High frequency of PDGFRA and MUC family gene mutations in diffuse hemispheric glioma, H3 G34-mutant: a glimmer of hope?

Pediatric-type diffuse high-grade glioma (pHGG) is a highly aggressive brain tumor with a poor prognosis that accounts for approximately 8–15% of all central nervous system tumors in children and adolescents [1]. The widespread use of high-throughput sequencing, genetic profiling and epigenetic analysis has greatly increased our understanding of the cell origin, pathogenesis and biological characteristics of pHGG. Two types of heterozygous somatic mutations in H3F3A, which encodes histone H3, were firstly identified in pHGG in 2012 [2]. These mutations lead to an amino acid substitution at key residues (K27M or G34R/G34V). Subsequent studies found that K27M mutation presented in both H3F3A and HIST1H3B/C genes, but G34R/V-mutant gliomas were characterized by recurrent glycine-to-arginine/valine alterations at codon 34 (G34R/V) only in the H3F3A gene. The fifth edition of the WHO Classification of Tumors of Central Nervous System in 2021 defined the standard name of G34R/V-mutant gliomas as diffuse hemispheric glioma H3 G34-mutant, classified as WHO grade 4.

Most diffuse hemispheric glioma H3 G34-mutant tumors (G34-DHGs) carry TP53 mutations and ATRX deletions [2], with hypermethylation at the OLIG2 locus, leading to the absence of OLIG2 expression [3]. The current treatment for these tumors is surgery, radiotherapy and chemotherapy. However, the efficacy of this therapeutic strategy is not satisfactory. In-depth understanding of the molecular changes and immune infiltration in G34-DHGs will help to establish personalized tumor treatment strategies.

In this study, we conducted a comprehensive analysis of G34-DHGs at the clinicopathological and genetic level. Our research uncovered new molecular characteristics of this unique tumor, which provide potential opportunities to develop individualized therapy.

Diffuse hemispheric glioma H3 G34-mutant is a newly recognized tumor type in the 2021 WHO Classification of Tumors of the Central Nervous System. This tumor is a malignant infiltrative glioma that typically occurs in the cerebral hemispheres and harbors a missense mutation in the H3F3A gene that results in a G34R/V substitution in histone H3. The histological heterogeneity of G34-DHGs has been previously reported [15]. In general, most cases show high-grade anaplastic features, with microscopic features of anaplastic astrocytoma, GBM or embryonic tumors (primitive neuroectodermal tumor-like features) or even anaplastic pleomorphic xanthoastrocytoma [16]. Therefore, misdiagnosis of cases is likely if clinicians and pathologists rely solely on histological diagnosis. Consistent with these findings, most cases in our center also presented as GBM-like and with/without a focal embryonal appearance. However, calcification, perivascular growth pattern and perineuronal satellitosis, which rarely appear in GBM or called ‘secondary structures’, were observed [3, 17]. We further analyzed the immunohistochemical and characteristics of H3G34-mutant diffuse gliomas. Lack of OLIG2 expression is one of the characteristics of this tumor and was previously reported in 100% of G34 DHGs [18]. We also observed loss of OLIG2 expression in all of our cases. In addition, loss of ATRX expression and p53 overexpression are also frequently observed in this tumor type. The H3F3A G34R/V mutation is associated with a high frequency of MGMT methylation, but this is mutually exclusive with IDH or TERT promoter mutation.

Significantly, we found frequent PDGFRA mutation (12/15) in the G34-DHGs in our study. PDGFRA is a classical receptor tyrosine kinase with five extracellular immunoglobulin-like structures and responsive to platelet-derived growth factor (PDGF) [19], and this protein plays an important role in cell determination, cell proliferation and migration during neural development and adult neurogenesis [19, 20]. In glioma, somatic alterations of PDGFRA include missense mutations, in-frame insertions or deletions, gene amplification and fusion. A high proportion of gliomas, especially TERTp wild-type GBM, are associated with somatic alterations in PDGFRA. Recurrent PDGFRA mutations have also been found in low-grade neurotumors associated with refractory seizures, called septal dysplasia neuroepithelial tumors or mucinous glial neuronal tumors [21]. PDGFRA gene amplification is present in approximately one-third of HGGs, and most tumors with PDGFRA gene amplification have genomic deletions of exons 8 and 9[22]. Ozawa et al. reported an adult case with a gene fusion between PDGFRA and VEGFR2 [23]. Point mutations are located in the extracellular domain, transmembrane domain and kinase domain of the PDGFRA receptor. These oncogenic mutations lead to constitutively activated PDGFRα, thereby activating downstream signal transduction [24]. Aberrant PDGFRA signaling in gliomas leads to activation of the Ras-Raf-MEK-ERK pathway [25]. Gain-of-function PDGFRA variants trigger multiple tumor-promoting signaling cascades, including phospholipase Cγ (PLCγ), phosphatidylinositol-3-kinase (PI3K), mitogen-activated protein (MAP) kinase, and signal transduction and Activators of Transcription (STATs) [26, 27]. Targeting PDGFRA by tyrosine kinase inhibitors (TKIs) or antibodies showed promising antitumor effects in patients with various PDGFR-driven extracranial tumors and pre-clinical models of gliomas [28‐31]. In a glioma subset with histone H3 mutation (H3K27M DMGs), PDGFRA amplification and mutation were reported together with histone H3.3 mutation. Li et al. found that 18 cases of 112 midline glioma patients had PDGFRA mutations [32]. Both our study and Chen’s study found high frequencies of PDGFRA mutations in G34-DHGs [10]. PDGFRA mutation was present in 12 out of 15 cases (80%) in our study group, and Chen et al. extensively analyzed 95 cases of HGGs with H3.3G34R/V mutation and found that PDGFRA mutations presented in 44% of all tumors and 81% of recurrent tumors, they speculated that PDGFRA mutant G34 tumors have expanded astrocytic compartments that are conducive to maintaining the active chromatin conformation state of the GSX2 enhancer to maintain high PDGFRA expression and continue to promote the carcinogenic state [10]. In GBM and in IDH mutant lower-grade (WHO Grades II/III) glioma, PDGFRA amplification is associated with shorter progression-free survival and OS and it is an independent prognostic factor [33]. PDGFRA amplification was also an indicator of poor prognosis in H3K27M DMGs [34, 35]. However, the prognostic implication of PDGFRA mutation has not been extensively studied. Our data suggested that PDGFRA mutation may be an indicator for poor prognosis in G34-DHGs, but the results were not statistically significant; this may because of the relatively small number of G34 cases in our cohort.

