Molecular Signature for Parthenolide-Induced Apoptosis in AML CD34+ Cells Reveals Targets for Improving Induction of Leukemia-Specific Cell Death.

Previous studies have demonstrated that the plant-derived compound parthenolide (PTL) induces AML stem cell specific apoptosis while sparing normal counterparts. We have observed that PTL strongly inhibits NF-kappaB activity. Moreover, anti-oxidants such as N-acetylcysteine block the apoptotic process. However, the molecular events involved in the process are not completely understood. Given the demonstrated efficacy of PTL at targeting AML stem cells (AML-SCs), we hypothesized that a molecular signature that captures events invoked during PTL-induced apoptosis will provide a powerful tool for improving AML-SC targeting. To this end, we have conducted pharmacogenomic studies using primary CD34-enriched cells obtained from 12 randomly selected AML specimens. To capture dynamics of the PTL response, primary CD34+ AML were cultured in vitro and exposed to 5 micromolar PTL for either 1 hour or 6 hours. Labeled mRNA was hybridized to Affymetrix HG-U133plus2 chips. Differential expression of individual genes was assessed using moderated and permutation-based paired tests with false discovery rate control. Additionally, to capture pathway-level events possibly representing drugable targets, we employed functional, network, and robust geneset-based approaches to mine the data for coordinated changes in the expression of groups of biologically related genes, e.g., those controlled by a common transcription factor. By 1 hour post-treatment, we were able to detect the beginning stages of a cascade of responses leading to activation of antioxidant enzymes, cytoskeletal reorganization and downregulation of adhesion molecules, modulation of protein synthesis, proteasome remodeling, and modulation of cell cycle progression genes (q < 0.05). Transcription factor (TF) binding motif enrichment studies show that several of these processes are governed by the involvement of transcription factors such as Nrf2, NFkappaB, and interferon regulatory factor-1 (IRF-1) and their respective upstream controllers. The biological relevance of the array findings was directly validated in independent studies demonstrating heightened oxidative state and the associated nuclear translocation of Nrf2 and upregulation of its target heme oxygenase I. We further hypothesized that a number of genes activated immediately following PTL exposure are important for promoting AML survival and/or drug resistance and that such genes would represent potentially useful drug targets. This rationale resulted in our selection of the target, eukaryotic initiation factor 5 (eIF5). Our array-based analyses revealed that the eIF5 gene is upregulated within 1 h of PTL exposure and is further upregulated at 6 h. eIF5 activation is regulated by hypusination and can be inhibited by the anti-fungal agent, ciclopirox. Therefore, we tested the biological effects of 2.5 micromolar PTL alone or in combination with ciclopirox on independent primary AML specimens. Exposure to either drug alone had a negligible effect on primary normal bone marrow or AML cells after 16 h growth in suspension culture. However, the combination of both resulted in ~20% viability for AML relative to ~80% viability for normal bone marrow. Importantly, the combination of PTL and ciclopirox completely eradicated the colony forming ability of AML cells while only modestly affecting normal bone marrow. These studies support the utility of high-throughput genomic studies as a conduit for drug discovery.