In depth study of pediatric gliomas has been hampered due to difficulties in accessing patient tissue and a lack of clinically representative tumor models. Over the last decade, however, profiling of carefully curated cohorts of pediatric tumors has identified genetic drivers that molecularly segregate pediatric gliomas from adult gliomas. This information has inspired the development of a new set of powerful in vitro and in vivo tumor models that can aid in identifying pediatric-specific oncogenic mechanisms and tumor microenvironment interactions. Single-cell analyses of both human tumors and these newly developed models have revealed that pediatric gliomas arise from spatiotemporally discrete neural progenitor populations in which developmental programs have become dysregulated. Pediatric high-grade gliomas also harbor distinct sets of co-segregating genetic and epigenetic alterations, often accompanied by unique features within the tumor microenvironment. The development of these novel tools and data resources has led to insights into the biology and heterogeneity of these tumors, including identification of distinctive sets of driver mutations, developmentally restricted cells of origin, recognizable patterns of tumor progression, characteristic immune environments, and tumor hijacking of normal microenvironmental and neural programs. As concerted efforts have broadened our understanding of these tumors, new therapeutic vulnerabilities have been identified, and for the first time, promising new strategies are being evaluated in the preclinical and clinical settings. Even so, dedicated and sustained collaborative efforts are necessary to refine our knowledge and bring these new strategies into general clinical use. In this review, we will discuss the range of currently available glioma models, the way in which they have each contributed to recent developments in the field, their benefits and drawbacks for addressing specific research questions, and their future utility in advancing biological understanding and treatment of pediatric glioma.

Brain tumors are the leading cause of cancer-related death in children [1]. Approximately half of all children’s brain tumors are gliomas, a tumor type composed primarily of heterogenous populations of glial cells. Pediatric high-grade gliomas (pHGGs) are universally fatal – no standard treatment options improve survival, and therapy is generally only palliative [2]. Low-grade gliomas, on the other hand, which include pediatric low-grade gliomas (pLGGs), can have favorable outcomes when complete resection is possible. However, this is often not the case, and 50% of pLGG patients require complex clinical management involving surgery and chemoradiotherapy, which severely impacts quality of life [3]. For both pHGGs and pLGGs, treatment options are severely limited.

Chemoradiotherapy regimens used as the standard of care in children were initially designed for adult glioblastoma (GBM), despite adult and pediatric gliomas having different driver mutations, developmental origins, and patterns of progression [4]. The use of these therapeutic strategies reflects the lack of knowledge at the time about the origin of pediatric gliomas and their distinguishing characteristics. Since 2008, however, when pediatric glioma-specific mutations were first identified, a plethora of new studies have begun to reveal the origins, mechanisms, and treatment vulnerabilities of these tumors [5‒11]. These discoveries are beginning to lay the foundation for more appropriate and effective treatment options for patients.

Genomic profiling of large cohorts of patient material has revealed many characteristics specific to pediatric gliomas. For example, whole genome and single-cell sequencing have identified spatiotemporally discrete neural precursor cells (NPCs), and in particular oligodendrocyte precursor cells (OPCs), as putative cells of origin in pHGGs, giving rise to tumors as a result of stalled, dysregulated developmental programs [12‒16]. Further, these tumors are driven by characteristic sets of histone mutations and subtype-specific genetic partner alterations associated with tumor location, age of onset, and proliferative and invasive phenotypes [17]. In pLGGs on the other hand, characteristic fusions and alterations involving single genes have been identified as drivers of tumorigenesis [7‒10]. In addition, the importance of the tumor microenvironment (TME) in glioma development and progression is beginning to be investigated, including investigation of the immunosuppressive microenvironmental landscapes within pediatric gliomas, integration of glioma cells into tumor-neural networks, and novel patterns of invasion and therapeutic resistance [18‒23].

In depth investigation of pediatric gliomas has been historically limited by a number of factors, including the relative scarcity of patient material, the expense of deep molecular characterization techniques, and the lack of in vitro and, most especially, in vivo models capable of accurately recapitulating patient disease [2]. However, recent developments have improved the outlook on all counts. Over the last decade, there has been a concerted effort to collect and characterize patient material, including optimized protocols to isolate tissue, generate cell lines, and establish patient-derived in vitro and in vivo models [24‒26]. Advances in stem cell engineering have offered innovative new approaches for evaluating tumor origin and progression [27]. Furthermore, powerful in vivo mouse models have been generated, yielding insights into the prenatal origins and microenvironmental interactions underpinning these developmentally arrested tumors [28‒33].

Pediatric glioma patients are in dire need of new therapeutic options, the development of which depends on deeper understanding of the developmental origins, mechanisms of progression, and microenvironmental interactions within these tumors. Importantly, therapy development requires well-characterized tools and models where mechanism and efficacy of treatment can be appropriately determined. Our aim in this review is to provide an overview of the strengths and weaknesses of currently available models of pediatric gliomas, examine their contributions to the field, and discuss opportunities to use these models to drive new biological and therapeutic discoveries.

High-Grade Glioma

Approximately 20% of brain tumors in children receive a pHGG diagnosis [1]. These tumors are aggressive and heterogenous, with no effective treatment and a median survival of only 9–15 months [34]. pHGGs are histologically similar to adult GBM but present with different anatomical and age-dependent patterns, and crucially, also harbor distinct genetic and epigenetic profiles [17, 35]. Whereas adult GBM are primarily restricted to hemispheric locations, about half of pediatric gliomas are found along the midline of the brain, in the cerebellum, thalamus, and pons, where they are collectively referred to as diffuse midline gliomas (DMG), formerly diffuse intrinsic pontine gliomas [35]. A majority of these midline tumors exhibit epigenetic dysregulation, harboring p.Lys27Met (K27M) mutations in either H3F3A, encoding histone H3.3, or in HIST1H3B, encoding histone H3.1. Additionally, approximately 15% of pHGGs and 30% of HGGs in young adult and adolescents (12–35 years old) carry a distinct mutation in H3F3A, a p.Gly34Arg/Val (G34R/V) substitution, which is restricted to hemispheric tumors [5, 6, 36, 37]. These anatomically and developmentally distinct histone mutant tumor subtypes are also accompanied by sets of co-segregating genetic alterations in genes such as TP53, ATRX, ACVR1, PDGFRA, PIK3CA, and many others [17]. Recent evidence indicates that these tumors arise from NPCs and OPCs during development and are driven by a combination of epigenetic dysregulation and concurrent or successive genetic partner alterations [13, 14, 38]. While wide-ranging efforts over the past decade have vastly expanded our understanding of the molecular characteristics and origins of these tumors, there is still a dire need to improve the standard of care for patients and provide viable therapeutic options.

Low-Grade Glioma

pLGGs are among the most common childhood brain tumors, representing approximately 30% of brain tumors in 0–19 year olds, and while treatments are available, they can come at a high cost [1]. Like pHGGs, pLGGs have distinct anatomical and age-dependent patterns and are often driven by alterations or fusions in single genes that impact the MAPK pathway or MYB gene family, which distinguishes them from adult LGGs [3, 39, 40]. Tumors in this category include WHO grade 1 tumors: pilocytic astrocytoma, which can be either sporadic or neurofibromatosis 1 (NF1) associated, optic tumors, ganglioglioma, and dysembryoplastic neuroepithelial tumor, and WHO grade 2 tumors: diffuse astrocytoma, pleomorphic xanthoastrocytoma, and oligodendroglioma. A proportion of anaplastic astrocytomas in young adults have been classified as low-grade gliomas as well [39]. pLGGs are also associated with disruptions in neurodevelopmental pathways like pHGGs, as the active programs of neuro- and gliogenesis are particularly susceptible to somatic mutation early in life [3, 41]. Genetic drivers of pLGGs include alterations in NF1, FGFR1, MYB/MYBL1, BRAF fusion or BRAF V600E mutation, and others, particularly along the MAPK pathway [7‒10, 39]. Additionally, LGGs in adolescents and young adults commonly exhibit mutations in IDH1, accompanied by either chromosomal 1p/19q deletions or mutations in p53 and ARTX. Importantly, these tumors are likely to transition to HGGs over time [39, 42]. These characteristic mutations have opened the door for targeted therapy in these tumors and support the use of clinical molecular diagnostics to best tailor care to individual patients.

Orthotopic xenografting of patient tissue (PDX) or patient-derived cell lines (CDXs) in immunodeficient mice is the most common and well-established strategy to generate brain tumor models for preclinical research. However, recent advances in the generation of glioma models from embryonic or induced pluripotent human stem cells (ESCs and iPSCs) are also enabling investigation of gliomagenesis mechanisms through a closer recapitulation of the human context. These models are also more compatible with high-throughput approaches for drug screening and target identification. On the other hand, research is also progressing to improve xenograft models and address their lack of an intact immune and stromal microenvironment, which is broadly considered to be these models’ primary failing. In the following sections, we will discuss recent developments in the generation of human-derived models of pediatric gliomas (Fig. 1).

Fig. 1.

Human and mouse models of pediatric glioma. (Left) Human models can be derived either from primary patient material or from engineered stem cells. Patient material can be cultured as tissue, as dissociated cells in 2D or 3D, or as organoids. Material can also be xenografted into mice or introduced into GLICOs. Stem cell-derived models use developmentally precise cerebral organoids as a base to either induce oncogenesis or pair with patient tumor material. (Right) Mouse models of pediatric glioma involve either the introduction of oncogenes and knockdown of tumor suppressors (IUE, RCAS, MADR, transgenic strategies), or the injection of syngeneic mouse-derived tumor cells. GBO: glioblastoma organoid; GLICO: glioblastoma cerebral organoid; neoCOR: neoplastic cerebral organoid; IUE: in utero electroporation; RCAS: replication-competent avian-like sarcoma virus; MADR: mosaic analysis with dual recombinase-mediated cassette exchange.

Fig. 1.

Human and mouse models of pediatric glioma. (Left) Human models can be derived either from primary patient material or from engineered stem cells. Patient material can be cultured as tissue, as dissociated cells in 2D or 3D, or as organoids. Material can also be xenografted into mice or introduced into GLICOs. Stem cell-derived models use developmentally precise cerebral organoids as a base to either induce oncogenesis or pair with patient tumor material. (Right) Mouse models of pediatric glioma involve either the introduction of oncogenes and knockdown of tumor suppressors (IUE, RCAS, MADR, transgenic strategies), or the injection of syngeneic mouse-derived tumor cells. GBO: glioblastoma organoid; GLICO: glioblastoma cerebral organoid; neoCOR: neoplastic cerebral organoid; IUE: in utero electroporation; RCAS: replication-competent avian-like sarcoma virus; MADR: mosaic analysis with dual recombinase-mediated cassette exchange.

Close modal

Patient-Derived Xenograft Models

Patient-derived PDX and CDX models are among the most common systems to study pediatric brain tumors. However, because low-grade gliomas are more difficult to subculture in vitro or expand in vivo, most such xenograft models exist only for pHGGs. Historically, access to patient material from pHGG tumors has been limited, on account of the rare incidence of these tumors and the location along the midline of the brain, which has made biopsies challenging [26]. Over the last decade, however, a combination of technical advances and close collaboration between patients, clinics, and laboratories has led to expanding availability of patient-derived xenografts and cell lines, derived both from autopsy and biopsy samples, and often accompanied by molecular characterization (online suppl. Table 1; for all online suppl. material, see https://doi.org/10.1159/000531040). As a result, valuable insights into the biology of these tumors and new candidates for therapeutic development are beginning to emerge for the first time in many years.

In general, most commonly used patient-derived xenograft models and engraftable cell lines have been shown to broadly recapitulate both the histology and molecular landscape of the original patient tumor, retaining driver histone mutations and partner alterations, and recapitulating phenotypic characteristics such as diffuse invasive patterns [24, 26, 43, 44]. It should be noted that models derived from patients can lose certain attributes, such as partner mutations, with increased time or passages in culture, and these changes may be inconsistent, depending on the method of establishment [43]. Additionally, studies have suggested that among patient-derived cell populations, only cells with certain stem-like characteristics have tumorigenic capacity and those transcriptional landscapes may differ when comparing primary tumor, PDX, and cell culture models [13]. Nevertheless, the ability of patient-derived material to model human disease in vitro and in vivo has proven invaluable to the field, revealing disease mechanisms and therapeutic vulnerabilities [12, 43, 45‒56].

Initially, patient-derived models were obtained at autopsy, opening the door to essential insights into disease origin and progression [24]. Recent improvements to biopsy protocols have both aided in tumor characterization mid-treatment and have facilitated the generation of additional patient-derived models [57‒59]. Additionally, advances in cell culture techniques have enhanced our ability to maintain patient material in vitro. pHGG cells are most commonly maintained in 3D culture as neurospheres because these conditions select for tumor-initiating glioma stem cells (GSCs) that are more efficient at generating tumors and better reflect the heterogeneity of the original tumor than the rapidly dividing cellular compartment selected for in 2D models [60‒65].

Both serially implanted PDX and in vitro-maintained patient-derived cell line (CDX) models have advanced the study of gliomas, and each strategy is accompanied by certain strengths. Cell lines are generally more cost-effective and efficient to maintain, easier to share with the research community, and allow for high-throughput screening and therapeutic testing. Cell lines also provide the opportunity for genetic manipulation to evaluate biological pathways or dependencies and to introduce fluorescent or bioluminescent reporters for experimental observation. On the other hand, PDX models offer the advantage of maintaining the immediate microenvironment and cellular heterogeneity of the patient tumor, which can be lost or selected against when tissue is dissociated and subcultured in vitro [66]. However, these models are expensive and time consuming to maintain, have limited high-throughput capacity, and do eventually lose the microenvironmental and genetic characteristics of their human tumor origins at high passages, similar to in vitro culture [67].

With an increasing suite of patient-derived cell lines available, it will also be possible to better investigate the effects of treatment on patient representative material. Molecular characterization of patient biopsies and therapeutic testing of cell lines is already enabling stratification of patients into clinical trials for targeted therapy [17, 35, 39, 43, 51, 54, 55]. Further, an expanding series of studies using patient-derived material has led to new strategies for targeted therapy in DMGs. These include inhibition of receptor tyrosine kinases mutated in distinct tumor subtypes, epigenetic inhibitors, inhibitors of the DNA damage response, and disruption of TME dependencies. Some of these strategies have led to promising clinical trials, as will be discussed below.

However, among currently available models, it is important to identify whether the original patient had undergone treatment, and if so, with which agent or modality, as chemoradiation is known to drive tumor evolution and the rise of resistant cell populations [68]. With proper annotation of patient history accompanying these models, it will be possible to precisely evaluate the effects of the current standard of care and reveal potentially new, therapeutically tractable vulnerabilities.

Human Stem Cell-Derived Models

While patient-derived xenograft models represent the most common system used to study pHGGs/DMGs, recent advances in human stem cell engineering have led to the development of new, precisely tuned models of pediatric gliomas [69‒73]. Human stem cell-derived models provide two significant benefits. First, they are particularly useful to model tumors from which patient-derived material is difficult to obtain or maintain as a cell line or PDX. This is true generally for DMGs, which are rare tumors, and applies to more common tumors such as pLGGs, which are difficult to establish in culture and for which there are very few human-derived cell lines [31]. These models may also help avoid the sometimes significant genetic drift seen in patient-derived models following high passage. The second major benefit of these models is the ability to precisely control cell identity and genetic landscape with developmental and regional specificity [72]. This allows evaluation of the cell-of-origin and the roles of mutations as drivers of tumorigenesis within the context of controlled stem cell-derived progenitor identity in vitro.

