Long non-coding RNAs as novel targets for therapy in Hepatocellular Carcinoma

Abstract

The recognition of functional roles for transcribed long non-coding RNA (lncRNA) has provided a new dimension to our understanding of cellular physiology and disease pathogenesis. LncRNAs are a large group of structurally complex RNA genes that can interact with DNA, RNA or protein molecules to modulate gene expression and to exert cellular effects through diverse mechanisms. The emerging knowledge regarding their functional roles and their aberrant expression in disease states emphasizes the potential for lncRNA to serve as targets for therapeutic intervention. In this concise review, we outline the mechanisms of action of lncRNAs, their functional cellular roles, and their involvement in disease. Using liver cancer as an example, we provide an overview of the emerging opportunities and potential approaches to target lncRNA dependent mechanisms for therapeutic purposes.

1. Introduction

The availability and integrated use of advanced genome analysis platforms in recent years by consortia such as the Human Genome Project, Functional Annotation of the Mammalian genome project (FANTOM), and the Encyclopedia of DNA elements (ENCODE) project, have revealed that only a very small proportion of the actively transcribed human genome corresponds to protein-coding genes (Dingemans et al., 2014; Fatemi, Velmeshev, & Faghihi, 2014; Z. Zhang et al., 2014). The recognition and identification of transcribed RNAs such as the microRNAs that do not encode for proteins but yet have specific biological functions have prompted a re-evaluation of the central dogma and emphasize a versatile role of RNA beyond that of a transitional role in translation of DNA to protein. The large numbers of pervasive non-coding RNA transcripts allows us to presume that many functionally active and biologically relevant non-coding RNA (ncRNA) remain to be identified. Considering their ubiquitous presence and diversity, these non-coding RNA could be major contributors to normal cellular physiological processes.

Classification of RNA genes

NcRNAs can be arbitrarily classified into two groups based on the length of their transcripts. Small RNAs, with less than 200 nucleotides, include several distinct types of RNAs such as microRNAs, small nucleolar RNAs, PIWI-interacting RNAs and endogenous small interfering RNAs. Long RNAs, greater than 200 nucleotides include long intergenic non-coding RNAs (lincRNAs), natural antisense transcripts (NATs), transcribed ultraconserved regions (T-UCRs), long enhancer ncRNAs, non-coding repeat sequences, and pseudogenes. In contrast to some types of small ncRNAs such as microRNAs (miRNAs) and small-interfering RNAs (siRNAs), the mechanism and functions of long ncRNAs (lncRNAs) are very poorly understood (Dingemans, et al., 2014; Kimura et al., 2015; Y. Yang et al., 2013).

The lncRNAs are a highly abundant and heterogeneous group, in part reflecting their enormous diversity and structural complexity. Pointers to a functional biological role include the recognition of tissue-specific expression patterns, emerging literature showing functional relevance in complex physiological and pathological processes, and evidence of their involvement in several human diseases. In contrast to the protein coding genes, the number of lncRNAs has increased tremendously along the evolutionary process and reflects the complexity of the organism. (Necsulea et al., 2014). Despite this, an effective and useful method to categorize this large and heterogeneous group is lacking. A descriptive classification is based on genomic proximity to protein-coding genes. Depending on genomic location they can be divided broadly into several categories (Figure 1): 1. Intronic lncRNAs are located within an intron of a coding gene; 2. Intergenic lncRNAs (lincRNAs) are present between two protein-coding genes; 3. Bidirectional lncRNAs are expressed within the vicinity (within 1kb) of a coding transcript of the opposite strand; 4. Enhancer lncRNAs (e-lncRNAs) are present in the enhancer regions close to the promoter; 5. Sense: or 6. Antisense lncRNAs are those that overlap with one or more introns and exons of a different transcript on the same or opposite strand, respectively (Devaux et al., 2015; Hirose et al., 2014; H. Takahashi & Carninci, 2014). Non-coding repeats are another important class of lncRNAs that can have biological functions.

Figure 1. Non-coding RNA.

Non-coding RNA comprise a much larger portion of the human genome than protein-coding RNA, which comprise of <3% of the genome. Non-coding RNAs are arbitrarily classified into short and long non-coding RNA based on transcript size. LncRNA can be designated as Intergenic, Intronic, Enhancer, Sense, Antisense, or Bidirectional based on their genomic location relative to that of nearby protein-coding genes.

While a location-based classification may be useful for their recognition and annotation, an accurate classification of lncRNAs with respect to their biological role or biological activity is lacking.

Efforts to develop lncRNA databases have used systematic identification and annotation of lncRNAs within the human genome. The cDNA clone analysis within the FANTOM project led to the identification of over 23,000 lncRNAs (Carninci et al., 2005), while the GENCODE-ENCODE project identified approximately 24,000 lncRNA producing loci (Derrien et al., 2012). An analysis of a compendium of 7,256 RNA-seq libraries from tumors, normal tissues, and cell lines and comprehensively delineated genome-wide lncRNA expression reported the identification of 58,648 lncRNA genes (Iyer et al., 2015).

