Deregulated Activity of AML1/RUNX1 Cooperates with BCR-ABL to Immortalize Hematopoietic Progenitor Cells and Induces Blast Crisis-Like Disease of Chronic Myelogenous Leukemia in Mice,

Abstract 3749

Since the introduction of imatinib treatment for chronic myeloid leukemia (CML), most patients in the chronic phase (CP) achieve long-lasting hematologic remission. However, some patients acquire resistance or intolerance to imatinib, which results in disease progression to blast crisis (BC). Although it is known that BCR-ABL induces genomic instability, which results in leukemia cells creating multiple gene defects, genes inducing the progression to BC in CML are largely unknown. Moreover, mutations in the AML1/RUNX1 gene and overrepresentation of AML1/RUNX1 in BC compared to CP have been revealed. We therefore hypothesized that deregulation of AML1/RUNX1 might be one of the mechanisms for disease progression of CML.

We initially sought to establish a culture system in vitro that allows rapid identification of genes possibly contributing to the emergence of BC. Given that cells leading to the origin of BC are known to resemble committed progenitor rather than stem cells, we employed Sca-1(-)Lin(-) progenitors derived from mouse fetal liver for the culture. Thus, cells were retrovirally transduced with BCR-ABL alone or in combination with a gene of interest, and cultured in vitro without cytokines. BCR-ABL alone-transduced cells grew initially, but ceased to do so after a week. In contrast, cells co-transduced with BCR-ABL and either one of NUP98-HOXA9, Hes1 or Bmi1 (genes known to induce BC-like disease in concert with BCR-ABL) kept growing and were immortalized in the cytokine-free culture. Importantly, cells cultivated in vitro caused fetal disease when transplanted into mice. These findings suggest the utility of our in vitro culture system as an assay of the co-operativity of a given gene with BCR-ABL to induce BC.

We therefore applied the culture system for the assay of AML1/RUNX1 mutants. We chose four AML1/RUNX1 mutants (K83Q, R139G, R80C, and D171N) for the analysis of their co-operativity with BCR-ABL. Wild-type AML1/RUNX1 and a vector-only control were also included. Results showed that K83Q and R139G mutants co-operated with BCR-ABL to immortalize progenitor cells in cytokine-free culture in vitro. In contrast, R80C and D171N mutants lacked such ability. The cells co-expressing BCR-ABL and either K83Q or R139G mutants cultivated in vitro caused fetal disease in mice following transplantation: all 5 mice transplanted with BCR-ABL/K83Q-transduced cells and 7 of 10 mice transplanted with BCR-ABL/R139G-transduced cells developed lethal hematologic diseases 7–79 days following transplantation. These findings are suggestive of variance among mutants of AML1/RUNX1 regarding the ability to induce BC.

Unexpectedly, further analyses revealed that wild-type AML1/RUNX1 was as potent as K83Q and R139G in inducing BC: all 4 mice transplanted with BCR-ABL/wild-type AML1-transduced cells developed fetal disease. Examination of various lengths of C-terminally truncated versions of AML1/RUNX1 for their co-operativity with BCR-ABL revealed that AML1L419X and AML1H377X were able to co-operate with BCR-ABL to immortalize progenitor cells, while AML1I337X and AML1A297X did not. These findings suggest that transcriptional regulatory domains residing at the carboxyl-terminal of AML1/RUNX1 play important roles for wild-type AML1/RUNX1 to induce BC-like disease.

Perhaps consistent with the increased expression of AML1/RUNX1 contributing to the occurrence of BC, shRNA-mediates silencing of AML1/RUNX1 inhibited growth of human CML cell lines (K562, MegA2, and MEG01).

The overall findings of the present study suggest that (1) all mutants of AML1/RUNX1 do not contribute equally to the induction of BC-like disease, but there is a wide range of variance among mutants regarding their ability to induce BC in collaboration with BCR-ABL, and (2) increased expression of wild-type AML1/RUNX1 can induce BC-like disease in collaboration with BCR-ABL in mice. AML1/RUNX1 mutants tested here are known to dominant-negatively inhibit wild-type AML1/RUNX1 functions, but only two of 4 mutants exhibited BC-inducing activity. The over-expression of wild-type AML1/RUNX1 also exhibited such BC-inducing activity. We are currently conducting a detailed investigation of differences in cell lineage affected and growth properties caused by mutant and wild-type AML1/RUNX1s, which may reveal the dichotomic functions of AML1/RUNX1 in the development of BC.