Long non-coding RNA HOTAIR, transcribed from the antisense strand of the homeobox C (HOXC) gene locus in chromosome 12, is involved in the regulation of specific gene transcription. A study by Tsai et al. (2010) demonstrated that HOTAIR regulates gene expression by interacting with polycomb repressive complex 2 (PRC2) and lysine-specific demethylase 1A (LSD1). The 5′- and 3′-domains of HOTAIR can bind to the PRC2 and LSD1/CoREST/REST complex, respectively. Serving as a scaffold, HOTAIR can tether two distinct complexes together and recruit specific histone modification enzymes, thereby resulting in H3K27 methylation and H3K4 demethylation and ultimately gene silencing (Tsai et al., 2010). In addition, HOTAIR could also serve as an inducer of ubiquitin-mediated proteolysis to control protein levels. HOTAIR facilitates the ubiquitination of Ataxin-1 through E3 ubiquitin ligases Dzip3, Snurportin-1, and Mex3b to further accelerate their degradation. Through the rapid decay of targets Ataxin-1 and Snurportin-1, HOTAIR can prevent cellular premature senescence (Yoon et al., 2013). In addition, aberrant HOTAIR expression has been extensively revealed to correlate with cancer metastasis and is characterized as a negative prognostic factor for cancer patients (Cai et al., 2014; Wu et al., 2014).

Hox transcript antisense intergenic RNA expression is up-regulated in glioma tissues and cell lines, and can serve as a potential biomarker or therapeutic target for human gliomas (Kiang et al., 2015; Zhou et al., 2015) (Figure 3). Recent studies have indicated that HOTAIR expression is a critical regulator of cell cycle progression in gliomas (Zhang J.X. et al., 2013). HOTAIR regulates cell cycle progression predominantly via the HOTAIR 5′-domain-PRC2 axis, which is EZH2 (predominant PRC2 complex component)-dependent in GBM cells (Zhang K. et al., 2015). In addition, bromodomain and extraterminal (BET) domain proteins are required for GBM cell proliferation. BET protein inhibitors can reduce the proliferation of gliomas, in part, through the induction of the cyclin-dependent kinase inhibitor p21^Cip1 in vitro and in vivo (Pastori et al., 2014). Pastori et al. (2014) found that the bromodomain protein BRD4 could directly control HOTAIR expression by binding to its promoter. The overexpression of HOTAIR in conjunction with the BET protein inhibitor I-BET151 abolishes the anti-proliferative activity of the BET bromodomain inhibitor (Pastori et al., 2015). Furthermore, the HOTAIR-miRNA axis has an important role in malignant biological behaviors of human glioma. Ke et al. (2015) found that fibroblast growth factor 1 (FGF1) mediates oncogenic effects by activating the PI3K/AKT and MEK 1/2 pathways. HOTAIR, one target of miR-326, has been confirmed to down-regulatemiR-326; then, it exerts its tumor-suppressive activities by reducing the expression of FGF1 (Ke et al., 2015). Similarly, HOTAIR can act as an endogenous “sponge” of miR-141, thereby promoting the promoter methylation of miR-141 by DNA methyltransferase 1 (DNMT1) in glioma cells. Then, the hypermethylated miR-141 can repress the expression of the spindle and kinetochore associated complex subunit 2 (SKA2), which results in a significant increase in tumor growth (Bian et al., 2016). Collectively, these results suggest that HOTAIR may potentiate glioma development in many facets; thus, it is worthy of further investigation.