Most stem cell models of pHGG are derived from either human embryonic stem cells (hESCs) or induced pluripotent stem cells (iPSCs), or alternatively fetal neural stem cells (NSCs). When models are developed with hESCs or iPSCs, it is necessary to induce differentiation into NPCs and choose the stage at which to introduce putative driver mutations, such as histone mutations and partner alterations associated with DMG [71]. Precise control over cell identity has contributed to better understanding of the developmental origins of these tumors. Using gene signatures of specific progenitor identities to induce targeted differentiation in vitro allows evaluation of driver mutations in different cell types, reflecting temporally, functionally, or anatomically distinct progenitors. For example, Funato and colleagues demonstrated that introducing H3.3K27M into NPCs along with partner mutations in TP53 and PDGFRA resulted in phenotypes associated with oncogenic transformation. However, the H3.3K27M mutation alone did not have a proliferative effect in either undifferentiated hESCs, differentiated hES-derived astrocytes, or primary human astrocytes [70]. Only when combined with partner alterations in TP53 and PDGFRA, these cells gained a proliferative advantage in vitro and formed glioma-like lesions when transplanted in vivo [70]. In an effort to model H3.3G34R tumors, the same group further differentiated hESCs into specific NPC types recapitulating forebrain and hindbrain progenitors, including forebrain interneuron precursor cells and hindbrain NPCs. This enabled the identification of forebrain interneuron precursor cells as the putative cells of origin for H3.3G34R tumors [69]. Additional studies using iPSCs to model low-grade pilocytic astrocytoma have identified neuroglial progenitor populations, but not differentiated astrocytes, as potential cells of origin for pLGG, driven by either NF1 loss or BRAF fusion [31].

Human fetal NSCs are derived with regional specificity during brain development, so there is limited need to further differentiate the cells. Additionally, regionally specific characteristics are maintained in vitro, providing a useful platform for the induction of oncogenic mutations [74]. Using fetal NSCs derived from either the forebrain or hindbrain, it was recently shown that regional identity is a determining factor in the oncogenic response to either the H3.3K27M or H3.3G34R mutation, and that the H3.3G34R mutation prevents differentiation of forebrain fetal NSCs by impairing recruitment of the ZMYND11 transcriptional repressor [72]. Similarly, fetal hindbrain NSCs were used to examine the mechanism by which H3.3K27M appears to maintain immature epigenetic programs to support tumorigenesis [73]. This approach has also been utilized to model other types of pediatric brain tumors such as medulloblastomas, in which MYCN-driven transformation elicited different latencies and invasion phenotypes in iPSCs versus fetal NSCs [75]. Although the insights provided through use of fetal NSCs are valuable, this material is somewhat rare, and can be accompanied by ethical concerns.

Together, these studies provide valuable insight into the oncogenic programs operating in pediatric gliomas and emphasize the importance of the cell-of-origin and developmental stage in glioma development. While fetal NSCs may be a rare resource, both these cells and more versatile hESCs/iPSCs serve as a valuable platform for mechanistic studies in genetically well-defined human lines.

Organoids

Although established cell lines are the most common approach for modeling pediatric gliomas, cell lines lack the full complement of stromal cells within the intact TME, which have a significant impact on tumor progression and treatment. Organoid models on the other hand attempt to circumvent this issue by incorporating aspects of the TME in a fully humanized system. Existing models vary between entirely patient-derived tissue, iPSC-engineered models, and a combination of the two strategies, each with particular benefits, including glioblastoma organoids (GBOs), bio-printed organoids, neoplastic cerebral organoids (neoCORs), and recently developed GLICO models [27] (Table 1).

Table 1.

Organoid glioma models

ModelOriginNon-tumor stromal componentsTime to maturityReference
GBO Primary material No 1–2 weeks (79) 
Bioprinted organoid Established cell lines/primary material Yes 1–2 weeks (80, 81) 
neoCOR iPSCs Yes 1–3 months (83, 84) 
GLICO iPSCs/primary material Yes 1–2 months (85, 86) 
ModelOriginNon-tumor stromal componentsTime to maturityReference
GBO Primary material No 1–2 weeks (79) 
Bioprinted organoid Established cell lines/primary material Yes 1–2 weeks (80, 81) 
neoCOR iPSCs Yes 1–3 months (83, 84) 
GLICO iPSCs/primary material Yes 1–2 months (85, 86) 

GBO, glioblastoma organoid; neoCOR, neoplastic cerebral organoid; GLICO, glioblastoma cerebral organoid; iPSC, induced pluripotent stem cell.

GBOs are derived directly from patient tumors, the majority of which have been taken from adult GBM patients. Although, previously, GBOs were established from a single-cell suspension of patient material in vitro, recently developed techniques are able to more efficiently establish GBOs and better maintain tumor stromal architecture without dissociation [76]. These technical advances have increased the speed at which GBOs are developed, allowing for rapid characterization and in vitro drug screening, and quicker turnaround for patients in a clinical setting. Additionally, these techniques maintain many histological and molecular characteristics of patient tumors, including their tumor heterogeneity, and have been shown to engraft efficiently as xenografts in mice. As a result, biobanks of patient-derived GBOs are beginning to expand, serving as a valuable resource for continued investigation.

However, GBOs also lack many aspects of the TME that influence tumor progression. Bio-printed organoids offer an alternative approach, allowing the design of complex, customized environments with multiple cell types, introducing vascular endothelial cells, astrocytes, neural precursor cells, and macrophages alongside patient-derived cells [27]. These models have been shown to recapitulate patient transcriptional profiles, invasion patterns, and therapeutic response [77, 78]. While powerful use of these models is limited on account of the specialized equipment and training required.

Early non-tumor cerebral organoids demonstrated notably distinct regions with self-organizing cell types that mimicked the stages of human brain development [79]. Recently, two groups concurrently used these iPSC-derived organoids to model brain tumors. Using CRISPR/Cas9 and piggyBac transposon techniques, researchers introduced oncogenic mutations into developing cerebral organoids, labeling these iPSC-derived brain tumor organoids as “neoCORs,” or neoplastic cerebral organoids [80, 81]. Ogawa and colleagues introduced TP53−/− and HRasG12V expression to induce oncogenic transformation. Bian and colleagues utilized a variety of mutations, some combinations of which reliably induced transformation, including MYC amplification; CDKN2A−/−/CDKN2B−/−/EGFROE; NF1−/−/PTEN−/−/TP53−/−; and EGFRvIIIOE/CDKN2A−/−/PTEN−/−. The strength of these models is the ability to introduce oncogenic mutations at precise developmental stages and model tumor initiation, which is impossible in material derived from later-stage patient tumors. These models may still require further validation, as introducing mutations in this way may not lead to the same degree of heterogeneity as seen in patient gliomas [27]. However, the precision of the technique is valuable for mechanistic investigation of tumor progression in a human context.

Using cerebral organoids as a foundation, an additional technique has combined iPSC-derived organoids with patient-derived tumor cells or tumor tissue, combining the strengths of the previously available organoid models [80, 82, 83]. Glioblastoma cerebral organoid (GLICO) models have so far proved versatile for modeling tumor tissue in the context of a “normal” microenvironment. In particular, these studies have shown that patient material is capable of invading early-stage cerebral organoids [83], that tumor cells form tumor microtubes within the organoid – a hallmark of complex tumor networks and therapeutic resistance [82], and that GSCs in GLICOs recapitulate the cell states and plasticity of the patient tumor of origin [84].

Organoid models have indicated that the TME may be required to maintain the variety of cellular states found in adult GBM, in contrast to neurosphere culture which has been shown to select for GSCs [27]. Additionally, the growing biobank of GBOs is a valuable resource for tumor characterization and therapeutic testing, as stromal components are known to affect tumor progression and treatment response. Recent studies show the utility of invasion assays to analyze anti-infiltrative effects of inhibiting targets such as ADAM10 and Bet alongside traditional proliferation analysis, as infiltration is a key challenge to effective glioma treatment [85, 86]. Further, GLICOs have been used to demonstrate patient-specific sensitivities to chemotherapy (TMZ and BCNU) and radiotherapy, and a case study using GBOs derived from a patient, in which the tumor material was characterized and submitted to a panel of FDA-approved mTOR inhibitors, resulted in switching the patient from TMZ to everolimus treatment, suggesting future opportunities for personalized medicine [82, 87].

Advances in the complexity of cerebral organoids, such as the joining of vascular-differentiated spheroids with cortical brain organoids, may further enhance the utility of GLICO models [88]. However, these models are still missing key elements of the TME, including perfused vasculature and dynamic immune landscapes. Thus, although organoids are valuable tools for both therapeutic testing and mechanistic biology, it is necessary to complement these approaches with tumor models in full biological systems, the most common of which are mice.

Immunocompromised Mice

Most in vivo work with human material is performed in mice that are immunocompromised to some degree, allowing human tumor cells or material to engraft. Unfortunately, this limits the ability to evaluate tumor-immune interactions in vivo. However, the advent of humanized mice is offering the opportunity to introduce elements of the human immune system to murine models, allowing investigation of some human immune response to tumors in an intact environment. As these techniques have been extensively reviewed [89, 90], here we will only outline the currently available models.

Immunocompromised mice come in a variety of strains, with different degrees of immune depletion. Severe-combined immunedeficiency (SCID) and non-obese diabetic-SCID mice harbor the PRKDCscid mutation (protein kinase, DNA-activated catalytic polypeptide) and polymorphic Sirpa gene, conferring arrested development of T and B lymphocytes and disrupted innate immune activity, respectively, and NK function can be depleted through truncating (NOG) or deleting (NSG) the IL2 receptor Il2rg[91‒93]. T- and B-cell deficiency can also be achieved through mutations of Rag1 or Rag2, preventing recombination during lymphocyte education, as used in BRG and NRG lines [90]. It is necessary to recognize that immunocompromised mice carry with them important caveats, including short life expectancies, incomplete immune suppression, and susceptibility to DNA damage which limits the modeling of currently used chemoradiotherapy [89, 90].

Recently, a novel approach has been developed to support engraftment of human adult and pediatric glioma material into immunocompetent mice using CTLA-4 and anti-CD154 T-cell costimulatory blockade [94‒96]. Even though immunosuppressive treatment was confined to the week following tumor cell injection, the study showed sustained tumor growth, and when compared to tumors in SCID mice, tumors in immunocompetent mice displayed phenotypes such as inflammatory infiltration and microvascular proliferation that were more representative of human tumors [94]. This model may prove to be a promising alternative to the more commonly used immunocompromised mice, particularly in the ability to model tumor-immune interactions.

Humanized Mice

Humanized mice offer the opportunity to reintroduce elements of the human immune system into immunocompromised murine models, enabling a degree of insight into human tumor-immune interaction. Most models include the introduction of human immune cells or human tissue to irradiated and/or genetically immunocompromised mice to enable engraftment. As reviewed by Cogels and colleagues, there are four general strategies for humanized mouse models: the peripheral blood mononuclear cell (PBMC), hematopoietic stem cell (HSC), spleen mononuclear cell (SPMC), and bone-liver-thymus (BLT) models (Table 2) [97].

Table 2.

Strategies for humanized mouse models

ModelSourceSupplied immune cellsNotesReference
PBMC Peripheral blood CD3+ T cells, some B cells, myeloid cells GvHD (98) 
HSC Umbilical cord blood, peripheral blood, bone marrow CD34+ HSCs In situ lymphocyte education, long establishment time (97, 99) 
SPMC Adult spleen Spleen mononuclear cells (CD45+ T cells, B cells) GvHD limited source material (100) 
BLT Fetal bone, liver, and thymus T cells, B cells, macrophages, dendritic cells Human thymic education of immune cells limited source material (101) 
NSG-SGM3/MI(S)TRG Transgenes encoding human cytokines Supports engraftment of CD34+ HSCs, BLT Innate immune infiltration (102, 103) 
ModelSourceSupplied immune cellsNotesReference
PBMC Peripheral blood CD3+ T cells, some B cells, myeloid cells GvHD (98) 
HSC Umbilical cord blood, peripheral blood, bone marrow CD34+ HSCs In situ lymphocyte education, long establishment time (97, 99) 
SPMC Adult spleen Spleen mononuclear cells (CD45+ T cells, B cells) GvHD limited source material (100) 
BLT Fetal bone, liver, and thymus T cells, B cells, macrophages, dendritic cells Human thymic education of immune cells limited source material (101) 
NSG-SGM3/MI(S)TRG Transgenes encoding human cytokines Supports engraftment of CD34+ HSCs, BLT Innate immune infiltration (102, 103) 

PBMCs, peripheral blood mononuclear cells; HSCs, hematopoietic stem cells; SPMCs, spleen mononuclear cells; BLT, blood-liver-thymus; NSG, NOD scid gamma; SGM3, SCF, GM-CSF, IL-3; MI(S)TRG, M-CSF, IL-3, Sirpα, TPO, GM-CSF.

The PBMC model involves engraftment of primary CD3+ T cells, along with some B cells and myeloid cells, from easily accessible adult peripheral blood and can allow matching of patient tumor material with the patient’s own immune cells [98]. However, graft versus host disease (GvHD) develops in these models after only about 4 weeks, severely limiting their use for long-term studies [98]. The HSC model seeks to address the issue of GvHD, introducing CD34+ hematopoietic stem cells which travel to the bone marrow, resulting in production of B cells, CD14+ monocytes, and T cells that are educated to be tolerant of mouse tissue [99]. This model can take several months to establish, but the suite of human immune cell types and the absence of GvHD make these models particularly suitable for long-term studies [97].

The recently developed SPMC model involves the engraftment of SPMCs, including B, T, and activated T cells, obtained from adult spleen organ donors [100]. However, the reliance on organ donors restricts accessibility, and GvHD occurs in this model, similar to PBMC. Finally, the BLT model most completely reconstitutes the human immune system without GvHD, including presence of B and T cells, macrophages, dendritic cells, human thymic education of pro-T cells, and B- and T-cell coordination. This model requires engraftment of the fetal bone, liver, and thymic tissue, and the primary disadvantage is tissue accessibility, which is both rare and accompanied by ethical challenges [101].

An additional set of models has been developed focusing on the myeloid compartment to support monocyte survival and activation. For example, the NSG-SGM3 model, expressing IL-3, GM-CSF, and SCF, and the MI(S)TRG models, expressing M-CSF, IL-3, Sirpα, TPO, and GM-CSF, both allow for the engraftment of HSCs and expansion of myeloid cells, as well as adaptive immune cells, in mice [102‒104]. Recently, this strategy was used in combination with the BLT model (NSG-SGM3-BLT) to investigate T-cell-associated immunosuppression in DMG [105]. This model may be one of the most complete reconstitutions of the human immune system in mice available. However, technical and ethical challenges hinder widespread use.

Benefits

The recent advances in the recovery and maintenance of human tissue and engineering of human stem cells have led to a remarkable expansion of resources, each with its own set of benefits. Patient-derived cell lines and tissue, for example, have the potential to best represent an individual patient’s disease, and these resources could eventually be leveraged for individualized therapy testing. The ability to engineer multiple NPC and pHGG cell types from iPSCs leads to a more representative model of tumor initiation than patient material, whereas patient-derived and engineered GBOs, neoCORS, and GLICOs are able to maintain TME within a more complex human setting. Finally, the developing field of humanized mice has the potential to recapitulate at least some aspects of the human immune system, allowing patient material to be engrafted in a living system.

