Adult-type diffuse gliomas are composed of IDH-mutant and 1p/19q-codeleted oligodendrogliomas, IDH-mutant astrocytomas, and IDH-wildtype GBMs, based on histological characteristics and explicit molecular markers. In adults, oligodendrogliomas with IDH mutation and 1p/19q-codeletion also present with mutations in the TERT promoter gene (Lee et al., 2018). Oligodendroglial tumors have generated great interest over the past decade due to its favorable response to chemotherapy (Engelhard et al., 2003; Reifenberger and Louis, 2003; Van den Bent et al., 2008) which may be attributed to the concurrent loss of chromosome arms 1p and 19q (Cairncross et al., 1998; Smith et al., 2000; Sasaki et al., 2002). IDH-mutant astrocytomas are now graded as WHO grade II, III, or IV (Louis et al., 2021) and also harbor ATRX and TP53 mutations (Marker et al., 2021). GBM is the most common and deadliest primary brain tumor. IDH-wildtype GBM demonstrates alterations in epidermal growth factor receptor (EGFR), and similar to oligodendrogliomas, exhibit TERT promoter mutations (Galbraith et al., 2020).

Because oligodendrogliomas arise from oligodendrocytes, it is not surprising that attempts to diagnose oligodendrogliomas have utilized oligodendrocyte markers. Mature oligodendrocyte markers, such as myelin basic protein (MBP) and proteolipid protein (PLP), however, are not expressed at detectable levels in oligodendrogliomas (Sung et al., 1996; Popko et al., 2002). Furthermore, immature oligodendrocyte markers, such as the chondroitin sulphate proteoglycan NG2 and platelet-derived growth factor receptor alpha (PDGFR-α), lack specificity (Popko et al., 2002) and have been unsuccessful in discerning between glioma types (Shoshan et al., 1999; Marie et al., 2001). Several earlier studies have observed marked Olig2 expression in oligodendrogliomas (Lu et al., 2001; Marie et al., 2001; Yokoo et al., 2004). Specifically, anaplastic oligodendrogliomas displayed intense nuclear Olig2 expression (Ohnishi et al., 2003). Morphologically, Olig positive cells were moderately to densely packed, and displayed round and homogeneous nuclei with perinuclear halos (Lu et al., 2001), characteristics consistent with oligodendroglial tumors (Kim et al., 2005). Others have also observed an upregulation of both Olig1 and Olig2 in these tumors (Ohnishi et al., 2003; Aguirre-Cruz et al., 2004; Riemenschneider et al., 2004). For example, one study found an astounding 87% (26/30) and 93% (28/30) of oligodendroglial samples were positive for Olig1 and Olig2, respectively (Aguirre-Cruz et al., 2004). Furthermore, the strong expression of Olig1 and Olig2 was shown to be correlated to WHO classification with their expression increasing incrementally from grades I to III (Ohnishi et al., 2003). However, one report did note varied expression of Olig1 and Olig2. Here, the authors found 3 grade III oligodendrogliomas did not express either Olig1 or Olig2 while another 3 grade III oligodendrogliomas expressed Olig1 only (Bouvier et al., 2003).

Compared to oligodendrogliomas, Olig expression in astrocytomas and GBMs has been inconsistent and varied. Generally, low levels of Olig1 and Olig2 have been observed (Ohnishi et al., 2003; Riemenschneider et al., 2004; Tanaka et al., 2008) with weak Olig2 intensity in the nuclei (Ohnishi et al., 2003). In one report, low Olig1 expression was detected along with a marked upregulation of Olig2 (Riemenschneider et al., 2004), while another study found an upregulation of both Olig1 and Olig2, although the sample size was small (4 cases of diffuse astrocytomas; Aguirre-Cruz et al., 2004). In another study astrocytomas were found to exhibit only weak or moderate Olig expression (Lu et al., 2001). Olig expression was not detected in a case of grade III astrocytoma (Bouvier et al., 2003). GBMs also displayed varying Olig2 expression. While one study rarely observed Olig2 in GBM (Ohnishi et al., 2003), another study demonstrated lower mean transcript levels of Olig1 and Olig2 (Riemenschneider et al., 2004). In one rare case of GBM, upregulation of both Olig1 and Olig2 were observed (Aguirre-Cruz et al., 2004). Interestingly, in a separate study, Olig2 protein levels were upregulated in all cases of GBM and appeared nuclear (Ligon et al., 2004).

