LncRNA XIST/miR-34a axis modulates the cell proliferation and tumor growth of thyroid cancer through MET-PI3K-AKT signaling

A total of 77 paired thyroid cancer and non-tumor tissue samples were collected from patients who underwent surgical resection Xiangya Hospital, Central South University (Changsha, China). All the tissue samples were snap-frozen and stored at − 80 °C.

A human follicular thyroid carcinoma cell line, FTC133 derived from a lymph node metastasis of a follicular thyroid carcinoma, a human thyroid cancer papillary cell line, BCPAP, and a normal human thyroid epithelium cell line (TEC) were purchased from the ScienCell Research Laboratories (Carlsbad, CA, USA). A human papilloma thyroid carcinoma cell line, TPC-1 cell line was purchased from the American Type Culture Collection (ATCC, Manassas, VA, USA). Human thyroid cell lines SW1736 and KAT18 were purchased from the Tumor Cell Bank of the Chinese Academy of Medical Science (Shanghai, China). The above cells were cultured in RPMI-1640 medium (Sigma, St. Louis, MO, USA) supplemented with 10% fetal bovine serum (Life Technologies Corp., USA), streptomycin (100 U/ml) and penicillin sodium (100 U/ml) at 37 °C in a humidified atmosphere with 5% CO₂.

XIST expression was achieved by transfection of si-XIST#1 or si-XIST#2 (GeneCopoecia, Guangzhou, China). The overexpression of miRNAs was achieved by transfection of specific miRNA mimics (Genepharma, Shanghai, China) with the help of Lipofectamine 2000 (Invitrogen). MiR-34a inhibition was achieved by transfection of miR-34a inhibitor (Genepharma, Shanghai, China).

Total RNA was extracted using Trizol reagent (Invitrogen) following the protocol. By using miRNA-specific primer, total RNA was reverse transcribed and the miScript Reverse Transcription kit (Qiagen, Germany) was used for miRNA qRT-PCR. The RNU6B expression was used as an endogenous control. The SYBR green PCR Master Mix (Qiagen) was used for mRNA expression detection following the protocol. The GAPDH expression was used as an endogenous control. 2^−ΔΔCT method was used to analyze the relative fold changes. All sequences used are provided in Additional file 1: Table S2.

A modified MTT assay was used to evaluate cell viability. 24 h after seeding into 96-well plates (5 × 10³ cells/well), cells were transfected and/or treated as described. 48 h after transfection, 20 μl MTT (at a concentration of 5 mg/ml; Sigma-Aldrich) was added, and the cells were incubated for an additional 4 h in a humidified incubator. 200 μl DMSO was added after the supernatant discarded to dissolve the formazan. OD_{490 nm} value was measured. The viability of the non-treated cells (control) was defined as 100%, and the viability of cells from all other groups was calculated separately from that of the control group.

DNA synthesis in proliferating cells was determined by measuring 5-Bromo-2-deoxyUridine (BrdU) incorporation. BrdU assays were performed at 24 h and 48 h after infection. After seeding the infected cells in 96-well culture plates at a density of 2 × 10³ cells/well, they were cultured for 24 h or 48 h, and incubated with a final concentration of 10 μM BrdU (BD Pharmingen, San Diego, CA, USA) for 2 h to 24 h. When the incubation period ended, we removed the medium, fixed the cells for 30 min at RT, incubated them with peroxidase-coupled anti-BrdU-antibody (Sigma-Aldrich) for 60 min at RT, washed them three times with PBS, incubated the cells with peroxidase substrate (tetramethylbenzidine) for 30 min, and measured the absorbance values at 450 nm. Background BrdU immunofluorescence was determined in cells not exposed to BrdU but stained with the BrdU antibody.

