AUTHORS
Mayuresh Konde, PhD, Philomina Sona Peramangalam2, Emin Onur Karakaslar3, Sebastian Wolf, MD4, Shikan Zheng5, Akmaljon Salimov6, Sridevi Surapally, PhD7, Marieke Griffioen, PhD8, Tongjun Gu, PhD9, Sridhar Rao, MD, PhD10, Daniel Tenen, MD11, Thomas Oellerich, MD12, Erik Van Den Akker13, Martin Carroll, MD14, Caner Saygin, MD15, John Pulikkan, PhD16
AFFILIATIONS
1 Versiti blood research institute, Milwaukee, Wisconsin, United States
2 Versiti blood research institute, Milwaukee, Wisconsin, United States
3 Leiden University Medical Center, Leiden, Netherlands
4 Goethe University Frankfurt, Frankfurt, Germany
5 Versiti blood research institute, Milwaukee, United States
6 University of Pennsylvania, Philadelphia, United States
7 Versiti blood research institute, Milwaukee, Wisconsin, United States
8 Leiden University Medical Center, Leiden, Netherlands
9 Versiti blood research institute, Milwaukee, Wisconsin, United States
10 Versiti blood research institute, Milwaukee, Wisconsin, United States
11 Harvard Medical School, Boston, United States
12 Goethe University Frankfurt, Frankfurt, United States
13 Leiden University Medical Center, Leiden, Netherlands
14 University of Pennsylvania, Philadelphia, United States
15 University of Chicago, Chicago, United States
16 Versiti blood research institute, Milwaukee, Wisconsin, United States
ABSTRACT
C/EBPα is a master regulator of granulopoiesis, and CEBPA mutations occur in 10–15% of acute myeloid leukemia (AML) cases. Based on mutation status, CEBPA-mutant AML is classified into monoallelic N-terminal (CEBPA-N-AML), which expresses the truncated 30-kDa C/EBPα-p30 isoform, monoallelic C-terminal (CEBPA-C-AML), and biallelic N- and C-terminal (CEBPA-NC-AML) subtypes. Despite the markedly inferior survival of patients with CEBPA-N-AML compared with the other subtypes, differences in leukemic engraftment and the identity of leukemic stem cells (LSCs) across these subtypes remain poorly defined. Here, we investigated the leukemogenic potential of primary human CEBPA-mutant AML samples in NSGS mice and identified a C/EBPα-p30-driven LSC population specific to CEBPA-N-AML.
Male and female NSGS mice (6–8 weeks of age) were sublethally irradiated and T-cell-depleted primary AML cells from patients with CEBPA-N-AML, CEBPA-C-AML, and CEBPA-NC-AML (n = 5 independent patient samples per subtype) were transplanted via tail vein injection (1 × 10⁶ cells per mouse; 5 mice per patient sample). Human AML engraftment was monitored longitudinally by flow cytometric analysis of human CD45⁺CD33⁺ cells in bone marrow aspirates, and disease progression was assessed by serial peripheral blood analysis.
CEBPA-N-AML samples consistently exhibited significantly greater engraftment than CEBPA-C-AML or CEBPA-NC-AML samples. All mice transplanted with CEBPA-N-AML patient samples (25/25) developed AML, with median disease latencies ranging from 20 to 70 days depending on the patient sample. In contrast, leukemia developed only in recipients of one CEBPA-C-AML sample (5/25 mice) and one CEBPA-NC-AML sample (3/25 mice), both harboring WT1 mutations. Patients whose AML samples successfully engrafted and propagated leukemia in mice had significantly worse overall survival than patients whose samples failed to establish leukemia in vivo.
CEBPA-N-AML recipients showed marked expansion of human myeloid blasts in the bone marrow and peripheral blood, with extensive leukemic infiltration of the liver, spleen, and lungs. Secondary transplantation efficiently propagated disease that faithfully recapitulated the primary leukemia, demonstrating durable leukemia-propagating activity and self-renewal.
To determine whether CEBPA-N AML samples cluster together based on their global transcriptomic profiles, we performed t-SNE analysis of a multi-cohort AML transcriptomic dataset comprising 1,224 patients. t-SNE analysis showed that CEBPA-N-AML samples were dispersed across multiple transcriptomic clusters, indicating substantial molecular heterogeneity. These findings prompted us to investigate whether the aggressive biology of CEBPA-N AML is driven by a shared leukemic stem cell (LSC) program rather than by global transcriptomic similarities.
To identify CEBPA-N-AML LSC markers, we integrated previously reported AML LSC markers with immunophenotypic markers associated with CEBPA-mutant AML from multiple independent studies. Candidate markers were further prioritized based on direct C/EBPα-p30 binding identified by CUT&Tag analysis of leukemic cells from a conditional C/EBPα-p30 knock-in mouse model generated in our laboratory. This integrative strategy identified a CD366⁺CD73⁺CD123⁺CD117⁺CD371⁺CD247⁺ immunophenotype enriched for LSC activity. Analysis of individual marker expression within the CD33⁺ blast population of primary human CEBPA-N AML revealed variable expression frequencies for each marker. In contrast, cells co-expressing all six markers comprised only 0.05–0.9% of the CD33⁺ blast population, consistent with the expected rarity of leukemic stem cells. Sorted human primary CEBPA-N-AML LSC⁺ (CD33⁺CD366⁺CD73⁺CD123⁺CD117⁺CD371⁺CD247⁺) cells efficiently initiated leukemia following transplantation into NSGS mice, whereas LSC⁻ (CD33⁺CD366⁻CD73⁻CD123⁻CD117⁻CD371⁻CD247⁻) cells failed to engraft. LSC⁺ cells also exhibited significantly greater colony-forming capacity than LSC⁻ cells. Ongoing RNA-seq studies are investigating the molecular pathways that regulate LSC function in CEBPA-N-AML.
In summary, CEBPA-N-AML exhibits superior leukemogenic potential and a functionally validated six-marker LSC immunophenotype that enriches leukemia-propagating cells. These findings establish a robust patient-derived xenograft platform and identify the C/EBPα-p30-driven stem cell program as a therapeutic target in this high-risk AML subtype.
DISCLOSURES
Mayuresh Konde, PhD: No relevant conflicts of interest to declare.