MicroRNAs (miRNAs) are a group of endogenous, non-coding, 21–25 nucleotide-long RNAs. As important gene expression regulators, miRNAs are commonly known to base pair with the 3′ untranslated regions (3′-UTRs) of their downstream mRNA targets, thus driving the posttranscriptional inhibition [25,26]. Depending on the molecular context, miRNAs can exert activating or inhibitory effects on various tumourigenic processes, such as proliferation, migration, invasion and apoptosis [27,28,29,30]. To date, many miRNAs have been identified to affect glioma progression both as oncogenes (oncomirs) and tumour suppressors, including miR-139-5p [31], miR-595 [32], miR-296-5p [33] and miR-429 [34], to mention only a few. As miRNAs are potent regulators of glioma development, it is clear that they constitute a potential tumour therapeutic target.

The miRNAs that are known to affect ZEB1/ZEB2 act as tumour suppressors and are significantly downregulated in various glioma samples and cell lines. Various biological functions of ZEB miRNAs have been described (Table 1). Many miRNAs are believed to play an important role in EMT-associated events, as they target ZEBs directly or by their upstream regulators and affect other EMT-related factors. For example, miR-590-3p blocks EMT and metastasis by targeting ZEB1 and ZEB2 in GBM cell lines [35]; the miR-940/ZEB2 axis constitutes an important regulator of glioma cell aggressive phenotype, which was additionally proven in an orthotopic GBM mouse model [36]; and miR-205 suppresses the oncogenic potential of ZEB1 via the Akt/mTOR signalling pathway, which inhibits motile phenotype and EMT in GBM cells [37]. MiR-200a exerts an inhibitory effect on ZEB1 in glioma cells, as noted in a study exploring the role of glioma-related EMT in reference to the Tissue factor (TF) [38].

It is widely known that the interplay between ZEBs and miR-200 family members constitutes a double negative feedback loop and largely affects EMT-related signalling pathways [52]. The miR-200 family (miR-200s) is a tumour-suppressive group of miRNAs which consists of miR-200c/miR-141 and miR-200a/miR-200b/miR-429 clusters [53]. According to reports, miR-200s can affect ZEB1 and/or ZEB2 expression in many types of cancer, such as gastric cancer [54], non-small lung cancer [55] and oral squamous cell carcinoma [56]. In gliomas, miR-200c as well as miR-141 can inhibit glioma cell growth and migration by targeting ZEB1 [42].

Notably, the interaction between miR-200c, ZEB1 and the epidermal growth factor receptor amplification pattern seems to be particularly interesting. Epidermal growth factor receptor (EGFR) amplification is observed in about 35–70% of GBM cases [43]. As different GBM samples are highly varied in terms of EGFR alteration type, three groups of EGFR amplification status can be distinguished. They consist of GBMs with double minutes (high EGFR amplification level), GBMs with insertions on chromosome 7 (low EGFR amplification level) and GBMs with no EGFR amplification [44]. Different EGFR gene and protein alterations have been proposed as potential prognostic indicators or therapy response predictors—that said, different types of aberrant EGFR expression patterns could be used to asses individual schedules of treatment in cancer patients. However, the clinical value of EGFR is not clear yet and requires further research [8,57]. Serna et al. [44] reported that miR-200c expression level differs between primary GBM tumour samples with and without EGFR amplification, and that miR-200c and E-cadherin are downregulated in the high-level EGFR amplification group. In turn, tumours without EGFR amplification showed significantly lower ZEB1 expression [44]. Considering the subset of migration-related miRNAs, miR-200c has the most varied expression in a group of GBM patients regarding EGFR amplification status [43]. Subsequent experiments conducted with three patient-derived GBM cultures, which were varied in terms of EGFR amplification status, additionally elucidated the impact of EGFR amplification on miR-200c/ZEB1 interaction; although miR-200c overexpression exerts an inhibitory effect on ZEB1 regardless of EGFR amplification status, the expression level of ZEB1 was upregulated by miR-200c inhibition only in non-EGFR-amplifying cells, which according to authors may suggest the existence of some additional mechanism affecting ZEB1 in the context of an EGFR amplification environment [43]. As the EGF/EGFR pathway has a broad impact on various cellular processes, it states an interesting research direction also in the context of other EMT-related TFs and ncRNAs. They include not only alterations in its amplification patterns but also other EGFR gene aberrations, such as the mutational variant EGFRvIII, commonly associated with aggressive GBM cell behaviour [57].

