Modifying risks of secondary leukemias: is drug scheduling important?

Abstract

Secondary cancers are a formidable consequence of our present efforts to cure cancer. With an increasing armamentarium of genotoxic drugs that damage DNA in diverse ways, the use of these agents in combination, and the resultant lengthening of survival among many patients with neoplastic diseases, the secondary leukemias are a growing complication in the setting of therapy for numerous cancers. To date, the agents that have been associated with the induction of secondary leukemias are those that damage DNA directly, such as alkylating agents, and those that damage DNA through interaction with other molecules, such as the nuclear enzyme topoisomerase II (topo-II), which affects DNA conformation and function (7). Best known among the latter category of agents are the topo-IIdirected epipodophyllotoxins that damage DNA by binding to topo-II, stabilizing the DNA-topo-II covalent linkage, and impeding religation of the broken DNA strands (2,3). Exposure of cells to topo-II-active agents increases the frequency of illegitimate recombinational events, a physiologic activity that may be related to both the cytotoxicity and the leukemogenicity of the topo-II inhibitors (3,4). The leukemogenic translations induced by the topo-II-active agents often involve the MLL (mixed lineage or myeloid lymphoid leukemia) gene located at human chromosome band Ilq23 (5-8). More than 20 different translocations involving 1 Iq23 have been identified, indicating that MLL has numerous translocation partners with which it can genetically recombine to produce leukemogenic fusion proteins. Molecular dissection of the breakpoints on 1 Iq23 is providing insight into the molecular pathogenesis of topo-II inhibitor-associated leukemias (1,6,9-12). Most of the MLL breakpoints occur in a 9-kilobase region of the genomic DNA, the MLL breakpoint cluster region (bcr), which spans exons 5 to 11 and contains a number of interesting DNA sequences that could be involved in illegitimate recombination; these sequences include Alu sequences, topo-IIbinding sequences, scaffold-associated regions (SARs), V(D)J (variable-diversity-joining) recombinase recognition sites, and sequences with homology to the chi recombinatorial element in Escherichia coli (10,13,14). Comparison of the specific MLL breakpoint sites in secondary and de novo leukemias suggests that the pathogenesis of secondary leukemia involving MLL may be distinctive in certain cases from that of the de novo leukemias associated with MLL abnormalities (9,10,12). Topo-II-directed agents can induce site-specific DNA cleavage at the consensus topo-II-binding site within the MLL bcr in normal peripheral blood lymphocytes and in malignant cell lines (predominantly those originating from Tand B-cell leukemias) (9). This action is consistent with the clinical finding that the MLL breakpoints in the majority of secondary acute myelogenous leukemia (AML) cases linked to topo-II-directed agents occur in the telomeric (5') half of the bcr, mainly in the high-affinity SARs and pinpointed to the consensus topo-II-binding sites that are located within the SARs (10,12). In contrast, MLL breakpoints in de novo leukemias occur in the centromeric half of the bcr and may involve Alu repeat sequences (12). Still, while the specific sites of genomic damage within MLL may differ according to the cause of the leukemia, the net result is similar, i.e., MLL gene breakage, illegitimate recombination, and subsequent expression of a leukemogenic fusion protein. The HPRT (hypoxanthine phosphoribosyltransferase) gene mutant frequency assay can detect somatic cell genetic damage by diverse mutagenic agents, including cytotoxic chemotherapeutic agents (15,16), and can identify the molecular signature resulting from a specific mutagenic challenge (17-20). Etoposide is capable of inducing site-specific deletions of HPRT exons 2 + 3 (21), an event that appears to be mediated by inappropriate V(D)J recombinase activity (20-22). In this issue of the Journal, Chen et al. (22) exploit this phenomenon and use the HPRT exon 2 + 3 deletion as a surrogate marker of etoposide-induced aberrant recombinase activity to examine the differentia] effects of etoposide on cytotoxicity versus potential leukemogenicity as related to duration of drug exposure. Using the human T-cell leukemia cell line CCRF-CEM, they demonstrate that illegitimate recombinational events, manifested by deletion of exons 2 + 3 in the HPRT gene, occur with greater frequency following brief (4-hour) than following prolonged (24-hour) exposure. Importantly, the ratio of the percentage of