Drawbacks

Despite these benefits, there are still drawbacks to existing human models of pHGG. Among them remains the limited patient material available, and the genetic and phenotypic drift that serial passage of any patient material can produce. In any in vitro culture, whether 2D, 3D, or organoid, there is still a lack of the entire stromal component. Because the TME is integral to tumor development, these models still lack the power to answer many important questions about disease etiology and progression. Even for humanized mice, most models are time restricted, labor intensive, and remarkably expensive, the most complete model is ethically challenging, and most are still underpowered to assess the role of innate immune elements such as resident microglia. Without a comprehensive TME, it is difficult to fully address questions including how tumorigenesis is initiated by driver mutations during normal development, what role the stroma plays and whether it produces therapeutic vulnerabilities, and how resistance to therapies emerge.

Mouse models of pediatric brain tumors offer the benefit of having a tumor developing within the context of a functional, fully competent brain environment. In particular, models in which the tumor develops alongside standard neurodevelopmental patterns, as is seen in children, offer valuable opportunities to investigate developmental drivers of tumor initiation and progression (Fig. 1).

Transgenic Models

The traditional approach to generating genetically engineered mouse models has been to crossbreed mouse strains carrying germline transgenes and mutations. However, transgenes can affect a wide range of tissues non-specifically or induce embryonic lethality, which can confound experimental questions. To avoid this, many models use conditional and/or inducible knockout strategies, most commonly the Cre-loxP system. Importantly, Cre can be expressed under a tissue-specific promoter and can also be fused with a hormone responsive element, enabling spatiotemporal induction [106].

For the modeling of pediatric tumors, many studies have used NPC-, astrocyte-, or oligodendrocyte progenitor-specific promoters to express Cre, such as the Nestin, GFAP, or Olig2 promoters, respectively. For example, to model pLGGs, BRAFV600E was expressed in hGFAP-Cre mice in a wildtype or Cdkn2a knockout background to show that BRAFV600E was insufficient to induce tumorigenesis on its own but could do so when combined with Cdkn2a-KO[107]. However, this combination produced high-grade tumors as opposed to the intended low grade, and truly representative models of pLGGs remain challenging to model in mice. Recently, two studies developed transgenic models of H3.3- and H3.1-mutant DMGs. In the first, mice with floxed H3.3K27M, p53-inactivating, and PDGFRA-activating mutations were crossed with Nestin-Cre mice. With this strategy, they found that the oncogenic role of H3.3K27M is restricted to a small temporal window, during which additional mutations co-operate to induce gliomagenesis [30]. In the second study, mutation of ACVR1 resulted in minor neurological defects, notably in Olig2-Cre mice but not in Nestin-Cre mice, but did not induce tumorigenesis. Expression of ACVR1G328V with PIK3CAH1047R, however, both common mutations in H3.1K27M patients, was sufficient for tumor development, and H3.1K27M was found to accelerate lesion formation [108].

While these models have yielded valuable insight into pHGG/DMG development, there are still very few available. Developing a transgenic model with complex sets of mutations can take months to years of crossbreeding, requiring significant resources and numbers of animals. Additionally, transgenic models of brain tumors tend to lack the heterogeneity often seen in patients [60]. Researchers have therefore sought to complement transgenics with other targeted techniques to introduce mutations more directly and locally, by using viral, mosaic recombination, and IUE approaches, discussed further below.

Avian Retrovirus-Mediated Gene Transfer Models

The replication-competent avian-like sarcoma (RCAS) virus and the tumor virus receptor-A (tv-a) system, developed by Eric Holland’s group, has become vital for the study of glioma with spatiotemporal specificity in immune-competent models. With this technique, genes of interest can be inserted into an RCAS retroviral vector, and virions or transfected avian DF1 cells are injected into precise brain locations in a recipient mouse carrying tv-a under a tissue-specific promoter. The viral envelope proteins then bind to the TV-A receptor to enter cells, integrating with host cell DNA and inducing gene expression with cell-type specificity [109]. An advantage to this technique is the anatomical precision of transgene expression, which cannot be achieved using traditional transgenics. Further, despite the relatively small insert size for RCAS vectors, multiple infections can be performed simultaneously, allowing the investigation of multiple genes and gene combinations.

RCAS was used as an early strategy to model gliomagenesis syngeneically in mice. These studies evaluated the necessity of multiple driving mutations, and, importantly, emphasized the role of the cell of origin; in EGFR models, for example, tumors arose more frequently when targeted to Nestin-expressing cells (Ntv-a) compared to GFAP-expressing cells (Gtv-a) [110‒112]. Becher and colleagues developed the first RCAS-based DMG model, injecting RCAS vectors encoding PDGFb into Ntv-a cells lining the fourth ventricle of neonatal mice. PDGFb alone induced low grade-like gliomas, while the same vector injected into Cdkn2a−/−/Ntv-a and p53fl/fl mice produced high-grade gliomas [113‒115]. Soon after, a key study added RCAS-H3.3K27M to the p53-deficient model and demonstrated that H3.3K27M reduced H3K27me3 by inhibiting PRC2 activity in vivo [116]. However, this approach was only able to induce ectopic proliferation, and later studies required PDGFb or constitutively active PDGFRAD842V/PDGFRAV544ins to promote high-grade brainstem gliomagenesis. In these models, H3.3K27M was not necessary for tumorigenesis as PDGFb/PDGFRA mutants drove transformation [21, 117, 118].

Subsequent studies have investigated the cell of origin of H3.3K27M tumors by targeting Pax3+ tv-a cells in addition to Ntv-a cells, resulting in greater heterogeneity of tumorigenesis similar to that seen in patients [119]. Additional work using RCAS has also demonstrated that H3.3K27M mediates repression of p16 to enhance cell proliferation, whereas ACVR1 mutations were shown to cooperate with H3.1K27M and p53 mutation to promote glioma-like lesions [117, 120]. Finally, RCAS models of IDH1-mutant gliomas have shown that the IDH1 mutation alone is not sufficient but requires co-occurring mutations for proneural glioma formation [121]. This approach was also used to investigate the role of neutrophils and macrophages in glioma progression, emphasizing the importance of immunocompetent syngeneic models [122].

One of the disadvantages of the RCAS model is the limited payload, restricting its use for larger sequences [110]. Virions or virion-containing cells can also induce an immune response, potentially confounding studies on immune infiltration [29]. Additionally, variability in copy number, integration sites, and integration efficiency can result in unpredictable patterns of tumor formation, as well as the possibility of transcriptional squelching, each of which are important confounding variables and may introduce heterogeneity in tumor models [29]. Finally, because many of these models still rely on crossing of Cre- and tv-a-bearing transgenic mice, developing a model remains time and labor intensive.

IUE-Based Models

An alternative strategy for somatic genetic engineering is in utero electroporation (IUE) [123, 124]. Most common methods utilize the piggyBac and Sleeping Beauty (SB) transposon systems to mediate gene transfer, which can be combined with short hairpin RNAs or CRISPR/Cas9 constructs carrying guide RNAs to knock down tumor suppressors [28, 52, 125‒128]. One of the key advantages of this strategy is that it is entirely mouse strain independent and does not require time-intensive transgenics. Additionally, because transposon vectors have a high cargo capacity, they offer expanded opportunities to introduce genes of interest, including carrying multiple genes within the same vector [33]. Finally, IUE enables high spatial, anatomical, and temporal precision, which can be limited in other models. These 3 factors combined make IUE one of the most efficient strategies for glioma modeling, and combining transposition with shRNA/CRISPR allows the modeling of the entire range of mutations observed in patient tumors.

An early use of these systems to model gliomas used CRISPR/Cas9 to generate medulloblastoma and GBM, either through Ptch1 knockdown or a combination of TP53, PTEN, and NF1 knockdown, respectively [32, 125, 126]. Loss of ATRX – a common feature of both adult and pediatric gliomas – was investigated using the SB system alongside shRNA against p53 and the constitutively active NRASG12V mutation, a combination which led to gliomagenesis and reduced survival [127].

Recently, to produce reliable models of H3.3K27M-driven pHGGs and resolve the role of co-operating mutations, we used IUE of piggyBac and CRISPR/Cas9 constructs to introduce H3.3K27M and wildtype PDGFRA and to downregulate ATRX and p53 in embryonic neural progenitors in vivo. This combination was able to induce 100% penetrant glioma development, and further, PDGFRA and ATRX loss were dispensable for gliomagenesis, demonstrating that although they reduced latency and increased tumor aggressiveness, H3.3K27M and p53 loss alone were sufficient for transformation [28]. These models also showed that H3.3K27M tumors could be induced both in the forebrain and brainstem, where they recapitulated human tumors both molecularly and histologically, showing global loss of H3K27me3, high Olig2 expression, and characteristic diffuse infiltration. Importantly, H3.3K27M and p53 loss drove tumorigenesis only within a defined window during fetal brain development.

Subsequent studies have used IUE to model a range of other alterations found in H3K27M tumors, and to improve the efficiency of targeting the brainstem, the anatomical compartment in which these tumors are most often found in patients. Similar to the previous model, Patel and colleagues developed an IUE-based model of DMG driven by H3.3K27M, dominant-negative p53, and PDGFRAD842V and showed that PDGFb signaling differentially influences vascular morphology in forebrain and brainstem tumors [128]. This model was used to evaluate the treatment efficacy of everolimus with dasatinib, which led to first-in-human studies of the combined treatment [52]. Models carrying PDGFRAD842V were also used to evaluate cannabidiol to target ID1 expression in DMG, the role of FOXR2 as a pan-cancer oncogene, vulnerabilities to MDM2 inhibition in tumors expressing PPM1D truncation, and the pro-tumorigenic role of ATRX loss and ATRX-associated vulnerabilities [50, 53, 129, 130]. However, it is important to note that the PDGFRAD842V mutation is not as common in glioma patients as focal amplification of the wild-type gene, and as this mutation is potently oncogenic on its own, its presence may eclipse the roles of other partner alterations [17].

Recently, we developed a suite of IUE models to recapitulate the range of co-segregating alterations observed in patient pHGG tumors. These models include many of the most commonly observed alterations including H3.3K27M, H3.1K27M, and H3.3G34R histone mutations, p53 knockdown, PDGFRA amplification for DMG tumors, and characteristic PDGFRA mutations common to H3.3G34R-bearing cortical tumors, as well as other common co-segregating alterations in genes such as ACVR, PIK3CA, PPM1D, FGFR1, NF1, and others. These models demonstrated different latencies, histology, invasiveness, and stemness traits, reflecting the spatiotemporal and molecular heterogeneity of patient disease and elucidating the contribution of various driving mutations to tumorigenesis [131].

The relatively quick proliferation of these IUE-based models is a testament to the versatility of this technique to develop a wide range of patient-representative in vivo models. As with every model, there are drawbacks to transposon- and CRISPR-based IUE. As in RCAS models, there is no control over copy number variability and sites of genome integration, which may increase transformation capacity through high mutational burden or insertional mutagenesis. Furthermore, there is the possibility of non-physiologically relevant levels of overexpression of certain genes, as well as epigenetic transgene silencing [29]. Nevertheless, these models have proven to be valuable, highly flexible systems to investigate gliomagenic mechanisms and treatment vulnerabilities in pHGGs.

Recombination-Mediated Cassette Exchange Models

Another novel strategy for genetically engineering DMG models was recently developed, utilizing the dual recombinase-mediated cassette exchange (MADR) method [29]. Kim and colleagues designed donor plasmids in which a cassette carrying reporters and transgenes as well as miR-E, CRISPR/Cas9, and guide RNAs are flanked by loxP and Flp recombinase target sites. These plasmids were delivered in vivo along with a Flp-Cre expression vector into subventricular zone (SVZ) NPCs derived from Rosa26WT/mTmG mice. Fluorescent reporting confirmed stable, single-copy integration of the donor genes at the Rosa26 locus in neonatal mice. These models allow the introduction of a single transgene carrying multiple mutations into a safe harbor locus while maintaining allelic, copy number control. Thus, this strategy can combine the anatomical and temporal control of electroporation-based approaches with the genetic control possible in transgenics.

This strategy was used to develop both ependymoma and pediatric glioma mouse models. Pediatric gliomas were driven by mutations including dominant-negative p53, PDGFRAD842V, and either H3.3K27M or H3.3G34R, delivered into the neonatal SVZ in the forebrain. Tumors driven by either K27M or G34R mutations arose within ventral or dorsal forebrain regions, respectively, despite both being introduced into the SVZ. Importantly, by comparing MADR-induced tumors to human single-cell sequencing data, the group showed that K27M MADR gliomas recapitulated the heterogeneity and developmental hierarchy described in patient tumors [13, 29].

MADR models have some advantages over those using RCAS or IUE strategies employing CRISPR/Cas9 and piggyBac transposition. Importantly, integration loci are precisely defined rather than random, and it is possible to avoid the variable copy number integration common in other models. This strategy cannot be used on any mouse strain, however, as recipient GEMM mice must have dual recombinase sites to target with MADR. Even so, these sites are commonly introduced in routine generation of transgenic lines and so there is a long list of mouse lines that could be leveraged with this technique. Further development of MADR to allow evaluation of other partner alterations, along with refinement of inducible strategies and targeting of discrete progenitors (hindbrain vs. forebrain) during embryonic, as opposed to neonatal development, will expand opportunities to investigate key questions in pHGG biology.

Syngeneic Models

While groundbreaking advances have been made to expand genetic engineering strategies and model pHGGs in mice, even the most efficient de novo models require 6–12 months for tumors to fully develop [28, 30, 128]. Although investigations of tumor initiation necessitate de novo modeling, syngeneically engraftable systems are more efficient for translational studies that seek to evaluate therapeutic efficacy in frank tumors. These models are developed through the engraftment of tumor cells that are already transformed from primary glioma-bearing models into secondary recipients. This produces tumor development at a much faster rate and with higher penetrance, while retaining the tissue, immune and microenvironmental components needed for evaluating the mechanistic bases of treatment efficacy or failure.

Syngeneic orthotopic models of glioma have, until recently, been restricted to carcinogen- or transposon-induced cell lines such as GL261, CT-2A, and SMA-560 that bear little similarity to human tumors, bearing very high mutational burdens, substantially greater immunogenicity, and non-representative invasive behavior. Furthermore, these lines are often used to study adult GBM, not pediatric gliomas [67]. However, special consideration should be paid to the recently developed transposon-induced cell line, SB28. This syngeneically engraftable line was generated in C57BL6/J mice through Sleeping Beauty-mediated delivery of p53 shRNA, PDGFb, and NRASG12V [132, 133]. Compared to other cell lines, SB28 has a mutational load similar to human GBM, is not as immunogenic as other carcinogen-induced tumor cell lines, is able to infiltrate and invade in a manner more like human GBM, and better recapitulates patient response to chemoradiotherapy and immunotherapy [67]. It should be noted that NRAS may be a potent oncogenic driver in this system, considering its tumorigenic activity in other models, and Ras mutations are rare in histone mutant pediatric gliomas [17, 126, 134]. Even so, this cell line merits more widespread use when a syngeneic system is needed, particularly in the context of adult glioma.

One of the earliest attempts to develop a pHGG-specific syngeneic cell line used primary mouse astrocytes derived from p53-null neonates [135]. Based on sequencing of patient tumors, either PDGFRA wildtype or one of six patient-identified PDGFRA mutants was introduced, and cells were implanted orthotopically into the mouse cortex, resulting in tumorigenesis. Although this study modeled cortical pHGGs only, not DMGs, it is nevertheless one of the first examples of a reproducible syngeneic system to specifically study pediatric gliomas [135].

