Long Non-Coding RNAs in Diagnosis, Treatment, Prognosis, and Progression of Glioma: A State-of-the-Art Review

Glioma is the most common malignant central nervous system tumor with significant mortality and morbidity. Despite considerable advances, the exact molecular pathways involved in tumor progression are not fully elucidated, and patients commonly face a poor prognosis. Long non-coding RNAs (lncRNAs) have recently drawn extra attention for their potential roles in different types of cancer as well as non-malignant diseases. More than 200 lncRNAs have been reported to be associated with glioma. We aimed to assess the roles of the most investigated lncRNAs in different stages of tumor progression and the mediating molecular pathways in addition to their clinical applications. lncRNAs are involved in different stages of tumor formation, invasion, and progression, including regulating the cell cycle, apoptosis, autophagy, epithelial-to-mesenchymal transition, tumor stemness, angiogenesis, the integrity of the blood-tumor-brain barrier, tumor metabolism, and immunological responses. The well-known oncogenic lncRNAs, which are upregulated in glioma, are H19, HOTAIR, PVT1, UCA1, XIST, CRNDE, FOXD2-AS1, ANRIL, HOXA11-AS, TP73-AS1, and DANCR. On the other hand, MEG3, GAS5, CCASC2, and TUSC7 are tumor suppressor lncRNAs, which are downregulated. While most studies reported oncogenic effects for MALAT1, TUG1, and NEAT1, there are some controversies regarding these lncRNAs. Expression levels of lncRNAs can be associated with tumor grade, survival, treatment response (chemotherapy drugs or radiotherapy), and overall prognosis. Moreover, circulatory levels of lncRNAs, such as MALAT1, H19, HOTAIR, NEAT1, TUG1, GAS5, LINK-A, and TUSC7, can provide non-invasive diagnostic and prognostic tools. Modulation of expression of lncRNAs using antisense oligonucleotides can lead to novel therapeutics. Notably, a profound understanding of the underlying molecular pathways involved in the function of lncRNAs is required to develop novel therapeutic targets. More investigations with large sample sizes and increased focus on in-vivo models are required to expand our understanding of the potential roles and application of lncRNAs in glioma.


INTRODUCTION
Glioma is the most common malignant central nervous system (CNS) tumor with significant mortality and morbidity (1). Glioblastoma is the most common and aggressive type of glioma with a median overall survival of less than two years (2). Notwithstanding substantial advances, the exact molecular pathways involved in tumorigenesis, tumor suppression, and treatment response are not fully elucidated in glioma, and patients commonly face a poor prognosis (3).
Non-coding ribonucleic acids (RNAs), comprising more than 97% of the human genome with various functions in physiological and pathological conditions, play a major role in glioma tumorigenesis (4). Non-coding RNAs are divided into the categories of short and long non-coding RNAs. The quintessential example of the former group are mi-RNAs, the role of which in glioma has been thoroughly investigated and reviewed (5,6). In the past decade, long non-coding RNAs (lncRNAs) have drawn extra attention. More than 95% of the articles on lncRNAs and glioma retrieved from PubMed were published after 2017.
Lack of optimal treatment options in addition to specific and sensitive biomarkers (7) necessitates investigation of molecular pathways involved in glioma progression in the hope of finding novel therapeutic and diagnostic targets. LncRNAs may stand as prospective candidates for this purpose.
In this review, after providing a brief background on lncRNAs and their functions, we reviewed their role in various oncogenic processes. We also assessed their role in determining treatment response, survival, and prognosis. Lastly, the diagnostic and prognostic value of circulatory lncRNAs and potential therapeutic applications of modulation of lncRNAs expression in-vivo were investigated.

