E2F8, a direct target of miR-144, promotes papillary thyroid cancer progression via regulating cell cycle

Two TCGA datasets named TCGA_THCA_miRNA_HiSeq-2015-02-24 and TCGA_THCA_exp_HiSeqV2-2015-02-24 were downloaded at the website of the UCSC cancer browser (https://​genome-cancer.​ucsc.​edu/​) [25], containing 59 paired PTC tissues and adjacent normal tissues. All normalized gene expression values can be obtained from “genomicMatrix” files.

A list of 143 genes with highest co-expression correlation (Pearson r value > 0.5) (Additional file 1: Table S1) with E2F8 were submitted to DAVID Bioinformatics Resources 6.7 (http://​david.​abcc.​ncifcrf.​gov/​) [26] for Gene Ontology (GO) enrichment analysis.

In this study, we collected 64 paired cases of PTC and adjacent normal tissue samples from patients who underwent surgical resection at The First Affiliated Hospital of Nanjing Medical University (Nanjing, China) from 2012 to 2015. Informed written consent for scientific use of biological material was obtained from each patient, and this study was approved by the Ethics Committee of Cancer Institute of Jiangsu Province. All patients’ clinicopathological parameters, including age, gender, primary tumor size, lymph node status, TNM stage, tumor location and focus type, were obtained from their medical records.

BCPAP and TPC-1 cells were cultured in RPMI1640 media (KeyGEN, Nanjing, China) supplemented with 10% fetal bovine serum and penicillin/streptomycin, and cultured at 37 °C in a humidified incubator containing 5% CO2. Transfection was performed following the small-interfering RNA (siRNA) sequences transfection protocol for Lipofectamine RNAi MAX (Invitrogen, USA). Nonsense RNAi (nsRNA) was used as a negative control. Transfection efficiency was evaluated by quantitative real-time RT-PCR and western blot. miR-144 mimic, control mimic, control inhibitor, miR-144 inhibitor and siRNAs against E2F8 were synthesized by Genechem. The sequences used were: siRNA-1 for E2F8: 5’-GGCCAAAGACUGUAUACACTT-3’(sense), 5’-GUGUAUACAGUCUUUGGCCTT-3’(antisense); siRNA-2 for E2F8: 5’-GCCCUAUCAAGACCAACAATT-3’(sense), 5’-UUGUUGGUCUUGAUAGGGCTT-3’(antisense). And the following nonsense siRNA was used as negative control (NC): 5’-UUCUCCGAACGUGUCACGUTT-3’(sense), 5’-ACGUGACACGUUCGGAGAATT-3’(antisense). miR-144 mimic: 5’-UACAGUAUAGAUGAUGUACU-3’. The human E2F8-targeting small hairpin RNA sequences were designed based on siRNA-1 and nsRNA. We generated recombinant lentiviral particles and cells were transfected with E2F8 or negative control recombinant lentivirus (shRNA-E2F8 or shRNA-NC, respectively). For overexpressing miR-144, recombinant lentiviruses containing miR-144 precursor or negative control sequences were purchased from Genechem. For overexpressing CCND1 and E2F8, CCND1 cDNA and E2F8 cDNA without 3’-UTR were cloned into a pEGFP-N1 vector (purchased from Genechem) to construct overexpression plasmid, and an empty vector (EV) was used as a negative control.

A wild-type 3’-UTR fragment of E2F8 cDNA was amplified by using PCR and cloned into XbaI and SacI site of pmirGLO dual-luciferase miRNA target expression vector (Promega, Madison, WI, USA) and named as WT-E2F8 3’-UTR. The mutant variant of E2F8 3’-UTR was generated based on WT-E2F8 3’-UTR by mutating six nucleotides that potentially bind to miR-144 and named as Mut-E2F8 3’-UTR. These vectors (WT-E2F8 3’-UTR or Mut-E2F8 3’-UTR were together with miR-144 mimic or miR-NC) were transiently transfected into BCPAP and TPC-1 cells using Lipofectamine 2000 reagent (Invitrogen). Luciferase activity was measured using the Dual-Luciferase Reporter Assay System (Promega, Madison, USA) after transfection at 48 h. Data are presented as the mean value ± SD for triplicate experiments.

Total RNA was extracted from cultured cells using TRIzol reagent (Invitrogen, Carlsbad, CA, USA) according to the manufacturer’s instruction. The qRT-PCR data collection was performed using a QuantStudioTM 6 Flex Real-Time PCR System and the qRT-PCR reaction included an initial denaturation step at 95 °C for 10 min, followed by 40 cycles of 92 °C for 15 s and 60 °C for 1 min. The primers are shown in Additional file 2: Table S2. Each sample was run in triplicate and the relative expression was calculated and normalized using the 2^-ΔΔCt method.

