This protocol was approved by the Ethics Committee of the Institutional Review Board at Zhejiang University School of Medicine, and written informed consent was collected before surgery. In total, 84 pairs of primary NSCLC (including adenocarcinoma and squamous cell carcinoma) and adjacent non-tumour tissues were surgically obtained from patients who underwent surgery at the hospital from 2010 to 2012. None of the patients received radiotherapy or chemotherapy prior to surgery. The adjacent non-cancerous tissues were obtained at least 3 cm away from the tumour margin. All of the samples were snap-frozen in liquid nitrogen and stored at −80°C.

NSCLC cell lines (H520, A549, H1299 and H1975) and the lung/bronchial normal epithelial cell line BEAS-2B were obtained from the American Type Culture Collection (ATCC; Manassas, VA, USA), and maintained in RPMI-1640 medium (Gibco, Karlsruhe, Germany) supplemented with 10% fetal bovine serum (FBS) (Invitrogen, Carlsbad, CA, USA) in a humidified atmosphere of 5% CO₂ at 37°C.

Total RNA was extracted using TRIzol reagent (Invitrogen) and NanoDrop 2000 (Thermo Fisher Scientific, Inc., Waltham, MA, USA) was used to assess the quality of RNA. Complementary DNA (cDNA) was synthesized using High Capacity cDNA Reverse Transcription kit (ABI, Carlsbad, CA, USA). qRT-PCR was carried out using 10 ng of cDNA with SYBR-Green Master Mix (ABI) on an LightCycler 480 (Roche, Basel, Switzerland). Primers were as follows: LARP1 forward, 5′-CATTAGCTAAGCACATGAAGTG-3′ and reverse, 5′-GTACGTACGAATCTTGCAA-3′. For the miRNA assay, the cDNA was synthesized using the One Step PrimeScript miRNA cDNA Synthesis kit (Qiagen, Hilden, Germany), and miRNA was quantified using TaqMan^® MicroRNA Assays under the ABI 7900 system (Thermo Fisher Scientific, Inc.). All samples were assessed in triplicates. β-actin and U6 were used as endogenous controls. Data were analysed using the 2^−ΔΔCt method.

To knock down LARP1, we used a lentivirus infection strategy. Lentivirus containing shRNA targeting the LARP1 gene or non-silencing control shRNA were constructed by GeneChem (Shanghai, China). Cells were plated at a density of 30% and were in good condition before the day of infection. The aforementioned lentivirus was used to treat the H520 and A549 cells according to the manufacturer's protocol.

LARP1 cDNA (XM_005268404) without its 3-untranslated region (3′-UTR) was amplified by PCR using the primers containing the MluI and KpnI restriction sites and was then inserted into the pcDNA 3.1(+) vector (Invitrogen) to generate the recombinant vector pcLARP1. The primers used were: 5′-GGATTTCCAAGACGCGTCCCATACCTAGCTGCCC-3′ (forward) and 5′-CGAGTGGTACCAGAGTGATGGGGCTGGTG-3′ (reverse). The empty vector served as a negative control. The miR-374a mimics, inhibitor and corresponding negative controls were purchased from GenePharma, and transfections were performed using Lipofectamine 2000 (Invitrogen).

For the knockdown of XIST, shRNA against lncRNA XIST or scrambled oligonucleotides was ligated into the LV-3 (pGLVH1/GFP + Puro) vector (GenePharma). The 293 cells were co-transfected with Lenti-Pac HIV Expression Packaging Mix and the lentiviral vectors (or the control lentivirus) using Lipofectamine 2000. Lentiviral particles in the supernatant were harvested. The H520 cells were transfected with lentivirus or control virus. The cells were treated with puromycin (2 µg/ml) for 14 days, and GFP-positive cells were selected as sh-XIST and sh-ctrl (control).

Cell proliferation was determined using the MTT method. Cells (2×10³) were seeded into 96-well plates and were maintained at 37°C. MTT (100 µl; 0.5 mg/ml; Sigma, St. Louis, MO, USA) was added to each well at the indicated time-points (24, 48, 72 or 96 h). Dimethyl sulfoxide (DMSO) (150 µl) was used to stop the reaction, followed by measurement of the absorbance at 490 nm.

For the wound-healing assay, 1×10⁶ cells were seeded into 6-well plates. A wound was made by scraping the cell monolayer with a 200-µl pipette tip. Cell motility was evaluated by observation at intervals of 0 and 24 h. For the cell invasion assay, cells (1×10⁵) in serum-free medium were placed into the upper chamber of a 24-well Transwell chamber (8-µm pore size; Corning, Cambridge, MA, USA) coated with Matrigel (BD Biosciences, San Jose, CA, USA). The bottom chamber was filled with medium containing 10% FBS. Cells that had invaded through the membrane to the lower surface were fixed, stained and counted.