AN OVERVIEW ON LncRNAs
LncRNAs are non-protein-coding RNAs with more than 200 nucleotides that are transcribed mainly by RNA polymerase II. As a result, lncRNAs, like messenger (m)RNAs, are typically polyadenylated and capped (8). However, compared to mRNAs, they are more nuclear-localized, more scarce, less evolutionary conserved, and contain fewer exons (9).
LncRNAs have various functions in the nucleus and cytoplasm. In the nucleus, they play a role in chromatin remodeling, modulating chromosomal interactions, transcription regulation, and regulation of gene expression at a post-transcriptional level by altering the function and integrity of nuclear bodies. In the cytoplasm, they are involved in mRNA turnover, translation, and post-translational modification regulation. To regulate mRNA stability, competing endogenous RNAs (ceRNA) can modulate mi-RNA availability via vying with mRNAs for mi-RNA and act as mi-RNA sponges. Moreover, lncRNAs can recruit mRNA degradation-associated proteins or act as decoys for RNA binding proteins involved in mRNA decay machinery. lncRNAs can affect translation through interacting with ribosomes or modifying mRNAs to activate their translation. lncRNAs are also involved in a variety of post-transcriptional modifications, most importantly phosphorylation and ubiquitination (9,10).

Overview -Expression Pattern
Metastasis-associated lung adenocarcinoma transcript 1 (MALAT1), also known as nuclear enriched abundant transcript (NEAT)2, is an intergenic lncRNA located on chromosome 11q13. Originally, MALAT1 was introduced as a prognostic marker in nonsmall cell lung cancer. It is associated with several cancers such as breast, ovarian, prostate, pancreatic cancers, and leukemia (11).
Both higher (12,13) and lower (14,15) MALAT1 expression are found in glioma than non-neoplastic tissue. Similarly, among glioma cell lines, glioma stem cell lines showed either lower (14) or higher (15) MALAT1 expression than parental cells. Cancer stem cells showed upregulation of MALAT1 compared to differentiated cancer cells in glioblastoma (16). Notably, between different glioblastoma cell lines, MALAT1 expression was higher in U87 than U251 (15).
In contrast to the above-mentioned mechanisms, Han et al. reported that knockdown of MALAT1 induced tumor proliferation in U87 and U252 cell lines, potentially by suppressing the extracellular signal-regulated kinases (ERK)/mitogen-activated protein kinase (MAPK) signaling pathway (22). In line with their finding, Cao et al. found that MALAT1 can have tumor-suppressing effects by reducing miR-155 expression and increasing expression of FBXW7 tumor suppressor, which interacts with several molecules involved in cellular growth, development, stemness, and cell cycle (14,23,24).
Apoptosis: MALAT1 knockdown increased apoptosis and expression of apoptotic regulators, including MYC and CCND1 (encoding cyclin D1) in glioma (13). Inhibition of apoptosis by MALAT1 can also be regulated via the MALAT1/ miR-101/Rap1B axis (19) and the miR-124/ZEB2 axis (18). Moreover, inhibition of MALAT1 by si-MALAT1 resulted in a significant decrease in the levels of several molecules involved in apoptosis, such as Bcl-2, inhibitors of apoptosis proteins family, and heat shock protein (HSP) 70 (16). Additionally, MALAT-1 knockdown resulted in lower expression of Bax and higher expression of Bcl-2 via regulating the miR-199a/ZHX1 axis (20).
Autophagy: Although autophagy may induce cytotoxic effects, it has also been suggested to promote the progression and viability of glioma in stressful environments (25). MALAT1 is found to promote tumor progression by enhancing autophagy. Sponging miR-101 not only enhanced tumor proliferation by upregulating Rap1b (19), but also induced higher expression of autophagy-associated genes (Stathmin 1, RAB5A, and ATG4D) (26). MALAT1 acts as a sponge for miR-384 as well. Inhibition of miR-384 activity induced autophagy by putatively interfering with Golgi membrane protein 1 (GOLM1) and led to increased migration and invasion of glioma cells (27,28).
Conversely, Han et al. found that MALAT1 knockdown increased invasion and proliferation of glioma cells in addition to inducing higher expression of matrix metalloproteinase (MMP)2 (22).
Stemness: MALAT1 overexpression promoted proliferation of glioma stem cells (30) by enhancing SRY-related HMG-box (SOX)-2 expression via inhibiting tumor suppressor miR-129, which led to increased tumor proliferation and viability (17). In addition to SOX-2, MALAT1 downregulation has shown inhibitory effects on the expression of Nestin (another stemness marker) and proliferation of glioma stem cell lines by activating the ERK/MAPK signaling pathway, which is a key pathway in tumor development (15,31).
Blood tumor barrier (BTB): MALAT1 knockdown led to enhanced BTB permeability and reduced expression of tight junction proteins in glioma endothelial cells via upregulating miR-140. The effect of miR-140 on BTB is mediated by inhibiting expression of nuclear factor YA (NFYA), a regulator of BTB integrity, resulting in increased expression of tight junction proteins (12). Immunology: In microglia, deactivating MALAT1 using si-MALAT1 modulated the miR-129-5p/high mobility group box 1 protein (HMGB1) axis resulting in a reduced inflammatory response (30).

