The Emerging Role of Non-Coding RNAs in Pituitary Gland Tumors and Meningioma

Simple Summary Non-coding RNAs have been recently attained attention because of their contribution in the pathogenesis of brain tumors. These transcripts have been shown to be dysregulated in pituitary gland tumors as well as meningiomas. In these two types of brain tumors, dysregulation of non-coding RNAs has been associated with some clinical features and response to therapeutic options. Different types of non-coding RNAs have been shown to interact with each other to promote progression of brain tumors. Further research is needed to find the possible application of non-coding RNAs as biomarkers for pituitary gland tumors as well as meningiomas, particularly in patients’ follow-up. Abstract Long non-coding RNAs (lncRNAs), microRNAs (miRNAs), and circular RNAs (circRNAs) are non-coding transcripts which are involved in the pathogenesis of pituitary gland tumors. LncRNAs that participate in the pathogenesis of pituitary gland tumors mainly serve as sponges for miRNAs. CLRN1-AS1/miR-217, XIST/miR-424-5p, H19/miR-93a, LINC00473/miR-502-3p, SNHG7/miR-449a, MEG8/miR-454-3p, MEG3/miR-23b-3p, MEG3/miR-376B-3P, SNHG6/miR-944, PCAT6/miR-139-3p, lncRNA-m433s1/miR-433, TUG1/miR-187-3p, SNHG1/miR-187-3p, SNHG1/miR-302, SNHG1/miR-372, SNHG1/miR-373, and SNHG1/miR-520 are identified lncRNA/miRNA pairs that are involved in this process. Hsa_circ_0001368 and circOMA1 are two examples of circRNAs that contribute to the pathogenesis of pituitary gland tumors. Meanwhile, SNHG1, LINC00702, LINC00460, and MEG3 have been found to partake in the pathogenesis of meningioma. In the current review, we describe the role of non-coding RNAs in two types of brain tumors, i.e., pituitary tumors and meningioma.


Introduction
Non-coding RNAs comprise heterogeneous types of transcripts in terms of functions, size, evolutionary conservation, and expression level. The integrated application of largescale sequencing methods and bioinformatics tools has facilitated the annotation of noncoding RNAs; thus, they are not considered as either junk portions of the genomes or byproducts of massive transcription. Soon after the completion of the Human Genome Project, several non-coding RNAs were detected in mammals [1]. With the advent of high-throughput sequencing strategies, the expression profile of non-coding RNAs has been more precisely identified [2]. The ENCODE project has stated that approximately

Pituitary Gland Tumors
Pituitary adenomas comprise a heterogeneous group of tumors historically classified based on their size into micro-and macroadenomas considering a threshold of 1 cm. More recently, immunohistochemical analyses and electron microscopy have been used to classify these tumors further [8]. These tumors can also be categorized as functional or nonfunctional, based on their hormonal activity. They might be identified either as an incidental finding in radiology or being associated with symptoms, particularly visual disturbance, based on their size, growth rate, or hormone secretion [9]. The overall prevalence of pituitary adenomas is estimated to be around 16% [10]. Recent studies have shown dysregulation of lncRNAs, miRNAs, and circRNAs in pituitary gland tumors. For instance, RPSAP52, an antisense lncRNA from the HMGA2 locus, has been found to be overexpressed in both gonadotroph-and prolactin-producing adenomas of this gland, where its expression has been correlated with the expression of HMGA2. Contrariwise, the expression of RPSAP52 has been variable among somatotroph adenomas. RPSAP52 has been shown to act as a molecular sponge for miR-15a, miR-15b, and miR-16 to increase the expression of HMGA2. In addition, RPSAP52 could enhance the expression of HMGA1. RPSAP52 has been shown to enhance cell growth through increasing G1/S transition [11]. XIST is another upregulated lncRNA in pituitary tumors whose expression is increased parallel with upregulation of bFGF and downregulation of miR-424-5p. Functionally, XIST acts as a sponge for miR-424-5p to increase bFGF levels. Silencing of XIST or bFGF or overexpression of miR-424-5p has been shown to inhibit proliferation, migration, and invasiveness of pituitary neuroendocrine tumor cells [12].

