Mutant RUNX1 and histone tales

In this issue of Blood, Lehnertz et al identify rare cases of histone H3 mutations at position K27 that alter methylation levels and cooperate with RUNX1 mutations,¹  further strengthening the association between RUNX1 and polycomb repressive complex 2 (PRC2) activity in leukemic transformation.

Chromatin structure is critical for transcription factor activity, and many enzymes modify the lysine (K) residues found within histone tails.²  These modifications regulate how the histone proteins pack DNA into chromatin. PRC2 comprises 4 core proteins: EZH2, SUZ12, EED, and RBBP4, along with its accessory proteins, including ASXL1, AEBP2, JARID2, and the polycomb-like proteins. This multiprotein complex catalyzes the methylation of K27 in the histone H3 tail (see figure, right side).²  This position is recognized by polycomb repressive complex 1 (PRC1), which performs additional histone modification steps to repress the chromatin (see figure, right side).² 

Disruption of the histone tail modifications can occur by several means, and the article presented in this edition of Blood demonstrates the cooperation between RUNX1 mutations and histone H3 K27M and K27I point mutations in leukemogenesis. Because the PRC2 complex can methylate only a lysine residue, these mutations block PRC2 activity at these sites (see figure, left side). Among 415 patients, Lehnertz and colleagues identified 2 who had H3 K27 mutations, one with K27M and one with K27I. In addition, the group identified a similar patient with a K27M mutation in the 200 patients with acute myeloid leukemia (AML) described in The Cancer Genome Atlas database. Lehnertz and colleagues elegantly demonstrate that although these mutations affect a subclone within the leukemia cell populations, they are functional, with lower levels of H27 methylation demonstrated within the mutated clones shown by fluorescence-activated cell sorter analysis. Importantly, each of these 3 leukemias had somatic mutations in addition to the H3 K27 mutation: all 3 had an alteration in RUNX1; the H3K27I mutant also had a t(8;21); and both of the H3K27M mutants possessed an accompanying RUNX1 truncating mutation and 1 also had a RUNX1 point mutation, although it could not be determined whether those were on the same allele. Moreover, each of the 3 leukemias had additional mutations in genes that encode chromatin modifiers: ASXL1, ASXL2, BCOR, TET2, and USP7. One also had heterozygous loss of EZH2 due to a chromosomal rearrangement. Importantly, several of these genes have been implicated previously in RUNX1-mediated malignancies, thereby providing strong circumstantial evidence that chromatin structure is critical for abnormal RUNX1 proteins in driving tumorigenesis.^3-5 

To test whether the H3K27 and RUNX1 mutations cooperate in leukemogenesis, Lehnertz et al turned to in vitro and mouse modeling. They demonstrate that the H3K27M/I mutations led to a proliferation advantage in the context of an AML-ETO9a fusion protein in vitro and accelerated leukemia development in vivo. The authors also show that on its own, the H3K27M mutation results in expansion of primitive hematopoietic cells. The results from this work fit within a broader literature that has demonstrated the oncogenic properties of H3K27 mutations in diffuse midline gliomas.^6-8  In these tumors, the H3K27M mutation also results in global loss of H3K27 methylation levels.

Given the importance of chromatin structure in transcription factor binding, several questions remain to define how H3K27 mutations and lower K27me levels cooperate with RUNX1 mutations: where are the mutant histones located in relation to RUNX1 binding sites? How do the other mutations in genes that encode chromatin-modifying proteins contribute to the altered chromatin structure in these leukemias? Do the RUNX1 mutants/fusion proteins bind DNA at loci that lack H3K27me or elsewhere? We await the next chapter in the tale of mutant RUNX1 and histones!