The Valproic Acid-Fludarabine Combination Induces a Synergistic Response in Chronic Lymphocytic Leukemia Via a Mechanism Involving the Lysosomal Protease Cathepsin B.

Abstract 2892

Chronic Lymphocytic Leukemia (CLL) is the most common haematological malignancy in the western world. Fludarabine, a nucleoside analogue, is commonly used to treat Chronic Lymphocytic Leukemia (CLL) in untreated and relapsed CLL. However, patients commonly develop resistance to fludarabine. We hypothesize that the addition of Valproic Acid (VPA), an inhibitor of histone deacetylases (HDACs), can improve fludarabine-based therapy. The VPA-Fludarabine combination induced a synergistic response in human leukemic cells and primary CLL cells. Fludarabine also interacted synergistically with three other HDAC inhibitors, suberoylanilide hydroxamic acid (SAHA), Trichostatin A, and sodium butyrate, while the synergy was not observed with valpromide, the VPA analogue that does not inhibit HDACs. We confirmed that fludarabine treatment activates caspases-8, -9 and caspase-3, and we also show that fludarabine treatment activates caspase-2, an upstream caspase that has been implicated in cell death associated with lysosome membrane permeabilization (LMP). Activation of all four caspases was enhanced by the addition of VPA. Enhanced activation of caspases was associated with down-regulation of two prominent anti-apoptotic proteins, Mcl-1 and XIAP. The down-regulation of Mcl-1 and XIAP was dependent on the lysosomes, as their alkalinization using either chloroquine or NH₄Cl partially stabilized both proteins, leading to reduced apoptosis. Chemical inhibition of a specific lysosomal protease, cathepsin B, using CA074-Me, was sufficient to stabilize Mcl-1 and XIAP, reduce caspase activation and apoptosis. Treatment with fludarabine or the VPA-fludarabine combination led to the loss of lysosome integrity, as visualized by fluorescent staining, thus suggesting a leakage of the lysosomal content into the cytosol in response to the drugs. Addition of purified cathepsin B to leukemic cell lysates led to the reduction in protein levels of Mcl-1, XIAP and pro-caspase-2, thus suggesting that the re-localization of cathepsin B into the cytosol is sufficient to drive cell death. VPA treatment enhanced cathepsin B levels in both leukemic cell lines and primary CLL cells. When cathepsin B activity was examined using zRR-AMC, a fluorogenic substrate of cathepsin B, VPA also increased cathepsin B activity, and this activity was abolished by the addition of CA074-Me. In parallel with the in vitro/ex vivo experiments, we had launched a phase II clinical trial at CancerCare Manitoba. Six relapsed CLL patients who had received at least one prior therapy with fludarabine were examined. No responses were seen after 28 days using VPA alone, in line with the in vitro observation of minimal cytotoxicity of VPA at low doses. However, in five patients who continued on VPA with fludarabine, three patients showed a >50% fall in lymphocyte/lymph node size after receiving five cycles of the combination. When the leukemic cells from VPA-treated CLL patients were examined, VPA administration induced increased levels of histone-3 acetylation and cathepsin B in vivo.

In summary, a novel mechanism for fludarabine cytotoxicity has been elucidated, where fludarabine induces a loss of lysosomal integrity, leading to cathepsin B-dependent cell death. VPA interacted with fludarabine synergistically, and this synergy was associated with the VPA-induced increase in VPA level and activity. VPA induced increase in histone-3 acetylation and cathepsin B in vivo, and this induction of cathepsin B is likely to be contributing to the clinical response observed in fludarabine-relapsed/refractory CLL patients.