Bmi-1 Expression Is Useful To Design Therapy and To Predict Prognosis in Patients with Acute Myeloid Leukemia.

Since prognosis of patients with acute myeloid leukemia (AML) is highly variable even in a single subpopulation in FAB classification, it would be useful to find prognostic molecular markers for AML. Thus, we investigated Bmi-1 expression in AML cells by flow cytometry and analyzed whether it predicts prognosis in AML patients and further it is helpful to choose therapies in the modalities of treatment options, because it is known to be required for self-renewal mechanism of leukemic stem cells. Bmi-1 expression in bone marrow or peripheral blood cells was analyzed in 49 patients with AML (M0(n=5), M1(n=7), M2(n=6), M3(n=5), M4(n=8), M5(n=5), M6(n=1)), granulocytic sarcoma(n=1), MDS-AML (n=9), and secondary AML(n=2). Freshly isolated AML cells were stained with a PE-conjugated anti-CD34-antibody followed by fixation and then with anti-Bmi-1-antibody-FITC. All of patients with low Bmi-1 positivity (<35%, n=11) except for de novo AML(M0) entered in complete remission (CR) with single induction chemotherapy(n=5) and accordingly had better overall survival, even though lower dose of chemotherapy (60% of standard dose) was given (n=3). Alternatively, patients with higher percentage of Bmi-1 expression in AML cells (> 70%, n=19) except for AML(M3) progressed to death within two years, unless they were treated with highly intensive therapy such as high dose AraC or allogeneic stem cell transplantation (n=3). Patients with intermediate degree of Bmi-1 expression (35–60%, n=5) responded to standard intensity of chemotherapy (n=2) and are alive for more than two years. Interestingly, patients with MDS-AML (n=9) had high Bmi-1 expression (79%) and all of them have died within 20 months. Binary logistic regression model showed that significant correlation was found among survival status as dependent variable, Bmi-1, and treatment intensity as independent variable (p = 0.004). On the other hand, Univariate analysis did not reveal any relation of Bmi-1 expression to karyotype, age, WBC count, or FAB subtype. In conclusion, Bmi-1 expression could be an independent prognostic marker and useful tool to design therapy for the AML patients.