Background
Hepatocellular carcinoma (HCC) is one of the most common causes of cancer mortality worldwide [
1,
2]. Currently, surgical treatment for HCC is only available in the initial stage of the disease, but most patients present with advanced disease upon diagnosis, at which point the efficacy of radiotherapy and chemotherapy is limited. The high rate of tumor recurrence and metastasis are major factors that contribute to the poor prognosis of patients with HCC. Therefore, novel insights into the mechanism of HCC are urgently needed to identify novel prognostic molecular markers and potential effective therapeutic targets to improve patient survival [
3,
4].
Long noncoding RNAs (lncRNAs) are RNA transcripts that are longer than 200 nt and exhibit limited or no protein-coding capacity, and many lncRNAs are uniquely expressed in differentiated tissues or specific cancer types [
5‐
8]. Recent discoveries indicate that lncRNAs drive many important cancer phenotypes by interacting with other cellular macromolecules, including DNA, RNA, and protein [
9‐
12]. Advanced studies have found that aberrant lncRNA expression plays critical roles in hepatocarcinogenesis and metastasis [
13‐
15]. Moreover, MALAT1 is associated with tumor metastasis and can predict recurrence after liver transplantation [
16,
17]. Furthermore, UFC1 promotes HCC cell proliferation by inhibiting cell apoptosis and inducing cell cycle progression [
18]. Long noncoding RNA low expression in tumor (lncRNA-LET) reduces hepatic invasion and abdominal metastases via the degradation of NF90 [
19], and MEG3 is a predictive biomarker for monitoring epigenetic therapy [
20]. In addition, ultraconserved noncoding RNAs (ucRNAs) are noncoding RNAs transcribed from regions that are highly conserved across many species, so-called ultraconserved regions [
21]. Evf-2, a ucRNA in the intergenic region between the Dlx-5 and Dlx-6 genes, activates transcriptional activity by directly interacting with Dlx-2 in a target- and homeodomain-specific manner [
22]. However, the molecular mechanisms of lncRNAs remain poorly understood and warrant further study.
A growing number of studies have focused on the widespread Hippo kinase signaling inactivation and nuclear localization of YAP in epithelial malignancies [
23‐
25]. Specifically, a gene expression analysis of HCC identified YAP as an important “driver oncogene” [
26], and the Hippo pathway is a highly conserved protein kinase chain that plays a pivotal role in restricting tumor cell proliferation and promoting apoptosis. By phosphorylating YAP, the serine/threonine-protein kinase LATS1 inhibits the translocation of YAP into the nucleus and decreases the expression of its downstream target genes, which are important for cell proliferation and migration [
27]. However, the mechanism by which lncRNAs regulate Hippo kinase signaling in HCC remains largely unclear.
In this study, we used a lncRNA microarray to identify that the expression of the ultraconserved lncRNA uc.134 was significantly decreased in the highly aggressive HCC cell line HCCLM3 compared with MHCC97L cells. We then first confirmed the full-length transcript of uc.134 using 5′- and 3′-rapid amplification of cDNA ends (RACE) analysis (GenBank accession no. KY355383). The high conservation of uc.134 implicates that its aberrant expression may play a critical role in HCC progression. Moreover, in situ hybridization (ISH) and quantitative real-time polymerase chain reaction (qRT-PCR) results showed that uc.134 expression significantly decreased in HCC and uc.134 expression directly correlated with patient survival. Furthermore, we demonstrated that the overexpression of uc.134 suppresses HCC cell proliferation and invasion by inhibiting CUL4A to ubiquitinate LATS1 and increasing pYAPS127 expression. We also confirmed that uc.134 inhibits the expression of YAP downstream target genes. Finally, a positive relationship among uc.134, LATS1, and pYAPS127 was also confirmed in 90 paraffin-embedded clinical samples. In conclusion, our study reveals that the novel lncRNA uc.134 represses HCC progression by inhibiting the CUL4A-mediated ubiquitination of LATS1 and increasing pYAPS127 expression. Thus, this lncRNA may offer a promising treatment approach by inhibiting YAP and activating Hippo kinase signaling.