Clinical Applications
Circulatory biomarker: Compared to healthy controls, glioma patients had lower serum MALAT1 levels (14). Serum levels of MALAT1 have also been used as diagnostic and prognostic biomarkers in some other cancers (32)(33)(34).
Prognostic value: Meta-analyses showed that increased MALAT1 expression could predict poor overall survival (35,36) and higher "tumor, node, metastasis" (TNM) stage (37) in glioma patients. Moreover, tissue expression levels of MALAT1 positively correlated with tumor grade according to the world health organization (WHO) classification and tumor size (18,38). Serum levels of MALAT1 were also positively associated with WHO grade, tumor size, functional impairment (14), and overall and recurrence-free survival (39). However, Shen et al. did not find a significant association between serum levels of MALAT1 and 2-year survival or disease-free survival (40).
Determining treatment response: MALAT1 plays a major role in tumor chemosensitivity with a higher expression in temozolomide (TMZ)-resistant glioblastoma cells. Its knockdown reduced TMZ resistance both in vivo and in vitro (16,41). Additionally, elevated serum levels of MALAT1 predicted chemoresistance (39). MALAT1 can modulate treatment response via several mechanisms. It can inhibit the miR-101 pathway through direct binding, resulting in increased apoptosis and suppressed cell growth (41), downregulate miR-203 leading to upregulation of thymidylate synthase (39), and regulate ZEB1 expression (42). Furthermore, inhibition of MALAT1 resulted in decreased expression of multi-drug resistance (MDR)-associated protein 1 (MRP1), a drug efflux pump associated with TMZ resistance (16).

Overview -Expression Pattern
H19 is an imprinted intergenic lncRNA located on chromosome 11p15.5, which is generally expressed by the maternal allele. H19 has well-known oncogenic effects in several cancers, such as hepatocellular carcinoma, bladder, breast, gastric, and colorectal cancers (44).
The expression of H19 was higher in glioma tissue (low and high grade) (45,46) and cell lines, including U251 and U87MG cells (47,48).
Apoptosis: Downregulation of H19 induced apoptosis and stopped the cell cycle (52) mainly by suppressing the Wnt/bcatenin signaling pathway (48), in addition to iASPP upregulation (52). Knockdown of H19 via siRNA resulted in increased TMZinduced apoptosis rate in U87MG and U251 cell lines in glioblastoma (47).
Epithelial-mesenchymal transition (EMT) process: EMT, a major role player in tumorigenesis by promoting metastasis, tumor stemness, and chemoresistance, is characterized by increased expression of epithelial markers, such as E-cadherin, and decreased mesenchymal markers, such as N-cadherin, vimentin, and ZEB1/2 (55).

