Year: 2020
Type: article
Abstract: The Creighton and Deneen labs recently developed a mouse model of glioblastoma involving in utero CRISPR/Cas9-mediated knockout of three known tumor suppressors Nf1, Trp53, and Pten (termed 3xCr).1 Here, Yu et al. modified this mouse model to screen for driver mutations in phosphatidylinositol 3-kinases (PI3Ks). Phosphatidylinositol (3,4,5)-trisphosphate [PtdIns(3,4,5)P3] is produced in the inner leaflet of the plasma membrane when Class I PI3Ks phosphorylate the 3’-position hydroxyl group of the D-myo-inositol head group.2 Proteins with plekstrin homology (PH) domains are recruited to the plasma membrane under such circumstances, such as PDK1 and AKT, and activate a number of cell signaling pathways, including those regulating survival and proliferation. One of the catalytic subunits of this class of PI3Ks, PIK3CA, is commonly mutated in glioblastoma. PTEN is a phosphatase that converts PtdIns(3,4,5)P3 back into PtdIns(4,5)P2. PTEN function is also frequently lost in glioblastoma. As such, Yu et al. left Pten function intact in their targeting strategy, but instead combined the loss of the two remaining tumor suppressors (2xCr) with overexpression of a panel of Pik3ca mutations observed in glioblastoma.3 This ingenious strategy allowed them to identify strong driver mutations of Pik3ca such as C420R and H1047R, as well as other less-active, albeit oncogenic ones such as E545A/K, M1043I, and R88Q. While Pik3ca C420R transcriptionally upregulated proliferation-associated genes and functionally enhanced proliferation, it, as well as H1047R, also affected synapse-associated genes and hyperexcitability in electroencephalography (EEG) tests. This appears to be the result of a net increase in excitatory synapses and loss of inhibitory synapses. Seizures are common in brain tumors, including in glioblastoma, and thus this new mouse model appears to reflect an important characteristic of the disease. The mechanism remains to be fully elucidated, but Pik3ca H1047R-expressing astrocytes induced this phenotype in a cell-autonomous manner. In contrast, C420R expression only resulted in hyperexcitability in the context of tumors. A key molecule in the C420R mouse model appears to be a proteoglycans GPC3, another usual suspect in cancer4 and a member of a well-studied group of glycoproteins in brain tumors.5 Heretofore, mouse models of glioblastoma have often modeled loss of tumor suppressors.6–8 The approach by Yu et al. should enable improved in vivo modeling of gain-of-function or oncogenic drivers of glioblastoma, such as EGFRvIII. One understudied aspect of PI3K biology has been its protein kinase activity,9 rather poignantly described in one review as ‘the protein kinase that time forgot’.10 Perhaps a screen of glioblastoma-associated PI3K mutations in the 3xCr mice might uncover this forgotten function, if any, in cancer. Even PTEN is endowed with phosphatase activity-independent scaffolding functions,11 and expression of such a version of Pten in the 2xCr background might be intriguing. Apart from candidate-based approach, Yu et al. demonstrate the use of pooled, individually bar-coded alleles that could be identified after successful tumor development – a powerful screening method that could lead to the discovery of a plethora of novel oncogenes in glioblastoma. Oncogenes are tractable targets as they can be disabled by inactivation, whereas lost tumor suppressors would require reconstitution. The pioneering work of Yu et al. might spur target identification and drug discovery in this difficult disease. The recent study by Ryall et al1 details a comprehensive molecular and clinical analysis of 1000 pediatric (<19 years) low-grade glioma (pLGG). pLGG represent the most common brain tumors in children (~30%) and comprise various histologic entities of glial, neuronal, and mixed glioneural as defined by the WHO classification. In contrast to LGG in adults, pLGG are located throughout the brain and rarely progress to higher malignancy. Nevertheless, there are large differences in prognosis and response to therapy. A major finding of the molecular characterization revealed two types of tumors: those characterized by gene rearrangements associated with lower risk, and those with single nucleotide variations (SNVs) associated with higher risk. The pLGG with