Methods
Microarray assay
Total RNA was isolated from cells using TRIzol Reagent (Takara, Dalian, China). Sample labeling and array hybridization were performed according to the Arraystar microarray-based gene expression analysis protocol (Arraystar, Rockville, MD). The Arraystar human lncRNA microarray is designed for the global profiling of human lncRNAs and protein-coding mRNA transcripts. The array detects a total of 40,173 lncRNAs. The lncRNAs were carefully constructed using the most highly respected public transcriptome databases (RefSeq, UCSC Known Genes, Ensembl, etc.), as well as landmark publications. The Arraystar mRNA microarray provided a global view of all known genes and transcripts in the human genome. A total of 27,958 Entrez Gene RNAs were detected by this microarray. The content was sourced from RefSeq, Ensembl, UniGene Build, and GenBank. Quantile normalization and subsequent data processing were performed using the GeneSpring GX v12.1 software package (Agilent Technologies). After quantile normalization of the raw data, the lncRNAs and mRNAs that were detected in at least three out of six samples were chosen for further data analysis. Differentially expressed lncRNAs and mRNAs with statistical significance between the two groups were identified through
P value/FDR filtering. KEGG pathway analysis and gene ontology (GO) analysis were applied to determine the roles of these differentially expressed mRNAs in the corresponding biological pathways or GO terms. The microarray data were uploaded in the Additional files
1,
2, and
3.
Tissue samples, ISH, immunohistochemical staining (IHC), and fluorescence in situ hybridization (FISH)
From January 2009 to December 2013, 170 human HCC samples were collected at Nanfang Hospital, Southern Medical University (Guangzhou, China). None of these patients had been pretreated with chemotherapy or radiotherapy before undergoing surgery. The study was approved by the Nanfang Hospital Institutional Ethical Review Board, and informed consent was obtained from each patient.
LncRNA uc.134 expression was measured in paraffin-embedded samples using an ISH optimization kit (Roche, Basel, Switzerland) according to the manufacturer’s instructions. The locked nucleic acid (LNA)-modified oligonucleotide probe targeting uc.134 was designed and synthesized at Exiqon (Vedbaek, Denmark). Briefly, HCC samples were treated with pepsin for 10 min at room temperature and incubated with 500 nM of probe at 55 °C for 4 h. The samples were incubated with blocking solution for 30 min, anti-digoxigenin (anti-DIG) reagent was applied for 60 min and the samples were incubated with AP substrate 4-nitro-blue tetrazolium and 5-bromo-4-chloro-3′-indolylphosphate (NBT-BCIP) for 2 h at 30 °C. The samples were then mounted with Nuclear Fast Red™ (BOSTER, Wuhan, China), and a blue stain in the nucleus indicated a positive signal by NBT-BCIP. IHC was performed as we previously described [
28]. For FISH, the signals representing the expression of LNA probes were determined using the tyramide signal amplification (PerkinElmer, USA) system. In brief, the signal was detected by incubation with horseradish peroxidase (HRP)-conjugated anti-DIG antibodies. Then, the signals were amplified using tetramethylrhodamine (TRITC)-conjugated tyramide. The images were acquired with a fluorescence microscope (IX70, Olympus, Japan).
The ISH and IHC results were evaluated by two individuals in a blinded fashion; the evaluators scored the samples using a quick scoring system from 0 to 12 by combining the intensity and percentage of the positive signal (signal: “0,” no staining; “1,” weak staining; “2,” intermediate staining; and “3,” strong staining; percentage: “0,” 0%; “1,” 1–25%; “2,” 26–50%; “3,” 51–75%; and “4,” >75%), and this was in good agreement with the initial quantification. An optimal cutoff value was identified. If the evaluated uc.134 score was higher than the average score, the uc.134 expression in those HCC samples was classified as high; otherwise, it was classified as low. To account for inconsistencies in the percentage of the ISH signals, an ImageJ software (National Institutes of Health, Bethesda, MD) was used for scoring signals. The data were statistically analyzed using t test to determine the differences in uc.134 expression levels between different groups of tissues. P < 0.05 was considered significant.
