Year: 2017
Type: article
Abstract: As the most common primary spinal tumor and despite responses to a few targeted agents,1,2 schwannomas remain poorly understood with regard to their genetic underpinnings. We do know that early diagnosis and treatment is key to preventing the long-term morbidity that occurs with a tumor intrinsic to the cranial and spinal nerves. Complete resections are often difficult and progression of residual or radiation-treated schwannomas challenging. A recent study by Agnihotri and colleagues applied whole-exome sequencing, RNA sequencing and methylation profiling to 125 spinal and cranial schwannomas.3 Not surprisingly, they identified LZTR1 mutations4 and NF2 as a frequently altered gene within this cohort as has been previously reported. In fact, loss of 22q was the only recurrent chromosomal aberration in this tumor, with a slightly higher percentage among cranial cases. DNA-methylation profiling further solidified the presence of two stable anatomically driven subgroups, with group 1 comprised mainly of vestibular schwannomas and group 2 of spinal cases. Recurrent, low-frequency mutations not previously reported across both groups included ARID1A, ARID1B, DDR1, TAB3, ALKP2, CAST, TSC1 and TSC2. Interestingly, 15% of schwannomas harbored mutations in either of the latter two, TSC1 or TSC2, identifying a potential subgroup of patients that may benefit from mTOR inhibitors. Furthermore, pathway analysis of non-silent mutations showed that they were enriched differentially among these groups for immune response, MEK signaling, and epithelial to mesenchymal transition, among others. RNA sequencing for recurrent gene fusions identified a novel, in-frame fusion involving SH3PXD2A and HTRA1 in a subset of patients, further confirmed by RT-PCR. Expression of this fusion promoted cell invasion and resistance to anoikis, unlike single or combined loss of wild-type SH3PXD2A and HTRA1. Confirmatory studies of injection of fusion-containing cells into the flank of NOD-SCID mice showed a higher penetrance of tumor formation compared with those without the fusion. The authors also examined fusion-positive human tumors, which were significantly larger than their fusion-negative counterparts, with evidence of bone erosion and remodeling among spinal cases, highlighting their potentially more aggressive nature. Gene expression profiling of fusion-positive human Schwann cells compared with controls also revealed enrichment for epithelial-to-mesenchymal transition as noted above and IL2-STAT5 signaling, the latter supporting activation of pMEK-pERK signaling and offering the possibility of therapeutic benefit from MEK and/or ERK inhibitors. Treatment of fusion-positive cells with the MEK1/2 inhibitor trametinib did, in fact, halt pERK expression, increase apoptosis, and reduce cell growth. The discovery of this fusion, whose effect on cell growth is not recapitulated by combined loss of both SH3PXD2A and HTRA1, raises questions as to the functional consequences of losing both fusion partners. The answers should come in the near future, opening other potential avenues for new therapeutic strategies in schwannomas and a better understanding of the underlying biology that drives these tumors. Glioblastoma (GBM) is thought to contain so-called glioblastoma stem cells (GSC) that are associated with tumor initiation, progression, recurrence, and resistance to therapy. Previous studies have shown that GSCs can differentiate into glial and neuronal-like cells as well as transdifferentiate into endothelial cells (ECs); however, the mechanism of transdifferentiation and its contribution to GBM malignancy are not well understood. A recently published report by Hu et al. uncovered a new mechanism of GSC transdifferentiation into ECs that involves WNT5 epigenetic activation and elucidated the role of this transdifferentiation in tumor progression and recurrence. The authors generated Myc-immortalized human neural stem cells (hNSC) transfected with dominant negative p53 and constitutively active AKT and showed that these cells (designated iGSC) exhibit GSC-like characteristics and are highly tumorigenic in mice, where they form GBM-like tumors. Transcriptomic comparison of hNSC and iGSC showed upregulation of EC pathways without an increase in hematopoietic stem cell numbers, suggesting a transdifferentiation of iGSCs into ECs. This transdifferentiation was experimentally verified and shown to be primarily driven by AKT activation using multiple complementary approaches. By intersecting AKT-regulated genes with genes displaying histone modifications associated with cell fate determination and subsequent experimental verifications, they identified WNT5A as the pivotal mediator of