Glioblastoma multiforme (GBM) and diffuse intrinsic pontine glioma (DIPG) are aggressive subtypes of brain tumors that both have a very poor prognosis and are almost always lethal. Two new studies in Nature and this journal today identify the same recurrent mutations in H3F3A in pediatric cases of glioblastoma multiforme and diffuse intrinsic pontine glioma. These are the first reports of human disease associated with mutations in histones, which play an extraordinarily important and conserved role in chromatin structure and gene regulation. With the recent spate of papers reporting somatic mutations in chromatin remodelers in various types of cancer (examples from this journal alone include the histone H3K27 demethylase UTX, transitional cell carcinoma of the bladder, histone methyltransferase EZH2 in follicular and diffuse large B-cell lymphoma and myeloid disorders, DNMT3A in AML, ARID2 in hepatocellular carcinoma, MLL2 in DLBCL, and ARID1a in gastric cancer) it is clear that targeting the chromatin remodeling machinery will be an important area in the development of new cancer drugs.

Around 3000 children are diagnosed with brain tumors each year in the United States. DIPG is a type of aggressive brainstem tumor that occurs almost only in children. The Nature study reports somatic mutations in the H3.3-ATRX-DAXX chromatin remodeling pathway in 44% (21/48) of tumors, while the Nature Genetics study reports mutations in either H3.1 or H3.3 in 60% (52/86) of tumors. Remarkably, the Nature Genetics study finds that 78% (39/50) of DIPG tumors display a p.Lys27Met change in either histone H3.1 or H3.3.

Suzanne Baker and colleagues report in this journal today whole-genome sequences of 7 DIPGs and matched normal tissue. Four of the tumors harbored a p.Lys27Met change in H3.3 and one of the tumors showed a p.Lys27Met change in the related histone variant H3.1. The authors subsequently sequenced the genes encoding H3.3 (H3F3A) and H3.1 (HIST1H3B) in 43 more DIPGs and 36 non-brainstem pediatric glioblastomas. In total, 39/50 DIPGs carried a p.Lys27Met change in H3F3A or HIST1H3B. 13/36 non-brainstem pediatric glioblastomas harbored a p.Lys27Met change in H3F3A or HIST1H3B or a p.Gly34Arg change in H3F3A. The authors also sequenced the16 histone H3 genes in other types of pediatric brain tumors but found no other histone H3 mutations in these lower-grade tumor subtypes.

Nada Jabado and colleagues report in Nature today 48 whole exomes of pediatric GBMs, as well as matched normal tissue for 6 of those samples. Two of these samples harbored the p.Lys27Met change in H3.3 and 2 samples harbored a p.Gly34Arg change in H3.3. After extending the analysis to 48 whole exomes, the authors found that 44% (21/48) of samples harbored mutations in H3F3A, ATRX or DAXX. It is particularly notable that the two amino acids affected in H3.3 (K27 and G34) are at or in close proximity to sites that are important post-translational modifications. Trimethylation of K27 (H3K27me3) is associated with silencing of genes whereas K36 is associated with transcriptional activation.

Mutations in histones have not been reported in cancer (or in any other human disease), although somatic mutations in genes regulating histone modifications have been reported in cancer. It is clear that different histone variants are associated with different chromatin and transcriptional states. In particular, H3.3 is enriched at sites of active gene transcription and regulatory elements. Jabado and colleagues speculate that the finding of the same mutations in different tumors, plus the lack of truncating mutations, suggests that these mutations here are gain-of-function. However, the precise mechanism of action is hard to predict. Analysis of gene expression in 27 of the whole-exome samples shows that tumors with the K27 or G36 mutations have distinct profiles, suggesting that each mutation leads to tumors in different ways. Regardless, both papers show a central role for the chromatin remodeling machinery in pediatric gliomagenesis and mutations in histones as another way that epigenetic events drive cancer.