Cells were harvested and treated with lysis buffer on ice (KeyGEN, Nanjing, China), and a BCA kit (KeyGEN, Nanjing, China) was used to quantify protein concentration. Equal amounts of protein were loaded in SDS–PAGE gels. After separation in the gel, the protein was transferred on a PVDF membrane. Membranes were blocked in 2% BSA in TBS-T for 1 h, and then incubated overnight (4 °C) with antibodies against E2F8 (Abcam, ab185727 1:1000), cyclin D1 (CST, 2978 1:1000) or β-actin (Cell Signaling, 8H10D10 1:1000). After being washed in TBS-T, membranes were incubated with goat anti-rabbit HRP-conjugated secondary antibody (1:10,000; Abcam) or goat anti-mouse HRP-conjugated secondary antibody (1:10,000; Abcam) for 2 h at room temperature. The blots were visualized by ECL detection (Thermo Scientific). All experiments were repeated at least three times independently.

A tissue microarray (TMA) containing 58 paired formalin-fixed paraffin-embedded (FFPE) PTC and adjacent normal tissue samples was used. The TMA was purchased from the Shanghai Biochip Co., Ltd., Shanghai, China. All tissues were re-examined by an experienced pathologist after they were transferred from a local hospital and the TNM stage was determined in each patient.

Immunohistochemistry(IHC) for E2F8 protein expression in samples was performed using standard methods. Briefly, tissue sections were deparaffinized and rehydrated through graded alcohol. Endogenous peroxidase activity was blocked by incubation in 3% H₂O₂. Antigen retrieval was carried out with 0.01 M citrate buffer (pH 6.0) and microwave heat induction. E2F8 staining was scored by blinded observers (including a pathologist) according to intensity and percentage of positive cells. The staining intensity was scored according to 4 grades: 0 (No staining), 1 (weak staining), 2 (intermediate staining), or 3 (strong staining). The product (percentage of positive cells and respective intensity scores) was used as the final staining score (a minimum value of 0 and a maximum of 300).

The cell proliferation was monitored using a Cell Counting Kit-8 (KeyGEN, Nanjing, China) or the xCELLigence system. After transfection, cells were plated in 96-well plates at a density of 2000 cells in 100ul per well and the absorbance was measured at 450 nm with an ELx-800 Universal Microplate Reader. Experiments were repeated at least three times with similar data. For the xCELLigence system, exponentially growing cells with corresponding treatment in complete media were seeded in E-plates at a density of 20,000 per well. The plates were then locked into the RTCA DP device in the incubator. The proliferative ability in each well was automatically monitored by the xCELLigence system and expressed as a “cell index” value. The cell growth was recorded in real-time for 90 h.

For colony formation assay, a total of 100 transfected cells were placed in a fresh 6-well plate and maintained in media containing 10% FBS, replacing medium every 3 or 4 days. After 2 weeks, cells were fixed with 4% paraformaldehyde and stained with 0.1% crystal violet. Visible colonies were then counted. For each treatment group, each well was assessed in triplicate.

Flow cytometry analysis was performed to detect cell cycle distribution. Cells were transferred and fixed in centrifuge tubes containing 4.5 mL of 70% ethanol on ice. The cells were kept in ethanol for at least 2 h at 4 °C. Then, the ethanol-suspended cells were centrifuged for 5 min at 300 g. Cell pellets were resuspended in 5 mL of PBS for approximately 30 s and centrifuged at 300 g for 5 min, then resuspended in 1 mL of PI staining solution and kept in the dark at 37 °C for 10 min. Samples were analyzed using a FACSCalibur flow cytometer. The percentage of the cells in G0–G1, S, and G2–M phases were counted and compared. All the samples were assayed in triplicate.

All animal studies were conducted in accordance with NIH animal use guidelines and protocols approved by Nanjing Medical University Animal Care Committee. Twelve male nude mice (ages 4–6 weeks) were purchased from Nanjing Medical University School of Medicine’s accredited animal facility. Briefly, in each group, 1.0 × 10⁶ exponentially growing TPC-1 cells were injected in axilla subcutaneously. Tumor volume was estimated using calipers every week as length × width² × 0.5. Five weeks after injection, mice were sacrificed, tumor weights were measured and tumors were collected for further analysis.

All statistical analyses were performed using SPSS Statistics (version 20.0, Chicago, Ill) and GraphPad Prism 6 software (GraphPad Software, Inc., La Jolla, CA, USA). The results were presented as mean ± S.D. Relative quantification of mRNA expression level was calculated with the 2^-ΔΔCt method. Student’s t-test was used to analyze difference between two groups. Correlation of E2F8 expression with other genes was analyzed using Pearson test. Association of E2F8 expression with clinicopathological parameters was analyzed using Chi-square test. p < 0.05 was considered statistically significant.

