Human NSCLC cells PC9 (parental) cells, PC9/GR (gefitinib-resistant) cells, and HCC827 cells were cultured in RPMI-1640 medium containing 10% fetal bovine serum (FBS) at 37°C. PC9/GR cells were induced using progressive concentrations of gefitinib. Briefly, PC9 cells in logarithmic growth were treated with 0.5 µmol/l of gefitinib. After 48 h, gefitinib was removed and the cells were cultured without gefitinib until they recovered. The same treatment was then performed again, and when the cells were resistant to the current concentration, the gefitinib concentration was gradually increased to 1, 2 µmol/l, and finally to 3 µmol/l. When the induced cells survived 3 µmol/l of gefitinib for ~2 months with normal activity, the cells were confirmed to be gefitinib-resistant and named PC9/GR. The PC9/GR cells were cultured with 1 µmol/l gefitinib.

Cells were transfected with miRNA mimics, miRNA inhibitors, siRNAs, and plasmids using Lipofectamine Transfection reagent (Invitrogen; Thermo Fisher Scientific, Inc.). In the present study, we used a mimic mixture of nine let-7 members or six miR-17 members and inhibitors mixture of nine let-7 members or six miR-17 members to enhance and inhibit the let-7 family or miR-17 family simultaneously (Qiagen GmbH). The sequences of miRNA mimics and miRNA inhibitors are listed in Table I.

Fifty-six NSCLC patients (Table II) were recruited for this study. Inclusion criteria were as follows: Patients with primary NSCLC; with a histological diagnosis of NSCLC with at least one measurable lesion; with a TNM clinical stage of IIIB to IV; who had undergone molecular-targeted therapy with gefitinib. Tissue samples were obtained and divided into two groups according to patient responses assessed using Response Evaluation Criteria in Solid Tumors (RECIST). Patients with a response or partial response to treatment were considered to be gefitinib sensitive (GS), and patients with stable or progressive disease were considered to be gefitinib resistant (GR).

All patients provided written informed consent. NSCLC tissues were collected from the Cancer Center of Guangzhou Medical University (Guangzhou, China) with permission from the Institutional Review Board. The study protocol was approved by the Ethics Committee of the Cancer Center of Guangzhou Medical University [approval no. (2014) 66].

Total RNA from PC9 and PC9/GR cells was isolated using a Total RNA Purification kit (Qiagen GmbH). Microarray hybridization assays were carried out in two experiments: PC9 cells (Cy5-labeled) compared with PC9/GR cells (Cy3-labeled). Data were analyzed using LOWESS filters and t-tests.

Total RNA in cells and tissue samples was extracted by TRIzol™ (Invitrogen; Thermo fisher Scientific, Inc.). For the detection of mRNA and miRNA, RT-PCR was performed using gene specific primers or the miScript™ Primer assay kit (Qiagen GmbH) after reverse transcription from RNA to cDNA. Relative expression of mRNA and miRNA was normalized by GAPDH and U6, respectively.

RT-PCR primers were designed as follows: MYC forward, 5′-CGTCCTCGGATTCTCTGCTC-3′ and reverse, 5′-GCTGGTGCATTTTCGGTTGT-3′; CDKN1A forward, 5′-TGCCGAAGTCAGTTCCTTGT-3′ and reverse, 5′-CATTAGCGCATCACAGTCGC-3′; CD44 forward, 5′-TTACAGCCTCAGCAGAGCAC-3′ and reverse, 5′-TGACCTAAGACGGAGGGAGG-3′; ALDH1 forward, 5′-CTGTGTTCCAGGAGCCGAAT-3′ and reverse, 5′-AGCATCCATAGTACGCCACG-3′; OCT4 forward, 5′-TGTCAGGGCTCTTTGTCCAC-3′ and reverse, 5′-TCTCCCCAGCTTGCTTTGAG-3′; GAPDH forward, 5′-TGACTTCAACAGCGACACCCA-3′ and reverse, 5′-CACCCTGTTGCTGTAGCCAAA-3′.

