The emerging roles of long non-coding RNA in gallbladder cancer tumorigenesis

1. Introduction

Gallbladder cancer (GBC) is the most frequent neoplasm of the biliary tract and represents almost 50% of all biliary tract cancer [1]. GBC has a great worldwide incidence variability in correlation with both geographic and ethnic features, which is relatively rare in most western countries, but is much widespread in some other regions of the world, including Chile, Indian subcontinent, Japan, and Korea [2]. The lack of symptoms at initial stage leads to advanced stage at diagnosis. The aggressive tumor biology, late presentation and complicated anatomic position make it difficult in treatment [3]. Taken into consideration that the lack of a serosal layer of gallbladder adjacent to the liver, the precipitated hepatic invasion and metastatic progression also causes its miserable prognosis [2, 4].

Non-coding RNAs (ncRNAs) refer to RNA molecu- les that do not encode information about proteins. Long non-coding RNAs (lncRNAs) represent one of the prevalent and functionally diverse class of ncRNAs, which are defined as transcripts longer than 200 nucleotides [5]. LncRNAs have been observed to perform various functions in a wide variety of important biological processes [6]. They work through modular domains that interact with other RNAs or DNA by nucleic-acid base pairing or with proteins through higher-order RNA structures [7]. In many cases, the lncRNA itself can exert gene-regulatory effects through changes in chromatin accessibility by binding to histone-modifying complexes, to DNA binding proteins (including transcription factors), and even to RNA polymerase II [8]. In other cases, the specific lncRNA transcripts themselves instead of the lncRNA products appear to be regulatory by tran- scription-factor trapping at target-gene promoters [9].

The role of lncRNAs in GBC is an emerging field of study. In this review, we summarize what is currently known about lncRNAs in GBC and focus on their key biological functions.

2. Alteration of lncRNAs in GBC

Previous studies have shown that lncRNAs appear to play vital roles in cancer biology and dysregulations of lncRNAs are associated with many types of human cancers [10, 11, 12, 13]. To date, numerous lncRNAs have been reported to be implicated in GBC. For example, the expression levels of lncRNA AFAP1-AS1 were significantly elevated in GBC, which was significantly associated with tumor sizes and correlated with poor prognosis in GBC patients [14]. A recent study group analyzed the expression profiles of lncRNAs between GBC tumor tissue and para-carcinoma tissue using microarray data. They assessed the fold-changes (tumor tissue vs. para-carcinoma tissue), P-values and FDRs to normalize the expression of genes and validated these data using RT-qPCR. Total 1,758 lncRNAs were identified differentially expressed in GBC, including a series of novel lncRNAs correlated with GBC like RP11-152P17.2-006 and CTA-941F9.9, which demonstrated that lncRNA expression in GBC was markedly altered [15]. The emerging roles of lncRNAs in GBC tumorigenesis are becoming more and more attractive [16] and their prognostic and therapeutic potentials provide a future significance in developing effective molecular biomarker(s) for gallbladder [17].

Different lncRNAs have been investigated to exert different roles in GBC, where they can act as a tumour suppressor or accelerator. For example, LINC00152 was identified as an oncogene involved in GBC and the inhibition of LINC00152 can suppress proliferation, migration and invasion of the GBC cells [18, 19]. Another lncRNA-LET, was demonstrated to be down-regulated and exhibits tumor-suppressive activity in GBC [20]. Meanwhile, patients with low expression of lncRNA-LET have significantly poorer prognosis than those with high expression, indicating that lncRNA-LET may represent a positive prognostic marker [20]. Knowing the positive or negative roles of lncRNAs, the delivery of oligonucleotides inhibitors of lncRNAs, for example, oligonucleotide inhibitors targeted to a natural antisense transcript (antagoNAT), can be a good attempt for gene therapy [21]. Table 1 listed the dysregulated lncRNAs in GBC and their effects may be owing to various interactions of lncRNAs with DNA, RNA and proteins [22], which will be illustrated in the following sections.

