In Vivo Analysis of the Role of C/EBPα in Acute Promyelocytic Leukemia Genesis.

Abstract Acute promyelocytic leukemia (APL) is characterized by the infiltration of bone marrow (BM) and peripheral blood (PB) by leukemic cells presenting a block of differentiation at the stage of promyelocytes. At the cytogenetic level, APL is associated with the t(15;17) which causes the fusion of two genes: Retinoic Receptor α (RARα) and Promyelocytic Leukemia (PML), on chromosomes 17 and 15, respectively. C/EBPα is a leucine zipper transcription factor essential to normal granulopoiesis. Mutations in C/EBPα gene are detected in 6–10% of Acute Myeloid Leukemia (AML) cases and C/EBPα activity is down-regulated by AML1-ETO fusion protein associated with FAB M2 subtype. We decided to test whether PML-RARα interferes with C/EBPα function, thus contributing to the blockage of differentiation characteristic of APL. We generated mutant mice expressing PML-RARα and haploinsufficient for C/EBPα (PR C/EBPα+/−) by crossing hCG PML-RARα transgenic mice (PR TM) and C/EBPα+/−. Leukemia was not detected in WT (n=415) and C/EBPα+/− (n=47) mice after 800 days of follow up. In contrast, 8.2% PR TM (19/233) and 14.9% (13/87) PR C/EBPα+/− mice developed a form of leukemia that closely resembled human APL and identical to that developed by PR TM. The leukemia-free survival was significantly shorter in PR C/EBPα+/− compared to PR TM (644.3 days; 95% Confidence Interval, 95%C.I.: 586.4 – 702.2 vs 718.4 days; 95%C.I.: 689.3 – 747.5, P=0.02). Both groups presented a long latency for the development of the disease with a mean age (95%C.I.) at diagnosis of 399.9 days (184–673) and 495.8 days (215–757) in PR and PR C/EBPα+/−, respectively. PR and PR C/EBPα+/− leukemic mice presented similar WBC counts (107.1 ± 82.65x103vs 63.55 ± 57.82x103cells/ml, P=0.26), hemoglobin concentrations (10.87 ± 3.69 vs 9.92 ± 2.39 g/dl, P =0.57) and platelet counts (283.8 ± 188.7x103vs 177.8 ± 149.5x103platelets/ml, P =0.24). In both groups, the leukemic cells resembled promyelocytes, and expressed the phenotype CD11b+ Gr1+ CD34± c-Kit+, which represented 46.65 ± 26.89%; 32.72 ± 15.16% and 1.91 ± 1.42% of the spleen cells from PR; PR C/EBPα+/− and WT, respectively. In order to isolate PML-RARα leukemic cells and their normal counterparts to gene expression assessment, BM samples from WT mice (pooled from 2 mice), non leukemic PR C/EBPα+/− mice (pooled from 2 mice), and spleen cells from WT and leukemic PR C/EBPα+/− mice were first submitted to red cells lyses, stained with previously conjugated antibodies. Cells expressing CD16/32, CD11b, c-kit and CD34, but neither CD3 nor CD45/B220 were isolated. These cells presented medium to large size with a granular cytoplasm. C/EBPα and PML-RARα expression was analyzed by qRT-PCR in the sorted cells. Compared to WT cells, promyelocytes from leukemic and non leukemic PR C/EBPα+/− mice expressed significantly less C/EBPα, with the lowest levels detected in leukemic samples. As expected, PML-RARα was not detected in WT samples. Comparison between cells suspensions containing similar numbers of promyelocytes revealed that PML-RARα expression was higher in leukemic compared to non leukemic PR C/EBPα+/− mice. In conclusion, our data strongly suggest that PML-RARα fusion protein acts as a dominant negative product on C/EBPα gene expression in a specific subset of early myeloid cells and contribute to the pathogenesis of APL.