Analysis of the Anti-Leukemia Mechanism of Parthenolide.

Abstract 2734

Poster Board II-710

We have previously demonstrated that parthenolide (PTL), a naturally occurring small molecule found in feverfew Chrysanthemum parthenium, induces apoptosis in primary acute myeloid leukemia (AML) cells, including the stem and progenitor cell compartment. Based on these preclinical findings, a PTL derivative (dimethylamino parthenolide) is currently being evaluated in a phase I clinical trial. However, despite the promising activity of PTL, its underlying mechanism of action remains poorly understood. Thus, we have undertaken biochemical studies to better characterize how PTL mediates leukemia-specific cell death. Chemically, the key structural feature of PTL is its alpha-metheylene-gamma-lactone moiety, which via Michael reaction is predicted to mediate potent free thiol scavenging, an activity readily observed in PTL-treated cells. Reported consequences of PTL chemical reactivity in a variety of cell types, represent a broad range of activities that include inhibition of NFkB, activation of p53, ubiquitination of MDM2, inhibition of DNMT1 and inhibition of HDAC1. To better define the specific activities responsible for parthenolide-mediated leukemia cell death, we have employed two general approaches. First, we generated a biotinylated analog of parthenolide (PTL-biotin), which was shown to retain the anti-leukemia activity of the parent compound. PTL-biotin was then used in biochemical pull-down assays to purify parthenolide target proteins, followed by liquid phase chromatography-mass spectrometry (LC-MS) for protein identification. To further verify molecular interactions, native non-biotinylated parthenolide was used to compete binding between candidate targets and PTL-biotin. These studies identified HSP70 as a direct target of PTL. Notably, cysteine-17 of HSP70 is exposed to the ADP/ATP binding site crevice and a molecular docking study indicates that the covalent attachment of PTL to this residue should disrupt the ATP hydrolysis function of the protein. These findings imply that inhibition of HSP70 may contribute to the cell death mechanism underlying PTL anti-leukemia based activity. As a second approach to characterizing PTL, we have performed comparative studies using the closely related compound costunolide (CSN). Since previous structure-activity studies with PTL analogs revealed that opening of the epoxide ring at C₄-C₅ of the molecule completely destroys the anti-tumor activity, we sought to utilize a compound lacking this feature. CSN lacks the epoxide group, but is otherwise identical to PTL, and retains the key alpha-metheylene-gamma-lactone moiety. Interestingly, at concentrations where PTL is highly cytotoxic, CSN does not induce leukemia-specific cell death (less than 10% death for primary AML cells at 7.5 microM). Analysis of CSN activity demonstrated that despite the lack of AML cell death, CSN still induced loss of free thiols and increased reactive oxygen species in a fashion comparable to PTL (as measured by mBBR and CM-H₂DCFDA based flow cytometry). However, CSN is markedly less effective as an inhibitor of NFkB activity (measured by phosphorylation level of NFkB p65). Taken together these findings indicate that oxidative stress alone is not sufficient for PTL-mediated cell death, and further extend previous molecular genetic data demonstrating that NFkB inhibition is an important component of the overall cell death mechanism. The data also show that the alpha-metheylene-gamma-lactone moiety alone is not sufficient to mediate all aspects of PTL activity, and that at least some activity/specificity is created by juxtaposition of the epoxide group. Based on these studies, as well as previous data, we propose that inhibition of NFkB and HSP70 are components of the parthenolide-mediated cell death mechanism, and that oxidative stress is a necessary but not sufficient aspect of its leukemia-specific activity.