In addition to ZEB-related double negative feedback loop and EGFR amplification status, there are also many other epigenetic mechanisms that can affect miR-200s in tumour progression [58,59,60]. One of the epigenetic regulators known as an important factor in various neoplasms (e.g., prostate cancer, colorectal cancer, breast cancer) is methyl CpG-binding protein 2 (MeCP2) [61,62,63], which belongs to the methyl CpG binding domain (MBD) family. In LN-25 and U251 glioma cell lines, knockdown of MeCP2 resulted in decreased expression of the EMT-related markers, i.e., ZEB1, ZEB2 and Twist1; a similar effect was also observed in a xenograft mouse model [45]. In addition, MeCP2 inhibited miR-200c expression in glioma cells and was negatively correlated with miR-200c expression in glioma tissues. As a mechanism, epigenetic repression caused by MeCP2 and the suppressor of variegation 39H1 (SUV39H1), resulting in miR-200c promoter repression, has been proposed [45]. As the MeCP2 has been reported in several “MeCP2-related disorders” and is considered “an essential reader of DNA methylation in the brain”, it remains a good point of interest in further research of molecular interactions in glioma pathogenesis and other CNS-related diseases [64].

The ability to self-renew, multi-potent differentiation as well as intrinsic tumourigenic potential are some of the key characteristics attributed to cancer stem cells (CSCs). It is well-known that GBM harbours a population of CSCs that are commonly known as glioblastoma stemlike cells (GSCs). GSCs largely contribute to tumour recurrence and constitute a valuable target for GBM treatment [65,66]. As CSCs are at the top of the ontogenic hierarchy not only in normal organs but also in tumours, they have been considered a valuable target for cancer therapy and recurrence for many years. That said, various strategies have been proposed in order to effectively combat GSCs, including targeting genetic and epigenetic regulation, a tumour’s microenvironment and immunotherapy-based approaches [67]. For example, valproic acid (histone deacetylase inhibitor) can exert an inhibitory effect on GSCs proliferation and motility because it induces DNA methylation changes in the Wnt/β-catenin signalling pathway and inhibits the expression of key EMT markers—Snail and Twist1 [68].

Another transcription factor which constitutes a hallmark signature of the occurrence and development of cancer is MYC. As a key epigenetic regulator, it induces stem cell phenotype, inhibits cellular senescence and blocks the differentiation of tumour cells. Thus, MYC represents a valuable target for molecular treatment in various types of human neoplasms [69,70]. Considering GSCs, MYC has been shown to possess a broad impact on a stem-related transcriptome by interacting with other transcription factors in promoter regions of many downstream genes [40]. Several studies reported that MYC inhibition can be effective in cancer treatment, due to the action of a 90-aa polypeptide Omomyc [71,72]. Accordingly, Galardi et al. [40] showed that in patient-derived GSC lines, Omomyc blocks the activity of key proteins with CSC phenotype and affects various miRNAs associated with tumour progression. For example, Omomyc increased the expression of the ZEB1-suppressing miR-200a/miR-200b/miR-429 cluster, which is well-known due to its anti-proliferative activity in cancer cells.

Notably, aggressive infiltration to adjacent tissue and increased motility of cancer cells is largely dependent on the proteolytic activity of numerous matrix metalloproteinases (MMPs), which are closely linked to EMT-associated tumour progression and can be released by tumourigenic cells themselves. MMPs are a crucial component of the tumour microenvironment in regulating angiogenesis and disseminating metastatic lesions outside the tumour zone [73]. One of the members of the A Disintegrin and Metalloproteinase (ADAM) family of zinc-dependent proteinases—ADAMDEC1 (which is secreted by GSCs)—is known to release Fibroblast growth factor 2 (FGF2) from the extracellular matrix (ECM) that stimulates FGFR1 activity. FGFR1 via ERK1/2 upregulates ZEB1, which rescues ADAMDEC1 expression from the inhibitory effect of miR-203. Hence, the pathway constitutes a positive feedback loop in the maintenance of stemness in the GSC population [48].