Mechanisms of action

As may be expected from their structural diversity, a wide range of functions have been described for lncRNAs. These include gene transcription, epigenomic regulation, translation of protein-coding genes, RNA turnover, chromatin organization, and genome defense (Ferraiuolo et al., 2010). Notably, most of the functions described to date involve regulation of gene expression, both of protein-coding genes or other non-coding RNAs, through a multitude of mechanisms as outlined in Table 1 and Figure 2. LncRNAs differ from small RNAs in both complexity and length. The structural complexity of lncRNAs offers multiple possibilities for interactions with DNA, RNA and/or proteins that depend on the secondary and tertiary structures. Likewise, the length of lncRNAs enables the formation of multi-component complexes such as those involved in epigenetic regulation of transcription. LncRNAs may be transcribed from DNA in several orientations, and can interact with genomic DNA elements, as well as with proteins in multiple configurations.

Table 1.

Long noncoding RNA-mediated regulatory mechanisms

lncRNA	Function	Mechanism	Reference
mRNA transcription
HOTAIR	↓ at the HOXD locus	Chromatin-mediated ↓	(Tsai et al., 2010)
HOTTIP	↑ at the HOXA locus	Chromatin-mediated ↑	(K. C. Wang et al., 2011)
XIST	X inactivation	Chromatin-mediated ↓	(Zhao, Sun, Erwin, Song, & Lee, 2008)
ANRIL	↓ at the INK4b-ARF- INK4a locus	Chromatin-mediated ↓	(Kotake, et al., 2011)
KCNQ1OT1	Imprinting at the KCNQ1 cluster	Chromatin-mediated ↓	(Pandey et al., 2008)
GAS5	↓ of glucocorticoid receptor- mediated transcription	DNA mimicry	(Kino, et al., 2010)
CCND1 promoter RNA	↓ of CCND1 transcription	Allosteric ↑ of TLS	(X. Wang et al., 2008)
NRON	↓ of NFAT-mediated transcription	↓ of transcription factor nucleocytoplasmic shuttling	(Willingham et al., 2005)
AIRN	Imprinting at the IGF2R cluster	Chromatin-mediated ↓, transcription interference	(Latos et al., 2012)
IME4 antisense	↓ of IME4 mRNA interference	Transcription	(Hongay, et al., 2006)
SRG1	↓ of SER3 mRNA	Nucleosome remodeling	(Hainer, Pruneski, Mitchell, Monteverde, & Martens, 2011)
ICR1 and PWR1	↓ and ↑ of FLO11mRNA, respectively	Modulation of transcription factor recruitment	(Bumgarner et al., 2012)
Alu repeat- containing RNA	Transcriptional ↓ during heat shock	↓ of Pol II	(Mariner et al., 2008)
HSR1	↑ of HSF1	Allosteric ↑ together with eEF1A	(Shamovsky, Ivannikov, Kandel, Gershon, & Nudler, 2006)
mRNA processing
MALAT-1	Serine/Arginine splice factor modulation	Scaffolding nuclear domain	(Tripathi et al., 2010)
ZEB2	↑ of ZEB2 translation	Controlled splicing of IRES-containing intron region	(Beltran et al., 2008)
Neuroblastoma MYC	↓ of neuroblastoma MYC intron 1 splicing	Controlled splicing via RNA-RNA duplex formation	(Krystal, Armstrong, & Battey, 1990)
Sas10mRNA 3′ UTR	↓ of Rnp4F mRNA	RNA editing	(Peters, Rohrbach, Zalewski, Byrkett, & Vaughn, 2003)
Post-transcriptional modulation
HULC	↓ of miRNA-mediate repression	Sequestration of miRNAs	(J. Wang, et al., 2010)
PTENP1 pseudogene	↑ of PTEN	Sequestration of miRNAs	(Poliseno et al., 2010)
CDR1as	↓ of miRNA-mediate repression	Sequestration of miRNAs	(Hansen et al., 2013)
BACE1AS	↑ of BACE1	Blocking miRNA- induced repression	(Faghihi et al., 2010)
Antisense UCHL1	↑ of UCHL1 protein level	Translational upregulation	(Carrieri et al., 2012)
Protein activity
EVF2	Transcriptional ↑ of DLX2 targets	↑ of DLX2	(Feng et al., 2006)
15q11-q13 sno- lncRNA	Alternative splicing	↓ of FOX2 function	(Yin et al., 2012)
mcs-1	Dicer-mediated ↓	Sequestration of Dicer or other RNA-binding proteins	(Hellwig & Bass, 2008)
sfRNA	Stabilization of viral and host mRNAs	↓ of XRN1- mediated mRNA degradation	(Moon et al., 2012)
gadd7	↓ of TDP43-mediated regulatory events	Sequestration of TDP43	(X. Liu, Li, Zhang, Guo, & Zhan, 2012)
Protein complex organization
HOTAIR	↓ at the HOXD locus	Recruitment of PRC2 and LSD1	(Tsai, et al., 2010)
KCNQ1OT1	Imprinting at the KCNQ1 cluster	Recruitment of PRC2 and G9A	(Pandey, et al., 2008)
ANRIL	↓ at the INK4b-ARF- INK4a locus	Recruitment of PRC1 and PRC2	(Kotake, et al., 2011; Yap et al., 2010)
TERC	Inclusion of telomeric repeats to chromosomal ends	Scaffolding for telomere components and template identification for repeat sequences	(Zappulla & Cech, 2004)
NEAT1	Paraspeckles formation	Nucleation of subnuclear domains	(Sasaki, et al., 2009)
SRP RNA	Guiding proteins to ER	Scaffolding of SRP components	(Halic et al., 2004)

Figure 2. lncRNA interactions and actions.