Recently, stable cell lines have been developed by excising tumor material from genetically engineered mouse and propagating cells in vitro. Kadiyala and colleagues used the SB system to study IDH1-mutant tumors, engineering mouse gliomas through knockdown of TP53 and ATRX and expression of NRASG12V, as well as either IDH1 wildtype or IDH1-R132H [136]. Additionally, IUE of piggyBac vectors have been used to induce tumors, and cells from these tumors can be grown ex vivo and engrafted orthotopically as syngeneic models. For example, cells from an IUE model carrying p53 mutation, PDGFRAD842V, and either wildtype H3.3, H3.3K27M, or H3.1K27M were injected orthotopically, where they recapitulated the histology, immune environment, and sensitivity to HDAC inhibitors seen in patient-derived DMG cell lines [137]. In our work, we have derived a set of sixteen orthotopically engraftable cell lines from IUE pHGG models, and these cell lines showed subtype-specific therapeutic vulnerabilities in vitro and in syngeneic models in vivo. In particular, these models were used to demonstrate specific efficacy of PDGFRA-targeted tyrosine kinase inhibitor avapritinib in a model driven by PDGFRA amplification, as well as vulnerability of G34R hemispheric tumors carrying PDGFRAC235Y to the tyrosine kinase inhibitor infigratinib [131].

These syngeneic models have so far proven useful for testing therapeutic strategies and for analysis of the relationship between driver and co-operating mutations. However, so far there are still only a handful of models available, representing only a few of the many co-segregating genetic alterations found in patients. As these techniques continue to develop, it will be important to model a wide range of alterations, or subtypes of this disease. These pHGG subtype models will allow testing of targeted therapy approaches and will also enable dissection of the precise roles of partner alterations in glioma development and TME interactions.

Until recently, detailed study of pHGG has been challenging due to the rarity of the tumors, lack of patient material, and absence of representative in vitro and in vivo models. However, the advances mentioned above have led to a dramatic expansion of knowledge surrounding the initiation, progression, and therapeutic opportunities for this disease. We will focus on a subset of these recent discoveries, discussing developments in (1) targeted precision therapy, (2) pHGG cellular origins, (3) tumor-neural networks, (4) the role of resident immune cells, and finally, (5) novel opportunities for immunotherapy (Fig. 2).

Fig. 2.

Novel insights into the underlying biology of pediatric gliomas and therapeutic strategies for targeting them. New biological insights into pediatric glioma include investigations of subtype-specific cells of origin, the existence of integrated tumor-neural networks, and their immune and microenvironmental landscapes. These insights, along with molecular characterization of patient tumors, have opened the door for novel therapeutic strategies. These strategies include CAR-T therapy, precision therapy targeting subtype-specific genetic and epigenetic alterations, and the combination of these approaches to sensitize tumors to standard chemoradiotherapy.

Fig. 2.

Novel insights into the underlying biology of pediatric gliomas and therapeutic strategies for targeting them. New biological insights into pediatric glioma include investigations of subtype-specific cells of origin, the existence of integrated tumor-neural networks, and their immune and microenvironmental landscapes. These insights, along with molecular characterization of patient tumors, have opened the door for novel therapeutic strategies. These strategies include CAR-T therapy, precision therapy targeting subtype-specific genetic and epigenetic alterations, and the combination of these approaches to sensitize tumors to standard chemoradiotherapy.

Close modal

Targeted Therapy

The expansion and growing accessibility of sequencing techniques has enabled deep evaluation of the genetic and transcriptomic landscapes of pHGG, shining a light on new opportunities for precision therapy. The use of TKIs has been a leading precision therapy strategy to target gliomas harboring specific tyrosine kinase alterations. For example, dasatinib, a BBB-penetrant PDGFRA inhibitor, was combined with either MEK inhibitor trametinib or mTOR inhibitor everolimus to reduce pHGG aggressiveness in vitro and in PDX and IUE models in vivo [45, 54]. Another highly specific TKI, avapritinib, was previously approved for PDGFRA-mutated gastrointestinal cancers, and as a result of successful repurposing in case studies of individual patients, the drug is currently undergoing phase 1/2 trials in pediatric patients with enhanced KIT and PDGFRA signaling, including those with K27M gliomas; this efficacy has also been confirmed in syngeneic models of PDGFRA-amplified H3K27M DMG in vivo (NCT04773782) [131, 138‒140]. Tumors harboring ACVR1 mutations have also been targeted with TKIs, in which the multi-kinase inhibitor vandetanib, combined with everolimus, was found to extend survival in ACVR1-mutant patient-derived DMG xenografts [54]. Suppressing ACVR1-mutant-associated ID1 expression was also found to reduce tumor cell invasion and proliferation in IUE and xenograft models [50].

There have also been recent advances in the specific targeting of histone mutant – or methylation-associated vulnerabilities. An early example is the combination of panobinostat, an HDAC inhibitor, with GSK-J4, a histone demethylase inhibitor, to sensitize H3K27M-xenografted tumors, although toxicity issues and development of resistance prevented widespread clinical use [45, 141, 142]. Chung and colleagues have recently shown that disrupting alpha ketoglutarate metabolism resulted in altered epigenetic programs and prolonged survival in H3K27M and IDH1R132H PDX glioma models. Similarly, reshaping the epigenetic landscape using antidiabetic drug metformin reduced patient-derived xenograft tumor growth in EZHIP-overexpressing posterior fossa group A ependymomas [143, 144]. An additional study by this group demonstrated an increased reliance on LIF/STAT3 signaling in H3.3G34R-mutant tumors, and targeting STAT3 signaling with a small molecule inhibitor showed promising results in patient-derived xenograft and syngeneic pHGG G34R models [51]. This is particularly promising as this molecule is currently in clinical trials for adult gliomas.

Another precision therapy strategy has been the targeting of DNA damage response (DDR) and cell cycle regulation pathways associated with sensitivity to radiotherapy and tumor progression. Recently, downregulation of DNA repair pathways was observed in H3.3G34R-driven tumors, and treatment with radiotherapy and the DDR inhibitors pamiparib and AZD7762 resulted in significantly improved survival in syngeneic and PDX models in vivo [49]. Enhanced sensitivity to radiation was likewise observed in tumors with ATRX loss, accompanied by increased activation of cell cycle regulator ATM. Adding an ATM inhibitor to standard radiotherapy markedly improved survival of PDX and syngeneic tumor-bearing mice specifically in the context of ATRX loss [130]. For tumors expressing truncation of PPM1D, inhibition of PPM1Dtrunc using shRNA or small molecular inhibitors in IUE and PDX preclinical models resulted in increased activity of DDR pathways and revealed a dependency on MDM2, leading to the inclusion of PPM1D-truncated tumors in a clinical trial of an MDM2 inhibitor for pediatric cancers [53, 145]; (NCT03654716). Alternatively, the PPM1Dtrunc reliance on the NAMPT pathway was targeted using the NAMPT inhibitor FK866, which resulted in restricted tumor growth and survival benefit in engineered isogenic xenograft models in vivo [56].

Targeting the TME has proven to be a promising orthologous approach to targeted tumor therapy. For example, it was recently shown that D-2-hydroxyglutarate (D-2-HG), an oncometabolite produced as a result of IDH1 mutation, disrupts transcriptional mechanisms in activated T cells, and that inhibition of D-2-HG results in increased survival and immunological memory against mutant IDH [136, 146]. Combining D-2-HG inhibition with PD-L1 immune checkpoint blockade resulted in 60% tumor regression in GEMM in vivo models [136]. An additional approach is the development of an IDH1R132H vaccine. Syngeneic models of vaccination indicated increased T-cell recruitment, higher levels of inflammatory cytokines, and survival benefit with IDH-mutant peptide-based vaccines, and one of these vaccines has progressed to clinical trials for patients with high-grade IDH1R132H glioma [146‒149].

As additional targets within the TME, macrophages and microglia rely on the macrophage colony stimulating factor-1 (CSF-1) and its receptor CSF-1R. Inhibition of CSF-1R has been shown to alter macrophage polarization, markedly reduce tumor size, and increase survival in syngeneic and PDX models of GBM [150‒152]. An additional study determined a synergistic effect of targeting the PD-1/PD-L1 axis along with CSF-1R inhibition in PDX and syngeneic models, recommending this as a potential approach for GBM [153]. While these studies have focused primarily on adult GBM, they may have implications for future developments in pediatric glioma as well. The wealth of knowledge generated as a result of new and expanded sequencing techniques, as well as faithful recapitulation of tumors in vitro and in vivo, supports a sustained push toward investigating targeted therapies based on genetic and transcriptional biomarkers.

Cell of Origin

Despite their histological similarities, different subtypes of pHGG possess distinct defining mutations, occur in different anatomical compartments, and derive from unique cells of origin [13, 16]. Recent studies for each of these pHGG subtypes have profoundly deepened our knowledge of the developmental origins of these tumors.

An early observation of pHGG etiology indicated that the presence of Nestin- and Olig2-positive NPCs in the ventral pons of normally developing children coincided precisely with the temporal incidence of diffuse intrinsic pontine gliomas, suggesting that a neural precursor-like population could be the cell of origin [24]. Subsequent work with iPSC-derived NPCs demonstrated that the H3.3K27M mutation increased proliferation of Nestin+/Sox2+/Olig2+ NPCs, but not hESCs or differentiated astrocytes, demonstrating that a specific temporal window was required for the H3K27M mutation to be oncogenic [70]. The development of IUE-based de novo mouse models of H3.3K27M glioma experimentally confirmed this hypothesis, showing that the H3.3K27M mutation is oncogenic when introduced specifically into prenatal NPCs [28].

Accordingly, H3.3K27M has been seen to accelerate tumor development when expressed in postnatal tumor models alongside p53 knockdown and constitutively active forms of PDGFRα, alterations which have been shown to drive tumorigenesis independent of the mutant histone [30, 117, 135]. H3.3K27M alone does not appear sufficient for oncogenesis and at minimum Trp53 loss is also required [28, 30]. However, in Larson et al. [30] 2019, although postnatal H3.3K27M and Trp53 loss were able to drive tumor development without the addition of a PDGFRA mutation, these tumors most often formed outside the brainstem – in the cerebellum and cortex. Therefore, to specifically induce brainstem gliomas driven by H3.3K27M, the current consensus is that prenatal NPCs are the likely cells of origin for H3.3K27M-driven tumors.

Subsequent genetic and transcriptional profiling provided new perspectives into specific cellular components of pHGGs. Identification of an OPC-like super-enhancer profile in DMG and their single-cell transcriptomic signatures indicated that these tumors were primarily composed of cycling, stem-like and OPC-like cells, and that only the OPC-like cells retained tumorigenic capacity [12, 13]. Mapping pHGG transcriptomic profiles onto a reference dataset of single-cell transcriptomic profiles of human prenatal tissue, researchers further demonstrated that the H3K27M mutation maintains glial-committed cells in a stem-like state and prevents differentiation along the glial lineages [14]. Together, these studies demonstrated that OPC-like cells are the putative cell of origin for H3K27M tumors.

Recently, an expansive profiling of 116 patient tumors compared DMG tumor profiles to the developmental waves of normal OPC specification. The study found that H3.1K27M tumors likely acquired the mutation during the earliest waves of OPC differentiation, in cells defined as NKX6-1+SHH-dependent ventral progenitors [38, 154]. In contrast, H3.3K27M tumors appeared to derive from a different PAX3+BMP-reliant dorsal progenitor, and thalamic tumors arose from more mature OPCs that had committed to the p2 stage during diencephalon development [38].

H3.3G34R/V tumors, on the other hand, were recently described as primarily neuronal malignancies, with a dual neuronal and astroglial identity, lacking OPC-like characteristics entirely [16]. The cell of origin was identified as a GSX2/DLX-expressing interneuron progenitor, clearly distinguishing H3.3G34R/V cortical tumors from their H3.3K27M midline counterparts [16]. The regional specificity of G34R/V mutant tumors was clarified using hESC-engineered and fetal-derived ventral forebrain and hindbrain NPCs, showing that the G34R/V mutation is oncogenic specifically in ventral forebrain interneuronal progenitor cells [69, 72]. Importantly, these results support the theory that pHGGs result from stalled development along NPC lineages, and further reinforce the spatiotemporal specificity of these specific cell vulnerabilities to oncohistone-mediated transformation.

Tumor-Neuronal Networks

Tumor-neuronal networks have recently been shown to contribute significantly to tumor development and progression. For decades, researchers have observed perineural growth of gliomas, and white matter tracts are one of the preferred routes for glioma cell invasion [155, 156]. Neuronal activity has a role in postnatal neurogenesis via GABA and glutamate signaling in the SVZ and subgranular zones, which are the locations for the cells of origin of adult GBM [157, 158], and where DMG cells are known to routinely spread [159]. Given that neurons can form synaptic connections with OPCs, the cells of origin for H3K27M pHGG, to promote mitosis and oligodendrogenesis [160, 161], it was anecdotally hypothesized that glioma cell:neuronal crosstalk could be occurring in vivo, and that this interaction had therapeutic implications.

However, it was not until Venkatesh, Monje and colleagues first investigated the role of neuronal activity in pHGG in 2015 that the extent of these glioma:neuron interactions came to be better appreciated. Their studies have shown that paracrine neuronal signaling promotes proliferation and growth of cortical, patient-derived pediatric glioma in vivo through secretion of the synaptic protein neuroligin-3 (NLGN3) and activation of the PI3K-mTOR pathway [47, 48]. Preventing NLGN3 shedding through ADAM10 inhibition was shown to block glioma growth in xenograft models [47]. Venkatesh and colleagues further demonstrated the existence of bona fide AMPA receptor-dependent synapses between neurons and glioma cells in H3K27M tumor models, confirmed upregulation of synaptic-associated genes in OPC-like cells from patient-derived cell lines, and observed that direct activation of glioma cell currents via AMPA receptors promoted glioma proliferation in vivo [22].

These findings were corroborated in adult gliomas, which harbor glutamatergic AMPA receptor-mediated synapses between neurons and glioma microtubes, suppression of which through anesthesia or AMPA receptor antagonists inhibited glioma invasion and proliferation in vivo [162]. Further, it has been shown that neuronal-glioma synapses can stimulate neuronal-like invasion of single pioneer glioma cells [163]. Subpopulations of glioma cells may also act as orchestrators of rhythmic intratumoral calcium transient oscillations, and targeting this activity through cell ablation or pharmacologic inhibition of potassium channels results in prolonged survival in vivo [164]. Interestingly, patient studies have shown glioma-mediated neuronal hyperexcitability, which could potentially act as a feedback mechanism further promoting tumor growth, and may have clinical implications for treatment, seizure-related complications, and supportive care strategies [22].

This seminal line of research, now termed cancer neuroscience, has deepened our understanding of the critical role of the brain microenvironment, and networked interactions with neurons in particular, in glioma progression. Further examining these intricate functional networks will hopefully offer actionable strategies for development of novel treatments.

Microglial Surveillance of TME

The immune axis is another critical element of the TME, and a concerted effort has been made in recent years to examine both the anti- and pro-tumor functions of immune cells in the context of high-grade glioma. Pediatric and adult HGGs are generally regarded as immunologically cold tumors, with limited T-cell infiltration [165]. Even so, myeloid cells including resident microglia and CNS-infiltrating macrophages, collectively referred to in the tumor context as tumor-associated macrophages (TAMs), can make up between 20 and 50% of glioma tumor tissue, accounting for a substantial proportion of the tumor bulk [165‒167].

