Clinical manifestations

Presentation and course

Symptoms vary based on involvement of different midline structures and develop over 1 to 2 months (36; 16; 50). Cranial neuropathies can cause expected symptomatology: diplopia, ptosis, facial pain or numbness, facial palsy, tinnitus, hearing loss, vertigo, dysarthria, and dysphagia. Involvement of motor and sensory pathways causes weakness, hypoesthesia, hyperesthesia, and gait abnormalities. High CSF pressure causes headache, nausea, and vomiting. Other CNS symptoms include seizures and persistent hiccups.

H3 K27M mutant diffuse midline gliomas have been extensively described in children and are rarely seen in adults (61). In children, H3 K27M-mutant gliomas occur frequently within the brainstem, particularly the pons, whereas in adults they occur more frequently within the thalamus and spinal cord. Rare adult cases have been reported in the corpus callosum, hypothalamus, pineal region, basal ganglia, and third ventricle. In children, they occur primarily at a median age of 7 years (5 to 11 years) in the pons and at 11 years in the thalamus. In adults, median age at diagnosis is 35 years (range 18 to 82 years). There is no gender preponderance.

In a retrospective study of adult midline gliomas, the H3 K27M mutation was identified in 15% cases (59). Most common midline locations for H3 K27M-mutated tumors were midbrain, pons, and cerebellum. In the pediatric population, this mutation is identified in approximately 27% of gliomas (42).

Prognosis and complications

H3 K27M-mutant midline gliomas are associated with poor prognosis and were designated as a grade 4 entity in the 2016 WHO Classification (36). However, there is heterogeneity in the behavior of gliomas with H3 K27M mutation. Prognosis for circumscribed gliomas, thalamic, or diffuse non-midline gliomas with H3 K27M mutation remains uncertain (20; 45; 47; 48).

Prognosis remains dismal even with timely and optimal treatment; overall survival is typically 7 to 11 months in children and 8 to 28 months in adults (31; 57; 61). Overall survival with standard care is slightly longer compared to patients with IDH wildtype glioblastoma but shorter compared to patients with WHO grade 4 IDH-mutant astrocytomas (38). K27M mutation is a significant independent prognostic marker for poor outcome among pediatric glioblastomas (32). As compared to H3.1K27M mutation, H3.3K27M mutation is associated with shorter survival, higher metastatic relapses, and poorer response to radiotherapy (66; 10).

Patients with NF1-associated high-grade midline gliomas (H3 K27M-mutant and nonhistone mutant) have extremely poor prognosis (progression-free survival 3 months; overall survival 5 to 7 months), worse than high-grade gliomas developing in non-NF1 patients (progression-free survival 6 months, overall survival 10 to 11 months) (23).

H3 K27M mutation is often associated with a poorer prognosis in children, but not necessarily in adults (41; 17; 19).

Additional biomarkers influence outcomes in midline gliomas harboring histone mutations. For example, enhancer of zeste homolog 2 (EZH2) overexpression is a poor prognostic factor in glioma. High-grade midline gliomas are associated with H3 K27M mutation, high EZH2 expression, and low P16 expression (19). Low-grade midline gliomas have high H3 K27me3 expression and low membrane-associated transporter protein (MATP) protein expression (19). Poor median survival is associated with H3 K27M mutation, high EZH2 expression, and low H3 K27me3 expression.

Biological basis

Etiology and pathogenesis

Frequency of IDH mutation in the midline region has been reported to be low (63).

H3 K27M-mutant gliomas are IDH-wildtype, lack 1p/19q co-deletion, and are defined by the presence of K27M mutation in the H3F3A or HIST1H3B/C genes, which encode the histone H3 variants H3.3 and H3.1 in both children and adults (67; 61). These histone mutations are mutually exclusive from the other common mutations that define distinct infiltrating glioma types.

Histones are basic nuclear proteins responsible for the nucleosome structure within eukaryotic cells. Among the 5 classes of histones, some are expressed only during the S phase, whereas others are replacement histones that are replication-independent and expressed in quiescent or terminally differentiated cells.