(A) LncRNAs can target genomic DNA loci, and can modulate gene transcription by associating with RNA pol II or pre-initiation complexes, or through epigenetic modulation by guiding chromatin-modifying complexes to target genomic DNA loci. (B) LncRNA can contribute to RNA inhibition or degradation, by association with mRNAs and miRNAs to regulate splicing, or by acting as endogenous sponges. (C) LncRNAs can modulate protein activity and localization by acting as molecular guides and scaffolds or as decoys for proteins such as transcription factors.

While the majority of lncRNA are present within the cytoplasm, some lncRNAs are primarily confined within the nucleus (Z. Zhang, et al., 2014). Trans-acting nuclear lncRNAs can act in a tissue-specific manner in conjunction with chromatin modifiers such as histone modifying complexes and DNA methyltransferases, to epigenetically regulate or to modulate transcription (Tao et al., 2013). However, the mechanism by which these lncRNAs can recognize specific genomic loci is not well understood. It is hypothesized that interactions with genomic regions may be dictated by secondary or tertiary structural elements. Efforts to define and categorize RNA structural modules may enable systematic approaches to evaluate this hypothesis in the future. Some lncRNAs have been shown to maintain the nuclear architecture by scaffolding the assembly of DNA-RNA-protein interface at specific locations. In contrast, cis-acting lncRNAs may be able to regulate gene expression in a locus and allele-specific manner because of their genomic proximity to their targets. Cis-acting lncRNAs have been implicated in genomic imprinting, dosage compensation, and enhancer functions (Chen, Xu, & Liu, 2013; Xu et al., 2013). In such cases, modulation of neighboring gene expression can occur via direct binding or through the recruitment of chromatin-binding proteins that can modify the chromatin structure in the vicinity. While most of the well-studied lncRNAs regulate cellular activities at the nuclear level, others are localized within the cytosol and mediate post-transcriptional control via RNA-RNA interactions (Chapman et al., 2014). They can act as sponges and interact with short non-coding RNAs, such as miRNAs to simultaneously influence the expression of several thousands of miRNA-target genes leading to an overall change in cellular physiology.

The functional heterogeneity of lncRNAs is further emphasized by their ability to contribute to multi-protein complex formation, and participate in protein localization and function. In addition, a role for lncRNA as mediators of inter-cellular signaling is also being recognized by studies that have demonstrated the presence of lncRNA within extracellular vesicles, their release during conditions of cellular stress and their ability to modulate cellular responses in recipient cells after uptake of these vesicles (Kogure, Yan, Lin, & Patel, 2013; K. Takahashi, Yan, Haga, & Patel, 2014; K. Takahashi, Yan, Wood, Haga, & Patel, 2014).

2. LncRNA in human cancers

The ability of lncRNA to alter cellular physiology by altering gene expression raises the possibility that deregulated lncRNA expression may contribute to disease pathophysiology. Altered expression of lncRNA has been reported in many cancers and in other disease settings. Based on the emerging literature, lncRNA may be useful in elucidating disease pathogenesis and pathophysiology. In settings where alterations in lncRNA are associated with disease states, lncRNA may have roles as therapeutic targets or as biomarkers of disease (Table 2). Recent studies implicate the potential role of lncRNAs in cancer development by virtue of their ability to interact with DNA, RNA, proteins and/or their combinations, and their potential to modulate many of the hallmarks of cancer (Figure 3). A specific group characterized by non-coding repeats can activate immune responses and contribute to disease progression (Chiappinelli et al., 2015; Leonova et al., 2013; Rooney, Shukla, Wu, Getz, & Hacohen, 2015; Roulois et al., 2015; Tanne et al., 2015). Examples include LINE-1. HERV, Satellites, and TERRA. In addition, the LINE-1 repeat (CoT-1) may affect chromatin formation. These repeats are actively reverse transcribed and re-incorporate in the genome in cancers, thus causing an expansion of specific genomic regions (Bersani et al., 2015; Hall et al., 2014; Iskow et al., 2010; E. Lee et al., 2012; Rodic et al., 2015). TERRA plays a role in telomere extension through the alternative lengthening of telomeres (ALT) pathway (Flynn et al., 2011).

Table 2.

