To investigate the potential role and the mechanism of PLZF-RARalpha/RARalpha-PLZF double fusion gene in the pathogenesis of acute promyelocytic leukemia (APL) in vivo at systematic biological level, PLZF-RARalpha/RARalpha-PLZF double transgenic mouse model was established by intercross; the integration and expression of fusion genes were analyzed by PCR and RT-PCR; the disease phenotype was detected by morphological and pathological examination of peripheral blood and bone marrow cells, as well as flow cytometry assays; the effects of ATRA with or without tricostatin A on bone marrow blast cells from PLZF-RARalpha/RARalpha-PLZF double TM were observed. The results showed that leukemia occurred in 5 PLZF-RARalpha/RARalpha-PLZF double TM 7, 7, 9, 11 and 11 months respectively, out of them two (40%) with classic APL features, the others (60%) with chronic myeloid leukemia through an observation period of 18 months. The leukemia occurrence of PLZF-RARalpha/RARalpha-PLZF TM was about 10%, which was similar to PLZF-RARalpha TM as that reported before. The latency was over 6 months, not earlier than PLZF-RARalpha TM only. No morphologic changes of PLZF-RARalpha/RARalpha-PLZF double TM blast cells to ATRA were observed, but increased cytoplasmic-nuclear ratio and nuclear condensation in bone marrow blast cells were found in combination of ATRA with tricostatin A. It is concluded that PLZF-RARalpha/RARalpha-PLZF double fusion gene transgenic mice have heterogeneity of pathogenesis. HDAC inhibitors such as trichostatin A, in combination with ATRA, induce differentiation of the blast/promyelocytic cells from PLZF-RARa/RARa-PLZF double TM, but not ATRA alone.