While diffuse high-grade gliomas (HGGs) are more common in adults, pediatric HGGs present with similar histopathological features and devastating prognosis (Jones and Baker, 2014; Wu et al., 2014). Pediatric diffuse HGGs can arise from various regions in the brain but most develop as diffuse intrinsic pontine glioma (DIPG) which occurs in the brainstem (Jones and Baker, 2014) during a restricted window of childhood (median age ~ 7 years; Hennika and Becher, 2016). DIPGs are the most common brainstem tumors in children with a median of survival of less than 1 year from diagnosis (Warren, 2012). Histopathologically, DIPG hosts a spectrum of features that is consistent with diffuse and anaplastic astrocytomas and GBMs (Buczkowicz et al., 2014). Because DIPG appears during development, neural stem cells (NSCs) and neural progenitor cells (NPCs), which are actively proliferating and differentiating, are highly impacted during disease progression (Anderson et al., 2017). Olig proteins are critical players in cellular specification and differentiation during development (Szu et al., 2021). Their expression in DIPGs have been investigated. Not surprising, a large number of cells in the pons were found to be positive for Olig2 with a subset of these cells also co-expressing Sox2 and Nestin (Monje et al., 2011; Ballester et al., 2013), markers of not only CNS embryogenesis (Vinci et al., 2016), but also tumorigenesis (Boumahdi et al., 2014; Neradil and Veselska, 2015).

Circumscribed astrocytic gliomas are astrocytic neoplasms with circumscribed growth (Riemenschneider et al., 2004). Pilocytic astrocytoma (PA) is a type of circumscribed astrocytic glioma and is considered a low-grade glioma (LGG). PAs occur mostly in children and young adolescent but can be observed in older patients as well. This brain tumor is commonly observed in the cerebellum, spinal cord, and optic pathways, but can occur anywhere in the brain (Riemenschneider et al., 2004; Ferris et al., 2017). Histologically, PAs display the classical biphasic pattern which is composed of compact areas containing Rosenthal fibers and loose microcystic areas (Ceppa et al., 2007).

Olig expression in PAs have been conflicting. Some studies have found low to moderate expression of Olig1 and Olig2 (Lu et al., 2001; Ohnishi et al., 2003) while others have reported high expression of these genes (Bouvier et al., 2003; Tanaka et al., 2008; Otero et al., 2011). One study observed greater immunoreactivity of Olig1 (97%; 62/64) compared to Olig2 (75%; 48/64; Takei et al., 2008). Diffuse staining patterns of Olig2 were observed (Otero et al., 2011) and similar to oligodendrogliomas, Olig immunoreactivity was found localized to the nuclei (Takei et al., 2008; Tanaka et al., 2008). Interestingly, double immunolabeling of Ki67 and Olig2 showed that most proliferating cells were also positive for Olig2, however, Ki67⁺ cells embodied a small portion of Olig2 expressing cells as PAs are LGGs and have a low rate of proliferation (Tanaka et al., 2008; Otero et al., 2011).

Glioneuronal tumors (GNTs) are exceptionally rare neoplasms composed of both mixed neuronal and glial cells. The majority of GNTs are classified as grade I and are associated with seizures (Gatto et al., 2020; Krauze, 2021). The pathological aspects of GNTs remain unclear however, case reports have found Olig2 commonly expressed in these tumors and thus lean more toward oligodendrogliomas. Three subtypes of GNTs that demonstrate Olig2 upregulation are dysembryoplastic neuroepithelial tumors (DNTs; Komori and Arai, 2013; Matsumura et al., 2013), papillary glioneuronal tumors (PGNTs; Tanaka et al., 2005; Chen et al., 2006; Gelpi et al., 2007; Iżycka-Świeszewska et al., 2008; Matsumura et al., 2013), and rosette-forming glioneuronal tumors (Wang et al., 2009; Luan et al., 2010; Xiong et al., 2012; Matsumura et al., 2014).

DNTs are highly heterogenous with varying morphological features. Histologically, these tumors display nuclear atypia, mitosis, endothelial proliferation, or increased cell density, however, these appearances provide no prognostic value (Daumas-Duport et al., 1999). DNTs are also subtyped as simple or complex which displays oligodendroglia-like cells (OLCs) and floating neurons (Suh, 2015). With these hallmarks, the definition of DNTs remain controversial. DNTs were found to be more similar to oligodendrogliomas rather than a glioneuronal tumor. In this same study 88% of OLCs were diffusely Olig2⁺ and 10% of these cells also colocalized with galectin3 in the nuclei of OLCs. Few OLCs were positive for PDGFRα and did not exhibit 1p/19q codeletion. Additionally, NeuN⁺ and Olig2⁺ cells were mutually exclusive, further suggesting that DNTs are clear glial tumors rather than glioneuronal tumors (Komori and Arai, 2013).

Similarly, Olig2 expressing cells were also found in PGNTs suggesting that these tumors may be oligodendroglial or oligodendroglial-like. Histologically, PGNTs exhibit two distinct architectures: (1) pseudopapillary structures surrounded by (2) compact regions consisting of neuronal elements under different maturation stages (Tanaka et al., 2005; Gelpi et al., 2007; Iżycka-Świeszewska et al., 2008; Marumo et al., 2013).

