The Epigenetic Evolution of Glioma Is Determined by the IDH1 Mutation Status and Treatment Regimen

Authors

Malta TM

Sabedot TS

Morosini NS

Datta I

Garofano L

Vallentgoed W

Varn FS

Aldape K

D’Angelo F

Bakas S

Barnholtz-Sloan JS

Gan HK

Hasanain M

Hau AC

Johnson KC

Cazacu S

de Carvalho AC

Khasraw M

Kocakavuk E

Kouwenhoven MCM

Migliozzi S

Niclou SP

Niers JM

Ormond DR

Paek SH

Reifenberger G

Sillevis Smitt PA

Smits M

Stead LF

van den Bent MJ

Van Meir EG

Walenkamp A

Weiss T

Weller M

Westerman BA

Ylstra B

Wesseling P

Lasorella A

French PJ

Poisson LM

The GLASS Consortium

Verhaak RGW

Iavarone A

Noushmehr H

Doi

10.1158/0008-5472.CAN-23-2093

PMID: 38117484 · DOI: 10.1158/0008-5472.CAN-23-2093 · Journal: Cancer Research (2024)

TL;DR

This study performed an epigenomic analysis of 132 matched initial and recurrent gliomas from the GLASS consortium to characterize longitudinal epigenetic evolution. IDH-wildtype gliomas showed a stable epigenome over time, while IDH-mutant gliomas exhibited progressive loss of DNA methylation at recurrence, particularly after treatment with temozolomide and/or radiotherapy. The study identified HOXD13 as a master regulator of IDH-mutant astrocytoma progression and demonstrated that treatment-associated epigenetic drift results in tumor microenvironment changes resembling treatment-naive IDH-wildtype glioblastoma.

Cohort & data

  • 132 patients with matched initial and first recurrent gliomas from the GLASS International consortium (difg_glass), totaling 354 DNA methylation samples profiled by Illumina 450K or EPIC BeadChip methylation arrays (450k-methylation-array).
  • Cancer types: IDH-mutant 1p/19q-codeleted oligodendroglioma (ODG; n=13), IDH-mutant astrocytoma (ASTR; n=59), IDH-wildtype glioblastoma (GB; n=60).
  • 54 patients with RNA sequencing data (rna-seq), 64 with DNA sequencing (whole-genome (whole-genome-seq) or whole-exome (whole-exome-seq) sequencing), 49 with all three modalities.
  • Independent validation cohort: GLASS-NL consortium with 100 IDH-mutant astrocytoma patients (36 treated, 64 untreated paired samples).
  • ChIP-seq for H3K27ac and H3K4me3 on 17 fresh-frozen GCIMP tumor samples (9 primary GCIMP-high, 4 primary GCIMP-low, 3 recurrent GCIMP-low).

Key findings

  • IDH-wildtype gliomas had a stable epigenome over time with low global methylation and zero CpG probes showing differential methylation > 15% between initial and recurrent tumors (N=60).
  • IDH-mutant gliomas showed genome-wide loss of DNA methylation at recurrence (674 CpG probes with > 15% difference; N=72), with the most prominent loss in patients progressing from GCIMP-high to GCIMP-low.
  • 71% (10/14) of IDH-mutant tumors that switched subtypes transitioned from GCIMP-high to GCIMP-low, associated with worse overall survival (log-rank P=0.06).
  • Treatment (temozolomide and/or radiotherapy) was associated with consistent loss of DNA methylation at recurrence compared to untreated patients: 620 differentially methylated probes (Kruskal-Wallis FDR < 0.01, absolute difference > 20%) distinguished treated from untreated groups.
  • 34% (10/29) of GCIMP-high tumors in the treatment group progressed to GCIMP-low vs. 4% (1/27) in the nontreatment group (Fisher test, P=0.005).
  • 68% (15/22) of IDH-mutant astrocytomas progressed to high-grade in the treatment group vs. 23% (6/26) in the nontreatment group (Fisher test, P=0.003) by Pan-CNS classification.
  • 60% (3/5) of IDH-mutant tumors that switched subtypes and had genomic data showed a hypermutator phenotype at recurrence (Fisher test, P=0.01).
  • HOXD13 was identified as a master regulator of IDH-mutant astrocytoma progression, showing increased H3K27ac and H3K4me3 peaks and overexpression in recurrent IDH-mutant noncodel gliomas (t-test, P=0.0019; N=24 paired samples).
  • CRISPR knockout of HOXD13 in a patient-derived IDH-wildtype cell line resulted in decreased cell proliferation in a time-dependent manner.
  • 83 of the epigenetically regulated genes in IDH-mutant noncodel progression were experimentally confirmed HOXD13 targets, including CENPF, PCNA, and HOXA7.
  • Treatment-associated DNA hypomethylation was enriched for the NEUROD1 motif, leading to upregulation of 18 unique genes including MYB and RSPO4.
  • Treated recurrent IDH-mutant tumors showed increased CD31+ endothelial cells and CD8+ T lymphocyte infiltration compared to untreated recurrences, validated by immunohistochemistry.
  • Previously treated IDH-mutant patients had worse survival from second surgery (log-rank P=0.0001) despite longer progression-free interval (40.5 vs. 27 months, log-rank P=0.009).

