Screens, iron, and leukemia

A typical modern strategy for therapeutic drug development in leukemia is identification of a molecular target indispensable for maintenance of the leukemia state by academic biologists followed by an intense industrial-based search for a novel (nonpatented) specific inhibitor. Such an approach is time consuming and expensive. Eberhard et al have “gone back to the future” by screening a large number of publicly available compounds for their ability to inhibit a cell-based target presumed vital to leukemic cell physiology.¹  This is analogous to the “old” National Cancer Institute screen of natural compounds against a variety of tumor cell lines that lead to the discovery of many important antineoplastic agents.²  The twist is that these investigators used off-patent chemicals and natural products (4800) against HeLa cells that overexpressed the survivin promoter driving luciferase, using an easily detectable readout. Determination of the target or mechanism of action of a drug positive in this screen would come later.¹  Ciclopirox (CPX), a compound formerly explored as an antifungal agent but never developed, was identified via this screen and found capable of specific leukemic cell killing.

The choice of the specific screening strategy is very important in this type of drug discovery. The investigators hoped that inhibiting transactivation of the survivin promoter would correlate with anticancer activity. Whereas such an assumption is reasonable, there was no guarantee the screen would be robust enough to not miss many positives and specific enough to exclude generally toxic compounds. In fact, of the 45 “best” compounds, 44 had a poor ratio of survivin repression to nonspecific killing. Another assumption was that a “positive” compound that worked in their model systems (lineage-deleted human core blood cells transduced with a TLS-ERG or MLL-ENL oncogene) would correlate with useful clinical activity. Although CPX did inhibit in this model, the compound also suppressed growth of normal hematopoietic cells at only slightly higher concentrations.

The investigators suggested that the mechanism of action of CPX was the result of the depletion of intracellular iron in a fashion distinct from that of the extracellular iron chelating agent deferoxamine. One aspect of CPX-mediated cytotoxicity was inhibition of ribonucleotide reductase, an iron-dependent enzyme. CPX was significantly more potent that hydroxyurea, a commonly used ribonucleotide reductase inhibitor that acts through a different mechanism. CPX was also synergistic with cytarabine in inducing leukemic cell death. So intracellular iron chelation, and thus CPX, could represent a novel approach to inhibit this enzyme. Another group has successfully used an antisense oligonucleotide approach to the R2 subunit of this enzyme.³  Whether ribonucleotide reductase inhibitors like CPX will be a viable antileukemic approach alone or in combination with other cytotoxic agents remains to be seen.

The Eberhard paper has important implications beyond the derivation of a potential interesting novel antileukemic agent. First, it suggests that screening large libraries of available compounds could yield potentially clinically effective agents at low developmental cost. Second, the notion that intracellular iron depletion might have a therapeutic role in acute myeloid leukemia (AML) is a relatively novel idea, possibly suggesting a pathway for rational drug development. Furthermore, the investigators chose a potentially relevant screen, survivin inhibition. Survivin is preferentially expressed in malignant cells compared with normal cells, caused by transactivation of the promoter rather than gene mutation. Perhaps a more elegant screening strategy uses expression-based screening to identify compounds that engender a desired pattern of gene expression.⁴  This approach has the advantage of looking at many genes simultaneously. Such screens have identified drugs that might be useful in the treatment of AML. For example, Stegmaier and colleagues have shown that dexamethasone, gefitinib, and dihydro-folate reductase inhibitors might be effective agents.^5,6 

Before prescribing CPX for patients with leukemia, a great deal of work has to be done. There are concerns that the therapeutic index might be too narrow because the drug also appears to suppress normal stem cells at slightly higher doses. Second, the drug could turn out to be significantly toxic in humans despite being able to achieve therapeutic concentrations that are tolerable in rodents and canines. Nonetheless, the general approach of Eberhard et al is to be applauded. Despite the many ongoing clinical trials with a variety of drugs in AML (eg, nucleoside analogs, alkylating agents, and inhibitors of FLT3, c-Kit, farnesyl transferase, MDR, cyclin-dependent kinase, DNA methyl transferases, histone deacetylases), we are in dire need of novel agents. This is especially true as we still commonly use only 2 drugs, anthracyclines and cytarabine, for the management of patients with this difficult disease.