INTRODUCTION
Diffuse hemispheric glioma, H3 G34‐mutant (DHG‐H3 G34) is a novel tumor type in the 2021 World Health Organization (WHO) classification of central nervous system (CNS) tumors [ 1 ]. DHG‐H3 G34 is diagnostically defined as a cellular, infiltrative glioma with mitotic activity that arises within cerebral hemispheres and harbors an H3.3 p.G35R (G34R) or p.G35V (G34V) mutation; abnormal OLIG2 and ATRX loss of expression, and p53 diffuse overexpression or loss of expression are desirable diagnostic criteria, and for unresolved lesions, a methylation profile of DHG‐H3 G34 is required [ 1 ]. This tumor type primarily occurs in adolescent and young adult (AYA) patients [ 2 , 3 , 4 , 5 , 6 ] and is infrequent in older adults [ 7 ]. Available data suggest that DHG‐H3 G34 in adults has similar clinicopathological characteristics when compared to tumors in pediatric patients [ 6 , 8 , 9 ].
Molecularly, DHG‐H3 G34 has a characteristic mutational profile with frequent co‐occurring TP53 and ATRX mutations in addition to the diagnostic hallmark histone H3.3 p.G35 (G34) mutation [ 3 , 4 , 5 , 6 , 10 ]. PDGFRA mutations have also been shown to play a key role in tumorigenesis and have been reported in approximately 30%–50% of cases at initial presentation [ 3 , 6 , 10 ] and in more than 80% of recurrent tumors [ 10 ]. Hypermutation has been rarely reported at tumor recurrence [ 10 , 11 ]. Recurrent copy number changes include multiple chromosome arm losses (3q, 4q, 5q, 10q, 13q, and 18q) and gains (1q and 7q), oncogene amplification ( PDGFRA , CDK6 , and CCND2 ), and CDKN2A/B deletion [ 4 , 5 , 6 ]. By methylation profiling, DHG‐H3 G34 forms a distinct epigenetic group, and the MGMT promoter is frequently methylated [ 4 , 6 , 12 ].
Reported copy number data for this tumor type have been largely derived from genome‐wide methylation profiling data [ 4 , 6 ]. Oncoscan chromosomal microarray is a single nucleotide polymorphism (SNP) array that delivers superior resolution for copy number analysis to detect unbalanced copy number changes as well as loss of heterozygosity (LOH), including copy‐neutral LOH (cnLOH). It leverages on a highly multiplexed genome‐wide probe design that includes copy number and polymorphic allele probes, where polymorphic probes provide additional evidence to support gains/losses/cnLOH and determine levels of gains and losses including homozygous versus heterozygous status [ 13 ]. Similar to copy number imbalances, cnLOH is a relevant oncogenic mechanism of inactivation of tumor suppressors like TP53 that has often been unappreciated and is not detectable by methylation profiling [ 14 ]. To date, high‐resolution copy number analysis for DHG‐H3 G34 is lacking.
In this work, we describe copy number abnormalities detected by high‐resolution SNP‐based chromosomal microarray and additional molecular alterations observed on routine clinical testing of DHG‐H3 G34 to further our understanding about the spectrum of genetic changes associated with this new 2021 WHO CNS tumor type.
MATERIALS AND METHODS
This study was conducted according to Institutional Review Board‐approved protocols.
Case selection
Cases with a reported H3‐3A (previously H3F3A ) mutation (i.e., clinically relevant variant; Tier I or II AMP/ASCO/CAP) [ 15 ] involving amino acid glycine at codon 35 (G35, also known as G34) ( n = 60) were retrieved through a retrospective search of cases clinically tested using the Mayo Clinic targeted neuro‐oncology next‐generation sequencing (NGS) panel (2018–2022). Data for demographics, tumor location and presentation (primary/recurrent), histopathological features (primitive neuronal‐like component, Grade 4 morphological features [necrosis/microvascular proliferation]), and immunohistochemical (IHC) results (OLIG2, ATRX, and p53) were collected from pathology reports and re‐review of hematoxylin and eosin‐stained sections from formalin‐fixed‐paraffin‐embedded (FFPE) tissue tested by NGS and of available IHC stains performed at the Mayo Clinic.
Genomic testing
Sequencing data were obtained using the Mayo Clinic targeted neuro‐oncology NGS panel consisting of amplicon‐based DNA mutation and RNA fusion subpanels that collectively evaluate for mutations and/or rearrangements involving 187 unique CNS tumor‐related genes as previously described [ 16 ]. Additional mutations other than the H3.3 p.G35 (G34) mutations were also extracted for all cases. Follow‐up tumor mutational burden (TMB) and microsatellite stability testing was performed using TruSight Oncology 500 High‐Throughput NGS panel (Illumina, San Diego, CA) for cases with >10 reportable sequence variants (mutations and variants of unknown/uncertain significance) that were suspected to be hypermutant.