A total of 30 eight-week-old athymic female nude mice with an average weight of 20 g were obtained from The Jackson Laboratory (Bar Harbor, ME USA), and anesthetized with an intraperitoneal injection of 2% 2,2,2-Tribromoethanol (200 μl/mouse; Sigma) before implantation of KAT18 or TPC-1 thyroid cancer cells (si-NC, si-XIST#1 or si-XIST#2 transfected). Flank tumors were established by injecting 1 × 10⁶ cells in 100 μL of ECM gel (Sigma) into the subcutaneous flanks of nude mice. Tumor dimensions were serially measured with electronic calipers, and the volumes were calculated by the following formula: a (the largest diameter) × b² (the perpendicular diameter) × 0.4. The body weight of each animal was followed as a marker of toxicity. The humane endpoints were a tumor diameter ≥ 2.0 cm, significant weight loss (20% of pre-experiment body weight), weight loss to a final weight of 16 g, very slow breathing rate, shallow or labored breathing pattern, decreased activity, poor response to handling, absence of grooming, social isolation, hunched posture, shivering, and muscle atrophy. The method of euthanasia was CO₂ exposure for 10 min, at a 20% fill rate of cage volume/min. The animal protocol for this experiment has been maintained in accordance with the guidelines of the animal care committee of Xiangya Hospital, Central South University. All procedures were conducted under sterile conditions. The mice were allowed access to sterile food and water ad libitum.

The fragment of XIST was amplified by PCR and cloned to the downstream of the Renilla psiCHECK2 vector (Promega, Madison, WI, USA), named wt-XIST. To generate the XIST mutant reporter, the seed region of the XIST was mutated to remove all complementarity to nucleotides 2–7 of miR-34a, named mut-XIST. HEK293 cells (ATCC, USA) were seeded into a 24-well plate. After cultured overnight, HEK293 cells were co-transfected with the indicated vectors and miR-34a mimics or miR-34a inhibitor, respectively. Luciferase assays were performed 48 h after transfection using the Dual Luciferase Reporter Assay System (Promega). Renilla luciferase activity was normalized to Firefly luciferase activity for each transfected well. All sequences used are provided in Additional file 1: Table S2.

RNA immunoprecipitation was performed using Magna RIP RNA-Binding Protein Immunoprecipitation Kit (17–700, Millipore) according to manufacturer’s instructions. RNA for in vitro experiments was transcribed using T7 High YieldRNA Synthesis Kit (E2040S, NEB) according to manufacturer’s instructions. IgG, XIST and miR-34a levels in the immunoprecipitates were measured by qRT-PCR.

The protein levels of MET, p-PI3K, PI3K, p-AKT and AKT in thyroid cancer cells were detected by performing immunoblotting assays. Cells were lysed by RIPA buffer (Sigma-Aldrich, USA) with Complete Protease Inhibitor Cocktail (Roche, USA). Cell lysates were transferred to 1.5 mL tube and kept at − 20 °C before use. SDS-PAGE was conducted to separate the cellular proteins. Proteins were loaded onto SDS-PAGE minigel, and then transferred onto PVDF membrane. The blots were probed with the following antibodies at 4 °C overnight: anti-MET (ab51067, Abcam, Cambridge, MA, USA), anti-p-PI3K (phospho Y607, ab182651, Abcam), anti-PI3K (ab86714, Abcam), anti-p-AKT (phospho S473, ab81283, Abcam), anti-AKT (ab179463, Abcam) and anti-GAPDH (ab8245, Abcam) and incubated with HRP-conjugated secondary antibody (1:5000). Signals were visualized using ECL Substrates (Millipore, USA). The protein expression was normalized to endogenous GAPDH.

Data are processed using SPSS17.0 statistical software and presented as the mean ± S.D. of results from at least three independent experiments. A Student t test was used for statistical comparison between means where applicable. Differences among more than two groups in the above assays were estimated using one-way ANOVA. *P < 0.05; **P < 0.01.