Long noncoding RNA associated with human diseases

Disease	Type of disease	lncRNA	Reference
Cardiovascular disease	Myocardial infarction	MIAT	(Ishii et al., 2006)
	Myocardial ischemia	HIF1A	(Zolk, Solbach, Eschenhagen, Weidemann, & Fromm, 2008)
Diabetes	Transient neonatal diabetes mellitus	HYMAI	(Mackay et al., 2002)
		PINK1-AS	(Scheele, Nielsen, et al., 2007)
Genetic disorders	X-fragile syndrome	FMR4	(Khalil, Faghihi, Modarresi, Brothers, & Wahlestedt, 2008)
	HELLP Syndrome	HELLPAR	(van Dijk et al., 2012)
	Silver-Russell, Beckwith- Wiedemann syndrome	KCNQ1OT1	(Chiesa et al., 2012)
	Duchenne muscular dystrophy	Linc-MD1	(Cesana et al., 2011)
	Facioscapulohu- meral muscular dystrophy	DBET	(Cabianca et al., 2012)
Neurological disorders	Alzheimer’s disease	BACE1-AS, NAT- Rad18, 17A	(Faghihi et al., 2008)
	Schizophrenia	DISC2, GOMAFU	(Chubb, Bradshaw, Soares, Porteous, & Millar, 2008)
	Huntington	TUNA	(N. Lin et al., 2014)
	Parkinson	naPINK1	(Scheele, Petrovic, et al., 2007)
Cancer	Female cancers (breast, ovary, uterine, cervical)	HOTAIR, H19, GAS5, ANRASSF1, UCA1, SRA, PVT1, CCAT2, MALAT-1, XIST	(L. Cui et al., 2013; Khalil, et al., 2008; Kino, et al., 2010; Mizrahi, et al., 2009; Z. Zhang, et al., 2014)
	Colon	MALAT-1/NEAT2, PTENP1, HOTAIR, CCAT1, CCAT2, CRNDE, CUDR, H19, PVT1, KCNQ1OT1, ncRAN/SNHG16	(L. D. Graham et al., 2011; M. Z. Ma et al., 2015; Michalik, et al., 2014; Nakano et al., 2006)
	Renal	CUDR, H19, UCA1, ncRAN/SNHG16,	(Fan, et al., 2014; Y. Wang et al., 2012)
	Liver	HULC, HOTAIR, TUC338, H19, MEG3, lncRNA-ATB, HEIH, PCNA-AS1, MALAT-1, HOTTIP, linc-RoR	(Anwar et al., 2012; Du et al., 2012; Geng, Xie, Li, Ma, & Wang, 2011; Ishibashi et al., 2013; K. Takahashi, Yan, Kogure, et al., 2014; Yuan et al., 2014)
	Prostate	PCA3, ANRIL/p15AS, MALAT-1/NEAT2, PCAT-1, PCGEM1, PTENP1, CTBP1-AS, ANRASSF1, NEAT1, PlncRNA-1, PRNCR1, SRA, SChLAP1/PCAT 114	(Chung et al., 2011; Z. Cui et al., 2013; Kotake, et al., 2011; G. L. Lee, et al., 2011; Petrovics et al., 2004; Prensner et al., 2011)
	Lung	MALAT-1/NEAT2, linc-p21, TUG1, CCAT2, CUDR, HOTAIR, PVT1, AK126698	(Dingemans, et al., 2014; Gutschner, et al., 2013; Hou, et al., 2014; Qiu et al., 2014; Y. Yang, et al., 2013)
	Leukemia	ANRIL/p15AS, uc.73a, MEG3, HOTTIP, XIST, PVT1, NEAT1	(Folkersen et al., 2009; M. Graham & Adams, 1986; Rack et al., 1994; Zeng et al., 2014)
	Gastrointestinal	CCAT1, H19, HOTAIR, PVT1, MALAT-1	(Ding et al., 2014; Fellig et al., 2005; J. H. Liu, Chen, Dang, Li, & Luo, 2014; C. Ma et al., 2014)

Figure 3. lncRNAs associated with hallmarks of cancer.

LncRNAs have been implicated in key hallmarks of cancer and can contribute to the onset and progression of cancer.

The emergence of comprehensive detection technology such as RNA-seq has enabled the cataloging of databases with information on several lncRNAs that are deregulated or aberrantly expressed in several types of cancers such as hepatocellular carcinoma (HCC), gastric cancer, renal cancer, colorectal cancer, glioma, prostate cancer, and others. High-throughput, comprehensive and multi-dimensional data analyses have extended our knowledge about molecular regulatory mechanisms of lncRNAs that can control specific tumor-related behavior, thus adding a new stratum of key mediators and regulatory complexity during the evolution of cancer.

To highlight these, we will use liver cancers as an illustrative example. Liver cancers such as HCC are ideal to study lncRNA targets for treatment because liver uptake and delivery may be feasible for some therapeutic approaches. Several lncRNAs have been identified to play a functional role in HCC initiation and progression, as well as associated with recurrence, metastases or prognosis. These RNA genes could regulate epigenetic events in conjunction with chromatin modifying factors such as chromatin, histone-modifying enzymes, and DNA methylating agents leading to altered expression of target genes. The downstream cellular functions of these lncRNA transcripts are better understood than the mechanisms by which the expression of these lncRNAs are regulated, or their relevance to genetic or somatic mutations related to tumorigenesis. Examples of some functionally active lncRNAs that are associated with HCC are listed in Table 3. We emphasize these may represent just the tip of the iceberg, as many other lncRNA related to HCC or other cancers remain to be identified. To illustrate the diversity of functional mechanisms, we discuss a few of these herein (Figure 4).

Table 3.