Clinical applications
Circulatory biomarker: While circulatory H19 levels were a reliable prognostic indicator, to the best of our knowledge, their diagnostic value has not been investigated in glioma (40,62). Moreover, H19 plasma levels are proposed as a diagnostic biomarker for gastric cancer (63).
Prognostic value: H19 overexpression in glioma tissue was associated with poor overall and progression-free survival and more advanced tumor stage (45,46,48,64). Moreover, serum levels of H19 showed a significant positive correlation with tumor grade (62). However, Shen et al. did not find a significant association between serum levels of H19 and 2-year or disease-free survival (40).
In-vivo therapeutic applications: H19 promoted proliferation, migration, and angiogenesis in tumor xenograft investigations, while its knockdown inhibited tumor progression (46,51,52). Modulating the miR-342/Wnt5a/b-catenin axis is one of the proposed mechanisms for the oncogenic effect of H19 on tumor growth, metastasis, and angiogenesis in vivo (67). Notably, in a study assessing the therapeutic effects of phenformin in glioblastoma, phenformin was found to inhibit tumor stemness through downregulating H19 and high mobility group A (HMGA)2 (68).

Role in Tumor Pathology
Cell cycle and proliferation: MEG3 plays a substantial role in glioma proliferation and cell cycle regulation. Deletion of MEG3 increased tumor cell growth and enhanced cell proliferation in normal human astrocytes (76). MEG3 overexpression led to cell cycle arrest in the G2/M phase in U251 cells (77) and inhibited cell proliferation of glioma cells (71).
Apoptosis: MEG3 overexpression induced apoptosis in glioma cell lines, mainly regulated by the interaction of MEG3 and p53 activation (72,73). Apoptosis was inhibited after silencing of MEG3 in U118 cells. At the same time, it was enhanced following MEG3 overexpression in U251 cells through induction of cell cycle arrest at G2/M phase and increasing mRNA levels of caspase 8/3 and TP53, both playing a crucial role in cell apoptosis (73,77).
Invasion and metastasis: Silencing of MEG3 increased migration and invasion in glioma (77). The interaction of MEG3 and tumor suppressors plays a key role in tumor invasion and metastasis. MEG3 upregulation inhibited metastasis via regulating the miR−96−5p/MTSS1 axis (71). Its downregulation promoted migration and invasion via modulating the miR-19a/PTEN axis through acting as a ceRNA for miR-19a (79). Downregulating miR-21-3p (70) and enhanced expression of SMARCB due to sponging miR-6088 (80) are among other proposed mechanisms by which MEG3 blocks tumor invasion and migration. Nevertheless, since MEG3 overexpression induced autophagy (72), it increased migration and invasion in U87 and U251 cells via this mechanism (81).

Clinical Applications
Circulatory biomarker: To the best of our knowledge, the biomarker value of circulatory MEG3 has not been investigated in glioma. Although, circulatory MEG3 has shown biomarker value in other cancers, such as colorectal (82), gastric (83), breast (84), bladder (34), and pancreatic (85).
Determining treatment response: MEG3 also determined chemoresponse in glioma. TMZ-resistant glioblastoma had a lower MEG3 expression compared to TMZ responders (39). M o r e o v e r , e n h a n c e d M E G 3 e x p r e s s i o n i n c r e a s e d chemosensitivity to cisplatin while MEG3 silencing via si-RNA induced chemoresistance (87).