Tumor-associated macrophages and microglia have been implicated in promoting gliomagenesis and tumor cell proliferation and invasion [167]. TAM-secreted factors have been shown to promote tumor growth and glioma cell invasion, with microglial depletion suppressing glioma proliferation, and, reciprocally, tumor cells have been suggested to secrete chemokines which attract macrophages [168‒171]. The precise roles of microglia versus infiltrating macrophages are continuing to be elucidated, as some studies suggest there may be differences in spatiotemporal tumor infiltration of the different cell populations [172, 173]. Importantly, there appear to be different infiltration and functional patterns of TAMs depending on the genetic characteristics of tumors. Among adult gliomas, studies show a higher degree of TAM infiltration in mesenchymal-categorized tumor samples and in tumors harboring mutations in NF1 [171, 174]. Additional studies indicate that IDH1-mutant tumors show reduced TAM infiltration and pro-inflammatory chemokine expression compared to IDH1-wildtype GBMs [122, 175].

Although the majority of research regarding glioma-associated macrophages has been focused on adult glioma, a growing number of studies suggest there is a difference in the role of TAMs between adult and pediatric brain tumors. Specifically, while enrichment of microglial and macrophage-associated genes was associated with poor survival in adults, this association was not found for pediatric patients [174]. Subsequent work has clarified that, while DMG tumors exhibit substantial TAM infiltration similar to adult tumors, the DMG-associated macrophages express fewer inflammatory cytokines and chemokines compared to adult GBM, which may explain the immunologically cold character of many DMGs [18, 176]. Interestingly, comparing patient samples of hemispheric pHGGs and DMGs, it was reported that DMGs express a more substantial inflammatory profile compared to hemispheric tumors, and targeting CCL3, a mediator of TAM infiltration, can suppress infiltration and extend survival [21]. These studies reinforce the imperative to treat pediatric HGGs independently of adult GBM and support future investigation of the role of driving mutations in tumor response to, and orchestration of, the microenvironment.

CAR-T Cells

Chimeric antigen receptor engineered T (CAR-T) cell therapy has resulted in significant benefit for patients with B-cell malignancies, and preliminary research suggesting new possibilities for CAR-T therapy in brain tumors, including in adult GBM, has led to several ongoing clinical trials [177‒180].

The use of CAR-T therapy could be particularly beneficial to pediatric glioma patients with diffuse or pontine tumors that cannot be resected, even if DMG tumors are generally classified as immunologically cold. To investigate this, Mount and colleagues identified the disialoganglioside GD2 to be uniformly highly expressed in both H3.3K27M and H3.1K27M populations [181]. They proceeded to develop GD2-CAR-T cells and showed complete tumor clearance in vivo in 2 patient-derived H3K27M xenografts. A follow-up experiment confirmed these results, though some mice suffered lethal toxicity effects and onset of GvHD. This work provided the basis for a phase 1 clinical trial. Of 4 patients with H3K27M or spinal cord DMG, three showed clinical and radiographic response to the therapy, and on-target, off-tumor toxicity was not observed [182]. This clinical response supports further development of GD2-CAR-T therapy for pHGG patients and is a significant milestone toward a hopeful, treatable future for DMG.

It should be noted, however, that CAR-T therapy can be accompanied by several challenges. In general, CAR-T therapy is patient specific, limiting large-scale development, increasing cost, and requiring a large, highly skilled team [177]. CAR-T therapy can also be accompanied by serious side effects. Patients receiving GD2-CAR-T therapy, for example, experienced tumor-adjacent CAR-T-mediated inflammation, a danger due to the sensitive location and existing swelling from the tumor itself, though this inflammation was managed with intensive inpatient care [182]. Additionally, one of the most significant challenges with CAR-T therapy is that, as with many targeted therapies, tumors may become resistant to therapy on account of antigen escape [177].

Despite these challenges, large-scale interest and investment has made CAR-T therapy the focus of several ongoing clinical trials in adult GBM, pHGG, as well as ependymoma, medulloblastoma, and neuroblastoma (NCT04510051, NCT05298995, NCT04099797, NCT04661384, NCT02442297). As this technique continues to be refined, CAR-T therapy may prove to be a valuable lifeline for patients with few treatment options and aggressive disease.

Dramatic developments in the models and resources to study pediatric gliomas have led to vital advances in our understanding of these tumors. These new insights have highlighted important therapeutic vulnerabilities and encouraged bold new treatment strategies. Despite this, the standard of care still fails to improve survival even by a few months for these patients. Expanded coordination and investment is required to make sure these research advances result in substantial clinical benefits. A deeper understanding of tumor origin will further elucidate developmental patterns of this disease and may inspire new strategies to reset normal developmental programs, or induce differentiation, to treat these tumors. Expanding our frameworks to include the roles of the TME will also provide new perspectives on tumor treatment strategies. Additionally, a more nuanced understanding of the role of co-segregating driving mutations will highlight targetable therapeutic vulnerabilities unique to each glioma subtype and will hopefully lead to better stratification of patients. This will also open the door to novel strategies for combinational therapy, which have already shown promising synergistic effects in recent preclinical studies and early clinical trials. It is necessary to further refine our models to accurately represent individual patient tumors in fully immunocompetent environments to identify targets and predict therapeutic efficacy. We also need continued commitment to inter-institutional collaboration to best share resources with a common focus on expanding patient care. We trust that a dedicated effort to developing and utilizing state-of-the-art technology and tumor models will lead to the development of novel, desperately needed treatments, and provide renewed hope for patients with pediatric glioma.

The authors have no conflicts of interest to declare.

A.F. is supported by the NIH Oxford-Cambridge Scholars Program and the Cambridge Trust. M.P. is supported by the Department of Oncology and the CRUK Children’s Brain Tumour Centre of Excellence, a Great Ormond Street Hospital Children’s Charity grant (V4020), a Brain Research UK grant (202021-28), and a CRUK Career Establishment Award (RCCCEA-May22\100003).

A.F. and M.P. wrote and edited the manuscript.