Both the wild-type (WT) and mutated (Mut) 3′-UTRs of LARP1 mRNA were subcloned into the XhoI and NotI sites of the psiCHECK-2 vector (Promega, Madison, WI, USA). The mutant constructs were generated using a QuikChange Site-Directed Mutagenesis kit (Stratagene, La Jolla, CA, USA). For the luciferase reporter assay, 293 cells (1×10⁵ cells/well) were cultured and co-transfected with 50 nM miR-374a mimics or negative control (Mim-NC), 200 ng of WT or Mut vector, and 2 ng of Renilla luciferase-expressing vector pRL-TK (Promega). Then, 48 h later, the cells were harvested and assayed using the Dual-Luciferase Reporter Assay system (Promega).

The fragment from XIST containing the predicted miR-374a binding site was amplified by PCR and cloned into a pmirGLO vector (Promega) to form the reporter vector XIST wild-type (XIST-WT). The mutant was generated by mutating the miR-374a seed region binding sequence, which was named XIST-Mut. 293T cells were co-transfected with the pmirGLO vector with either wild-type fragments or mutation fragments and miR-374a using Lipofectamine 2000. Luciferase activity was detected as aforementioned.

Total protein was extracted by lysing cells in RIPA buffer (Beyotime Institute of Biotechnology, Haimen, China). Cell protein lysates were separated by 10% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), and then transferred onto a polyvinylidene difluoride (PVDF) membrane (Millipore, Boston, MA, USA). After blocking with 5% fat-free milk, the membranes were incubated with primary antibodies for LARP1 (ab86359; Abcam, Cambridge, MA, USA), p-STAT3 (Y705, ab76315), cyclin D1 (ab134175), E-cadherin (ab40772) and N-cadherin (ab18203) overnight at 4°C. PVDF membranes were washed in TBST and incubated with horseradish peroxidase-conjugated secondary antibodies.

All of the animal protocols were approved by the Institutional Animal Care and Use Committee of Zhejiang University School of Medicine. Male BALB/c nude mice (4–6 weeks old) were purchased from SLAC Laboratory Animal Ltd. (Shanghai, China). The H520 cells (2×10⁶) were subcutaneously implanted into the right flank (control shRNA) and left flank (shLARP1) of nude mice. Five weeks later, tumours were harvested, weighed, excised, fixed and embedded in paraffin and sectioned for histological examination of p-STAT3 and Ki67 expression.

Mouse tumour tissues were fixed and embedded in paraffin, treated for 2 h at 65°C and then deparaffinized. In addition to standard haematoxylin and eosin (H&E) staining, the tumour sections were subjected to IHC staining to detect Ki67 (ab92742) and p-STAT3. The sections were incubated with biotinylated secondary antibody, followed by incubation with diaminobenzidine and counterstaining with haematoxylin.

Graphical depictions and statistical analyses were conducted using GraphPad Prism 5.0 (GraphPad Software Inc., San Diego, CA, USA) and SPSS 16.0 (SPSS, Inc., Chicago, IL, USA). One-way ANOVA was performed for data involving three or more groups, and sets of two groups were analysed using Student's t-test (two-tailed). The Kaplan-Meier method was used for overall survival analysis. A P-value of <0.05 was considered to be statistically significant.

To explore the relationship between LARP1 and the biological behaviour of NSCLC cells, we first detected endogenous LARP1 in NSCLC and BEAS-2B cells. The mRNA levels of LARP1 were higher in NSCLC cells than those in the normal control cells (Fig. 1A). H520 and A549 cells, which expressed relatively high levels of LARP1, were used in the knockdown analysis. Compared with the control cells, shLARP1-transfected H520 cells expressed low levels of LARP1 protein (Fig. 1B). Compared with the negative control group, cell growth was markedly suppressed in the LARP1-silenced group (Fig. 1C). In addition, LARP1 downregulation decreased the migration of H520 and A549 cells compared with that in the control-transfected cells (Fig. 1D). shLARP1-transfected NSCLC cells also exhibited a significant reduction in invasive ability (Fig. 1E).

As the activated STAT3 pathway is known to be an effector of proliferation and metastasis, we used western blotting to explore the relationship between LARP1 and STAT3 signalling. We found that the protein levels of p-STAT3, cyclin D1 (a cell growth marker), and N-cadherin (an EMT marker) were significantly decreased in shLARP1-transfected cells (Fig. 2A), while the protein level of E-cadherin was upregulated. These data indicated that inactivated STAT3 signalling and EMT were involved in shLARP1-induced cell growth and invasion inhibition.