Role in Tumor Pathology
Cell cycle and proliferation: HOTAIR is required for the formation of glioblastoma (93) and influenced the cell cycle (96) by regulating molecules having a role in its different phases (97). Several mechanisms have been suggested for the involvement of HOTAIR in the cell cycle. HOTAIR can promote cell growth by suppressing EZH2 [the catalytic component of polycomb repressive complex 2 (PRC2)] activity, which leads to chromatin condensation by binding to the PRC2 complex (98). HOTAIR is found to have reciprocal interactions with miR-15-b and p53. miR-15-b positively regulated p53. Both of these molecules inhibit tumor proliferation and invasion, while HOTAIR activity can suppress their impact (99). Moreover, HOTAIR suppressed the b-catenin pathway, leading to cell cycle arrest and repression of invasion, putatively by downregulating Nemo-like kinase (NLK) in glioblastoma (93). HOTAIR silencing also decreased cyclin D1 expression by upregulating miR-219 (100). Additionally, HOTAIR promoted tumor proliferation by acting as a ceRNA for miR-218, resulting in upregulation of PDE7A (101). HOTAIR also activated the mTOR pathway via regulating miR-125a, resulting in increased tumor viability (102). HOTAIR also downregulated tumor suppressor programmed cell death 4 (PDCD4), leading to increased growth and proliferation of glioma stem cells (103).
Invasion and metastasis: HOTAIR downregulation inhibited tumor invasiveness and migratory abilities. Several molecular mechanisms have been suggested for the positive effect of HOTAIR on tumor progression. Downregulating NLK, resulting in increased activation of the b-catenin pathway (93), suppressing miR-125a, leading to increased activity of the mTOR pathway (102), and inhibiting the tumor suppressor miR-15b/ p53 axis (99) are among these mechanisms. Moreover, overexpression of HOTAIR was also associated with higher levels of MMP-7 and MMP-9 (106). Regulating the miR-218/ PDE7A axis (101) and glutamine metabolism via downregulating miR-126-5p, which resulted in glutaminase upregulation (105), in addition to inhibiting tumor suppressor PDCD4 (103), are other proposed mechanisms.

Clinical Applications
Circulatory biomarker: Serum HOTAIR levels were significantly higher in glioblastoma patients than controls and correlated with HOTAIR expression within the glioblastoma tissue and glioma grade (108). Circulatory HOTAIR has also been proposed as a potential biomarker for other cancers (109)(110)(111).
Prognostic value: In addition to the higher circulatory HOTAIR levels in higher glioma grades (108), several investigations, including those with large datasets, found that HOTAIR is far more expressed in high-grade than low-grade glioma tissue (91,93,96,106). Moreover, IDH-wild type cases, which typically have a poor prognosis, had higher expression levels of HOTAIR (91). Higher HOTAIR expression was an independent predictor of reduced overall survival in glioblastoma (91). Additionally, two single-nucleotide polymorphisms (SNP) of HOTAIR (rs920778 CT and rs12826786 CT genotypes) were also associated with more prolonged overall survival in patients with WHO grade III anaplastic oligodendroglioma (112).
Determining treatment response: Expression of HOTAIR was higher in non-TMZ responder glioblastoma patients compared to responders (39). HOTAIR downregulation induced increased chemosensitivity to TMZ treatment, the underlying mechanism of which may be HOTAIR acting as ceRNA for miR-126-5p (105). Moreover, HOTAIR was found to induce higher expression of HK2 via downregulating miR-125. Increased hexokinase 2 (HK2) expression is associated with chemoresistance putatively through HK-2 mediated lactate production and mitochondria permeability transition pore opening (113). Notably, in addition to HOTAIR, the expression of HK-2 is also related to other lncRNAs, including MALAT1, UCA1, and PVT1.
In-vivo therapeutic applications: In vivo, knockdown of HOTAIR using shRNA inhibited tumor growth and invasiveness and enhanced chemosensitivity (93,113). Notably, promoter demethylation using 5-Aza-2'-deoxycytidine, which is typically used in the treatment of leukemia, affected the expression of HOTAIR (91). The decreased HOTAIR expression was associated with inhibition of invasiveness, angiogenesis, and chemoresistance (105). Of note, HOTAIR downregulation can mediate the tumor-suppressive effects of some miRNAs, such as miR-326 (90).

Overview -Expression Pattern
Plasmacytoma variant translocation 1 (PVT1) is an intergenic lncRNA located on chromosome 8q24, a well-known cancerassociated region. The role of PVT1 has been explored in several cancers such as leukemia, colon, hepatocellular, breast, lung, and ovarian cancers (114).