1.
Ostrom
QT
,
Price
M
,
Ryan
K
,
Edelson
J
,
Neff
C
,
Cioffi
G
et al
.
CBTRUS statistical report: pediatric brain tumor foundation childhood and adolescent primary brain and other central nervous system tumors diagnosed in the United States in 2014-2018
.
Neuro Oncol
.
2022
24
Suppl 3
iii1
38
.
2.
Sturm
D
,
Pfister
SM
,
Jones
DTW
.
Pediatric gliomas: current concepts on diagnosis, biology, and clinical management
.
J Clin Oncol
.
2017
;
35
(
21
):
2370
7
.
3.
Baker
SJ
,
Ellison
DW
,
Gutmann
DH
.
Pediatric gliomas as neurodevelopmental disorders
.
Glia
.
2016
;
64
(
6
):
879
95
.
4.
Wu
G
,
Diaz
AK
,
Paugh
BS
,
Rankin
SL
,
Ju
B
,
Li
Y
et al
.
The genomic landscape of diffuse intrinsic pontine glioma and pediatric non-brainstem high-grade glioma
.
Nat Genet
.
2014
;
46
(
5
):
444
50
.
5.
Sturm
D
,
Witt
H
,
Hovestadt
V
,
Khuong-Quang
DA
,
Jones
DT
,
Konermann
C
et al
.
Hotspot mutations in H3F3A and IDH1 define distinct epigenetic and biological subgroups of glioblastoma
.
Cancer Cell
.
2012
;
22
(
4
):
4425
37
.
6.
Schwartzentruber
J
,
Korshunov
A
,
Liu
XY
,
Jones
DT
,
Pfaff
E
,
Jacob
K
et al
.
Driver mutations in histone H3.3 and chromatin remodelling genes in paediatric glioblastoma
.
Nature
.
2012
;
482
(
7384
):
226
31
.
7.
Zhang
J
,
Wu
G
,
Miller
CP
,
Tatevossian
RG
,
Dalton
JD
,
Tang
B
et al
.
Whole-genome sequencing identifies genetic alterations in pediatric low-grade gliomas
.
Nat Genet
.
2013
;
45
(
6
):
602
12
.
8.
Jones
DT
,
Hutter
B
,
Jager
N
,
Korshunov
A
,
Kool
M
,
Warnatz
HJ
et al
.
Recurrent somatic alterations of FGFR1 and NTRK2 in pilocytic astrocytoma
.
Nat Genet
.
2013
;
45
(
8
):
927
32
.
9.
Bandopadhayay
P
,
Ramkissoon
LA
,
Jain
P
,
Bergthold
G
,
Wala
J
,
Zeid
R
et al
.
MYB-QKI rearrangements in angiocentric glioma drive tumorigenicity through a tripartite mechanism
.
Nat Genet
.
2016
;
48
(
3
):
273
82
.
10.
Jones
DT
,
Kocialkowski
S
,
Liu
L
,
Pearson
DM
,
Backlund
LM
,
Ichimura
K
et al
.
Tandem duplication producing a novel oncogenic BRAF fusion gene defines the majority of pilocytic astrocytomas
.
Cancer Res
.
2008
;
68
(
21
):
8673
7
.
11.
Capper
D
,
Jones
DTW
,
Sill
M
,
Hovestadt
V
,
Schrimpf
D
,
Sturm
D
et al
.
DNA methylation-based classification of central nervous system tumours
.
Nature
.
2018
;
555
(
7697
):
469
74
.
12.
Nagaraja
S
,
Vitanza
NA
,
Woo
PJ
,
Taylor
KR
,
Liu
F
,
Zhang
L
et al
.
Transcriptional dependencies in diffuse intrinsic pontine glioma
.
Cancer Cell
.
2017
;
31
(
5
):
635
52.e6
.
13.
Filbin
MG
,
Tirosh
I
,
Hovestadt
V
,
Shaw
ML
,
Escalante
LE
,
Mathewson
ND
et al
.
Developmental and oncogenic programs in H3K27M gliomas dissected by single-cell RNA-seq
.
Science
.
2018
;
360
(
6386
):
331
5
.
14.
Jessa
S
,
Blanchet-Cohen
A
,
Krug
B
,
Vladoiu
M
,
Coutelier
M
,
Faury
D
et al
.
Stalled developmental programs at the root of pediatric brain tumors
.
Nat Genet
.
2019
;
51
(
12
):
1702
13
.
15.
Behjati
S
,
Gilbertson
RJ
,
Pfister
SM
.
Maturation block in childhood cancer
.
Cancer Discov
.
2021
;
11
(
3
):
542
4
.
16.
Chen
CCL
,
Deshmukh
S
,
Jessa
S
,
Hadjadj
D
,
Lisi
V
,
Andrade
AF
et al
.
Histone H3.3g34-mutant interneuron progenitors Co-opt PDGFRA for gliomagenesis
.
Cell
.
2020
;
183
(
6
):
1617
33.e22
.
17.
Mackay
A
,
Burford
A
,
Carvalho
D
,
Izquierdo
E
,
Fazal-Salom
J
,
Taylor
KR
et al
.
Integrated molecular meta-analysis of 1,000 pediatric high-grade and diffuse intrinsic pontine glioma
.
Cancer Cell
.
2017
;
32
(
4
):
520
37.e5
.
18.
Lin
GL
,
Nagaraja
S
,
Filbin
MG
,
Suva
ML
,
Vogel
H
,
Monje
M
.
Non-inflammatory tumor microenvironment of diffuse intrinsic pontine glioma
.
Acta Neuropathol Commun
.
2018
;
6
(
1
):
51
.
19.
Mackay
A
,
Burford
A
,
Molinari
V
,
Jones
DTW
,
Izquierdo
E
,
Brouwer-Visser
J
et al
.
Molecular, pathological, radiological, and immune profiling of non-brainstem pediatric high-grade glioma from the HERBY phase II randomized trial
.
Cancer Cell
.
2018
;
33
(
5
):
829
42.e5
.
20.
Plant
AS
,
Koyama
S
,
Sinai
C
,
Solomon
IH
,
Griffin
GK
,
Ligon
KL
et al
.
Immunophenotyping of pediatric brain tumors: correlating immune infiltrate with histology, mutational load, and survival and assessing clonal T cell response
.
J Neuro Oncol
.
2018
;
137
(
2
):
269
78
.
21.
Ross
JL
,
Chen
Z
,
Herting
CJ
,
Grabovska
Y
,
Szulzewsky
F
,
Puigdelloses
M
et al
.
Platelet-derived growth factor beta is a potent inflammatory driver in paediatric high-grade glioma
.
Brain
.
2021
;
144
(
1
):
53
69
.
22.
Venkatesh
HS
,
Morishita
W
,
Geraghty
AC
,
Silverbush
D
,
Gillespie
SM
,
Arzt
M
et al
.
Electrical and synaptic integration of glioma into neural circuits
.
Nature
.
2019
;
573
(
7775
):
539
45
.
23.
Wei
X
,
Meel
MH
,
Breur
M
,
Bugiani
M
,
Hulleman
E
,
Phoenix
TN
.
Defining tumor-associated vascular heterogeneity in pediatric high-grade and diffuse midline gliomas
.
Acta Neuropathol Commun
.
2021
;
9
(
1
):
142
.
24.
Monje
M
,
Mitra
SS
,
Freret
ME
,
Raveh
TB
,
Kim
J
,
Masek
M
et al
.
Hedgehog-responsive candidate cell of origin for diffuse intrinsic pontine glioma
.
Proc Natl Acad Sci U S A
.
2011
;
108
(
11
):
4453
8
.
25.
Tsoli
M
,
Shen
H
,
Mayoh
C
,
Franshaw
L
,
Ehteda
A
,
Upton
D
et al
.
International experience in the development of patient-derived xenograft models of diffuse intrinsic pontine glioma
.
J Neuro Oncol
.
2019
;
141
(
2
):
253
63
.
26.
He
C
,
Xu
K
,
Zhu
X
,
Dunphy
PS
,
Gudenas
B
,
Lin
W
et al
.
Patient-derived models recapitulate heterogeneity of molecular signatures and drug response in pediatric high-grade glioma
.
Nat Commun
.
2021
;
12
(
1
):
4089
.
27.
Rybin
MJ
,
Ivan
ME
,
Ayad
NG
,
Zeier
Z
.
Organoid models of glioblastoma and their role in drug discovery
.
Front Cell Neurosci
.
2021
;
15
:
605255
.
28.
Pathania
M
,
De Jay
N
,
Maestro
N
,
Harutyunyan
AS
,
Nitarska
J
,
Pahlavan
P
et al
.
H3.3 (K27M) cooperates with Trp53 loss and PDGFRA gain in mouse embryonic neural progenitor cells to induce invasive high-grade gliomas
.
Cancer Cell
.
2017
;
32
(
5
):
684
700 e9
.
29.
Kim
GB
,
Rincon Fernandez Pacheco
D
,
Saxon
D
,
Yang
A
,
Sabet
S
,
Dutra-Clarke
M
et al
.
Rapid generation of somatic mouse mosaics with locus-specific, stably integrated transgenic elements
.
Cell
.
2019
;
179
(
1
):
251
67.e24
.
30.
Larson
JD
,
Kasper
LH
,
Paugh
BS
,
Jin
H
,
Wu
G
,
Kwon
CH
et al
.
Histone H3.3 K27M accelerates spontaneous brainstem glioma and drives restricted changes in bivalent gene expression
.
Cancer Cell
.
2019
;
35
(
1
):
140
55.e7
.
31.
Anastasaki
C
,
Chatterjee
J
,
Cobb
O
,
Sanapala
S
,
Scheaffer
SM
,
De Andrade Costa
A
et al
.
Human induced pluripotent stem cell engineering establishes a humanized mouse platform for pediatric low-grade glioma modeling
.
Acta Neuropathol Commun
.
2022
;
10
(
1
):
120
.
32.
Zuckermann
M
,
Hovestadt
V
,
Knobbe-Thomsen
CB
,
Zapatka
M
,
Northcott
PA
,
Schramm
K
et al
.
Somatic CRISPR/Cas9-mediated tumour suppressor disruption enables versatile brain tumour modelling
.
Nat Commun
.
2015
;
6
:
7391
.
33.
Gronych
J
,
Korshunov
A
,
Bageritz
J
,
Milde
T
,
Jugold
M
,
Hambardzumyan
D
et al
.
An activated mutant BRAF kinase domain is sufficient to induce pilocytic astrocytoma in mice
.
J Clin Invest
.
2011
;
121
(
4
):
1344
8
.
34.
Jones
C
,
Perryman
L
,
Hargrave
D
.
Paediatric and adult malignant glioma: close relatives or distant cousins
.
Nat Rev Clin Oncol
.
2012
;
9
(
7
):
400
13
.
35.
Jones
C
,
Baker
SJ
.
Unique genetic and epigenetic mechanisms driving paediatric diffuse high-grade glioma
.
Nat Rev Cancer
.
2014
;
14
(
10
):
651
61
.
36.
Wu
G
,
Broniscer
A
,
McEachron
TA
,
Lu
C
,
Paugh
BS
,
Becksfort
J
et al
.
Somatic histone H3 alterations in pediatric diffuse intrinsic pontine gliomas and non-brainstem glioblastomas
.
Nat Genet
.
2012
;
44
(
3
):
251
3
.
37.
Khuong-Quang
DA
,
Buczkowicz
P
,
Rakopoulos
P
,
Liu
XY
,
Fontebasso
AM
,
Bouffet
E
et al
.
K27M mutation in histone H3.3 defines clinically and biologically distinct subgroups of pediatric diffuse intrinsic pontine gliomas
.
Acta Neuropathol
.
2012
;
124
(
3
):
439
47
.
38.
Jessa
S
,
Mohammadnia
A
,
Harutyunyan
AS
,
Hulswit
M
,
Varadharajan
S
,
Lakkis
H
et al
.
K27M in canonical and noncanonical H3 variants occurs in distinct oligodendroglial cell lineages in brain midline gliomas
.
Nat Genet
.
2022
;
54
(
12
):
1865
80
.
39.
Jones
DTW
,
Bandopadhayay
P
,
Jabado
N
.
The power of human cancer genetics as revealed by low-grade gliomas
.
Annu Rev Genet
.
2019
;
53
:
483
503
.
40.
Manoharan
N
,
Liu
KX
,
Mueller
S
,
Haas-Kogan
DA
,
Bandopadhayay
P
.
Pediatric low-grade glioma: targeted therapeutics and clinical trials in the molecular era
.
Neoplasia
.
2023
;
36
:
100857
.
41.
Rowitch
DH
,
Kriegstein
AR
.
Developmental genetics of vertebrate glial-cell specification
.
Nature
.
2010
;
468
(
7321
):
214
22
.
42.
Masui
K
,
Mischel
PS
,
Reifenberger
G
.
Molecular classification of gliomas
.
Handb Clin Neurol
.
2016
;
134
:
97
120
.
43.
Izquierdo
E
,
Carvalho
DM
,
Mackay
A
,
Temelso
S
,
Boult
JKR
,
Pericoli
G
et al
.
DIPG harbors alterations targetable by MEK inhibitors, with acquired resistance mechanisms overcome by combinatorial inhibition
.
Cancer Discov
.
2022
;
12
(
3
):
712
29
.
44.
Smith
KS
,
Xu
K
,
Mercer
KS
,
Boop
F
,
Klimo
P
,
DeCupyere
M
et al
.
Patient-derived orthotopic xenografts of pediatric brain tumors: a St. Jude resource
.
Acta Neuropathol
.
2020
;
140
(
2
):
209
25
.
45.
Grasso
CS
,
Tang
Y
,
Truffaux
N
,
Berlow
NE
,
Liu
L
,
Debily
MA
et al
.
Functionally defined therapeutic targets in diffuse intrinsic pontine glioma
.
Nat Med
.
2015
;
21
(
7
):
827
9
.
46.
Anastas
JN
,
Zee
BM
,
Kalin
JH
,
Kim
M
,
Guo
R
,
Alexandrescu
S
et al
.
Re-Programing chromatin with a bifunctional LSD1/HDAC inhibitor induces therapeutic differentiation in DIPG
.
Cancer Cell
.
2019
;
36
(
5
):
528
44.e10
.
47.
Venkatesh
HS
,
Tam
LT
,
Woo
PJ
,
Lennon
J
,
Nagaraja
S
,
Gillespie
SM
et al
.
Targeting neuronal activity-regulated neuroligin-3 dependency in high-grade glioma
.
Nature
.
2017
;
549
(
7673
):
533
7
.
48.
Venkatesh
HS
,
Johung
TB
,
Caretti
V
,
Noll
A
,
Tang
Y
,
Nagaraja
S
et al
.
Neuronal activity promotes glioma growth through neuroligin-3 secretion
.
Cell
.
2015
;
161
(
4
):
803
16
.
49.
Haase
S
,
Banerjee
K
,
Mujeeb
AA
,
Hartlage
CS
,
Nunez
FM
,
Nunez
FJ
et al
.
H3.3-G34 mutations impair DNA repair and promote cGAS/STING-mediated immune responses in pediatric high-grade glioma models
.
J Clin Invest
.
2022
;
132
(
22
):
e154229
.
50.
Messinger
D
,
Harris
MK
,
Cummings
JR
,
Thomas
C
,
Yang
T
,
Sweha
SR
et al
.
Therapeutic targeting of prenatal pontine ID1 signaling in diffuse midline glioma
.
Neuro Oncol
.
2023
;
25
(
1
):
54
67
.
51.
Sweha
SR
,
Chung
C
,
Natarajan
SK
,
Panwalkar
P
,
Pun
M
,
Ghali
A
et al
.
Epigenetically defined therapeutic targeting in H3.3G34R/V high-grade gliomas
.
Sci Transl Med
.
2021
13
615
eabf7860
.
52.
Miklja
Z
,
Yadav
VN
,
Cartaxo
RT
,
Siada
R
,
Thomas
CC
,
Cummings
JR
et al
.
Everolimus improves the efficacy of dasatinib in PDGFRα-driven glioma
.
J Clin Invest
.
2020
;
130
(
10
):
5313
25
.
53.
Khadka
P
,
Reitman
ZJ
,
Lu
S
,
Buchan
G
,
Gionet
G
,
Dubois
F
et al
.
PPM1D mutations are oncogenic drivers of de novo diffuse midline glioma formation
.
Nat Commun
.
2022
;
13
(
1
):
604
.
54.
Carvalho
DM
,
Richardson
PJ
,
Olaciregui
N
,
Stankunaite
R
,
Lavarino
C
,
Molinari
V
et al
.
Repurposing vandetanib plus everolimus for the treatment of ACVR1-mutant diffuse intrinsic pontine glioma
.
Cancer Discov
.
2022
;
12
(
2
):
416
31
.
55.
Clarke
M
,
Mackay
A
,
Ismer
B
,
Pickles
JC
,
Tatevossian
RG
,
Newman
S
et al
.
Infant high-grade gliomas comprise multiple subgroups characterized by novel targetable gene fusions and favorable outcomes
.
Cancer Discov
.
2020
;
10
(
7
):
942
63
.
56.
Fons
NR
,
Sundaram
RK
,
Breuer
GA
,
Peng
S
,
McLean
RL
,
Kalathil
AN
et al
.
PPM1D mutations silence NAPRT gene expression and confer NAMPT inhibitor sensitivity in glioma
.
Nat Commun
.
2019
;
10
(
1
):
3790
.
57.
Gupta
N
,
Goumnerova
LC
,
Manley
P
,
Chi
SN
,
Neuberg
D
,
Puligandla
M
et al
.
Prospective feasibility and safety assessment of surgical biopsy for patients with newly diagnosed diffuse intrinsic pontine glioma
.
Neuro Oncol
.
2018
;
20
(
11
):
1547
55
.
58.
Roujeau
T
,
Machado
G
,
Garnett
MR
,
Miquel
C
,
Puget
S
,
Geoerger
B
et al
.
Stereotactic biopsy of diffuse pontine lesions in children
.
J Neurosurg
.
2007
107
1 Suppl
1
4
.
59.
Cage
TA
,
Samagh
SP
,
Mueller
S
,
Nicolaides
T
,
Haas-Kogan
D
,
Prados
M
et al
.
Feasibility, safety, and indications for surgical biopsy of intrinsic brainstem tumors in children
.
Childs Nerv Syst
.
2013
;
29
(
8
):
1313
9
.
60.
Li
Z
,
Langhans
SA
.
In vivo and ex vivo pediatric brain tumor models: an overview
.
Front Oncol
.
2021
;
11
:
620831
.
61.
Lee
J
,
Kotliarova
S
,
Kotliarov
Y
,
Li
A
,
Su
Q
,
Donin
NM
et al
.
Tumor stem cells derived from glioblastomas cultured in bFGF and EGF more closely mirror the phenotype and genotype of primary tumors than do serum-cultured cell lines