Supratentorial high-grade gliomas have mutations in H3F3A (located on chromosome 1q) encoding histone H3.3. H3.3 is a replacement histone subclass that is encoded by 2 distinct genes, H3.3A (H3F3A) and H3.3B. This leads to the creation of oncohistones.

Oncohistones in high-grade gliomas show specific associated mutations and brain location. G34R/V is restricted to hemispheric tumors, and K27M occurs in the midline (31; 63; 66; 13). K27M mutations occur in 2 main histone variants H3.3 and H3.1 and result in lysine to methionine exchange at position 27 of H3F3A or HIST1H3B/C genes. G34 mutations of H3.3 result in glycine to arginine exchange at codon 34 of H3F3A (60; 66).

The H3 K27M-mutated diffuse intrinsic pontine gliomas group is further divided into two subgroups, H3.1 K27M and H3.3 K27M (10).

H3.3 K27M mutation. H3.3 K27M mutation occurs at median age 7.5 years. Patients with this mutation have poor prognosis and frequent leptomeningeal dissemination; they are likelier to have an oligodendroglial histologic phenotype, poor response to radiation, and median survival of 9 months (13). These tumors are seen all over the midline structures (67; 10). Approximately 20% of tumors previously categorized as pediatric glioblastomas harbor this mutation (25).

H3.1 K27M mutation. H3.1 K27M mutation occurs at median age 5.5 years. Patients with this mutation are likelier to have an astroglial differentiation, a good response to radiation, and a median survival of 15 months (13). These tumors are almost exclusive to the pons (67; 10).

H3 K27 wild type. Occurs at median age of 6 years with median survival of 9 months (13).

H3 K27M-mutant diffuse midline gliomas frequently share genetic alterations, including tumor protein 53 (TP53) overexpression, alpha-thalassemia/mental retardation X-linked (ATRX) loss, monosomy 10, activin A receptor type 1 (ACVR) mutation, poly ADP-ribose polymerase (PARP1) overexpression, cyclin D1-3(CCND1-3), and cyclin-dependent kinase (CDK4/CDK6) mutations (61; 16). TP53 mutations occur in about 40% of diffuse intrinsic pontine gliomas and represent the second most frequent mutation, correlated with worse overall survival (07).

H3 K27M mutation is mutually exclusive or rarely co-occurs with IDH1 mutation, telomerase reverse transcriptase (TERT) promoter mutation, epidermal growth factor receptor (EGFR) amplification, and BRAF-V600E mutation (61; 41; 16). Prototypic alterations of adult primary glioblastoma IDHwt (eg, EGFR amplification, CDKN2A/B homozygous deletions, PTEN mutations) are infrequent in childhood (51; 63; 52; 32). O6-methylguanine-DNA methyltransferase (MGMT) promoter methylation is found in 75% of G34-mutant subtype tumors and in 3% of K27 mutant tumors (32; 41).

Diffuse glioma, H3.3 G34-mutant is included as a distinct grade 4 entity (38).

Diagnostic workup

Diagnosis is made with imaging and histologic confirmation. Radiographic criteria alone are used for tumors with midline locations associated with serious procedural complications (28). Imaging involves use of brain MRI with and without gadolinium. MRI abnormalities include diffuse, T2 expansile signal changes in the midline structures.

In most cases, little or no enhancement is noted after the administration of gadolinium at initial diagnosis. Occasionally, areas of necrosis with surrounding serpiginous contrast enhancement are present. A retrospective study to characterize imaging features of 24 diffuse midline gliomas found distinguishing features in H3K27-mutant midline gliomas as compared to their H3wild-type counterparts (01). In H3K27-mutant gliomas, there was an absence of perilesional edema at low frequency of strong contrast enhancement. Contrast enhancement of intermediate or higher and rim enhancement was associated with microscopically detected microvascular proliferation. Necrosis on MRI is less frequently seen than in the H3wt high-grade midline gliomas (01).

PET imaging with amino acid tracers, such as 18-F-dihydroxy-phenylalanine (F-DOPA) is able to discriminate H3 K27M-mutant from wild-type diffuse midline gliomas and has demonstrated correlation of higher tracer uptake with worse prognosis (69; 53).