Functional role of long noncoding RNAs in Hepatocellular Carcinoma

lncRNA	Classification	Genome Location	Effects	Reference
HULC	LincRNA	Chr6	Promotes proliferation, positive correlation with histological grade or HBV	(Du, et al., 2012; J. Wang, et al., 2010; Xie, Ma, & Zhou, 2013)
HOTAIR	Antisense	Chr12	Promotes metastasis, negative association with chemosensitivity and positively associated with invasion and recurrence	(Geng, et al., 2011; Ishibashi, et al., 2013; Z. Yang et al., 2011)
HOTTIP	Bidirectional	Chr7	Promotes proliferation, Predictor of disease outcome and tumor progression	(Quagliata et al., 2014)
HEIH	LincRNA	Chr5	Promotes proliferation by modulating cell cycle, associated with HBV-HCC	(F. Yang, et al., 2011)
H19	LincRNA	Chr11	Increased in HCC, suppresses metastasis via miR- 220 dependent pathway and associated with drug resistance	(Iizuka et al., 2002; Tsang & Kwok, 2007; L. Zhang et al., 2013)
LALR1	Antisense	Chr17	Promotes proliferation	(Xu, et al., 2013)
LET	Intronic	Chr15	Suppresses metastasis	(F. Yang et al., 2013)
MALAT-1	LincRNA	Chr11	Promotes metastasis and recurrence, migration, invasion and synaptogenesis	(Lai, et al., 2012; R. Lin, et al., 2007)
MEG3	LincRNA	Chr14	Induces apoptosis, epigenetic modulator associated with methylation	(Anwar, et al., 2012; Braconi, Kogure, et al., 2011)
MVIH	Sense	Chr10	Promotes angiogenesis, important predictor of recurrence-free survival	(Yuan et al., 2012)
RERT	Sense	Chr19	Not understood	(Zhu et al., 2012)
uc002mb e.2	LincRNA	Chr19	Induce apoptosis	(L. Zhang, et al., 2013)
Linc-ROR	LincRNA	Chr18	Tumor cell survival during hypoxia	(K. Takahashi, Yan, Haga, et al., 2014)
Dreh	Sense	Chr17	Promotes metastasis	(J. F. Huang et al., 2013)
PCNA-AS1	Antisense		Regulates proliferation and cell cycle	(Yuan, et al., 2014)
UCA1/CUDR	Intronic	Chr19	Chemotherapeutic resistance	(Tsang, Wong, Cheung, Co, & Kwok, 2007)
PVT1	Variant	Chr8	Regulates proliferation, cell cycle, and stem cell- like properties	(F. Wang, et al., 2014)
TUC338	Ultra-conserved	Chr12	Enhanced levels in liver cirrhosis and HCC, regulates cell growth	(Braconi, Valeri, et al., 2011)

Figure 4. Examples of mechanisms of lncRNAs associated with liver cancers.

(A) Epigenetic silencing: lncRNA HEIH binds to EZH2 subunit of PRC2 complex, directing this complex to specific gene locus and regulate the expression of its target genes. (B) Splicing modulation: MALAT-1 transcript within the nuclear speckles modulates the activity of serine/arginine (SR) splicing factors to generate spliced variants. (C) lncRNA-miRNA interactions: lncRNA HULC acts as a molecular sponge to sequester miRNA-372 and de-repress PRKACB expression to enable phosphorylation and activation of CREB (cAMP response element-binding protein), thus stimulating HULC expression to activate an autoregulatory loop. (D) lncRNA-protein interactions: lncRNA HULC binds to oncofetal protein IGF2BP1 and guides it to its associating partner CCR4-NOT1 complex.

A classic example of lncRNA gene networking in HCC is Highly up-regulated lncRNA in liver cancer (HULC). Overexpression of HULC is associated with low-stage and low-grade tumors, implying a functional role in early stage of cancer progression. The cAMP response element binding protein (CREB) binds to a binding site in the promoter region of HULC, and this association leads to an autoregulatory loop of HULC with miR-372 via its conserved target site causing inhibition of translational repression of the catalytic subunit of PKA (PRKACB), a miR-372 target. This in turn increases the phosphorylation of CREB, making it more available to bind to proximal promoter of HULC. Overall, this autoregulatory loop leads to increased expression of HULC in HCC (J. Wang et al., 2010). HULC can also promote proliferation of HCC by targeting and suppressing p18. IGF2BP1, an oncofetal protein acts in trans to regulate HULC stability and expression by acting as a guide by bringing HULC into close proximity to the CCR4-NOT1 deadenylase complex (Hammerle et al., 2013). Metastasis associated in lung adenocarcinoma transcript 1 (MALAT-1) is another lncRNA that is upregulated in HCC; Although the mechanistic contribution of MALAT-1 to HCC is not known, MALAT-1 is associated with alternative splicing, and interacts with SF2/ASF and CC3 antigen to modulate SR proteins (serine/arginine-rich family of nuclear phosphoproteins) that regulate splicing of various pre-mRNAs (Lai et al., 2012; R. Lin, Maeda, Liu, Karin, & Edgington, 2007). LncRNA high expression in HCC (HEIH) can contribute to cell cycle arrest by association with enhancer of zeste homolog 2, a part of polycomb repressive complex 2 that leads to its recruitment to p16 promoter region, thereby preventing the expression of p16 gene (J. J. Huang et al., 2015; F. Yang et al., 2011). Non-coding repeat LINE-1 insertions occur with high frequency in HCC (Shukla et al., 2013) LncRNAs can also bind with miRNAs as sponges to regulate gene expression. These examples serve to illustrate the diversity of many unique lncRNA-based mechanisms in HCC.