H520 cells infected with shLARP1 were injected into the flanks of nude mice. Mouse tumours derived from the shLARP1-transfected cells were significantly smaller than those derived from the vector control (Fig. 2B and C). The weights of the tumours induced by LARP1 downregulation were significantly lower than those induced by the control (Fig. 2D). To determine whether the reduction of tumour growth was a result of the ablation of LARP1, the xenograft tumours were subjected to qRT-PCR for LARP1 mRNA expression. As shown in Fig. 2E, the expression of LARP1 was significantly reduced in shLARP1-treated mice. The expression levels of Ki67 (a cell proliferation marker) and p-STAT3 were decreased in tumours treated with shLARP1 (Fig. 2F).

We used publicly available databases (TargetScan, microRNA, miRDB) to identify candidate miRNAs that regulated LARP1. miR-374a had a binding site in the 3′-UTR of LARP1 mRNA (Fig. 3A). The luciferase reporter assay revealed that the overexpression of miR-374a significantly suppressed the luciferase activity of the wild-type LARP1 3′-UTR, but failed to affect the mutant 3′-UTR (Fig. 3B). The protein expression of LARP1 was also decreased in miR-374a-overexpressing cells (Fig. 3C).

To ascertain whether miR-374a regulated cell growth and migration through LARP1, we rescued LARP1 expression by introducing the pcDNA3.1-LARP1 plasmid without the 3′-UTR (pcLARP1) or the empty vector (pcDNA3.1) in the presence or absence of ectopic miR-374a expression in H520 cells. The expression of LARP1 was confirmed by western blotting (Fig. 3D). The restoration of LARP1 markedly abrogated the miR-374a-induced proliferation, migration and invasion inhibition in H520 cells (Fig. 3E and F). The protein levels of p-STAT3, cyclin D1 and N-cadherin were upregulated in LARP1-overexpressing cells, while E-cadherin was downregulated (Fig. 3G). miR-374a mimics led to downregulated p-STAT3, cyclin D1 and N-cadherin and enhanced E-cadherin, which was inconsistent with the phenomenon induced by shLARP1. These data suggested that LARP1 was a functional target of miR-374a.

The expression of LARP1 was significantly upregulated in NSCLC (3.567±1.743) tumours compared with that in matched non-tumour tissues (2.210±1.217; P<0.001; Fig. 4A). Upregulated LARP1 was significantly correlated with the advanced tumour-node-metastasis (TNM) stage (P=0.001), lymph node (P=0.017) and distant metastases (P=0.039; Table I). According to the mean level of LARP1 in NSCLC tumours, we separated cases into low or high expression groups. NSCLC patients with low LARP1 expression had better overall survival than those with high LARP1 expression (P=0.022; Fig. 4B). In addition, to determine the appropriate cut-off for a potential biomarker application, we performed a receiver operating characteristic curve (ROC) analysis. The area under the ROC (AUC) was 0.742 (P<0.001). The optimal cut-off provided 56% sensitivity and 82% specificity for the LARP1 overexpression status as a diagnostic marker for NSCLC (Fig. 4C).

Moreover, the expression of miR-374a was significantly decreased in primary tumours compared to that in non-tumour tissues (1.892±0.878 vs. 2.164±1.073; P=0.0022; Fig. 4D). Downregulated miR-374a was related to advanced tumour stage and distant metastasis (P=0.012, and 0.005, respectively; Table I). The mean level of miR-374a in NSCLC tissues was used as a threshold to classify NSCLC cases into low or high expression groups; and low expression of miR-374a indicated poor survival in patients (P=0.024; Fig. 4E). There was an inverse correlation between the expression of miR-374a and LARP1 in tumour tissues (Pearson r=−0.225; P=0.039; Fig. 4F), suggesting that the miR-374a/LARP1 regulatory pathway was related to the progression of NSCLC.

Using the starBase v2.0 database, miR-374a was found to potentially bind to XIST (Fig. 5A). The qRT-PCR assay showed that miR-374a expression was increased in the sh-XIST group (Fig. 5B). We cloned the predicted miR-374a binding site of XIST (XIST-WT) and a mutated binding site (XIST-Mut) into a reporter plasmid. The results showed that co-transfection of miR-374a mimics and XIST-WT decreased the luciferase activity (Fig. 5C). As shown in Fig. 5D, XIST expression was reduced in cells treated with miR-374a mimics, whereas the expression in cells treated with the miR-374a inhibitor was increased. Finally, we determined whether XIST can regulate the expression of LARP1. Knockdown of XIST reduced the mRNA and protein levels that were increased by the miR-374a inhibitor (Fig. 5E and F). Functionally, the knockdown of XIST significantly inhibited proliferation in H520 cells and abolished the miR-374a inhibition-induced increase in cell proliferation (Fig. 5G). Likewise, knockdown of XIST suppressed the migration and invasion ability increased by miR-374a inhibition in H520 cells (Fig. 5H and I). These results indicated that XIST regulates miR-374a to modulate cell viability and motility in NSCLC cells. We also measured the expression of XIST in H1975 cells derived from female and in H520, A549 and H1299 cells derived from male. There was no significant difference between the expression of XIST in female and male cancer cells (data not shown).