We report for the first time the use of a selective small-molecule inhibitor of DNA repair to potentiate topoisomerase II (topo II) poisons, identifying DNA-dependent protein kinase (DNA-PK) as a potential target for leukemia therapy. Topo II poisons form cleavable complexes that are processed to DNA double-strand breaks (DSBs). DNA-PK mediates nonhomologous end joining (NHEJ). Inhibition of this DSB repair pathway may sensitize cells to topo II poisons. We investigated the effects of a novel DNA-PK inhibitor, NU7026 (2-(morpholin-4-yl)-benzo[h]chomen-4-one), on the response to topo II poisons using K562 leukemia cells. NU7026 (10 μM) potentiated the growth inhibition of idarubicin, daunorubicin, doxorubicin, etoposide, amsacrine (mAMSA), and mitroxantrone with potentiation factors at 50% growth inhibition ranging from approximately 19 for mAMSA to approximately 2 for idarubicin (potentiation of etoposide was confirmed by clonogenic assay). In contrast, NU7026 did not potentiate camptothecin or cytosine arabinoside (araC). NU7026 did not affect the levels of etoposide-induced topo IIα or β cleavable complexes. NU7026 alone had no effect on cell cycle distribution, but etoposide-induced accumulation in G2/M was increased by NU7026. A concentration-dependent increase in etoposide-induced DSB levels was increased by NU7026. The mechanism of NU7026 potentiation of topo II poisons involves inhibition of NHEJ and a G2/M checkpoint arrest. (Blood. 2004;103:4659-4665)

We have shown that both DNA topoisomerase (topo) IIα and β are in vivo targets for etoposide using a new assay which directly measures topo IIα and β cleavable complexes in individual cells after treatment with topo II targeting drugs. CCRF-CEM human leukemic cells were exposed to etoposide for 2 hr, then embedded in agarose on microscope slides before cell lysis. DNA from each cell remained trapped in the agarose and covalently bound topo II molecules from drug-stabilized cleavable complexes remained associated with the DNA. The covalently bound topo II was detected in situ by immunofluorescence. Isoform-specific covalent complexes were detected with antisera specific for either the α or β isoform of topo II followed by a fluorescein isothiocyanate-conjugated second antibody. DNA was detected using the fluorescent stain Hoechst 33258. A cooled slow scan charged coupled device camera was used to capture images. A dose-dependent increase in green immunofluorescence was observed when using antisera to either the α or β isoforms of topo II, indicating that both isoforms are targets for etoposide. We have called this the TARDIS method, for trapped in agarose DNA immunostaining. Two key advantages of the TARDIS method are that it is isoform-specific and that it requires small numbers of cells, making it suitable for analysis of samples from patients being treated with topo II-targeting drugs. The isoform specificity will enable us to extend our understanding of the mechanism of interaction between topo II-targeting agents and their target, the two human isoforms.

Poly(ADP-ribose) polymerase (PARP) inhibitors enhance DNA topoisomerase I (topo I) poison-induced cytotoxicity and antitumor activity in vitro and in vivo, but the mechanism has not been defined. We investigated the role of PARP-1 in the response to topo I poisons using PARP-1^−/− and PARP-1^+/+ mouse embryonic fibroblasts and the potent PARP-1 inhibitor, AG14361 (K_i < 5 nmol/L). PARP-1^−/− mouse embryonic fibroblasts were 3-fold more sensitive to topotecan than PARP-1^+/+ mouse embryonic fibroblasts (GI₅₀, 21 and 65 nmol/L, respectively). AG14361 caused a >3-fold sensitization of PARP-1^+/+ cells to topotecan compared with a <1.4-fold sensitization in PARP-1^−/− cells. In human leukemia K562 cells, AG14361 caused a 2-fold sensitization to camptothecin-induced cytotoxicity. AG14361 did not affect the cellular activity of topo I as determined by measurement of cleavable complexes and topo I relaxation activity, showing that sensitization was not due to topo I activation. In contrast, repair of DNA following camptothecin removal, normally very rapid, was significantly retarded by AG14361, resulting in a 62% inhibition of repair 10 minutes after camptothecin removal. This led to a 20% increase in the net accumulation of camptothecin-induced DNA strand break levels in cells coexposed to AG14361 for 16 hours. We investigated the DNA repair mechanism involved using a panel of DNA repair–deficient Chinese hamster ovary cells. AG14361 significantly potentiated camptothecin-mediated cytotoxicity in all cells, except the base excision repair–deficient EM9 cells. Therefore, the most likely mechanism for the potentiation of topo I poison-mediated cytotoxicity by AG14361 is via PARP-1-dependent base excision repair.