LncRNAs
Several lncRNAs have been found to act as suppressors of tumorigenesis processes in pituitary adenomas. Among these, lncRNAs is clarin 1 antisense RNA 1 (CLRN1-AS1), which is underexpressed in prolactinomas. This lncRNA has a role in the inactivation of Wnt/β-catenin signaling. Being mainly located in the cytoplasm, CLRN1-AS1 has been shown to suppress cell proliferation, promote apoptosis, and inhibit autophagy. In addition, CLRN1-AS1 sponges miR-217 to increase the expression of the dickkopf WNT signaling pathway inhibitor 1 (DKK1). The FOXP1 transcription factor could suppress the expression of CLRN1-AS1 [13]. Additionally, H19 has been found to be commonly downregulated in primary pituitary adenoma samples. Downregulation of H19 has been correlated with tumor progression. Overexpression of H19 could suppress the proliferation of pituitary adenoma cells in vitro and their growth in vivo. Functionally, H19 could regulate tumorigenesis via suppressing the activity of mTORC1 but not mTORC2. In fact, H19 blocks mTORC1-mediated phosphorylation of 4E-BP1 without changing the activity of S6K1. H19 also interacts with 4E-BP1 and inhibits 4E-BP1 binding to Raptor. Notably, H19 could suppress pituitary tumors more effectively than cabergoline [14]. Moreover, H19 has been found to have a synergistic effect with dopamine agonists in prolactinoma. From a mechanistical point of view, H19 increases the expression of ATG7 through sequestering miR-93a [15]. Table 1 shows the list of lncRNAs participating in the pathogenesis of pituitary gland tumors. Figure 1 represents the role of several ncRNAs in regulating the MAPK/ERK, PI3K/AKT, Wnt/β-Catenin, and BMP cascades in pituitary gland tumors and meningiomas.