AKT-driven EC transdifferentiation. Subsequently, using epigenetic analyses and experimentation, the authors found that WNT5A is regulated by an AKT-activated epigenetic switch that involves opposing roles of the transcription factors PAX6 and DLX5. To determine whether EC transdifferentiation plays a functional role in gliomagenesis in vivo, they employed a patient-derived GSC orthotopic tumor model. Overexpression of WNT5A in these GSCs generated rapidly growing tumors with short latency, high vascular density, and highly invasive peritumoral satellite lesions containing transdifferentiated ECs. The transdifferentiated ECs appeared to recruit host ECs to form the peritumoral lesions, an observation that was subsequently confirmed experimentally and shown to be mediated by WNT5A. To ascertain the clinical relevance of their findings, the authors then investigated whether the above findings are operative in human GBM specimens. Using markers and genetic signatures, they demonstrated the presence of transdifferentiated ECs in human GBM as well as their vicinity to host ECs and their positive association with WNT5A expression. The transdifferentiated ECs were enriched in peritumoral satellite regions and associated with shorter patient survival. Finally, using paired primary/recurrent GBM tumors, they found that recurrent tumors have higher levels of WNT5A and greater numbers of transdifferentiated ECs than their primary counterparts, thus implicating EC transdifferentiation in human GBM recurrence. Altogether, this study showed that epigenetic activation of WNT5A expression contributes to glioblastoma tumor progression and recurrence by promoting differentiation of glioblastoma stem cells into endothelial cells. Consequently, the authors propose inhibition of WNT5A as a promising new GBM therapy. Gliomas are heterogeneous tumors, with both genetic and cellular determinants accounting for their heterogeneity. Cancer stem cells, in part, explain the intra-tumoral heterogeneity in that a hierarchy is dictated by stem/progenitor populations at the apex, which give rise to more differentiated cell types similar to normal stem cells. Oligodendroglioma is a low-grade glioma that is predominantly characterized by mutations in the isocitrate dehydrogenase 1 or 2 (IDH1 and 2) and co-deletion of chromosomes 1p and 19q, but well-established models to define cancer stem cell hierarchy in these tumors are lacking. In a recent publication, Tirosh et al1 use RNA-sequencing of freshly resected oligodendrogliomas to address the cellular architecture at the single-cell level. After separating normal from tumor cells using copy number variation (CNV) estimates, they found that almost all cells (~4000) from six tumors showed similar genetic alterations within the tumor. Two tumors, however, contained sub-clones with differing CNVs. They show that irrespective of the genetic alterations, all cells segregated into two prominent groups in principle component analyses (PCA). One group of cells was enriched for oligodendrocytic and the other enriched for astrocytic markers, implying that these populations were differentiated progeny. Interestingly, a small subset of the population with an intermediate PC1 score overexpressed genes involved in stemness and showed similarities to pre-natal and fetal human brain. Given the previous studies showing a progenitor cell-like origin for oligodendroglioma2,3, the authors, however, did not observe similarities of the stemness group with oligodendrocyte progenitor cell signatures; rather, this group was similar to neural stem cells. The cells with higher stemness scores also had higher expression of cells involved in the G1/S, G2/M phases of the cell cycle and expressed CCND2, a marker highly expressed in cycling cells. In depth clonal studies from the RNA-seq data also showed that genetic subclones spanned all compartments of cell populations (stem and differentiated). This study provides a genome-wide transcriptional profile of the cellular architecture of oligodendroglioma and provides a compelling argument that cancer stem cells may play a role in the biology of this tumor type. To establish this concept, future efforts must be directed toward prospective identification and propagation of cancer stem cells from oligodendroglioma. Malignant gliomas arise from tumor-initiating cells that share features of somatic stem cells such as sustained self-renewal1. Adult neural stem cells (NSCs), recognized as a major cell of origin of gliomas, retain stemness while exposed to persistent signals in their niche of origin2 but differentiate into neural lineages when they migrate away from this niche. This raises the question of how glioma stem cells (GSCs) can retain