Abstract
Hydroxamic acid analogue pan-histone deacetylase (HDAC) inhibitors (HA-HDIs), e.g., vorinostat, LAQ824 and LBH589, induce in vitro growth arrest, differentiation and apoptosis of human acute leukemia cells. Continuous and protracted use of HA-HDI, as currently used in the clinic against hematologic malignancies is likely to result in the emergence of HA-HDI resistance in leukemia cells. By continuous in vitro exposure of the AML HL-60 cells to the cinnamic acid analogue HA-HDI LAQ824, we have generated an in vitro and in vivo model of HA-HDI-resistant HL-60/LR cells, which are capable of growth in high concentrations (200 nM) of LAQ824. HL-60/LR versus the parental HL-60 cells have a shorter doubling time (12 versus 24 hours), increased % of cells in the S phase of the cell cycle (62.4 versus 40.0) and exhibit shorter interval to generation of leukemia and survival in NOD/SCID mice. As compared to HL-60, HL-60/LR cells have a resistance index of 100 for LAQ824, and are cross-resistant to other antileukemia agents exhibiting resistance index for LBH589: 50; trichostatin A: 15; vorinostat: 30; sodium butyrate: 10; etoposide: 5.0; Ara-C: 3.3 and TRAIL: 31.3. As compared to HL-60, HL-60/LR cells express higher levels of Bcl-xL and XIAP but lower levels of MCL-1. HL-60/LR versus HL-60 cells also express markedly reduced levels of Bim and Bak but higher levels of Bax. Although expressing higher levels of the death receptors (DR) 4 and 5 and lower levels of c-FLIP, HL-60/LR cells lack expression of caspase-8 and show barely detectable levels of FADD. Additionally, HL-60/LR versus HL-60 cells have markedly higher levels of AKT, c-RAF, and p-STAT5. Although expressing higher levels of HDAC1, HDAC2, and HDAC4, HL-60/LR cells lack detectable expression of HDAC6, with increased expression of hyper-acetylated hsp90 and α-tubulin- two of the substrates deacetylated by HDAC6. As compared to hsp90 in HL-60 cells, hyper-acetylated hsp90 in HL-60/LR cells exhibits less binding to ATP and p23. Utilizing a polyclonal antibody generated against acetylated hsp90α, confocal immunofluorescence microscopy showed higher and mostly cell surface expression of acetylated hsp90α in HL-60/LR versus HL-60 cells. As compared to HL-60, treatment of HL-60/LR cells with LAQ824 failed to induce p21 and hsp70, or increase the levels of hyper-acetylated hsp90 and α-tubulin. Notably, although cross-resistant to several anti-leukemia drugs, HL-60/LR cells are collaterally sensitive to the hsp90-inhibiting geldanamycin analogues 17-allylamino-demothoxy geldanamycin (17-AAG) and 17-DMAG with a four and five-fold increased sensitivity to 17-AAG and 17-DMAG, respectively. This was associated with a lack of both a 17-AAG mediated induction of hsp70 and a lesser decline in the levels of AKT and c-RAF in HL-60/LR versus HL-60 cells. Taken together, these findings elucidate several notable in vitro and in vivo biologic characteristics and drug-sensitivity profile of the first fully-characterized HA-HDI-resistant human AML cells. Our findings clearly demonstrate that in vitro resistance to HA-HDIs is associated with loss of HDAC6 expression, hyperacetylation of hsp90, aggressive leukemia phenotype, but cross-sensitivity to 17-AAG. These findings also suggest that hsp90 inhibitors should be tested for overriding de novo or acquired HA-HDI resistance in AML.
Binding of inhibitors to active-site mutants of CD1, the enigmatic catalytic domain of histone deacetylase 6