By using TargetScan prediction system (version 7.1; http://www.targetscan.org/vert_71/), we identified a number of potential miRNA targets. The wild-type (WT) and mutant-type (MUT) 3′untranslated region (3′-UTR) of MYC and CDKN1A were synthesized chemically and inserted into the psiCHECK™-2 vector (Promega Corp., Madison, WI, USA) to obtain psiCHECK™-2-MYC-3′-UTR (WT or MUT plasmids) and psiCHECK™-2-CDKN1A-3′-UTR (WT or MUT plasmids). Cells were transfected with WT plasmids or MUT plasmids in the presence of miRNA mimics or a non-target control (NC). At 48 h after transfection, the luciferase activity of cells was assessed according to the Dual-Luciferase reporter assay system (Promega Corp.).

Proteins were extracted from cells by RIPA buffer (Thermo Fisher Scientific, Inc.) for 30 min at 4°C. A total of 50 µg proteins per lane were determined by BCA method, loaded into 15% SDS-PAGE and blotted onto polyvinylidene fluoride (PVDF) membranes for analysis. Blocking buffer (5% skim milk), was then added and shaken gently for at least 1 h at 25°C. The primary antibody was rabbit polyclonal anti-CD44 (cat. no. 37259), anti-ALDH1 (cat. no. 36671) and anti-OCT4 (cat. no. 2890) (1:1,000 dilution; shaken gently overnight at 4°C; Cell Signaling Technology, Inc., Danvers, MA, USA). The secondary antibody was anti-rabbit IgG, HRP-linked antibody (cat. no. 7074) (1:1,000 dilution; shaken gently for at least 1 h at 25°C; Cell Signaling Technology, Inc.). Proteins were detected by SuperSignal chemiluminescence substrate (Pierce, Rockford, IL, USA). Actin was used as an internal control.

Cells were incubated in DMEM F12 serum-free medium with 20 ng/ml of EGF, 20 ng/ml of bFGF, 2% B27 and 1%methylcellulose (Invitrogen; Thermo Fisher Scientific, Inc.). After 4–7 days, microsphere-like structures were visible, and images of the microspheres were captured using a light microscope.

The cell cytotoxicity assay was performed using a CCK-8 kit (Beyotime Institute of Biotechnology, Haimen, China). Cells were incubated with different concentrations of gefitinib. After 48 h, the medium was removed, and 90 µl of medium and 10 µl of CCK-8 solution were added to each well of the plate. The plate was incubated for 3 h at 37°C. The absorbance at a wavelength of 450 nm was measured by an automated reader. Gefitinib-induced cytotoxicity was represented as the IC₅₀ value (µmol/l).

Cell apoptosis was determined using the Apoptosis Analysis kit (Beyotime Institute of Biotechnology). Cells were incubated with gefitinib for 48 h and then collected and fixed in 70% ethanol overnight at 4°C. The cells were labeled with Annexin V-FITC and propidium iodide (PI) and analyzed using flow cytometry. The cell apoptosis ratio was analyzed using FlowJo software (FlowJo LLC, Ashland, OR, USA).

The cell cycle assay was performed using the Cell Cycle Analysis kit (Beyotime Institute of Biotechnology). Cells were collected and fixed in 70% ethanol overnight at 4°C. The cells were stained with PI and analyzed using flow cytometry. The cell cycle was analyzed using FlowJo software (FlowJo LLC).

Values were presented as the mean ± standard deviation (SD) of at least three separate experiments. The IC₅₀ values were assessed by linear regression analysis. The Student's unpaired t-test, one-way analysis of variance (ANOVA) and receiver operating characteristic (ROC) curves were performed using SPSS, version 21.0 (IBM Corp., Armonk, New York, USA). Multiple comparisons between the groups was performed using Bonferroni method. P<0.05 (two-tailed) was considered to indicate a statistically significant difference.

miRNA microarray chip analysis was used to detect the miRNA expression profiles of gefitinib-resistant PC9/GR cells and gefitinib-sensitive PC9 cells. The results of microarray analyses and RT-PCR confirmed that in comparison with PC9 cells, the expression levels of nine members of the let-7 family (let-7a, let-7b, let-7c, let-7d, let-7e, let-7f, let-7g, let-7i and miR-98) were downregulated in PC9/GR cells, and the expression levels of six members of the miR-17 family (miR-17, miR-20a, miR-20b, miR-93, miR-106a and miR-106b) were upregulated in PC9/GR cells (Fig. 1A and B). Therefore, these results indicated that the let-7 family and miR-17 family were involved in NSCLC gefitinib resistance.