Clinical Applications
Circulatory biomarker: To the best of our knowledge, the biomarker value of circulatory PVT1 has not been investigated in glioma. Nevertheless, circulatory PVT1 levels had diagnostic and prognostic value in some cancers (125,126).
Prognostic value: Higher expression of PVT1 was an indicator of poor prognosis (116) and survival (127) in glioma. Patients with higher glioma grade, metastasis, or IDH wild type glioma had higher tissue expression of PVT1 (116,119,121,123,128). Higher PVT1 expression positively correlated with Ki-67 level and the number of TP53 mutations (127). However, PVT1 expression was not associated with gender, age, KPS score, or tumor size (116). Only a few studies have evaluated the prognostic role of PVT1 SNPs in glioma. Ding et al. reported that while rs13255292 and rs4410871 increased susceptibility to glioma in the Chinese Han population, they do not have a prognostic value (129).

Role in Tumor Pathology
Cell cycle and proliferation: UCA1 interacted with the cell cycle. Its knockdown inhibited tumor proliferation, and its overexpression had the opposite effect both in vitro and in vivo (135,136). UCA1 knockdown induced G0/G1 cell cycle arrest and downregulation of cyclin D1 (132). Cyclin D1 is also involved in the Wnt/b-catenin signaling, which is inhibited following UCA1 knockdown (135). Additionally, sponging tumor suppressor miR-122 (133), miR-135a (136), and enhancing iASSP expression via downregulating miR-182 (137) also contribute to the underlying mechanism of the positive effect of UCA1 expression on tumor proliferation.

Clinical Applications
Circulatory biomarker: Diagnostic values of circulatory UCA1 has been reported in other cancers, such as bladder, gastric cancer, and colorectal cancer (143)(144)(145). However, we did not find any reports in patients with glioma.
Determining treatment response: UCA1 overexpressioninduced chemoresistance (shown by an increased IC50) to cisplatin and TMZ in U87 and SHG139 cells. This effect was attenuated by inhibiting the Wnt/b-catenin signaling pathway. Notably, TMZ sensitivity increased after UCA1 knockdown (135).