rearranged drivers were more common in younger children and enriched for grade I histology, whereas the SNVs were more likely in the older children, an observation that may hint at differences in the biological evolution, according to the authors. However, in most cases the alterations of either type point towards an activation of the RAS-mitogen-activated protein kinase (RAS/MAPK) pathway. In eighty-four percent of the tumors a driver mutation was identified. The most common alterations comprised somatic alterations of BRAF (KIAA1549-BRAF, BRAF pV600E) or germline mutations of NF1 (68%). Rarer alterations involved FGFR1/2/3, NTRK2, RAF1, ALK, and ROS1. Less than 5% of the driver mutations identified concerned genes not directly associated with the RAS/MAPK pathway, including H3F3A, IDH1, MYB, or MYBL1. Interestingly, the latter pLGG or molecularly undetermined pLGG also exhibited an activation of the pathway. This claim was supported by means of determining expression levels of phospho-ERK or analysis of gene expression and by using gene set enrichment scores for MAPK-upregulated and -repressed genes, respectively, as a measure of pathway activation. Among the NF1-associated tumors that generally have a good prognosis, the non-optic nerve gliomas (non-OPGs) had a worse prognosis. Of the high-risk, recurrent tumors, 20% were associated with additional driver mutations such as BRAF p.V600E, FGFR1, and/or H3F3A (H3.3) p.K27M. Concerning clinical implications, the authors suggest that non-OPG NF1 tumors may require particular management with early determination of such additional mutations, followed by respective treatment strategies in line with previous reports.2 Based on the molecular and clinical data, a molecular-based risk stratification was developed for pLGG, placing the tumors with fusion genes and germline NF1 mutations into the low-risk group (67%), with little progression in the first 10 years and few deaths at 20 years of follow-up. The intermediate risk group (25%) comprises the tumors with BRAF pV600E mutations, but without CDKN2A deletions, whereas those with the deletion belong to the high-risk group (4%) together with H3.3 pK27M tumors that invariably progress and therefore require immediate and more aggressive treatment. For a number of rare alterations affecting 4% of pLGG, no risk stratification could be established. This comprehensive analysis of pLGG established that activation of the KRAS/MAPK pathway is a key prerequisite in the development of basically all pLGG and enabled molecular-based risk stratification. This work is of great importance in that it provides guidance for risk-based treatments, with surveillance for low-risk tumors and more aggressive treatments for high-risk groups, including strategies for targeted therapies based on the individual driver mutations detected. Despite the identification of important oncogenic drivers in the development of diffuse intrinsic pontine glioma (DIPG) from biopsy and autopsy patient-derived tumor samples, we have yet to identify long-term effective therapies for this tumor, which accounts for the majority of pediatric brainstem gliomas and carries a medial survival time of 9 months from diagnosis with only about 10% of patients surviving 2 years. Due to the anatomic location and infiltrative nature of DIPG, these tumors are not amenable to surgical resection. Radiation therapy is palliative at best. Recent super-enhancer analysis1 and single-cell transcriptomic studies2 of DIPG, however, suggest oligodendroglial precursors as the cell of origin, fueling the backdrop for further studies in DIPG, a tumor once only described by means of magnetic resonance imaging. Furthermore, alterations in key genetic drivers, including histone H3, ACVR1, TP53, PDGFRA, and Myc, are helping to direct preclinical investigations and offer the possibility of identifying molecularly targeted treatments for affected patients. Using a conditional knock-in murine model of ACVR1 mutant DIPG, Fortin and colleagues3 demonstrated mutant ACVR1 arrests glial cell differentiation of oligodendroglial-like precursors in murine models, suggesting ACVR1 mutations as an early oncogenic event in a subset of DIPG, and when combined with mutant histone H3K27M and Pik3ca, they produce high-grade diffuse gliomas most commonly in the mid- and forebrain, unlike