3. LncRNAs regulate gene expression epigenetically in GBC

Epigenetic regulation has been widely accepted as stably inherited modulations in the expression of genes without changes in their DNA sequence [23]. Increasing emphasis is placed on the ability of lncRNAs to regulate gene expression as epigenetic modifiers, either depending on the function of themselves or their interactions with DNA and/or proteins [7, 24]. To date, a large number of lncRNAs have been shown to function on gene expression to activate or silence transcription, either by targeting chromatin-modifying activities to particular genomic sites, or by recruit histone-modifying enzymes [25].

With respect to their locations in the genome relative to protein coding genes, lncRNAs are subdivided into intergenic and intragenic lncRNAs. Intragenic lncRNAs can be further classified as exonic, intronic and overlapping lncRNAs [26]. Their influence on gene expression is by either cis-acting or trans-acting manner, involving transcriptional interference, initiation of chromatin remodeling, epigenetic silencing of gene clusters, and epigenetic repression of genes [27, 28, 29]. For example, urothelial carcinoma associated 1 (UCA1), was significantly overexpressed in GBC and positively correlated with tumor size, lymph node metastasis, TNM stage and short survival time [30]. Researchers revealed that UCA1 promoted GBC progression by a “sponge” role to physically bind with enhancer of EZH2 and further recruited EZH2 to the promoter of p21 and E-cadherin, with subsequently suppressed their expressions at transcriptional level in nucleus [30].

4. LncRNAs regulate GBC progression by interaction with miRNA

MicroRNAs (miRNAs), another class of ncRNAs, are endogenous ∼ 22 nt RNAs. MiRNAs negatively regulate mRNA translation and stability originate from long primary transcripts with local hairpin st- ructures [31]. Increasing literature has revealed the associations of lncRNAs with microRNAs. The interplay between lncRNAs and microRNAs has been the most convincing evidence linking lncRNAs and tumorigenesis [12]. In GBC, the lncRNAs-miRNA crosstalk represented most of the lncRNAs regulation mechanism [32, 33, 34, 35, 36, 37]. LincRNA can function as a key competing endogenous RNA (ceRNA) by sharing miRNA

response elements with the core transcription factors [38]. Several examples of lincRNA-miRNA interactions in GBC will be presented as follows.

Taurine Upregulated Gene 1 (TUG1), a 7598-bp lncRNA located at chromosome 22q12, was first identified as a transcript up-regulated by taurine in murine retina cells [39]. TUG1 has been found to play important roles in tumor proliferation and metastasis, and correlated with prognosis and metastasis in various human cancers [40]. A recent literature showed that the expression of TUG1 was significantly increased in GBC tissues and knockdown of TUG1 obviously inhibited GBC cell proliferation and metastasis [34]. The study further revealed that TUG1 is upregulated by TGF-b1 and can negatively regulate miR-300. However, the detailed functional mechanism is still unclear.

Metastasis-associated lung adenocarcinoma transcript 1 (MALAT1), a length of more than 8000 nt expressed from chromosome 11q13, was initially discovered as a prognostic marker for cancer metastasis in early-stage non-small cell lung cancer [41]. Increasing evidence showed that MALAT1 have important roles in human cancers, including osteosarcoma [42], oral squamous cell carcinoma [43], triple-negative breast cancer [44], gastric cancer [45] and GBC [35, 46]. The expression of MALAT1 was up-regulated in GBC tissues and cells and the high MALAT1 levels were correlated positively with tumor size and lymphatic metastasis, while negatively with overall survival [35]. Furthermore, MALAT1 was identified to sponge miR-206 [35] and miR-363-3p [46] by functioning as a ceRNA in GBC.

LncRNA H19 is located on chromosome 11p15.5, encoding a 2.3 kb long, spliced, and polyadenylated ncRNA and acts as an imprinted gene transcribed from the intergenic locus of the maternal allele [47]. H19 has been reported to promote GBC cell proliferation by modulating miR-194-5p targeting AKT2 [48]. However, the direct interaction of H19 with miR-194-5p was not interpreted. Another publication revealed that H19 competitively binds endogenous miR-342-3p and thus regulates FOXM1 expression to promote tumor development in GBC [32]. Additionally, the authors argued that H19 was identified to be a direct target of miR-342-3p and participated in the regulatory network as a ceRNA [32].

