The Impact of Long Non-Coding RNAs in the Pathogenesis of Hepatocellular Carcinoma

Hepatocellular carcinoma (HCC) is among the utmost deadly human malignancies. This type of cancer has been associated with several environmental, viral, and lifestyle risk factors. Among the epigenetic factors which contribute in the pathogenesis of HCC is dysregulation of long non-coding RNAs (lncRNAs). These transcripts modulate expression of several tumor suppressor genes and oncogenes and alter the activity of cancer-related signaling axes. Several lncRNAs such as NEAT1, MALAT1, ANRIL, and SNHG1 have been up-regulated in HCC samples. On the other hand, a number of so-called tumor suppressor lncRNAs namely CASS2 and MEG3 are down-regulated in HCC. The interaction between lncRNAs and miRNAs regulate expression of a number of mRNA coding genes which are involved in the pathogenesis of HCC. H19/miR-15b/CDC42, H19/miR-326/TWIST1, NEAT1/miR-485/STAT3, MALAT1/miR-124-3p/Slug, MALAT1/miR-195/EGFR, MALAT1/miR-22/SNAI1, and ANRIL/miR-144/PBX3 axes are among functional axes in the pathobiology of HCC. Some genetic polymorphisms within non-coding regions of the genome have been associated with risk of HCC in certain populations. In the current paper, we describe the recent finding about the impact of lncRNAs in HCC.


INTRODUCTION
Liver cancer is among the most lethal malignancies among both sexes. More than 8% of cancerrelated mortalities are due to this type of cancer (1). Hepatocellular carcinoma (HCC) includes more than 75% of the primary liver neoplasms (1). Several factors have been related with elevated risk of HCC among them are chronic infection with hepatitis B virus (HBV) B or hepatitis C virus (HCV), dietary exposure with aflatoxin, excessive alcohol use, obesity, and smoking (2). The cirrhosisinduced carcinogenic alterations have been detected in 90% of HCC patients (3). High throughput sequencing methods have shown the occurrence of several genetic changes in the HCC samples (4) among the early events are inactivating mutations in insulin-like growth factor 2 receptor (5). Catenin Beta 1 (CTNNB1) and Tumor Protein P53 (TP53) are the utmost recurrently mutated oncogene and tumor suppressor gene in HCC, respectively (4). In addition to these somatic mutations, several epigenetic factors partake in the evolution of HCC. Such involvement is further highlighted by the fact that liver is an organs that is continuously adapting to extremely various environmental factors (6). Non-coding RNAs are among epigenetic elements that contribute in the pathogenesis of HCC. Long non-coding RNAs (lncRNAs) can affect expression of genes via diverse mechanisms including recruitment of regulatory protein complexes, acting as a decoy, changing genome organization and modulating the distribution of posttranslational modifications (7). These transcripts have sizes longer than 200 nucleotides and are comparable with mRNAs in the terms of chromatin state of genome loci, their transcription by RNA polymerase II, polyadenylation, 5' capping and being spliced, yet they do not produce large-sized polypeptides (8). However, there are several reports demonstrating the presence of stable, functional micropeptides being translated from lncRNAs (9). Several lines of evidence indicates that these transcripts contribute in the pathophysiology of HCC (10). In the present manuscript, we review the current knowledge about the partake of lncRNAs in the pathogenesis of HCC.

