Normal human dermal fibroblast (NHDF) cells were purchased from Lonza (Basel, Switzerland) and cultured in Dulbecco’s modified Eagle’s medium (DMEM, Gibco, Life Technologies, Grand Island, NY, USA) containing 10% fetal bovine serum (FBS, Sigma-Aldrich, St. Louis, MO, USA) at 37°C in a humidified chamber with 5% CO₂. To evaluate the patterns of cell cycle effects and miRNA expression, 7×10⁵ cells were seeded in a 60-mm culture plate and grown for 24 h. These cells then were irradiated with 0.1 Gy of γ-radiation using a MDI-KIRAMS 137 irradiator (¹³⁷Cs γ-ray source, KIRAMS, Seoul, Korea).

A BMI1 stable knockdown cell line was established using lentiviral gene transfer. The plasmid shBMI1-pLKO.1 puro was purchased from Sigma-Aldrich. The 293T cells were co-transfected with shBMI1-pLKO.1 puro lentiviral transfer vector, pCMV-dR8.2 (Addgene, Cambridge, MA, USA), and pCMV-VSV-G plasmid (Addgene). Approximately 48 h after transfection, recombinant lentivirus particle-containing medium was collected and used to infect NHDFs. Infected NHDFs were incubated with the growth culture medium containing puromycin (1 μg/ml) to select BMI1-knockdown cells.

Cell lysates were prepared in SDS lysis buffer [1% (w/v) SDS, 20 mM Tris-HCl pH 7.4, 20 mM EDTA]. Protein samples were subjected to SDS-PAGE and then transferred to a nitrocellulose membrane (Whatman International Ltd., Maidstone, UK). Membranes were incubated in a solution of 5% (w/v) skim milk in Tris-buffered saline and Tween-20 (TBST) buffer and probed with anti-β-actin IgG (Sigma-Aldrich) or anti-BMI1 IgG antibodies (Santa Cruz Biotechnology, Santa Cruz, CA, USA).

The effect of γ-radiation (0.1 Gy) on viability was determined using a 3-[4,5-dimethylthiazol- 2-yl]-2,5 diphenyl tetrazolium bromide (MTT) assay (Sigma-Aldrich), according to the manufacturer’s instructions. Briefly, seeded cells were irradiated with 0.1 Gy of γ-radiation. After 6 and 24 h of additional incubation, MTT solution was added to the irradiated cells, and the samples were incubated for 1 h. Media was removed, and the blue formazan crystals trapped in cells were dissolved in dimethyl sulfoxide (DMSO, Sigma-Aldrich). Cell viability was measured using an iMark plate reader (Bio-Rad, Hercules, CA, USA) at 590 nm with a reference filter of 620 nm. All results are presented as the mean percentages ± standard deviation (SD) of three independent experiments. A p-value of <0.05, as determined by Student’s t-test, was considered significant.

Cell cycle distribution was determined using FACS (Fluorescence Activated Cell Sorting) analysis. Irradiated cells were fixed by the addition of cold 70% ethanol overnight. Following fixation, cells were washed with cold PBS and then stained with propidium iodide (PI) staining solution (50 μg/ml PI, 0.5% Triton X-100, and 100 μg/ml RNase) at 37°C for 1 h. The PI fluorescence intensity was detected using a BD FACSCalibur flow cytometer (BD Biosciences, San Jose, CA, USA). The mean PI fluorescence intensity was obtained from 10,000 cells using the FL2-H channel.

Total RNAs were purified using TRIzol reagent (Life Technologies) according to the manufacturer’s protocol. The purity and integrity of the RNA sample was assessed using the ratio of absorbance at 230, 260 and 280 nm by MaestroNano^®, a micro-volume spectrophotometer (Maestrogen, Las Vegas, NV, USA) and Agilent 2100 Bioanalyzer^® (Agilent Technologies, Santa Clara, CA, USA). Microarray analyses were performed using SurePrint G3 Human V16 miRNA 8×60K microarray kit (Agilent Technologies), as previously described (29). Briefly, 50 ng of purified RNA was treated with calf intestine alkaline phosphatase prior to labeling with cyanine 3-cytidine bisphosphate (3-pCp). The labeled RNA was hybridized with the microarray kit in the Agilent Microarray Hybridization Chamber (Agilent Technologies) for 20 h. The fluorescence intensities of the labeled miRNA samples on the microarray were measured using Scanner and Feature Extraction software (Agilent Technologies). The digitalized fluorescence intensities were analyzed using GeneSpring GX version 11.5 (Agilent Technologies). The raw data were filtered using FLAG and t-tests and analyzed using the fold-change analysis, which was conducted based on a factor of a 2.0-fold difference between the two groups (i.e., non-irradiated control cells and irradiated cells).