In vivo model
All animal studies were performed with the approval from the Institutional Animal Care and Use Committee of Nanfang Hospital. Male BALB/c nude mice (age, 4–6 weeks; Guangdong Medical Laboratory Animal Center, China) were raised under specific pathogen-free conditions. All in vivo experiments were performed according to our institution’s guidelines for the use of laboratory animals. For the subcutaneously injected tumor model, 3 × 106 cells were subcutaneously injected into the left flanks or right flanks of mice. After 4 weeks, the tumors were embedded in paraffin for ISH or IHC. For the lung metastasis model, 2 × 106 cells were injected into the tail veins. We monitored lung metastasis at 6 weeks to quantify lung colonization by histology examination.
RACE analyses
The 5′ and 3′-RACE experiments were performed using the SMARTer® RACE 5′/3′ Kit (Clontech, Mountain View, CA) according to the manufacturer’s instructions. Briefly, at least two sets of primers were designed and synthesized for the nested PCR. The RACE PCR products were separated on a 1.5% agarose gel. The results of electrophoresis were confirmed, and the amplified bands were sequenced bi-directionally using the indicated primers. The gene-specific RACE primers used for mapping each end are listed in Additional file
4.
RNA isolation and qRT-PCR analysis
Cytoplasmic and nuclear RNA fractionation was performed using the PARIS™ Kit (Life Technologies, Carlsbad, California) according to the manufacturer’s instructions. The yield and quality of the RNA samples were evaluated prior to qRT-PCR. For qRT-PCR analysis, total RNA was isolated from the cells using TRIzol Reagent (Takara) following the manufacturer’s protocol. First-strand cDNA synthesis from 1 μg of total RNA was performed using a reverse transcriptase cDNA synthesis kit (Takara). The resulting cDNA was then analyzed by qRT-PCR using a SYBR Green PCR Kit (Takara) and a 7500 Fast real-time PCR system (AB Applied Biosystems). In brief, the reaction mixture containing 500 ng cDNA, the forward primer and the reverse primer, was used to amplify the PCR product corresponding to the human gene. The experiments were repeated at least three times independently to ensure the reproducibility of the results. Human GAPDH gene was amplified as an internal control. β-actin and U6 were used as cytoplasmic and nuclear controls, respectively. Comparative quantification was done by using the 2
−ΔΔCt method. The primer sequences are listed in Additional file
4.
RNA immunoprecipitation (RIP)
A RIP assay was performed using the Magna RIP RNA-Binding Protein Immunoprecipitation Kit (Millipore, MA, USA) according to the manufacturer’s instructions. Briefly, whole-cell extracts prepared in lysis buffer containing a protease inhibitor cocktail and RNase inhibitor were incubated on ice for 5 min, followed by centrifugation at 10,000g and 4 °C for 10 min. Magnetic beads were preincubated with 5 ug of IP-grade antibody for 30 min at room temperature with rotation. The supernatant was added to bead-antibody complexes in immunoprecipitation buffer and incubated at 4 °C overnight. Finally, the RNA was purified and quantified by qRT-PCR. Input controls and normal rabbit IgG controls were assayed simultaneously to ensure that the signals were detected from RNA that was specifically bound to protein.