Other lncRNAs
More than two hundred lncRNAs have been identified to be associated with glioma (146). Providing a detailed review of all of them would be beyond the scope of this review. Therefore, in this section, we give an overview of the other most investigated lncRNAs in glioma.
Multiple studies reported increased expression and an oncogenic effect for NEAT1 in glioma (147)(148)(149)(150)(151)(152)(153)(154)(155)(156)(157)(158)(159)(160). NEAT1 increased the activity of several signaling pathways with key roles in the cell cycle, including WNT/b-Catenin and mTOR signaling, leading to increased proliferation, invasion, and metastasis and decreased apoptosis. The interaction of NEAT1 and EZH2, which mediates the trimethylation of H3K27 in their promoters, results in the activation of the WNT/b-Catenin pathway (156). Moreover, NEAT1 activated the mTOR signaling by acting as a ceRNA for miR-185-5p (158). NEAT1 also altered the activity of some cell cycle regulators, including CDK6 and CDK14, via regulating the expression of miR-107 and miR-139-5p, which led to promotion of tumor proliferation, invasion and stemness, and inhibition of apoptosis (148,149,154). NEAT1 also enhanced EMT via sponging miR-185-5p (158), and its knockdown resulted in increased BTB permeability by binding to miR-181d-5p. In contrast to these reports, Liu and colleagues found lower NEAT1 expression in glioma tissue compared to the adjacent tissue. They found that NEAT1 overexpression inhibited tumor proliferation and promoted apoptosis via regulating the miR-92b/DKK axis (161).
XIST, another oncogenic lncRNA with increased expression in glioma, plays a major role in regulating the cell cycle leading to increased tumor proliferation and invasion and decreased apoptosis. Some of the underlying molecular mechanisms include regulating the expression of Bcl-2 via cross-talk with miR-204-5p (164), upregulating CREB1 via sponging miR-329 (167), and regulating the insulin receptor substrate 1 (IRS1)/ PI3K/Akt pathway via acting as a ceRNA for miR-126. Additionally, XIST also promoted EMT and tumor stemness via regulating the miR-133a/SOX4 and miR-152-Krüppel-like factor 4 (KLF4) axes, respectively. XIST also induced tumor proliferation and angiogenesis via inversely regulating miR-429 (169) and miR-137. As a result of targeting miR-137, XIST knockdown also increased BTB permeability (168).
Second to MEG3, the anti-tumoral effects of GAS5 are well investigated in glioma. GAS5 has a major role in cell cycle regulation with several mechanisms. For instance, GAS5 regulated the expression of tumor suppressors Bcl-2-modifying factor (Bmf) and Plexin C1 via targeting miR-222, which led to inhibition of tumor progression (215). GAS5 also inhibited tumor inducer miR-196-5p, which led to suppressed tumor growth by positive regulation of tumor suppressors forkhead box protein O1 (FOXO1) and phosphotyrosine interaction domain containing 1 (PID1) (213). Direct interaction of GAS5 and EZH2, in addition to the promotion of miR-424 expression, are among other putative mechanisms for the tumor-suppressing effect of GAS5. miR-424 inhibited AKT3 and regulated the expression of cyclin D1, c-Myc, Bax, and Bcl-2 (214). Furthermore, GAS5 was found to inhibit excessive autophagy in glioma (216).
Moreover, the function of MATN1-AS1 is controversial in glioma. Han et al. found lower expression of MATN1-AS1 in glioblastoma tissue compared to the adjacent tissue, acting as a tumor suppressor (226). Its overexpression inhibited proliferation and promoted apoptosis (226). On the other hand, Zhu and colleagues found higher MATN1-AS1 expression in glioma tissue and cell lines (227). They reported oncogene effects for this lncRNA, with its silencing inhibiting proliferation and promoting apoptosis./ As depicted in Figure 2, lncRNAs are involved in almost all stages of tumorigenesis (228). In addition to the lncRNAs described in Tables   In addition to diagnosis, circulatory levels of lncRNAs can also aid in determining prognosis. Lower circulatory levels of GAS5 (40) and TUSC7 (240) and higher circulatory levels of AWPPH (238) were associated with poor prognosis. Moreover, levels of circulating lncRNAs may also illuminate treatment response. For instance, higher levels of lncSBF2-AS1 in serum exosomes were associated with poor TMZ-response (241). The most investigated lncRNAs with either diagnostic or prognostic value for their circulatory levels are described in Figure 3.
Notably, controversial findings were reported for some lncRNAs.  (245). In contrast to HOXA11-AS, for which several studies reported a positive correlation with poor prognosis, low expression of HOXA11, which is within the same family, was associated with poor outcome in glioblastoma (246). Additionally, higher levels of CCAT2 (247), HOTTIP, HANR, and lower levels of DRAIC and HCG11 were also associated with poor prognosis (248). In addition to the expression level, SNPs of some lncRNAs may also provide prognostic information. For example, specific ANRIL SNPs were related to the susceptibility of glioma and patients' overall survival (249,250).