It was then determined whether let-7 and miR-17 were associated with the outcome of gefitinib therapy in NSCLC patients. Using RT-PCR, the expression levels of let-7 were found to be significantly lower in gefitinib-resistant patients (GR samples) compared with gefitinib-sensitive patients (GS samples) (Fig. 1B). In contrast to let-7, the expression levels of miR-17 were significantly higher in GR samples than in GS samples (Fig. 1B). These results indicated that low expression levels of let-7 and high expression levels of miR-17 were positively associated with a poor response to gefitinib therapy in NSCLC patients. To determine the potential of let-7 and miR-17 as biomarkers, the expression levels of let-7 and miR-17 were analyzed as continuous or categorical variables in ROC analyses. Compared with separate analyses of let-7 and miR-17, combined analyses of let-7 and miR-17 produced a higher area under the curve (AUC) score (Fig. 1C). Moreover, according to the association analysis between the expression levels of the let-7 and miR-17 and the clinicopathological factors of NSCLC patients, the clinical outcome of gefitinib therapy was associated with the expression levels of the let-7 in NSCLC tissues (P=0.057, Table III) and significantly correlated with the expression levels of the miR-17 in NSCLC tissues (P=0.015, Table IV). These results indicate that let-7 and miR-17 could be potential biomarkers for predicting the clinical outcome of gefitinib therapy in NSCLC.

In order to investigate the regulatory functions of let-7 and miR-17 in gefitinib resistance, we used a miRNA mimics mixture and a miRNA inhibitors mixture of the let-7 family or the miR-17 family to determine the influence of let-7 and miR-17 on gefitinib resistance (Fig. 2A). Cell cytotoxicity and apoptosis assays were used to compare the gefitinib resistance of PC9/GR cells and PC9 cells. Compared with PC9 cells, PC9/GR cells revealed increased resistance to gefitinib (the IC₅₀ increased from 0.49 to 5.66 µmol/l) and decreased gefitinib-induced cellular apoptosis (Fig. 2B and C). In PC9/GR cells, upregulation of let-7 or downregulation of miR-17 restored sensitivity to gefitinib (the IC₅₀ decreased from 5.37 to 3.92 µmol/l and from 5.37 to 3.85 µmol/l, respectively) and increased gefitinib-induced cellular apoptosis (Fig. 2D and E). Conversely, in PC9 cells, downregulation of let-7 and upregulation of miR-17 protected PC9 cells from gefitinib (the IC₅₀ increased from 0.51 to 1.13 µmol/l and from 0.51 to 0.96 µmol/l) and decreased gefitinib-induced cellular apoptosis (Fig. 2F and G). Moreover, combined regulation of let-7 and miR-17 influenced gefitinib resistance more significantly compared with single regulation of let-7 or miR-17 (the IC₅₀ decreased from 5.37 to 2.39 µmol/l in PC9/GR cells, and the IC₅₀ increased from 0.51 to 2.07 µmol/l in PC9 cells) (Fig. 2D and F). Using another NSCLC cell line, HCC827, the effects of let-7 and miR-17 on gefitinib resistance were also observed (Fig. 2H and I). Therefore, these findings indicated that both let-7 and miR-17 may play important roles in the regulation of gefitinib resistance in NSCLC.

The self-renewal capacity of PC9/GR cells was compared with that of PC9 cells using the microsphere formation assay. It was found that the number and size of microspheres formed by PC9/GR cells were greater than the microspheres formed by PC9 cells (Fig. 3A and B). The effects of let-7 and miR-17 on the self-renewal capability were then investigated. The results revealed that the number of microspheres formed by PC9/GR cells decreased when let-7 was upregulated or miR-17 was downregulated, and the number of microspheres formed by PC9 cells increased when let-7 was downregulated or miR-17 was upregulated (Fig. 3A and B). Moreover, consistent with the results of the cytotoxicity analyses, the combined regulation of let-7 and miR-17 expression increased the number of microspheres more significantly than single regulation of let-7 or miR-17 (Fig. 3A and B). Using another NSCLC cell line, HCC827, the effects of let-7 and miR-17 on self-renewal were also observed (Fig. 3C). Therefore, these data indicated that let-7 and miR-17 influenced gefitinib resistance by regulating the self-renewal capability of NSCLC cells.