UP-REGULATED LNCRNAS IN HCC
The LINC01138 is located in a frequently amplified region in HCC. This lncRNA transcript is stabilized by IGF2BP1/ IGF2BP3. Over-expression of LINC01138 in HCC confers malignant characteristics and is associated with poor survival of patients. Mechanistically, this lncRNA interacts with arginine methyltransferase 5 and increases the stability of this protein through inhibiting ubiquitin-mediated degradation in proteasomes (11). Over-expression of the lnc-Epidermal Growth Factor Receptor (EGFR) regulatory T cells (Tregs) has been related with tumor size and levels of EGFR/Foxp3. Its overexpression has also been negatively correlated with the levels of interferon (IFN)-g in HCC patients and animal models. This lncRNA promotes Treg differentiation, inhibits function of cytotoxic T cells and increases HCC growth. These effects are exerted through binding of lnc-EGFR with EGFR, increasing its stability and activation of the AP-1/NF-AT1 axis (12). The oncogenic lncRNA HULC has been shown to exert its effects via modulation of phosphorylation pattern of YB-1. Notably, upregulation of this lncRNA in HCC has been correlated with pathological grade and patients' outcome. HULC can also increase cell proliferation, migration, and invasion and suppress cisplatin-associated cell apoptosis (13). LncRNA-MUF is another over-expressed lncRNA in HCC tissues whose upregulation has been correlated with poor clinical outcome. This lncRNA has an indispensable impact in epithelial-mesenchymal transition (EMT). Such effects have been exerted through binding with Annexin A2 and induction of the Wnt/b-catenin signaling. Mechanistically, lncRNA-MUF serves as a competing endogenous RNA (ceRNA) for miR-34a, resulting in upregulation of Snail1 induction of EMT process (14). GHET1 over-expression in HCC sections has been associated with vascular invasion, cirrhosis, size of tumor, histological grade, and poor clinical outcome. GHET1 silencing has suppressed cell proliferation and prompted both cell cycle arrest and cell apoptosis. GHET1 can suppress expression of KLF2 in HCC cells through recruitment of PRC2 into its promoter (15). MALAT1 is another up-regulated lncRNA in HCC, which affect neoplastic transformation through several mechanisms among them is its role as a ceRNA. Figure 1 depicts this mechanism in HCC.

Down-Regulated lncRNAs in HCC
Through a high throughput approach, Ni et al. have identified uc.134 as a novel lncRNA which is under-expressed in a highly aggressive HCC cell line. They further verified its downregulation in clinical HCC samples compared with paired nearby tissues. Notably, down-regulation of uc.134 has been related with poor prognosis of HCC patients. Functionally, this lncRNA suppresses cell proliferation, invasion, and metastasis through binding with CUL4A suppressing its nuclear export. Besides, uc.134 suppresses the CUL4A-associted ubiquitination of LATS1 and enhances YAPS127 phosphorylation which results in down-regulation of YAP target genes of YAP (223). LncRNA-PRAL has been shown to suppress HCC growth and stimulate apoptosis via a p53-dependent route. Certain motifs at the 5' end of this lncRNA have been identified that participate in competitive inhibition of MDM2-dependent p53 ubiquitination (224). Expression of the lncRNA-LET has been decreased in HCC. Further experiments have shown the role of hypoxiainduced histone deacetylase 3 in down-regulation of this lncRNA. Notably, repression of lncRNA-LET has been identified as an important step in the stabilization of nuclear factor 90 protein and subsequent hypoxia-associated tumor cell invasion. The association between down-regulation of lncRNA-LET and metastatic potential of HCC has also been verified in clinical samples (225). TSLNC8 is also down-regulated in HCC samples. Down-regulation of this lncRNA in HCC has been shown to confer malignant phenotype. TSLNC8 competitively interacts with transketolase and STAT3 and alters the phosphorylation patterns and transcriptional activity of STAT3 leading to suppression of the IL-6-STAT3 signaling (226). CASC2 is another down-regulated lncRNAs in HCC samples, particularly in the samples obtained patients with aggressive and recurrent forms of HCC. CASC2 suppresses migration and invasive properties of HCC cells and inhibits EMT program in these cells. Mechanistically, it serves as a competing endogenous RNA for miR-367 to increase expression of its target gene FBXW7. Notably, CASC2 down-regulation and miR-367 upregulation have been associated with the metastasis-associated characteristics in the clinical samples (227). Table 2 displays the impact of down-regulated lncRNAs in HCC.

