Novel mutations target distinct subgroups of medulloblastoma

PMID: 22722829 · DOI: 10.1038/nature11213 · Journal: Nature (2012)

TL;DR

Whole-genome sequencing of 37 medulloblastomas (plus validation in 56 additional tumours; total n=93) identified 41 novel recurrently mutated genes that segregate by molecular subgroup. Subgroup-3 and 4 tumours harbour inactivating mutations in histone modification regulators (KDM6A, CHD7, ZMYM3) and gain/overexpress EZH2, maintaining a stem-like H3K27me3 epigenetic state. WNT-subgroup tumours carry mutations in CTNNB1-associated chromatin remodellers (SMARCA4, CREBBP) and the RNA helicase DDX3X. Functional modelling in mouse lower rhombic lip progenitors (LRLPs) demonstrated that CDH1 loss phenocopies mutant CTNNB1, DDX3X mutations expand the LRLP pool, and PIK3CA mutations accelerate but do not initiate WNT-subgroup medulloblastoma.

Cohort & data

• Discovery cohort: 37 medulloblastomas with matched normal blood, whole-genome sequenced (WNT n=5, SHH n=5, subgroup-3 n=6, subgroup-4 n=19, unclassified n=2).
• Validation cohort: 56 medulloblastomas subjected to targeted sequencing of 136 candidate genes (WNT n=6, SHH n=8, subgroup-3 n=11, subgroup-4 n=19, unclassified n=12).
• Total analysed: 93 tumours across all four subgroups.
• Cancer type: MBL (medulloblastoma).
• Dataset: mbl_pcgp (St. Jude/Washington University Pediatric Cancer Genome Project).
• Platforms: Whole-genome sequencing, Affymetrix U133v2 expression arrays, Affymetrix SNP 6.0, custom capture arrays, Sanger sequencing.
• Data deposition: dbGaP phs000409; SRA SRP008292.

Key findings

• WGS detected 22,887 validated somatic sequence mutations (SNVs/indels), 536 SVs, and 5,802 CNVs across the discovery cohort.
• 49 genes were recurrently mutated across all 93 tumours; 41 (84%) were novel for medulloblastoma.
• Mutation burden and amplification patterns varied significantly among subgroups (P values reported per gene in Figure 1).
• Fusion proteins were rare (<2% of SVs encode potential in-frame fusions); inherited cancer-predisposition mutations were uncommon (1/37 confirmed biallelic hit).
• Subgroup-3/4 epigenetics: KDM6A inactivating mutations in 8/93 tumours (predominantly subgroup-4 males); co-deletion of UTY on chrY in 57% of KDM6A-mutant males vs. 6% of KDM6A-WT males (P<0.005). CHD7 nonsense/frameshift in 4 subgroup-3/4 tumours. ZMYM3 mutations in 3 males with co-occurring KDM6A or KDM1A mutations. EZH2 gained/overexpressed via chr7q gain (P<0.005).
• WNT subgroup: CTNNB1 stabilising mutations in 8/11 WNT tumours (P<0.0001). SMARCA4 helicase-domain missense mutations in 4 WNT tumours (P<0.002). CREBBP nonsense in 2 WNT tumours (P<0.02). DDX3X DEAD-box domain mutations enriched in WNT (P<0.0001). CDH1 missense mutations in 2 WNT tumours lacking CTNNB1 mutations (P<0.05).
• PIK3CA: Activating mutations in 3 tumours (one each WNT, SHH, subgroup-4). Mouse modelling showed PIK3CA E545K accelerated but did not initiate WNT-medulloblastoma (100% penetrance by 3 months when combined with mutant Ctnnb1 vs. 4% by 11 months without).
• SHH subgroup: Biallelic PTCH1/SUFU inactivation in 4/13; GABRG1 T48M mutation in 2 SHH tumours.