RNA pulldown assay
Biotin-labeled RNA uc.134 was transcribed in vitro with the Biotin RNA Labeling Mix (Roche) and T7 RNA polymerase (Roche) and then treated with RNase-free DNase I (Roche) and 0.2 M EDTA to stop the reaction. Biotinylated RNAs were mixed with streptavidin agarose beads (Life Technologies, Gaithersburg, MD) at 4 °C overnight. Total cell lysates and RNase inhibitor were added to each binding reaction and incubated on ice for 1 h. The RNA–protein binding mixture was boiled in SDS buffer, and the eluted proteins were detected by Western blotting or mass spectrometry. The full-length transcript of uc.134 is 1867 bp in length; Δ1, Δ2, and Δ3 correspond to the 1–718 bp, 719–1407 bp, and 1408–1867 bp sequence fragments of uc.134 until the end of the uc.134 sequence. CUL4A was cloned into the eukaryotic expression vector pcDNA3.1(+) with a C-terminal Myc tag and translated a 87.7-kilodalton (kDa) protein. CUL4A lacking the 55–401 amino acid (aa) region was cloned into pcDNA3.1(+) to afford the pcDNA3.1(+)-Cul4a-△1-myc construct, which translated a 46.94-kDa protein; CUL4A lacking the 400–671 aa region encoding a cullin homolog was cloned into pcDNA3.1(+) to afford the pcDNA3.1(+)-Cul4a-△2-myc construct, which translated a 56.53-kDa protein; CUL4A lacking the 688~753 aa region, which encodes a neddylation domain, was cloned into pcDNA3.1(+) to afford the pcDNA3.1(+)-Cul4a-△3-myc construct, which translated a 79.81-kDa protein; CUL4A lacking the 592–759 aa region, which encodes the winged helix-turn-helix DNA-binding domain (WHDD), was cloned into pcDNA3.1(+) to afford the pcDNA3.1(+)-Cul4a-△4-myc vector, which translated a 68.03-kDa protein.
Cycloheximide (CHX) chase measurements of LATS1 half-life
The CUL4A and uc.134 plasmids were transiently transfected into HCC cells using jetPRIME (Polyplus, Strasbourg, France). After 24 h, CHX (10 ug/ml) was added to the DMEM culture medium, and incubation was continued for 0, 3, 6, or 9 h. The cell lysates were submitted to Western blotting using rabbit anti-LATS1 monoclonal antibody (Cell Signaling Technology, Beverly, MA), and Western blot data were quantified using the ImageJ software.
Immunoprecipitation (IP) and ubiquitination assay
Ubiquitin, LATS1, and uc.134 plasmids were transfected into cells using jetPRIME (Polyplus). Thirty-six hours after transfection, 10 nM MG132 was added to the DMEM culture medium and incubation was continued for 8 h. The lysates were immunoprecipitated with the indicated antibodies on protein A/G beads (Life Technologies) overnight at 4 °C with rotation and then boiled in SDS buffer. The eluted proteins were detected by Western blotting.
Statistical analysis
All statistical analyses were performed using the SPSS software (Chicago, IL, USA). Survival curves were plotted based on the Kaplan–Meier and log-rank tests. Pearson’s chi-square test was used to analyze the relationship between uc.134 expression and the clinicopathologic features of HCC. Student’s t test was used to detect significance differences in data obtained from qRT-PCR experiments and colony formation assays. A multi-way classification analysis of variance test was performed to assess data obtained from the CCK8 assays and tumor growth. Correlations among uc.134 expression, LATS1, and pYAPS127 were analyzed with a Spearman rank correlation. P < 0.05 was considered to indicate a significant difference.
Discussion
HCC is a leading cause of cancer-related death worldwide, and tumor recurrence and metastasis are major factors that contribute to the poor prognosis of patients with HCC. Moreover, recent research has greatly advanced our understanding of the essential role of lncRNAs in HCC [
10,
12]. Although thousands of lncRNAs have been functionally characterized, the vast majority of members of this RNA class have not been thoroughly described. UcRNAs are noncoding RNAs transcribed from regions that are highly conserved across humans, mice, and rats, so-called ultraconserved regions (UCRs). UCRs are usually located at genomic regions that are involved in cancer and differentially expressed in carcinomas. Here, we used lncRNA microarrays to identify that the expression of an ultraconserved lncRNA, uc.134, was significantly decreased in the highly aggressive HCC cell line HCCLM3 compared with MHCC97L cells. We first confirmed the full-length transcript of uc.134 using 5′- and 3′-RACE analyses, and qRT-PCR showed that uc.134 expression was reduced in both tissue samples and HCC cell lines. Thus, the decreased expression of uc.134 in HCC, especially in highly aggressive cell lines, indicates that uc.134 may be a promising marker for HCC. ISH results showed that the expression of uc.134 was significantly downregulated in 170 paraffin-embedded HCC specimens and decreased uc.134 expression was related to poor survival in patients with HCC. Moreover, an analysis of clinical follow-up data and clinicopathologic parameters demonstrated that the expression of uc.134 significantly correlated with lymph node metastasis (
P = 0.005) and TNM classification (
P < 0.001) in patients with HCC.