Long non-coding RNA H19, produced from the imprinted gene H19, is one of the most highly conserved transcripts involved in mammalian development. Studies have also demonstrated that H19 could potentially serve as an oncogenic lncRNA in different types of cancers, including gliomas (Kiang et al., 2015; Chen et al., 2016). Mechanistically, the product of the MYC oncogene, c-Myc, induces the expression of the H19 non-coding RNA, thereby potentiating gliomagenesis (Barsyte-Lovejoy et al., 2006). Furthermore, serving as a miRNA precursor, H19 could modulate glioma progression by generating miR-675. The oncogenic function of H19/miR-675 is dependent on the expression of cancer-associated cadherin 13 (CDH13), which is the direct target of miR-675 (Shi et al., 2014). Li C. et al. (2015) found that miRNA-675, which is derived from the first exon of H19, could regulate the immoderate proliferation and migration of glioma cell lines by inhibiting the expression of CDK6, which is a pivotal regulator of the cell cycle and involved in glioma development (Rader et al., 2013; Sherr et al., 2016). These findings agree with another study showing that H19 overexpression can promote the cell-cycle progression of cancer cells (Berteaux et al., 2005). Moreover, the knockdown of H19 by siRNA displays higher therapy efficiency when induced by the chemotherapy drug TMZ in GBM cells (Li W. et al., 2016). Thus, LncRNA H19 could be increasingly recognized as a potential target for glioma treatment.

Accumulating evidence has identified that tumorigenic CSCs, with self-renewing capability, contribute to tumor initiation and therapeutic resistance (Lathia et al., 2015). Intriguingly, H19 overexpression could maintain the stem cell properties of GBM cells. Li W. et al. (2016) found that the markers of CSCs, including CD133, NANOG, Oct4, and Sox2, are significantly down-regulated in H19-deficient cells. This conclusion was further confirmed by Jiang et al. (2016), who found that the increased level of H19 promotes invasion, angiogenesis, and stemness of GBM cells. H19 is significantly overexpressed in CD133-positive GBM cells, and higher H19 expression levels are associated with increased tumor growth (Jiang et al., 2016). In spite of the critical role of H19 in the maintenance of glioma stemness, its exact mechanism is still unclear and needs to be further investigated.

Colorectal neoplasia differentially expressed was initially identified by Derrien et al. (2012) as a putative non-coding RNA; it is highly expressed in developmental neurobiology and neuropathology. Studies have found that CRNDE expression is also elevated in many colorectal cancers and brain cancers, such as GBM, astroblastomas, and astrocytomas (Ellis et al., 2012; Kiang et al., 2015). Of note, among the 129 lncRNAs differentially expressed in glioma tissues, CRNDE is consistently identified as the most up-regulated lncRNA by 32-fold up (Zhang X. et al., 2012).

Colorectal neoplasia differentially expressed potentiates glioma development possibly by maintaining the stemness of the tumor cells, as it functions in neural precursors (Ellis et al., 2012; Watkins and Sontheimer, 2012). In support of this notion, a previous study by Zheng J. et al. (2015) demonstrated a direct link between the overexpression of CRNDE and GSCs. Mechanistically, CRNDE could negatively regulate miR-186 and depress the expression of the downstream target genes XIAP (X-linked inhibitor of apoptosis) and PAK7 [p21 protein (Cdc42/Rac)-activated kinase 7], thus contributing to the malignant characteristics of human GSCs (Zheng J. et al., 2015). In addition to these observations, Wang Y. et al. (2015) showed that the overexpression of the CRNDE transcript promotes glioma cell growth in vitro and in vivo through mammalian target of rapamycin (mTOR) signaling. Intriguingly, epigenetic modifications, including histone acetylation in the promoter region, can also promote CRNDE expression (Wang Y. et al., 2015). More recently, CRNDE was reported to promote malignant behavior by attenuating the miR-384/PIWIL4 (piwi-like RNA-mediated gene silencing 4) axis. Briefly, CRNDE knockdown can decrease the protein level of PIWIL4, a target of miR-384, which leads to glioma regression in vivo (Zheng J. et al., 2016). Overall, these results revealed that CRNDE could potentiate glioma via multiple signaling pathways and may be a promising novel therapeutic target for glioma therapy.