To assess the biological significance of the altered miRNA expression, three bioinformatic analyses were performed: determination of putative target genes of the miRNAs using the DIANA-microT bioinformatic software tool (http://diana.imis.athena-innovation.gr/DianaTools/index.php?r=microT_CDS/index) (30), prediction of the target-related cellular pathways using the Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways and DAVID (Database for Annotation, Visualization and Interrogate Discovery, http://david.abcc.ncifcrf.gov/home.jsp) bioinformatics resources v6.7 according to the developer’s protocol (31), and categorization of target genes with specific biological functions using the AmiGO Gene Ontology (GO) analysis tool (http://amigo.geneontology.org/amigo). The prediction of target genes was limited by setting the value of the threshold to 0.8 in the DIANA-microT tool. The ‘KEGG_pathway’ category was processed by setting the threshold of EASE score, a modified Fisher Exact p-value, to 0.1, and involved KEGG pathways that displayed a value >1% (percentage of involved target genes/total target genes in each pathway) were selected. GO analysis was performed in eight categories for positive or negative regulation of apoptotic processes, cell growth, cell proliferation and cell cycle.

To evaluate the possible effect of BMI1 in resistance to LDR, we first generated BMI1 shRNA- or control shRNA-expressing NHDFs using a lentiviral system. After infection, the cells were treated with selection marker (puromycin) to obtain stable clones. The BMI1 protein level was clearly reduced in shBMI1-expressing clones, thereby confirming efficient knockdown (Fig. 1A). Using these stable clones, we then examined alterations in cell viability at 6 and 24 h after irradiation with 0.1 Gy of γ-radiation. Unexpectedly, no significant changes in cell viability were detected (data not shown), indicating that BMI1 may not be an essential regulator of LDR resistance in NHDFs. We next performed PI staining-based cell cycle analysis following irradiation. The cell cycle patterns of the irradiated cells at 6 and 24 h after irradiation were not significantly different from that of control cells (Fig. 1B). Although the proportion of cells in G1 and G2/M phase were slightly changed after 6 h by 5.88 and −3.21%, respectively, these values returned to baseline (0 h) at 24 h post-irradiation. Therefore, these results suggest that cells were only slightly sensitive to LDR even in the absence of BMI1 expression in NHDFs.

Although our results (Fig. 1) suggest a dispensable role for BMI1 in radioresistance to LDR, these studies did not fully exclude the possibility that such a modest radiosensitivity was caused by alteration of cellular pathways in BMI1-knockdown NHDFs. Recent evidence supports the role of BMI1 as a transcription factor engaged in cell proliferation, cell cycle, cell immortalization, chemoresistance and radioresistance (1–4,12). Notably, numerous cDNA microarray studies have sought to identify BMI1 target genes (14,32,33); however, miRNA microarray-based analysis of BMI1 putative target miRNAs has not yet been described. We first compared the miRNA expression profiles between control and BMI1-knockdown NHDFs not exposed to LDR. As shown in Fig. 2, 108 and 43 miRNAs are up- and downregulated more than 2-fold, respectively, in BMI1- knockdown NHDFs compared to control cells, indicating that BMI1 regulates the expression of specific miRNAs. Notably, expression levels of miR-17-3p, miR-1825 and miR-33b-3p were significantly increased by 81.71-, 61.18- and 30.62-fold, respectively, while expression levels of miR-328, miR-885-5p and let-7d-3p were significantly downregulated by 23.53-, 23.12- and 21.69-fold, respectively.

Next, we analyzed the LDR-induced alterations in miRNA expression profiles in BMI1-knockdown NHDFs. At 6 and 24 h after irradiation with 0.1 Gy of γ-radiation, total RNA was purified, and then miRNA microarray assays were performed. Although LDR induced minimal changes in cell viability and cell cycle distribution in BMI1-knockdown NHDFs (Fig. 1), we unexpectedly found that the expression of numerous miRNAs was altered more than 2-fold in the irradiated BMI1-knockdown NHDFs (Fig. 3). Specifically, LDR upregulated 21 specific miRNAs and downregulated 4 miRNAs at the 6-h time point in BMI1-knockdown NHDFs (Fig. 3A). Also, 4 and 16 miRNAs were significantly upregulated and downregulated, respectively, by LDR at 24 h after irradiation (Fig. 3B). The full data are shown in Tables I and II. Among these miRNAs, miR-202-3p and miR-4323 were the most upregulated miRNAs by 81.90- and 66.47-fold at 6 and 24 h after irradiation, respectively, and miR-758 and miR-1224-5p were the most downregulated miRNAs (21.20- and 106.85-fold, respectively) (Tables I and II). Overall, these data suggest that BMI1 altered miRNA expression in response to LDR.