Analysis of the TCGA_THCA_exp_HiSeqV2-2015-02-24 dataset showed that the mean expression value of E2F8 was increased in PTC tissues compared to adjacent normal tissues (3.389 ± 0.1505 vs 2.257 ± 0.2352; p < 0.001) (Fig. 1a). Moreover, Pearson correlation test in TCGA PTC dataset showed that E2F8 was positively correlated with Ki-67(r = 0.7901, p < 0.0001, n = 506), a frequently used proliferation marker (Fig. 1b).

The expression profile of E2F8 was then further validated by qRT-PCR in 64 paired PTC patients’ tissues (tumor and adjacent normal tissues). As shown in Fig. 1c, E2F8 mRNA was upregulated in 84.4% (54 out of 64) of PTC patients. To associated E2F8 mRNA expression with clinicopathological features (including age, gender, primary tumor size, lymph node status, TNM stage, tumor location and focus type), patients were divided into two groups: low-E2F8 and high-E2F8 group, according to the median expression level. We found that high E2F8 level was significantly associated with bigger tumor size(p = 0.0025), positive lymph node status(p = 0.0244) and advanced TNM stage(p = 0.0166) (Table 1). As shown in Fig. 1d, Student’s t-test also exhibited similar results. However, no significant association was observed between E2F8 expression with age, gender, tumor location or focus type.

To evaluate the protein expression of E2F8 in PTC tissues from a cohort of 58 papillary thyroid cancer patients, immunohistochemistry(IHC) was performed using specific anti-E2F8 antibody. As results, we observed that E2F8 was predominantly localized in the nuclei of PTC cells. Significantly higher expression of E2F8 was observed in PTC tissues compared with matched adjacent normal tissues (staining score: 202.8 ± 6.182 vs 96.90 ± 6.940; p < 0.001) (Fig. 1e and f). Consistent with mRNA results as mentioned above, the score of E2F8 staining was also significantly increased along with more advanced T stage in PTC tissues (Fig. 1e and g).

To investigate the biological function of E2F8 in vitro, two different effective siRNAs were used to knockdown(KD) E2F8, and the transfection efficiency was measured by western blot (Fig. 2a). As shown in Fig. 2b and c, cell-counting kit-8 (CCK-8) assay revealed that knockdown of E2F8 markedly inhibited proliferation of both BCPAP and TPC-1 cells. Moreover, the siRNA-E2F8 transfected groups had significantly fewer colonies than the matched siRNA-NC group (Fig. 2d and e).

Finally, the effect of E2F8 on cell cycle distribution was evaluated by flow-cytometry analysis. As shown in Fig. 2f and g, the percentage of BCPAP and TPC-1 cells transfected with siRNA-E2F8 was significantly increased in G1 phase compared to siRNA-NC.

To assess the oncogenic role of E2F8 in vivo, we established xenograft tumor models using TPC-1 cells transfected with shRNA-NC and shRNA-E2F8. First, we observed that all the nude mice developed xenograft tumors at the injection sites, and xenograft tumors were harvested 5 weeks after injection (Fig. 3a and b). Second, as shown in Fig. 3c and d, average volume and weight of tumors in the shRNA-E2F8 group were significantly lower than those in the shRNA-NC group. Finally, IHC analysis revealed that tumors derived from shRNA-E2F8 transfected cells showed weaker staining of Ki-67 than those in the shRNA-NC group (Fig. 3e). These data suggested that increased E2F8 might promote tumor growth in vivo.

To explore how E2F8 exerts its oncogenic activity, a list of 143 genes that have highest correlation values with E2F8 were selected from TCGA PTC dataset. By GO enrichment analysis on the 143 genes, as shown in Fig. 4a, the data revealed that most of the genes were enriched in the “cell cycle” pathway. The result indicated that E2F8 might play a pivotal role in the cell cycle.

Considering that knockdown of E2F8 induces G1 phase arrest in vitro, we sought to determine the mRNA levels of CCND1, CCNE1, CCNE2, CDK1, CDK4, CDK6, p21 and p27 which were closely related to G1 phase or G1/S transition in both shRNA-E2F8 or shRNA-NC cells by qRT-PCT. Interestingly, compared with shRNA-NC transfected cells, the result showed that mRNA expression levels of CCND1, CCNE1, CCNE2 and CDK6 were significantly decreased, p21 was significantly increased, while CDK1, CDK4 and p27 were not influenced in shRNA-E2F8 transfected cells (Fig. 4b). Among these genes, considering CCND1 was a key factor to control G1 phase and it was also the most change in mRNA level, we validated its protein level in shRNA-E2F8 cells. By western blot, the data showed that CCND1 expression was markedly decreased in shRNA-E2F8 group compared to shRNA-NC group (Fig. 4c). To examine whether E2F8 regulated the proliferation of TPC-1 cells via altering CCND1, TPC-1 cells with shRNA-E2F8 was transfected with pEGFP-N1-CCND1 plasmid. Transfection efficiency was determined by western blot (Fig. 4e). By the xCELLigence system analysis, the data showed that the proliferation ability was partially recovered in TPC-1 cells with shRNA-E2F8 after enforced overexpression of CCND1(oe-CCND1) (Fig. 4d). Moreover, flow-cytometry analysis showed that oe-CCND1 significantly alleviated shRNA-E2F8-mediated G1-phase arrest of TPC-1 cells (Fig. 4f and g). Our results demonstrated that E2F8 might promote PTC cells proliferation by regulating cell cycle, especially by influencing CCND1.