DISCUSSION
Given the mounting and emerging evidence on the roles of lncRNAs in different cancers, including glioma, this review provided a comprehensive summary of the mechanisms of action and clinical relevance of the most investigated lncRNAs in glioma. A profound understanding of the underlying molecular pathways involved in the function of lncRNAs is required to develop novel therapeutic targets. As described earlier, several lncRNA/miRNA/mRNA axes have been  proposed to mediate the oncogenic or tumor suppressor effects of lncRNAs. In addition to well-known roles and associations with prognosis and treatment response in various cancers, lncRNAs can provide clinical clues in several non-neoplastic diseases, such as neurodegenerative disease (256) and cardiovascular disease (257,258). The disrupted pattern of lncRNAs in a wide spectrum of cancers raises the question of whether the roles and associations identified in a particular type of malignancy can be expanded to the other cancer types. A recent study found that while some lncRNAs are consistently associated with better or poorer prognosis across different cancer types, some other lncRNAs show different associations in various cancers (259). LncRNAs hold promise for developing novel biomarkers and therapeutic targets. To the best of our knowledge, prostate cancer antigen 3 (PCA3) is the only lncRNA approved as a diagnostic biomarker for prostate cancer in clinical practice (260). All in all, given the high specificity of tissue/serum lncRNAs in glioma, they may be excellent candidates for novel biomarkers. LncRNAs secreted as exosomes in body fluids, especially serum, can provide novel non-invasive assessment tools (255). Two main approaches are commonly utilized to modulate the expression of lncRNAs, namely antisense oligonucleotides (ASOs) and duplex RNAs, such as siRNA. ASOs may be preferred for lncRNAs functioning in the nucleus, while siRNA may be selected for lncRNAs functioning in the cytoplasm (261). While oligonucleotide therapeutics provide an opportunity to target any gene of choice, there are several obstacles in their clinical use. As a consequence of their chemical structure, they are susceptible to rapid enzymatic and nonenzymatic degradation. Moreover, their relatively large size hinders their penetration through the blood-brain barrier and cellular uptake. Therefore, novel delivery systems, such as chemical engineering and nanoparticles, are needed to overcome these challenges (262, 263). Despite considerable attention drawn to lncRNAs in cancer and the growing evidence, there are several gaps in the literature. The findings of some studies are not sufficiently reliable due to small sample sizes. Moreover, the majority of investigations are in-vitro assessments highlighting the need for further validation by nude animal models and clinical trials. However, the low conserveness of lncRNAs among different species may hinder using animal models since the function of a certain lncRNA can be different between humans and animal models. As a result, engineered models may be required (261).
Our study shed light on several directions for future studies. In addition to the need for increased in-vivo investigations and studies with larger sample sizes, it is not elucidated whether disruption in expression of lncRNAs is a culprit or consequence of the malignancy. Future studies need to address this question, particularly by investigating upstream regulators of the expression of lncRNAs. Moreover, the development of specific panels of lncRNAs for diagnostic or prognostic applications can lead to increased specificity and sensitivity. Several studies have already taken this approach and sought to find specific signatures of lncRNAs for glioma (264)(265)(266). Additionally, more studies are required on the therapeutic effects of combining the lncRNAtargeted therapies and conventional chemotherapy. For instance, a better outcome was achieved after administrating si-MALAT1 in addition to TMZ in vitro and in animal models (16). Additionally, detecting the disease-associated lncRNAs and their regulatory pathways can also lead to finding novel putative drugs (267). Lastly, the advancement of computational technologies and bioinformatics have provided novel opportunities for identifying new lncRNAs and their potential molecular mechanism (268). Applying artificial intelligence technology, including machine-learning and deep-learning models, can also aid in the identification of novel lncRNAs associated with a specific disease mainly via classification models (269). To conclude, regulating diverse cellular signaling pathways and the expression of various proteins involved in different stages of tumor formation, proliferation, and invasion, lncRNAs are potential candidates for developing novel diagnostic, prognostic, and therapeutic approaches. The so far discovered associations between their expression in the tissue or circulatory exosomes with treatment response and prognosis boost hopes for their potential use in clinical practice. However, despite substantial advances, the role of lncRNAs in glioma remains fairly unknown. It is not well known whether the disruption in the expression of lncRNAs has a causal effect on glioma or is a consequence of the malignant process. Moreover, some controversial reports hinder drawing a concrete conclusion. More investigations with larger sample sizes and increased focus on in-vivo models are required to expand our understanding of the potential roles and application of lncRNAs in glioma.

AUTHOR CONTRIBUTIONS
SM: Conceptualization, Investigation, Writing -Original Draft, and Visualization. NR: Conceptualization, Writing -Review & Editing, Supervision. All authors contributed to the article and approved the submitted version.