miRNAs
Dysregulation of miRNAs is also implicated in the pathogenesis of pituitary gland tumors. While almost all related miRNAs have been assessed in a single study, the expression of miR-34a and miR-145 has been evaluated in independent studies. Notably, the results of these studies are contradictory. Yang et al. have assessed the expression of miR-34a in rat pituitary tumor cells versus normal pituitary cells. The expression of this miRNA is lower in tumor cells compared with that in normal ones. Moreover, upregulation of miR-34a could suppress the proliferation of tumor cells and promote cell apoptosis via the regulation of expression of SOX7 [37]. On the other hand, another experiment in AIPmut+ cells as a genetic model of hereditary somatotropinoma has shown upregulation of miR-34a and miR-145 in AIPmut+ compared with AIPmut− somatotropinomas. Ectopic expression of AIPmut (p.R271W) in Aip −/− mouse embryonic fibroblasts led to enhancement of miR-34a and miR-145 levels, demonstrating a pivotal correlation between AIPmut and miRNAs signature. Upregulation of miR-34a enhances proliferation, colony-forming ability, and migration of rat pituitary cells and inhibits their apoptosis. miR-145 could also affect the proliferation and apoptosis of these cells to a lesser extent. Overexpression of miR-34a enhances intracellular levels of the mitogenic factor cAMP and reduces octreotide-mediated growth hormone inhibition and antiproliferative impacts. miR-34a has been found to target the Gnai2 gene, a gene that encodes a G protein subunit suppressing cAMP synthesis. Taken together, miR-34a has been identified as a downstream target of mutant AIP that confers a cellular phenotype reflecting the aggressive manifestations of AIPmut+ acromegaly [38].
Contrary to this report, the expression of miR-145-5p has been shown to be lower in bromocriptine-resistant prolactinoma clinical specimens and cell lines compared with that in sensitive samples and cells. TPT1 has also been identified as the direct target of miR-145-5p. Forced overexpression of miR-145-5p has increased the sensitivity of prolactinoma cells to bromocriptine through enhancing the expression of TPT1. Thus, miR-145-5p has been identified as a critical modulator of drug resistance in prolactinoma [39]. Figure 1. A schematic representation of the role of several non-coding RNAs in regulating the MAPK/ERK, PI3K/AKT, Wnt/β-Catenin, and BMP signaling pathways in pituitary gland tumors and meningiomas. The figure represents the potential crosstalk between various signaling cascades modulated via several ncRNAs in triggering the development of tumor cells. WNT-signaling is a crucial part of the crosstalk between key oncogenic cascades involved in pituitary gland tumors. Elements of the WNT cascades both could be regulated through diverse pathways, including MAPK/ERK, PI3K/AKT, and BMP, as well as transcriptional regulators containing p53 and MITF [32]. In addition, an accumulation of β-catenin in the cytoplasm could, in turn, lead to its translocation to the nucleus, where it could create a complex with TCF/LEF to trigger the transcription of RUNX2, Cyclin D1, and PIT1/2. Additionally, RUNX2 could modulate the transcription of various targets, including OCN, OSX, OPN, MMP13, and ALP [33]. Furthermore, BMP receptors could phosphorylate receptor-SMADs upon ligand binding. TCF12 and TWIST1 are basic helix-loop-helix transcription factors that could play an effective role in heterodimerization and suppressing transcription downstream of the BMP cascade [34]. According to the current report, lncRNA AFAP1-AS1 could enhance growth and inhibit apoptosis in pituitary adenomas through promoting PTEN expression and suppressing the expression levels of PI3K and AKT in tumor cells [29]. Moreover, another research has denoted that LINC00460 could elevate meningioma progression and metastasis through promoting the expression levels of MMP-2, MMP-9, and ZEB1 by sponging miR-539 and thereby acting as an oncogenic RNA in the meningioma malignancy [35]. Moreover, based on recent study, lncRNA SNHG1 via targeting miR-556-5p could elevate TCF12 expression, thereby promoting tumorigenesis of meningioma through the Wnt signaling cascade. In fact, TCF12 expression was positively modulated via SNHG1, and TCF12 could, in turn, enhance transcription of SNHG1 through binding with the promoter region of SNHG1 [36]. Green arrows indicate the upregulation of target genes mod- Figure 1. A schematic representation of the role of several non-coding RNAs in regulating the MAPK/ERK, PI3K/AKT, Wnt/β-Catenin, and BMP signaling pathways in pituitary gland tumors and meningiomas. The figure represents the potential crosstalk between various signaling cascades modulated via several ncRNAs in triggering the development of tumor cells. WNT-signaling is a crucial part of the crosstalk between key oncogenic cascades involved in pituitary gland tumors. Elements of the WNT cascades both could be regulated through diverse pathways, including MAPK/ERK, PI3K/AKT, and BMP, as well as transcriptional regulators containing p53 and MITF [32]. In addition, an accumulation of β-catenin in the cytoplasm could, in turn, lead to its translocation to the nucleus, where it could create a complex with TCF/LEF to trigger the transcription of RUNX2, Cyclin D1, and PIT1/2. Additionally, RUNX2 could modulate the transcription of various targets, including OCN, OSX, OPN, MMP13, and ALP [33]. Furthermore, BMP receptors could phosphorylate receptor-SMADs upon ligand binding. TCF12 and TWIST1 are basic helix-loop-helix transcription factors that could play an effective role in heterodimerization and suppressing transcription downstream of the BMP cascade [34]. According to the current report, lncRNA AFAP1-AS1 could enhance growth and inhibit apoptosis in pituitary adenomas through promoting PTEN expression and suppressing the expression levels of PI3K and AKT in tumor cells [29]. Moreover, another research has denoted that LINC00460 could elevate meningioma progression and metastasis through promoting the expression levels of MMP-2, MMP-9, and ZEB1 by sponging miR-539 and thereby acting as an oncogenic RNA in the meningioma malignancy [35]. Moreover, based on recent study, lncRNA SNHG1 via targeting miR-556-5p could elevate TCF12 expression, thereby promoting tumorigenesis of meningioma through the Wnt signaling cascade. In fact, TCF12 expression was positively modulated via SNHG1, and TCF12 could, in turn, enhance transcription of SNHG1 through binding with the promoter region of SNHG1 [36]. Green arrows indicate the upregulation of target genes modulated via ncRNAs (miRNAs and lncRNAs); red arrows depict inhibition regulated by them. All the information regarding the role of these ncRNAs in modulating pituitary gland tumors can be seen in Tables 1-5.
The expression of miR-378 has also been found to be decreased in pituitary adenoma tissues. This miRNA has been shown to downregulate the expression of RNF31. RNF31 silencing has remarkably inhibited the proliferation and migration of GH3 pituitary adenoma cells [40]. On the other hand, the expression of miR-543 has been found to be upregulated in pituitary adenoma tissues parallel with downregulation of Smad7. Smad7 has been verified as a target gene of miR-543. Overexpression of miR-543 has enhanced the proliferation, migration, and invasiveness of HP75 cells. This miRNA also reduces the apoptosis of these cells and decreases the expression of cleaved caspases 3 and 8 [41]. Table 2 shows the role of miRNAs in pituitary gland tumors. Figure 2 illustrates the role of various ncRNAs in pituitary gland tumors through regulating the TGF-β/SMAD signaling pathway.
Cancers 2021, 13, x FOR PEER REVIEW 10 of 18 Figure 2. A schematic diagram of the role of various ncRNAs in modulating the TGF-β/SMAD signaling pathway in pituitary gland tumors. According to this cascade, it could be triggered through the binding of active TGF-β with TβRII and forming the TβRI-TβRII heteromeric complex, resulting in phosphorylation of Smad2/3, oligomerization with Smad4, and consequent nuclear translocation to modulate the transcription of ECM genes. Furthermore, Smad7 could play a remarkable role as a negative modulator of the TGF-β cascade. In addition, TGF-β has a significant part in triggering the activation of downstream signaling pathways containing MAPK, modulated by the Ras-Raf-MEK-ERK cascade, and TAK1, regulated by the TAB1 pathway. This could also lead to mediating the activation of MKK4-JNK and MKK3-p38 cascades and upregulation of AP-1 and ATF-2, respectively, and the overexpression of NF-κB to modulate profibrotic responses [61]. Previous studies have authenticated that several ncRNAs could have a significant part in regulating the TGF-β/SMAD cascade in pituitary gland tumors. As an illustration, recent literature has detected that overexpression of lnc-SNHG1 could considerably elevate the expression level of TβRII through activating TβRII/SMAD3 in invasive pituitary tumor cells via sponging miR-302/372/373/520 [31]. Furthermore, other research has indicated that upregulation of miR-93-5p could downregulate the expression level of Smad7, thereby activating the TGF-β1/Smad3 signaling-mediated fibrosis of prolactinoma cells [51]. Green arrows indicate the upregulation of target genes modulated via ncRNAs (miRNAs and lncRNAs); red arrows depict inhibition regulated by them.