Cancer susceptibility candidate 2, located at chromosome 10q26, is a lncRNA originally identified as a tumor suppressor gene in endometrial cancer. CASC2 consists of three alternatively spliced transcript isoforms, CASC2a, CASC2b and CASC2c, which contain identical first three exons and diverse downstream exons (Zhao et al., 2014). CASC2a expression is down-regulated at the transcription level in endometrial cancer. Baldinu et al. (2007) revealed that the exogenous expression of CASC2a in undifferentiated endometrial cancer cells significantly inhibits the clonal growth. Using a positional candidate approach, 7% CASC2a mutations in tumor DNA from 44 endometrial cancer patients were identified (Baldinu et al., 2004), suggesting that inactivation of CASC2a might probably be due to mechanisms different from genetic alterations. In non-small cell lung cancer (NSCLC) tissues and cell lines, He X. et al. (2016)) reported that CASC2 expression is involved in the development and progression of NSCLC. However, little is known about the role and function of CASC2 in human gliomas.

Recently, Wang P. et al. (2015) reported that CASC2 expression is decreased in glioma tissues as well as glioma cell lines (U251 and U87). Consistent with previous studies in other tumors, the overexpression of CASC2 could inhibit the malignancy of glioma cells through an arrest of proliferation and migration, correspondingly promoting cellular apoptosis. RIP and RNA pull-down assays confirmed that the tumor suppressive role of CASC2 is mainly mediated via the down-regulation of miR-21, one potential direct target of CASC2, in a sequence-specific manner (Wang P. et al., 2015). A growing body of literature has shown that miR-21 serves as an oncogene, and the inhibition of miR-21 is a novel therapeutic strategy for specific and effective action against gliomas (Harmalkar et al., 2015; Belter et al., 2016). Mechanistically, miR-21 promotes gliomagenesis by regulating multiple oncogenesis-related processes, including proliferation, apoptosis, migration and invasion. Therefore, targeting the CASC2-miR-21 axis may be an effective strategy for the treatment of malignant gliomas.

X-chromosome inactivation (XCI) ensures dosage compensation between the sexes in mammals and is a paradigm for allele-specific gene expression on a chromosome-wide scale. The lncRNA XIST, a product of the XIST gene, is located within the 500 kb stretch of XCI DNA at Xq13, which is known as the X-inactivation center (XIC); XIST is the master regulator of X chromosome inactivation in mammals (Furlan and Rougeulle, 2016; Maduro et al., 2016). The current model proposes that XIST induces epigenetic silencing of multiple genes by recruiting the chromatin modifier, the PRC2 complex, to the XIC (Goodrich et al., 2016). With the help of the high-affinity RNA-binding protein ATRX (alpha thalassemia/mental retardation syndrome X-linked), a growing number of XIST RNAs accumulate and are tethered to the X chromosome. Afterward, the XIST RNA spreads and forms a RNA “cloud” coating the XIC in cis. To recruit PRC2, the XIST RNA first associates with approximately 150 intense PRC2 binding sites (CpG islands), followed by its association with 3,000–4,000 moderate-strength binding sites of PRC2. Finally, XIST RNA spreads to both gene-rich and poor regions in distinct stage-specific forms on the X chromosome (Simon et al., 2013; Sarma et al., 2014).

X-inactive specific transcript has been found to be dysregulated in a variety of human cancers (Yildirim et al., 2013; Tantai et al., 2015). Specifically, a recent study showed that XIST expression is abnormally up-regulated in glioma tissues and GSCs. The knockdown of XIST by short-hairpin RNA exerts a tumor suppressive function in GSCs. Furthermore, as XIST and miR-152 may form a reciprocal repression feedback loop and are located in the same RNA induced silencing complex (RISC), miR-152 can mediate the promotion of GSCs by XIST (Yao et al., 2015). In addition, XIST can inhibit hepatoma cell proliferation and metastasis by targeting miR-92b (Zhuang et al., 2016). Moreover, XIST has been identified to directly bind tomiR-210 (Fasanaro et al., 2009). Consistently, other miRNAs, such as miR-92b and miR-210, may also regulate the expression of XIST in gliomas. Altogether, further studies should focus on the XIST-miRNA axis in glioma research and treatment.

