Abstract
Congenital neutropenia (CN) is a heterogeneous inherited preleukemia bone marrow syndrome, with autosomal dominant mutations in ELANE, encoding the neutrophil elastase protein, being the most common genetic cause while mutations in other genes are less frequent. Despite this genetic heterogeneity, CN patients who progress to myelodysplastic syndrome (MDS) or acute myeloid leukemia (AML) exhibit a similar pattern of somatic mutation acquisition. CN/AML patients frequently acquired somatic truncating mutations in CSF3R, which encodes the granulocyte colony-stimulating factor (G-CSF) receptor, followed by the co-acquisition of RUNX1 mutations, typically located within the RUNT-homology domain. Here, we investigate the molecular mechanisms involved in step-wise leukemogenic transformation in ELANE-CN patients using an in-vivo mouse model.
We transduced lineage-depleted hematopoietic cells from mice carrying a CN/AML-associated truncated Csf3rd715 mutation (Liu et al., 2008) with lentiviral constructs expressing the hot-spot ELANE mutation p.C151R, either in combination with wildtype (WT) or RUNX1 p.R139G or p.R174L mutations frequently detected in CN/AML patients. Genetically modified cells were injected intravenously into lethally irradiated Ly5.1 recipient mice. Transplanted mice were chronically treated with recombinant human G-CSF (rhG-CSF) at a dose of 100µg/kg three times a week for 52 weeks to mimic CN patient condition. Engraftment and clonal expansion were regularly monitored in peripheral blood by tracking GFP+RFP+ cells, co-expressed with ELANE and RUNX1, respectively. At experimental endpoint, transplanted cells from bone marrow, peripheral blood and spleen were examined by flow cytometry and morphological analysis. Furthermore, barcoding analysis was performed to track mutant clones in vivo, and RNA-seq was used to identify deregulated signaling pathways potentially involved in leukemogenesis.
All populations successfully engrafted, but expansion varied depending on the mutational status. The highest expansion was observed in mice with the triple mutation (Csf3r d715/ELANEC151R/RUNX1R174L), showing 3.5 to 20-fold increase in expansion independently of treatment. In contrast, a low 1.46-fold increase was observed in triple mutants harboring RUNX1R139G when treated with rhG-CSF. Cells expressing Csf3r d715/ELANEC151R/RUNX1WT did not expand. Immunophenotypic analyses of mature and immature populations revealed a significant increase of LK (Lin-ckit+) cells and myeloid lineages in triple mutants along with a decrease in lymphoid cells. Of note, rhG-CSF treatment alone showed a similar trend in those populations when compared to placebo. Morphological examination of bone marrow confirmed the presence of immature blast in mice transplanted with cells carrying all three mutations, as compared to WT control. Barcoding analysis revealed the dynamic evolution of clonality in the peripheral blood, showing traceable unique clones over time and matching clone distributions between bone marrow and peripheral blood. We observed that RUNX1 mutations provide a proliferative advantage to transduced cells, thereby increasing clone size. Interestingly, rhG-CSF treatment strongly reduced the clonal diversity of RUNX1R139G clones, compare to placebo, which may promote the pre-selection of fewer dominant expanding clones. RNA-seq data revealed deregulation of immune response-activating signaling pathway, calcium ion transport, intrinsic apoptotic signaling pathway, and chemotaxis in RUNX1R174L compared to RUNX1WT cells. Additionally, mRNA expression of the immune reaction-controlling CD48 and the HSPA1a and HSPA1b, as well as the transcription factors GFI1b and Irf8 was significantly downregulated, while the expression of the signal-transducing adapter protein 2 (STAP2), Follistatin (FST), and Chil3 was elevated.
Taken together, our group has established an in-vivo mouse model to investigate the development of leukemia emerging from an ELANE-CN background. The pre-leukemia predisposition of the cells carrying leukemogenic mutations can be efficiently observed and monitored in-vivo over the long-term. Furthermore, our model provides a robust system to investigate the role of additional leukemia-associated mutations and evaluate therapeutic strategies to eliminate early oncogenic cells not fully leukemic but dangerously close to full leukemic transformation.