Molecular mechanisms of acute erythroid leukemia: learning from rare chromosomal translocations in pediatric patients

Two novel fusion genes, NFIA-ETO2 and ZMYND8-RELA, have been identified in pediatric
patients suffering from pure erythroid leukemia (PEL). Based on the normal function of the
fusion partners, we hypothesized that both fusions might initiate PEL by interfering with
erythroid differentiation. I studied the role of these two fusion genes in erythropoiesis and found
that ZMYND8-RELA significantly reduced the viability of erythroid progenitor cells. On the other
hand, expression of the NFIA-ETO2 fusion significantly impaired terminal differentiation of
murine erythroblasts and slightly increased their self-renewal capacity. However,
transplantation of these cells into lethally irradiate mice did not induce any disease. Mutations
of the tumor suppressor TP53 have been reported to be molecular hallmarks of PEL. I therefore
investigated whether NFIA-ETO2 cooperates with one of the most prevalent PEL-associated
TP53 mutation (R248Q). I found that NFIA-ETO2 not only increased self-renewal of TP53R248/+
erythroblasts, but also induced a PEL-like disease upon transplantation of the cells into lethally
irradiated recipients. To understand how NFIA-ETO2 expression interferes with erythroid
differentiation, I performed transcriptome analysis and found that NFIA-ETO2 expression was
associated with up-regulation of proto-oncogenes such as Myb, Notch2, Myc and Stat5a and
down-regulation of GATA1 and NFIA target genes related to erythropoiesis. Surprisingly the
presence of the TP53R248Q mutation in NFIA-ETO2 erythroblasts only resulted in
downregulation of few genes, mainly previously characterized TP53 targets. Chromatin
analysis (ATAC-seq) showed that these genes presented a positive enrichment in their
promoter and predictive transcription factor analysis suggested that they could be repressed
by the cooperative activity between the TP53R248Q and the PRC2 (Polycomb complex 2)
epigenetic key regulator.
As the molecular drivers of acute erythroleukemia are mostly unknown, we decided to
study the epigenomic landscape of the disease in more detail. In collaboration with Thomas
Mercher, we collected patient samples worldwide for RNA-seq and Exome-seq. We found that
erythroleukemia patients can be genetically grouped into at least 3 groups, those with TP53
mutations, those with mutations in epigenetic/metabolic regulators and patients with very few
mutations. Interestingly, we found that the transcriptome, while poorly correlating with the
mutation status, reflected the maturation stage of the erythroblasts, and was clearly different
from signatures from most MDS or other AML. Overall, the results from my PhD project
provide some functional insight into molecular mechanisms of acute erythroleukemia that may
in the long run help to improve diagnosis and therapy for this rare but aggressive disease.