Taurine up-regulated gene 1, a 7.1 kb lncRNA located at chromosome 22q12, is a cancer-related lncRNA in some tumors, including NSCLC (Zhang et al., 2014), bladder cancer (Tan et al., 2015) and gliomas (Li J. et al., 2016). TUG1 was first identified in a genomic screen for genes differentially regulated by taurine in developing mouse retinal cells. Furthermore, TUG1 is found to play crucial roles in the formation of photoreceptors and retinal development (Young et al., 2005).

Recent investigations have reported that in human glioma cell lines, TUG1 is down-regulated, in response to necrosis induced by a high dose of DOX (Liu Q. et al., 2015). Li J. et al. (2016) showed that TUG1 acts as a tumor suppressor in glioma tumorigenesis, and is negative correlated with glioma grade, tumor size, and overall survival. Further studies via gain- and loss-of-function assays revealed that TUG1 induces glioma cell apoptosis through caspases-mediated intrinsic pathways, rather than the Bcl-2-mediated anti-apoptotic pathway (Li J. et al., 2016). However, the precise mechanism of TUG1 in cell proliferation, as well as invasion, in glioma development is still unclear. The BTB limits the effect of conventional chemotherapy by restricting drug delivery to brain tumor tissues (Hendricks et al., 2015). Using a co-culture assay with glioma and endothelial cells, Cai et al. (2015) revealed that the knockdown of TUG1 could reduce tight junction protein expression in endothelial cells by down-regulating heat shock transcription factor 2 (HSF2), the target of miR-144, increasing BTB permeability of chemotherapeutic agents. Thus, there may be potential role of TUG1 in anti-glioma therapy, and BTB function may represent a useful therapeutic intervention strategy in the future.

Nuclear enriched abundant transcript 1 (NEAT1) is an essential lncRNA for the formation of paraspeckles, which are nuclear bodies named for their close proximity to nuclear speckles (Yu and Shan, 2016). NEAT1 is an unusual RNA polymerase II (pol II) transcript that lacks introns, and it is widely expressed in many types of mammalian cells (Naganuma and Hirose, 2013). NEAT2/MALAT1 is a highly conserved lncRNA associated with tumorigenesis and plays a prognostic role in various cancers (Wei and Niu, 2015).

Up to date, lncRNAs have been demonstrated to be involved in the DNA damage response, thus contributing to the process of cellular defense against genotoxic agents (Zhang and Peng, 2015). Upon treatment with the DNA damage-inducing agent resveratrol, NEAT1 is up-regulated in the glioma cell lines U251 and U87 (Liu Q. et al., 2015). An increase in NEAT1 expression has also been reported in human glioma tissues compared with non-cancerous brain tissues. NEAT1 promotes glioma pathogenesis by regulating glioma cell proliferation, invasion, and migration. Zhen et al. (2016) demonstrated that functioning as a molecular sponge for miR-449b-5p, NEAT1 could up-regulate the expression of c-Met, a direct target of miR-449b-5p, thus promoting glioma oncogenesis. Furthermore, clinical investigations revealed that aberrant NEAT1 expression is negatively associated with clinical outcome in high-grade glioma patients (He C. et al., 2016).

Recent works have illustrated the tumor-suppressive role of NEAT2 in the development of glioma cells. NEAT2 expression is lower in glioma tissues than in normal brain tissues. Mechanistically, NEAT2 inhibits the proliferation and invasion of glioma cells (U87 and U251) by inactivating ERK/MAPK signaling and down-regulating MMP2 (matrix metalloproteinase 2; Han et al., 2016a). In contrast, Xiang et al. (2016) showed an opposite role of NEAT2 in gliomas. According to their observations, NEAT2 expression is significantly increased in glioma tissues, as well as in U87 and U251 cells (Xiang et al., 2016). Remarkably, GSCs of the U87, SHG44 and SHG139 cell lines expressed higher levels of NEAT2 than their parental lines (Han et al., 2016a). In addition, Han et al. (2016b) found that the down-regulation of NEAT2 suppresses the expression of stemness markers Sox2 and Nestin in SHG139S cells, while NEAT2 down-regulation promotes the proliferation of SHG139S cells. Therefore, NEAT2 plays a complex role in gliomagenesis as both a positive and a negative regulator, possibly based on its specific cellular context.