High‐resolution copy number profiling was performed using the OncoScan™ CNV Plus Assay (Thermo Fisher Scientific, Waltham, MA) at Mayo Clinic ( n = 26; 44%). Any partial or whole‐arm chromosomal losses and gains were recorded. Presence of cnLOH was identified when loss of heterozygous allele tracks occurred in the absence of copy number gains or losses. Chromosomal copy number complexity was calculated by determining the number of clinically relevant copy number abnormalities and the percentage of the genome altered. The percentage of genome altered was calculated as the sum of the sizes of clinically relevant copy number changes divided by the total genomic span covered by the OncoScan array probes (2,970,951,373 base pairs). The copy number profile of a group of IDH‐mutant astrocytomas CNS WHO Grade 4 ( n = 92) was used for comparison with the DHG‐H3 G34 group.
Methylation profiling was performed at the Mayo Clinic ( n = 8; 13%) using the Infinium MethylationEPIC v1.0/2.0 array (Illumina, San Diego, CA). Briefly, residual genomic DNA extracted from FFPE tumor tissue after clinical testing was bisulfite converted using the EZ DNA Methylation Kit (Zymo Research, Irvine, CA). Bisulfite converted DNA was processed through the Infinium HD FFPE Methylation Assay automated protocol per manufacturer's instructions. Data were processed using the NCI/Bethesda v.2 classifier. Classifier results were interpreted as “match” for a methylation class when superfamily and class calibrated scores were ≥0.9. MGMT promoter methylation status was evaluated using the MGMT‐STP27 prediction model [ 17 , 18 ]. Copy number profiles were inferred using the R “conumee” package ( http://bioconductor.org/packages/conumee/ ) as implemented in the classifier package. Using the genome‐wide copy number plot generated by the classifier report, gains and losses were visually estimated relative to the overall baseline.
MGMT promoter methylation status was assessed by the Mayo Clinic real‐time methylation‐specific polymerase chain reaction (PCR) ( n = 43; 72%), which interrogates eight CpG sites (CpG sites 79–86), following analytically validated protocols [ 19 ].
Immunohistochemical studies
IHC stains performed at the Mayo Clinic were obtained using validated protocols on the Ventana Benchmark Ultra platform (Ventana Medical Systems, Tucson, AZ, USA). Immunostains included OLIG2 ( n = 31; clone EPR2673; 1:100; Abcam, Cambridge, UK), ATRX ( n = 35; clone#D‐5: sc‐55,584; 1:1000; Santa Cruz, Dallas, TX, USA), and p53 ( n = 32; clone DO‐7; no dilution; Ventana Medical Systems, Tucson, AZ, USA). All immunostains have a nuclear staining pattern. IHC expression pattern was considered preserved for OLIG2 and ATRX when definitive nuclear staining similar to internal positive control was present in at least 25% tumor cells, and for p53, when definitive nuclear staining was variably present in less than 50% tumor cells. For OLIG2 and ATRX immunostain, abnormal expression pattern was considered valid only in the presence of a convincing internal positive control in endothelial and non‐neoplastic cells. IHC expression pattern was considered abnormal for p53 in cases with diffuse strong or complete lack of nuclear staining in at least 50% tumor cells.
Statistical analysis
Pearson's chi‐squared test, linear model analysis of variance (ANOVA), and chi‐squared test for given probabilities were used to compare clinicopathological and molecular features between pediatric (<18 years) and adult (≥18 years) patients. Fisher's exact test was used to assess the statistical significance of selected recurrent copy number categorical variables between DHG‐H3 G34 and astrocytoma, IDH‐mutant CNS WHO Grade 4.
RESULTS
Clinicopathological findings
As shown in Figure 1 and detailed in Table S1 , there were 60 unique patients ( n = 3, operated at Mayo Clinic; n = 32, neuropathology consultation; n = 25, cases received for genetic testing only). Median age at testing was 21 years (range, 12–50). Adult patients (≥18 years; 38/60, 63%) were more frequent than pediatric patients (<18 years; 22/60, 37%), and 72% (43/60) of patients were within the AYA range (15–39 years). Male: female ratio was 1.6:1. All tumors with available data (58/60) were described as involving the cerebral hemispheres, with a single multifocal tumor that also involved the midline; the two cases with unavailable tumor location had otherwise typical IHC and mutational profile for DHG‐H3 G34. Most tumors (58/60; 97%) were reportedly primary at clinical presentation. Tumor specimens were derived from resection (35/60; 58%) or biopsy procedure (25/60; 42%).
Histologically, all tumors were highly cellular infiltrating gliomas with evident mitotic activity (Figure 2A ). A primitive neuronal‐like component was seen in 48% (31/60) of cases, and Grade 4 morphological features (necrosis and/or microvascular proliferation) were observed in 65% (39/60) of tumors (Figure 2B ). Grade 4 morphological features were more frequent in resection than biopsy specimens (28/35 [80%] vs. 11/25 [44%]; p ‐value 0.004), suggesting that such features may be underrepresented in biopsy specimens because of sampling. By immunohistochemistry, cases with available results were universally negative for OLIG2 (42/42; 100%) (Figure 2C ) and had frequent abnormal expression patterns for p53 (46/48; 96%) (Figure 2D ) and loss of expression of ATRX (46/53; 87%) (Figure 2E ).