the human counterpart where such tumors arise in the pons, potentially reflecting the targeted cell of origin and/or interspecies variations. Nonetheless, their results highlight the interplay of these genes in tumorigenesis. Understanding how these genes cooperate to impose an oligodendroglial precursor cell phenotype in DIPG cells, however, has yet to be fully elucidated. The authors further characterized the in vivo activity of a potential therapeutic agent in the form of a dual ACVR and MEK1/2 inhibitor, called E6201, which shows adequate brain penetration in mice4 and has been well tolerated in phase I trials for human solid tumors.5 This drug decreased the growth of DIPG cells lines carrying mutant ACVR1 and HIST1H3B or H3F3A in a dose-dependent manner. In light of the association between ACVR1 and PIK3CA mutations in DIPG, the authors also tested combination therapy with E6201 and CNS-penetrant PI3K inhibitor buparlisib in DIPG cell lines, documenting primarily synergistic effects at clinically relevant IC50 concentrations. The question remains whether these findings will translate into effective therapeutic options for patients with DIPG in the future. So, as Brain Tumor Awareness Month comes to a close, we recognize and honor those affected by brain cancer and cannot stress enough the continued support of ongoing research such as this in these diseases. High tumor mutational burden (TMB) could be a surrogate marker of neoantigen expression. Recognition of neoantigens is a key factor in immunotherapy activity, and therefore neoantigen load may be a biomarker for immunotherapy benefit. Tumors with a very high TMB, of which tumors with microsatellite instability (MSI) represent the highest TMB fractions, respond favorably to programmed cell death-1 (PD-1) inhibitors, and this has led to the approval of pembrolizumab in MSI-high tumors, regardless of the organ of origin. Individuals with germline Lynch syndrome mismatch repair (MMR) defects have long been recognized to be at increased risk of central nervous system tumors. Patients with biallelic dMMR in glioblastoma respond to nivolumab.1 It is not well understood how hypermutation develops in gliomas and it is less clear whether it predicts response to immunotherapy in gliomas. Indeed, retrospective series have shown disappointing results in patients with high TMB without germline Lynch syndrome or MSI-high status treated with PD-1 inhibitors.2 In a large collaborative effort, Touat et al3 analyzed molecular determinants of mutational burden and signatures in 10,294 gliomas. Two main pathways to hypermutation were described, one associated with constitutional defects in DNA polymerase and MMR genes, and a more common post-treatment pathway associated with acquired resistance driven by MMR defects in chemotherapy-sensitive gliomas that relapse following temozolomide. Experimentally, mutational signature of post-therapy hypermutated gliomas was reintroduced by temozolomide-induced damage in cells with MMR deficiency. MMR-deficient gliomas that lacked prominent T cell infiltrates had extensive intratumoral heterogeneity, poor patient survival, and a low rate of response to PD-1 blockade. Bulk analyses did not detect MSI in MMR-deficient gliomas, but single-cell, whole-genome sequencing analysis of post-treatment hypermutated glioma cells identified microsatellite mutations. This study shows that chemotherapy may introduce hypermutated populations without promoting a response to PD-1 inhibitors. Touat et al defined high TMB as ≥17.0 mut. per Mb.1 However, the cutoff for TMB is not well defined, and response to PD-1 inhibitors in inflamed tumors may differ from that in non-inflamed tumors.4 Intratumoral heterogeneity may have a greater impact on immune response and may predict immunotherapy outcomes better than mutational burden.5 All of these complexities make the value of TMB as a biomarker questionable with respect to highly heterogeneous tumors such as GBM, even in those with MMR defects. The interesting study by Touat et al generates many unanswered questions, eg, whether nitrosoureas should be the preferred chemotherapy for patients with temozolomide-induced high TMB.
Source: Neuro-Oncology
Language: English
Subfield: Genetics
Field: Medicine
Domain: Health Sciences
Open Access status: bronze
OpenAlex ID: W4242500316