Abstract
HDACs epigenetically control gene transcription by modulating acetylation of lysine residues on histones. HDAC6 (a class IIB histone deacetylase) also deacetylates cytosolic proteins, e.g., α tubulin and hsp90, a molecular chaperone involved in folding nascent proteins into active conformation. Inhibition by pan-HDAC inhibitors or depletion by HDAC6 siRNA induces hyperacetylation of hsp70 and hsp90. This inhibits ATP binding and chaperone function of hsp90, also seen following treatment with geldanamycin analogues such as 17-AAG. This results in polyubiquitylation and proteasomal degradation of hsp90 client proteins, including BCR-ABL, FLT-3, AKT and c-Raf, in human leukemia cells. Here, we determined whether a combination of 17-AAG and hsp90 hyperacetylation would exert superior anti-hsp90 and anti-leukemia effects than treatment with either agent alone. Co-treatment of K562 cells with HDAC6 siRNA (for 48 hours) markedly increased 17-AAG (3 to 5 μM for 16 hours) mediated attenuation of the binding of hsp90 to ATP and the co-chaperone p23 but not cdc37, resulting in polyubiquitylation and depletion of hsp90 client proteins. Combined treatment with pan-HDAC inhibitor LBH589 and 17-AAG also induced more inhibition of hsp90 chaperone function and depletion of BCR-ABL (in K562 cells) and FLT-3 (in MV4-11 cells), as well as synergistically induced apoptosis of K562 and MV4-11 cells. Notably, treatment with 17-AAG reduced HDAC6 binding to hsp90, increased hsp90 acetylation and depleted HDAC6 levels. This indicates that HDAC6 has chaperone association with hsp90. HDAC6 also binds to and shuttles misfolded, polyubiquitylated proteins into the aggresome- a perinuclear inclusion of polyubiquitylated proteins formed as the ‘first-line’ protective response to unfolded proteins (UPR). While treatment with bortezomib (100 nM for 16 hours) increased the levels of misfolded proteins and aggresome formation, depletion of HDAC6 by siRNA or 17-AAG markedly inhibited aggresome formation, as demonstrated by immunofluorescent confocal and electron microscopy. Additionally, co-treatment with 17-AAG or HDAC6 siRNA enhanced loss of clonogenic survival of K562 cells. Combined treatment with LBH589 or 17-AAG and bortezomib also inhibited more primary leukemia cell survival than either agent alone. Thus, depletion of aggresome formation accentuates cellular toxicity due to misfolded proteins. Cellular endoplasmic reticulum (ER) stress response to unfolded proteins includes a) activation of the PKR-like ER kinase (PERK), which phosphorylates eIF2 and inhibits protein translation, and b) activation of IRE1 RNase, which splices and activates XBP1 and enhances ER-associated degradation of unfolded proteins. Since IRE1 RNase has chaperone association with hsp90, 17-AAG treatment inhibits IRE1 mediated XBP1 activation. Our studies also demonstrated that treatment with 17-AAG and LBH589 depleted PERK and inhibited eIF2 phosphorylation in K562 cells. Thus, through mechanisms cited above, inhibition of hsp90 (by 17-AAG and LBH589) abrogates the ‘first-line’ protective aggresome formation and ‘second-line’ ER stress response, thereby increasing the cytotoxic effects of bortezomib in K562 cells. By highlighting the biologic consequences of HDAC6 and hsp90 inhibition, these findings support the rationale for therapeutically targeting HDAC6 and hsp90 for enhancing the antileukemia effects of proteasome inhibitors.
Disclosures: Author PA is employed by Novartis Pharmaceutical Company.; Author KB has received research funding from Novartis.; Author KB has received honoraria from Novartis.
Author notes
Corresponding author