Molecular testing
Molecular findings are summarized in Figure 1 and detailed in Table S2 . H3‐3A mutations consisted of G34R (c.103G>A p.Gly35Arg, n = 48; c.103G>C p.Gly35Arg, n = 8; 94%), G34V (c.104G>T p.Gly35Val, n = 3; 5%), and a G34E (c.104G>A p.Gly35Glu, n = 1; 1%). Most tumors also had mutations in TP53 (55/60; 92%) and/or ATRX (50/60; 83%), which co‐occurred in 80% (48/60) of cases. TP53 and ATRX mutations were concordant with respective immunostains in 96% (43/45) and 95% (41/43) of cases. When combining IHC and mutation testing, the frequency of TP53 and ATRX ‐altered tumors increased to 97% (58/60) for TP53 and 92% (55/60) for ATRX . TERT promoter mutations were detected in 7% (4/60) of cases, and were mutually exclusive with ATRX mutations and had normal ATRX IHC expression. Overall, 98% (59/60) of cases had either ATRX inactivation or TERT promoter mutation hypothesized to function as a mechanism for telomere maintenance; the remaining case (case 57) had an ATRX variant of unknown significance with preserved IHC protein expression.
Among other recurrent mutations, PDGFRA mutations were the most frequent with 57% of cases (34/60), including one (of two) recurrent tumor. All PDGFRA mutations were previously observed missense mutations presumed or shown to result in constitutive activation of downstream signaling pathways [ 10 ]. These mutations clustered within the PDGFRA extracellular immunoglobulin‐like domains, and most mutations (30 [of 35], one case had two mutations interpreted as subclonal based on their variant allele frequencies) involved codons C235 ( n = 8), E229 ( n = 5), Y288 ( n = 5), P343 ( n = 4), C290 ( n = 3), G313 ( n = 3), and P250 ( n = 2). At lower frequency, there were mutations involving PTEN (7/60; 12%), BCOR (5/60; 8%), PIK3CA (5/60; 8%), NF1 (4/60; 7%), CDKN2A (3/60; 5%), and RB1 (2/60; 3%). Four cases (including one recurrent tumor) had an increased number of sequence variants suggestive of hypermutation/high TMB. Follow‐up TMB testing could be performed in two (of four) cases that were primary non‐recurrent tumors. High TMB was confirmed in one (of two) tested tumor (case 39); this tumor harbored a PMS2 mutation and had stable microsatellite status. The recurrent tumor (case 28), suspected to be hypermutant, had an MSH6 mutation suggesting the presence of mismatch repair deficiency but had insufficient material for any additional TMB or mismatch repair IHC testing.
A single pediatric case (case 7) showed an in‐frame FGFR3::FAM184B (exon 17::exon 2) fusion (Figure S1 ). This fusion has been previously reported in bladder cancer [ 20 ] and is predicted to lead to loss of the FGFR3 regulatory 3′ untranslated region similar to FGFR3::TACC3 fusions [ 21 ]. These findings suggest that the FGFR3::FAM184B fusion is likely clinically relevant, although it has not been functionally characterized. This tumor also harbored TP53 , ATRX , and PIK3CA mutations.
High‐resolution copy number profiling
Chromosomal microarray data were available mainly in adult patients (21/26; 80%), and all cases were tumors at first presentation. Genomes were unbalanced with frequent polyploidy (12/26; 46%) and multiple copy number abnormalities, including partial and whole chromosomal gains and losses and regions of cnLOH (Figure 3 ). Chromosomal copy number complexity, defined as at least 20 alterations involving at least 20% of the genome, was observed in 54% (14/26) of cases (Figure 1 ). The most common recurrent copy number abnormalities, present in at least 50% of cases, were losses or cnLOH involving 3q, 4q, 5q, 9p, 10q, 13q, 17p, and 18q (Figures 1 and 3 ). Copy number abnormalities of 17p encompassing TP53 were detected in 54% (14/26) of tested cases. Notably, 86% (12/14) of 17p copy number alterations were cnLOH; the remaining cases had 17p gain. Collectively, 17p cnLOH occurred in 46% (12/26) of the tested cases (Figures 3 and 4A,B ). While loss or cnLOH of 9p, including CDKN2A/B genes, was frequent (18/26; 69%), homozygous deletion of these genes was relatively uncommon (4/26; 15%). Oncogene amplifications were detected in 11% (4/26) of cases and involved PDGFRA (in a tumor lacking PDGFRA mutation), EGFR , KRAS , and PIK3CA in a single case each.
The pattern of partial chromosomal arm losses associated with 17p cnLOH encompassing TP53 , loss or cnLOH of 9p, including CDKN2A/B genes, and presence of copy number complexity observed in DHG‐H3 G34 is reminiscent of the copy number profile routinely observed in Grade 4 IDH‐mutant astrocytomas (Figure 3 ). Both tumor types had similar frequencies of losses of 5q, 9p (including CDKN2A/B ), 10q (including MGMT ), and 13q, and of copy number complexity (Table S2 ). Losses of 3q, 4q, and 18q were more frequent in DHG‐H3 G34 than in CNS WHO Grade 4 IDH‐mutant astrocytomas ( p ‐value <0.004 for all comparisons), whereas 17p cnLOH (including TP53 ) was more frequent in CNS WHO Grade 4 IDH‐mutant astrocytomas even though it was also common in DHG‐H3 G34 ( p ‐value 0.034) (Table S2 ).
