Loss of ASXL1 Triggers an Apoptotic Response in Human Hematopoetic Stem and Progenitor Cells

In recent clinical studies, it has been demonstrated that ASXL1 is frequently mutated in myelodysplastic syndrome (MDS), in particular in high-risk MDS patients who have a significant chance to progress to acute myeloid leukemia (AML). Mutation of ASXL1 leads to truncation of the protein and thereby to a loss of its chromatin interacting and modifying domain, possibly facilitating malignant transformation. However, the functions of ASXL1 in human hematopoietic stem and progenitor cells are not well understood. In this study, we addressed whether manipulation of ASXL1-expression in hematopoietic system in vitro mimics the changes observed in MDS-patients.

We down regulated ASXL1 in CD34⁺ cord blood (CB) cells using a lentiviral approach and obtained a 40-50% reduction of ASXL1 expression. Colony forming (CFC) assays revealed that erythroid colony formation was significantly impaired (p=0.01) and, to some extent, granulocytic and macrophage colony formation (p=0.09, p=0.05 respectively). As MDS can affect all hematopoietic lineages, we first stimulated cell differentiation along the myeloid or erythroid lineage in liquid culture. Upon culturing shASXL1 CB CD34⁺ cells in suspension, we observed a modest reduction in expansion (two-fold at week1) under myeloid conditions. In erythroid conditions, shASXL1 CB CD34⁺ cells showed a strong four-fold growth disadvantage, with a more than two-fold delay in erythroid differentiation. The reduced expansion was partly due to a significant increase in apoptosis (5.9% in controls vs. 14.0% shASXL1, p=0.02). The increase of cell death was restricted to differentiating cells, defined as CD71 bright- and CD71/GPA-double positive. This phenotype is similar to what has been observed in patients, where increased cell death of progenitors occurs, and suggests that ASXL1 loss may reflect an MDS-like phenotype in this culture setting. Furthermore, as MDS is considered a hematopoietic stem cell (HSC)-driven disorder, we tested whether HSCs were affected by ASXL1 loss. Long-term culture initiating cell (LTC-IC) assays revealed a two-fold decrease in stem cell frequency. To test dependency of shASXL1 CB 34⁺ cells on the microenvironment, we performed cultures on stromal layers with or without cytokines. shASXL1 CB CD34⁺ cells cultured on MS5 stromal layer showed a modest, two-fold reduction in cell growth at week 4. In the presence of EPO and SCF, we detected a growth disadvantage (three-fold at week 2) and a delay in erythroid differentiation, similar to what was observed in liquid culture. In patients, mutations in ASXL1 are frequently accompanied by a loss of p53. Possibly, loss of p53 is necessary to allow ASXL1-mutant induced transformation thereby bypassing the apoptotic response. Therefore, we modeled simultaneous loss of ASXL1 and TP53 using shRNA lentiviral vectors. Our first data showed that while in primary CFC cultures shASXL1/shTP53 did not give rise to more colonies compared to shASXL1/shSCR cells, an increase in colony-forming activity was observed upon replating of the cells. Furthermore, when using erythroid liquid conditions, a decrease in apoptosis compared to the ASXL1 single mutation could be observed. Nevertheless, no transformation occurred and ASXL1 mutated cells were eventually lost in the double hit model despite reduced apoptosis, suggesting that the p53 axis might not be sufficient as the second hit for full transformation.

In conclusion, our data indicate that mutations in ASXL1 may lead to an increase in cell death and reduced progenitor output in vitro, which may reflect disease development and progression as seen in patients. Unexpectedly, MS5 stromal did not alter the negative phenotype caused by ASXL1 knock down. Therefore, studies are ongoing to investigate whether an already established MDS microenvironment will influence ASXL1 mutation positively. To this end, we are using healthy human mesenchymal stem cells (MSC) and patient derived MDS MSCs.