The Spectrum of Epigenetic Mutations in Myeloma and Their Clinical Impact

Dysregulation of the epigenome plays an important role in the pathogenesis of the plasma cell malignancy myeloma (MM). For example the H3K36 methyltransferase, MMSET, is overexpressed as a result of t(4;14) in 15% of patients and associated with a distinct DNA methylation pattern and shorter survival. Epigenetic modifiers may also be deregulated due to somatic mutations, seen in the histone demethylase, KDM6A/UTX (Van Haaften et al, Nat.Genet. 2009) and histone methyltransferase, MLL(Chapman et al, Nature 2011). We analysed the spectrum and clinical implications of epigenetic gene mutations in the largest series of newly diagnosed MM patients sequenced to date.

Whole exome sequencing was performed on DNA extracted from tumour (CD138+) and peripheral blood samples from patients entering the NCRI Myeloma XI trial (n=463) using SureSelect (Agilent) with extra baits to cover IGH, IGK, IGL and MYC loci, median depth 60x. Data were processed to identify acquired variants, copy number, indels and translocation breakpoints and annotated for potentially deleterious mutations. Significantly mutated genes were detected using MutSigCV (v1.4) inputting all SNV and short indels (q-value threshold 0.1). Survival from initial randomization had median follow up of 25 months.

The gene encoding the linker histone protein Histone 1.4, HIST1H1E was mutated in 2.8% of samples and one of the most significantly mutated genes in myeloma (p<1x10^-10, q<1x10^-10). Average cancer cell fraction in HIST1H1E mutated samples was close to 100%, suggesting these mutations may be an early event in MM pathogenesis. There were also recurrent mutations in genes encoding variants of the Histone 1 protein, HIST1H1B (0.22%), HIST1H1C (2.59%), HIST1H1D (0.65%); the percentage of patients with a mutation in any variant totals 6% (28/463). Mutations clustered in the globular domain and multiple sequence alignment revealed sites of recurrent mutation across variants. Along with the absence of mutations in the fifth common protein variant HIST1H1A,this suggests that these are not passenger mutations and may carry some significance to MM pathogenesis. Histone 1 mutations have been demonstrated to play an important role in other haematological malignancies but have not been previously characterized in MM.

Potentially deleterious mutations in histone methyltransferase/demethylase enzymes were also seen in 24% of patients, though the percentage of patients with each gene mutated was low. The most frequently mutated gene family in the methyltransferases was MLL/2/3/4/5 (7% of patients). There were no mutations in EZH2, recurrently mutated in other B cell malignancies, and none of the MMSET activating mutations p.E1099K described in the MM1.S myeloma cell line were seen. The most frequently mutated demethylase gene was KDM3B in 1.5% of patients. KDM6A/UTX mutations occurred in 1.3% of patients and targeted analysis for deletion of whole exons increased the number of patients affected by a potentially inactivating lesion to 3%. Patients carrying a KDM6A mutation or deletion appear to have a shorter OS at current follow up than wild type (medians NR, log-rank p=0.0498, % alive at 2 years 51% CI 30-85 vs 80% CI 77-84). Data suggest that EZH2 inhibitors, currently in development for lymphoma, could be investigated for these patients as inhibiting the H3K27 methyltransferase may counteract the increased H3K27 methylation resulting from inactivation of the demethylase.

DNA methylation modifiers were found to be mutated in 4% of patients. These include mutations previously reported in glioma (p.R132C in IDH1) and AML/MDS (p.R140W in IDH2 and p.C1378Y in TET2). Collectively, mutations in any DNA methylation modifier (TET1/2/3 n=9, IDH1/2 n=2 or DNMT1/3A/B n=6) are associated with a shorter OS (medians NR, p=0.045, % alive at 2 years 58% CI 39-88 vs 80% CI 76-84). Patients with these mutations might be amenable to demethylating agents such as azacytidine and newer agents such as IDH inhibitors currently in early stages of development.

This is the first extensive analysis of the spectrum of mutations in epigenetic modifiers in a uniformly treated population in MM. An association with clinical outcome is suggested in our dataset but will need validation due to the low overall frequency of the mutations. This data further emphasises the importance of epigenetics in MM and provides new potential targets for personalised therapeutic strategies for patients.