circRNAs
Du et al. have assessed circRNAs signature in growth hormone-secreting pituitary adenoma using a circRNA microarray. They have reported upregulation of more than 1900 circRNAs and downregulation of about 1600 circRNAs in this type of adenoma compared with that in normal control. Ten most overexpressed circRNAs have been shown to be mainly enriched in the mTOR and the Wnt signaling pathway. Upregulation of hsa_circ_0001368 has also been confirmed by qRT-PCR. This circRNA has been found to be specifically overexpressed in this type of adenoma in correlation with the invasive Figure 2. A schematic diagram of the role of various ncRNAs in modulating the TGF-β/SMAD signaling pathway in pituitary gland tumors. According to this cascade, it could be triggered through the binding of active TGF-β with TβRII and forming the TβRI-TβRII heteromeric complex, resulting in phosphorylation of Smad2/3, oligomerization with Smad4, and consequent nuclear translocation to modulate the transcription of ECM genes. Furthermore, Smad7 could play a remarkable role as a negative modulator of the TGF-β cascade. In addition, TGF-β has a significant part in triggering the activation of downstream signaling pathways containing MAPK, modulated by the Ras-Raf-MEK-ERK cascade, and TAK1, regulated by the TAB1 pathway. This could also lead to mediating the activation of MKK4-JNK and MKK3-p38 cascades and upregulation of AP-1 and ATF-2, respectively, and the overexpression of NF-κB to modulate profibrotic responses [61]. Previous studies have authenticated that several ncRNAs could have a significant part in regulating the TGF-β/SMAD cascade in pituitary gland tumors. As an illustration, recent literature has detected that overexpression of lnc-SNHG1 could considerably elevate the expression level of TβRII through activating TβRII/SMAD3 in invasive pituitary tumor cells via sponging miR-302/372/373/520 [31]. Furthermore, other research has indicated that upregulation of miR-93-5p could downregulate the expression level of Smad7, thereby activating the TGF-β1/Smad3 signaling-mediated fibrosis of prolactinoma cells [51]. Green arrows indicate the upregulation of target genes modulated via ncRNAs (miRNAs and lncRNAs); red arrows depict inhibition regulated by them.

