Expression Profiles of VDR-Induced Differentiation of atRA Refractory AML Cell Lines.

Whatever the success of all-trans retinoic acid (atRA) therapy in acute promyelocytic leukemia (FAB AML3), other AML are poorly responsive to atRA and don’t benefit of any differentiation therapy. Moreover, primarily responsive patients often acquire resistance. Among molecules that control normal myelopoiesis, 1,25(OH)₂-vitaminD₃ (VD3), through its receptor VDR, allows maturation of myelomonocytic precursors. If VD3 has low effects on AML cells differentiation, its activity can be enhanced by various molecules. So, we tested several differentiation activators in combination with the non calcemiant VDR agonist EB1089 (EB) (Leo Pharmaceutics) on two atRA-refractory cell lines. Thus, U937 cells reproduce myelomonocytic acute leukemias that are strongly sensitive to VDR-induced differentiation and with a poor atRA sensitivity. Parallel, NB4-derived LR2 model mimics promyelocytic leukemia with acquired atRA resistance and has lower response to VD3.

First, we confirm VDR-sensitivity of those cell lines. Moreover, as expected, U937 cells respond to EB1089 at lower dose (1nM) than NB4-LR2 (10nM) as showed by CD11b myeloid marker level.

Among tested activators, the cytokine TGFb induces the most complete differentiation for both cell lines. After a 72 hour-treatment, only 5% of cells are still clustered in S phase. In addition, all are positive for the monocyte marker CD14 which expression reaches to 10 and 6 time fold the base level for U937 and NB4-LR2 respectively. At our knowledge, it is the first model of complete differentiation of NB4-LR2. Even if this agent can’t be used in therapy, it provides a reference for comparing other potential differentiating molecules. In fact, Histones Deacetylases Inhibitors (HDI) and arsenate (ATO) both benefit to VDR-induced differentiation of atRA refractory cell lines.

Specifically, ATO benefits to EB-induced differentiation, as both U937 and NB4-LR2 cells acquire monocytic phenotype (CD14 expression) even though growth arrest is weaker. Indeed there are still 10–15% of S-phase clustered cells.

Concerning the association of HDI valproic acid (VPA) and trichostatine A (TSA) with EB, growth arrest is strongest for both cell lines (5–10% S phase clustered cells) at 1mM and 100nM respectively. For phenotype markers, responses vary: U937 cells only respond to TSA combination whereas NB4-LR2 cells differentiate with TSA or VPA. Moreover, only TSA-EB treated cells become positive for CD14 even if expression is low. The addition of LGD1069, agonist of the Retinoid X receptor, a obligatory partner of VDR, improves HDI effects in all measured parameters.

In case of NB4-LR2 cell line, flow cytometry data were reinforced by expression profile analyses. A hundred messengers were selected from bibliography and SAGE libraries of the promyelocytic cell line NB4 (unpublished data) and profiles were established by semi-quantitative real time PCR on low density array. Manual clusterization after comparison to granulocytic and monocytic differentiation (atRA-treated NB4 and VDTGFb-treated NB4-LR2 cells respectively) allows selection of about 10 messengers sufficient to predict patients’ response and to specify differentiation. In addition, the clinical use of VDR agonist, HDI and ATO is conceivable because of low side effects.