.
Cancer Cell
.
2006
;
9
(
5
):
391
403
.
62.
Laks
DR
,
Masterman-Smith
M
,
Visnyei
K
,
Angenieux
B
,
Orozco
NM
,
Foran
I
et al
.
Neurosphere formation is an independent predictor of clinical outcome in malignant glioma
.
Stem Cell
.
2009
;
27
(
4
):
980
7
.
63.
Wakimoto
H
,
Mohapatra
G
,
Kanai
R
,
Curry
WT
Jr
,
Yip
S
,
Nitta
M
et al
.
Maintenance of primary tumor phenotype and genotype in glioblastoma stem cells
.
Neuro Oncol
.
2012
;
14
(
2
):
132
44
.
64.
Singh
SK
,
Hawkins
C
,
Clarke
ID
,
Squire
JA
,
Bayani
J
,
Hide
T
et al
.
Identification of human brain tumour initiating cells
.
Nature
.
2004
;
432
(
7015
):
396
401
.
65.
Gimple
RC
,
Bhargava
S
,
Dixit
D
,
Rich
JN
.
Glioblastoma stem cells: lessons from the tumor hierarchy in a lethal cancer
.
Genes Dev
.
2019
33
11–12
591
609
.
66.
Akter
F
,
Simon
B
,
de Boer
NL
,
Redjal
N
,
Wakimoto
H
,
Shah
K
.
Pre-clinical tumor models of primary brain tumors: challenges and opportunities
.
Biochim Biophys Acta Rev Cancer
.
2021
;
1875
(
1
):
188458
.
67.
Haddad
AF
,
Young
JS
,
Amara
D
,
Berger
MS
,
Raleigh
DR
,
Aghi
MK
et al
.
Mouse models of glioblastoma for the evaluation of novel therapeutic strategies
.
Neurooncol Adv
.
2021
3
1
vdab100
.
68.
McAbee
JH
,
Rath
BH
,
Valdez
K
,
Young
DL
,
Wu
X
,
Shankavaram
UT
et al
.
Radiation drives the evolution of orthotopic xenografts initiated from glioblastoma stem-like cells
.
Cancer Res
.
2019
;
79
(
23
):
6032
43
.
69.
Funato
K
,
Smith
RC
,
Saito
Y
,
Tabar
V
.
Dissecting the impact of regional identity and the oncogenic role of human-specific NOTCH2NL in an hESC model of H3.3G34R-mutant glioma
.
Cell Stem Cell
.
2021
;
28
(
5
):
894
905 e7
.
70.
Funato
K
,
Major
T
,
Lewis
PW
,
Allis
CD
,
Tabar
V
.
Use of human embryonic stem cells to model pediatric gliomas with H3.3K27M histone mutation
.
Science
.
2014
;
346
(
6216
):
1529
33
.
71.
Haag
D
,
Mack
N
,
Benites Goncalves da Silva
P
,
Statz
B
,
Clark
J
,
Tanabe
K
et al
.
H3.3-K27M drives neural stem cell-specific gliomagenesis in a human iPSC-derived model
.
Cancer Cell
.
2021
;
39
(
3
):
407
22.e13
.
72.
Bressan
RB
,
Southgate
B
,
Ferguson
KM
,
Blin
C
,
Grant
V
,
Alfazema
N
et al
.
Regional identity of human neural stem cells determines oncogenic responses to histone H3.3 mutants
.
Cell Stem Cell
.
2021
;
28
(
5
):
877
93.e9
.
73.
Brien
GL
,
Bressan
RB
,
Monger
C
,
Gannon
D
,
Lagan
E
,
Doherty
AM
et al
.
Simultaneous disruption of PRC2 and enhancer function underlies histone H3.3-K27M oncogenic activity in human hindbrain neural stem cells
.
Nat Genet
.
2021
;
53
(
8
):
1221
32
.
74.
Sun
Y
,
Pollard
S
,
Conti
L
,
Toselli
M
,
Biella
G
,
Parkin
G
et al
.
Long-term tripotent differentiation capacity of human neural stem (NS) cells in adherent culture
.
Mol Cell Neurosci
.
2008
;
38
(
2
):
245
58
.
75.
Čančer
M
,
Hutter
S
,
Holmberg
KO
,
Rosen
G
,
Sundstrom
A
,
Tailor
J
et al
.
Humanized stem cell models of pediatric medulloblastoma reveal an oct4/mTOR Axis that promotes malignancy
.
Cell Stem Cell
.
2019
;
25
(
6
):
855
70.e11
.
76.
Jacob
F
,
Salinas
RD
,
Zhang
DY
,
Nguyen
PTT
,
Schnoll
JG
,
Wong
SZH
et al
.
A patient-derived glioblastoma organoid model and biobank recapitulates inter- and intra-tumoral heterogeneity
.
Cell
.
2020
;
180
(
1
):
188
204 e22
.
77.
Yi
HG
,
Jeong
YH
,
Kim
Y
,
Choi
YJ
,
Moon
HE
,
Park
SH
et al
.
A bioprinted human-glioblastoma-on-a-chip for the identification of patient-specific responses to chemoradiotherapy
.
Nat Biomed Eng
.
2019
;
3
(
7
):
509
19
.
78.
Tang
M
,
Xie
Q
,
Gimple
RC
,
Zhong
Z
,
Tam
T
,
Tian
J
et al
.
Three-dimensional bioprinted glioblastoma microenvironments model cellular dependencies and immune interactions
.
Cell Res
.
2020
;
30
(
10
):
833
53
.
79.
Lancaster
MA
,
Renner
M
,
Martin
CA
,
Wenzel
D
,
Bicknell
LS
,
Hurles
ME
et al
.
Cerebral organoids model human brain development and microcephaly
.
Nature
.
2013
;
501
(
7467
):
373
9
.
80.
Bian
S
,
Repic
M
,
Guo
Z
,
Kavirayani
A
,
Burkard
T
,
Bagley
JA
et al
.
Genetically engineered cerebral organoids model brain tumor formation
.
Nat Methods
.
2018
;
15
(
8
):
631
9
.
81.
Ogawa
J
,
Pao
GM
,
Shokhirev
MN
,
Verma
IM
.
Glioblastoma model using human cerebral organoids
.
Cell Rep
.
2018
;
23
(
4
):
1220
9
.
82.
Linkous
A
,
Balamatsias
D
,
Snuderl
M
,
Edwards
L
,
Miyaguchi
K
,
Milner
T
et al
.
Modeling patient-derived glioblastoma with cerebral organoids
.
Cell Rep
.
2019
;
26
(
12
):
3203
11.e5
.
83.
da Silva
B
,
Mathew
RK
,
Polson
ES
,
Williams
J
,
Wurdak
H
.
Spontaneous glioblastoma spheroid infiltration of early-stage cerebral organoids models brain tumor invasion
.
Slas Discov
.
2018
;
23
(
8
):
862
8
.
84.
Pine
AR
,
Cirigliano
SM
,
Nicholson
JG
,
Hu
Y
,
Linkous
A
,
Miyaguchi
K
et al
.
Tumor microenvironment is critical for the maintenance of cellular states found in primary glioblastomas
.
Cancer Discov
.
2020
;
10
(
7
):
964
79
.
85.
Krieger
TG
,
Tirier
SM
,
Park
J
,
Jechow
K
,
Eisemann
T
,
Peterziel
H
et al
.
Modeling glioblastoma invasion using human brain organoids and single-cell transcriptomics
.
Neuro Oncol
.
2020
;
22
(
8
):
1138
49
.
86.
Jermakowicz
AM
,
Rybin
MJ
,
Suter
RK
,
Sarkaria
JN
,
Zeier
Z
,
Feng
Y
et al
.
The novel BET inhibitor UM-002 reduces glioblastoma cell proliferation and invasion
.
Sci Rep
.
2021
;
11
(
1
):
23370
.
87.
Loong
HH
,
Wong
AM
,
Chan
DT
,
Cheung
MS
,
Chow
C
,
Ding
X
et al
.
Patient-derived tumor organoid predicts drugs response in glioblastoma: a step forward in personalized cancer therapy
.
J Clin Neurosci
.
2020
;
78
:
400
2
.
88.
Kook
MG
,
Lee
SE
,
Shin
N
,
Kong
D
,
Kim
DH
,
Kim
MS
et al
.
Generation of cortical brain organoid with vascularization by assembling with vascular spheroid
.
Int J Stem Cell
.
2022
;
15
(
1
):
85
94
.
89.
Martinov
T
,
McKenna
KM
,
Tan
WH
,
Collins
EJ
,
Kehret
AR
,
Linton
JD
et al
.
Building the next generation of humanized hemato-lymphoid system mice
.
Front Immunol
.
2021
;
12
:
643852
.
90.
Mian
SA
,
Anjos-Afonso
F
,
Bonnet
D
.
Advances in human immune system mouse models for studying human hematopoiesis and cancer immunotherapy
.
Front Immunol
.
2020
;
11
:
619236
.
91.
Ishikawa
F
,
Yasukawa
M
,
Lyons
B
,
Yoshida
S
,
Miyamoto
T
,
Yoshimoto
G
et al
.
Development of functional human blood and immune systems in NOD/SCID/IL2 receptor {gamma} chain (null) mice
.
Blood
.
2005
;
106
(
5
):
1565
73
.
92.
Shultz
LD
,
Schweitzer
PA
,
Christianson
SW
,
Gott
B
,
Schweitzer
IB
,
Tennent
B
et al
.
Multiple defects in innate and adaptive immunologic function in NOD/LtSz-scid mice
.
J Immunol
.
1995
;
154
(
1
):
180
91
.
93.
Bosma
GC
,
Custer
RP
,
Bosma
MJ
.
A severe combined immunodeficiency mutation in the mouse
.
Nature
.
1983
;
301
(
5900
):
527
30
.
94.
Lan
X
,
Kedziorek
DA
,
Chu
C
,
Jablonska
A
,
Li
S
,
Kai
M
et al
.
Modeling human pediatric and adult gliomas in immunocompetent mice through costimulatory blockade
.
Oncoimmunology
.
2020
;
9
(
1
):
1776577
.
95.
Li
S
,
Oh
BC
,
Chu
C
,
Arnold
A
,
Jablonska
A
,
Furtmuller
GJ
et al
.
Induction of immunological tolerance to myelinogenic glial-restricted progenitor allografts
.
Brain
.
2019
;
142
(
11
):
3456
72
.
96.
Semenkow
S
,
Li
S
,
Kahlert
UD
,
Raabe
EH
,
Xu
J
,
Arnold
A
et al
.
An immunocompetent mouse model of human glioblastoma
.
Oncotarget
.
2017
;
8
(
37
):
61072
82
.
97.
Cogels
MM
,
Rouas
R
,
Ghanem
GE
,
Martinive
P
,
Awada
A
,
Van Gestel
D
et al
.
Humanized mice as a valuable pre-clinical model for cancer immunotherapy research
.
Front Oncol
.
2021
;
11
:
784947
.
98.
Verma
B
,
Wesa
A
.
Establishment of humanized mice from peripheral blood mononuclear cells or cord blood CD34+ hematopoietic stem cells for immune-oncology studies evaluating new therapeutic agents
.
Curr Protoc Pharmacol
.
2020
;
89
(
1
):
e77
.
99.
Holyoake
TL
,
Nicolini
FE
,
Eaves
CJ
.
Functional differences between transplantable human hematopoietic stem cells from fetal liver, cord blood, and adult marrow
.
Exp Hematol
.
1999
;
27
(
9
):
1418
27
.
100.
Matas-Cespedes
A
,
Brown
L
,
Mahbubani
KT
,
Bareham
B
,
Higgins
J
,
Curran
M
et al
.
Use of human splenocytes in an innovative humanised mouse model for prediction of immunotherapy-induced cytokine release syndrome
.
Clin Transl Immunol
.
2020
;
9
(
11
):
e1202
.
101.
Tonomura
N
,
Habiro
K
,
Shimizu
A
,
Sykes
M
,
Yang
YG
.
Antigen-specific human T-cell responses and T cell-dependent production of human antibodies in a humanized mouse model
.
Blood
.
2008
;
111
(
8
):
4293
6
.
102.
Jangalwe
S
,
Shultz
LD
,
Mathew
A
,
Brehm
MA
.
Improved B cell development in humanized NOD-scid IL2Rγnull) mice transgenically expressing human stem cell factor, granulocyte-macrophage colony-stimulating factor and interleukin-3
.
Immun Inflamm Dis
.
2016
;
4
(
4
):
427
40
.
103.
Rongvaux
A
,
Willinger
T
,
Martinek
J
,
Strowig
T
,
Gearty
SV
,
Teichmann
LL
et al
.
Development and function of human innate immune cells in a humanized mouse model
.
Nat Biotechnol
.
2014
;
32
(
4
):
364
72
.
104.
Chiorazzi
M
,
Martinek
J
,
Krasnick
B
,
Zheng
Y
,
Robbins
KJ
,
Qu
R
et al
.
Autologous humanized PDX modeling for immuno-oncology recapitulates the human tumor microenvironment
.
BioRxiv
.
2022
2022
.
105.
Lenzen
A
,
Lauing
KL
,
Zhai
L
,
Ladomersky
E
,
Raman
P
,
Rathi
K
et al
.
Immu-01. Novel rna-targeting strategy for treating T cell-driven immunosuppression in human diffuse intrinsic pontine glioma
.
Neuro Oncol
.
2019
21
Suppl_2
ii92
3
.
106.
Kim
H
,
Kim
M
,
Im
SK
,
Fang
S
.
Mouse Cre-LoxP system: general principles to determine tissue-specific roles of target genes
.
Lab Anim Res
.
2018
;
34
(
4
):
147
59
.
107.
Huillard
E
,
Hashizume
R
,
Phillips
JJ
,
Griveau
A
,
Ihrie
RA
,
Aoki
Y
et al
.
Cooperative interactions of BRAFV600E kinase and CDKN2A locus deficiency in pediatric malignant astrocytoma as a basis for rational therapy
.
Proc Natl Acad Sci U S A
.
2012
;
109
(
22
):
8710
5
.
108.
Fortin
J
,
Tian
R
,
Zarrabi
I
,
Hill
G
,
Williams
E
,
Sanchez-Duffhues
G
et al
.
Mutant ACVR1 arrests glial cell differentiation to drive tumorigenesis in pediatric gliomas
.
Cancer Cell
.
2020
;
37
(
3
):
308
23.e12
.
109.
Kanvinde
PP
,
Malla
AP
,
Connolly
NP
,
Szulzewsky
F
,
Anastasiadis
P
,
Ames
HM
et al
.
Leveraging the replication-competent avian-like sarcoma virus/tumor virus receptor-A system for modeling human gliomas
.
Glia
.
2021
;
69
(
9
):
2059
76
.
110.
Holland
EC
,
Hively
WP
,
DePinho
RA
,
Varmus
HE
.
A constitutively active epidermal growth factor receptor cooperates with disruption of G1 cell-cycle arrest pathways to induce glioma-like lesions in mice
.
Genes Dev
.
1998
;
12
(
23
):
3675
85
.
111.
Holland
EC
,
Varmus
HE
.
Basic fibroblast growth factor induces cell migration and proliferation after glia-specific gene transfer in mice
.
Proc Natl Acad Sci U S A
.
1998
;
95
(
3
):
1218
23
.
112.
Shih
AH
,
Dai
C
,
Hu
X
,
Rosenblum
MK
,
Koutcher
JA
,
Holland
EC
.
Dose-dependent effects of platelet-derived growth factor-B on glial tumorigenesis
.
Cancer Res
.
2004
;
64
(
14
):
4783
9
.
113.
Becher
OJ
,
Hambardzumyan
D
,
Walker
TR
,
Helmy
K
,
Nazarian
J
,
Albrecht
S
et al
.
Preclinical evaluation of radiation and perifosine in a genetically and histologically accurate model of brainstem glioma
.
Cancer Res
.
2010
;
70
(
6
):
2548
57
.
114.
Hambardzumyan
D
,
Amankulor
NM
,
Helmy
KY
,
Becher
OJ
,
Holland
EC
.
Modeling adult gliomas using RCAS/t-va technology
.
Transl Oncol
.
2009
;
2
(
2
):
89
95
.
115.
Barton
KL
,
Misuraca
K
,
Cordero
F
,
Dobrikova
E
,
Min
HD
,
Gromeier
M
et al
.
PD-0332991, a CDK4/6 inhibitor, significantly prolongs survival in a genetically engineered mouse model of brainstem glioma
.
PLoS One
.
2013
;
8
(
10
):
e77639
.
116.
Lewis
PW
,
Muller
MM
,
Koletsky
MS
,
Cordero
F
,
Lin
S
,
Banaszynski
LA
et al
.
Inhibition of PRC2 activity by a gain-of-function H3 mutation found in pediatric glioblastoma
.
Science
.
2013
;
340
(
6134
):
857
61
.
117.
Cordero
FJ
,
Huang
Z
,
Grenier
C
,
He
X
,
Hu
G
,
McLendon
RE
et al
.
Histone H3.3K27M represses p16 to accelerate gliomagenesis in a murine model of DIPG
.
Mol Cancer Res
.
2017
;
15
(
9
):
1243
54
.
118.
Tomita
Y
,
Shimazu
Y
,
Somasundaram
A
,
Tanaka
Y
,
Takata
N
,
Ishi
Y
et al
.
A novel mouse model of diffuse midline glioma initiated in neonatal oligodendrocyte progenitor cells highlights cell-of-origin dependent effects of H3K27M
.
Glia
.
2022
;
70
(
9
):
1681
98
.
119.
Misuraca
KL
,
Hu
G
,
Barton
KL
,
Chung
A
,
Becher
OJ
.
A novel mouse model of diffuse intrinsic pontine glioma initiated in pax3-expressing cells
.
Neoplasia
.
2016
;
18
(
1
):
60
70
.
120.
Hoeman
CM
,
Cordero
FJ
,
Hu
G
,
Misuraca
K
,
Romero
MM
,
Cardona
HJ
et al
.
ACVR1 R206H cooperates with H3.1K27M in promoting diffuse intrinsic pontine glioma pathogenesis
.
Nat Commun
.
2019
;
10
(
1
):
1023
.
121.
Philip
B
,
Yu
DX
,
Silvis
MR
,
Shin
CH
,
Robinson
JP
,
Robinson
GL
et al
.
