4448 (Figure 5; Table 3). We also examined the expression of P65 using qRT-PCR. IFN-γ treatment upregulated P65 mRNA expression levels by 1.57 ± 0.27 fold, while the overexpression of miRNA4448 downregulated P65 mRNA levels by 50.9% compared to the control group (Figure 5). P65 mRNA expression levels were partially reversed to 75% in the miRNA4448 overexpression group following IFN-γ treatment (Figure 5). Overall, these results suggest that IFN-γ activated NFκB/P65 through the regulation of miRNA4448 in HCECs.

When HCECs was treated with PDTC for 2 h, the expression of P65 mRNA decreased significantly to 41% compared to the control group (p < 0.05, Figure 6). As time progressed, there was no significant difference between the 4-, 6-, 8-, and 10-h treatment groups compared to the 2-h treatment group (p > 0.05, Figure 6). Meanwhile, the P-P65 protein levels decreased the most in the 2-h treatment group. The qRT-PCR analysis showed that when the PDTC concentration increased from 0 to 50 ng/ml, P65 mRNA expression was gradually inhibited, but there was no statistical difference among the 50, 100, and 200 ng/ml groups (p > 0.05, Figure 6). P65 mRNA levels decreased to 30.8% in the 50 ng/ml PDTC group compared to the control group. Moreover, the downregulation of the P-P65 protein was the most significant in the 50 ng/ml group (Figure 6). Next, 500 ng/ml IFN-γ was used to treat cells for another 24 h. This showed that IFN-γ could significantly upregulate the expression of MICA mRNA levels by 1.58 times compared to the PDTC group.

cDNA3.1-NFκB-3Flag and either pGL4.10-MICA promoter-WT or pGL4.10-MICA promoter-mutant were co-transfected to investigate the potential function of NFκB on MICA. The results show that NFκB significantly increased the luciferase activity of pGL4.10-MICA promoter-WT in the 293 T cells (p < 0.001; Figure 7), whereas it had no significant effect on the mutant group (p > 0.05; Figure 7). This indicates that NFκB can act on the MICA promoter region and promote MICA gene expression.

pcDNA3.1-GFP-MICA or pcDNA3.1-GFP-NC was transfected into HCECs for 24, 48, or 72 h. The expression of MICA mRNA was 1.836 ± 0.155 times greater in the pcDNA3.1-GFP-MICA line than in the NC group, as determined by qRT-PCR (p < 0.05). According to the ELISA results, the MICA protein levels were significantly upregulated by 1.54 times compared to the NC group. Overall, the transfection of pcDNA3.1-GFP-MICA in the HCEC line resulted in increased MICA mRNA and protein levels.

To confirm the apoptosis effect of MICA, Annexin V/PI staining was used to quantify the number of apoptotic cells after HCECs were cocultured with allergic NK cells and CD8+ T cells. As shown in Figure 8, the ratio of late apoptotic cells increased from 4.45 to 8.20 and 8.25% when HCECs were incubated with IFN-γ or transfected with MICA-plasmid compared with the control group. On the contrary, when HCECs were pre-incubated with NKG2D-mAb, the percentage of late apoptotic cells decreased significantly to 3.29%. Collectively, these results demonstrate that the interaction between HCECs and NK (CD8+ T) cells upon MICA stimulation results in the late apoptosis of HCEC cells and is mediated by MICA–NKG2D interaction.

As determined by RT-qPCR analysis, in pcDNA3.1-GFP-MICA-transfected or IFN-γ treatment groups, the expression level of Survivin was obviously decreased and that of Caspase3 was significantly increased when HCECs were cocultured with NK cells and CD8+ T cells (p < 0.01; Figure 9). The overexpression of MICA increased NK cells and CD8+ T cell-induced HCEC cell apoptosis via the regulation of Survivin and Caspase3.