Colon cancer-associated transcript-1 (CCAT1), a lncRNA first identified in colon cancer, maps to chromosome 8q24.2 with a 5200 nt long isoform and a 2628 nt short isoform [49]. CCAT1 was reported to be significantly upregulated in GBC tissues compared with adjacent normal tissues. The increased expression of CCAT1 was highly associated with tumor status, lymph node invasion and advanced tumor node metastasis stage of GBC patients. Further experiments demonstrated that CCAT1 knockdown impaired GBC cell proliferation and invasion [36]. The silence of CCAT1 repressed the expression of Bmi1, which is a miRNA-218-5p target gene, whereas CCAT1 overexpression enhanced its expression. Moreover, CCAT1 transcript level was correlated with Bmi1 mRNA level in GBC tissues.

HOTAIR is a 2158-bp lnRNA located at the antisense strand of the HOXC gene locus on chromosome 12q13.13 [50]. Previous literature showed that HOTAIR was significantly upregulated in GBC and it is a direct target of c-Myc by interacting with putative c-Myc target response element in the upstream region of HOTAIR [51]. Researchers predicted a miRNA-130a binding site of HOTAIR and observed a negative correlation between HOTAIR and miRNA-130a in GBC, indicating that the oncogenic activity of HOTAIR is exerted in part through its repression of miRNA-130a [51]. Furthermore, HOTAIR was demonstrated to act as sponge to miRNAs in a variety of human cancers, involving miR-148a (esophageal cancer) [52], miR-613 (pancreatic cancer) [53], miR-545 (colorectal cancer) [54] and miRNA-1 (hepatocellular carcinoma) [55].

Antisense non-coding RNA at the INK4 locus (ANRIL), located on chromosome 9q21.3, showed significantly higher expression in GBC tissues than in normal tissues. ANRIL can improve the GBC cell proliferation and inhibit apoptosis [56]. Several miRNAs have been reported to interplay with ANRIL, including miR-125a [57], miR-186 [58], miR-323 [59], miR-125a [60], miR-199a [61] and let-7a [62]. However, links of ANRIL to miRNAs were still unrevealed in GBC up to now.

Gallbladder cancer-associated suppressor of pyruvate carboxylase lncRNA (lncRNA GCASPC) [37], a modestly conserved lncRNA on human chromosome 6, was first identified by Ma et al. GCASPC was observed significantly downregulated in GBC tissues and the expression level of GCASPC was negatively associated with tumor size, tumor stage, overall survival and disease-free survival rates. Functionally, GCASPC was identified as a target of miR-17-3p, and negatively regulates pyruvate carboxylase-dependent cell proliferation in GBC [37].

Prognosis associated with gallbladder cancer lncRNA (lncRNA-PAGBC) was identified as an independent prognostic marker in GBC. PAGBC can competitively bind to miR-133b and miR-511 and activate the AKT/mTOR pathway to promote tumour growth and metastasis, serving as an independent prognostic factor for overall survival of GBC patients [33].

As mentioned above, MALAT1, was reported to crosstalk with miR-206 and miR-363-3p to serve as an oncogenic lncRNA in GBC [35, 46]. An earlier publication demonstrated that MALAT1 also promotes GBC proliferation and metastasis by activating the ERK/MAPK pathway, which is often aberrantly activated in human cancers [63]. However, the underlying molecular mechanism of MALAT1 with ERK/MAPK pathway is still unclear. MALAT1 inhibits expression of endogenous miR-206 by directly binding to miR-206 and thus increases levels of miR-206 target oncogenes KRAS and ANXA2 [35]. MALAT1 also acts as a ceRNA targeting MCL-1 by binding miR-363-3p [46]. These data suggest that MALAT1 may regulate tumor-related signaling pathways by its sponge effects as well.