MiRNAs have been found to play diverse, key biological roles in cancer development and have been widely used for cancer diagnosis, prognosis, and as therapeutic targets. To identify potential miRNAs targeting E2F8, we used a combination of three algorithms, Targetscan, miRDB and miRanda. Among candidate miRNAs, miR-144 was identified by all the three programs as targeting E2F8 with highest predictive score (Fig. 5a). As shown in Fig. 5b, 91–98 of E2F8 3’-UTR was a predicted target of miR-144. Next, we performed luciferase reporter assays to determine whether miR-144 regulates E2F8 expression through binding to the predicted site in 3’-UTR of E2F8 mRNA. As expected, data demonstrated that miR-144 inhibited luciferase activity by around 46% in BCPAP cells and 55% in TPC-1 cells when the reporter plasmid carried the wild-type(WT)-E2F8 3’-UTR, but no significant inhibition was observed at the reporter plasmid carried a mutant(Mut)-E2F8 3’-UTR (Fig. 5c). We then examined the effect of miR-144 on the mRNA and protein expression of E2F8 in BCPAP and TPC-1 cells, respectively. As shown in Fig. 5d and e, results of qRT-PCR and western blot demonstrated that miR-144 inhibited expression of E2F8 compared with negative control group. Our results reveal that miR-144 targets E2F8 by directly binding to the predicted site in 3’-UTR of E2F8 mRNA.

Finally, we investigated the biological role of miR-144 in vitro. CCK-8 assay showed that the proliferation ability of BCPAP and TPC-1 cells transfected with miR-144 was markedly decreased compared with cells transfected with miR-NC (Fig. 6a and b). Moreover, the colony number of cells with miR-144 was significantly reduced compared to negative control (Fig. 6c and d). Furthermore, the effect of miR-144 on cell cycle distribution was evaluated by flow-cytometry analysis. As shown in Fig. 6e and f, miR-144 transfection treatment significantly caused more BCPAP and TPC-1 cells arrested in G1 phase compared with negative control group.

To test whether restoration of E2F8 could reverse miR-144-mediated inhibition of proliferation of PTC cells, we performed the rescue experiments in TPC-1 cells. We found that enforced overexpression of E2F8(oe-E2F8) by transfection of a cDNA that lacked the miR-144-binding site in the 3’-UTR partially abrogated miR-144-mediated suppression of proliferation in TPC-1 cells (Fig. 6g). As shown in Fig. 6h and i, flow-cytometry analysis also showed that oe-E2F8 partially alleviated miR-144-mediated G1-phase arrest. Expression of E2F8 and CCND1 was verified by western blot in each group (Fig. 6j).

To confirm the growth-inhibitory effect of miR-144 on PTC cells in vivo, the subcutaneous growth of tumors derived from TPC-1 cells transfected with miR-144 or miR-NC was assessed. All the nude mice developed xenograft tumors at the injection sites, and xenograft tumors were harvested 5 weeks after injection (Fig. 7a and b). In all mice, the proliferation of tumor cells stably overexpressing miR-144 was dramatically suppressed compared to that with miR-NC. Tumor volumes and weights were significantly reduced in miR-144 transfected tumors compared with NC (Fig. 7c and d). IHC analysis showed that levels of E2F8 and CCND1 protein were greatly decreased in miR-144 transfected tumors (Fig. 7e).

To study the expression pattern of miR-144 in PTC, we first referred to the TCGA dataset named TCGA_THCA_miRNA_HiSeq-2015-02-24. In TCGA dataset, miR-144 was found to be significantly downregulated in PTC tissues compared with paired adjacent normal tissues (Fig. 8a). Then, 64 collected paired tissues were used to investigate the association between miR-144 expression and T stage. As shown in Fig. 8b, miR-144 expression was found to be significantly decreased along with more advanced T stage in PTC tissues. Moreover, levels of miR-144 and levels of E2F8 mRNA in PTC tissues exhibited a significant inverse correlation calculated by Pearson correlation test (Fig. 8c).