The roles of lncRNAs in GBM, especially in cell migration and invasion, have not been extensively explored. Recent studies reported that LINC00470 sponges with miR-101, which causes high expression and epigenetic silencing of ELFN2, leading to decreased GBM cell autophagy5. In another study by the same authors demonstrated that the same lncRNA interacts with the protein FUS which activates the AKT pathway, leading to inhibited GBM cell autophagy6. This suggests that lncRNAs may operate simultaneously via different targets and mechanisms within a single biological process that can alter the cell characteristics. Studies have demonstrated that the progression of lower-grade gliomas to GBM can be regulated by the AKT pathway21. This has driven the need to validate if lncRNAs could contribute towards GBM invasion by activating cell-migration and invasion related biological pathways through their targets such as proteins and/or miRNAs.
Our microarray data indicates the expression of multiple lncRNAs is significantly upregulated in LN18 cells than in anaplastic astrocytoma cells (SW 1783) compared to normal astrocytes. Among them, LOC644656 and LOC101928837 demonstrated high difference in fold-change between LN18 and SW 1783. However, there is limited information available on LOC644656 and LOC101928837, and we could not find any involvement of both lncRNAs with cell invasion and migration. Future studies could investigate the reasons for their high expression (difference in fold-change) observed since GBM is more invasive than anaplastic astrocytoma. Studying the different upregulated lncRNAs in both cell lines could shed light on the possible biological pathways that are potentially regulated and influence the GBM characteristics. However, most of these lncRNAs are uncharacterised; hence, further bioinformatics analysis is not possible. From the list of lncRNAs identified from microarray, only one lncRNA, SRGAP3, is found in the database to be associated with cell migration and invasion. SRGAP3 is predicted to be associated with several cancers such as lung adenocarcinoma22, pancreatic ductal adenocarcinoma23, and breast cancer24. However, there are no available literatures on SRGAP3 related to GBM or molecular function that could be found. Hence, no further information could be drawn from functional analysis as most of the lncRNAs are uncharacterised.
At the current stage, most of the functions of the lncRNAs, specifically in GBM, are still undefined at both the cellular and molecular levels. There are very limited studies that have extensively studied any specific lncRNAs that can allow the use of GO-annotated lncRNA database to perform cluster analysis as compared with other types of cancers. Hence, to predict the lncRNA functions, both miRNA and proteins that interact with these lncRNAs were first obtained and subsequently clustered to roughly predict the specific lncRNA's function based on the miRNA or proteins they are reported to interact with, in the processes relevant to GBM migration and invasion. Angiogenesis is a critical event in the progression of GBM since the tumours exhibits a high degree of vascular proliferation and endothelial cell hyperplasia25. Endothelial cells-associated with angiogenesis are among the critical inducers of cell invasion26. On the other hand, hypoxia can trigger the invasive phenotype of GBM by upregulating the levels of invasion proteins that drive the degradation and remodelling of the extracellular matrix and EMT in GBM27.
Four lncRNAs, LINC00221, LINC00265, LINC001564, and LOC100240735, were predicted to be associated with these specific processes. Among them, LINC00221 promotes cisplatin resistance in non-small-cell lung cancer, leading to non-effective chemotherapy28. Additionally, LINC00221 suppresses the malignancy in acute lymphoblastic leukaemia (ALL) where its overexpression leads to an anti-proliferative and pro-apoptotic effect in ALL cells29. Within the context of GBM, the literature regarding these four lncRNAs is very scarce. This raises the need to investigate the role of these lncRNAs in GBM. Among the lncRNAs predicted based on miRNA functions, CHKB-AS1 has been reported as a potential oncogenic lncRNA30 while LINC00482 regulates angiogenesis in bladder cancer through FOXA131. LINC00239 promotes cell migration and invasion through regulation of PI3K/Akt pathway and enhancing KLF12 expression via sponging with mi-484 in acute myeloid leukaemia and colorectal cancer, respectively32,33. LINC1003 predicted to be involved in the EMT process in this study, was previously reported as oncogene and tumour-suppressor in colorectal cancer and multiple myeloma34,35.
CRNDE has been reported to promote malignancy in GBM through miR-384/PIWIL4/STAT3 axis36. The analysis performed in our study also shows that CRNDE can act on miR-145-5p, which subsequently regulates tumour suppressor BTG1. Studies demonstrated that PUM2 could promote GBM cell migration through repression of BTG137. This can be further validated if the lncRNA CRNDE could play any role in this interaction. Our data indicates that NEAT1 is significantly upregulated in GBM cells than in astrocytoma, with a difference in fold-change than normal astrocytes at 2.95 times. In GBM, NEAT1 promotes its progression through the WNT/β-catenin pathway38, and upregulation of SOX2 expression39. Our analysis predicts that NEAT1 affects SMAD4 and LOXL2, which were both established to be upregulated in GBM40,41. The miRNA miR124 was predicted to interact with NEAT1, and will subsequently target LOXL2, which also has been reported to target SMAD442. SMAD4 is an important mediator for transforming growth factor β (TGFβ) signalling, where its abnormal activity often leads to tumourigenesis43. Similarly, the current data predicts NEAT1 to regulate CITED1 and SMAD4, which is part of TGFβ-signalling. Our prediction analysis suggests that both lncRNAs and miRNAs interaction is essential as they may play significant roles in regulating cell migration and invasion, where multiple proteins/mRNA involved are targeted simultaneously. Several studies have demonstrated SNHG1 contributes to GBM malignancy through several mechanisms. SNHG1 binds with miRNA-154-5p, affecting FOXP2, leading to enhanced expression of oncogene KDM5B44. Another study shows that SNHG1 upregulation promotes glioma progression through sponging with miR-194 that regulates PHLDA145. Recently, SNHG1 is shown to regulate PI3K/AKT pathway and miR140, leading to GBM malignant progression46. In this study, we predict that SNHG1 in the regulation of FBN2 expression through miR101. FBN2 overexpression is recently studied in lung cancer tissues, contributing to the proliferation, invasion, and migration abilities of lung cancer cells47.
The complete lncRNA-miRNA-protein network were only able to be established from the lncRNAs predicted based on protein function. This is because there are various lncRNA-miRNA interactions and, subsequently, miRNA-mRNA interactions in the bioinformatics database. Establishing the complete network for the upregulated lncRNAs identified from microarray would be a huge challenge as there are millions of miRNA-mRNA associated interactions with these lncRNA. Thus, to perform network analysis on such a huge dataset would be impossible to visualise for a specific biological axis. In this predictive analysis, there will be loss of potential lncRNA candidates as predictions are based on established interactions from other cancer studies compiled into the bioinformatic databases. Thus, where lncRNA-miRNA-protein/mRNA interactions are concerned, only lncRNA which was shown in both lncRNA-miRNA and lncRNA-protein/mRNA interaction databases, will be shown within the study output. LncRNAs interactions reported in only one of either network will not yield results in the complete network analysis. In this study, our prediction of the lncRNAs in GBM migration and invasion is limited to lncRNA and based on the available experimental data in other types of cancers. Additionally, the prediction depends on the new available literature and the frequency the database is updated. Although our protocol prediction study may identify novel interaction/axis, with the limitation of information, further novel interactions between lncRNAs and their target with no prior studies cannot be predicted.