AUTHORS
Xiaodi Wu, MD, PhD, Maya C. Bagish2, Matthew G. Wereski3, Michael R. Waarts4, Rebecca G. Gelfer, PhD5, Wenbin Xiao, MD, PhD6, Sheng F. Cai, MD, PhD7, Ross L. Levine, Prof. Dr.8
AFFILIATIONS
1 Memorial Sloan Kettering Cancer Center, New York, New York, United States
2 Memorial Sloan Kettering Cancer Center, New York, New York, United States
3 Memorial Sloan Kettering Cancer Center, New York, New York, United States
4 Memorial Sloan Kettering Cancer Center, New York, New York, United States
5 Memorial Sloan Kettering Cancer Center, New York, New York, United States
6 Memorial Sloan Kettering Cancer Center, New York, New York, United States
7 Memorial Sloan Kettering Cancer Center, New York, New York, United States
8 Memorial Sloan Kettering Cancer Center, New York, New York, United States
ABSTRACT
In-frame indels in the basic leucine zipper (bZIP) domain of transcription factor C/EBPα, encoded by CEBPA, characterize a subset of acute myeloid leukemia (AML) with distinct prognosis (Georgi et al., 2024) but no specific targeted therapies. For preclinical studies, patient-derived xenografts of this disease subset are difficult to establish and select for additional co-mutations (Kawashima et al., 2022). By contrast, Cebpa mutations that do not encode bZIP in-frame indels are sufficient to generate leukemias in mice and have been extensively studied (Kirstetter et al., 2008; Heyes et al., 2023). With increasing refinements in prognostic stratification, however, it is unclear whether leukemias carrying “other” mutations can adequately model the biology of human disease with favorable prognosis.
We devised a strategy to knock into the native Cebpa locus an inducible bZIP in-frame indel by employing a synthetic intron to divide the wild-type coding sequence and embed, in the reverse orientation, a codon-optimized 3′-fragment of mutant coding sequence; LE/RE mutant lox sites were positioned to enable Cre recombinase–mediated induction of the mutant allele. Expression of HA-tagged wild-type protein at baseline and FLAG-tagged mutant protein upon Cre-mediated recombination was demonstrated by western blot. Crosses between mice heterozygous for the allele produced offspring in a genotypic ratio not significantly different from Mendelian expectation. Yet homozygotes exhibited splenomegaly and defects in bone marrow progenitor development consistent with baseline hypomorphic expression of C/EBPα, likely attributable to incomplete splicing of the synthetic intron; intracellular C/EBPα staining confirmed decreased protein abundance at baseline among peripheral blood monocytes by flow cytometry. Therefore, the targeted allele disaggregates the effects of decreased wild-type protein activity from any residual or neomorphic effects specific to mutant protein activity.
We used our inducible allele to interrogate the cistrome in lineage-depleted bone marrow progenitors with or without Cre-mediated indel induction, finding a subset of wild-type binding sites bound by mutant protein. Among retained binding sites were those at the Gata2 distal hematopoietic enhancer (−77 kb; Grass et al., 2006) and the Zeb2 −165 kb enhancer (Huang et al., 2021). Known-motif analysis revealed that a bipartite C/EBP motif was most substantially enriched in the mutant cistrome as compared to the wild-type cistrome, suggesting a possible composite motif hijacking mechanism. However, with the protein expressed at sub-physiologic abundance, differential gene expression analysis in primary lineage-depleted bone marrow progenitors could not identify transcriptomic correlates of such activity in bulk RNA.
When crossed to a constitutive amino-terminal frameshift mutant to recapitulate biallelic CEBPA-mutated AML, compound heterozygotes exhibited bone marrow dysplasia at baseline that transformed reliably to leukemia, with or without Cre-mediated indel induction. Disease latency in primary transplant recipients was similar to that of published constitutive models (Bereshchenko et al., 2009). These leukemias showed stereotyped acquisition of RAS pathway activating mutations, particularly in Kras exon 4 and Ptpn11, a genetic profile also found in human AML with “other” CEBPA mutations—that is, those that do not encode bZIP in-frame indels. Moreover, among recipients of induced bZIP in-frame indel donor cells, leukemic outgrowth was concomitant with decreasing frequency of the induced allele, suggesting that the indel may impair competitive fitness in the context of insufficient C/EBPα. Thus, increased mutant protein abundance, cooperating mutations, or both may be required to recapitulate bona fide bZIP in-frame indel AML and identify its unique vulnerabilities.
DISCLOSURES
Xiaodi Wu, MD, PhD: No relevant conflicts of interest to declare.