Since miRNAs function by silencing target genes, we used the TargetScan prediction system and Dual-Luciferase reporter assay to determine whether let-7 binds to the 3′-UTR of MYC (Fig. 4A). RT-PCR revealed that MYC was higher in PC9/GR cells than in PC9 cells, and upregulation of let-7 decreased MYC expression in PC9/GR cells and downregulation of let-7 increased MYC expression in PC9 cells (Fig. 4B and C). Furthermore, we used overexpression plasmids and siRNA to determine the influence of MYC on gefitinib resistance (Fig. 4D). The results revealed that in PC9/GR cells, transfection with si-MYC decreased gefitinib resistance and microsphere formation capacity, and transfection with MYC plasmid inhibited the reducing effect of let-7 upregulation on gefitinib resistance and microsphere formation capacity (Fig. 4E and F). Transfection with MYC plasmid in PC9 cells increased gefitinib resistance and microsphere formation capacity, and transfection with si-MYC inhibited the promoting effect of let-7 downregulation on gefitinib resistance and microsphere formation capacity (Fig. 4E and F). These results indicated that let-7 influenced gefitinib resistance and self-renewal capacity by directly regulating MYC in NSCLC cells.

MYC is an important transcription factor in stem cells, and both RT-PCR and western blotting assays were used to detect the effect of let-7 and MYC on the expression of stem cell markers CD44, ALDH1 and OCT4. The expression of CD44, ALDH1, and OCT4 in PC9/GR cells increased compared with PC9 cells (Fig. 4G). In PC9/GR cells, upregulation of let-7 or transfection with si-MYC decreased CD44, ALDH1, and OCT4 expression, and transfection with MYC plasmid inhibited the reducing effect of let-7 upregulation on CD44, ALDH1 and OCT4 expression (Fig. 4G). In PC9 cells, downregulation of let-7 or transfection with MYC plasmid increased CD44, ALDH1 and OCT4 expression, and transfection with si-MYC inhibited the promoting effect of let-7 downregulation on CD44, ALDH1 and OCT4 expression (Fig. 4G). Using another NSCLC cell line, HCC827, the effects of let-7 on stemness maintenance were also observed (Fig. 4H). Therefore, these findings indicated that let-7 regulated self-renewal by targeting MYC essential for stemness maintenance.

The TargetScan prediction system and dual-luciferase reporter assay were used determine whether miR-17 binds to the 3′-UTR of CDKN1A (Fig. 5A). It was revealed that CDKN1A was lower in PC9/GR cells compared with PC9 cells, and downregulation of miR-17 increased CDKN1A expression in PC9/GR cells and upregulation of miR-17 decreased CDKN1A expression in PC9 cells (Fig. 5B and C). Moreover, we used overexpression plasmids and siRNA to determine the influence of CDKN1A on gefitinib resistance (Fig. 5D). In PC9/GR cells, transfection with CDKN1A plasmids decreased gefitinib resistance and microsphere formation capacity, and transfection with si-CDKN1A inhibited the reducing effect of miR-17 downregulation on gefitinib resistance and microsphere formation capacity; in PC9 cells, transfection with si-CDKN1A increased gefitinib resistance and microsphere formation capacity, and transfection with CDKN1A plasmid inhibited the promoting effect of miR-17 upregulation on gefitinib resistance and microsphere formation capacity (Fig. 5E and F). These results indicated that miR-17 influenced gefitinib resistance and self-renewal capacity by directly regulating CDKN1A in NSCLC cells.

CDKN1A is a cyclin-dependent kinase inhibitor, and flow cytometric assays were used to determine the effect of miR-17 and CDKN1A on the cell cycle. It was found that the G2/M phase increased and the G1 phase decreased in PC9/GR cells compared with PC9 cells (Fig. 5G). In PC9/GR cells, downregulation of miR-17 or transfection with CDKN1A plasmid decreased the G2/M phase and increased the G1 phase, and transfection with si-CDKN1A inhibited the reducing effect of miR-17 downregulation on the increased G2/M phase (Fig. 5G). In PC9 cells, upregulation of miR-17 or transfection with si-CDKN1A increased the G2/M phase and decreased the G1 phase, and transfection with CDKN1A plasmid inhibited the promoting effect of upregulation of let-7 on the decreased G2/M phase (Fig. 5G). Using another NSCLC cell line, HCC827, the effects of miR-17 on the cell cycle were also observed (Fig. 5H). Therefore, these findings indicated that miR-17 regulated self-renewal by targeting CDKN1A.