Mutant IDH1 promotes glioma formation in vivo
.
Cell Rep
.
2018
;
23
(
5
):
1553
64
.
122.
Amankulor
NM
,
Kim
Y
,
Arora
S
,
Kargl
J
,
Szulzewsky
F
,
Hanke
M
et al
.
Mutant IDH1 regulates the tumor-associated immune system in gliomas
.
Genes Dev
.
2017
;
31
(
8
):
774
86
.
123.
Saito
T
.
In vivo electroporation in the embryonic mouse central nervous system
.
Nat Protoc
.
2006
;
1
(
3
):
1552
8
.
124.
Saito
T
,
Nakatsuji
N
.
Efficient gene transfer into the embryonic mouse brain using in vivo electroporation
.
Dev Biol
.
2001
;
240
(
1
):
237
46
.
125.
Chen
F
,
Rosiene
J
,
Che
A
,
Becker
A
,
LoTurco
J
.
Tracking and transforming neocortical progenitors by CRISPR/Cas9 gene targeting and piggyBac transposase lineage labeling
.
Development
.
2015
;
142
(
20
):
3601
11
.
126.
Chen
F
,
Becker
AJ
,
Loturco
JJ
.
Contribution of tumor heterogeneity in a new animal model of CNS tumors
.
Mol Cancer Res
.
2014
;
12
(
5
):
742
53
.
127.
Koschmann
C
,
Calinescu
AA
,
Nunez
FJ
,
Mackay
A
,
Fazal-Salom
J
,
Thomas
D
et al
.
ATRX loss promotes tumor growth and impairs nonhomologous end joining DNA repair in glioma
.
Sci Transl Med
.
2016
8
328
328ra28
.
128.
Patel
SK
,
Hartley
RM
,
Wei
X
,
Furnish
R
,
Escobar-Riquelme
F
,
Bear
H
et al
.
Generation of diffuse intrinsic pontine glioma mouse models by brainstem-targeted in utero electroporation
.
Neuro Oncol
.
2020
;
22
(
3
):
381
92
.
129.
Tsai
JW
,
Cejas
P
,
Wang
DK
,
Patel
S
,
Wu
DW
,
Arounleut
P
et al
.
FOXR2 is an epigenetically regulated pan-cancer oncogene that activates ETS transcriptional circuits
.
Cancer Res
.
2022
;
82
(
17
):
2980
3001
.
130.
Qin
T
,
Mullan
B
,
Ravindran
R
,
Messinger
D
,
Siada
R
,
Cummings
JR
et al
.
ATRX loss in glioma results in dysregulation of cell-cycle phase transition and ATM inhibitor radio-sensitization
.
Cell Rep
.
2022
;
38
(
2
):
110216
.
131.
McNicholas
M
,
De Cola
A
,
Bashardanesh
Z
,
Foss
A
,
Lloyd
CB
,
Hebert
S
et al
.
A compendium of syngeneic, transplantable pediatric high-grade glioma models reveals subtype-specific therapeutic vulnerabilities
.
Cancer Discov
.
2023
CD-23-0004
.
132.
Genoud
V
,
Marinari
E
,
Nikolaev
SI
,
Castle
JC
,
Bukur
V
,
Dietrich
PY
et al
.
Responsiveness to anti-PD-1 and anti-CTLA-4 immune checkpoint blockade in SB28 and GL261 mouse glioma models
.
Oncoimmunology
.
2018
;
7
(
12
):
e1501137
.
133.
Kosaka
A
,
Ohkuri
T
,
Okada
H
.
Combination of an agonistic anti-CD40 monoclonal antibody and the COX-2 inhibitor celecoxib induces anti-glioma effects by promotion of type-1 immunity in myeloid cells and T-cells
.
Cancer Immunol Immunother
.
2014
;
63
(
8
):
847
57
.
134.
Breunig
JJ
,
Levy
R
,
Antonuk
CD
,
Molina
J
,
Dutra-Clarke
M
,
Park
H
et al
.
Ets factors regulate neural stem cell depletion and gliogenesis in Ras pathway glioma
.
Cell Rep
.
2015
;
12
(
2
):
258
71
.
135.
Paugh
BS
,
Zhu
X
,
Qu
C
,
Endersby
R
,
Diaz
AK
,
Zhang
J
et al
.
Novel oncogenic PDGFRA mutations in pediatric high-grade gliomas
.
Cancer Res
.
2013
;
73
(
20
):
6219
29
.
136.
Kadiyala
P
,
Carney
SV
,
Gauss
JC
,
Garcia-Fabiani
MB
,
Haase
S
,
Alghamri
MS
et al
.
Inhibition of 2-hydroxyglutarate elicits metabolic reprogramming and mutant IDH1 glioma immunity in mice
.
J Clin Invest
.
2021
;
131
(
4
):
e139542
.
137.
du Chatinier
A
,
Meel
MH
,
Das
AI
,
Metselaar
DS
,
Waranecki
P
,
Bugiani
M
et al
.
Generation of immunocompetent syngeneic allograft mouse models for pediatric diffuse midline glioma
.
Neurooncol Adv
.
2022
4
1
vdac079
.
138.
von Mehren
M
,
Heinrich
MC
,
Shi
H
,
Iannazzo
S
,
Mankoski
R
,
Dimitrijević
S
et al
.
Clinical efficacy comparison of avapritinib with other tyrosine kinase inhibitors in gastrointestinal stromal tumors with PDGFRA D842V mutation: a retrospective analysis of clinical trial and real-world data
.
BMC Cancer
.
2021
;
21
(
1
):
291
.
139.
Mayr
L
,
Madlener
S
,
Weiler-Wichtl
L
,
Rosenmayr
V
,
Furtner-Srajer
J
,
Guntner
A
et al
.
DIPG-60. Avapritinib for targeting PDGFRA in H3K27M: mutated diffuse midline glioma
.
Neuro Oncol
.
2022
24
Suppl 1
i32
.
140.
Mayr
L
,
Madlener
S
,
Verena
R
,
Schmook
M
,
Stepien
N
,
Baumgartner
A-C
et al
.
Innv-17. Response to avapritinib in a pediatric spinal cord H3k27m-mutant glioma patient
.
Neuro Oncol
.
2021
23
Suppl 6
vi
108
.
141.
Hennika
T
,
Hu
G
,
Olaciregui
NG
,
Barton
KL
,
Ehteda
A
,
Chitranjan
A
et al
.
Pre-clinical study of panobinostat in xenograft and genetically engineered murine diffuse intrinsic pontine glioma models
.
PLoS One
.
2017
;
12
(
1
):
e0169485
.
142.
Meel
MH
,
de Gooijer
MC
,
Metselaar
DS
,
Sewing
ACP
,
Zwaan
K
,
Waranecki
P
et al
.
Combined therapy of AXL and HDAC inhibition reverses mesenchymal transition in diffuse intrinsic pontine glioma
.
Clin Cancer Res
.
2020
;
26
(
13
):
3319
32
.
143.
Panwalkar
P
,
Tamrazi
B
,
Dang
D
,
Chung
C
,
Sweha
S
,
Natarajan
SK
et al
.
Targeting integrated epigenetic and metabolic pathways in lethal childhood PFA ependymomas
.
Sci Transl Med
.
2021
13
614
eabc0497
.
144.
Chung
C
,
Sweha
SR
,
Pratt
D
,
Tamrazi
B
,
Panwalkar
P
,
Banda
A
et al
.
Integrated metabolic and epigenomic reprograming by H3K27M mutations in diffuse intrinsic pontine gliomas
.
Cancer Cell
.
2020
;
38
(
3
):
334
49.e9
.
145.
Xu
C
,
Liu
H
,
Pirozzi
CJ
,
Chen
LH
,
Greer
PK
,
Diplas
BH
et al
.
TP53 wild-type/PPM1D mutant diffuse intrinsic pontine gliomas are sensitive to a MDM2 antagonist
.
Acta Neuropathol Commun
.
2021
;
9
(
1
):
178
.
146.
Bunse
L
,
Pusch
S
,
Bunse
T
,
Sahm
F
,
Sanghvi
K
,
Friedrich
M
et al
.
Suppression of antitumor T cell immunity by the oncometabolite (R)-2-hydroxyglutarate
.
Nat Med
.
2018
;
24
(
8
):
1192
203
.
147.
Schumacher
T
,
Bunse
L
,
Pusch
S
,
Sahm
F
,
Wiestler
B
,
Quandt
J
et al
.
A vaccine targeting mutant IDH1 induces antitumour immunity
.
Nature
.
2014
;
512
(
7514
):
324
7
.
148.
Pellegatta
S
,
Valletta
L
,
Corbetta
C
,
Patane
M
,
Zucca
I
,
Riccardi Sirtori
F
et al
.
Effective immuno-targeting of the IDH1 mutation R132H in a murine model of intracranial glioma
.
Acta Neuropathol Commun
.
2015
;
3
:
4
.
149.
Platten
M
,
Bunse
L
,
Wick
A
,
Bunse
T
,
Le Cornet
L
,
Harting
I
et al
.
A vaccine targeting mutant IDH1 in newly diagnosed glioma
.
Nature
.
2021
;
592
(
7854
):
463
8
.
150.
Pyonteck
SM
,
Akkari
L
,
Schuhmacher
AJ
,
Bowman
RL
,
Sevenich
L
,
Quail
DF
et al
.
CSF-1R inhibition alters macrophage polarization and blocks glioma progression
.
Nat Med
.
2013
;
19
(
10
):
1264
72
.
151.
Przystal
JM
,
Becker
H
,
Canjuga
D
,
Tsiami
F
,
Anderle
N
,
Keller
AL
et al
.
Targeting CSF1R alone or in combination with PD1 in experimental glioma
.
Cancers
.
2021
;
13
(
10
):
2400
.
152.
Quail
DF
,
Bowman
RL
,
Akkari
L
,
Quick
ML
,
Schuhmacher
AJ
,
Huse
JT
et al
.
The tumor microenvironment underlies acquired resistance to CSF-1R inhibition in gliomas
.
Science
.
2016
352
6288
aad3018
.
153.
Antonios
JP
,
Soto
H
,
Everson
RG
,
Moughon
D
,
Orpilla
JR
,
Shin
NP
et al
.
Immunosuppressive tumor-infiltrating myeloid cells mediate adaptive immune resistance via a PD-1/PD-L1 mechanism in glioblastoma
.
Neuro Oncol
.
2017
;
19
(
6
):
796
807
.
154.
Kessaris
N
,
Fogarty
M
,
Iannarelli
P
,
Grist
M
,
Wegner
M
,
Richardson
WD
.
Competing waves of oligodendrocytes in the forebrain and postnatal elimination of an embryonic lineage
.
Nat Neurosci
.
2006
;
9
(
2
):
173
9
.
155.
Cuddapah
VA
,
Robel
S
,
Watkins
S
,
Sontheimer
H
.
A neurocentric perspective on glioma invasion
.
Nat Rev Neurosci
.
2014
;
15
(
7
):
455
65
.
156.
Scherer
HJ
.
Structural development in gliomas
.
Am J Cancer
.
1938
;
34
(
3
):
333
51
.
157.
Pathania
M
,
Yan
LD
,
Bordey
A
.
A symphony of signals conducts early and late stages of adult neurogenesis
.
Neuropharmacology
.
2010
;
58
(
6
):
865
76
.
158.
Lee
JH
,
Lee
JE
,
Kahng
JY
,
Kim
SH
,
Park
JS
,
Yoon
SJ
et al
.
Human glioblastoma arises from subventricular zone cells with low-level driver mutations
.
Nature
.
2018
;
560
(
7717
):
243
7
.
159.
Caretti
V
,
Bugiani
M
,
Freret
M
,
Schellen
P
,
Jansen
M
,
van Vuurden
D
et al
.
Subventricular spread of diffuse intrinsic pontine glioma
.
Acta Neuropathol
.
2014
;
128
(
4
):
605
7
.
160.
Bergles
DE
,
Roberts
JD
,
Somogyi
P
,
Jahr
CE
.
Glutamatergic synapses on oligodendrocyte precursor cells in the hippocampus
.
Nature
.
2000
;
405
(
6783
):
187
91
.
161.
Gibson
EM
,
Purger
D
,
Mount
CW
,
Goldstein
AK
,
Lin
GL
,
Wood
LS
et al
.
Neuronal activity promotes oligodendrogenesis and adaptive myelination in the mammalian brain
.
Science
.
2014
;
344
(
6183
):
1252304
.
162.
Venkataramani
V
,
Tanev
DI
,
Strahle
C
,
Studier-Fischer
A
,
Fankhauser
L
,
Kessler
T
et al
.
Glutamatergic synaptic input to glioma cells drives brain tumour progression
.
Nature
.
2019
;
573
(
7775
):
532
8
.
163.
Venkataramani
V
,
Yang
Y
,
Schubert
MC
,
Reyhan
E
,
Tetzlaff
SK
,
Wißmann
N
et al
.
Glioblastoma hijacks neuronal mechanisms for brain invasion
.
Cell
.
2022
;
185
(
16
):
2899
917.e31
.
164.
Hausmann
D
,
Hoffmann
DC
,
Venkataramani
V
,
Jung
E
,
Horschitz
S
,
Tetzlaff
SK
et al
.
Autonomous rhythmic activity in glioma networks drives brain tumour growth
.
Nature
.
2023
;
613
(
7942
):
179
86
.
165.
Price
G
,
Bouras
A
,
Hambardzumyan
D
,
Hadjipanayis
CG
.
Current knowledge on the immune microenvironment and emerging immunotherapies in diffuse midline glioma
.
EBioMedicine
.
2021
;
69
:
103453
.
166.
Simmons
GW
,
Pong
WW
,
Emnett
RJ
,
White
CR
,
Gianino
SM
,
Rodriguez
FJ
et al
.
Neurofibromatosis-1 heterozygosity increases microglia in a spatially and temporally restricted pattern relevant to mouse optic glioma formation and growth
.
J Neuropathol Exp Neurol
.
2011
;
70
(
1
):
51
62
.
167.
Gutmann
DH
,
Kettenmann
H
.
Microglia/brain macrophages as central drivers of brain tumor pathobiology
.
Neuron
.
2019
;
104
(
3
):
442
9
.
168.
Hambardzumyan
D
,
Gutmann
DH
,
Kettenmann
H
.
The role of microglia and macrophages in glioma maintenance and progression
.
Nat Neurosci
.
2016
;
19
(
1
):
20
7
.
169.
Chen
P
,
Zhao
D
,
Li
J
,
Liang
X
,
Li
J
,
Chang
A
et al
.
Symbiotic macrophage-glioma cell interactions reveal synthetic lethality in PTEN-null glioma
.
Cancer Cell
.
2019
;
35
(
6
):
868
84.e6
.
170.
Wallmann
T
,
Zhang
XM
,
Wallerius
M
,
Bolin
S
,
Joly
AL
,
Sobocki
C
et al
.
Microglia induce PDGFRB expression in glioma cells to enhance their migratory capacity
.
iScience
.
2018
;
9
:
71
83
.
171.
Guo
X
,
Pan
Y
,
Gutmann
DH
.
Genetic and genomic alterations differentially dictate low-grade glioma growth through cancer stem cell-specific chemokine recruitment of T cells and microglia
.
Neuro Oncol
.
2019
;
21
(
10
):
1250
62
.
172.
Chen
Z
,
Feng
X
,
Herting
CJ
,
Garcia
VA
,
Nie
K
,
Pong
WW
et al
.
Cellular and molecular identity of tumor-associated macrophages in glioblastoma
.
Cancer Res
.
2017
;
77
(
9
):
2266
78
.
173.
Hutter
G
,
Theruvath
J
,
Graef
CM
,
Zhang
M
,
Schoen
MK
,
Manz
EM
et al
.
Microglia are effector cells of CD47-SIRPα antiphagocytic axis disruption against glioblastoma
.
Proc Natl Acad Sci U S A
.
2019
;
116
(
3
):
997
1006
.
174.
Engler
JR
,
Robinson
AE
,
Smirnov
I
,
Hodgson
JG
,
Berger
MS
,
Gupta
N
et al
.
Increased microglia/macrophage gene expression in a subset of adult and pediatric astrocytomas
.
PLoS One
.
2012
;
7
(
8
):
e43339
.
175.
Han
CJ
,
Zheng
JY
,
Sun
L
,
Yang
HC
,
Cao
ZQ
,
Zhang
XH
et al
.
The oncometabolite 2-hydroxyglutarate inhibits microglial activation via the AMPK/mTOR/NF-κB pathway
.
Acta Pharmacol Sin
.
2019
;
40
(
10
):
1292
302
.
176.
Caretti
V
,
Sewing
AC
,
Lagerweij
T
,
Schellen
P
,
Bugiani
M
,
Jansen
MH
et al
.
Human pontine glioma cells can induce murine tumors
.
Acta Neuropathol
.
2014
;
127
(
6
):
897
909
.
177.
Sterner
RC
,
Sterner
RM
.
CAR-T cell therapy: current limitations and potential strategies
.
Blood Cancer J
.
2021
;
11
(
4
):
69
.
178.
Brown
CE
,
Alizadeh
D
,
Starr
R
,
Weng
L
,
Wagner
JR
,
Naranjo
A
et al
.
Regression of glioblastoma after chimeric antigen receptor T-cell therapy
.
N Engl J Med
.
2016
;
375
(
26
):
2561
9
.
179.
Ahmed
N
,
Brawley
V
,
Hegde
M
,
Bielamowicz
K
,
Kalra
M
,
Landi
D
et al
.
HER2-Specific chimeric antigen receptor-modified virus-specific T cells for progressive glioblastoma: a phase 1 dose-escalation trial
.
JAMA Oncol
.
2017
;
3
(
8
):
1094
101
.
180.
O’Rourke
DM
,
Nasrallah
MP
,
Desai
A
,
Melenhorst
JJ
,
Mansfield
K
,
Morrissette
JJD
et al
.
A single dose of peripherally infused EGFRvIII-directed CAR T cells mediates antigen loss and induces adaptive resistance in patients with recurrent glioblastoma
.
Sci Transl Med
.
2017
9
399
eaaa0984
.
181.
Mount
CW
,
Majzner
RG
,
Sundaresh
S
,
Arnold
EP
,
Kadapakkam
M
,
Haile
S
et al
.
Potent antitumor efficacy of anti-GD2 CAR T cells in H3-K27M(+) diffuse midline gliomas
.
Nat Med
.
2018
;
24
(
5
):
572
9
.
182.
Majzner
RG
,
Ramakrishna
S
,
Yeom
KW
,
Patel
S
,
Chinnasamy
H
,
Schultz
LM
et al
.
GD2-CAR T cell therapy for H3K27M-mutated diffuse midline gliomas
.
Nature
.
2022
;
603
(
7903
):
934
41
.