Briefly, the expression of XIST in thyroid cancer and adjacent non-tumor tissues was first examined. As shown in Fig. 1a, XIST expression was significantly increased in thyroid cancer tissue samples, compared to adjacent non-tumor tissue samples, based on the data from TCGA-THCA. In a total of 77 paired thyroid cancer and non-tumor tissue samples, we detected the expression of XIST using real-time PCR assays. Consistent with the above data, XIST expression was remarkably up-regulated in thyroid tissues compared to that in non-tumor tissues (Fig. 1b-c). These 77 cases were divided into two groups according to XIST expression: a high XIST group (XIST expression above the median value, n = 39) and a low XIST group (XIST expression below the median value, n = 38). The correlation of clinical parameters with XIST expression was analyzed and shown in Table 1. As shown in Fig. 1d, XIST expression was higher in tissue samples derived from patients in advanced stages (III + IV), compared to that in earlier stages (I + II). Moreover, higher XIST expression was correlated with larger tumor size, advanced N stages and TNM stages (Table 1). A COX risk proportional regression model was used to analyze the overall survival time and the clinical parameters of 77 patients. Univariate analysis revealed that both TNM stages and XIST expression could cause differences in overall survival time; furthermore, multivariate analysis revealed that the differences in overall survival time caused by XIST expression were significantly significant (Table 2). The survival time of the patients with thyroid cancer possessing a lower XIST expression was longer (Fig. 1e). To validate the sensitivity and specificity of XIST expression as a biomarker for thyroid cancer, receiver operating characteristic (ROC) curve was employed. The area under the curve (AUC) was 0.7360 (P < 0.0001), suggesting the potential of XIST expression as a novel biomarker for thyroid cancer and might help with diagnosis and monitor therapeutic efficacy.

To further investigate the detailed role and molecular mechanism of XIST in thyroid cancer, five thyroid cancer cell lines, as well as a normal human thyroid epithelium cell line were subjected to real-time PCR assays for XIST expression. Consistent with that in tissue samples, XIST expression was significantly increased in thyroid cancer cell lines, more up-regulated in KAT18 and FTC113 cells, compared to TEC (Fig. 2a). To evaluate the effect of XIST, si-XIST#1 or si-XIST#2 was transfected into KAT18 and FTC113 cell lines to conduct XIST expression, as confirmed by real-time PCR assays (Fig. 2b). Next, MTT and BrdU assays were used to examine the cell growth of transfected KAT18 and FTC113 cells. XIST knockdown significantly suppressed the cell proliferation of KAT18 and FTC113 cells (Fig. 2c-f). Moreover, xenograft tumor models were also constructed in mice with KAT18 and TPC-1 cell lines. We monitored the tumor volume and weight in response to XIST knockdown. As shown in Fig. 2g-l, the volume and weight of tumors derived from both cell lines could be significantly reduced by XIST knockdown. The above indexes were more down-regulated by si-XIST#2, thus si-XIST#2 was chosen as siRNA for XIST for further experiments.

LncRNAs commonly exert their effects through interacting with downstream target miRNAs [27]. To investigate whether XIST modulates the cell proliferation and tumor growth of thyroid cancer through interacting with miRNAs, we searched online database Gene Expression Omnibus (GEO) for dysregulated miRNAs. We selected 14 down-regulated miRNAs (log (FC) < − 0.5 and P < 0.05) based on microarray data (GSE73182) [28] (Fig. 3a). Next, lnc Tar online tool (http://​www.​cuilab.​cn/​lnctar) [26] was used to identify candidate target miRNAs of XIST. Of the above 14 miRNAs, 9 were predicted to be possible downstream targets of XIST (Fig. 3b). Next, the mimics of the above candidate miRNAs were transfected into FTC33 cells to achieve miRNA expression, as confirmed by real-time PCR assays (Fig. 3c). Next, XIST expression in the above transfected cells was detected by real-time PCR assays. XIST expression was obviously down-regulated by the above miRNAs, more suppressed by miR-34a (Fig. 3c). Thus, miR-34a was chosen as further object. Next, miR-34a expression in 77 paired thyroid cancer and adjacent non-tumor tissue samples was examined using real-time PCR assays; consistent with microarray results, miR-34a expression was significantly reduced in thyroid cancer tissue samples (Fig. 3d). Moreover, miR-34a was negatively correlated with XIST, indicating that they might negatively regulate each other.