In total, we found 108 upregulated and 43 downregulated miRNAs in BMI1- knockdown NHDFs and 21 upregulated and 4 downregulated miRNAs in BMI1-knockdown cells 6 h post-irradiation. In addition, 4 and 16 miRNAs were upregulated and downregulated, respectively, in these cells 24 h post-irradiation (Fig. 4A). Seven miRNAs were upregulated in both BMI1-knockdown cells and in cells 6 h post-irradiation (Fig. 4B, upper left). In particular, 2 of the 21 miRNAs were found to be differentially expressed in either BMI1-knockdown cells or in BMI1- knockdown cells 6 h post-irradiation (Fig. 4B, upper left). These results also indicated that 2 upregulated miRNAs were specific to BMI1-knockdown NHDFs at 6 h post-irradiation (Fig. 4B, upper left). Interestingly, no downregulated miRNAs were specific to the profile of these cells at 6 h post-irradiation (Fig. 4B, lower left). Similarly, the differentially expressed miRNAs in BMI1-knockdown NHDFs at 24 h post-irradiation were also shared, in part, with BMI knockdown-induced miRNA alterations (Fig. 4B, upper and lower right). The group of differentially expressed miRNAs in BMI1-knockdown NHDFs and the miRNAs upregulated in BMI1-knockdown NHDFs 24 h post-irradiation had 2 miRNAs in common, and thus, only a subset of 150 and 2 miRNAs were unique to each group, respectively (Fig. 4B, upper). Furthermore, the sets of altered miRNAs in BMI1-knockdown NHDFs 24 h post-irradiation and BMI1-knockdown NHDFs also had 13 miRNAs in common, and only a subset of 3 and 139 miRNAs were unique to each group, respectively (Fig. 4, lower). Overall, these results indicate that knockdown of LDR and BMI1 affects both unique and shared miRNAs in NHDFs.

We next examined the biological influence of the observed altered miRNAs. Because biological functions of miRNAs are mainly determined by the regulation of their specific target mRNAs, the putative target of each miRNA was first predicted using the DIANA-microT-CDS (v5.0) web-based bioinformatics tool. The prediction threshold, which is a cut-off value for presented prediction and ranges from 0.3 to 1.0, was fixed at 0.8. Following prediction, a Gene Ontology and KEGG pathway annotation of the miRNA targets were performed using the AmiGO and DAVID bioinformatics tools. Using AmiGO, we categorized the target genes into eight types according to biological function: positive and negative regulation of the apoptotic process, cell growth, cell proliferation and cell cycle. The target genes for the miRNAs that were upregulated in BMI1-knockdown NHDFs at 6 h post-irradiation (written as ‘6PB-NHDFs’) are involved in both positive and negative regulation of the four biological functions. Similarly, the target genes for the altered miRNAs in the BMI1-knockdown NHDFs at 24 h post-irradiation (written as ‘24PB-NHDFs’) are also involved in both positive and negative regulation of the four biological functions (Fig. 5). Interestingly, the targets of the upregulated miRNAs in 6BP-NHDFs are largely involved in these processes; however, the targets of the downregulated miRNAs in 24BP-NHDFs are largely involved in these processes. These results indicate that the miRNA-mediated LDR response is functionally engaged with both positive and negative biological processes.

A KEGG pathway annotation of the miRNA targets was then performed based on scoring of the pathways collected in the KEGG databases using the DAVID tool. The EASE Score, which is a modified Fisher Extract p-value, was fixed at 0.1, and meaningful KEGG pathways showing a value >1% (percentage of involved target genes/total genes involved in each pathway) were selected. From these analytic processes, unique and shared pathways were found to be involved with either the 6BP- or 242BP-NHDFs. Axon guidance, cell adhesion molecules, MAPK signaling pathway, and pathways involved in cancer were the main overrepresented pathways in the targets of the upregulated miRNAs in 6BP-NHDFs (Table III), whereas, Wnt, mTOR, MAPK and ErbB signaling pathways were common in the targets of the downregulated miRNAs in 6BP-NHDFs (Table IV), suggesting that these pathways are significantly regulated after 6 h of post-irradiation with LDR. The full lists for the identified pathways are presented in Tables III and IV. We also found that the Wnt, MAPK and ErbB signaling pathways were unique specific to several upregulated miRNAs in 24BP-NHDFs (Table V); however, these pathways were mostly shared involving the targets of the downregulated miRNAs in 24BP-NHDFs (Table VI). Overall, the results of the pathway analysis demonstrate possible roles for the differentially expressed miRNAs in the LDR response.