Cancer is fundamentally a genetic disease with numerous alterations in DNA, RNA, and proteins that support tumor growth and development [
38]. The molecular and cellular characteristics of cancer-associated lncRNAs are under distinct regulatory regimes that are different from physiological conditions [
12]. Although numerous studies have used gene overexpression plasmids or lentiviral vectors to evaluate the functions of tumor-suppressor genes, the cellular toxicities of these gene carriers are inevitable. Therefore, a knockdown model such as RNAi is necessary. In our study, we performed gain-of-function and loss-of-function studies both in vitro and in vivo to demonstrate that uc.134 plays a critical role in the inhibition of cell proliferation, invasion, and metastasis in HCC. Ideally, Tet-Off and Tet-On systems will fulfill the requirements for the quantitative and temporal control of gene expression via nontoxic effector molecules, and these will be used in the future [
39].
A growing number of studies confirm that the Hippo kinase signaling, which is critical for tumor progression and invasion, is inactivated in many cancers. The core components of Hippo kinase signaling pathway are conserved in mammals and have a complex network of crosstalk with other important signaling pathways, such as the TGFβ/SMAD, WNT/β-catenin, PI3-kinase/AKT, Hedgehog, Jak/Stat, and Notch pathways [
40]. A previous study confirmed that YAP increases resistance to RAF- and MEK-targeted cancer therapies [
41]. Moreover, the serine/threonine kinase LATS1 is a core kinase of Hippo kinase signaling pathway and plays important roles in tumor proliferation, apoptosis, and stem cell differentiation. A previous study showed that Merlin activates Hippo kinase signaling by inhibiting CRL4
DCAF1, an E3 ubiquitin ligase of the CRL4 complex [
42]. CUL4A is a core component of the CRL4 complex, and its N-terminus associates with a cullin-specific adaptor protein to recruit a large number of substrate proteins. CUL4A can interact with LATS1 protein and enhance its proteasomal degradation. Another study demonstrated that the de-repression of CRL4
DCAF1 inhibited the activation of the Hippo pathway by directly binding to and ubiquitinating LATS1/2 in NF2-mutant tumors in the nucleus [
37]. However, the mechanism that regulates LATS1 at the lncRNA level remains unknown. We performed RNA pulldown and RIP assays, which showed that uc.134 bound the E3 ligase CUL4A. Protein domain mapping and deletion mutation analyses identified a 592–759 aa region of CUL4A, which encodes a WHDD domain, that binds a 460-nt region at the 3′ end of uc.134 (Δ3-1408–1867). In addition, we demonstrated that uc.134 inhibited the translocation of CUL4A from the nucleus to the cytoplasm, which inhibits the CUL4A-mediated ubiquitination and degradation of LATS1 in the cytoplasm and increases YAP
S127 phosphorylation. LncRNAs regulate many important activities in cancer through their interactions with DNA, RNA, and proteins. The mechanisms through which lncRNAs interact with proteins to regulate protein-protein interactions or modulate the subcellular transport of proteins are largely unknown [
43]. It remains to be investigated whether the uc.134-CUL4A RNP can be retained in the nucleus through binding with other nuclear proteins or by forming an RNA–DNA triplex.