Another important finding included the MUC family genes were also frequently mutated in G34-DHGs. The MUC gene family encodes mucins, a type of high molecular weight glycoprotein [36]. There are 21 MUC genes in the human genome that encode secreted and membrane-type mucins [37]. Increasing evidence has shown that MUC proteins play an important role in regulating tumor cell proliferation, growth, apoptosis and chemical tolerance [38, 39]. MUC gene mutation analysis showed that MUC16 (OMIM 606154) has the highest mutation frequency in G34-DHGs, followed by MUC17. A pan-cancer analysis involving somatic mutations of 10,195 samples and mRNA expression profiles of 9850 samples for 30 solid tumors found a greater abundance of immune cells in the microenvironment of MUC16-mutated tumors, and MUC16 mutation was associated with factors associated with response to immune checkpoint inhibitor therapy[14]. Tumors with MUC16 mutations exhibited a higher tumor mutational burden and more abundant neoantigen compared with that of MUC16 wild-type tumors, indicating increased tumor immunogenicity. MUC16-mutated tumors were characterized by upregulated expression of T-effector and interferon-γ gene signature, a hallmark of preexisting immunity associated with pronounced benefit from checkpoint blockade. An additional hallmark of MUC16-mutated tumors is the augmented expression of multiple inhibitory checkpoints including LAG3 and others, suggesting potential adaptive immune resistance to anti-PD-1/PD-L1 therapies and that additional inhibitory pathways beyond the PD-1/PD-L1 axis might be targets. Although we discovered a high frequency of MUC16 mutation in G34-DHGs, only one case in the MUC16 mutation group showed abnormal high immune infiltration with abundant tumor-infiltrating lymphocytes and this patient survived up to 75 months. We did not find a relationship between MUC16 and immune infiltration in G34-DHGs, which may be from the small sample numbers and immune-cold characteristics of G34-DHGs.

Regarding survival analysis, one study reported that the survival of patients with G34-DHGs was as poor as survival of patients with K27M DMGs [13]. However, all the cases were children in the study and the sample size was small, which may have resulted in bias. Our conclusion is consistent with most studies, showing that the OS of patients with H3 G34-mutant DHGs was worse than the survival of patients with IDH-mutant GBMs, but better than the survival of patients with H3 K27M-mutant DMGs; this may be partly because of the high frequency of MGMT promoter methylation in these tumors and favorable response to TMZ chemotherapy [8].

It is clear that histone mutations represent clearly defined entities within an umbrella HGG classification, and will require therapeutic development and clinical trials distinct. Compared with H3K27M DMGs, G34-DHGs were not sensitive to ONC-201, an effective therapeutic drug for H3K27M-DMGs [40]. Most G34-DHG patients had MGMT promoter methylation, and the rate was higher than the low incidence rate in H3K27M-DMGs (only one case had MGMT promoter methylation in our cohort, 1/34; data not shown), indicating that GTR and long-term maintenance TMZ treatment might benefit patients with G34-DHGs. Most cases also had PDGFRA mutations, and therefore the PDGFRA signaling pathway could be a potential therapeutic target. In addition, immunotherapy may be an option for cases with substantial immune infiltration and MUC16 mutation. These important findings will lay the foundation for the future personalized treatment of G34-DHGs, thereby helping to improve the current scenario in which existing treatment methods have not significantly improved the median survival of these patients.

Our study has some limitations. First, the overall sample size was small, mainly because G34 DHG is a brand-new type of pediatric-type diffuse high-grade gliomas in the latest 2021 WHO classification of tumors of the central nervous system, thus, total number of confirmed cases was small and some G34 DHG cases could be going undiagnosed which lack molecule detection. In addition, the study is mainly descriptive and the significance is possible only for survival analysis of different types of glioma but not yet for the definition of prognostic genetic markers, like PDGFRA and MUC16 and 17 mutations, studies on large sample are needed to confirm the prognostic results of this study. Last, prospective studies such as a follow-up clinical trial is needed to access the effects of the various PDGFRA targeted drugs combined with or without TMZ in G34 DHGs, like phase I study using crenolanib to target PDGFR kinase in children and young adults with newly diagnosed DIPG [41].

G34-DHG is a new high-grade glioma with high frequency of PDGFRA and MUC gene family mutations. PDGFRA may serve as a poor prognostic indicator and an effective therapeutic target for G34-DHGs. Moreover, MUC16 mutation tends to be a favorable prognostic factor and indicates high immune infiltration in certain patients, which may provide a new direction for immunotherapy of G34-DHGs.