TP53 abnormalities are a unifying feature
A TP53 abnormality predicted to negatively impact its function was detected at copy number (12/26, all cnLOH), sequence (55/60), and protein expression (46/48) level in all 60 cases (Figure 4C ).
Of the 12 cases with 17p cnLOH, 10 cases had a TP53 mutation; among the remaining two cases without a TP53 mutation, 1 case had a homozygous deletion involving TP53 (Figure S2 ). All 14 cases without 17p cnLOH had a TP53 mutation: 12 cases had two mutations and balanced 17p ( TP53 ) copy number state, and 2 cases had gain of 17p ( TP53 ) and tetramerization domain mutations predicted to be non‐functional. There were an additional 31 cases with a TP53 mutation that lacked copy number data, of which 8 had two TP53 mutations. Collectively, biallelic TP53 loss‐of‐function alterations, consistent with complete gene inactivation, could be demonstrated in 31 (of 60; 52%) cases with available copy number and/or sequence data.
Overall, abnormal p53 protein expression correlated with the presence of 17p cnLOH and/or at least one TP53 mutation (43/48; 90%). Two cases had normal p53 protein expression despite a TP53 loss‐of‐function genetic alteration. Case 33 had TP53 cnLOH and a nonsense mutation (R342*) resulting in premature protein truncation at the tetramerization domain; this nonsense mutation was located at the penultimate exon and escaped the predicted nonsense mediated decay resulting in expression of a functionally abnormal protein that was not resistant to proteolytic degradation and appeared normally stained by IHC. Case 47 had a TP53 missense mutation (C176Y) that presumably translated to a mutant protein that did not acquire resistance to protein degradation with a normal expression pattern. Conversely, three cases had abnormal p53 protein expression and lacked a TP53 mutation but copy number data were unavailable precluding exclusion of an underlying copy number abnormality.
Methylation profiling
Using the NCI/Bethesda v.2.0 classifier, all eight tested cases, including tumors with G34R ( n = 6), G34V ( n = 1), and G34E ( n = 1) mutation, matched with high confidence and calibrated scores >0.9 to the superfamily of pediatric‐type high‐grade gliomas (Table S1 ). For the methylation class, seven (of eight) cases matched with high confidence and calibrated scores >0.9 to diffuse hemispheric glioma, G34 (DHG_34), while the tumor with the H3.3 p.G35E (G34E) mutation also aligned with this methylation class but had a score of 0.831. In the Uniform Manifold Approximation and Projection (UMAP), all eight cases, including the H3.3 p.G35E (G34E)‐mutant tumor (case 47), clustered within the diffuse hemispheric glioma, H3 G34‐mutant (DHG_34) group (Figure 5 ).
The copy number profile derived from the generated genome‐wide plot also showed frequent losses involving 3q, 4q, 5q, 9p (including CDKN2A/B ), 10q, 13q, and 18q, with at least three (of eight cases) cases each (Figure 1 ). Among the five cases with CDKN2A/B loss, three were suggestive of a homozygous deletion. A single case (case 41) also had chromosomal microarray data. This case had a complex genome with a tetraploid state and multiple copy number abnormalities, including relative losses of 4q, 5, 10, 13q, and 18q as well as 17p cnLOH involving TP53 (Figure 6A ). Although the overall copy number profile was similar, copy number abnormalities such as the 4q, 5, and 10q relative losses, and 17p cnLOH (as expected based on the array design) were not evident in the methylation array‐derived copy number plot (Figure 6B ).
MGMT promoter methylation and copy number status
MGMT promoter methylation testing by real‐time PCR (RT‐PCR) was positive in 63% (27/43), indeterminate in 4% (2/43), and negative in 33% (14/43) of cases (Figure 1 ). By methylation profiling, MGMT promoter status was reported as methylated in 87% of cases (7/8) and was concordant with results from RT‐PCR in all six (of six; 100%) cases with available data (Figure 1 ). In 18 (of 19, 94.7%) cases with 10q loss, which was heterozygous in all cases, the loss included MGMT . For the 14 cases where MGMT promoter methylation and copy number status were available, 8 had copy number loss with promoter methylation and 6 had copy number loss without promoter methylation.
Pediatric versus adult tumors
No significant difference was noted in clinical, histopathological, and mutational profile between pediatric and adult patients in this series.