Growth arrest-specific 5, localized at chromosome 1q25.1, could transcribe a tumor-suppressive lncRNA in human cancers. To date, GAS5 has been considered to act as a “riborepressor” or “miRNA sponge” that modulates the transcriptional activity of cancer-associated genes (Kino et al., 2010; Zhang Z. et al., 2013). Recent studies have reported that GAS5 negatively regulates the growth of cancer cell lines in vitro and in vivo, including gliomas (Pickard and Williams, 2015). GAS5 exerts complementary effects on cell proliferation (inhibitory) and apoptosis (stimulatory), and taken together, these cellular mechanisms likely form the basis of its tumor-suppression action (Yin et al., 2014; Shi X. et al., 2015). Mechanistically, the up-regulation of Gas5 increases the expression of tumor suppressor bmf (Bcl-2-modifying factor) and Plexin C1 via directly reducing the expression of miR-222 (Zhao X. et al., 2015). In addition, the overexpression of GAS5 could enhance the cellular response to erlotinib, a tyrosine kinase inhibitor used as a second line treatment for glioma (Garcia-Claver et al., 2013). The induction of GAS5 is apparently detected during DOX-induced apoptosis in human glioma cell lines (Liu Q. et al., 2015). The above examples suggest that GAS5 may be used as diagnostic markers or therapeutic targets for gliomas, but much work needs to be done before such applications become clinically practical.

The ADAMTS family has been implicated in essential physiological processes, such as angiogenesis and organ development (Ho et al., 2016). ADAMTS9-AS2 is the antisense transcript of ADAMTS9, a member of the ADAMTS family. Walsh et al. (2016) pointed that ADAMTS9-AS2 plays a critical role in epigenetic regulation, affecting early stage digit development. Recently, the ADAMTS9-AS2 locus has been revealed as a potential therapeutic target and prognostic marker in gliomas. ADAMTS9-AS2 serves as a tumor suppressor, which is significantly down-regulated in glioma tissues, and its expression is negatively correlated with tumor grade and prognosis. Meanwhile, DNMT1 knockdown remarkably enhances ADAMTS9-AS2 expression, inhibiting cell migration in gliomas (Yao et al., 2014).

SPRY4-IT1, a 708 bp intron-retained lncRNA localized at chromosome 5q31.3, is found to be significantly expressed in breast cancer (Shi Y. et al., 2015), osteosarcoma (Ru et al., 2016) and bladder cancer (Zhao X.L. et al., 2015), and its suppression can inhibit proliferation and induce apoptosis in cancer cells. SPRY4-IT1 was originally reported by Khaitan et al. (2011) to play an important role in the molecular etiology, modulation of cell apoptosis and invasion of human melanoma. Recently, the expression of SPRY4-IT1 is shown to be significantly expressed in glioma tissues and glioma cell lines compared with normal donors (Liu H. et al., 2015). The epithelial-to-mesenchymal transition (EMT), as a relevant molecular event in malignant gliomas, is an essential process in tumor dissemination and metastatic behavior (Kahlert et al., 2013). Liu H. et al. (2015) showed that the knockdown of SPRY4-IT1 by siRNA could suppress the EMT phenotype in glioma cells (U251 and SF295). However, the exact mechanism underlying the role of SPRY4-IT1 in glioma pathology still remains to be elucidated.

Highly up-regulated in liver cancer has pro-oncogenic activity in many human malignancies, such as B-cell lymphoma (Peng et al., 2016), hepatocellular carcinoma (Huang et al., 2016), and osteosarcoma (Sun et al., 2015). Recently, Zhu’s et al. (2016) reported that HULC has important biological function in human gliomas. HULC can promote the angiogenesis, one hallmark of malignant gliomas, by inhibiting the expression of angiogenesis-related molecule ESM-1 (endothelial cell specific molecule 1). In addition, the PI3K/AKT/mTOR signaling pathway is involved in the response induced by HULC (Zhu et al., 2016). These intriguing findings will help pave the way for exciting functional studies of HULC in gliomagenesis.