Evaluating the genotoxicity of topoisomerase-targeted antibiotics

Abstract

Antibiotics like fluoroquinolones (FQs) that target bacterial type II topoisomerases pose a potential genotoxic risk due to interactions with mammalian topoisomerase II (TOPO II) counterparts. Inhibition of TOPO II can lead to the generation of clastogenic DNA double-strand breaks (DSBs) that can in turn manifest in mutagenesis. Thus, methods that allow early identification of drugs that present the greatest hazard are warranted. A rapid, medium-throughput and predictive genotoxicity screen that can be applied to bacterial type II topoisomerase inhibitors is described herein. Maximal induction of the DSB biomarker serine139-phosphorylated histone H2AX (γH2AX) in L5178Y cells was quantified via flow cytometry and correlated with data derived from the mouse lymphoma screen (MLS), a default assay used to rank genotoxic potential. When applied to a class of novel bacterial type II topoisomerase inhibitors (NBTIs) in lead-optimisation, maximal γH2AX induction >1.4-fold (relative to controls) identified 22/27 NBTIs that induced >6-fold relative mutation frequency (MF) in MLS. Moreover, response signatures comprising of γH2AX induction and G(2)M cell cycle arrest elucidated using this approach suggested that these NBTIs, primarily of the H class, operated via a TOPO II poison-like mechanism of action (MoA) similar to FQs. NBTIs that induced ≤6-fold relative MF, which were mainly A class-derived, had less impact on γH2AX (≤1.4-fold) and also evoked G(1) arrest, indicating that their cytotoxic effects were likely mediated through a non-poison MoA. Concordance between assays was 86% (54/63) when 1.4- and 6-fold 'cut offs' were applied. These findings were corroborated through inspection of human TOPO IIα IC(50) data as NBTIs exhibiting equivalent inhibitory capacities had differing genotoxic potencies. Deployed in an early screening capacity, the γH2AX by flow assay coupled with structure-activity relationship evaluation can provide insight into MoA and impact medicinal chemistry efforts, ultimately leading to the production of inherently safer molecules.

Fig. 1

Exemplars of the bacterial type II topoisomerase drug class; NBTI GSK299423 and FQs ciprofloxacin and moxifloxacin.

Fig. 2

Induction of genotoxicity by 63 NBTIs and 4 comparator TOPO II poisons. Compounds that induced relative MF at RTG 10–20% cytotoxicity >6-fold (6-fold threshold indicated by small dashed line) were more likely to induce γH2AX levels >1.4-fold (1.4-fold threshold indicated by large dashed line). In addition, NBTIs that exceeded both genotoxicity cut offs mainly induced G2M arrest, whereas less genotoxic NBTIs generally evoked G1 arrest. Circles represent G2M phase-arrested nuclei; squares, G1 arrest; triangles, S arrest; star, sub-G1 DNA.

Fig. 3

NBTIs (63) binned according to their effects in γH2AX (1.4-fold threshold) and MLS (6-fold threshold). Expression of data in this manner illustrates the high concordance (86%; 54/63) and low incidence (14%; 9/63) of false predictions (five false negatives and four false positives) produced in the γH2AX by flow assay with respect to MLS. The number in the right-hand corner of each pie chart represents the total number of compounds in each bin. The number of NBTIs that induced G1, S, G2M and sub-G1 phase arrest in each bin is represented as both an integer and %.

Fig. 4

SAR with respect to genotoxicity for 63 NBTIs and 4 comparator TOPO II poisons. H class-derived NBTIs were determined to be TOPO II poison-like compounds with a greater propensity to induce genotoxicity, with A and E class-derived NBTIs exhibiting less genotoxicity via a non-TOPO II poison MoA. Circles represent G2M phase-arrested nuclei; squares, G1 arrest; triangles, S arrest; star, sub-G1 DNA.

Fig. 5

Effect of hTOPO IIα inhibition on genotoxicity for 26 NBTIs. Strong inhibition of hTOPO IIα led to increased genotoxicity for some NBTIs, however, in some cases no or only weak genotoxicity was observed, suggesting the existence of at least two MoA. Spheres represent G2M phase-arrested nuclei; cubes, G1 arrest; cone, S arrest; star, sub-G1 DNA. It should be noted that markers are ‘jittered’ for clarity, however, because of this their position within the graph may not be entirely accurate.