GENOMIC VARIANTS WITHIN LNCRNAS AND RISK OF HCC
Genetic polymorphisms include at least four type of variations namely, single nucleotide polymorphisms, small insertion/deletion polymorphisms, polymorphic repetitive elements and microsatellites. The importance of somatic copy number variations (SCNVs) loci in non-coding regions in the development of HCC has been assessed by  (16). MALAT1 also reduces expression of miR-204 in HCC leading to upsurge in SIRT1 levels. SIRT1 up-regulation enhances expression of Vimentin and Twist and inhibits E-cadherin, thus facilitating epithelial-mesenchymal transition (EMT) (17). MALAT1 can also sequester miR-143-3p, thus up-regulating FGF1, N-cadherin, Vimentin, Snail, and Slug while down-regulating E-cadherin. These effects are associated with enhancement of EMT (18). Similarly, through downregulation of miR-30a-5p, MALAT1 enhances Vimentin levels and EMT process (19). Via sequestering miR-200a, MALAT1 increases aspartate-b-hydroxylase (ASPH) levels, thus elevating expression of proteins which are involved in EMT or cell proliferation (20). MALAT1-mediated down-regulation of miR-124-3p leads to upregulation of Slug, therefore increasing cell proliferation and EMT (21). MALAT1 can also sponge miR-195 resulting in over-expression of FGFR, activation of PI3K/ AKT and enhancement of cell proliferation and invasion (22). Finally, MALAT1-mediated down-regulation of miR-22 increases Snail levels and facilitates EMT. Moreover, MALAT1 recruits EZH2 to the of promoter E-cadherin and miR-22 to decrease their expression (23). Table 1 enlists function of over-activated lncRNAs in HCC.               (224). The lncRNA TSLNC8 on 8p12 is another tumor suppressor lncRNA which is commonly deleted in HCC tissues (226). Table 5 shows the summarized results of studies which assessed association between lncRNAs insertion/deletion or tetranucleotide repeat polymorphisms and HCC.

DISCUSSION
LncRNAs contribute in the pathogenesis of HCC through diverse mechanisms including modulation of oncogenes and tumor suppressor genes as well as modification of tumor microenvironment. The latter route of action has been best exemplified by the lnc-EGFR which enhances differentiation of    Tregs therefore increasing immune evasion (12). Moreover, certain lncRNAs such as MUF and SNHG7 facilitate EMT process through modulation of Wnt/b-catenin signaling pathway (14,114). Other lncRNAs can modulate EMT through sponging a number of miRNAs. MAPK, PI3K/AKT and JAK/STAT signaling pathways are other cancer-related pathways that are modulated by several lncRNAs in HCC. The interactions between lncRNAs, miRNAs and mRNAs have functional importance in the pathogenesis of HCC. Examples of such trios include H19/miR-15b/CDC42, H19/miR-326/ TWIST1, NEAT1/miR-485/STAT3, MALAT1/miR-124-3p/ Slug, MALAT1/miR-195/EGFR, MALAT1/miR-22/SNAI1 and ANRIL/miR-144/PBX3. Functional roles of lncRNAs in HCC have been appraised in animal models. These models have facilitated identification of lncRNAs targets and related pathways (304), which can be used as therapeutic candidates in HCC. HCC-associated lncRNAs can affect gene expression via recruiting epigenetic factors (305), regulation of transcription factors (306), modulation of protein degradation (307) and alteration of phosphorylation of proteins (308).
Genomic alterations and polymorphisms within lncRNAcoding regions have been shown to confer risk of HCC. Such variations might also predict survival of these patients. However, the observed association between these variants and HCC should be verified in independent samples from different ethnic groups. Integration of the results of genome-wide association studies with high throughput sequencing data obtained from microarray and RNA seq experiments would help in discovery of HCCrelated single nucleotide polymorphisms within lncRNAs.
The biomarker role of lncRNAs in HCC has been verified by several studies indicating their importance both in the diagnosis and in the prognosis of this cancer. Expression levels of lncRNAs can differentiate HCC patients from inactive HBs Ag carriers, patients with chronic hepatitis and those with liver cirrhosis. In addition, the high diagnostic power values of peripheral levels of a number of lncRNAs such as UCA1 and NEAT1 have potentiated them as methods for non-invasive diagnosis of HCC. Moreover, lncRNAs can be regarded as therapeutic targets in HCC. The importance of lncRNAs as therapeutic targets in HCC has been noted by several experiments in animal models of HCC. Yet, such experiments wait approval in clinical settings. In vivo delivery of a number of lncRNAs such as lncRNA-PRAN, uc.134 and TSLNC8 has been shown to attenuate tumor growth and enhance lifespan of xenograft models of HCC (223,224,226). Moreover, a number of  Its elevated level is related with low OS.
an independent prognostic factor for OS (31) UBE2CP3 46 HCC specimens and ANTs Its elevated level is related with poor OS. --

LINC00461
87 HCC specimens and paired ANTs Its elevated level is related with decreased OS. --