DISCUSSION
Prior series have reported the strong association of DHG‐H3 G34 with TP53 inactivation through mutations and/or abnormal IHC protein expression [ 3 , 4 , 5 ]. In this series of DHG‐H3 G34 that underwent routine NGS and high‐resolution copy number analysis, we detected a TP53 abnormality at copy number, sequence, and/or protein expression level in all tested cases, with two combined alterations indicating complete inactivation in half of these cases. This finding further supports TP53 inactivation as another universal characteristic of DHG‐H3 G34. Notably, we found that TP53 copy number abnormalities were primarily represented by 17p cnLOH. Copy number data associated with DHG‐H3 G34 have been primarily reported from methylation array data [ 4 , 6 ]. Methylation array does not include polymorphic allele probes precluding detection of 17p cnLOH by this testing platform. Therefore, the presence of 17p cnLOH including TP53 in DHG‐H3 G34 has been unrecognized. Our high‐resolution copy number analysis expands the spectrum of TP53 genetic abnormalities in this tumor type to include cnLOH, a mechanism whereby tumors often utilize to inactivate this tumor suppressor [ 14 , 22 ].
In addition to detecting the previously reported recurrent partial chromosomal losses of 3p, 4p, 5q, 10q, 18q as well as oncogene amplifications [ 4 , 6 ], we were also able to definitively distinguish heterozygous from homozygous CDKN2A/B losses based on supporting polymorphic allele probe patterns, which were more often heterozygous in our series. By methylation array, a definitive distinction between heterozygous and homozygous deletions is often not possible because of the absence of polymorphic allele probes. In the single case with methylation and SNP‐array copy number data, we observed copy number changes that were not clearly seen in the methylation‐based data. Methylation array‐based copy number data use the total methylation signal (i.e., unmethylated + methylated) to calculate copy number changes relative to reference samples whereas SNP‐array also leverages on polymorphic allele probes to assist in estimating tumor ploidy. Average concordance using Infinium EPIC methylation array and Conumee package when compared to SNP‐based array copy number data has been reported to be approximately 60% [ 23 ]. Inferring copy number data from methylation array is a cost‐effective strategy to obtain genome‐wide copy number footprint [ 24 ], but it is important to note that the lack of polymorphic allele probes precludes (a) accurate baseline ploidy determination when genomes are not diploid leading to normalization bias, (b) distinction between homozygous versus heterozygous deletion, and (c) detection of cnLOH.
High‐resolution copy number analysis also showed that DHG‐H3 G34 often has complex genomes, in keeping with the previously reported higher genomic instability in DHG‐H3 G34 relative to the other pediatric‐type diffuse high‐grade gliomas [ 25 ]. The increased genome complexity may be associated with inactivation of TP53 and ATRX [ 22 , 26 , 27 , 28 ], and with mismatch repair dysfunction because of impaired interaction of H3K36 with MutSα (MSH2–MSH6 heterodimer) caused by the mutant H3.3 p.G35 (G34) protein [ 29 , 30 , 31 ]. It is noteworthy that DHG‐H3 G34 has a mutational/copy number footprint reminiscent of that of IDH‐mutant astrocytoma, CNS WHO Grade 4. Both tumor types typically show TP53 and ATRX (less commonly TERT promoter) mutations, and 17p cnLOH (encompassing TP53 ). They also often have complex genomes with CDKN2A/B losses. Epigenetically, however, DHG‐G34 and IDH‐mutant astrocytoma, CNS WHO Grade 4 form distinct methylation classes, with the former showing global hypomethylation and the latter showing global hypermethylation characteristic of the glioma CpG island methylator phenotype [ 2 , 5 , 12 ].
Among additional molecular alterations was a non‐canonical histone H3.3 p.G35E (G34E) mutation detected in a tumor that fulfills the clinicopathological criteria and shows a mutational and epigenetic profile in keeping with a DHG‐H3 G34. In silico coding and amino acid modeling predict that this mutation has a significant impact on most models (Alamut Visual Plus v.1.11, SOPHiA GENETICS; created by on January 2, 2025). The H3.3 G34E mutation has also been reported in two diffuse gliomas without additional sample‐level information [ 32 ] and a secondary malignant tenosynovial giant cell tumor (GCT) [ 33 ], which is another tumor type where histone H3.3 G35 (G34) mutations are driver events [ 34 ], supporting its oncogenic role. H3.3 p.G35L (G34L) and p.G35M (G34M) mutations [ 6 , 32 ], and an H3‐3B , instead of H3‐3A , H3.3 p.G35R (G34R) mutation, have also been described in tumors with clinicopathological, genetic, and/or epigenetic features compatible with DHG‐H3 G34 [ 35 ]. These findings suggest that a broader mutational spectrum of H3.3 p.G35 (G34) is associated with DHG‐H3 G34 than originally defined [ 1 ]. We also detected a single DHG‐H3 G34 with an FGFR3 fusion, further expanding the spectrum of molecular alterations and potential therapeutic targets associated with this tumor type.
