Abstract 2070

Poster Board II-47

Despite recent progress, the majority of AML patients are refractory or relapse following standard cytarabine/anthracycline-based treatments. Studies have demonstrated that leukemia stem cells (LSCs) are able to initiate and perpetuate AML and are reported to be resistant to standard chemotherapeutics. Therefore, it has been hypothesized that these cells contribute to determination of the poor outcome of AML patients. The discovery of novel agents that directly target LSCs may potentially improve these disease outcomes. Previous studies have demonstrated the utility of screening public microarray data for new gene expression patterns mimicking the gene expression patterns of known anti-LSC compounds. Indeed, in screening the NCBI Gene Expression Omnibus (GEO), we found that compounds mimicking the gene expression pattern of a known anti-LSC compound, parthenolide (PTL), could impair the leukemogenic activity of phenotypically- and functionally-defined LSCs. Using this approach, we have reported two novel PTL-like compounds: celastrol and 4-hydroxy-2-nonenal. However, as investigators continue to deposit new data into GEO, we have re-interrogated the GEO database to identify additional compounds active against LSCs . One of our most recent “hits” was the HDACi, AR-42. This compound represents a novel class of HDACis that are structurally similar to phenylbutyrate, but with improved HDACi activity within sub-micromolar concentrations. To validate our screening results, we then tested the efficacy of AR-42 in AML cell lines and primary human AML samples evaluating bulk, progenitor/stem cell populations. Indeed, as predicted by the gene expression-based screen, exposure to 1 microM AR-42 produced greater than 50% cell death at 24 hours post-treatment in phenotypically defined AML stem/progenitor cells. Next, we tested the ability of AR-42 to impair the growth of AML LSC-derived colonies in methylcellulose media. 1 microM AR-42 significantly impaired the formation of AML colonies compared to relative to untreated controls (5% colony formation in AR-42-treated vs untreated; p < 0.05). Importantly, impairment of colony formation of normal myeloid and erythroid progenitors was markedly less pronounced than that of leukemic progenitors, suggesting a preferential activity of this compound on leukemogenic cells. Surprisingly, we observed that AR-42 mechanism of induction of cell death differs from other anti-LSC compounds hitherto identified (parthenolide, celastrol, 4-hydroxy-2-nonenal), which involved NFkappaB inhibition coupled to induction of oxidative stress. Interesting, we found that AR-42 failed to induce strong NFkappaB inhibition and, unlike parthenolide, the AR-42-dependent leukemia-specific death mechanism involved activation of the caspase-3, caspase-8 and PARP cleavage Altogether, these findings indicate the utility of gene expression-based screens to identify novel compounds that are active on LSC, including AR-42 which is being tested in patients with AML and other hematologic malignancies.


No relevant conflicts of interest to declare.

Author notes


Asterisk with author names denotes non-ASH members.