C-Terminal BRE Overexpression in 11q23-Rearranged and t(8;16) Acute Myeloid Leukemia Is Caused By Intragenic Transcription Initiation

Abstract Text

Overexpression of the BRE ( Brain and Reproductive organ-Expressed ) gene defines a distinct pediatric and adult acute myeloid leukemia (AML) subgroup (Balgobind et al, Leukemia 2010, and Noordermeer et al, Blood 2011 and 2012). BRE overexpression specifically associates with 11q23/MLL/KMT2A-rearranged AML, occurring in 70% and 50% of 11q23 childhood and adult AML cases, respectively. In adult 11q23-rearranged AML, BRE overexpression significantly associates with younger age. BRE functions in BRCA1-A (Breast Cancer 1) complex mediated DNA damage repair, BRISC (BRCC3 isopeptidase complex) complex mediated mitotic spindle assembly and in protection from apoptosis and senescence. Another constituent of the BRCA1-A/BRISC complexes is BRCC3 which is recurrently mutated in AML, especially in cases with a TET2 mutation (Huang et al., Haematologica, 2015). BRE is relatively highly expressed in blood cells in general but its exact function in hematopoiesis is unknown.

By performing genome-wide ChIP-sequencing analyses on primary KMT2A-MLLT3 patient material (n=9) we identified a promoter enriched for active chromatin marks (H3K4me3/H3K27Ac) in BRE intron 4, exclusively in AML cases with BRE overexpression. Both 5' RACE end RNA-sequencing identified strong biallelic expression of a previously unknown C-terminal BRE transcript. This transcript starts near the intragenic promoter, with BRE intron 4 sequences spliced to exon 5 and downstream sequences. RNA-seq and qRT-PCR showed that the downstream transcript is expressed 5-8-fold higher compared to the full-length transcript, explaining BRE overexpression. Of note, two start codons in exon 5 of the downstream BRE transcript may cause the formation of N-terminally truncated BRE isoforms.

By performing a qRT-PCR that specifically recognizes the new BRE transcript on samples from a large cohort of pediatric and adult AML (n=195) we observed that it was highly expressed in 60% of 11q23/ KMT2A -rearranged (n=58) and 70% of t(8;16)/ KAT6A-CREBBP cases (n=20), while it was virtually absent from other AML subsets (n=101) and normal hematopoietic tissues (n=14). In gene reporter assays, the AML-specific fusion protein KMT2A-MLLT3 transactivated the intragenic BRE promoter. Further epigenome analyses revealed 97 additional intragenic promoter marks frequently bound by KMT2A in AML with C-terminal BRE overexpression. The bound genes may be part of a context-dependent KMT2A-MLLT3-driven oncogenic program, because they were higher expressed in this AML subtype compared to other cytogenetically-defined AML groups (p<0.01). C-terminal BRE might be a contributor to this oncogenic program because in a case with relapsed AML, we observed an ins(11;2) fusing CHORD1 in frame to BRE exon 5. The CHORDC1 fusion to BRE exon 5 is remarkable because the alternate BRE transcript starts with exon 1b also fused to exon 5. This means that if the CHORDC1-BRE and alternate BRE transcripts would be translated, they would contain a similar C-terminal part of the BRE protein. The CHORDC1 disruption may also contribute to AML pathogenesis because it was recently shown that heterozygous Chord1 (a.k.a. Morgana) inactivation in mice results in the spontaneous development of atypical chronic myeloid leukemia (Di Savino, Blood, 2015). We conclude that abnormal expression of C-terminal alternate BRE sequences specific for AML is caused by distinct, non-random events; by a chromosomal abnormality and by intragenic transcription initiation in 11q23/ KMT2A -rearranged and t(8;16)/ KAT6A-CREBBP AML. These findings warrant further studies to determine whether leukemia-specific alternate BRE contributes to disease pathogenesis and how abnormal intragenic transcription activation in AML is brought about.