KAT18 and FTC33 cells were subjected to miR-34a mimics or miR-34a inhibitor transfection to achieve miR-34a expression, as confirmed by real-time PCR assays (Fig. 4a); XIST expression in these cells was examined. As hypothesized, XIST expression was remarkably increased by miR-34a inhibition whereas suppressed by ectopic miR-34a expression (Fig. 4b). Moreover, si-XIST-induced XIST knockdown also significantly increased miR-34a expression (Fig. 4c), indicating that XIST and miR-34a could indeed negatively regulate each other.

To confirm the interaction between XIST and miR-34a, luciferase and RIP assays were performed. Wild-type and mutant-type XIST reporter gene vectors were constructed and named wt-XIST and mut-XIST; mut-XIST vector contained a 6 bp mutation at the predicted miR-34a binding site (Fig. 4d). HEK293 cells were then co-transfected with the above vectors and miR-34a mimics or miR-34a inhibitor; and then the luciferase activity was examined. As shown in Fig. 4e, the luciferase activity of wild-type luciferase reporter gene vector could be dramatically suppressed by miR-34a mimics whereas amplified by miR-34a inhibitor; after mutating the putative miR-34a binding site predicted by online tool, the alternation of luciferase activity caused by miR-34a overexpression or miR-34a inhibition was eliminated. Furthermore, RIP assay showed that XIST and miR-34a were associated with the AGO2 in HEK293 cells. XIST and miR-34a levels were dramatically higher in the RNA derived from precipitated AGO2 protein than those in IgG (Fig. 4f). We also performed RIP assay in HEK293 cells transfected with control miRNA (miR-NC) or miR-34a followed by real-time PCR to detect XIST associated with AGO2; the results shown in Fig. 5g confirmed the interaction between XIST and miR-34a.

Previously, miR-34a has been regarded as a tumor suppressor through targeting MET [29‐32], a RTK (receptor tyrosine kinase) playing a key role in promoting cell proliferation via transducing extracellular stimuli to intracellular signalling circuits [33, 34], through PI3K/AKT signaling [31]. Here, we investigated whether MET and PI3K/AKT signaling could be modulated by XIST/miR-34a axis. KAT18 and FTC33 cells were co-transfected with miR-34a inhibitor and si-XIST; next, the protein levels of MET, p-PI3K, PI3K, p-AKT and AKT were examined using Immunoblotting assays. As shown in Fig. 5a and b, MET protein levels were significantly decreased by XIST knockdown whereas increased by miR-34a inhibition; the MET protein levels reduced by XIST knockdown could be significantly rescued by miR-34a inhibition (Fig. 5a and b). Furthermore, the ratio of p-PI3K/PI3K and p-AKT/AKT in these two cell lines could be also reduced by XIST knockdown and increased by miR-34a inhibition; the effect of XIST knockdown on PI3K and AKT phosphorylation was partially attenuated by miR-34a inhibition (Fig. 5a and c), indicating that XIST knockdown could decrease MET protein and the phosphorylation of PI3K and AKT, while miR-34a inhibition could partially reverse the effect of XIST knockdown.

In order to further confirm the above findings, we analyzed the expression of MET mRNA based on TCGA-THCA database containing 512 thyroid cancer and 337 non-tumor tissues; as shown in Fig. 6a, MET mRNA expression could be remarkably promoted in thyroid cancer tissue samples, compared to that in non-tumor tissue samples. In 77 paired thyroid cancer and non-tumor tissues, MET mRNA expression was also remarkably up-regulated, consistent database results (Fig. 6b). In three paired randomly-selected tissue samples, MET protein levels were increased in cancer tissues, compared to those in non-tumor tissue samples (Fig. 6C). Furthermore, the Spearman’s rank correlation analysis revealed that MET mRNA expression was positively correlated with XIST whereas negatively correlated with miR-34a (Fig. 6d-e). These data further indicate that XIST targets miR-34a to modulate the cell proliferation and tumor growth through miR-34a downstream MET and PI3K/AKT signaling.