As a co-activator of transcription, YAP binds to the TEAD complex and elevates the levels of numerous target genes to promote tumor proliferation and metastasis. CYR61 is overexpressed in various cancers and has been reported to be involved in tumor growth and vascularization. Previous reports showed that the YAP-dependent expression of CYR61 increased the invasive activity of the glioblastoma cells [
44]. Inactivation of c-Myc results in tumor dormancy and pluripotent differentiation of tumor cells. YAP has been reported to promote the transcriptional activity of c-Myc via interaction with c-Abl in HCC [
45]. diap1 directly binds to and inhibits caspases to suppress apoptosis. Activation of YAP increased the transcription of diap1, cell proliferation, and tissue overgrowth [
46,
47]. A previous study has shown that Yorkie promotes the overgrowth of drosophila neuroepithelial cells and delays their differentiation through the regulation of the cell cycle regulator E2F1, which plays key roles in cell cycle progression and cell differentiation in HCC [
48,
49]. We performed microarray analysis and qRT-PCR indicated that overexpression of uc.134 significantly decreased the expression of YAP target genes. In addition, when we co-transfected the HCC cells with plasmids expressing YAP and those expressing uc.134, the decrease in the mRNA level of CYR61, c-Myc, diap1, and E2F1 was reversed by YAP, indicating that YAP is responsible for the uc.134-mediated gene regulation. Finally, a positive correlation between uc.134, LATS1, and pYAP
S127 was confirmed in 90 paraffin-embedded samples by ISH and IHC. Taken together, our results suggest that uc.134 increased Hippo kinase activity and repressed the downstream target genes of YAP by inhibiting CUL4A-mediated ubiquitination and degradation of LATS1. To the best of our knowledge, this study is the first to investigate the critical role of lncRNAs in the regulation of LATS1 in HCC. Thus, this lncRNA may offer a promising approach for HCC therapy by inhibiting YAP and activating Hippo kinase signaling.
Previous studies have clarified that the major risk factor for HCC in China is hepatitis B virus (HBV) infection, although hepatitis C virus infection and exposure to toxic chemical substances also contribute to the incidence of HCC. The lncRNA MEG3 is regulated by miR-29a in a methylation-dependent, tissue-specific manner, and it contributes to the growth of HCC [
20]. Histone deacetylase 3 (HDAC3) is involved in the suppression of HCC-related lncRNA-LET [
19]. The aberrant expression of lncRNA uc.134 in HCC may be mediated by HBV infection at the transcriptional level, post-transcriptional level, or by epigenetic regulations such as DNA methylation or histone deacetylation [
50]. In our study, we report that the ultraconserved lncRNA uc.134 suppressed the progression of HCC by inhibiting CUL4A-mediated ubiquitination of LATS1. Hippo kinase signaling is highly conserved among diverse species and regulates tissue overgrowth and development [
51]. The classic tumor suppressor miRNA let-7 is also conserved among species and promotes cell cycle exit and cell differentiation both during normal development and in cancer [
52,
53]. However, does uc.134 have a similar developmental expression pattern as the miRNA let-7? It will be interesting to unravel the underlying mechanisms that control the expression of uc.134 in both cancer and normal development in the future.
Conclusions
Taken together, our data indicated that the expression of a novel lncRNA, uc.134, was repressed in HCC specimens and the expression of uc.134 was significantly correlated with the overall survival of patients. Notably, our results showed that the overexpression of uc.134 suppressed HCC cell proliferation, invasion, and metastasis in vitro and in vivo. In addition, we demonstrated that uc.134 plays a crucial role in activating Hippo kinase signaling by inhibiting CUL4A ubiquitination of LATS1 and increasing YAPS127 phosphorylation. Furthermore, we confirmed that uc.134 repressed the downstream target genes of YAP. Finally, ISH and IHC showed positive relationships among uc.134, LAST1, and pYAPS127 in 90 paraffin-embedded samples. Consequently, these results identified that lncRNA uc.134 activates Hippo kinase signaling by blocking CUL4A, suggesting that it may serve as a tumor suppressor and prognostic biomarker in HCC.
Acknowledgements
We are grateful to all laboratory members of the Department of Pathology of Southern Medical University for their support and comments.