circRNAs
Du et al. have assessed circRNAs signature in growth hormone-secreting pituitary adenoma using a circRNA microarray. They have reported upregulation of more than 1900 circRNAs and downregulation of about 1600 circRNAs in this type of adenoma compared with that in normal control. Ten most overexpressed circRNAs have been shown to be mainly enriched in the mTOR and the Wnt signaling pathway. Upregulation of hsa_circ_0001368 has also been confirmed by qRT-PCR. This circRNA has been found to be specifically overexpressed in this type of adenoma in correlation with the invasive properties and serum levels of growth hormone. Hsa_circ_0001368 silencing suppresses proliferation, invasion, and secretion of growth hormone from primary cultured cells. Additionally, levels of hsa_circ_0001368 have been positively correlated with the pituitary-specific transcription factor Pit-1 [62]. CircOMA1 is an upregulated circRNA in nonfunctioning pituitary adenomas which sponges miR-145-5p, a miRNA that inhibits the growth of this type of tumor through targeting TPT1 [63]. Table 3 shows the role of circRNAs in pituitary adenomas.

Prognostic/Diagnostic Value of Non-Coding RNAs in Pituitary Gland Tumors
The prognostic/diagnostic value of non-coding RNAs has been assessed in pituitary gland tumors ( Table 4). Downregulation of SNHG7 [17], CCAT2 [27], and IFNG-AS1 [30] lncRNAs has been shown to increase the survival of patients with pituitary adenomas. miR-193a-3p upregulation results in a lower relapse-free survival rate [44]. On the other hand, upregulation of miR-137 expression is related to a higher recurrence-free survival rate [49]. - [27] IFNG-AS1 ENSG00000255733/12q15 20 tumors and ANTs -Higher IFNG-AS1 expression is an indicator of a lower survival rate.

Meningioma
Meningiomas are another group of brain tumors that are mostly encapsulated lesions. These benign lesions are typically associated with few types of genetic aberrations, yet their intracranial location can result in serious and possibly fatal consequences [64]. The expression of a number of lncRNAs and miRNAs has been dysregulated in meningioma.

LncRNAs
SNHG1 is an upregulated lncRNA in meningioma cell lines. SNHG1 silencing blocks cell growth and induces their apoptosis. SNHG1 could function as a sponge for miR-556-5p and enhance the expression of TCF12. In fact, the SNHG1/miR-556-5p/TCF12 axis could promote the proliferation of meningioma cells and suppress their apoptosis by enhancing the activity of Wnt signaling. Moreover, TCF12 has been shown to increase the expression of SNHG1 via binding with its promoter [36]. LINC00702 is another upregulated lncRNA in meningioma which regulates proliferation and migration of these cells via the miR-4652-3p/ZEB1 axis [65]. In addition, LINC00460 has been found to increase cell invasion and proliferation and decrease apoptosis rate through sponging miR-539 [35]. On the other hand, MEG3 is a possible tumor suppressor lncRNA in meningioma, which modulates invasive properties of these cells through the miR-29c/AKAP12 axis [66]. Table 5 shows the role of lncRNAs in meningiomas.