Traditional WHO grading for diffuse intrinsic pontine gliomas was not prognostic. Outcomes for grade 2, 3, and 4 tumors are equally poor (07). Next-generation sequencing helps accurately classify these tumors, improves understanding of unique tumor biology, and provides important insights into prognosis and management. Reports have shown that biopsy of these tumors is safe (28). Whether a biopsy at the time of diagnosis should be obtained in the absence of a clinical trial remains contentious. Modern neuronavigation tools allow accurate 3D localization of deep structures in the brain and identification of major vessels prior to surgery, which greatly minimizes surgical complications (28).

Features of glioblastoma like microvascular proliferation and necrosis may be seen. Immunophenotyping shows expression of neural cell adhesion molecule 1 (NCAM1), S100, and oligodendrocyte transcription factor 2 (OLIG2) on tumor cells (36). Microtubule-associated protein 2 (MAP2) expression is common. The immunoreactivity of glial fibrillary acidic protein (GFAP) is variable, and synaptophysin can be focal. Chromogranin A and neuronal nuclei (NeuN) are not typically expressed (36).

Once tissue is obtained, the macroscopic and microscopic evaluation can provide insight into the underlying disease. Macroscopically, these tumors cause enlargement and distortion of anatomical structures by diffuse infiltration. Microscopically, they typically have an astrocytic morphology and rarely an oligodendroglial pattern. Mitotic activity is present in most cases but is not necessary for diagnosis (36).

Management

Anatomic location of the tumor severely limits any opportunity for meaningful surgical resection. Treatment usually consists of standard fractionated radiation to a dose of 54-59Gy (over 30 fractions) and temozolomide (13).

Radiation is the mainstay of treatment. Photon-based radiotherapy to a range of 54 to 59.4 Gy given in 30 to 33 fractions of 1.8 Gy daily (6 weeks), as well as hypo-fractionated radiation (39 to 45 Grays given in 13 fractions of 3.0 Grays over 3 to 4 weeks), has been utilized with similar survival rates (44; 29; 68). Temozolomide has limited to no value in pediatric H3 K27M-mutant gliomas, likely due to a lack of MGMT promoter methylation (12; 58). In adult H3 K27M-altered gliomas, an aggressive multimodality approach may be beneficial. This requires further study.

Pre-radiation chemotherapy has shown improvement in overall survival and progression-free survival in a small series of patients (65; 26). Median survival benefit after re-irradiation (30 to 36 Grays) for recurrent disease ranges from 3 to 4 months (30; 08).

Future directions

Novel ongoing treatment trials include convection-enhanced therapies, immunotherapy (vaccine therapies, CAR-T cells), and targeted therapies.

Convection-enhanced delivery. Convection-enhanced delivery is a therapeutic strategy that allows for targeted treatment of a specific region via a cannula that can be placed in difficult-to-access areas and allows for direct intraparenchymal infusion of drugs at a steady rate over a prolonged period of time. Studies have shown safe placement of convection-enhanced delivery catheters into the brainstem in humans (03; 62).

Immunotherapy.

Peptide-based vaccine therapy. Glioma-associated antigens interleukin-13 receptor alpha 2 (IL-13Rα2), EphA2, and survivin are commonly overexpressed in pediatric gliomas (46). A pilot study of subcutaneous vaccination with glioma-associated antigen epitope peptides in HLA-A2-positive children with newly diagnosed brainstem and high-grade gliomas was well tolerated and has preliminary evidence of immunologic and clinical responses (55).

Dendritic cell vaccines reactivate tumor-specific T cells in both clinical and preclinical settings. Autologous dendritic cell vaccine administered intradermally is safe and able to generate a diffuse intrinsic pontine glioma-specific immune response detected in peripheral blood mononuclear cells and CSF (05).