MALAT1
56 HCC specimens and paired ANTs Its elevated level is related with decreased OS. -- MNX1-AS1 81 HCC specimens and paired ANTs Its elevated level is related with poor OS. -- MCM3AP-AS1 80 HCC specimens and paired ANTs Its elevated level is related with shorter OS.
-- -- XIST 52 HCC tissues and paired ANTs Its elevated level is related with poor survival of HCC patients. -- TRPM2-AS 108 HCC tissues and paired ANTs Its elevated level is related with poor OS. --

LSINCT5
126 HCC tissues and paired ANTs Its elevated level is related with poor OS. --

HCC tissues and paired ANTs
Its elevated level is related with poor OS.
an independent prognostic marker for OS (93) (Continued)  an independent prognostic factor for survival (180)

HCC tissues and paired ANTs
Its elevated level is related with poor OS. -- AGAP2-AS1 137 HCC tissues and paired ANTs Its elevated level is related with poor DFS and OS. --

AK002107
134 HCC tissues and paired ANTs Its elevated level is related with poor DFS and OS.
an independent prognostic factor for DFS and OS (184) DDX11-AS1 40 HCC tissues and paired ANTs Its elevated level is related with poor OS. -- GATA3-AS1 80 HCC tissues and paired ANTs Its elevated level is related with low OS. --

HCC tissues and paired ANTs
Its elevated level is related with low OS. -- KTN1-AS1 80 HCC tissues and paired ANTs Its elevated level is related with low OS. -- Linc-GALH 108 HCC tissues and paired ANTs, 12 normal liver tissues Its elevated level is related with poor RFS and OS.
-- an independent prognostic factor for OS (15) OR3A4 78 HCC tissues and paired ANTs Its elevated level is related with poor OS.
an independent prognostic factor for HCC (213) PITPNA-AS1 60 HCC tissues and paired ANTs Its elevated level is related with poor OS. -- lnc-FTX 129 HCC tissues and paired ANTs Its low expression is related with poor OS and RFS. --

HCC tissues and 35 ANTs
Its expression is correlated with short OS. -- (Continued) --

LIN00607
159 HCC tissues and paired ANTs Its low expression is related with low OS. --

AOC4P
108 HCC tissues and paired ANTs Its low expression is related with low DFS and OS.
an independent prognostic factor for DFS and OS (246) uc. 134 170 paraffin-embedded samples of HCC tissues and ANTs Its low expression is related with low OS. -- GAS8-AS1 82 HCC tissues and paired ANTs Its low expression is related with poor OS. -- LINC00657 49 HCC tissues and paired ANTs Its low expression is related with poor OS. -- MAGI2-AS3 88 HCC tissues and paired ANTs Its low expression is related with poor OS.
an independent prognostic factor for OS an independent prognostic factor for OS (282) TCONS_00027978 241 HCC tissues and paired ANTs Its low expression is related with low DFS and OS.
an independent prognostic factor for DFS and OS (283) Expression levels of lncRNAs can differentiate HCC tissues from non-tumoral tissues indicating the role of these transcripts as diagnostic biomarkers for HCC. The best diagnostic power values have been reported for NEAT1, PANDAR, CCHE1 and SNHG1. Most notably, serum or plasma levels of a number of lncRNAs such as LINC-ITGB1, LINC00978, LncDQ, PAPAS, MEG3, UCA1 and NEAT1 could be used as diagnostic markers for this kind of cancer ( Table 4).  Heterozygote subjects with one allele 10 and those without allele 10 compared with subjects with homozygote 10-10 genotype have decreased risk of HCC.
-Cell lines without allele 10 have higher expression of KCNQ1OT1. (303) lncRNAs such as HULC confer resistance to chemotherapeutic agents (13), indicating the potential of targeted therapies against these transcripts in enhancement of response of HCC patients to conventional therapeutic options. Antisense oligonucleotides and small interfering RNAs are putative methods for suppression of expression of lncRNAs (309, 310) whose efficacies have been verified in animal models and cell line experiments. Yet, this knowledge has not been translated into clinical practice. Taken together, lncRNAs as important class of regulatory transcripts can influence pathogenesis of HCC from different aspects and can be used as suitable markers for differentiation of HCC from related pathogenic conditions.

AUTHOR CONTRIBUTIONS
SG-F and MT wrote the draft and revised it. BH and MG designed the tables and figures. All authors contributed to the article and approved the submitted version.