Throughout CNS development, NSCs and NPCs transform into distinct cell types in a spatiotemporal manner. A central function of Olig2 is to direct cell fate and specification, particularly into oligodendrocytes and neurons, in distinct regions of the brain and spinal cord during development (Szu et al., 2021). Olig2 is induced by Sonic hedgehog (Shh; Ortega et al., 2013) where its pathways are known to regulate cellular patterning and cell fates (Dessaud et al., 2008). The interplay between Shh and fibroblast growth factor (FGF) promotes Olig2 transcription (Tsigelny et al., 2016; Farreny et al., 2018). Increasing evidence has associated Shh signaling pathway with CNS tumors, however its relationship with Olig2 in gliomas is only beginning to be elucidated.

Several lines of evidence have associated Shh signaling with gliomas. For example, overexpression of Shh was observed in CD133⁺ cells and accelerated tumor growth while inhibition of Shh or shRNA knockdown of Shh delayed tumor growth and downregulated Ptch1 and Gli1 (Hung et al., 2020). Shh is activated via binding to the Ptch1 receptor while Gli1 is transcriptionally induced by Shh signaling (Cohen et al., 2015). Aberrant activation of Gli1 (Avery et al., 2021) and mutations in Ptch1 (Wang et al., 2019) are correlated with various cancers. In another study, expression of Shh and Ptch1 levels were significantly higher in brainstem astrocytomas compared to supratentorial astrocytomas (Yu et al., 2011). Increased levels of Notch receptors and its ligands were observed in astrogliomas and GBMs. Interestingly, glioma cell lines expressing the active form of Notch1 proliferated faster than those that did not (Zhang et al., 2008). Furthermore, overexpression of Notch1 further increased formation of Nestin⁺ neurosphere colonies (Zhang et al., 2008) and its expression in GBM cells (Shih and Holland, 2006). Similarly, overexpression of Notch1, its ligands, and downstream targets (Hes1 and Hes2) have been detected in GBM. Notch activation has also been shown to contribute to Ras-mediated transformation of glial cells to glioma growth, proliferation, and survival (Kanamori et al., 2007).

Because Olig2 activity is regulated by Shh (Ortega et al., 2013), it is plausible that increased levels of Olig2 in gliomas are contributed by Shh deregulation. Recently, Olig2 was shown to behave as an oncogenic activator in Shh medulloblastoma (Shh-MB; Zhang et al., 2019), a malignant pediatric brain tumor characterized by activation of Shh signaling (Skowron et al., 2021). Olig2⁺ progenitors were identified as the rapidly dividing Type C cells at the onset of tumorigenesis. Surprisingly, a substantial increase in Olig2⁺ progenitors was found in recurrent Shh-MB indicating that Olig2⁺ progenitors are reactivated during recurrence or metastasis. Finally, enhanced Olig2⁺ expression was also detected in Shh-MB and was significantly correlated with decreased survival.

Studies have also illustrated Olig2 participation in positive feedback loops with the EGFR receptor tyrosine kinase (RTK; Kupp et al., 2016; Tsigelny et al., 2016). EGFR signaling is known to activate the oncogenic PI3K-AKT–mTOR and RAS–RAF–MEK–ERK pathways (Asati et al., 2016). Exposure to EGF leads to proliferation of Olig2⁺ type C cells (Hack et al., 2005; Menn et al., 2006; Ligon et al., 2007) and inhibition of EGFR signaling results in Olig2 depletion indicating that EGFR signaling is responsible for sustaining Olig2 expression in progenitor cells (Kupp et al., 2016). Furthermore, Olig2 directly targets EGFR (Meijer et al., 2014; Mateo et al., 2015) and overexpression of Olig2 leads to significant upregulation of EGFR and transcripts (Kupp et al., 2016). Additionally, phosphorylated Olig2 leads to differentially regulated genes associated with RTKs (Meijer et al., 2014; Kupp et al., 2016).

Gene network analysis has identified potential roles of Olig2 involvement in gliomas (Figures 1, 2). One such network entails cell cycle regulation (Tsigelny et al., 2016). p53 is a tumor suppressor gene that functions in growth arrest and apoptosis in response to cellular stress. An effector of p53 and cell cycle inhibitor is p21 (Haupt et al., 1997). Chromatin immunoprecipitation (ChIP) analysis demonstrated that p21 is a direct target of Olig2 repression in NPCs and gliomas (Ligon et al., 2007). Malignant gliomas that are resistant to radiation and genotoxic drugs are associated with reduced p53 functions as a result of Olig2 expression. However, in the absence of Olig2, even attenuated p53 functions were shown to be sufficient to activate radiation-induced apoptosis and growth arrest. Olig2 opposes p53 functions by suppressing acetylation of p53. Therefore, Olig2 acts as post-translational modifier of p53 to repress its downstream biological activities (Mehta et al., 2011).