Many miRNAs are indicated as potential biomarkers in cancer, and the fact that they can be utilised in terms of their diagnostic and/or prognostic value is still extensively studied [39,74,75,76,77]. Among the ZEB family-related miRNAs, there are also a few examples. Downregulation of the ZEB2 suppressor miR-622 correlated with advanced WHO pathological grade and low Karnofsky performance score (KPS) in 108 tissue samples obtained from glioma patients (Pearsons’s Chi-square test). Notably, glioma specimens were varied in terms of WHO grades, comprising grade I–IV tumours (including 30 patients with GBMs) and receiving no treatment prior to tumour resection [49]. As for survival analysis, Kaplan–Meier curves also indicated a significant association between the downregulation of miR-622 expression and a low overall survival rate. To conclude, miR-622 downregulation, advanced WHO tumour grade and low KPS could be utilised as independent poor prognostic factors, as was indicated by Cox regression analysis [49]. Likewise, miR-769-3p is another ZEB2-suppressing molecule with potential prognostic value. Wang et al. [50] showed that miR-769-3p expression was significantly associated with the WHO grade and KPS both in tumour tissues and serum samples obtained from 113 primary glioma patients (58 patients with low grade gliomas and 55 patients with high grade gliomas) with no prior treatment (Chi-square test). Kaplan–Meier analysis revealed a close correlation between low miR-769-3p tumour expression level and poor overall survival; in addition, miR-769-30 tumour expression was identified as an independent prognostic factor (Cox regression analysis). Based on receiver operating characteristic (ROC) curves, serum miR-769-3p level can be utilised to distinguish between healthy individuals and glioma patients [50].

To date, temozolomide (TMZ) remains a first-choice chemotherapeutic agent in the treatment of GBMs and astrocytoma. However, patients do not respond to this alkylating agent due to a variety of factors such as MGMT methylation status, drug efflux transporters overexpression and the presence of highly resistant cancer stem cells [78]. Among them, methylation in the promoter of the MGMT gene is one of the most valuable epigenetic markers considering potential treatment response predictors. As MGMT is a DNA repair enzyme, it has an ability to remove methyl residues form nucleotides altered by alkylating drugs. Thus, the hypermethylation of its promoter confers good prognosis in TMZ therapy [5,8]. One direction to fight chemoresistance in cancer is using molecules that impede EMT and EMT-inducing signalling agents, such as those associated with tumour microenvironment. Notably, miRNAs themselves can act as therapeutic molecules due to their tumour suppressive activity [79]. In gliomas, several miRNAs with an impact on classical EMT transcription factors have been implicated as important regulators of TMZ resistance development. For example, miR-181b decreases TMZ resistance by impeding EGFR activity in GBM cells [80]. In turn, miR-128-3p can negatively affect EMT-related proteins, such as c-Met, PDGFRα, Notch1 and Slug, which enhanced glioblastoma TMZ chemosensitivity both in vitro and in vivo [81]. The decreased expression of miR-186-5p contributes to the proliferation and TMZ resistance of GBM cells due to the attenuation of its Twist1-targeting activity [82].

Although many papers point to the therapeutic potential of miRNAs, one particularly interesting study in the field of EMT TF-related molecules in gliomas is by Sadeghipour et al. [51], which aimed to identify a panel of miRNAs targeting different oncogenic pathways in GBMs; these selections could potentially sensitise cancer cells prior to exposure to chemotherapy. A review of the Gene Expression Omnibus (GEO) and The Cancer Genome Atlas (TCGA) databases identified a group of miRNAs and their target genes with essential roles in GBM pathways. Then, different combinations of these miRNAs were studied in GBM cells treated with TMZ or doxorubicin (DOX) by utilising a synthetic miRNA delivery system. Based on cell survival and apoptosis analysis, a panel of five miRNAs consisting of miRNA-138, miRNA-139, miRNA-218, miRNA-490 and miRNA-21 was assessed. The soundness of this particular group was additionally supported by the fact that downstream targets of those rationally selected miRNAs (CDK6, ZEB1, STAT3, TGIF2 and SMAD7) were related to specific, oncogenic pathways (growth and apoptosis, invasion and metastasis, cytokine signalling and stemness). Finally, the increased therapeutic effect of TMZ and DOX was confirmed in vivo, due to this rationally selected miRNAs panel [51]. This approach seems to be particularly relevant, as it reflects the complexity of oncogenic processes and highlights the significance of reciprocal interactions between specific GBM pathways.