The disialoganglioside GD2 is highly expressed in H3 K27M-mutant glioma cells. Preclinical studies have demonstrated GD2 as a novel immunotherapy target in H3 K27M mutant diffuse midline gliomas, and antitumor efficacy of GD2-CAR T-cells delivered systemically (43). The first-in-human phase I clinical trial included four patients with H3 K27M-mutant diffuse intrinsic pontine glioma or diffuse midline glioma treated with GD2-CAR T cells administered intravenously, with subsequent intraventricular infusions (39). GD2-CAR T therapy produced toxicities that were largely related to tumor location and reversible with intensive supportive care. Radiographic response was seen in three patients but was not durable. Proinflammatory cytokines were increased in plasma and CSF. These preliminary cases suggest that GD2-CAR T-cell therapy holds promise for H3K27M+ diffuse intrinsic pontine gliomas or diffuse midline gliomas.

A single-center trial of newly diagnosed pediatric patients with diffuse intrinsic pontine glioma who underwent oncolytic viral therapy (DNX-2401, an oncolytic adenovirus that selectively replicates in tumor cells) followed by radiation revealed changes in T-cell activity and stabilization to the reduction of tumor size (22). Adverse events related to virotherapy were largely grade 1 or grade 2, with only one event of grade 3 and no grade 4 or 5 events (22).

Targeted therapies. K27M mutation drives gliomagenesis by alteration of an important site of posttranslational modification in the histone H3 variants and leads to altered DNA methylation and gene expression profiles (63; 04). Ongoing efforts aim to study the efficacy of therapeutics targeting histone-modifying enzymes for midline gliomas with histone H3 mutations.

Panobinostat. Panobinostat is a general histone deacetylase inhibitor that has shown good in vitro efficacy against diffuse intrinsic pontine gliomas (27; 54). Panobinostat synergizes with histone demethylase inhibitor GSKJ4 in H3.3K27M mutant diffuse intrinsic pontine gliomas cells (27). Together, these data suggest a promising therapeutic strategy for diffuse intrinsic pontine gliomas.

ONC201. ONC201 is a small-molecule selective antagonist of dopamine receptor D2/3 (DRD2/3) that crosses the blood-brain barrier and exhibits P53 independent anticancer efficacy. DRD2 expression within the central nervous system is highest in midline structures of the brain. DRD2 is a G protein-coupled receptor that promotes tumor growth and has emerged as a therapeutic target for gliomas that overexpress this receptor (33; 09). H3 K27M-mutant diffuse midline gliomas have been reported to exhibit elevated expression and dependency on DRD2 as a downstream epigenetic consequence of the mutation (11). It has an exceptional safety profile and has been reported to have clinical and radiographic responses in both recurrent and frontline settings (11). ONC206 is another drug in the same class actively undergoing investigation.

Cdk4/6 inhibitors. CDK alterations are described in about 30% to 40% of diffuse midline gliomas. Cdk4/6 inhibitors (palbociclib, ribociclib, and abemaciclib), tyrosine kinase inhibitors, and other agents that more specifically target these mutations are being explored in patients with diffuse midline gliomas (15; 64; 49). Other potential therapeutic targets tested with in vivo studies include BET bromodomain inhibitors (54).

mTOR inhibitors. PI3K/AKT/mTOR pathway is dysregulated in more than 50% of diffuse midline gliomas. A single-center study from Italy reported significant benefit in overall survival with utilization of personalized treatment based on tumor molecular alterations. mTOR inhibition with everolimus was associated with the best overall survival (14).

EGFR inhibitors. EGFR is overexpressed in 80% to 85% of high-grade gliomas. Anti-EGFR therapies have shown limited benefit in a small subset of patients (06; 24; 56; 02; 18). A phase 3 trial of concomitant radiation and outpatient nimotuzumab (IgG1 antibody that targets EGFR) demonstrated survival outcomes similar to radiotherapy and intensive chemotherapy in a hospitalized setting (21). Another phase 2 trial of nimotuzumab with concomitant radiation and vinorelbine revealed a median overall survival of 15 months (40). The interpretation of these studies is somewhat difficult as the WHO classification system has evolved since their completion.

Multiple phase I and II trials are currently ongoing to test novel agents to treat midline gliomas with histone mutation in a frontline and recurrent setting (Table 2). An understanding of molecular pathology and ongoing efforts to improve diagnostic tools and therapy holds promise for future advances in prognosis.

Table 2. Ongoing Trials to Test Novel Agents for Diffuse Midline Glioma in a Frontline and Recurrent Setting