Another lncRNA, maternally expressed gene 3 (MEG3), an imprinting gene located at the imprinted DLK-MEG3 locus with length of 1.6 kb and ten exons located at chromosome 14q32.3 [64], was demonstrated to be reduced in GBC tissues compared to normal tissue samples [56]. MEG3 can induce the accumulation of p53 protein and thus activate p53-dependent signaling pathways to suppress GBC tumorigenesis [56]. Moreover, MEG3 also decreases the expression of the cyclin D1 gene and ubiquitin ligases and may play a vital role in apoptosis pathways in GBC [56]. ANRIL was found together with MEG3 in GBC but showed significantly high levels, serving as an oncogene [56]. The inhibition of ANRIL showed significantly repression of phosphorylated PI3K and Akt, suggesting that ANRIL induced proliferation and metastasis of cancer cells through the PI3K/Akt pathway [65]. LINC00152, identified significantly upregulated in GBC and correlated positively with tumor status, lymph node invasion and TNM stage advancement, can also participate in the PI3K/AKT signaling pathway to promote GBC tumorigenesis [19].

6. LncRNAs participate in epithelial-mesenchymal transition (EMT) process

EMT is a process of epithelial cells trans-diffe- rentiate into mesenchymal cells and acquire mesenchymal features including losing cell adhesion and increasing cell motility, which is a main step of tumor metastasis [66]. Several lncRNAs, involving H19 [67], long intergenic non-coding RNA (Linc-ITGB1) [68], HOXA cluster antisense RNA2 (HOXA-AS2) [69] and NmrA-like family domain containing 1 pseudogene (Loc344887) [70] were revealed to induce EMT in GBC. For example, evidence showed that the upregulation of H19 in GBC was induced by TGF- β 1 and IL-6 treatment, followed by downregulation of E-cadherin and increased expression of vimentin, indicating an EMT phenotype [67]. Furthermore, the overexpression of H19 accelerated GBC invasion and promoted EMT by enhancing transcription factor twist, while the interference of H19 suppressed GBC cell invasion and reversed the EMT process (also means mesenchymal-epithelial transition) by suppressing twist expression [67]. These results indicated that H19 might play important regulatory effects on the EMT process in GBC.

7. LncRNAs implicated in metabolic reprogramming in GBC

Reprogrammed energy metabolism has drawn increasing attention since it facilitates tumor cell growth and proliferation through increased glycolysis and glucose uptake. Even in sufficient oxygen state, most cancer cells employ aerobic glycolysis coupling with reduced mitochondrial oxidative phosphorylation for energy instead of oxidative phosphorylation [31, 71]. Recently, lncRNAs have been demonstrated as key regulators in cancer metabolism [72, 73, 74]. For instance, lncRNA breast cancer anti-estrogen resistance 4 (BCAR4) was reported to coordinate the Hedgehog signaling to enhance the transcription of HK2 and PFKFB3, two key glycolysis activators, and thus reprograms the glucose metabolism in breast cancer [73]. In GBC, lincRNA GCASPC can downregulate the expression level and activity of pyruvate carboxylase, an enzyme that converts pyruvate to oxaloacetate, by limiting protein stability, and may play an important role in GBC cell metabolism [37].

8. Conclusions and future perspective

Table 1 listed fifteen lncRNAs involved in GBC, and twelve of them harbor oncogenic properties and only three of them repress tumorigenesis, suggesting that lncRNAs tend to exhibit a carcinogenic function in GBC. This phenomenon has also been reported in prostate cancer. Most of lncRNAs involved in prostate cancer are over-expressed and only few exhibit a decreased expression [29]. Identification of GBC specific lncRNAs, like GCASPC and PAGBC, offers potential strategies to overcome the present limitations for GBC diagnosis and prognosis. Understanding the effects of lncRNAs in cancer development and progression, especially the crosstalk between targets and lncRNAs themselves, helps scientists to develop strategies to prevent and/or treat GBC. For example, targets for the oncogenic (e.g. UCA1, TUG1, MALAT1, H19) and tumor suppressive (e.g. LET, GCASPC, MEG3) lncRNAs could be exploited for therapeutic purposes by reversing these lncRNA expression levels. From a clinical perspective, the dysregulation lncRNAs may also provide a promising clue to improve GBC patient outcome.