Genes & alterations

  • IDH1 / IDH2: Mutations define the GCIMP phenotype with high DNA methylation; IDH-mutant gliomas undergo progressive epigenetic demethylation at recurrence especially after treatment.
  • HOXD13: Epigenetically activated master regulator in IDH-mutant astrocytoma progression; gained H3K27ac and H3K4me3 marks at recurrence; CRISPR KO reduced proliferation.
  • CDKN2A: Homozygous deletion is a diagnostic molecular marker for CNS WHO grade 4 IDH-mutant astrocytoma; associated with radiotherapy-related acquisition.
  • MGMT: Promoter methylation status assessed by DNA methylation array; relevant to temozolomide response.
  • MYB: Upregulated after treatment-associated DNA hypomethylation; controls cell survival and proliferation.
  • RSPO4: Upregulated after treatment-associated DNA hypomethylation; Wnt/beta-catenin pathway agonist.
  • CENPF: HOXD13 target gene; associated with glioma prognosis and cell proliferation.
  • PCNA: HOXD13 target gene; prognostic indicator in gliomas and potential therapeutic target.
  • HOXA7: HOXD13 target gene; promotes tumor growth and metastasis.
  • MSH6: Somatic mutations in recurrent gliomas linked to hypermutation phenotype after alkylator chemotherapy.
  • NEUROD1: DNA sequence motif enriched at treatment-associated hypomethylated CpGs in IDH-mutant gliomas.

Clinical implications

  • Standard treatment (temozolomide and/or radiotherapy) in IDH-mutant gliomas delays progression (PFI 40.5 vs. 27 months, P=0.009) but is associated with epigenetic evolution toward an aggressive, IDH-wildtype-like phenotype at recurrence.
  • Survival from second surgery is markedly worse in previously treated IDH-mutant patients (log-rank P=0.0001), consistent across the GLASS-International and GLASS-NL validation cohorts.
  • Histologic progression to higher-grade astrocytoma (WHO grade 4) is significantly enriched in treated vs. untreated IDH-mutant gliomas, both by molecular and histopathological criteria.
  • GCIMP subtype transition (GCIMP-high to GCIMP-low) may serve as an epigenetic biomarker for monitoring disease progression and treatment response.
  • The identification of HOXD13 as a master regulator and its downstream targets (CENPF, PCNA) suggests potential therapeutic targets in recurrent IDH-mutant glioma.
  • Increased neoangiogenesis (CD31+) and T-cell infiltration (CD8+) in treated recurrent IDH-mutant tumors suggest the TME changes could inform immunotherapy strategies.

Limitations & open questions

  • Retrospective, nonrandomized cohort design: the observed worse survival after treatment may reflect inherent tumor aggressiveness rather than treatment-induced changes.
  • HOXD13 functional validation was limited to an IDH-wildtype cell line due to the lack of available IDH-mutant GCIMP-low or recurrent cell lines; the biology may differ in the IDH-mutant context.
  • Treatment groups were small (temozolomide only N=12, RT only N=18, RT+TMZ N=6, untreated N=33), limiting statistical power for distinguishing the individual effects of radiotherapy vs. chemotherapy.
  • Deconvolution of tumor microenvironment used in silico methods (methylCIBERSORT) from bulk data; spatial and single-cell resolution would strengthen conclusions about TME changes.
  • The study cannot establish causality between treatment and epigenetic changes; prospective studies are needed.
  • GLASS-NL validation cohort did not confirm the association between treatment and histologic progression to WHO grade 4 (Fisher test, P=0.3), suggesting possible cohort-specific effects.

Citations from this paper used in the wiki

  • “IDHwt gliomas showed a stable epigenome over time with relatively low levels of global methylation. The epigenome of IDHmut gliomas showed initial high levels of genome-wide DNA methylation that was progressively reduced to levels similar to those of IDHwt tumors.” (Abstract)
  • “HOXD13 [was identified] as a master regulator of IDHmut astrocytoma evolution.” (Abstract)
  • “34% (10/29) of GCIMP-high tumors progressed to GCIMP-low in the treatment group versus 4% (1/27) in the nontreatment group (Fisher test, P = 0.005).” (Results, Treatment section)
  • “patients receiving treatment after initial surgery had a worse survival than patients who did not receive treatment beyond surgery (log-rank P = 0.03)” and “we found a markedly worse survival in previously treated patients (log-rank P = 0.0001)” (Results, Clinical implications section)
  • “recurrent IDHmut tumors that were treated after surgery were marked by increased infiltration of endothelial cells (CD31+) and CD8 T lymphocytes” (Results, TME section)

This page was processed by crosslinker on 2026-05-04.