This series of DHG‐H3 G34 had multiple findings in keeping with previous studies. ATRX and TP53 mutations highly correlated with abnormal IHC patterns, and OLIG2 IHC expression was universally absent, further supporting the utility of this IHC panel in the diagnostic workup of DHG‐H3 G34. ATRX inactivation or TERT promoter activation hypothesized to function as a mechanism for telomere maintenance could be demonstrated in almost all cases. PDGFRA mutations were recurrent, underscoring the role of PDGFRA activation in DHG‐H3 G34 [ 5 , 10 , 36 ]. No significant differences at the histopathological or genomic level were noted between DHG H3‐G34 in adult and pediatric patients [ 6 ]. Hypermutation was observed in our series of DHG‐H3 G34, with at least one confirmed hypermutant tumor. Hypermutation has been rarely described at presentation as well as recurrence in DHG‐H3 G34 [ 11 ], and may be a therapeutic opportunity for some patients. MGMT promoter methylation by methylation‐specific PCR was frequent with a slightly lower rate relative to the reported rates from methylation array [ 4 , 6 , 37 ] and pyrosequencing [ 8 ], but in line with data from other methylation‐specific PCR assays [ 36 ]. This difference may be due to the fact that the CpG sites assessed by different testing platforms only partially overlap. Heterozygous loss of MGMT as part of 10q loss was also common [ 4 , 6 ], and associated with MGMT promoter methylation in a subset of cases [ 6 ], but we did not observe any case with MGMT homozygous deletion.
Our study is limited by the lack of follow‐up and survival data because of the nature of our dataset derived from routine clinical genetic testing in a referral laboratory. There is, however, a growing body of literature on the clinical course and outcomes for DHG‐H3 G34 patients [ 6 , 38 ].
In summary, integrated high‐resolution copy number and histomolecular analysis of DHG‐H3 G34 revealed recurrent cnLOH involving TP53 and universal TP53 abnormalities as another hallmark of DHG‐H3 G34, expanding the spectrum of genetic changes associated with this new 2021 WHO CNS tumor type.
AUTHOR CONTRIBUTIONS
Jorge A. Trejo‐Lopez : funding acquisition, data collection/analysis, writing—original draft, writing—review and editing. Cinthya Zepeda Mendoza : data collection/analysis, writing—review and editing. Thomas M. Kollmeyer : data collection/analysis/summary, review. Kristen L. Drucker : data collection/analysis/summary, review. Vanessa M. Pazdernik : data analysis/summary, review. Paul A. Decker : data analysis/summary, review. Corinne E. Praska : data collection/analysis/summary, review. Jayson Hardcastle : data collection/analysis, technical assistance, review. Stephanie Smoley : data collection/analysis, technical assistance, review. Matthew Isaacson : data collection/analysis, technical assistance, review. Mallika Gandham : data analysis, technical assistance, review. Surendra Dasari : data analysis, technical assistance, review. Aditya Raghunathan : data collection, review. Caterina Giannini : data collection, review. Gang Zheng : data collection, review. Aaron Wagner : data collection, review. M. Beatriz Lopes : data collection, review. Kar‐Ming A. Fung : data collection, review. JoElle G. Peterson : data collection, review. Andreana L. Rivera : data collection, review. Christopher Rossi : data collection, review. Meggen A. Walsh : data collection, review. Zied Abdullaev : data analysis, review; Kenneth Aldape : data analysis, review. Martha Quezado : data analysis, review. Drew Pratt : data analysis, review. Patrick Joseph Cimino : data analysis, review. Robert B. Jenkins : data collection/analysis, review. Cristiane M. Ida : conceptualization, methodology, data analysis/summary, writing—original draft, writing—review and editing, supervision.
CONFLICT OF INTEREST STATEMENT
None of the authors have a conflict of interest to disclose.
Supporting information
Table S1. Detailed clinical, immunohistochemical and genomic findings of DHG‐ H3 G34.
Table S2. Copy number abnormalities for DHG‐H3 G34 versus astrocytoma, IDH‐mutant, CNS WHO Grade 4.
Figure S1. Case 7‐ DHG‐H3 G34 with FGFR3::FAM184B fusion characterized by cDNA sequencing reads spanning fusion junction between FGFR3 exon 17 and FAM184B exon 2.
Figure S2. (A–C) Case 40‐ DHG‐H3 G34 with 17p cnLOH and homozygous deletion involving portion of TP53 indicated by arrows in (A) the chromosomal microarray genome‐wide view and detailed at a closer view in the ChAS (chromosome analysis suite software), with (B) the 17p cnLOH characterized by a balanced copy number state at level 0 in the smooth signal and weighted Log2 ratio tracks with absence of heterozygous probes in the allele difference track and the homozygous deletion characterized by the discrete dip in the smooth signal and probe intensities in the weighted Log2 ratio tracks associated with return to heterozygosity in the allele track, (C) disrupting part of the TP53 gene (arrowhead).
ACKNOWLEDGMENTS
Portions of this manuscript were presented at the Society for Neuro‐Oncology 2022 Annual Meeting, Tampa, Florida, November 16–20, 2022, and published in abstract form: Neuro‐Oncology 2022 Vol. 24 Issue Supplement_7 Pages vii152–vii152.
Trejo‐Lopez JA, Mendoza CZ, Kollmeyer TM, Drucker KL, Pazdernik VM, Decker PA, et al. Integrated high‐resolution copy number and histomolecular analysis of diffuse hemispheric glioma, H3 G34‐mutant reveals universal TP53 abnormalities. Brain Pathology. 2026;36(4):e70076. 10.1111/bpa.70076
Jorge A. Trejo‐Lopez and Cinthya Zepeda Mendoza are co‐authors.