The detection of certain lncRNAs in body fluids, like urine and plasma make them potential biomarkers of disease, similar to circulating miRNA. To explore the potential role of lncRNA as biomarkers, it will be necessary to establish the expression and stability of circulating lncRNAs in healthy and diseased states. For example, plasma levels of lncRNA HULC are higher in patients with HCC patients than in healthy controls. Similarly, PCA3, a prostate-specific lncRNA that is significantly overexpressed in prostate cancer may be a promising candidate for a diagnostic biomarker for prostate cancer (G. L. Lee, Dobi, & Srivastava, 2011). Establishing the relationship between circulating lncRNA and the disease course may yield prognostic markers, as suggested by a report of ncRNA MALAT-1, as an early prognostic indicator of poor survival rate in HCC.

Although the correlation between lncRNAs and human cancers is recognized, detailed mechanistic information regarding the contribution of lncRNA to these cancers is currently lacking. Nevertheless, the rapidly growing interest and emerging data are likely to provide many new mechanistic insights into the contributions of lncRNA in cancer, and thereby are expected to identify novel candidates for therapeutic intervention or biomarkers.

3. Therapeutic approaches to targeting lncRNA in cancers

Differential expression of lncRNA in tumors compared with normal tissues has been reported for many different cancers. The identification of lncRNA that are involved in cellular processes contributing to oncogenesis, tumor suppression, or tumorigenesis provides opportunities to develop novel therapeutics for cancer that are based on targeting lncRNA (Kimura, et al., 2015; Sasaki, Ideue, Sano, Mituyama, & Hirose, 2009). These genes differ from protein-coding RNA or miRNA in several ways that need to be considered in analyzing their effects or in exploring their therapeutic potential. First, their structural complexity and participation in multi-component complexes provides several potential targetable critical residues to modulate structure-based interactions. Second, they can subserve key regulatory roles in gene expression at absolute expression levels that are low in comparison with protein-coding genes. Third, they demonstrate tissue or cell-type specificity in expression. In addition, their effects can result from very diverse mechanisms of action, and they may also participate in inter-cellular communication. Finally, our current understanding of RNA gene function emphasizes the combinatorial nature of their action, which can result in complex interactions that incorporate multiple associated effectors. As a consequence, the functional impact of targeting lncRNA needs to be analyzed in a systemic context rather than as an individual target.

Several different types of approaches can be considered to target lncRNA and to modulate their expression for therapeutic purposes (Figure 5).

Figure 5. Strategies to target lncRNA.

(A) Gene transcription: DNA-binding elements can target the genomic locus to alter lncRNA transcription. (B) Transcript destabilization/degradation: siRNAs (19–30 nt long double-stranded RNAs) can bind to complementary lncRNA sequences through RISC (RNA-induced silencing complex). ASO and Gapmers (8–50 nt long single-stranded DNAs or RNAs can carry out sequence specific and RNase H-mediated lncRNA degradation. Ribozymes (single-stranded RNA in neutral condition) can undergo cellular processing to expose the hammerhead structure of the binding arms that bind with target sites and result in cleavage of the target lncRNA. (C) Block interactions. Small synthetic molecules can block binding of lncRNAs with protein, DNA, RNA or other interacting complexes by associating with specific binding pockets. (D) Functional disruption: Aptamers (3- dimensional short RNA or DNA oligonucleotides) can bind at specific structural regions to target lncRNA and antagonize their association with their binding partners.

Silencing of lncRNAs

Similar to other genes, lncRNA can be targeted by using specific small interfering RNAs (siRNAs) that associate with RISC (RNA-induced silencing complex) complex and bind to specific target sequence based on complementarity leading to argonaute-mediated degradation of target sequence. For example, In a study by Hung et al, siRNA against promoter of CDKN1A antisense DNA damage activated RNA (PANDA) significantly reduced its expression and sensitized human fibroblasts to doxorubicin-induced apoptosis (Kimura, et al., 2015; Regha, Latos, & Spahn, 2006; F. Wang et al., 2014).

Natural antisense transcripts (NATs)

NATs are a type of lncRNA and their use illustrates an example of extending knowledge of lncRNA to design effective therapeutics. Inhibition of NATs can increase sense mRNA transcript levels (Modarresi et al., 2012) and the use of AntagoNATs is being evaluated as a therapeutic approach to upregulate the expression of specific sense mRNA (H. Takahashi & Carninci, 2014).

Antisense oligonucleotide (ASO)

Therapeutic strategies based on ASO could be used to target the expression and function of specific lncRNAs. ASOs are single stranded oligonucleotides that offer specific complementarity and RNase H-mediated degradation of target sequence. The feasibility of ASO-mediated knockdown to modulate MALAT-1 function has been demonstrated in lung cancer cells and murine xenografts where metastasis was restricted. ASO may be more effective than siRNAs for lncRNA targets in which the secondary structure of the lncRNA might preclude or limit optimal association with a siRNA. However, as with other ASO-targeting strategies, the adequacy of cellular uptake and potential for off-target effects needs to be considered (Gutschner et al., 2013; Kotake et al., 2011; Lebo, Niederer, & Zappulla, 2015; Zappulla & Cech, 2006).

Locked Nucleic Acid (LNA) GapmeRs

The use of these provides an alternative to RNAi-knockdown of upregulated lncRNAs. These single-stranded oligonucleotide consist of a DNA stretch flanked by LNA nucleotides and, similar to ASOs, form base pairs with target lncRNA to induce degradation via RNAse H-dependent mechanism (Yang, Lu, & Yuan, 2014). Inhibition of MALAT-1 using GapmeRs resulted in a functional reduction in blood flow recovery and capillary density after hind limb ischemia (Michalik et al., 2014). However, this approach is currently limited only to nuclear localized oncogenic lncRNAs.

