In this issue of Blood, Piqué-Borràs et al1 investigate how the NFIA-ETO2 fusion perturbs erythropoiesis and show how this oncoprotein cooperates with mutant TP53 in leukemogenesis.
Pure erythroid leukemia (PEL) is a rare and aggressive subtype of acute myeloid leukemia (AML), and driver NFIA-ETO2 fusions have been restricted to pediatric PEL patients to date. The transcription factor NFIA is a key regulator of erythroid differentiation during early hematopoiesis.2 Similarly, ETO2 (also known as CBFA2T3) is a transcriptional corepressor that appears to also play a role in erythroid differentiation through the coregulation of GATA1 targets.3 In addition, the ETO2 gene is a frequent fusion partner with GLIS2 and RUNX1 in megakaryoblastic and myeloblastic forms of AML, respectively. This suggests that ETO2 broadly impacts cell fate decisions in hematopoietic stem and progenitor cells (HSPCs) in a cell context–dependent manner. Although the NFIA-ETO2 oncoprotein is exclusively associated with PEL, how it functions in leukemogenesis remains poorly understood. In this study, the authors used multiple complementary model systems to show that NFIA-ETO2 impairs erythroid differentiation while simultaneously driving proliferation but is not sufficient to induce full leukemic transformation. By contrast, enforced NFIA-ETO2 expression in primary murine hematopoietic cells harboring a recurring point mutation (Tp53R248Q/+) induced a highly penetrant erythroleukemia in mice.
Piqué-Borràs et al first showed that expressing NFIA-ETO2 in mouse erythroleukemia (MEL) cells and in primary murine erythroblasts impaired terminal erythroid differentiation in vitro. Mutagenesis experiments demonstrated that the DNA-binding domain of NFIA and the protein interaction NHR2-4 motifs of ETO2 are essential for inducing this differentiation block. Paired chromatin immunoprecipitation sequencing and transcriptome (RNA-seq) analyses showed NFIA-ETO aberrantly binds near key erythroid differentiation regulators such as Tal1,Gfi1b, and Klf1 and that this is associated with decreased messenger RNA expression. By contrast, Myc and Myb expression was markedly up-regulated. Mechanistically, the use of RNA interference to reduce Myb expression was sufficient to overcome the erythroid differentiation block. The authors went on to validate these general observations in MEL cells by performing RNA-seq in primary mouse fetal liver–derived erythroblasts engineered to express either “wild-type” NFIA-ETO2 or an inactive control fusion lacking the ETO2 NHR4 domain. Altogether, these data support a model whereby NFIA-ETO2 acts, in part, by enforcing an immature and proliferative state in cells that are partially committed to the erythroid lineage through aberrant expression of MYC, MYB, and perhaps other myeloid transcription factors.
Despite these widespread effects on cell fates and key transcriptional networks, exogenous NFIA-ETO2 expression neither enhanced colony replating potential ex vivo nor initiated leukemia in vivo in a transduction/transplantation assay. These data are consistent with an essential role for cooperating mutations in leukemic transformation, which is a biologic property of other AML-associated transcription factor fusions. An intriguing aspect of NFIA-ETO2–driven PEL is that it has only been reported in pediatric patients. This suggests that fetal/neonatal HSPCs are particularly (and perhaps uniquely) vulnerable to transformation by NFIA-ETO2 as has been observed for some KMT2A (also known as MLL) fusions in infant AML and for mutant Ras pathway genes in juvenile myelomonocytic leukemia. Future studies might address the effects of expressing NFIA-ETO2 in HPSCs isolated from human umbilical cord blood vs adult bone marrow.
TP53 is frequently altered in PEL, most often owing to point mutations with or without concurrent TP53 deletions (often manifesting as chromosome 17 partial deletions).4 A key question for addressing the effects of mutant TP53 in mouse models of hematologic malignancies involves the relative merits of using a “first-generation” Tp53 null allele or subsequent knockin models encoding highly prevalent point mutations found in human cancers. These respective strains spontaneously develop a different spectrum of primary malignancies.5,6 Although there remains some debate about the precise mechanisms underlying TP53-mediated oncogenesis in different tissues contexts, these in vivo data and studies in AML cell lines support dominant-negative biologic activity of mutant TP53.7 In this context, it is notable that Piqué-Borràs et al compared the transforming activity of expressing NFIA-ETO2 in bone marrow cells from both heterozygous Tp53 knockout (Tp53+/−) and TP53R248Q/+ knockin mice. Remarkably, recipients of NFIA-ETO2; TP53R248Q/+ “double mutant” cells developed a fully penetrant transplantable erythroleukemia, whereas mice transplanted with NFIA-ETO2; Tp53+/− cells did not. These data have implications for characterizing potential dominant-negative and gain-of-function activities of mutant TP53 in leukemogenesis and for future studies of how it cooperates with other leukemia-associated transcription factor oncoproteins.
Mechanistically, the authors’ studies support a model whereby NFIA-ETO2 promotes aberrant proliferation of early erythroid cells and impairs their differentiation with mutant TP53 cooperating by promoting self-renewal (see figure). These data are broadly similar to previous studies that investigated the cooperating and competing effects of oncogenic Nras/Kras alleles and Tp53 inactivation in leukemogenesis.8,9 Although somatic TP53 mutations are highly prevalent in other subtypes of PEL, it should be noted that they have not been detected in patients with NFIA-ETO2 fusions. This raises potential questions about using the new transplantable model reported here for future biologic and preclinical studies. With this said, few model systems for pure erythroleukemia exist, and the authors’ findings that transcriptional rewiring induced by NFIA-ETO2 expression in primary bone marrow cells cooperate with mutant TP53 to induce erythroleukemia represent a novel contribution to the field.
Fetal liver–derived TP53R248Q/+ murine erythroblasts retrovirally transduced with the NFIA-ETO2 fusion and transplanted into recipients induces a highly penetrant, transplantable erythroleukemia. By contrast, expressing NFIA-ETO2 in other genetic contexts (wild-type or Tp53+/− mutant cells) and transducing TP53R248Q/+ mutant cells with the control vector did not cause leukemia. Ctrl, control; FL, fetal liver; WT, wild type. Professional illustration by Somersault18:24.
Fetal liver–derived TP53R248Q/+ murine erythroblasts retrovirally transduced with the NFIA-ETO2 fusion and transplanted into recipients induces a highly penetrant, transplantable erythroleukemia. By contrast, expressing NFIA-ETO2 in other genetic contexts (wild-type or Tp53+/− mutant cells) and transducing TP53R248Q/+ mutant cells with the control vector did not cause leukemia. Ctrl, control; FL, fetal liver; WT, wild type. Professional illustration by Somersault18:24.
Conflict-of-interest disclosure: The authors declare no competing financial interests.