Genes & alterations

Gene	Alteration(s)	Subgroup	Finding
KDM6A	Inactivating (splice site, nonsense, frameshift)	Subgroup-3/4	Two-hit tumour suppressor model with UTY loss on chrY
EZH2	Copy-number gain (7q), overexpression	Subgroup-3/4	Maintains H3K27me3; anti-correlated with KDM6A/CHD7/ZMYM3 mutations
DDX3X	Missense (DEAD-box domain)	WNT	Alters RNA binding; expands LRLP lineage rather than blocking migration
PIK3CA	Activating missense (Q546K, H1047R, N345K)	WNT/SHH/subgroup-4	Accelerates but does not initiate tumourigenesis via AKT pathway
CTNNB1	Stabilising missense (exon 3)	WNT	Present in >70% of WNT tumours; canonical driver
SMARCA4	Missense (helicase domain)	WNT	Chromatin remodeller recruited to WNT-responsive genes by CTNNB1
CREBBP	Nonsense	WNT	Histone acetyltransferase; co-mutated with SMARCA4 in one tumour
CHD7	Nonsense, frameshift	Subgroup-3/4	H3K4me3 chromatin remodeller
ZMYM3	Frameshift, nonsense, missense	Subgroup-4 (males)	Co-mutated with KDM6A/KDM1A; H3K4me3 regulation
CDH1	Missense (R63G, V329F)	WNT	Disrupts CTNNB1 sequestration; phenocopies mutant CTNNB1 in LRLPs
MYC	Amplification	Subgroup-3	Mutually exclusive with epigenetic mutations
MYCN	Amplification (incl. chromothripsis)	Subgroup-3/4	Co-occurs with OTX2 amplification
OTX2	Amplification	Subgroup-3/4	Cooperative with DDX31/TSC1 deletion at 9q34
TP53	Mutations	SHH	Previously reported; co-occurs with chromothripsis
PTCH1	Biallelic inactivation, focal deletion	SHH	Expected SHH-pathway driver
SUFU	Biallelic inactivation	SHH	Expected SHH-pathway driver
PTEN	Focal deletion, mutation	SHH	Previously reported
APC	Germline nonsense + somatic deletion	WNT	Turcot syndrome case
KMT2D	Mutations	SHH/other	Previously reported (as MLL2)
TRRAP	Missense	WNT	CTNNB1-recruited histone acetyltransferase complex
MED13	Missense	WNT	Mediator complex; coordinates RNA Pol II at WNT targets

Clinical implications

• EZH2 inhibitors may be therapeutic for subgroup-3 and 4 medulloblastoma, which rely on aberrant H3K27me3 maintenance and represent the most aggressive disease with limited treatment options.
• PIK3CA/AKT pathway inhibitors could be considered for WNT-subgroup tumours harbouring PIK3CA activating mutations; however, these mutations progress rather than initiate tumours.
• Subgroup-stratified clinical trials are essential: the study demonstrates that mutations are subgroup-specific, meaning drug trials must recruit the appropriate molecular subtype.
• DDX3X mutations represent a potential novel therapeutic target in WNT-subgroup disease.

Limitations & open questions

• The discovery cohort (n=37) is small; some subgroups (WNT, SHH) had only 5 cases each in the discovery set.
• Functional validation was performed only for WNT-subgroup genes (DDX3X, CDH1, PIK3CA) in mouse LRLPs; subgroup-3/4 epigenetic mutations lack equivalent in vivo modelling.
• The cell of origin for subgroup-3 medulloblastoma remains undefined, limiting interpretation of H3K27me3 progenitor data.
• Some mutated genes (DDX31, GABRG1) were detected at low frequency and await independent validation cohorts.
• The study population is pediatric; generalizability to adult medulloblastoma is unclear.
• Chromothripsis was rare (2/37) but its contribution to subgroup-3/4 pathogenesis warrants further study.

Citations from this paper used in the wiki

• “Recurrent mutations were detected in 41 genes not yet implicated in medulloblastoma: several target distinct components of the epigenetic machinery in different disease subgroups” (Summary).
• “Six subgroup-4, one subgroup-3, and one unclassified medulloblastoma contained novel inactivating mutations in KDM6A” (p. 5).
• “57% (n=4/7) of KDM6A-mutant male medulloblastomas deleted chromosome Y, compared with only 6% (n=3/51) of male, KDM6A wild-type tumours (P<0.005)” (p. 5).
• “Four WNT-subgroup tumours contained heterozygous missense mutations in the helicase domain of SMARCA4 (P<0.002)” (p. 6).
• “100% (n=11/11) of Blbp-Cre; Ctnnb1+/lox(Ex3); Tp53+/flx; Pik3caE545K mice developed WNT-subgroup medulloblastomas by 3 months of age” (p. 8).
• “Inhibitors of the epigenetic machinery, especially those that maintain H3K27me3 e.g., EZH2 methylases, may be useful treatments of subgroup-3 and 4 disease” (Discussion).

This page was processed by paper-compiler on 2026-05-06.