Ribozymes or deoxyribozmes

Catalytic degradation is another therapeutic strategy. These enzymes can target RNA and cleave in a site-specific manner via a protein-independent mechanism. ‘Hammerhead’-ribozymes are designed to have ~20 nt long arms flanking a central loop and these “arms” facilitate the binding with its targets. A significant inhibition of liver metastasis was observed by anti-VEGFR-1/2 ribozymes in preclinical studies in colon cancer model (Pavco et al., 2000). However, secondary structures of lncRNAs often disrupt efficient binding of these molecules to specific targets.

Aptamers

The use of single stranded oligonucleotide or peptide aptamers could overcome limitations imposed by structural features as they could bind by forming tertiary structures with their targets. Thus, they may provide a greater specificity than siRNAs, ASOs, or ribozymes. Moreover, the chemical structure of aptamers could be modified to enhance their stability and half-life. Aptamer-based -therapeutics are promising and undergoing clinical trials for many different indications such as non-small cell lung cancer, renal cell carcinoma, and acute myeloid leukemia (Vitiello, Tuccoli, & Poliseno, 2015). In 2004, an anti-VEGF aptamer (Eyetech Pharmaceutics/ Pfizer) was approved by FDA for macular degeneration.

Small molecule inhibitors

As discussed above, lncRNAs exhibit dynamic binding affinities to various cellular components. Small molecules can mask the binding site for lncRNAs and prevent the association with its binding partners or antagonistic oligonucleotide sequences to disrupt interactions that are linked to altered function in disease-related lncRNA. For example, a small molecule inhibitor that prevents the binding of an oncogenic lncRNA, HOTAIR with its binding partners could restrict tumor cell invasion, angiogenesis, and metastatic potential (Tsai, Spitale, & Chang, 2011). Small molecules and oligonucleotides can also be used to interfere with the formation of RNA structures and disrupt folding patterns that modulate function.

RNA destabilizing elements (RDEs)

These can regulate lncRNA expression at the genomic level, and confer a knockout-like effect on gene function. RDEs have been successfully employed to reduce MALAT-1 expression (Miller et al., 2007).

Synthetic lncRNA mimics and other constructs

The use of synthetic mimics are being developed and tested for their ability to act as a decoy to block transcription of target genes or to restore their function. Alternatively, a plasmid BC-819 (DTA-H19) has been designed to specifically mediate its action through high abundance lncRNA H19 promoter to induce high levels of diphtheria toxin in tumors. The use of this construct has been shown to lead to an overall reduction in tumor size in human trials in several carcinomas although the underlying mechanisms need to be investigated (Mizrahi et al., 2009).

Several approaches to target lncRNA for therapeutic purposes can be considered once critical disease-relevant contributions of these genes have been identified. A major challenge of all of these approaches is to accomplish target-specific delivery. The use of synthetic nanoparticles for delivery of biologically active constructs has been extensively investigated. An alternative approach involves the use of extracellular vesicles, biologically derived particles that are involved in inter-cellular communication, and could be packaged with RNA genes for successful targeted therapy (Hongay, Grisafi, Galitski, & Fink, 2006; Kimura, et al., 2015). We have recently shown that lncRNAs within extracellular vesicles retain functional activity after they are taken up by recipient cells (K. Takahashi, Yan, Kogure, Haga, & Patel, 2014). The use of extracellular vesicles as delivery vehicles for lncRNA is thus highly attractive. Effective, target-specific therapy can be designed by fusing exosome membrane protein with specific peptides. Other potential advantages may include immune tolerance with the use of autologous vesicles, ability to traverse the blood-brain barrier and more effective tissue penetration into target tissues such as tumors.