miRNAs
Negroni et al. have reported that the expression of the miR-497~195 cluster in meningioma is reduced with increasing tumor grade. Notably, Cyclin D1 upregulation has been correlated with a decrease in the levels of the miR-497~195 cluster. In fact, the effect of GATA binding protein 4 in enhancing the viability of meningioma cells is exerted through the regulation of expression of this miRNA cluster, which finally results in enhancement of Cyclin D1 level. Finally, serum exosome levels of miR-497 are lower in patients with high-grade meningioma compared to those with benign lesions [67].
Expression of miRNAs and mRNAs in benign and malignant meningiomas has also been assessed by RNA sequencing and miRNA microarray. This study has led to identifying upregulation of fatty acid synthase (FASN) in malignant lesions compared with benign ones. This gene has been shown to be targeted by miR-195. Overexpression of miR-195 could inhibit proliferation, migration, and invasiveness of malignant meningioma cells. Taken together, miR-195 has been verified as a tumor-suppressive miRNA in malignant meningioma by targeting FASN. NUP210, SPIRE2, SLC7A1, and DMTN are also among competing endogenous RNAs that modulate the expression of FASN through sponging miR-195 [68].
In another study, Katar et al. have shown significant enhancement in miR-21 levels with increasing grade of meningiomas, whereas there was a remarkable decrease in miR-107 levels with the increasing grade. Expressions of miR-137 and miR-29b have not been different in different histopathologic grades [69]. Table 6 shows the role of miRNAs in meningioma. Finally, a number of studies have shown associations between genetic polymorphisms of lncRNAs or miRNAs and risk of meningioma (Table 7). For instance, the rs619586 A > G polymorphism of MALAT1 has been shown to affect expression levels of MALAT1 and COL5A1, resulting in lower invasiveness of meningioma [73]. Expression of MALAT1 has been shown to be reduced in a stepwise manner with enhancing levels of miR-145 in tumor/serum specimens having AA, AG, and GG genotypes of rs619586, respectively. In addition, levels of COL5A1 have been reduced in a similar stepwise manner in relation to the rs619586 genotypes. Thus, the rs619586A > G of the MALAT1 gene can decrease the expression of this lncRNA, influencing the impact of miR-145 on COL5A1. Consistently, meningioma cells harboring the G genotype of the rs619586 had higher levels of COL5A1 [73]. Moreover, certain haplotype blocks of miR-146a, miR-149, miR-196a2 and miR-499 have been shown to be associated with risk of meningioma [74]. qRT-PCR evaluated let-7d expression. Proliferation, invasion, and viability are effectively inhibited by let-7d, whereas the apoptosis rate is elevated.
The contribution of lncRNAs in meningioma is less studied. However, similar to pituitary adenomas, lncRNAs mainly serve as sponges for miRNAs. SNHG1/miR-556-5p, LINC00702/miR-4652-3p, LINC00460/miR-539, and MEG3/miR-29c are identified lncRNA/miRNA pairs in this type of lesion. Functional effects of miRNAs sponging are diverse, ranging from the enhancement of invasiveness of tumors to induction of epithelial-mesenchymal transition in these cells.
Exosomal levels of miRNAs in serum samples of patients have been associated with the presence of brain malignancies in some cases. However, the diagnostic and prognostic applications of these exosomes have not been assessed comprehensively. In other types of cancer, exosomal levels of non-coding RNAs represent an applicable source of biomarkers for prediction of course of cancer as well as response to therapies [75].
Expression of non-coding RNAs might also affect response to therapeutic regimens [76]. For instance, H19 increases ATG7 expression and influences resistance to dopamine agonists [15].
Theoretically, the expression of non-coding RNAs can be used for molecular classification of meningiomas. However, this approach has not been implemented yet. A previous study has used the DNA methylation signature for this purpose. This approach has captured clinically more homogenous groups. Moreover, it has been proved to be superior to WHO classification in prediction of tumor recurrence and prognosis [77].
Genetic variants within non-coding regions have the potential to affect the function of these transcripts or the regulatory impacts of other transcripts on non-coding ones. A number of these variants have been demonstrated to be associated with the risk of meningioma, yet their impacts on the risk of pituitary tumors have not been revealed.
Since aberrant expressions of lncRNAs/miRNAs can affect the response of tumor cells to therapeutic modalities, it is possible that targeted therapies for modulation of expression of lncRNAs/miRNAs not only reduce the invasiveness of these tumors but also increase their response to conventional therapies.

Conclusions
Cumulatively, non-coding RNAs represent an emerging class of transcripts with putative effects on the pathogenesis of pituitary adenomas as well as meningiomas. Future studies are needed to find possible specific markers for each of these tumors to help in the identification of tumors in a less invasive manner.