stemness when they invade brain tissue away from their germinal niche(s). To address this question Shinghu et al. have focused on QKI, a candidate tumor suppressor that affects GSC stemness3. QKI (“Quaking”) is a multifunctional RNA-binding protein highly expressed in NSCs that is involved in normal myelination. By crossing transgenic models driven by Nestin-CreERT2 the authors conditionally deleted QKI together with the tumor suppressors Pten and Trp53 in postnatal, Nestin-positive NSCs and followed the fate of the resulting pre-malignant stem cells. Interestingly, deletion of Pten and Trp53 in QKI+/+ animals was insufficient to sustain stemness of pre-malignant cells when they migrated away from the subventricular zone (SVZ) and, accordingly, no gliomas formed in this model. Deletion of QKI in Pten+/+Trp53+/+ mice also failed to induce glioma formation, but allowed NSCs to retain self-renewal and an undifferentiated state after migrating away from the SVZ. When QKI was finally co-deleted with Pten and Trp53 in NSCs the result was a highly reproducible glioma model with >90% penetrance and a median survival of 105 days after tamoxifen-induced gene deletion. Mechanistically, loss of QKI decreased endocytosis and degradation of cell membrane receptors. Therefore, receptors such as EGFR, Notch1, and Frizzled, which respond to extracellular signals that instruct stem cell behavior, remained active in the membrane of pre-malignant (QKI-/-Pten-/-Trp53-/-) NSCs. This sustained receptor signaling kept the cells undifferentiated, despite suboptimal levels of pro-stemness signals away from the original NSC niche. In agreement with these findings, QKI expression inversely correlated with levels of several cell-surface receptors in clinical specimens and with patient survival. This study uncovers QKI deletion as a major event that promotes stemness in GSCs by reducing the recycling of cell-surface receptors and making the cells “hyper-sensitive” to pro-stemness cues. The results suggest that glioma stemness is highly dependent on this sensing activity, opening a new area for potential therapeutic targeting. The IDH-1 mutation results in the production of the oncometabolite 2-HG, which appears to initiate a cascade of metabolic and intracellular signaling changes leading to tumor phenotypes that can be distinct from IDH-1 wildtype glioma. Work to characterize the underlying tumor biology that separates these entities is ongoing, and the potential global differences are only starting to be carefully characterized. Invasiveness is a key driver of malignancy in gliomas, resulting in high rates of morbidity and mortality. Yet invasiveness can be difficult to assess histopathologically due to under-sampling errors and also limitations on the extent of surgical resection of brain parenchyma adjacent to the main tumor mass. One way imaging can help provide global surveillance of tumor invasion is through the use of diffusion tensor imaging (DTI). White matter tracks show preferential diffusion parallel to neural fibers; so-called directional or anisotropic diffusion. This diffusion property is degraded by tumor infiltration so that water movement becomes less directional, i.e., more isotropic, in areas of bulk tumor invasion. In a recent publication, Price et al1 leveraged DTI analysis to compare invasiveness in IDH-1 mutant versus wildtype glioblastoma. Invasion was determined using previously published criteria based on plots of anisotropic versus isotropic diffusion within regions of interest in the tumor-adjacent white matter. Of the cohort of 70 glioblastoma patients, 9 were IDH-1 mutant and 61 were wildtype. The authors report that all patients with the IDH-1 mutation were minimally invasive by DTI. For the wildtype tumors, only 8% were minimally invasive, with the rest either diffusively (69%) or locally (23%) invasive. Although the study was limited by a small number of IDH-1 mutants, the results support the hypothesis that there are global phenotypic distinctions between the mutants and their IDH-1 wildtype counterparts. The reduced invasion of IDH-1 mutant glioblastoma may help explain why these tumors have better outcomes, particularly after complete surgical resection2. The authors also note that this technique could provide a better target for tumor removal, one delineated by the region of isotropic diffusion rather than the contrast-enhancing lesion alone. The potential survival benefits of such an approach remain to be determined.
Source: Neuro-Oncology
Authors: Karen Wright, Roger Abounader, Krishna Bhat, Mariano S. Viapiano, Whitney Pope, Kenneth Aldape
Language: English
Subfield: Neurology
Field: Medicine
Domain: Health Sciences
Open Access status: bronze
OpenAlex ID: W4237382647