DATA AVAILABILITY STATEMENT
The data that support the findings of this study are available from the corresponding author upon reasonable request.
REFERENCES
1. WHO Classification of Tumours Editorial Board . Central nervous system tumours. WHO classification of tumours series. Vol 6. 5th ed. Lyon: International Agency for Research on Cancer; 2021.
2. 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;22:425–437.
3. 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;482:226–231.
4. Korshunov A, Capper D, Reuss D, Schrimpf D, Ryzhova M, Hovestadt V, et al. Histologically distinct neuroepithelial tumors with histone 3 G34 mutation are molecularly similar and comprise a single nosologic entity. Acta Neuropathol. 2016;131:137–146.
5. 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;32:520–537.e5.
6. Le Rhun E, Bink A, Felsberg J, Gramatzki D, Brandner S, Benhamida JK, et al. The clinical and molecular landscape of diffuse hemispheric glioma, H3 G34‐mutant. Neuro Oncol. 2025;27:1519–1535.
7. Trejo‐Lopez JA, Praska CE, Zepeda Mendoza C, Kollmeyer TM, Raghunathan A, Giannini C, et al. H3 G34 mutation assessment for diffuse gliomas in adults: when would testing be most diagnostically useful? J Neuropathol Exp Neurol. 2022;82:93–95.
8. Picart T, Barritault M, Poncet D, Berner LP, Izquierdo C, Tabouret E, et al. Characteristics of diffuse hemispheric gliomas, H3 G34‐mutant in adults. Neurooncol Adv. 2021;3:vdab061.
9. Wang L, Shao L, Li H, Yao K, Duan Z, Zhi C, et al. Histone H3.3 G34‐mutant diffuse gliomas in adults. Am J Surg Pathol. 2022;46:249–257.
10. 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;183:1617–1633.e22.
11. Wood MD, Beadling C, Neff T, Moore S, Harrington CA, Baird L, et al. Molecular profiling of pre‐ and post‐treatment pediatric high‐grade astrocytomas reveals acquired increased tumor mutation burden in a subset of recurrences. Acta Neuropathol Commun. 2023;11:143.
12. 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;555:469–474.
13. Wang Y, Cottman M, Schiffman JD. Molecular inversion probes: a novel microarray technology and its application in cancer research. Cancer Genet. 2012;205:341–355.
14. Ciani Y, Fedrizzi T, Prandi D, Lorenzin F, Locallo A, Gasperini P, et al. Allele‐specific genomic data elucidate the role of somatic gain and copy‐number neutral loss of heterozygosity in cancer. Cell Syst. 2022;13:183–193.e7.
15. Li MM, Datto M, Duncavage EJ, Kulkarni S, Lindeman NI, Roy S, et al. Standards and guidelines for the interpretation and reporting of sequence variants in cancer: a joint consensus recommendation of the Association for Molecular Pathology, American Society of Clinical Oncology, and College of American Pathologists. J Mol Diagn. 2017;19:4–23.
16. Blessing MM, Blackburn PR, Krishnan C, Harrod VL, Barr Fritcher EG, Zysk CD, et al. Desmoplastic infantile ganglioglioma: a MAPK pathway‐driven and microglia/macrophage‐rich neuroepithelial tumor. J Neuropathol Exp Neurol. 2019;78:1011–1021.
17. Bady P, Sciuscio D, Diserens AC, Bloch J, van den Bent M, Marosi C, et al. MGMT methylation analysis of glioblastoma on the Infinium methylation BeadChip identifies two distinct CpG regions associated with gene silencing and outcome, yielding a prediction model for comparisons across datasets, tumor grades, and CIMP‐status. Acta Neuropathol. 2012;124:547–560.
18. Bady P, Delorenzi M, Hegi ME. Sensitivity analysis of the MGMT‐STP27 model and impact of genetic and epigenetic context to predict the MGMT methylation status in gliomas and other tumors. J Mol Diagn. 2016;18:350–361.
19. Ida CM, Butz ML, Jenkins RB, Sarkaria JN, Kitange GJ, Giannini C, et al. Real‐time methylation‐specific polymerase chain reaction for MGMT promoter methylation clinical testing in glioblastoma: an alternative detection method for a heterogeneous process. Am J Clin Pathol. 2017;148:296–307.
20. Nassar AH, Lundgren K, Pomerantz M, Allen EV, Harshman L, Choudhury AD, et al. Enrichment of FGFR3‐TACC3 fusions in patients with bladder cancer who are young, Asian, or have never smoked. JCO Precis Oncol. 2018;2:PO.18.00013.
21. Parker BC, Annala MJ, Cogdell DE, Granberg KJ, Sun Y, Ji P, et al. The tumorigenic FGFR3‐TACC3 gene fusion escapes miR‐99a regulation in glioblastoma. J Clin Invest. 2013;123:855–865.
22. Donehower LA, Soussi T, Korkut A, Liu Y, Schultz A, Cardenas M, et al. Integrated analysis of TP53 gene and pathway alterations in the cancer genome atlas. Cell Rep. 2019;28:1370–1384.e5.