Therapeutic resistance is a major challenge that limits the efficacy of cancer treatments. The role of lncRNAs in conferring differential drug-resistance phenotype is becoming appreciated, and several mechanisms for therapeutic resistance are now recognized. These include involvement of lncRNA in modulating survival signaling pathways, drug transporter expression and elimination, sensitization to apoptosis, DNA repair and cell cycle progression, and intercellular communication mediated by extracellular vesicles (Figure 6). A proposed mechanism involves trans-regulatory modulation of P-glycoprotein (P-gp) that regulates drug elimination in resistant cells. P-gp is in turn regulated by multi-drug resistance-related and upregulated lncRNA (MRUL) with inhibition of MRUL leading to increased chemotherapeutic drug-mediated apoptosis via reduction of P-gp expression and followed by impaired chemotherapeutic expulsion (Thiebaut et al., 1987; Y. Wang et al., 2014). Some lncRNAs have the ability to re-sensitize chemoresistant cancer cells to therapeutic agents like cisplatin by modulating DNA damage response pathways. Cisplatin-mediated upregulation of HOTAIR in human lung adenocarcinoma cells suppressed p21 (WAF1/CIP1) signaling pathway and caused a G₀/G₁ arrest by modulating the p53 expression (Z. Liu et al., 2013). Another lncRNA, urothelial cancer-associated 1 (UCA1) increased cisplatin-mediated resistance in human bladder cancer cells by regulating the Wnt signaling pathway. A proposed mechanism is upregulation of wingless-type MMTV integration site family member 6 (Wnt6), making this a potential target to overcome chemoresistance (Fan et al., 2014). Similarly, lncRNA AK126698 modulated cisplatin-resistance in non-small cell lung cancer cells via Wnt pathway (Hou et al., 2014; Y. Yang, et al., 2013). We identified a novel lncRNA, lincRNA-ROR to be involved in mediating chemoresistance in human hepatocellular cells. Interestingly, this lncRNA was highly enriched within extracellular vesicles derived from HCC tumor cells and caused an increased lincRNA-ROR expression and reduced chemotherapy-induced cell death in recipient cells by modulating levels of TGFβ (tumor growth factor-beta) (K. Takahashi, Yan, Kogure, et al., 2014). LncRNA GAS5 renders glucocorticoid resistance by assuming a secondary structure that competes with the sequence within the response gene promoter region (Kino, Hurt, Ichijo, Nader, & Chrousos, 2010). PANDA contributes to anthracycline resistance in a subset of breast cancer cells by coordinating with NF-YA, a nuclear transcription factor that regulates apoptosis. Down regulation of PANDA led to increased chemotherapy sensitivity thus, suggesting a crucial role in drug resistance (Malek, Jagannathan, & Driscoll, 2014).

Figure 6. LncRNA-mediated mechanisms of therapeutic resistance.

Mechanisms of tumor cell resistance to chemotherapy or hypoxia involving lncRNA include modulation of drug transporters and elimination, survival signaling pathways, cell cycle progression and DNA repair, sensitization to apoptosis, and modulation of intercellular communication mediated by extracellular vesicles.

Although in vitro studies may provide promising results with regards to lncRNA function and their use as therapeutic targets, in vivo validation is extremely challenging and requires optimization at various levels. The poor conservation of lncRNA across species can confound the interpretation of data from animal studies. Our understanding of the mechanisms of action is rudimentary, and many intriguing questions remain unanswered such as the recruitment criteria for chromatin-remodeling complexes, whether or not the binding partners of lncRNAs are constant, or the impact of environment, localization, or associated disease states on lncRNA expression and function. The advent of genome engineering capabilities such as TALEN or CRISPR/Cas9 methodologies will enable the detailed study of lncRNA function (Hainer, Charsar, Cohen, & Martens, 2012; Hongay, et al., 2006). Definition of the repertoire of lncRNA in normal tissues or cells and in disease conditions will be necessary. There is a need to develop adequately sensitive detection techniques, and to comprehensively define structural and functional motifs and modules. Even with established targets for intervention, careful in vivo analyses will be necessary, followed by validation in appropriate patient cohorts. As the pace of research in lncRNAs progresses, addressing these challenges and gaps will provide opportunities for the development of novel therapeutic strategies based on targeting lncRNA for human cancers and other diseases.

4. Conclusions

Our current knowledge regarding lncRNA provides the basis and justification to further develop applications of lncRNA as disease markers and therapeutic targets. As discussed above, lncRNAs can interact with DNA, mRNA, ncRNAs and proteins to alter cellular physiology, can be deregulated in diseases such as cancer, and can form the basis for therapeutic intervention. Although a holistic and integrated approach that incorporates lncRNA with other proteins and genes will be necessary to successfully develop effective therapies, adding these RNA genes to our future armamentarium of therapeutics is likely to be very beneficial by providing new opportunities for intervention.

Table 4.

LncRNA-based therapeutic approach in cancer

Agent	Description and Mechanism	Progress made
Small interfering RNAs (siRNAs)	Double-stranded RNAs that associate with RISC complex and exhibit argonaute- mediated degradation depending on perfect sequence complementarity	Preclinical studies
Antisense oligonucleotides (ASOs) Example: LNA Gapmers	Single-stranded oligonucleotide sequence (13–25 nucleotides long) with perfect complementarity for target lncRNA. Effective binding with secondary structure of lncRNAs, RNase-H mediated degradation or blocking of the translational apparatus	Preclinical studies, Phase I and IIa
Ribozymes and deoxyribozymes (Hammerhead ribozymes)	A RNA (ribozyme) or DNA (deoxyribozyme) molecule that exhibits catalytic properties to cleave RNA in a site-specific manner. ~20 nucleotide long “arms” flanking a central loop target specific lncRNAs for cleavage	Preclinical studies
Aptamer	Single-stranded RNA or DNA oligonucleotides capable of efficiently targeting small molecules, peptides, proteins, RNA, lncRNA, and live cells by virtue of transforming to well-defined 3- dimensional shapes	Preclinical studies and early stage clinical trials
Small-molecule drugs	Chemical compounds that block the activity of target lncRNAs by structure-specific docking to regulate activity	Preclinical studies

Abbreviations

ASO

antisense oligonucleotide

HULC

highly upregulated in liver cancer

HCC

hepatocellular carcinoma

lncRNA

Long non-coding RNA

lincRNA

Long intergenic non-coding RNA

MALAT-1

metastasis associated in lung adenocarcinoma transcript 1

miRNAs

microRNAs

mRNAs

messenger RNAs

NATs

natural antisense transcript

ncRNA

non-coding RNA

siRNAs

small interfering RNAs