Human cells express two genetically distinct isoforms of DNA topoisomerase II, α and β, which catalyze ATP-dependent DNA strand passage and are an important antitumor drug target. Here we report for the first time the successful overexpression of human topoisomerase IIβ in yeast by cloning a topoisomerase IIβ cDNA in a yeast shuttle vector under the control of a galactose-inducible promoter. Recombinant human topoisomerase IIβ (residues 46-1621 fused to the first 5 residues of yeast topoisomerase II) was purified to homogeneity, yielding an enzymatically active polypeptide in sufficient quantity to allow analysis of its domain structure and comparison with that of recombinant human topoisomerase IIα. Partial digestion of β with either trypsin or protease SV8 generated fragments of approximately 130, 90, 62, and 45-50 kDa, arising from cleavage at three limited and discrete regions of the protein (A, B, and C) indicating the presence of at least four structural domains. Recombinant human topoisomerase IIα and β induced DNA breakage which was promoted by a variety of agents. Isoform differences in drug-induced DNA breakage were observed. These studies of human topoisomerase IIβ in concert with α should aid the determination of their individual roles in cancer chemotherapy and should facilitate the design, targeting, and testing of cytotoxic antitumor agents.

There are two major pathways leading to induction of NF-κB subunits. The classical (or canonical) pathway typically leads to the induction of RelA or c-Rel containing complexes, and involves the degradation of IκBα in a manner dependent on IκB kinase (IKK) β and the IKK regulatory subunit NEMO. The alternative (or non-canonical) pathway, involves the inducible processing of p100 to p52, leading to the induction of NF-κB2(p52)/RelB containing complexes, and is dependent on IKKα and NF-κB inducing kinase (NIK). Here we demonstrate that in primary human fibroblasts, the alternative NF-κB pathway subunits NF-κB2 and RelB have multiple, but distinct, effects on the expression of key regulators of the cell cycle, reactive oxygen species (ROS) generation and protein stability. Specifically, following siRNA knockdown, quantitative PCR, western blot analyses and chromatin immunoprecipitation (ChIP) show that NF-κB2 regulates the expression of CDK4 and CDK6, while RelB, through the regulation of genes such as PSMA5 and ANAPC1, regulates the stability of p21WAF1 and the tumour suppressor p53. These combine to regulate the activity of the retinoblastoma protein, Rb, leading to induction of polycomb protein EZH2 expression. Moreover, our ChIP analysis demonstrates that EZH2 is also a direct NF-κB target gene. Microarray analysis revealed that in fibroblasts, EZH2 antagonizes a subset of p53 target genes previously associated with the senescent cell phenotype, including DEK and RacGAP1. We show that this pathway provides the major route of crosstalk between the alternative NF-κB pathway and p53, a consequence of which is to suppress cell senescence. Importantly, we find that activation of NF-κB also induces EZH2 expression in CD40L stimulated cells from Chronic Lymphocytic Leukemia patients. We therefore propose that this pathway provides a mechanism through which microenvironment induced NF-κB can inhibit tumor suppressor function and promote tumorigenesis. Although the classical NF-κB pathway is frequently associated with the induction of cellular senescence and the senescence associated secretory phenotype (SASP), the role of the alternative NF-κB pathway, which is frequently activated in hematological malignancies as well as some solid tumors, has not been defined. We therefore investigated the role of the alternative NF-κB pathway in this process. Here we report that NF-κB2 and RelB, the effectors of the alternative NF-κB pathway, suppress senescence through inhibition of p53 activity. Using primary human fibroblasts, we demonstrate that this is accomplished through NF-κB2/RelB dependent control of a previously unknown pathway, incorporating regulation of CDK4 and 6 expression as well as regulators of p21WAF1 and p53 protein stability. Loss of NF-κB2/RelB results in suppression of retinoblastoma (Rb) tumour suppressor phosphorylation, which in turn leads to inhibition of EZH2 expression and de-repression of p53 activity. Interestingly, we find that CD40 ligand stimulation of cells from Chronic Lymphocytic Leukemia patients, which strongly induces the alternative NF-κB pathway, also induces EZH2 expression. We propose that the alternative NF-κB pathway can promote tumorigenesis through suppression of p53 dependent senescence, a process that may have relevance to cancer cells retaining wild type p53.