23. Kilaru V, Knight AK, Katrinli S, Cobb D, Lori A, Gillespie CF, et al. Critical evaluation of copy number variant calling methods using DNA methylation. Genet Epidemiol. 2020;44:148–158.
24. Feber A, Guilhamon P, Lechner M, Fenton T, Wilson GA, Thirlwell C, et al. Using high‐density DNA methylation arrays to profile copy number alterations. Genome Biol. 2014;15:R30.
25. Fontebasso AM, Papillon‐Cavanagh S, Schwartzentruber J, Nikbakht H, Gerges N, Fiset PO, et al. Recurrent somatic mutations in ACVR1 in pediatric midline high‐grade astrocytoma. Nat Genet. 2014;46:462–466.
26. Taylor AM, Shih J, Ha G, Gao GF, Zhang X, Berger AC, et al. Genomic and functional approaches to understanding cancer aneuploidy. Cancer Cell. 2018;33:676–689.e3.
27. Wang Y, Yang J, Wild AT, Wu WH, Shah R, Danussi C, et al. G‐quadruplex DNA drives genomic instability and represents a targetable molecular abnormality in ATRX‐deficient malignant glioma. Nat Commun. 2019;10:943.
28. Andersson N, Saba KH, Magnusson L, Nilsson J, Karlsson J, Nord KH, et al. Inactivation of RB1, CDKN2A, and TP53 have distinct effects on genomic stability at side‐by‐side comparison in karyotypically normal cells. Genes Chromosomes Cancer. 2023;62:93–100.
29. Yadav RK, Jablonowski CM, Fernandez AG, Lowe BR, Henry RA, Finkelstein D, et al. Histone H3G34R mutation causes replication stress, homologous recombination defects and genomic instability in S. pombe . Elife. 2017;6:e27406.
30. Lowe BR, Yadav RK, Henry RA, Schreiner P, Matsuda A, Fernandez AG, et al. Surprising phenotypic diversity of cancer‐associated mutations of Gly 34 in the histone H3 tail. Elife. 2021;10:10.
31. Haase S, Banerjee K, Mujeeb AA, Hartlage CS, Núñez FM, Núñez 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;132:132.
32. Williams EA, Brastianos PK, Wakimoto H, Zolal A, Filbin MG, Cahill DP, et al. A comprehensive genomic study of 390 H3F3A‐mutant pediatric and adult diffuse high‐grade gliomas, CNS WHO grade 4. Acta Neuropathol. 2023;146:515–525.
33. Righi A, Mancini I, Gambarotti M, Picci P, Gamberi G, Marraccini C, et al. Histone 3.3 mutations in giant cell tumor and giant cell‐rich sarcomas of bone. Hum Pathol. 2017;68:128–135.
34. Behjati S, Tarpey PS, Presneau N, Scheipl S, Pillay N, van Loo P, et al. Distinct H3F3A and H3F3B driver mutations define chondroblastoma and giant cell tumor of bone. Nat Genet. 2013;45:1479–1482.
35. Nazem A, Lavezo J, Abdullaev Z, Aldape K, Quezado M, Cimino PJ, et al. Diffuse hemispheric glioma with H3‐3B G34R mutation: expanding the spectrum of histone H3 genes in diffuse hemispheric glioma, H3 G34‐mutant. J Neuropathol Exp Neurol. 2023;83:58–60.
36. 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;33:829–842.e5.
37. Alnahhas I, LaHaye S, Giglio P, Mardis E, Puduvalli V. An evaluation of MGMT promoter methylation within the methylation subclasses of glioblastoma. Neurooncol Adv. 2020;2:vdaa117.
38. Williams A, Neth B, Webb M, Schwartz J, Daniels D, Caron S, et al. Diffuse hemispheric glioma, H3 G34‐mutant: defining the clinical picture through a single institution case series and literature analysis. J Neurooncol. 2025;173:635–643.
Associated Data
Supplementary Materials
Table S1. Detailed clinical, immunohistochemical and genomic findings of DHG‐ H3 G34.
Table S2. Copy number abnormalities for DHG‐H3 G34 versus astrocytoma, IDH‐mutant, CNS WHO Grade 4.
Figure S1. Case 7‐ DHG‐H3 G34 with FGFR3::FAM184B fusion characterized by cDNA sequencing reads spanning fusion junction between FGFR3 exon 17 and FAM184B exon 2.
Figure S2. (A–C) Case 40‐ DHG‐H3 G34 with 17p cnLOH and homozygous deletion involving portion of TP53 indicated by arrows in (A) the chromosomal microarray genome‐wide view and detailed at a closer view in the ChAS (chromosome analysis suite software), with (B) the 17p cnLOH characterized by a balanced copy number state at level 0 in the smooth signal and weighted Log2 ratio tracks with absence of heterozygous probes in the allele difference track and the homozygous deletion characterized by the discrete dip in the smooth signal and probe intensities in the weighted Log2 ratio tracks associated with return to heterozygosity in the allele track, (C) disrupting part of the TP53 gene (arrowhead).
Data Availability Statement
The data that support the findings of this study are available from the corresponding author upon reasonable request.